sign in sign up
<<
Home , Erosion and tectonics

Ellen Wohl IN THE LANDSCAPE Science and Management

Rivers in the Landscape

Rivers in the Landscape Science and Management

Ellen Wohl Department of Geosciences, Colorado State University, Colorado, USA This edition first published 2014 © 2014 by John Wiley & Sons, Ltd Registered office: John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial offices: 9600 Garsington Road, Oxford, OX4 2DQ, UK The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK 111 Street, Hoboken, NJ 07030-5774, USA For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell. The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. Limit of Liability/Disclaimer of Warranty: While the publisher and author(s) have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Library of Congress Cataloging-in-Publication Data Wohl, Ellen E., 1962– Rivers in the landscape / Ellen Wohl. pages cm Includes bibliographical references and index. ISBN 978-1-118-41483-5 (cloth : alk. paper) – ISBN 978-1-118-41489-7 (pbk. : alk. paper) 1. Rivers. 2. River channels. I. Title. GB1201.7.W65 2014 551.48′3–dc23 2013044791 A catalogue record for this book is available from the British Library. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Set in 10.5/13pt Minion by Aptara Inc., New Delhi, India

1 2014 Contents

Acknowledgements ix About the Companion Website xi

Chapter 1 Introduction 1 1.1 Connectivity and inequality 1 1.2 Six degrees of connection 3 1.3 Rivers as integrators 6 1.4 Organization of this volume 7 1.5 Understanding rivers 9 1.5.1 The Colorado Front Range 9 1.6 Only connect 19

Chapter 2 Creating channels and channel networks 21 2.1 Generating water, solutes, and sediment 21 2.1.1 Generating water 21 2.1.2 Generating sediment and solutes 22 2.2 Getting water, solutes, and sediment downslope to channels 23 2.2.1 Downslope pathways of water 23 2.2.2 Downslope movement of sediment 29 2.2.3 Processes and patterns of water chemistry entering channels 32 2.2.4 Influence of the on fluxes into channels 32 2.3 Channel initiation 34 2.4 Extension and development of the drainage network 37 2.4.1 Morphometric indices and scaling laws 37 2.4.2 Optimality 40 2.5 Spatial differentiation within drainage basins 41 2.6 Summary 43

Channel processes I 45

Chapter 3 Water dynamics 47 3.1 Hydraulics 47 3.1.1 Flow classification 48 3.1.2 Energy, flow state, and hydraulic jumps 51 3.1.3 Uniform flow equations and flow resistance 53 3.1.4 Velocity and turbulence 60 3.1.5 Measures of energy exerted against the channel boundaries 65

v vi Contents

3.2 Hydrology 67 3.2.1 Measuring, indirectly estimating, and modeling discharge 67 3.2.2 Flood frequency analysis 71 3.2.3 Hydrographs 73 3.2.4 Other parameters used to characterize discharge 75 3.2.5 Hyporheic exchange and hydrology 77 3.2.6 River hydrology in cold regions 77 3.2.7 Human influences on hydrology 78 3.3 Summary 79

Channel processes II 81

Chapter 4 Fluvial sediment dynamics 83 4.1 The channel bed and initiation of motion 84 4.1.1 Bed sediment characterization 84 4.1.2 Entrainment of non-cohesive sediment 85 4.1.3 of cohesive beds 89 4.2 91 4.2.1 Dissolved load 91 4.2.2 Suspended load 94 4.2.3 Bed load 98 4.3 104 4.3.1 Readily mobile bedforms 105 4.3.2 Infrequently mobile bedforms 108 4.3.3 Bedforms in cohesive sediments 115 4.4 In-channel depositional processes 115 4.5 Bank stability and erosion 117 4.6 Sediment budgets 120 4.7 Summary 124

Chapter 5 Channel forms 125 5.1 Cross-sectional geometry 125 5.1.1 Bankfull, dominant, and effective discharge 125 5.1.2 Width to depth ratio 127 5.1.3 Hydraulic geometry 128 5.1.4 Lane’s balance 130 5.1.5 Complex response 132 5.1.6 Channel evolution models 133 5.2 Channel planform 133 5.2.1 Straight channels 135 5.2.2 Meandering channels 136 5.2.3 Wandering channels 139 5.2.4 Braided channels 139 5.2.5 Anabranching channels 142 Contents vii

5.2.6 Compound channels 143 5.2.7 Karst channels 144 5.2.8 Continuum concept 144 5.2.9 River metamorphosis 146 5.3 Confluences 147 5.4 River gradient 149 5.4.1 Longitudinal profile 151 5.4.2 Stream gradient index 153 5.4.3 154 5.5 Adjustment of channel form 156 5.5.1 Extremal hypotheses of channel adjustment 157 5.5.2 Geomorphic effects of floods 157 5.6 Downstream trends 160 5.6.1 Grain size 160 5.6.2 Instream wood 161 5.7 Summary 163

Chapter 6 Extra-channel environments 165 6.1 165 6.1.1 Depositional processes and stratigraphy 167 6.1.2 Erosional processes and floodplain turnover times 172 6.1.3 Downstream trends in floodplain form and process 174 6.1.4 Classification of floodplains 175 6.2 Terraces 175 6.2.1 Terrace classifications 176 6.2.2 Mechanisms of terrace formation and preservation 176 6.2.3 Terraces as paleoprofiles and paleoenvironmental indicators 179 6.3 Alluvial Fans 181 6.3.1 Erosional and depositional processes 182 6.3.2 Fan geometry and stratigraphy 183 6.4 Deltas 185 6.4.1 Processes of erosion and 186 6.4.2 Delta morphology and stratigraphy 187 6.4.3 Paleoenvironmental records 190 6.4.4 Deltas in the Anthropocene 191 6.5 192 6.6 Summary 194

Chapter 7 Humans and rivers 197 7.1 Indirect impacts 198 7.1.1 Climate change 198 7.1.2 Altered land cover 200 7.2 Direct impacts 205 7.2.1 Flow regulation 205 7.2.2 Altered channel form and connectivity 208 viii Contents

7.3 River management in an environmental context 215 7.3.1 Reference conditions 215 7.3.2 Restoration 217 7.3.3 Instream, channel maintenance, and environmental flows 221 7.4 River health 223 7.5 Summary 224

Chapter 8 Rivers in the landscape 225 8.1 Rivers and topography 225 8.1.1 Tectonic influences on river geometry 226 8.1.2 Effects of on 228 8.1.3 Indicators of relations between rivers and landscape evolution 228 8.1.4 Tectonics, topography, and large rivers 229 8.2 Geomorphic process domains 230 8.3 Connectivity 232 8.4 Climatic signatures 234 8.4.1 High latitudes 234 8.4.2 Low latitudes 235 8.4.3 Warm drylands 236 8.5 Rivers with a history 237 8.5.1 Upper South Platte River drainage, Colorado, USA 240 8.5.2 Upper Rio Chagres, Panama 242 8.5.3 Mackenzie River drainage, Canada 244 8.5.4 Oregon Coast Range, USA 246 8.5.5 Yuma Wash, Arizona, USA 248 8.6 The greater context 250

References 255 Index 311 Acknowledgements

Although I had contemplated writing a fluvial geo- greater environmental context, from the atmosphere morphology text prior to starting work on this text to the regional tectonics, from history over in 2012, I put off the task as being intimidating in hundreds of millions of years to human history of its complexity and requirements of time and energy. the past few decades, and from interactions between I owe the initiation of this text to the conjunction plants, animals, and rivers to human manipulation of several things. I used David Knighton’s excellent of rivers. Fluvial Form and Process as a graduate student in Thefirstdraftofthevolumewasalsosub- Vic Baker’s fluvial seminar at the stantially modified in response to comments from University of Arizona in 1984, and when I began DeAnnaLaurelandinresponsetoanextensiveand teachingmyowngraduatecourseinfluvialgeomor- thorough review by Mike Church. I thank them phology at Colorado State University in 1989 I used bothforthetimeandeffortthattheyputintocriti- this text and then the second edition published in cally reading the draft. Any errors or omissions that 1998, Fluvial Form and Process: A New Perspective. remain are of course mine. Asthevolumeofresearchbeingpublishedinfluvial In addition to those who contributed directly geomorphology steadily increased, the 1998 text to this volume, I would like to acknowledge the became dated. Other texts with strong coverage influence of my location at Colorado State Uni- of certain aspects of fluvial geomorphology were versity. I am fortunate to be at an institution that published, but none of them reflected the integrated puts a great deal of emphasis on water and that approach on which the content of my graduate class hosts multiple strong programs in water. Because increasingly relied. Consequently, the need to write oftheeaseofinteractingwithfacultyandgradu- anewtextwasnaggingatthebackofmymind ate students in civil and environmental engineer- when Martin Doyle suggested to Rachael Ballard ing, aquatic and riparian ecology, watershed sci- of Wiley-Blackwell that she contact me about a ence, and geography, my approach to rivers has new volume. I think Rachael was surprised at how been transdisciplinary for many years. I would also readily I agreed to undertake the new volume, but like to acknowledge the continual nudges toward a the timing was right. transdisciplinary approach from my graduate stu- The organization and content of the volume were dents, who are likely to be more intellectually nim- modified in response to comments from six anony- ble in thinking about rivers outside of traditional mous reviewers who examined an initial extended disciplinaryboundaries.Hereagain,mylocationat outline. One of the reviewers commented on the Colorado State is critical, as I have been able to easily proposed title, Rivers in the Landscape, “Where else accept graduate advisees with undergraduate majors would they be?” Indeed, but we too commonly forget as diverse as mathematics, environmental science, thelandscapecontextwhenweisolatesegmentsof biology, geology, and engineering. This breadth of rivers for the purposes of research or management. approach is better suited to the real breadth of river I chose the title to emphasize the importance of the process and form, which we as scientists partition

ix x Acknowledgements into intellectual disciplines at our own peril of artifi- I would like to see all river scientists begin with cially constraining our ability to ask and answer fun- the attitude expressed on a sticker on my refrigerator damental questions. at home, “Yay! Rivers!,”and then follow their curios- ity wherever it may lead them. About the Companion Website

This book is accompanied by a companion website:

www.wiley.com/go/wohl/riverslandscape

The website includes: r Powerpointsofallfiguresfromthebookfordownloading r PDFs of all tables from the book for downloading r Supplementary materials

xi

Chapter 1 Introduction

Rivers are the great shapers of terrestrial landscapes. unintended consequences, and effectively mitigat- VeryfewpointsonEarthabovesealeveldonotlie ing these consequences requires that we understand within a . Even points distant from the riversinthebroadestsense,asshapersandintegra- nearest channel are likely influenced by that chan- tors of landscape. nel. raises rock thousands of meters Geomorphologist Luna Leopold once described above sea level. Precipitation falling on the uplifted rivers as the gutters down which flow the ruins of concentrates into channels that carry sed- continents (Leopold et al., 1964). His father, Aldo iment downward to the oceans and influence the Leopold, described the functioning of an ecosystem steepness of adjacent hillslopes by governing the as a round river to emphasize the cycling of nutri- rate at which the whole landscape is incised. Rivers ents and energy. Rivers can be thought of as gutters, migrate laterally across lowlands, creating a com- with a strong unidirectional and linear movement plex topography of terraces, floodplain wetlands, ofwater,sediment,andothermaterials.Riverscan and channels. Subtle differences in elevation, grain also be thought of as more broadly connected sys- size, and soil moisture across this topography con- temswithbidirectionalfluxesofenergyandmatter trol the movement of groundwater and the distribu- between the channels of the river network and the tion of plants and animals. greaterenvironment.Thisvolumeemphasizesthe Throughout human history, people have settled latter viewpoint. disproportionately along rivers, relying on the rivers for water supply, transport, fertile agricultural soils, wastedisposal,andfoodfromaquaticandriparian 1.1 Connectivity and organisms. People have also devoted a tremendous amount of time and energy to altering river pro- inequality cess and form. We are not unique in this respect: ecologists refer to various organisms, from beaver Contemporary research and conceptual models of to some species of riparian trees, as ecosystem engi- riverformandprocessincreasinglyexplicitlyrec- neers in recognition of the ability of these organ- ognize the important of connectivity. Connectivity, isms to alter the surrounding environment. People sometimes referred to as coupling (Brunsden and are unique in the extent to and intensity with which Thornes, 1979), is multi-faceted. Hydrologic connec- we alter rivers. In many cases, river engineering has tivity can refer to the movement of water down

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

1 2 Rivers in the Landscape a hillslope in the surface and/or subsurface, from thecresttothetoeofahillslope,butnotintothe hillslopes into channels, or along a channel net- adjacent channel because the sediment is trapped work (Pringle, 2001; Bracken and Croke, 2007). and stored in alluvial fans perched on stream ter- River connectivity refers to water-mediated fluxes races? Recent research focuses on quantifying con- within the channel network (Ward, 1997). Sediment nectivity or developing indices of connectivity using connectivity can refer to the movement, or stor- tools such as high-resolution digital terrain models age, of sediment down hillslopes, into channels, or derived from aerial LiDAR (Cavalli et al., 2013) or along channel networks (Harvey, 1997; Fryirs et al., direct measurements of fluxes (Jaeger and Olden, 2007; Kuo and Brierley, 2013). Biological connec- 2012). tivity refers to the ability of organisms or plant These dimensions of connectivity are important propagules to disperse between suitable habitats or for adequately characterizing fluxes within a land- between isolated populations for breeding. Land- scape, and for understanding how human activi- scape connectivity can refer to the movement of ties alter those fluxes (Kondolf et al., 2006). Many water, sediment, or other materials between indi- human actions substantially reduce connectivity vidual landforms (Brierley et al., 2006). Structural withinarivernetwork.Damsalterhydrologiccon- connectivity describes the extent to which land- nectivity and may effectively interrupt or elimi- scape units—which can range in scale from <1m nate connectivity of sediment and some organisms for bunchgrasses dispersed across exposed soil to along a river. Levees and bank stabilization inter- the configuration of hillslopes and bottoms rupt or prevent connectivity between the channel across thousands of meters—are physically linked and adjacent floodplain. Flow diversions, in con- to one another. Functional connectivity describes trast, may increase connectivity between drainage the process-specific interactions between multiple networks, allowing exotic organisms to migrate with structural characteristics, such as runoff and sedi- the diverted water and colonize a river network. ment moving downslope between the bunchgrasses Dredging, channelization (Figure S1.1), straighten- and exposed soil patches (Wainwright et al., 2011). ing, or other activities that reduce geomorphic com- Temporal variability, or connectedness of rainfall, plexity and storage of fine sediment and nutrients can create spatial variability, or connectedness of typically increase longitudinal connectivity of rivers flow paths, and thus functional connectivity along andassociateddownstreamfluxesofsedimentand the slope (Wainwright et al., 2011). solutes. By limiting overbank flows, however, these Whatever form of connectivity is under discus- alterations reduce lateral connectivity between the sion, the magnitude, duration, and extent of that channel and floodplain. Effective mitigation of connectivity are each important. Magnitude can be undesirable human alterations of rivers requires thoughtofasthevolumeofflux:isonlyatrickleof understanding the details of connectivity. water moving down a channel network, or a flood? Inextricable from connectivity is the idea of reser- Duration describes the time span of the connectiv- voirs, sinks, or storage: components of a river chan- ity: can fish disperse along a river network through- nel, river network, or other landscape feature in out an average flow year, or only during certain whichconnectivityisatleasttemporarilylimited. seasons of high flow? Closely associated with dura- Being able to quantify the magnitude and average tionistheideaofstorage.Ifsedimentstopsmov- storagetimeofmaterialinfluxiscriticaltounder- ing downstream during periods of lower discharge, standing connectivity, as is being able to predict thenthesedimentisatleasttemporarilystoredin the thresholds that define upper and lower lim- the streambed and banks. Organic matter can be itsofstorage.Sedimentmovingdownslopefroma stored on a floodplain until overbank flows or bank weathered bedrock outcrop toward a stream chan- erosion transport the organic matter back into the nel might remain in storage on a debris-flow fan for active channel. Extent is the spatial characteristic 2000 years before reaching the stream channel, for of connectivity: does sediment move readily from example, so that the fan limits connectivity between Chapter 1 Introduction 3 the slope and the channel at time spans of 100–103 More than 75% of the long-term sediment flux years. Or the sediment might progressively accumu- from mountain rivers in Taiwan occurs in less than late on the hillslope until a precipitation or seismic 1% of the time, during typhoon-generated floods triggercausestheslopetocrossathresholdofstabil- (Kao and Milliman, 2008). Approximately 50% of ity and fail in a mass movement that instantaneously the suspended sediment discharged by rivers of introduces much of the sediment into the stream. the Western Transverse Ranges of California, USA, Or the sediment might move quickly downslope and comes from the 10% of the basin underlain by into the channel as soon as the sediment is phys- weakly consolidated bedrock (Warrick and Mertes, ically detached from the bedrock outcrop, because 2009). Somewhere between 17% and 35% of the total theslopeangleistoohighforsedimentstorage. particulate organic carbon flux to the world’s oceans Focusing on coarse sediment transport in streams, comes from high-standing islands in the southwest Hooke (2003) distinguished: Pacific, which constitute only about 3% of Earth’s landmass (Lyons et al., 2002). One-third of the total amount of stream energy generated by the Tapi River r unconnected channel reaches with local sinks for of India during the monsoon season is expended sediment and lack of transport between reaches; on the day of the peak flood (Kale and Hire, 2007). r partially connected reaches with sediment transfer Along bedrock channels with large knickpoints, the only during large floods; great majority of channel incision occurs at the r connected reaches with coarse sediment transfer . during frequent floods; These are but a few of the many examples that r potentiallyconnectedreachesthatarecompetent are mentioned in the remainder of this volume. to transfer sediment, but lack a sediment supply; Becausenotallmomentsintimeorspotsonaland- and scape are of equal importance in shaping rivers, r disconnected reaches that were formerly con- effective understanding and management of rivers nected but are now obstructed by a feature such requiresknowledgeofhow,when,andwherefluxes as a dam. occur.

The point is that most natural and engineered river systems have some degree of retention of water, 1.2 Six degrees of sediment, solutes, and organisms, and understand- ing net and long-term fluxes of these quantities connection involves quantifying both movement and storage. Connectivity, storage, and fluxes are thus a cen- Any river network or segment of a single river exists tral component of river process and form. Connec- in a rich and complicated context that reflects fluxes tivity does not imply that all aspects of a connected of matter and energy between the river and the valley segment, river network, or landscape are of greaterenvironment,aswellasthehistoryofthese equal importance to fluxes of matter and energy. fluxes.Atanygivenmomentintime,theonlyfluxes Biogeochemists coined the phrases hot moment and that are likely to be obvious are longitudinal fluxes hot spot. Hot moment describes a short period of as water and sediment move downstream. Longitu- time with disproportionately high reaction rates rel- dinal fluxes, however, are only one of six degrees ative to longer intervening time periods. Hot spot of connection between a river and the environment describes a small area with disproportionately high (Wohl, 2010). reaction rates relative to the surroundings (McClain et al., 2003) (Figure S1.2). These ideas can be extrap- 1. The longitudinal connection is the most obvi- olated to rivers, because any aspect of river process ous and intuitive (Figure 1.1; Figure S1.3a). or form reflects inequalities in time and space. Water, sediment, and solutes move downstream. 4 Rivers in the Landscape

Upstream−downstream Atmosphere− channel Hillslope– channel

Floodplain− channel Hyporheic− channel

Groundwater− channel

Figure 1.1 Schematic illustration of the six degrees of connection between rivers and the greater landscape. The segment of channel (lighter gray) shown here is connected to: upstream and downstream portions of the river network; adjacent uplands; the floodplain; groundwater; the hyporheic zone (darker gray); and the atmosphere. The photograph for upstream–downstream connection was taken during a flood on the Paria River, a tributary ofthe Colorado River that enters just downstream from Glen Dam in Arizona, USA. In this view, the Paria is turbid with suspended sediment whereas the Colorado, which is released from the base of the dam, is clear. The photograph for the hillslope–channel connection shows a large landslide entering the Dudh Kosi River in Nepal. The photograph for the floodplain–channel connection was taken along the Rio Jutai, a blackwater tributary of the Amazon River, during the annual flood in early June. In this view the “flooded forest” is submerged by several meters ofwater.The photograph for hyporheic–channel connection shows a larval aquatic insect (macroinvertebrate) as an example of the organisms that can move between the channel and the hyporheic environment. The photograph for atmosphere– channel connection shows a mayfly emerging from the river prior to entering the atmosphere as a winged adult (image courtesy of Jeremy Monroe, Freshwaters Illustrated).

Globally, rivers transport an estimated 7819 mil- travel tens to hundreds of kilometers upstream lion tons of sediment to the oceans (Milliman from the ocean to spawn. and Syvitski, 1992), and approximately 0.9 Pg 2. The lateral connection between the river channel (1 Pg = 1015 g) of carbon per year (Aufdenkampe and adjacent floodplain is most obvious during et al., 2011). Organisms move actively up- and periods of flow with sufficient volume to over- downstream to new habitat and passively drift top the banks and spread across the unchanneled downstream with the current. Both European valley bottom (Figure S1.3c). Water, sediment, (Anguilla anguilla) and American eels (Anguilla solutes, and organisms disperse from the chan- rostrata) migrate from rivers to the Sargasso Sea nel onto the floodplain during the rising and peak off Bermuda for spawning, covering a distance of stages of a flood, and some of these materials as much as 5600 km, and numerous species of concentrate once more in the channel during the salmon (Salmo and Oncorhynchus spp.) typically falling stage of the flood. Chapter 1 Introduction 5

High rates of primary production by photo- an extensive network of macropores and pipes, synthetic organisms occur during the rising limb facilitates rapid downslope transmission of runoff of the flood, providing food for the consumer early in the wet season. Water repellency declines organisms that follow the flood pulse onto the as the wet season continues, allowing infiltration floodplain. High rates of decomposition occur toincreaseandrunofftodecrease.Bythepeakof duringthefloodpeak,andtheresultingnutri- the wet season, however, saturated soils can pro- ents concentrate back in the channel during the mote rapid, abundant (Niedzialek descending limb of the flood. Sediment moves and Ogden, 2005). Rivers fed by snowmelt typ- onto the floodplain during the rising limb, typi- ically exhibit an ionic pulse when the release cally remaining in storage within the floodplain of solutes from the snowpack and the flush- until bank erosion returns the sediment to the ingofweatheringproductsfromthesoilcreate channel (Dunne et al., 1998). Tropical river ecol- the highest solute concentrations in the stream ogists refer to the regular annual fluxes between water at the initiation of snowmelt (Williams the channel and the floodplain as the flood pulse, and Melack, 1991). Mineral sediment and organic a phrase now used to refer to fluxes during floods matter coming from the uplands can originate in of any recurrence interval or magnitude sufficient episodic,pointsourcessuchaslandslides(Hilton to create overbank flow (Junk et al., 1989; Bayley, et al., 2008a, 2008b) or via more diffuse, gradual 1991). Flow pulses—fluctuations in surface water erosion. below the bankfull level—create similar processes 4.Verticalfluxeslinkthechanneltothezoneof within secondary channels or areas of flow sepa- subsurface flow immediately below the channel, ration along a single, confined channel (Tockner with flow paths that originate and terminate at et al., 2000). the stream (Figure S1.3e). This subsurface region Levees, channelization, and flow regulation is known as the hyporheic zone,fromtheGreek have so restricted overbank flooding along most roots“hypo”forunderorbeneathand“rheo” of the world’s large and medium rivers that it for flow or current. Water, sediment, solutes, and is now easy to underestimate the spatial extent small organisms such as microbes and macroin- and duration of flooding once present along low- vertebrates moving between the surface and landrivers.TheAmazon,byfartheworld’slargest subsurface can strongly influence the volume, river and still one of the least engineered, can temperature, and chemistry of flow in the river extend across 50 km of floodplain during the channel, and hyporheic habitat can account for seasonal flood, which can last more than three a fifth of the invertebrate production in a river months. ecosystem (Smock et al., 1992). The hyporheic 3. Another lateral connection is that between adja- zone can extend more than 2 km laterally from cent uplands and the river channel (Figure S1.3b). the channel in wide valleys and to depths of 10 m This is more likely to be a one-way flux, with (Stanford and Ward, 1988). water, sediment, and solutes moving downslope 5. Deeper vertical fluxes between the river and the at the surface and subsurface into the channel. saturated zone of the groundwater can also occur The pathways, rates and magnitudes of flux from in both directions, with water and solutes mov- the uplands typically exhibit substantial spatial ingintothechannelinagaining stream or into and temporal variability. During an individual the groundwater in a losing stream (Figure S1.3d). rainstorm, for example, water flowing across sat- Human activities can create gaining and losing urated ground may become a progressively more streams. Groundwater withdrawal that lowers the importantsourceofrunoffasinfiltrationcapac- water table sufficiently to prevent groundwater ity declines (Dykes and Thornes, 2000). Dur- flow into the channel, for example, can substan- ing the dry season, soils in the seasonal trop- tially reduce stream flow in dryland rivers (Falke ics can develop water repellency that, along with et al., 2011). 6 Rivers in the Landscape

As in exchanges between the hyporheic zone river water into the air. Although essentially invis- and surface flow, exchanges between ground and ible, these fluxes are widespread and important. surface water can influence the temperature and chemistry of river water. Solute concentrations Conceptualizing a river as having six degrees of typically increase toward saturation as ground- connection with the greater environment empha- water moves relatively slowly through sediment sizes how diverse aspects of connectivity influence or bedrock (Constantz and Stonestrom, 2003), so river process and form. This conceptualization also groundwater inputs can strongly influence river emphasizes the diversity of temporal and spatial solute concentrations. The flow of rivers originat- scales across which connectivity occurs. ing from large springs in carbonate or landscapes with layered basalt flows, for exam- ple, can be almost entirely from groundwater 1.3 Rivers as integrators (Gannett et al., 2003) (Figure S1.4). Hydraulic conductivity, a measure of perme- Thanks to the extensive and sometimes subtle fluxes ability and groundwater flow rate, can range over between a river and the greater environment, the 12 orders of magnitude (Domenico and Schwartz, formsandprocessesofariverintegratethephysical, 1998). Consequently, the travel times of ground- chemical, and biotic processes—contemporary and water from areas of recharge to areas of dis- historical—within the environment. This may seem chargeinspringsorriverscanrangefromless obvious when considering Figure 1.1, but represents than a day to more than a million years (Alley the most profound summation possible regarding et al., 2002). This means that vertical connectiv- rivers,becauseoftheimplications. ity between groundwater and channels typically If a river integrates diverse and seemingly unre- influences river dynamics over long timescales lated processes within the greater environment, for relative to hyporheic flow. example, then attempting to manage the river or 6. The vertical connection between the river and some segment of the river in isolation from those the atmosphere (Figure S1.3f) can be obvious processesisabsurd. when precipitation falls directly on the river or If a river integrates … then human activities far an aquatic insect emerges from the river for fromthephysicalboundariesofthechannelmay the winged, terrestrial, adult phase of its life. strongly influence the river, as when increasing Otherfluxesinvolvedinthisconnectionarelikely atmospheric dust transport from the deserts of the to be much less visible. Water evaporates into southwestern United States alters snowpack melt- the atmosphere, especially from the oceans, and ing and the resulting spring snowmelt hydrograph moveslongdistancesbeforefallingontoland- and water chemistry in rivers of the Rocky Moun- scapes that drain into rivers. En route, the water tains (Clow et al., 2002). Another example comes vapor acquires very fine particulates, including: from the Mississippi River, where concentrations of dust that may have traveled from a different hemi- nitratehaveincreasedbytwotofivetimessincethe sphere (Prospero, 1999); nitrates from vehicles, early 1900s as farmers applied increasing quantities industrial emissions, and agricultural sources— of nitrogen fertilizers to upland crop fields across the nitrates are deposited with rain and snow, and the Mississippi’s huge drainage basin. The resulting as particles and gases, in rivers hundreds of kilo- flux of nitrate down the river to the Gulf of Mexico meters away (Heuer et al., 2000); and mercury tripled during the last 30 years of the twentieth cen- released by vehicles and by coal-burning power tury, resulting in massive algal blooms that cover a plants (Grahame and Schlesinger, 2007). Volatile swath of the Gulf as big as New Jersey (∼20,000 km2) organic compounds—solvents such as tetra- each year and in some years move out of the Gulf and chloroethylene, chlorinated compounds such as up the eastern coast of the United States (Goolsby chloroform, and others—volatilize from polluted et al., 1999). Chapter 1 Introduction 7

If a river integrates … then historical resource reduced sediment yield to the world’s oceans by uses of which most people are now unaware may 1.4billionmetrictonsperyearbecauseofretention continue to strongly influence contemporary river behind dams (Syvitski et al., 2005). The result of this process and form (Macklin and Lewin, 2008). reduced coastal sediment yield has been widespread Meandering gravel-bedded streams in the east- delta and near-shore erosion (Crossland et al., 2005; ern United States are typically bordered by fine- Yang et al., 2011). grained deposits that were formerly interpreted as In summary, a river integrates fluxes across a self-formed floodplains. Prior to European settle- much larger and more diverse environment than ment, however, these river networks consisted of the channel itself. Consequently, understanding and small anabranching channels within extensive veg- effectively managing river process and form is much etated wetlands that were buried by up to 5 m of more challenging than is likely to be recognized if a slackwater sedimentation behind tens of thousands river segment is manipulated as though it was spa- of seventeenth- to nineteenth-century milldams tially and temporally isolated. (Walter and Merritts, 2008). The ubiquitous fine sediments are thus fill terraces that reflect ongoing adjustment as the milldams breached and the chan- 1.4 Organization of nels incised. Another example comes from rivers in the Carpathian Mountains of Poland. Agriculture this volume began in the region during the thirteenth and four- teenth centuries, and the increased sediment yield The title of this book, Rivers in the Landscape, resulted in overbank along - reflects the inherent connections between a river ing rivers draining the mountains (Klimek, 1987). and the landscape. Landscape is defined here as the When the proportion of crop lands that remained physical, chemical, and biotic environment of the bare for some portion of the year increased with critical zone—Earth’s outer layer, from the top of more widespread cultivation of potatoes during the the vegetation canopy to the base of the soil and second half of the nineteenth century, the further groundwater, that supports life. The critical zone increases in sediment yield caused some of the represents the intersection of atmosphere, water, meandering rivers to assume a braided planform. soil, and ecosystems. Recent research increasingly If a river integrates … then altering river process reminds us of what perhaps should always have been andformatonepointintherivernetworkmayaffect obvious: rivers do not merely flow through a land- other portions of the network in unforeseen ways. scape in isolation, but rather interact with that land- ThetwoDjerdapdamsontheDanubeRiverwhereit scape in complex and fascinating ways. Riverine flows through Romania were built in 1970 and 1984. vegetation, for example, does not just increase the These massive dams, along with dozens of smaller hydraulic resistance of overbank flow—the vegeta- upstream dams, have reduced sediment yields to the tioncanalterthedefaultriverplanformfrombraid- river’s delta by 70% and silica export to the Black ing to meandering (Tal and Paola, 2007). Rivers do Sea by two-thirds relative to fluxes of these materi- not flow passively down steep topography created als prior to the last third of the twentieth century. by tectonic uplift—removal of mass through river- The reduced fluxes have caused erosion of the delta ine erosion can increase the upward flux of molten and a shift in the Black Sea’s phytoplankton commu- rock and tectonic uplift (Zeitler et al., 2001). nities from siliceous diatoms to non-siliceous coc- Recognition of the connections between rivers colithophores and flagellates. These changes have and landscapes implies that the topics tradition- stimulated algal blooms and destabilized the Black ally covered in a fluvial geomorphology text— Sea ecosystem (Humborg et al., 1997). Globally, hydraulics, sediment transport, river geometry—be humans have increased sediment supplied to and treated in a manner that explicitly recognizes the transported by rivers as a result of , yet influences exerted on river process and form by 8 Rivers in the Landscape entities beyond the channel boundaries. Con- focusofChapter7,whichsummarizesthedepthand sequently, this book builds from traditional breadth of human alterations of rivers. understanding of rivers toward the larger, more Chapter 8 metaphorically steps back to use the comprehensive viewpoint. knowledgeofprocessandformdevelopedinpre- Chapter 2 covers the development of channels ceding chapters as a means to understand rivers in and channel networks, including how water, sed- a landscape context. This chapter starts with a dis- iment, and solutes are produced; how they move cussion of how topography influences the spatial from uplands into channels; how channel heads distribution of river networks and energy expen- form; and how channel networks extend across the diture within rivers, how rivers influence rates of landscape. This chapter addresses the processes by landscape denudation, and the indicators used to whichwatermovesacrossandthroughunchan- infer relations between rivers and landscape evo- nelized hillslopes and concentrates sufficiently to lution. Spatial differentiation of geomorphic pro- create channels. cess and form within river basins is discussed, Chapter 3 covers channel processes, with a focus followed by a re-examination of connectivity. Dis- on energy (hydraulics) and quantities (hydrology). tinctive river characteristics associated with high Knowledge of the basic mechanics of channelized and low latitudes and arid regions provide examples flow is integral to understanding sediment ero- oftheimportanceoflandscapecontext,asdoaseries sion, transport and deposition, and adjustment of of case studies. These case studies illustrate how channel form. place-specific details of geology, climate, and land Chapter 4 covers the movement of sediment in usehistoryinfluenceriverprocessandform,aswell channels. The discussion begins with the sediment as the management implications of these influences. texture of channel beds and the processes that initi- One of the challenges in writing a reasonably con- ate motion of non-cohesive and cohesive sediment. cise fluvial geomorphology text is the tremendous Once sediment is mobilized from the streambed and volumeofresearchconductedonriverswithinthe banks, it can be transported in solution, in suspen- past century. Scientists from diverse backgrounds in sion,orincontactwiththebed,andcanbeorga- geology, geography, civil engineering, and other dis- nized into bedforms. ciplines study river process and form via Chapter 5 addresses channel form, exploring how water and sediment movement shape channel geom- r direct measurements and experimental manipula- etry through time and space. Interactions between tionsofrealrivers; process and form are implicit throughout Chapters r indirect measurements using remote sensing 3 and 4, but Chapter 5 explicitly examines feed- imagery from space-based (e.g., aerial pho- backs between process and form at increasingly tographs, satellite imagery, airborne LiDAR) larger spatial scales, from cross-sectional geometry, and ground platforms (e.g., ground penetrating through channel planform, and longitudinal gradi- radar); ent, to downstream trends along a river and across a r physical experiments in a laboratory; river basin. r numerical models; and Chapter 6 summarizes process and form of flu- r integration of all of these approaches. vially created and maintained features outside of the immediate channel—floodplains, terraces, allu- Another fundamental challenge is the diversity vial fans, deltas, and estuaries. These river landforms of rivers. Water flows downslope under the influ- both reflect and influence channel process and form. ence of gravity. The basic physics are the same Humans and human influences on the landscape in any environment, but the ability to generalize are now ubiquitous. Although human influences on beyond the most basic level is typically obscured rivers are mentioned where appropriate in each of by the local, place-specific details and history of a thefirstsixchapters,theseinfluencesaretheexplicit particular river. As fluvial geomorphology continues Chapter 1 Introduction 9 to develop as a discipline, there remains an underly- conceptual models and methods of inquiry in gov- ing tension within the community between investi- erning the questions that scientists ask. If we view gators who emphasize quantification as a means of rivers as complex systems with multiple interac- identifying physical principles and mechanisms act- tions between different components, we are more ing across a range of specific landscapes (Dietrich likely to focus on the factors that control those et al., 2003) and investigators who emphasize the use interactions and on ways to quantify and predict of historical and sedimentary records as a means of the interactions. If we view rivers as predominantly identifying the role of contingency and site-specific physical systems, we are more likely to neglect the characteristics in river process and form. interactions among hydraulics, sediment dynam- Until perhaps the 1960s or 1970s, the great major- ics, and aquatic and riparian organisms. Even when ity of river research focused on medium-sized, low- not explicitly recognized, our conceptual models of to medium-gradient, sand-bed rivers. These were rivers tend to constrain the questions that we con- the most accessible rivers for scientists living pri- sider interesting and important and the methods marily in the temperate latitudes, and the founda- we use to examine these questions (Grant et al., tional research conducted on these rivers gave rise 2013). Studies of sediment transport, for example, to widely used conceptual models and equations for that employ a Eulerian framework focus on the flux hydraulics, sediment transport, and channel geome- of sediment within a spatially bounded area—a very try. As investigators have subsequently spent more useful approach for developing a sediment budget, time quantitatively examining rivers with steeper whereas a Lagrangian framework in which specific gradients and more resistant boundaries (gravel-bed objects are tracked through time can provide more rivers, bedrock rivers, mountain rivers) and greater insight into actual mechanisms of sediment move- hydrologic variability (seasonal tropics, drylands), ment (Doyle and Ensign, 2009). as well as rivers at higher (boreal, arctic) and lower A conceptual model results from assumptions (tropical) latitudes, the ability of the “classic” mod- about how a river functions. The conceptual model els and equations to adequately describe process and can be qualitative or quantitative. A quantitative form across the known spectrum of rivers becomes model can be more precise than a qualitative model, weaker. Throughout this volume, I explicitly address but is not necessarily more accurate. Drawing on the some of the unique characteristics of rivers beyond second chapter of Leopold et al.’s (1964) fluvial geo- temperatezonesand-bedchannels. morphology text for inspiration, the remainder of My intent in this text is to maintain conciseness this section uses a landscape with which I am very whilereflectingthediversityofnaturalriversand familiar to explore the different conceptual models the methods of studying rivers. More extensive dis- and approaches that investigators employ to under- cussions of this diversity are provided in supplemen- stand river segments, river networks, and the greater talmaterialontheaccompanyingwebsite.Along landscape. these lines, the references cited in the text are not an exhaustive list, but rather a starting point that com- bines some foundational studies and particularly 1.5.1 The Colorado Front Range integrativeorinsightfulrecentstudies.Thewebsite provides a more thorough list of references for many Atop the Precambrian-age crystalline rocks that of the topics covered in the text. form the continental divide in Colorado, you can stand shivering in the cold wind even at the height of summer. Here, 4000 m above sea level, bedrock 1.5 Understanding rivers topography crests in a series of ridges and peaks that divide water flowing west to the Pacific Ocean Recent emphasis on connectivity in landscapes and water flowing east to the Atlantic (Figure 1.2). and river networks illustrates the importance of Insomeplacesthedivideisasharp-edgedridgeof 10 Rivers in the Landscape

Figure 1.2 Landscapes and river corridors in and adjacent to the Colorado Front Range. Upper left: view east from the summit surfaces at the continental divide. The coarse blocks in the foreground are periglacially weathered boulders and bedrock. The surface drops steeply into a glaciated valley that transitions downstream (out of sight) into a fluvial valley. Upper right: view northwest from a hogback, an asymmetrical hill of sandstone and limestone strata dipping steeply to the right in this view, with an intervening valley formed in shales. Lower right: the South Platte River near Fort Morgan, Colorado, in the low relief environment of the Great Plain. This sand-bed channel was historically much wider and had a braided planform, but flow regulation has resulted in encroachment of riparian vegetation and transformation to a single relatively narrow channel. This river heads high in the mountains. Lower left: view of smaller drainages that head on the Great Plains, here at Pawnee National Grassland. These channels have downcut within the past few decades, largely via piping erosion. bedrock and periglacial boulders with talus chutes Climate grows progressively warmer and drier and waterfalls. In other places, small alpine streams at lower elevations, and subalpine forest gives way meander across broad, gently undulating surfaces. to more open montane forest with more frequent Sharp or broad, the heights drop precipitously wildfires and associated debris flows. Warm, moist down to glacial cirques and troughs. Rivers alter- masses of air moving inland from the south- nate between paternoster lakes and steep cascades as east during summer are forced upward as they they flow through subalpine conifer forests. Beyond near the Colorado Rockies, and the water vapor the terminal glacial moraine, each valley contin- being transported with the air masses cools, con- ues downward, alternating between steep, narrow denses, and falls as rain. Most of this moisture gorges in which the river flows turbulent and aer- iswrungfromthecloudsatthelowertomid- ated or relatively wide with gentler gradi- dle elevations of the mountains, which can expe- ents along which the river flows through pools and rienceflashfloodsfromconvectivestorms,aswell . asthelatespringsnowmeltfloodsthatflowdown Chapter 1 Introduction 11 from the highest portions of the river network each energy exceeds a critical level defined by the average year. grain size of the sediment and (ii) is proportional to Atthebaseofthemountainsontheeasternside, the level of excess energy beyond the critical energy. the rivers gradually change from boulder- to cobble- The second assumption is illustrated by a generic bedchannelsastheyflowthroughaseriesofsteeply equation for bedload transport rate qb tilted sedimentary rocks forming asymmetrical hills. 𝜏 𝜏 n Beyond the hills lies the gently undulating topog- qb = k( − c) (1.1) raphy and steppe vegetation of the semiarid Great Plains, where sand-bed channels shrink back to a where k is an empirical constant, 𝜏 is boundary shear 𝜏 trickleaftertheannualsnowmeltpeakflow. stress, c is critical boundary shear stress for entrain- The dramatic topography and strong elevational ment, and n is an empirically derived exponent. This contrasts in climate and vegetation dominate ini- equation implies that bedload transport is propor- tial impressions of the Colorado Front Range. This tional to the amount of shear stress above the crit- leads to questions about how river process and ical level for moving sediment. Equation 1.1 is an form change moving downstream, and what fac- example of a flux equation.Forrivers,fluxequations tors influence these changes. At a basic level, we usually refer to flow–sediment interactions and pro- can address these questions using empirical or theo- cesses such as sediment flux within a channel retical approaches. Empirical approaches are largely The relatively narrow grain-size range of sand- inductive. In logic, to induce is to conclude or infer bed channels makes it easy to specify average grain general principles from particular examples. In an size,andtherelativeeaseofmobilityofsandgrains empirical approach, data are collected and analyzed makes the second assumption above reasonable. In a in order to establish relationships between vari- cobble- or boulder-bed channel, however, the wider ables. A fundamental challenge to empirical under- range of grain sizes means that larger grains can standing of rivers lies in generalizing from empir- smaller grains from the force of the flow and ical results defined by using a restricted database. limit the movement of the smaller grains. Conse- If I measure bedload transport along a cobble-bed quently, average grain size may not be a particularly mountain river segment for a year and demonstrate useful parameter for specifying the start of bedload that the majority of transport occurs when flow transport. Larger grains at the streambed surface can equals or exceeds half of the bankfull depth, can I prevent the movement of underlying smaller grains extrapolatefromthissitetootherrivers?WhatifI and create turbulence, so that bedload transport is repeat the measurements on a sand-bed river of the notlikelytohavealinearrelationshipwithflow plains and find that bedload transport begins at a energy. much lower level of flow? The problem of characterizing bedload transport Theoretical approaches formulate and test spe- in mountains and plains rivers can also be described cific statements based on established principles. To using the dichotomy of deterministic versus proba- deduce is to reason from the general to the particu- bilistic. Deterministic approaches assume that phys- lar. Theoretical approaches are more deductive, but ical laws control river process and form. Once these are typically hampered by a relative lack of estab- lawsareknown,riverbehaviorcanbepredictedfor lished geomorphic theory. Consequently, theoreti- agivensetofconditions. calapproachestoriverprocessandformcommonly Deterministic modeling of river processes relies draw heavily on related fields such as hydraulic engi- on five basic equations: neering in which the theory represents a system much more simple than most natural river channels. r continuity equations for water and sediment, Theoretical approaches to bedload transport r the flow momentum equation, developed by hydraulic engineers, for example, r a flow resistance equation, and assume that bedload transport (i) begins once flow r a sediment transport equation. 12 Rivers in the Landscape

Conservation equations or continuity equations ment transport coefficients in boulder-bed moun- are based on the fact that mass, momentum, and tain streams, for example, is limited by the extreme energy cannot be created or destroyed in any pro- spatialvariabilityinbedgrainsize,aswellasirregu- cess. The continuity equation for flow is simply larities in cross-sectional geometry caused by pieces of wood and lateral constrictions from bedrock out- Q = wdv (1.2) crops or very large boulders. Under these circum- stances, it is more effective to acknowledge a sub- where Q is discharge, w is flow width, d is flow stantial level of uncertainty in predicting bedload depth, and v is mean velocity. An example of a sed- transport: bedload movement may be described as iment version is the for sediment occurring when discharge falls within upper and continuity, lower bounds, rather than as a direct relationship between discharge and bedload transport. ( ) 𝜕𝜂 𝜕q Another approach to predicting bedload trans- 1 − 𝜆 =− (1.3) p 𝜕t 𝜕x portwouldbetouseaforceequation.Force equa- tions, typically the balance of forces involved in 𝜆 𝜂 where p is bed porosity, is bed elevation, t is time, erosional and depositional processes, describe a crit- q is volume transport rate of bed material load per ical level beyond which a process such as move- unit width, and x is direction of flow (Parker et al., ment of sediment on the streambed or erosion of the 2000). Another example of a continuity equation is stream bank begins. An example of a simple force asedimentbudgetthatequatessedimentstorageto equation for entrainment of a sediment particle in a sediment input minus output. river is The flow momentum equation is based on 𝜏 𝛾 Newton’s second law of motion and is well defined = f DS (1.4) theoretically. Momentum is a vector defined by the 𝜏 𝛾 product of mass and velocity. Momentum per unit where is shear stress acting on the sediment, f is time of water in a channel is 𝜌Qv,where𝜌 is water unit weight of the fluid, D is flow depth, and S is density, Q is discharge, and v is average velocity water-surface slope (Andrews, 1980). Again, the less (Robert, 2003). spatial and temporal variation there is in a system— The flow resistance and sediment transport equa- think sand bed (relatively uniform grain sizes) rather tions used in deterministic modeling of river pro- than boulder bed—the simpler it is to specify the cesses will include empirically derived coefficients. forces at work and to accurately assign average val- Deterministic modeling can thus use both empiri- ues to parameters such as flow depth and water- cal and theoretical understanding of a system, but surface slope in Equation 1.4. assumesthatriverprocessandformcanbedirectly Becausenaturalriversarecommonlyquitespa- predicted based on knowledge of existing parame- tially and temporally variable, geomorphologists try ters. to simplify process and form using physical exper- As the particular component of a river sys- iments in which one or more variables are directly tem being modeled increases in complexity, the and systematically manipulated to observe the effect interactions are increasingly difficult to represent on the whole system. Such manipulations are typ- using a set of closed equations, and predictions ically conducted in a laboratory setting or, more become less reliable (Knighton, 1998). Probabilistic rarely, in the field. approaches reflect an assumption that natural sys- Bedload transport in the boulder-bed mountain tems are so complex that complete deterministic rivers of the Colorado Front Range occurs 24 hours explanations are unrealistic because natural systems a day during the snowmelt peak. Much of the trans- include inherent randomness. The ability to spec- port actually occurs in the early hours of the morn- ify appropriate empirical flow resistance and sedi- ing when the previous afternoon’s snowmelt runoff Chapter 1 Introduction 13 comes down the river. Instead of attempting to another commonly used conceptual model of rivers directly measure bedload movement, and perhaps as complex and nonlinear systems. missing some of the sediment movement by not A complex system is composed of interconnected sampling the entire channel width or sampling con- partsthatasawholeexhibitoneormoreproper- tinuously, useful insights into sediment dynamics ties, including behavior, not obvious from the prop- canbegainedbycreatingascaled-downriverin erties of the individual parts. A complex system a flume and then measuring changes in bedload displays self-organization over time and emergence transport as discharge is varied. Physical experi- over scale. Emergence is defined as patterns that ments present challenges of scaling forces (can you arise from a multiplicity of relatively simple interac- effectively simulate the turbulence and associated tions. A tree topples into a river, for example, with hydraulic forces of a flow several meters deep?) and a portion of the roots still attached to the bank. of including all relevant variables (can you effec- The downed tree extends into the river, trapping tively simulate fluctuations in upland or tributary smaller pieces of wood in transport and forming a sedimentsupplytothemainchannel?).Experiments logjam. The logjam blocks flow, creating a backwater can nonetheless provide useful insights into process of lower velocity where sediment in transport settles and form in real channels. out. As the elevation of the streambed increases, flow Rivers can also be investigated by developing begins to spill over the channel banks and erodes a numerical simulations in which those variables and secondary channel that branches away before rejoin- interactions considered to be relevant are quan- ing the main channel downstream. Bank erosion tified (Coulthard and Van de Wiel, 2013). Simu- during formation of the secondary channel under- lation outcomes are then compared to real rivers mines more trees that fall into the river, forming to evaluate the accuracy of parameterization and, additional logjams that facilitate further channel once such accuracy is established, to test scenar- branching.Eventually,anetworkofbranchingand ios such as the effect of altering water or sediment rejoining channels that enhance wood recruitment yields to a river. Numerical simulations can be based andstorageispresent.Thetreefallanditsconse- on some combination of theoretical and empiri- quent effects that result in a multi-thread channel cally derived equations that can be deterministic segment are an example of a complex system (Fig- or probabilistic. A numerical simulation of bedload ure S1.5). transport, for example, might specify channel geom- In a nonlinear system,outputisnotdirectlypro- etry, streambed grain-size distribution, discharge, portional to input such that, mathematically, the and sediment input from upstream, and then use variable to be solved for cannot be written as a linear an equation such as Equation 1.3 to predict bed- combination of independent components because loadflux.Amongthechallengesofnumericalsim- of interactions among the components. For exam- ulation are identifying the relevant variables and ple, pieces of wood floating downstream in a river processes and parameterizing these variables and areinfluencednotonlybythehydraulicforceof processes. the flowing water, for example, but also by the In addition to downstream differences in movement of adjacent pieces of wood or the trap- streambed substrate and bedload dynamics, some ping effect of stationary instream wood. Because of the more obvious changes along river networks in the movement of wood down the channel does not the Colorado Front Range are the transitions from depend only on hydraulic force, this movement is an alpine meadows, to relatively dense subalpine forest example of a nonlinear system. of spruce and fir, more open montane pine forest, Although phrases such as nonlinear and complex and finally semiarid steppe. Along the forested systems were not commonly used until the 1990s, portions of the river networks, wood recruited the behavior described by these phrases was recog- from riparian forests can strongly influence channel nized decades earlier in descriptions of river pro- process and form. These interactions illustrate cess and form within the work of G.K. Gilbert (1877) 14 Rivers in the Landscape and, in the mid-twentieth century, the work of Luna Thresholds involve abrupt changes in process or Leopold, Stanley Schumm, and others (Supplemen- form.Externalorextrinsicthresholdsarecrossedas tal Section 1.5). a result of changes in external controls. Internal or Rivers are also viewed as open systems,char- intrinsic thresholds can be crossed in the absence of acterized by a continual exchange of matter and changes in external variables (Schumm, 1979). An energy with the surrounding environment (Chor- example of an external threshold comes from hill- ley, 1962). Such exchanges might be obvious at the slope hydrology, when the early stages of precipi- scale of a channel segment with fluxes of water and tation cause shallow infiltration and relatively slow sediment from upstream and upland sources and downslope movement of water via subsurface dif- to downstream portions of the river. As empha- fusion. When sufficient water infiltrates to reach sized in the opening discussion of connectivity, deeper layers with preferential flow paths in the form even the largest river networks also experience of soil pipes, downslope water delivery to channels continual inputs of matter and energy from the abruptlybecomesmuchmorerapid.Inthisexam- atmosphere and the lithosphere, sometimes accom- ple, the abrupt change in downslope water delivery plished by the activities of organisms. Snow falling is externally forced by increasing volumes of precip- in the Colorado Front Range reflects the dynam- itation. ics of cold, dry Arctic air masses moving southward The ephemeral tributaries that head on the dry and interacting with slightly warmer air carrying eastern steppe of Colorado provide an example of much more moisture and moving inland from the internal thresholds. Over hundreds to thousands of Pacific Ocean. The melting of the resulting snow- years, these channels alternately incise to form steep- pack,andconsequentlythetimingandvolumeof sided gullies or arroyos, and then aggrade to form snowmelt runoff in the rivers, is influenced not relatively shallow swales (Figure 1.2). These alternat- only by air temperature, but also by deposition of ingepisodesofcutandfillcanrepresentcrossingof wind-blown dust that can come from nearby sources anexternalthresholdinresponsetofluctuationsin such as the deserts of the southwestern United precipitation, vegetation, and runoff. States, and also from very distant sources in Asia Alternating cut and fill can also occur in response (Painter et al., 2010). to crossing an internal threshold of sediment trans- Viewing a river as a complex, open system implies portwithinthechannel.Streamflowinthesechan- that, at whatever scale the river is considered, it con- nels is brief and infrequent, and sediment can be tains multiple, interacting components. Interactions deposited midway down a channel as discharge between components include feedbacks, thresholds, declines because of evaporation and infiltration into and lag times, and can create equifinality. the streambed. Repeated deposition of sediment Feedback refers to interactions among vari- partway along the stream’s longitudinal profile can ables. Self-enhancing feedbacks promote continuing develop a steeper section of the bed at which a head- change, as when sand grains saltating across bedrock cuteventuallyforms,triggeringawaveofupstream- are preferentially trapped on accumulations of sand migrating incision. All of this can occur in the that increase with time. The fallen log that initi- absence of any change in external variables such as ates a logjam and eventually a network of secondary precipitation, runoff, or sediment inputs. channels that promote additional wood recruitment Lag time typically refers to the delay between a and logjams is another example of a self-enhancing change in an external variable, such as an increase feedback. Self-arresting feedbacks limit change. For in water yield, and the response of the river, such example, a lateral channel constriction causes an as bank erosion. The cobble-bed streams of the sub- increase in flow velocity, which results in erosion alpine forest in the Front Range provide an exam- of the constriction until the velocity drops below a ple. Commercial ski resorts in this region divert river magnitude capable of causing erosion of the chan- watertomakesnowfortheirskiruns.Whenthis nel boundaries. artificially created snow melts the next spring, the Chapter 1 Introduction 15 runoff commonly goes into a different channel than cult to infer processes from form (Chorley, 1962). thesourceofthewater.Thesereceivingchannelscan Channel incision leading to terrace formation along have peak flows more than 200% larger than would theprimaryriversoftheFrontRangemighthave result from natural runoff. Channels along which resulted from (i) lowering of , or (ii) fluc- dense riparian vegetation and cohesive silt and clay tuations in water and sediment supply to the river increase bank resistance take longer to respond to associated with the advance and retreat of Pleis- increased peak flows than channels with less ero- tocene valley glaciers, or with widespread defor- sionally resistant banks (David et al., 2009) (Fig- estation and mining during the nineteenth century ure S1.6). (Schumm, 1991). Data on the age, spatial extent, and Where an external disturbance is very intense or stratigraphyoftheterracesaswellasindependent widespread, lag times can be minimal. An intense information on the timing and nature of base level wildfire in the montane zone of the Colorado Front change, glaciations, and historical land use are nec- Range during summer 2012 completely consumed essary to explain terrace formation. hillslope vegetation over hundreds of hectares of Underlying conceptualizations such as feedbacks pine forest underlain by weathered granite. The first and thresholds is one of the most widely used flu- rainstorms following the fire resulted in widespread vial conceptual models: the idea that a river can erosion of hillslopes and headwater channels, and exhibit various forms of stability, or equilibrium aggradation of larger channels (Figure S1.7). (Gilbert, 1877). Equilibrium typically refers to a con- Sediment accumulation in the larger channels dition with no net change, and is thus very depen- is of particular concern because these rivers sup- dent on the time span being considered. A river that ply municipal drinking water to communities along undergoessubstantialchannelchangeduringashort thebaseoftheFrontRange.Watermanagerstry- duration flood can nonetheless be in equilibrium ing to maintain storage capacity at intake structures when considered over a decade because subsequent and limit turbidity associated with suspended sedi- smaller flows rework the erosional and depositional ment and organic matter could use the force equa- features created by the flood. Consequently, differ- tion for sediment entrainment, the continuity equa- ent forms of equilibrium can be distinguished with tions for flow and sediment, and the flux equation respect to time span (Figure 1.3). for bedload transport mentioned earlier to quantify Equilibrium implies that multiple interacting sediment transport. They could also use diffusion variables within a river can reach a state of stabil- equations, which describe the movement of matter, ity. This is reflected in the widely used definition momentum, or energy in a medium in response to of a graded river as a channel in which streambed some gradient, such as the turbulent mixing of sus- slope is adjusted to prevailing water and sediment pended sediment driven by gradients in flow energy discharges, such that the channel neither aggrades (Robert, 2003). An example particularly relevant to nor degrades and the slope remains constant over thedepositionofhillslopesedimentmobilizedafter the time interval of interest (Mackin, 1948). wildfire comes from unit sediment flux q in a river Equilibrium also implies that a river will change depositional system in response to changes in the supply of energy or material. Pleistocene valley glaciation in the upper 𝜕h portion of the Front Range changed water and sed- q = 𝜈 (1.5) 𝜕x iment yields to downstream portions of the river network. Thinking of these river networks within a where 𝜈 is diffusivity, h is elevation, and x is distance framework of equilibrium raises questions regard- downstream (Voller and Paola, 2010). ing how, and how rapidly or over what time span, Equifinality, also known as convergence, refers to the rivers responded to altered water and sedi- the fact that different processes and causes can pro- ment supplies during glacial advance and retreat. duce similar effects. This condition makes it diffi- Onewaytoassesthisistoevaluatedownstream 16 Rivers in the Landscape

Types of equilibrium Examples

Static • Pool‐ or step‐pool bedforms in cobble‐bed channel • Reach‐scale bed gradient • Channel planform Increasing time scales

Steady state • Average velocity over several minutes • Width/depth ratio of flow over several days • Planform over decades with alternating large and small floods

Dynamic • Progressive change in Dynamic channel width, bed metastable elevation, gradient, etc • Episodic increase in similar

State of system channel parameters

Time Time Figure 1.3 Schematic illustration of different types of equilibrium. Over the shortest time intervals, any particular variable representing state of the river system (e.g., channel planform or gradient) is static and unchanging. At progressively longer time intervals, the variable may be in steady state, with fluctuations about a consistent mean. At the longest time intervals, the mean value of the variable is likely to change, either progressively through time in dynamic equilibrium or in a stepped manner that reflects the crossing of thresholds, as in dynamic metastable equilibrium. The latter two cases are not, strictly speaking, equilibrium because the system exhibits net change over the time span being considered. The phrases are, however, widely used in this manner.

hydraulic geometry (Leopold and Maddock, 1953) valley glaciers altered discharge downstream from relations for rivers in the Front Range. Downstream the glaciers. hydraulic geometry relations are empirical equa- A river in equilibrium is expected to have tions in the form of power functions derived from well-developed downstream hydraulic geometry linear regressions of log-transformed data. These such that variations in discharge explain most equations relate dependent variables of channel of the observed downstream pattern of variation geometry to the independent variable of discharge. in width (Wohl, 2004b). Headwater rivers within For example, theglaciatedportionoftheFrontRangeexhibit less well-developed downstream hydraulic geome- w = aQb (1.6) try relations, as indicated by lower values of the regression coefficient for w-Q regressions, than where w is channel width, Q is discharge, and the headwater rivers at lower elevations beyond the coefficient a and exponent b are determined from extent of Pleistocene glaciations. This suggests that the linear regression. riversintheglaciatedzonearestilladjusting,more Equation 1.6 implies that discharge is the primary than 10,000 years after glacial retreat, to local influence on channel width. One implication is that variations in gradient, substrate resistance, sedi- valuesofchannelwidthintheFrontRangehavefluc- ment supply, and other factors that are affected by tuated through time as the advance and retreat of glaciation and that can influence channel width. Chapter 1 Introduction 17

These rivers may be farther from equilibrium than ables other than time remain relatively constant. otherwise analogous channels at lower elevations in Forexample,drainagenetworksdevelopedonother- the mountain range. wise comparable basalt flows of widely differing ages Equilibrium, or its absence, can also be described within a limited region can be compared to exam- in terms of steady-state versus transient landscapes. ine network development through time. The chal- A steady-state landscape can be defined with respect lenge of ergodic approaches is that variables other to denudation and topography as a landscape in than time likely differ between sites being compared. whicherosionandrockupliftarebalancedsuch Even if the basalt flows are identical in composition, that a statistically invariant topography and con- for example, fluctuations in climate through time stant denudation rate are maintained over a spec- might cause the older networks to represent at least ified time interval (Whipple, 2001). A steady-state preliminary development under a climate different landscape thus exhibits equilibrium between uplift than that acting during development of networks on and erosion. Transient landscapes are those expe- younger basalt flows. riencing relatively brief (on a geological timescale) Returning to the example of instream wood in increases in erosion rate in response to, for exam- forested streams of the Front Range, one way to ple, active tectonic uplift (Attal et al., 2008). Ongoing investigate the importance of forest stand age on change indicates that a transient landscape has not river–wood dynamics is to compare otherwise anal- yet reached equilibrium following an external per- ogousstreamsegmentsflowingthroughforestsof turbation. diverse age. Study design can be challenging: ide- Exhumation of the Denver Basin at the eastern ally, all other important parameters—drainage area, margin of the Colorado Front Range within the last streamflow,sedimentsupply,valleygeometry,and few million years caused relative base level fall for so forth—are similar between the stream segments, the major rivers of the Front Range. Base level fall and only the age of the riparian forest varies. Com- triggered a wave of incision that has been migrat- parisons using this ergodic approach suggest that ing upstream at an estimated rate of 0.15 mm/year old-growth forests with average tree age greater than (Anderson et al., 2006c). The location of contem- 200 years have more instream wood, larger wood porary active response to base level fall appears as pieces, more closely spaced channel-spanning log- a steepening—either a waterfall or steep section of jams, and consequently greater abundance of sec- —inthelongitudinalprofileofeachriver.Por- ondary channels and greater channel–floodplain tions of the river network upstream and downstream connectivity (Beckman, 2013). from this steeper zone are presently in steady state Stepping back to consider the river networks of with respect to the base level fall, whereas the gradi- theFrontRangeataregionalscale,manyoftheques- ent of the steeper portion of the longitudinal profile tions posed by Leopold et al. (1964) remain highly is transient. relevant today: Contrasting river process and form between dif- ferent portions of a region such as the Front Range underlies another approach to understanding rivers. r What factors control hillslope process and form, Data for understanding rivers can be obtained and the initiation of channelized flow? from direct measurements in a field setting or r What is the rate of bedrock and from remote sensing imagery (Oguchi et al., 2013). regolith production? Because of the long temporal scales over which river r How and how rapidly does regolith move processessuchasdevelopmentofdrainagenetworks downslope into channel networks? or longitudinal profiles act, ergodic reasoning is also r What variables influence the location of channel commonly used. Ergodic reasoning substitutes space heads? for time by comparing features in different stages r What processes result in the formation of chan- of development, under the assumption that vari- nel heads? 18 Rivers in the Landscape r What factors govern the longitudinal profile of the r What river-related geomorphic processes can be rivers? quantified in a manner applicable to diverse land- r In particular, what is the relative importance of scapes? landscape-scale denudation in response to con- r In an influential paper, Dietrich et al. (2003) tinuing adjustment to uplift, Pleistocene glacia- highlighted the importance of developing geo- tions, and Quaternary relative base level fall? morphic transport laws (GTLs) in the form of r What is the relative importance of longitudinal mathematical statements derived from a physi- variations in bedrock erosional resistance, sedi- cal principle or mechanism, which express the ment supply, and flow regime? mass flux or erosion caused by one or more processes. In order to be useful, it is important Examining rivers in the context of the greater that such laws can be parameterized from field landscape, we can also add a series of new questions. measurements, can be tested in physical mod- Examples include: els, and can be applied over relevant spatial and temporal scales. Existing GTLs include those r How do diverse types of connectivity vary for soil production from bedrock, linear slope- throughout these river networks? dependent transport of colluvium, and debris r Are the alpine summit surfaces storing flow and river incision into bedrock. periglacialsediment,forexample,orarethey r What fundamental processes in the Colorado stronglycoupledtoadjacentglaciatedvalleys? FrontRangecanusefullybeexpressedviaGTLs r Channel–floodplain and channel–hyporheic- (Figure 1.4)? groundwater connectivity increase within lower r What processes are not yet adequately described gradient, wider valley segments, and then by such laws? decrease in steep, narrow segments. What are r How can we integrate GTLs with quantitative the specific processes governing these down- measures of connectivity? stream variations in connectivity? r What components of river process and form are r Whatarethemagnitudeandextentofhuman significantly influenced by biota? alteration of river networks? r Beaver were much more abundant in the r When people of European descent settled the Colorado Front Range prior to the nineteenth Front Range during the nineteenth century, century. Have historical reductions in beaver theyinitiatedlodeandplacermining,extensive populations and beaver dams influenced rivers deforestation, and widespread flow regulation in regionally, or only local segments of rivers? theformofdamsanddiversions.Someofthese r How do instream wood volume and associ- activities ceased a century ago. Do river process ated geomorphic effects differ between sub- and form still differ between networks in which alpine and montane forests, or between steep, these historical activities occurred and networks narrow valley segments and wide, lower gradi- that were not altered in this way? ent valley segments? r How does contemporary flow regulation alter r The extent and species diversity of riparian physical and ecological functions of Front Range vegetation differ markedly between steep, nar- rivers? rowvalleysegmentsandwide,lowergradi- r Warming climate is resulting in changes in ent segments. How do these differences influ- precipitation, soil moisture, wildfire regime, ence valley-bottom sediment storage, hyporheic outbreaks of native insects that kill trees, and and groundwater exchange, and water chem- forest blowdowns. How do these pervasive alter- istry along Front Range rivers? ations of forest dynamics and precipitation- runoff-stream flow influence channel process To quote Leopold et al. (1964, p. 18): “Partial and form? explanations of these problems can be offered, but Chapter 1 Introduction 19

Precipitation and dry deposition onto uplands 1

Water and solutes moving downslope surface and subsurface 3 6 Water and dissolved, suspended and bed load moving downstream Sediment moving downslope 4 diffusion, mass movements

2 Wood and finer 5 Bedrock weathering to organic matter sediment and solutes moving into rivers

7 Debris flows and rivers incising into bedrock Bedrock

Figure 1.4 Schematic illustration of fundamental fluxes in any landscape. Among those for which some formof geomorphic transport law (GTL) has been proposed are 2 (bedrock weathering), 4 (downslope movement of sediment via diffusion), and 7 (incision into bedrock), although these GTLs require additional testing and parameterization for specific field settings. Empirical or theoretical equations have also been proposed for 1 (precipitation anddrydepo- sition), 3 (downslope movement of water and solutes), and 6 (water and sediment movement downstream). Again, these equations require testing and parameterization. All of the equations proposed for these fluxes assume that average values can be quantified based on prevailing conditions. Using bedrock weathering as an example, chemical reactions are a function of factors such as temperature and the amplitude of temperature oscillations (Supplemental Section 2.1.2). Explicitly incorporating connectivity requires quantifying variations in prevailing conditions through time that limit or enhance fluxes, such as short-term variations in weather and longer-term variations in climate that influence temperature and thus chemical weathering rate. more complete explanations require much more classes within a society, but this phrase is also partic- knowledge of processes than is presently available.” ularly apt for understanding rivers. If we can extend our understanding sufficiently and

1.6 Only connect r connect rivers to landscapes r connectcontemporaryriverconfigurationto E.M. Forster took “only connect” as the epigraph for human and geological history his novel “Howard’s End.” Forster referred to con- r connect site-specific river characteristics to uni- nections between individual people and different versal river process and form 20 Rivers in the Landscape r connect field observations to physical experi- r social science knowledge of human history and ments and numerical simulations, and decisions regarding resource use, and r connect geomorphic knowledge of river process r biogeochemical knowledge of aqueous and form to chemistry r ecologicalknowledgeofaquaticandriparian organisms, … then we will be making progress in understand- r geological knowledge of rock substrates and ing the complex and fascinating world of rivers tectonics, (Figure S1.8). Chapter 2 Creating channels and channel networks

2.1 Generating water, The characteristics of precipitation strongly influence the amount of water reaching channel solutes, and sediment networks and the pathways that water follows downslope into channels. Most river networks 2.1.1 Generating water are influenced by persistent types of precipitation, such as predominantly convective storms, although Rain is the most common form of precipitation moderate-tolarge-sizednetworkscommonly (Barry and Chorley, 1987). Precipitation includes all experience multiple forms of precipitation, such liquidandfrozenformsofwater—rain,snow,hail, as winter snowfall and summer convective storms. dew, hoar frost, fog drip, and rime—moving from Supplemental Section 2.2.1 describes human influ- theatmospheretowardthegroundsurface,butrain ences on precipitation in the context of warming and snow constitute by far the greatest contributors climates. to precipitation worldwide. Rainfall at a measure- Precipitation reaching Earth’s surface can be ment site is typically characterized in terms of max- stored for periods of hours to months in a solid imum or average intensity (amount per unit time), form as snow, or for decades to thousands of years duration, total amount, spatial extent of a rainstorm, as glacial ice. Melting of this solid, stored water and frequency of rainstorms. strongly influences rivers at high altitudes and lati- The major types of precipitation can be differ- tudes.Therelativeimportanceofglaciermelt,snow entiated into convective, cyclonic, and orographic, melt, and rainfall typically vary by latitude and based on the primary mode of uplift of the air that location with respect to atmospheric circulation causes water vapor to condense into water droplets patterns and, within a region, by elevation. Precipita- or ice crystals. Supplemental Section 2.1 describes tion and runoff patterns within the Himalayan mas- the characteristics of each type of precipitation, as sif exemplify the effect of elevation. Up to 60%–80% well as global patterns of precipitation. of river discharge originates as monsoon rainfall at

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

21 22 Rivers in the Landscape lower elevations on the southern side of the mas- Attheregionalscale,lithology,tectonics,and sif, whereas glacier and snow melt contribute 50%– climate influence processes and rates of bedrock 70% of river discharge at higher elevations and on weathering. Minerals that crystallize from molten the northern side of the massif (Gerrard, 1990; Wohl material at high temperatures are typically less resis- and Cenderelli, 1998). tant to weathering than minerals such as quartz that Worldwide, glacier and snow melt are progres- crystallize at lower temperatures, so the mineralog- sively more important at higher latitudes and higher ical composition of bedrock influences weathering elevations, and the seasonal melt contribution is (Anderson and Anderson, 2010; Ritter et al., 2011). delayed later into the summer with increasing lati- Tectonic stresses and regional-scale deformation can tude and elevation. As the proportion of the basin fracturebedrockatvariousdepthsinthecrust, area covered by ice and snow increases, progres- increasing surface area and making the rock more sively more of the total runoff occurs during sum- susceptible to chemical alteration and to other phys- mer (Chen and Ohmura, 1990; Collins and Taylor, icalweatheringprocessessuchasfreeze–thawcycles 1990). Interannual variation in runoff also declines (Anderson and Anderson, 2010). Chemical weath- with greater snow and ice coverage, although the ering results in chemical alteration of the regolith, relation is not linear (Collins, 2006b). Human influ- and in dissolution of more reactive minerals, which ences on glacier and snow melt in the context of release ions that are transported into ground and warming climates are discussed in Supplemental surface waters as solutes. Chemical weathering is Section 2.2.1. facilitated by warm, wet conditions, whereas physi- cal weathering is greatest in moderately wet climates with low temperatures that promote frost action (Ritter et al., 2011). 2.1.2 Generating sediment and At smaller spatial scales, factors such as soil solutes erosional flux, hillslope morphology, and biota influence rates of weathering. Bedrock must Thebedrockunderlyingadrainagebasinisthestart- weather at a rate equal to or greater than the rate of ingpointformuchofthesedimentandsolutes erosionifsoilistopersist,andmanysoilprofiles that eventually move downslope and into rivers, appear to reach a steady-state thickness such that although both particulate and dissolved materials soil production is balanced by removal (Lebedeva canalsoenterriversviaeoliantransportandwet et al., 2010). Chemical weathering rates, soil pro- and dry atmospheric deposition (e.g., Schuster et al., ductionrates,andhillslopecurvaturedecrease 2002; Galloway et al., 2004). An idealized vertical with increasing soil depth (Heimsath et al., 1997; profile through the weathering zone (Figure S2.1) Burke et al., 2007). Consequently, any process has unweathered bedrock at the base, overlain by that alters soil depth by moving soil downslope— regolith, defined as rock that is weathered to any mass movements, creep, bioturbation—also degree. Regolith is subdivided into weathered rock alters rates of bedrock weathering and soil that is fractured and/or chemically weathered but production. Humans, in particular, move tremen- has not been mobilized by hillslope processes or dous amounts of sediment globally (Hooke, 2000), bioturbation, overlain by saprolite,whichretains resulting in long-term and spatially extensive the original rock structure yet has been sufficiently changes in bedrock weathering and soil distribution altered that it can be augured through or dug with and development (Montgomery, 2007). Hillslope ashovel,andmobile regolith (Anderson and Ander- morphology is closely connected to erosional fluxes. son, 2010). Mobile regolith includes soil,whichis Upper, convex portions of a hillslope are likely to organized into horizons by soil-forming processes. have steady removal of weathered products, for In practice, mobile regolith and soil are commonly example, whereas lower, concave portions of the used interchangeably, as in this chapter. hillslope can accumulate weathered materials. At Chapter 2 Creating channels and channel networks 23 largerscales,hillslopecurvature(andsoilproduc- ways followed by the water entering a river exert a tion) varies inversely with soil depth (Heimsath strong influence on the volume and timing of flow in et al., 1997). Plants and animals influence rates thechannel.Precipitationfallingtowardtheground of weathering by excreting organic acids, and does not necessarily reach a river, however. physically disrupting weathered materials, as when Precipitation can be intercepted by plants and plant roots expand into bedrock joints or burrowing either directly evaporate from the plants or be taken animals churn the regolith. Supplemental Section up by the plant tissues and then released back to 2.1.2 provides more details on soil production. theatmospherethroughtranspiration.Evaporation As noted in the first chapter, sediment produc- and transpiration are strongly influenced by energy tion is rarely uniform across even a small catchment availability at the surface and water availability in the as a result of differences in lithology, climate, tec- subsurface. Transpiration also reflects plant physi- tonics, biotic communities, hillslope morphology, or ology (Kramer and Boyer, 1995) and the ability of land use. Investigating the tectonically active, semi- plant roots to take up, redistribute, and even selec- arid Western Transverse Ranges of California, USA, tively extract water from different subsurface depths Warrick and Mertes (2009) found that about half of (Lai and Katul, 2000). Recent estimates suggest that the suspended sediment discharge from the region transpiration represents the largest water flux from came from a very small proportion of the land- continents (i.e., larger than rivers), composing 80%– scape underlain by weakly consolidated bedrock and 90% of combined evaporation and transpiration— associated with the highest rates of tectonic uplift. commonly known as evapotranspiration (Jasechko Similarly, disproportionate sediment generation has et al., 2013). been demonstrated for the Amazon and Mississippi Interception losses from evaporation and tran- River basins (Meade et al., 1990; Meade, 2007). Mil- spiration can reach 10%–20% beneath grasses liman and Syvitski (1992) estimate that mountain and crops and up to 50% beneath forests (Selby, andhighmountain(>1000 m elevation) drainages 1982). Interception can vary substantially during the collectively cover 70% of the global land area, but course of the year in regions with seasons during contribute 96% of total river sediment yields. Anal- which plants go dormant or during which soil mois- ogouspatternsalsoappeartogovernsoluteproduc- ture declines and plants store more water (Link et al., tion. Lyons et al. (2002) estimate that high-standing 2005). Plants can also directly intercept cloud water. ocean islands in the southwest Pacific, for exam- In cloud forests, this interception can reach 35% of ple, make up only about 3% of Earth’s landmass, but mean annual precipitation (Bruijnzeel, 2005). Plant contribute 17%–35% of particulate organic carbon stems can concentrate the movement of precipita- entering the world’s oceans annually. tion toward the ground via stemflow,whichhasbeen described for environments as diverse as tropical montane forest, semiarid loess terrain, and a sea- 2.2 Getting water, solutes, sonal cloud forest in which stemflow accounted for 30% of total precipitation (Hildebrandt et al., 2007). and sediment downslope to Precipitation that does reach the ground sur- face can be evaporated from bodies of standing or channels flowing surface water, or from the ground surface. The pathways taken by precipitation reaching the 2.2.1 Downslope pathways ground surface are strongly dependent on the soil of water cover. Soil cover changes continuously in response to water movement and processes of weathering, bio- The great majority of water flowing in a river passes turbation, sediment transport, and land use or other over or through an adjacent upland and its regolith changes in land cover (Brooks, 2003). Development before reaching a channel (Kirkby, 1988). The path- of crusts, compaction, and sealing (for example, with JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

24 Rivers in the Landscape

clay particles) decrease surface permeability (Slat- is most common where vegetation is sparse, slope tery and Bryan, 1994; Kampf and Mirus, 2013), gradients are steep, regolith is thin or of low perme- as does development of soil water repellency, or ability (e.g., clay), and precipitation intensities are hydrophobicity, from burnt plant litter (Martin and high: characteristics of arid and semiarid regions. Moody, 2001). Precipitation is equally dynamic, and The great majority of most natural watersheds do theintensityandvolumeofprecipitationreaching not produce Hortonian overland flow, but human the land surface vary substantially through time and land uses that compact the soil or cause erosion of space. Consequently, downslope movement of water permeable, near-surface soil layers promote Horto- on the surface and in the subsurface is rarely in nian overland flow. The most extreme example of a steady-state condition (Wainwright and Parsons, this effect is paved, impermeable urban surfaces. 2002). Overland flow can also move downslope as satu- Water can flow down slope at the surface as over- ration overland flow,alsoknownassaturation excess land flow, which is used to describe surface flow overland flow (Dunne and Black, 1970a, 1970b). outside the confines of a channel. Overland flow This is a combination of direct precipitation onto includes Hortonian overland flow,alsoknownas saturated areas and return flow from the subsurface infiltration excess overland flow (Horton, 1945), if as saturation occurs. The moisture content of the infiltration capacity is low relative to precipita- the regolith before, during, and after precipitation tion intensity (Figure 2.1). Hortonian overland flow exerts a particularly important control on saturation

Precipitation (rainfall, clouds, fog, snow) Infiltration (2–500 mm/h) Hortonian overland flow Evapotranspiration (up to 50% for forests) (50 Transpiration Ground – wetted canopy evaporation surface 500 mm/h)

Saturation overland flow

(convergent zones)

Vadose zone

Soil evaporation

Throughflow* Evaporation (diffuse or conduit)

Water table

Phreatic zone Stream Ground water (> 1 × 10–8 mm/h)

* Pipe flow 50–500 mm/h; diffuse flow 0.005–0.3 mm/h Figure 2.1 Schematic illustration of different types and rates of downslope water movement. Downward arrows reflect movement of water toward (precipitation), into (infiltration), across (overland flow) or beneath (throughflow, groundwater) Earth’s surface. Upward arrows reflect water movement back into the atmosphere. JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 25

overland flow, which is influenced more by mote zones that saturate readily (e.g., layered basalt antecedent soil moisture and subsurface transmis- flows) (Mirus et al., 2007), exclusion from frozen sivity than by slope steepness (Montgomery and soil,orsnowmeltoversaturatedsoil(Kampfand Dietrich, 2002). As prolonged precipitation allows Mirus, 2013). Saturation overland flow can also deeper and less permeable regolith layers to become occurintropicalrainforestswhereasharpdropin saturated, subsurface flow expands to include areas permeability with depth and high rain inputs lead progressively closer to the surface (Knighton, 1984). to perched water tables (Bonnell and Gilmour, 1978; Conditions that favor saturation overland flow Elsenbeer and Vertessy, 2000). include high permeability near the surface, a humid Overland flow commences when water ponds on climate with high cumulative water input, and the land surface to sufficient depth to begin flow- gentler slopes with shallow soils that cannot drain ing downslope. Surface roughness from microto- as easily as steep slopes in which hydraulic gradients pography and vegetation generate flow resistance approximately parallel the steep surface topography that influences overland flow pathways (Bergkamp, (Kampf and Mirus, 2013) (Figure 2.2). 1998), particularly in humid environments with Saturation overland flow is rare outside of con- dense vegetation and organic litter, but the flow also vergent flow zones such as hillslope concavities modifies surface roughness. Vegetation also influ- (Dietrich et al., 1992). Several scenarios can cause ences overland flow by altering infiltration. Shrub saturation flow at other points along a hillslope, mounds in semiarid regions facilitate infiltration however, including: topographic breaks, permeabil- and generate less overland flow than intervening ity contrasts (e.g., roads or pavement), low subsur- spaces, creating downslope pathways strongly cou- face storage capacity, geologic structures that pro- pled to vegetation patterns (Dunne et al., 1991). Reported infiltration capacities from diverse locations range from 0 to 2500 mm/h (Selby, 1982). Infiltration can be extremely variable through space Subsurface permeability Low high and time within a single small catchment because Low of the numerous factors that influence infiltration Subsurface Steep stormflow (Figure 2.3). The variable source area concept (Hewlett and Hibbert, 1967) reflects the fact that Flow convergence * the area of a catchment actually contributing water Infiltration Humid forest to a channel extends during precipitation and con- excess tracts after the precipitation ends, with contributing overland flow * Temperate grassland area varying between 5% and 80% of the catchment (Dunne and Black, 1970b; Selby, 1982). Much of this Surface topography variability likely reflects thresholds of activation for Saturation lateral subsurface flow (McDonnell, 2003). * excess Arid shrubland overland flow High Saturation overland flow typically occurs first

Flat in downslope portions of a catchment and then Arid Humid Climate expands upslope (Dunne, 1978). In cold regions with low topographic relief, upslope areas can be Figure 2.2 Environmental controls on dominant runoff generation mechanisms. Runoff generation mechanisms the first sources of runoff as a result of varied local represented by patterned fields: overlapping fields indi- topography and water table position in relation to cate that no single runoff mechanism dominates hydro- frozen ground (Spence and Woo, 2003). Spence logic response. Asterisks represent approximate concep- and Woo (2006) use the element threshold concept tualizations, for example environments (from Kampf and to describe a catchment in which runoff is gov- Mirus, 2013, Figure 10). erned by the function of spatially and hydrologically JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

26 Rivers in the Landscape

ity,oreaseofwaterflowthroughthemedium,isat its maximum value (Kampf and Mirus, 2013). When Precipitation intensity Surface porosityand duration the void space is not completely filled with water, the and permeability connectivity of pore space and the hydraulic con- ductivity decrease. Hillslopes are typically unsaturated at the ground Antecedent moisture surface and water movement is predominantly ver- Soil ice tical during infiltration. Moisture is redistributed Soil vertical and lateral conductivity Water retention in the subsurface through both vertical and lateral movements that reflect the interaction of infiltra- tion and percolation (wetting) with evaporation and Surface porosity and permeability are a function of hillslope gradient transpiration (drying) (Kampf and Mirus, 2013). vegetation Watermovingintothesoilcanpropagateintheform regolith grain size, compaction, depth and areal extent of diffuse wetting fronts, fingered flow paths that Figure 2.3 Schematic illustration of the variables result from wetting front instabilities, or preferential influencing infiltration capacity. Line represents a side flow along conduits (Wang et al., 2003). Although view of a sloping ground surface. Precipitation character- lateral unsaturated flow can occur, lateral flow is istics listed at upper right, surface characteristics listed most common under saturated conditions (Kampf along the line, and subsurface characteristics listed at and Mirus, 2013). lower left all help to create changes in infiltration capac- Diffuse throughflow depends on the general ity over small spatial and temporal scales. porosity and permeability of the unsaturated zone, whereas concentrated throughflow moves in pref- erential flow paths such as pipes or macropores distinct areas (elements), such as bedrock uplands, (Dunne, 1980; Jones, 1981). Macropores are open- soil-filled valleys, and lakes, and the hydrologic link- ings sufficiently large that capillary forces have an ages between elements. Each element can store, con- insignificant effect on the water running through tribute, or transmit water, and the occurrence of the pores (Germann, 1990). Preferential flow within thesefunctionsreflectsthewaterbalanceofthe the macropore does not have time to equilibrate element as well as connections to adjacent ele- with slower flow through the surrounding matrix ments. The primary difference between the variable (Simˇ unek˚ et al., 2003) and is likely to be turbulent source area and element threshold concepts involves and thus not adequately described by Darcy’s law. the assumed greater hydrologic connectivity within Lateral flow through macropores requires pore con- catchments operating under the variable source area nectivity and a rate of water supply that exceeds concept. thelossratetothesurroundingsoil,whichsug- Infiltrating water that remains in the subsurface gests that macropore flow is triggered at a threshold can flow downslope in the vadose (unsaturated) wetness level (Beven and Germann, 1982) (Figure zone above the water table as throughflow,orinthe S2.2). Once this threshold is exceeded, flow is much phreatic (saturated) zone below the water table as faster than diffuse, matrix flow. Rapid subsurface groundwater. In either case, subsurface water flows stormflow is particularly widespread on densely veg- through interconnected voids between solid materi- etated hillslopes in steep terrain where dense biolog- als. Where the interconnected pores are small, flow ical activity creates high concentrations of macro- through the porous medium is laminar and of low pores, although the phenomenon has also been velocity (Kampf and Mirus, 2013) and is commonly documented in semiarid climates (Kampf and described using Darcy’s law (Darcy, 1856). When the Mirus, 2013). void space in a porous medium is filled with water Substantial preferential flow through macropores under saturated conditions, the hydraulic conductiv- can facilitate the formation of soil pipes. Soil pipes JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 27

are larger than macropores and are typically formed mountainous regions can respond to seasonal by subsurface erosion that enlarges animal burrows, processes such as snow melt runoff (McDonnell root channels, or cracks from desiccation or unload- et al., 1998; Clow et al., 2003). Deep infiltration into ing (Bryan and Jones, 1997). Pipes can be only a few bedrock can be important where bedrock is close to centimeters in length and diameter, or 2 m in diame- the surface or where highly porous and permeable ter and hundreds of meters long (Selby, 1982). Pipes or highly fractured bedrock can take up substantial tend to form just above a zone of lower porosity and volumes of water (Flint et al., 2008) (Figure S2.4). permeability or along a cavity created by a burrow- This type of deep subsurface flow can be reflected in ing animal or the decay of plant roots. Piping can slow or doubly peaked runoff response (Onda et al., occur in any region, but is particularly associated 2001), in which the two peaks reflect unsaturated withdrylands.Insomecatchments,pipingcancon- andsaturatedzoneflow. tribute up to nearly 50% of stormwater flow (Jones, Subsurface flow typically dominates slopes with 2010). full vegetative cover and thick regolith (Dunne and Pipe networks can exist at multiple levels in the Black, 1970a). Thicker soils increase the mean res- regolith, with each level being activated by precip- idence time of water on slopes and damp the tem- itation of different magnitude (Gilman and New- poral fluctuations of water movement in response son, 1980; Kim et al., 2004). Rapid lateral flow via to precipitation inputs (Sayama and McDonnell, soil pipes or via discontinuities at the soil–bedrock 2009). Water moving downslope via Hortonian interface appears to depend on thresholds, such that overland flow typically has the most rapid rate of hillslopes “turn on” when sufficient water infiltrates movement (50–500 m/h), with progressively slower (Uchida et al., 2001; McDonnell, 2003). This may rates of movement during saturation overland flow, help to explain the old water paradox in which pre- throughflow, and groundwater flow, which can event water largely dominates storm runoff, sug- move as slowly as 1 × 10−8 m/h (Selby, 1982). The gesting that catchments store water for considerable exception to this generally slower movement from periods of time but then promptly release the water the surface to progressively deeper paths is concen- during storms (Kirchner, 2003). The existence of trated throughflow in pipes or macropores, which thresholds also helps to explain abrupt changes in can move quite rapidly (Figure 2.4). hydrologic response and water delivery to channels Thedistributionofwateramongdifferentdowns- with different hillslope wetness states (Kampf and lope pathways can alter in relation to precipitation Mirus, 2013) (Figure S2.3). magnitude, intensity or duration during a storm, or Downslope water movement below the water on a regular annual basis in strongly seasonal cli- tablecanalsobequitecomplexasaresultofspa- matic regimes. The dominant runoff processes also tial variations in depth and rate of movement of change with spatial scale (McDonnell et al., 2005). groundwater. The water table responds separately in The manner in which moisture is released from riparian and hillslope zones, for example (Seibert pores within soil might condition runoff within a et al., 2003), so that upslope area groundwater can soil column, whereas partitioning between prefer- be falling during the early portion of runoff while the ential and non-preferential flow governed by soil riparian water table is rising. The hillslope ground- structure and rain intensity become more impor- water can be slowly falling as part of the recession tant at the plot scale. Spatial variation in soil depth from rainfall several days earlier. Groundwater close strongly influences lateral water movement in tran- tothestreamismorelikelytobeinphasewith sient subsurface saturated areas at the hillslope scale, runoff. and the proportion of water derived from differ- Other sources of complexity in groundwater ent catchment geomorphic units such as hillslopes, dynamics include regional groundwater flow riparian zones, and bedrock outcrops influence the between watersheds (Genereux and Jordan, 2006). river hydrograph at the catchment scale (McDonnell Small groundwater reservoirs such as those in et al., 2005). JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

28 Rivers in the Landscape

1000 of a surface river’s discharge. Precipitation falling on karst terrains can percolate down to the groundwa- 100 Horton overland flow terviadiffuseinfiltrationinthezoneofaeration,a process known as vadose seepage.Precipitationcan Saturation overland flow 10 Ephemeral also move downward via a highly permeable zone— pipe flow Perennial pipe flow typically formed by the intersection of vertical joints 1 in the bedrock or cylindrical solution openings—

Peak runoff rate (mm/h) runoff Peak Throughflow at the base of a closed depression, a process known 0.1 as vadose flow or internal runoff (Ritter et al., 0.000010.0001 0.001 0.01 0.1 1 10 100 2011). 2 Drainage basin area (km ) The epikarstic zone within the vadose zone is Figure 2.4 Generalized graph of hillslope hydrologic a heterogeneous interface between unconsolidated processes in relation to size of contributing area (from material and solutionally altered carbonate rock Jones, 2010, Figure 7b). The drainage basin area in (Jones et al., 2003). The epikarstic zone is par- this figure reflects total contributing area of the chan- tiallysaturatedandcandelayorstoreandlocally nel upstream from the point at which runoff enters a channel, but does not necessarily reflect the continuous reroute vertically infiltrating water into the deeper, extent of the area generating runoff. A 100 km2 drainage regional phreatic zone of the underlying karstic basin area, for example, does not mean that Horton over- aquifer.Aquifersformedinkarstterrainsdiffer land flow has to persist for tens of km before entering a from other types of aquifers in that karstic aquifers channel. typically have larger variations in porosity and permeability. Groundwater in these aquifers can move through intergranular pores within unfrac- Hillslopeflowpathscanbeconceptualizedas tured bedrock (matrix permeability), through joints occurring along a spectrum from predominantly and bedding planes created in the rock after depo- vertical to predominantly lateral (Elsenbeer, 2001). sition and lithification (fracture permeability), or Lateral flow, in particular, is highly nonlinear and throughconduitswithwidthsexceeding1cmthat exhibits threshold behavior that influences downs- have been enlarged by solutional weathering (con- lope connectivity as small depressions in the under- duit permeability) (White, 1999). lying bedrock topography fill and spill or lat- The extent of frozen soil strongly influences erally discontinuous soil pipes and macropores downslope pathways of water in cold regions. In self-organize into larger preferential flow systems as very cold regions, an active layer that thaws season- sites become wetter (Sidle et al., 2001; Hopp and allyandrangesfrom15cmto5mthickoverlies McDonnell, 2009). permanently frozen ground, or permafrost.Frozen Karst terrains, cold regions underlain by perma- soil impedes infiltration and limits percolation, with nently frozen ground, and the humid tropics form theresultthatalargeportionofsnowmeltmoves three distinctive subsets in terms of regional pat- downslope as overland flow and is quickly delivered terns of downslope pathways of water. Karst terrains to channels (Vandenberghe and Woo, 2002). As the have distinctive landforms and drainage as a result active layer thaws, the depth and importance of infil- of greater rock solubility in the presence of natu- tration can change, but the presence of permafrost ral water; typically, these terrains are associated with fundamentally limits deep infiltration and ground- carbonate rocks and evaporites. Water flowing at the water flow. Retreat of permafrost in response to surface in karst terrains over rocks of low solubility global warming is creating major changes in downs- can move very abruptly downward to the groundwa- lope pathways of flow in cold regions. ter via swallow holes,whichareopencavitiesonthe Many catchments in the humid tropics have high channel floor where the channel flows onto carbon- runoff coefficients and quick hydrologic responses aterocks.Swallowholescandivertaportionorall to precipitation. Infiltration rates in the humid JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 29

tropics range from 0 to over 200 mm/h (Harden in seasonal variations in the magnitude of rainfall- and Scruggs, 2003), however, and runoff genera- induced runoff (Collins, 1998). tion reflects numerous influences, as in temperate Supplemental Section 2.2.1 discusses measuring environments (Scatena and Gupta, 2013). A vari- andmodelingthedownslopepathwaysofwater,and ety of shallow subsurface flow paths contribute to describes human influences on downslope move- the typically flashy response and high discharge per ment of water. unit drainage area common in humid tropical catch- Although knowledge of downslope pathways of ments. Among these shallow subsurface flow paths water has increased substantially during the past few are tension cracks that develop during the dry season decades, the fundamental questions remain: where and abundant macropores (Niedzialek and Ogden, doeswatergowhenitrains?Whatflowpathdoesit 2005). take to streams? And, how long does water reside in The downslope pathways of meltwater through the catchment? Knowledge of downslope pathways snow and glacial ice are as complex as those through of water is critical to fluvial geomorphology because slopes of sediment and bedrock. Outflow from a theresultingstreamflowistheprimarydrivingforce glacier includes a base flow component supplied by in river networks. groundwater discharge, runoff from storage zones withintheice,runofffromthefirnwateraquiferat theglacier’ssurface,andregulardrainagefromlakes (Gerrard, 1990). The melting of seasonal snow cover 2.2.2 Downslope movement on the glacier and on non-glacial surfaces can pro- of sediment duce an initial runoff peak, with subsequent melt- ingoftheglacialiceproducinganotherrunoffpeak Sediment moves downslope through mass move- (Aizen et al., 1995). Runoff from the glacier can also mentsasslides,flows,orheave,andthroughgrad- vary during the melt season as channels develop on ualdiffusiveprocessesofrainsplashandoverland top, within, and beneath the glacier (Fenn, 1987; flow. Diffusive processes typically involve individ- Nienow et al., 1998) and as air temperatures fluctu- ualgrainsratherthanaggregatesofgrains.What- ate (Hodgkins et al., 2009). ever the transport mechanism, the portion of the soil Runoff from snowpack depends on percolation profile actively transported downslope in steeper, timesfromthemeltatthesurfaceofthesnowpack forested terrains can reflect the rooting depth and to the ground, as well as distance downslope from consequent root wad thickness of fallen trees, which the melt source. Travel times through the snow- in turn reflects depth to the soil/saprolite boundary pack dominate runoff in watersheds < 30 km2 in (Jungers et al., 2009). the Sierra Nevada of California, USA, for exam- Mass movements involve downslope transport of ple, whereas snowpack heterogeneity results in more aggregates rather than individual particles and are consistent timing of peak runoff in watersheds > 200 typically strongly seasonal as a function of moisture km2 (Lundquist et al., 2005). availability and freeze–thaw processes (Hales and Ice- and snow-covered portions of a catchment Roering, 2009). Mass movements are categorized can sometimes retain meltwater and rainfall to a based on the characteristics of the moving mass into greater extent than snow-free portions of the catch- slides, flows, and heave, each of which results from ment. The ability of the snowpack to retain or trans- a decrease in the shear strength of the material or mit rainfall partly depends on its structure because an increase in the shear stress acting on the material thepresenceoficelayerscansubstantiallyincrease (Carson and Kirkby, 1972) (Figure S2.5). water retention compared to a homogeneous snow- A slide occurs when a mass of unconsolidated pack (Singh et al., 1998). The proportion of a basin material moves without internal deformation along covered by ice and snow typically varies through the a discrete failure plane that can be curved and pro- seasons as the transient snow line shifts, resulting duce the rotational movement of a slump, or can JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

30 Rivers in the Landscape

be relatively straight. Slides typically result from a liding in patterns influenced by groundwater seep- decrease in the shear strength of the soil as a result of age, and entrenchment by river incision of land- weathering, increased water content, seismic vibra- slide scars and deposits. Large mass movements tions, freezing and thawing, or human alterations can cause substantial sediment accumulation in val- such as deforestation and road construction. An ley floors, interrupting progressive channel incision increase in shear stress caused by additions of mass into bedrock, as documented along the Indus River or removal of lateral or underlying support can also in Pakistan by Burbank et al. (1996). The down- trigger a slide. Slides typically transition downslope stream end of mass movement deposition in a valley into flows, which occur when the moving mass is bottom can create a knickpoint, or steeper segment, sufficiently liquefied or vibrated to create substantial alongariver’scourse.Landslidedamscanpondriver internal deformation during downslope movement. waterupstream,butaretypicallybreachedwithina Mass movements can also begin as flows initi- few days (Costa and Schuster, 1988), resulting in an ated by runoff-dominated processes that cause pro- outburst flood with a peak discharge and sediment gressive sediment entrainment or by infiltration- transportcapacitythatgreatlyexceedsprecipitation- dominated processes that trigger discrete failures generated floods along the river (Cenderelli and (Cannon et al., 2001). Mass movements resulting Wohl, 2001, 2003). Small catchments can form in the from runoff occur when sheetwash and rills entrain detachment area of a mass movement. Large land- progressivelymoresedimentdownslopeuntilthe slidescanmovesomuchmassthattheyreducelocal runoff concentrates in gullies and channels, eroding elevation, cause drainage divides to shift location additional sediment from these conduits. This suc- abruptly, or truncate headwater streams and cause cession of events can occur during intense rainfall in stream piracy. Landslides can even cause drainage arid environments, or after a wildfire that removes reversal. Mass movements also strongly influence surface vegetation and plant litter, as described by spatial and temporal variations in sediment deliv- Wohl and Pearthree (1991) for burned sites in Ari- ery to rivers in steep terrains (Korup et al., 2004, zona, USA, (Figure S2.6) and Meyer and Wells 2010). Mass movements can also occur in low-relief (1997) for burned sites in Wyoming, USA. Infiltra- areas, particularly where an impermeable layer such tion can initiate local slope failures in the form of a as frozen ground or shale promotes saturation of slide that entrains more material as it moves downs- overlying materials. lope. This can occur during intense precipitation or Creep occurs when particles displaced by bio- rapid melting of a snowpack, or following a wildfire. turbation and in wetting–drying or freeze–thaw Mass movements can also result from what Johnson cycles move downslope under the influence of grav- and Rodine (1984) termed the “firehose effect,” in ity (Figure S2.7) (Kirkby, 1967). Bioturbation via whichconcentratedpeakdischargeattheoutletofa tree throw (Figure S2.8a) and burrowing by soil- fanorthebaseofasteep,bedrockchutecanmobilize dwelling rodents (Figure S2.8b), in particular, can accumulated sediment. displace substantial amounts of sediment (Heimsath Abrupt mass movements recur frequently— et al., 1997; Roering et al., 2002; Gallaway et al., almost annually—in many high-relief terrains, and 2009). Creep in cold climates is strongly influenced transport the majority of sediment to or along by freeze–thaw processes (Hales and Roering, 2007) low-order stream channels (Jacobson et al., 1993; and, with increasing water or ice content, creep Guthrie and Evans, 2007). Mass movements can grades into solifluction or gelifluction (Figure S2.9), exert diverse influences on valley geometry and the respectively, which involve the very slow downslope spatial arrangement of channels in drainage net- flow of partially saturated regolith. Creep is greatest works (Korup et al., 2010; Korup, 2013). Hovius et al. in the upper meter of the soil and is proportional to (1998) describe three phases of valley development surface gradient (Selby, 1982; McKean et al., 1993). in Papua New Guinea: initial incision of isolated This results in soil flux proportional to the depth- gorges, lateral expansion and branching by lands- slope product (Furbish et al., 2009b). JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 31

Gradual diffusive processes in which individual released from microtopographic mounds into the grains move downslope typically occur via rain- sheet flow, however, so that the sediment supply is splash or overland flow. Rainsplash occurs when sufficient to prevent rill incision on the upper por- rain falling on a surface loosens or detaches indi- tion of the hillslope (Dunne et al., 1995). vidual particles, making the particles more suscep- At some point downslope, surface irregularities tible to entrainment by overland flow (Furbish et al., concentrate overland flow into slight depressions 2009a; Dunne et al., 2010) (Figure S2.10a). Over- that then enlarge as the increasing water depth land flow may be capable of eroding measurable increases the shear stress acting on the substrate at quantities of sediment where unvegetated, unfrozen thebaseoftheflow.Thiscangiverisetorills and slope surfaces are exposed (Dingwall, 1972; Rus- gullies, typically described as parallel channels with tomji and Prosser, 2001). Thread flow occurs when few or no tributaries. Sometimes these names are overland flow goes around individual roughness ele- used interchangeably, but rills can be distinguished ments. Sediment can be stripped evenly from a slope as channels sufficiently small to be smoothed by crest and upper zone during sheet flow that sub- ordinary farm tillage, whereas gullies are sufficiently merges individual roughness elements and forms deep not to be destroyed by ordinary tillage. Rills a fairly continuous sheet of water across the slope and gullies can form effective conduits for sedi- (Figure S2.10b). Erosion by sheet flow is particu- ment erosion down slopes and into river networks larly effective where inter-particle cohesion has been (Sutherland, 1991), and rill-channel networks can reduced by needle ice, trampling, or disturbance to exhibit equilibrium scaling characteristics for bifur- vegetation (Selby, 1982). Microbiotic soil crusts, very cation and channel length ratios similar to those of coarse particles at the surface (Figure S2.10c), or river networks (Raff et al., 2004). Downslope move- vegetation cover can substantially reduce sediment ment of sediment tends to be extremely spatially and detachment and erosion by rainsplash or sheet flow temporally variable as a result of local changes in (Uchida et al., 2000). The presence of seasonally or slope gradient, ground cover, vegetation and micro- permanently frozen soil can enhance overland flow topography (Saynor et al., 1994). Most hillslopes are and soil erosion during the melt season by impeding shaped through time by some combination of mul- infiltration (Ollesch et al., 2006). tiple processes (Jimenez Sanchez, 2002). Just as sub- Horton (1945) proposed the existence of a critical surface water in a hillslope can be “old water” stored

length of slope, Xc,abovewhichnoerosionoccurs for relatively long periods of time between precipi- via sheet flow. The critical length is a function of tation inputs, sediment that begins to move downs- surface resistance and slope gradient. As a simple lope and into a channel network can be stored for 103 approximation, sheet flow erodes particles as a func- years even in mountainous uplands. Investigating tion of the shear stress 𝜏 exerted on the surface by theOregonCoastRange,forexample,Lancasterand the flowing water (𝜏 = 𝛾 HS,where𝛾 is the spe- Casebeer (2007) found that sediment transit times cific weight of water, H is flow depth, and S is hill- along colluvial valley segments dominated by debris slope gradient) relative to the resistance of the sur- flows averaged 440 years, whereas the average tran- face over which the water is flowing. For a given sit time rose to 1220 years within the headwaters flu-

surface resistance and flow depth, Xc will be shorter vial portion of the channel network. In both debris- on steeper slopes up to about 40 degrees. Sheet flow flow and fluvial channels, significant volumes of becomes competent to transport sediment within a sediment remain in storage for thousands of years. few meters of the drainage divide on long hillslopes Consequently, terraces, fans, and some floodplain subject to Hortonian overland flow, but microtopo- deposits near tributary junctions are effectively sedi- graphic mounds (Figure S2.11) generated mostly by ment reservoirs at time spans of 102–103 years (Lan- biotic processes can force the sheetwash to converge caster et al., 2010). and diverge, increasing depth, velocity, and trans- Supplemental Section 2.2.2 includes more infor- port capacity in the converging zones. Sediment is mation on measuring and modeling the downslope JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

32 Rivers in the Landscape

pathways of sediment. Section 8.1.2 examines (Berner and Berner, 1987). These factors include: how human-induced changes in land cover influ- topographic relief, climate, glaciation, snow cover, ence downslope sediment movement by alter- landuse,andepisodiceventssuchasmassmove- ing topography, vegetation, and soil depth and ments or volcanic eruptions. Supplemental Section compaction. 2.2.3 contains further information on controls on uplandwaterchemistry,aswellasasummaryof measuring and modeling the chemistry of water 2.2.3 Processes and patterns of entering channels. water chemistry entering channels Although water chemistry has traditionally been neglected by fluvial geomorphologists, understand- ing influences on the chemistry of water before the The chemistry of precipitation falling over a water enters the channel network and once the water drainage basin varies with distance from the ocean, isinthechannelnetworkisimportantforatleastone with pollution inputs, and through time. The reason: a significant proportion of management of precipitation then reacts with plants, soil, regolith, river process and form is now undertaken in order to and bedrock. The resulting chemistry of water meet legislated water quality standards. River char- entering a river is usually more influenced by the acteristics such as flow depth, hyporheic exchange, hillslope flow paths followed by the water than by substrate grain-size distribution and stability, as well the chemistry of the original precipitation. And, as the characteristics of aquatic and riparian com- just as runoff during a storm can be dominated by munities, can strongly influence the chemistry of oldwaterthathasbeenstoredinthesubsurfacefor water in a river. These characteristics can poten- some period prior to the storm, runoff chemistry tially be managed to reduce or remove contaminants can be dominated by old water (Anderson and or other undesirable traits of water chemistry that Dietrich, 2001). result from the downslope pathways that water fol- Waters entering a river via precipitation falling lows before entering a channel. directlyontheriver,overlandflow,soilwater, and groundwater typically have distinctly different chemistries (McDonnell et al., 1991). The primary constituents of river chemistry (Berner and Berner, 2.2.4 Influence of the riparian 1987; Allan, 1995) are zone on fluxes into channels

r − 2+ 2− the dissolved ions HCO3 ,Ca ,SO4 ,H4SiO4, Riparian is a Latin word designating something “of − + 2+ + Cl ,Na ,Mg ,andK ; or belonging to the bank of a river” (Naiman et al., r dissolvednutrientsNandP; r 2005). The riparian zone is the interface between dissolved organic matter; terrestrial and aquatic ecosystems, and includes r dissolved gases N ,CO ,andO ;and r 2 2 2 sharp gradients of environmental factors, ecolog- trace metals. ical processes, and biotic communities (Gregory et al., 1991). Riparian zones can be difficult to delin- The sum of the concentrations of the dissolved eate because they include components as diverse major ions is known as the total dissolved solids as depressions that create floodplain wetlands and (TDS), and is typically highly temporally and spa- higher-elevation natural levees (Figure 2.5). tially variable in response to factors such as precip- The presence of a riparian zone can strongly itation input and downslope flow path, discharge, influence the characteristics of water, sediment and lithology, the growth cycles of terrestrial vegeta- solutes entering a river. Hillslope waters tend to be tion, and any factor that influences rates and pro- chemicallyandisotopicallydistinctfromriparian cesses of bedrock weathering and soil development zone waters and the degree of expression of hillslope JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 33

Lateral extent of hyporheic zone

Extent of flooding

Mixing of stream, hyporheic and

Vertical extent of hyporheic zone hyporheic of extent Vertical ground water

Figure 2.5 Schematic illustration of the spatial boundaries of a riparian zone. The boundaries of the riparian zone extend: outward to the limits of flooding—typically defined for a relatively frequent recurrence interval of flood, such as 10 years; upward into the canopy of streamside vegetation; and downward into the zone where groundwater, stream water, and hyporheic water mix.

waterintherivercanbeminimalbecauseofchem- to grow successive sets of near-surface roots as the icalreactionsasthewaterpassesthroughtheripar- plant’s base is progressively buried by sedimentation ian zone (Burns et al., 2001; McDonnell, 2003). (Figure S2.12). Subsurface water coming from the riparian zone Numerous studies indicate that riparian zones commonly leads the river hydrograph, with hills- can serve as sinks or buffers for nitrates in agricul- lope input either negligible or dominating the reces- tural watersheds (Mitsch et al., 2001). The capacity sion limb of the river hydrograph after the thresh- of the riparian zone to retain dissolved and particu- old for activating hillslope pathways is exceeded latenutrientssuchasnitrogen,phosphorus,calcium, (McGlynn and McDonnell, 2003). Observations and magnesium is controlled by two factors. The such as these led McDonnell (2003) to suggest that first is the hydrologic characteristics of the riparian a catchment should be conceptualized as a series of environment, including water table depth, water res- reservoirs with coupled unsaturated and saturated idence time, and degree of contact between soil and zones, and explicit dimensions and porosities, which groundwater. The second level of control involves connect laterally and vertically through time and biotic processes, including plant uptake and deni- space in linear and nonlinear ways. trification. Denitrification is microbially facilitated Vegetated riparian zones can effectively trap and nitrate reduction that ultimately produces molecular

store sediment of sand size or finer that is being nitrogen, N2 (Naiman et al., 2005). As an important transported by overland flow or by overbank flows nutrient and component of greenhouse gases, nitro- from a channel (Naiman et al., 2005). Sediment gen has received particular attention, and the com- deposition reflects the greater hydraulic roughness, plexities of nitrogen dynamics illustrate the impor- and consequently lower velocities and transport tance of riparian hydrology and biology (Supple- capacities, of flow passing over and through the mental Section 2.2.4). stems and leaves of the vegetation (Hickin, 1984; Biogeochemical hot spots for dissolved nutri- Griffin et al., 2005). Once the sediment is deposited, ents include anoxic zones beneath riparian envi- plant roots facilitate stabilization and continued ronments because of the microbial communities storage of the sediment (Allmendinger et al., 2005; present in these zones (Lowrance et al., 1984). Ripar- Tal and Paola, 2007). Many riparian species are able ian communities in wet temperate and tropical JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

34 Rivers in the Landscape

environments can mediate nitrate fluxes into rivers persists sufficiently to produce a discrete channel. because denitrification and plant uptake remove The upstream boundary of concentrated water flow allochthonous nitrate (nitrate that originated else- and sediment transport between definable banks where, such as from groundwater) within a few is the channel head (Figure S2.14) (Montgomery meters of travel along shallow riparian flow paths and Dietrich, 1988, 1989), which separates the pro- (McClain et al., 1999). The effectiveness of this cess domains of hillslopes or unchanneled hol- nitrogen removal can vary seasonally from very high lows (Figure S2.15) from channel networks (Diet- values in late summer to much lower values in win- rich and Dunne, 1993). Banks can be defined in ter in temperate regions (Maitre et al., 2003). In con- the field based on the presence of sediment trans- trast, little nitrogen processing may occur during port (wash marks, small bedforms, armored sur- transport from the uplands through riparian areas faces)andanobservablesharpbreakinslope.The in arid regions if precipitation moves rapidly across channel head does not necessarily coincide with theriparianzoneassurfacerunoff(McClainetal., the location where perennial flow occurs, which is 1999). the stream head (Figure S2.16). Channel segments In addition to uptake of nutrients by vascular of ephemeral and intermittent flow can be present plants and microbial communities, alluvial surfaces above the stream head and below the channel head of different ages within the riparian zone can have even in wet regions. different soil and water chemistry and infiltration The locations of individual channel heads in because vegetation influences evaporation and tran- even a small channel network can have substan- spiration that control soil water potential and con- tially different drainage areas. This is not surprising, centration of salts. Vegetation also creates litter, a given the multiple factors that influence the loca- surface layer of dead plant parts such as leaves or tion of a channel head. These factors include gra- conifer needles (Figure S2.13) that influences infil- dient, drainage area, infiltration capacity, porosity tration and evaporation from the soil, as well as con- and permeability, and cohesion, each of which influ- tributing organic matter to the soil. Some plants also ences surface and subsurface flow paths and erodi- directly fix nitrogen in the soil (Van Cleve et al., bility of near-surface materials (Figure S2.17). The 1993). location of a channel head can also vary through People have substantially reduced the extent of time as a result of changes in climate or land cover riparian vegetation and floodplain wetlands world- that affect runoff, surface erodibility, and sediment wide, causing associated changes in fluxes of water, supply (Montgomery and Dietrich, 1992). Follow- sediment, and dissolved and particulate materials ing a wildfire that killed vegetation and burned the from these environments into channels. Although surface layer of litter and duff on forested hillslopes no global estimates of total loss or of in the semiarid Colorado Front Range, USA, sur- riparian ecosystems have been published, numer- face runoff became more common, causing chan- ous regional or basin-wide estimates from industri- nelheadstomigrateupslopeandformatminimum alized countries suggest losses of much greater than drainage areas two orders of magnitude smaller than half of the riparian zones along rivers in industrial- pre-fire minimum drainage areas (Wohl, 2013a) ized countries, with consequent loss of flood attenu- (Figure S2.18). ation, sediment storage, and biological processing of Thedistributionofchannelheadsacrossa nutrients. drainage basin can reflect primarily surface or sub- surface flow, some combination of the two, or mass movements (Figure 2.6). Regardless of the mecha- 2.3 Channel initiation nisms, flow convergence facilitated by topography and/or stratigraphy promotes the concentration of Channel initiation is a threshold phenomenon in flow that initiates channels (Dunne, 1990; Dietrich which surface or subsurface flow concentrates and and Dunne, 1993). JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

Chapter 2 Creating channels and channel networks 35

SCALE OF CHANNEL HEAD Gradual Small step Large step Small headcut Large headcut <0.1 m 0.1−1.0 m 1−10 m >10 m

concentrated sediment transport Horton overland overland flow Horton

Saturation overland flow overland Saturation Slump DOMINANT RUNOFF DOMINANT PROCESS blocks Tunnel Water table Subsurface flow

Figure 2.6 Classification of channel heads based on incision depth and dominant runoff process. Sketches indicate flow paths for Hortonian overland flow and subsurface flow. Smooth arrows indicate saturated flow; wiggly arrows indicate unsaturated percolation, including flow through macropores. Even at sites with substantial Hortonian over- land flow, the face of a large headcut can allow the emergence of erosive seepage. Saturation overland flow drives erosion that includes features from both of the other runoff types (from Dietrich and Dunne, 1993, Figure 7.6).

Channel heads that reflect primarily surface pro- an increase in slope or the lowering of base level cesses can form via Hortonian or saturation over- (Dunne, 1980). land flow that leads to rilling. The greatest sur- Erosional hot spots occur where topographic face irregularities and erodibility result in flow con- constrictions or locally steep gradients amplify centrated in a master rill to which adjacent slopes hydraulic forces sufficiently to overcome surface are cross-graded (Horton, 1945). Steep slopes, high resistance and initiate headcuts (Tucker et al., 2006), rainfall intensity, low infiltration rates, and high which may coincide with the channel head or occur erodibility favor rilling (Dietrich and Dunne, 1993). downstream. Headcuts that coincide with the chan- These conditions are characteristic of arid, semiarid, nelheadareverticalfacesthatseparateupslope or disturbed landscapes dominated by Hortonian unchanneled environments from downslope chan- overland flow (Kampf and Mirus, 2013). nels. Headcuts downstream from the channel head Rills can develop nearly simultaneously across a separate upslope presently stable channel segments terrain and then integrate into a network (Dunne, from downslope recently incised channel segments 1980), although drylands typically have discontin- (Figure 2.7). uous headwater networks of short, actively erod- Channel heads dominated by subsurface pro- ing channel reaches separated by unchanneled or cesses can form via piping or sapping, or from weakly channeled, vegetated, stable reaches (Tucker shallow landsliding on steep slopes that creates a et al., 2006). Alternatively, channels can extend topographiclowwheresubsurfaceflowcanbegin downstream during slow warping or intermittent to exfiltrate (Montgomery et al., 2002; Kampf and exposure of new land on a rising land surface, Mirus, 2013). Piping occurs in the unsaturated zone or channels can extend upstream in response to when flow is sufficiently concentrated to erode or JWST426-c02 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:49 246mm×189mm

36 Rivers in the Landscape

(Henkle et al., 2011). Empirical, site-specific rela- tions between topographic parameters typically use bounding equations to quantify the range in channel head locations (Montgomery and Dietrich, 1989, 1992). An example comes from the work by Prosser and Abernethy (1996) in Gungoandra Creek, Australia

A = 30 tan휃−16 (2.1)

In this case, A is specific catchment area (the ratio of upslope catchment area to lower contour width), Figure 2.7 View upstream to a headcut along an and tan 휃 is hillslope gradient. ephemeral channel in the grasslands of eastern Colorado, The inverse slope–area relationship is not always USA. The headcut is just over 2 m tall. Approximately 1 m to the left of the survey rod and circled is a buried present, however, because of differences in runoff tree stump indicating a past cycle of incision followed processes. Low-gradient hollows with convergent by aggradation, prior to the present phase of incision. topography and seepage erosion can differ from The channel upstream of this headcut is stable now, but steeper topography where channel initiation is more has incised in the past. likely to reflect Hortonian or saturation overland flow or even landsliding (Montgomery and Diet- rich, 1989; Montgomery and Foufoula-Georgiou, dissolve subsurface materials and create physical 1993). In terrains with substantial flow through conduits for preferential flow. If a pipe enlarges suf- fractured bedrock, bedrock topography is likely to ficiently to cause collapse of overlying materials, the exert a greater influence on channel head locations pipe can have a surface expression as a channel head than does surface topography (McDonnell, 2003; (Figure S2.19). Piping typically occurs in unconsol- Adams and Spotila, 2005; Jaeger et al., 2007). Het- idated material. erogeneities in the bedrock, such as spatial variation Sapping occurs in the saturated zone, which can in joint density, can influence both the location of intersect the surface to form a spring. The return channel heads and the spatial distribution of chan- of subsurface flow to the surface enhances phys- nels within a river network, with channels tending ical and chemical weathering and creates a pore- to follow more densely jointed bedrock (Loye et al., pressure gradient that exerts a drag on the weath- 2012). ered material (Dunne, 1980). Sapping can occur in If channel heads form where saturation overland unconsolidated material or bedrock, but is not likely flow exerts a boundary shear stress that exceeds the to be an effective mechanism of erosion in bedrock critical value for substrate erosion, the channel ini- (Lamb et al., 2006; Jefferson et al., 2010) unless it is tiation threshold, C, can be expressed as the product accompanied by chemical dissolution in carbonate of contributing catchment area, A, and hillslope gra- rocks. dient, S (Dietrich et al., 1992, 1993) Early field-based studies of channel initiation indicated that the source area above the channel ASa ≥ C (2.2) head decreases with increasing local valley gra- dient in steep, humid landscapes with soil cover Substantial variability in values of A and S reflects (Montgomery and Dietrich, 1988). For slopes of the influence of factors such as vegetation, slope equal gradient, source area can vary in relation to aspect, surface versus subsurface flow paths, and total precipitation or precipitation intensity, as these substrate grain size (Montgomery and Foufoula- characteristics influence concentration of runoff Georgiou, 1993; Prosser et al., 1995; Istanbulluoglu Chapter 2 Creating channels and channel networks 37

(a) 1 Human-induced changes in land cover can alter 0.94 –0.23 –0.15 –0.27 –0.24 –0.94 the location of channel and stream heads by chang- Colluvial Bedrock Alluvial ing infiltration and the balance of water conveyed downslope in surface and subsurface flow paths. 0.1 The most typical scenario involves decreased infil- tration and smaller contributing areas for channel

Slope (m/m) and stream heads (e.g., Montgomery, 1994). Hillslopes Valley heads

0.01 102 103 104 105 106 107 108 109 1010 Drainage area (m2) 2.4 Extension and

(b) development of the drainage

Fluvial? network 101

Hillslope? Glock (1931) proposed that networks go through Transition five stages of development. Channel heads form zone 10–1 Debris flows? and rills integrate through cross-grading during the

Slope stage of initiation. The stage of extension by elonga- tion features headward growth of channels. During stage three, extension by elaboration, tributaries are added. This results in a stage of maximum extension, –3 0 3 10 10 10 followedbythefinalstageofabstractionaslocal 2 Drainage area (m ) relief is reduced and tributaries are lost (Figure 2.9). Figure 2.8 (a) Plot of log-bin averaged drainage area Subsequent field-based studies that substitute sur- versus slope relationship for the Olympic Mountains of faces of different ages for time (e.g., successive basalt Washington, USA. Plot shows the mean slope of individ- flows or glacial deposits in the same region) tend to ual 10 m grid cells for each 0.1 log interval in drainage support this general model of network development, area. Numbers at the top are the exponents for a power as do physical experiments (Schumm et al., 1987). function regression of values in the segments of the plot indicated by horizontal lines below. Dashed verti- Supplemental Section 2.4 describes different cal lines divide the plot into areas considered to reflect types of numerical models used to study network different geomorphic zones of the landscape, or process development through time. domains (from Montgomery, 2001, Figure 5A). (b) Hypo- thetical topographic signatures for hillslope and valley processes. Area and slope are measured incrementally up 2.4.1 Morphometric indices and valley mainstem to the valley head (from Stock and Diet- rich, 2003, Figure 1a; Wohl, 2010, Figure 2.4). scaling laws

Initial descriptions of drainage networks focused et al., 2002; Yetemen et al., 2010) (Figure 2.8). In on their relationship to topography. An example is the absence of field-based data, many investiga- J.W. Powell’s nineteenth century characterization of tors assume that channel heads lie near reversals antecedent and superimposed drainage networks. or inflections in averaged hillslope profiles (Ijjasz- Powellinferredthatarivercutsacrossamountain Vasquez and Bras, 1995), although this can result in range rather than flowing down and away from both significant over- or under-estimations of contribut- sides of a mountain drainage divide because the river ing area (Tarolli and Dalla Fontana, 2009; Henkle had maintained its location and cut downward as the et al., 2011) (Figure S2.20). surface was deformed around it (antecedent), or the 38 Rivers in the Landscape

order segments form a third-order segment, and so Initiation Elongation Elaboration on (Strahler, 1952). Bifurcation ratio is the ratio of the number of streams of a given order to the num- ber of streams of the next higher order. Length ratio is ratio of the average length of streams of a given order to the average length of streams of the next higher order. Drainage density is the ratio of total Maximum length of streams to basin area. Relief ratio is the ratio extension Abstraction of total basin relief to basin length. The various morphometric indices were designed to facilitate comparisons across channel networks of diverse sizes, and to provide insight into underly- ing, fundamental characteristics of energy distribu- tion, erosion, and the distribution of channels across Figure 2.9 Illustration of Glock’s stages of drainage network development. During initiation, a few, poten- a landscape. Horton (1945), for example, proposed a tially longitudinally discontinuous channels are present. law of stream numbers and a law of stream lengths With elongation, these channels grow upslope and inte- and interpreted their consistency across networks as grate into a continuous network. Additional, shorter, suggesting independence of the specific geomorphic first-order streams are added during elaboration and max- processes in any particular network. Hack (1957) imum extension. As secondary slopes and small-scale proposed that consistent scaling relations could be relief are reduced by continued channel erosion, some used to describe network characteristics such as the first-order streams are lost along the downstream mar- relation of stream length L to drainage area A gins of the main channel during abstraction. . L = 1.4 A0 6 (2.3) river had eroded down to a buried structure (super- imposed). Early descriptions also emphasized the Shreve (1966) and Smart (1968) compared statis- categorization of networks based on their planform tical properties among populations of natural chan- appearance (Zernitz, 1932; Parvis, 1950). Networks nels not strongly influenced by geologic controls. with a dendritic appearance typically indicate rela- They interpreted statistical similarity as indication tively low relief, homogeneous substrate, for exam- that the networks were topologically random and ple, whereas a rectangular network could indicate thus a consequence of random development accord- strong regional joint control that facilitates right- ing to the laws of chance. Milton (1966) and Kirch- angle channel junctions where joints intersect. ner (1993), however, proposed that, as all branching Starting in the mid-twentieth century with phenomena could equally satisfy the Horton crite- work by Horton (1932, 1945), descriptions of ria, the laws of drainage network configuration were drainage networks became increasingly quantita- geomorphologically irrelevant. tive and nondimensional. Horton, Schumm (1956), Similar controversy has arisen over more recent and Strahler (1957) modified earlier morphometric work describing drainage network patterns using indices and developed new indices, helping to dis- fractal approaches (e.g., Rodr´ıguez-Iturbe and seminatetheideaofcharacterizingrivernetworks Rinaldo, 1997). Fractals are geometrical structures using stream order, bifurcation ratio, length ratio, with irregular shapes that retain the same degree drainage density, relief ratio, and other indices. of irregularity at all scales and are thus self-similar. Stream order isanumberassignedtoastream Investigators who have observed fractal properties segment. A segment with no tributaries is a first- in networks argue that this reflects a scale inde- order stream. Where two first-order segments join, pendence to landscape dissection. Beauvais and they form a second-order segment; two second- Montgomery (1997), however, note that a minimum Chapter 2 Creating channels and channel networks 39 contributing area is needed to concentrate surface Stream order provides a convenient means to or subsurface flow sufficiently to form channels, indicate the relative size of a channel segment (e.g., indicating a scale dependence to landscape dis- a first-order stream) or of a drainage network (e.g., section. Beauvais and Montgomery found that a third-order drainage). Stream order is sometimes drainage networks in the western United States replacedbyalinkasthebasicunitofnetworkcom- do not exhibit the scaling properties required for position. A link is an unbroken section of channel fractal geometry. between successive nodes (sources, junction, or out- Regardless of the insights that may or may not let) (Shreve, 1966). An exterior link corresponds to a result from statistical self-similarity or consistent first-order stream in that it extends from a source to power functions between discrete variables, indi- the first junction downstream. An interior link con- vidual morphometric indices remain useful for nects two successive junctions, or the last junction inter-drainage comparisons and for understanding and the outlet (Knighton, 1998). drainage network history. Stream order, drainage Drainage density reflects the degree to which a density, and relief ratio are especially widely used basin is dissected by channelized flow, and this in (Figure 2.10). turn reflects the relative influences of substrate resis- tance (higher resistance typically results in lower density), rainfall-runoff-land cover (high drainage (a) (b) densities typically occur in semiarid regions that 1 1 receivemorerainfallthanaridregions,butlackthe 1 1 1 1 1 2 1 continuous vegetation cover and associated higher 2 2 infiltration of humid regions), and age of the net- 1 1 work (as reflected in Glock’s stages of drainage 111 development). Relief ratio indicates the steepness of topography 2 D = ∑ L/A 3 within a basin. Steepness can provide insight into tectonic uplift and rock resistance, as well as down- slopepathwaysandstoragetimeofwater,sediment, (c) and solutes. A fundamental limitation on being able to char- acterize any drainage network is the ability to accu- L Rh = H/L0 0 ratelymapactualchannellocations.Thisisnotpar- ticularly difficult for large channels, but can be very problematic for the smallest, headwater tributaries. The default assumption is typically to use the extent Figure 2.10 Examples of commonly used stream mor- of “blue lines” used to indicate channels on topo- phometric indices. (a) Stream order from Strahler (1952). graphic maps or digital elevation models (DEMs). First-order streams have no tributaries. A higher stream The correspondence between these lines and real order occurs where two streams of equal order join. This channels varies with the spatial resolution and type stream network is third order. Each of the first-order of data used to generate the map or DEM, and the streams is also an exterior link, and each of the second- consistency of channel extent through time (e.g., and third-order streams is an interior link (Shreve, 1966). (b) Drainage density is the ratio of the sum of stream channel extent can vary substantially over a period lengths per unit drainage area (area shaded gray). (c) ofyearsindrylandsorinareaswithrapidlyerod- ingheadwaters).Anotherapproachistousecon- Relief ratio Rh is the ratio of the maximum topographic relief (elevation difference) H to the longest horizon- tour crenulations (the degree to which contour lines tal distance of the basin measured parallel to the main are distorted when crossing a valley) as indicators of stream L0. sufficient valley incision to create channelized flow, 40 Rivers in the Landscape but the threshold value chosen for crenulation varies nels represents a quasi-equilibrium state that is most between studies and the correspondence between probable to occur because it balances the oppos- crenulation and channelized flow varies between ing tendencies of minimum total rate of work in regions. As noted in Section 2.3, reversals or inflec- the whole river system and uniform distribution of tionsinaveragedhillslopeprofilescanalsobeused energy expenditure throughout the system. to indicate the location at which channelized flow The tendency toward uniform distribution of begins across a drainage basin. Each of these meth- energy expenditure can be understood by consid- odsissubjecttoerrors,anditisimportanttobe ering a portion of channel with higher rates of aware of these errors as a source of uncertainty in energy expended against the channel boundaries network analysis. becauseofalateralconstrictionorlargergradient that results in locally higher velocity. Assuming that thechannelboundaryiserodible,thegreaterrateof 2.4.2 Optimality energy directed against the boundary should enlarge theboundaryuntiltherateofenergyexpenditure Various aspects of river form and process have been declines. An example comes from the Colorado described using extremal hypotheses that character- River in the Grand Canyon, USA. Flash floods and ize tendencies toward which rivers evolve based on debris flows along steep, ephemeral tributaries to the the balance between energy available to move water Colorado River create tributary debris fans that con- and sediment and shape channels, and the resistance strict the main channel. During large floods on the of the channel boundaries. The extremal hypothe- ColoradoRiver,theseconstrictionsforcetheflowto sis of optimal channel networks has been applied to become critical or supercritical (Section 3.3.1). The understanding the spatial arrangement of channels higher energy directed against the channel bound- in a drainage network. Optimal channel networks aries through the constriction during floods erodes display three principles of optimal energy expendi- thetoeofeachdebrisfanuntilthemainchannelis ture: (1) minimum energy expenditure in any link sufficiently widened to permit subcritical flow even of the network; (2) equal energy expenditure per during large floods (Kieffer, 1989). Erosion then unit area of channel anywhere in the network; and decreases and the channel cross-sectional geometry (3) minimum total energy expenditure in the net- becomes stable until the next tributary input again work as a whole (Rodr´ıguez-Iturbe et al., 1992). creates a constriction. In other words, sites of very The effect of the tendencies underlying these prin- high energy expenditure are transient in an erodi- ciples can be considered in terms of local and global ble channel. Channel geometry will adjust to reduce optimal energy expenditure. In local optimal energy energy expenditure at the site and create more uni- expenditure,thechannelproperties(e.g.,channel form distribution of energy expenditure relative to cross-sectional geometry and gradient) of river net- upstream and downstream sites. Numerous adjust- works adjust toward a constant rate of energy dissi- ments of the type described for the Grand Canyon pation per unit channel area. In global optimal energy along the course of a river or throughout a drainage expenditure,thetopologicalstructureofnetworks network can result in uniformity and minimiza- (e.g., drainage density, bifurcation ratio) adjusts to tion of energy expenditure, given sufficient time and minimize total energy dissipation rate (Molnar´ and energy relative to boundary resistance. Ram´ırez, 1998; Molnar,´ 2013). Energy expenditure One of the points of discussion regarding can be quantified using variables such as shear stress extremal hypotheses is whether most natural or stream power. channels ever actually attain an optimal state, given These principles largely grew out of work by the potential for highly resistant boundaries or Leopold and Langbein (1962), Langbein (1964), and continual perturbations such as tectonic uplift Langbein and Leopold (1964). Leopold and Lang- or sediment inputs. An assumption underlying bein proposed that the mean form of river chan- hypotheses of optimal channel networks is that Chapter 2 Creating channels and channel networks 41 place-specific details such as lithology, structure, uum are completely buffered links along which flood- or tectonics that influence substrate resistance are plains or valley-fill deposits protect hillslopes from absent or negligible. This is often not the case. basal erosion and limit direct sediment supply from Consequently, extremal hypotheses may describe hillslopes to channels. Schumm’s headwater produc- the conditions toward which a system is evolving, tion zone would be best characterized as composed rather than the actual state of the system. Many of strongly coupled links, whereas the proportion of natural river networks, however, do approximate the completely buffered links would increase progres- structure of optimal networks (Rodr´ıguez-Iturbe sively downstream in the transfer and depositional and Rinaldo, 1997; Molnar´ and Ram´ırez, 1998). zones. Again, these are general patterns: headwater channels can be buffered and downstream channel segments can be strongly coupled. Most importantly, 2.5 Spatial differentiation thedegreeofcouplingforanychannelsegmentcan vary through time. within drainage basins Spatial differentiation of geomorphic process domains can be used to distinguish six basic process Schumm (1977) conceptualized a drainage basin as domains (Montgomery et al., 1996; Sklar and Diet- being longitudinally zoned with respect to sediment rich, 1998; Montgomery, 1999) (Figure 2.11). dynamics. He distinguished a headwater produc- tion zone, mid-basin transfer zone, and downstream r Hillslopes: Hillslope in this context refers to the depositional zone (Figure S2.21). These zones rep- unchannelized, largely straight or convex portion resent predominant processes: the headwater zone of slopes, typically dominated by diffusive trans- does include transfer and deposition, for example, port or slopewash. but is most characterized by sediment production. r Unchanneled hollows or zero-order basins: Subsequent research has elaborated on and quanti- Unchanneled hollows are hillslope concavities fied these distinctions. Benda et al. (2005), for exam- thatdonothavechannelizedflow,butserveas ple, describe headwater streams as sediment reser- sediment storage sites and are important points voirs at timescales of 101–102 years, along which for the initiation of mass movements (Dietrich sediment is episodically evacuated by debris flows and Dunne, 1978; Montgomery et al., 2009). or gully erosion. Sediment yield per unit area, or r Debris-flow channels: Debris-flow channels can be sediment delivery ratio, decreases and sediment resi- influenced by , but non-fluvial dence time increases as stream order and drainage processes dominate sediment dynamics and chan- area increase as a result of increasing storage on nel geometry. hillslopes or valley bottoms (Schumm and Hadley, r Bedrock-fluvial channels: Bedrock-fluvial channels 1961; Dietrich and Dunne, 1978). These changes have bedrock exposed along the channel bound- mayalsobeastepfunction,withrelativelylargerates ariesoratsufficientlyshallowdepththatoverly- of change in sediment storage volume and time at ing alluvium is readily mobilized during higher transitions in geomorphic processes, such as where flows, so that the underlying bedrock limits chan- debris flows give way to predominantly fluvial pro- nel boundary erosion. The distinctions among cesses (Lancaster and Casebeer, 2007). bedrock-fluvial channels and coarse- and fine- Rice (1994) described a continuum of sediment bed alluvial channels reflect differences in the bal- transfer to channels, which represents another way ance between sediment supply and river transport of describing the sediment connectivity and storage capacity. Where transport capacity exceeds sedi- discussed in Chapter 1. At one end of the continuum ment supply, bedrock is exposed in the channel are strongly coupled links in which sediment is trans- bed. ferred from hillslopes to channels relatively rapidly r Coarse-bed alluvial channels: Coarse-bed allu- and continuously. At the other end of the contin- vial channels are channels with unconsolidated 42 Rivers in the Landscape

1

Debris flow region

0.1 Fluvial–debris flow transition

Bedrock–fluvial region Bedrock–alluvial transition 0.01

Channel gradient Coarse bed alluvial region 0.001 Gravel–sand transition

Fine bed alluvial region 0.0001

1100010 100 10000

Drainage area (km2)

Figure 2.11 Hypothesized distribution of channelized erosional process domains (fluvial vs. debris flow) and chan- nel substrate types in relation to drainage area and slope. (From Sklar and Dietrich, 1998, Figure 1, p. 241.)

sediment coarser than sand size. Sediment mobil- 40%–70% sediment by weight and a bulk density of ity in these channels is likely to be limited by sed- 1.3–1.8 g/cm3 (Costa, 1984; Pierson, 2005). A water iment supply, rather than flow energy. flow carries only 1%–40% sediment by weight and r Fine-bed alluvial channels: Fine-bed alluvial chan- has a bulk density in the range of 1–1.3 g/cm3 (Costa, nels have non-cohesive sand-sized sediment form- 1984). ing the channel bed. Sediment mobility is more Differences in bulk density are important because likely to be limited by flow energy than by sedi- bedload transport rate and maximum clast size ment supply. can increase with increasing fluid density if the flowaroundthegrainsisnotlaminar(Ricken- The transition between debris-flow and com- mann, 1991). Debris flows can be highly erosive pletely fluvial channels is marked by changes in in steep or confined channels, and can create sub- channel process and form as a result of the dif- stantial aggradation in lower gradient, less confined ferences in flow mechanics. Flow conditions can channel segments or at channel junctions (Benda, vary among debris flow, hyperconcentrated flow, 1990; Wohl and Pearthree, 1991). Because they are and water flow downstream and with time during a capable of mobilizing clast sizes and volumes of flow as a result of changes in sediment concentration sediment greater than those mobilized by fluvial (O’Connor et al., 2001). Water and sediment move processes, debris flows drive cycles of aggradation together in a debris flow as a single viscoplastic body and degradation in the mountainous portions of thatcanbeupto90%sedimentbyweight,witha many river networks (Benda and Dunne, 1987), bulk density of 1.8–2.6 g/cm3 (Costa, 1984; Iverson, as well as strongly influencing the gradient of side 2005). A hyperconcentrated flow is a water flow with slopes,disturbanceregime,andvalleyandchannel Chapter 2 Creating channels and channel networks 43 morphology,includingaquatichabitat,organicmat- duced through bedrock weathering does not move ter,andthestructureandcompositionofriparian uniformly downslope into rivers, but instead moves vegetation (Swanson et al., 1987; Florsheim et al., abruptly during mass movements or diffusive creep 1991; Hewitt, 1998; Benda et al., 2003b). that occurs primarily during precipitation, with Valleys consistently subject to different flow pro- sediment coming disproportionately from steep or cesses can have distinctly different geometries (De otherwise more readily erodible portions of a catch- Scally et al., 2001). Valleys dominated by debris ment.Solutesenteringtherivercomedispropor- flows and fluvial erosion are more incised and con- tionately from more readily weathered minerals or tain closely spaced channels, for example, relative lithologies, and are strongly influenced by move- to valleys dominated by landslides. Valley gradi- ment through biochemically active portions of the ent can decrease abruptly at the transition from catchment, such as riparian zones. debris flow to fluvial channels (Stock and Dietrich, Channel and stream heads that mark the start 2006). Downstream from this transition strath ter- of a river network reflect the processes that gov- races (Section 7.2.1) can form and drainage area– ern downslope movement of water and sediment, stream gradient relations (A–S)followfluvialpower as these processes cause water to concentrate laws such that S varies as an inverse power law of A sufficiently to create definable, persistent chan- (Stock and Dietrich, 2003). Debris flows in forested nels. Threshold conditions of drainage area and regions can entrain substantial volumes of wood that hillslope gradient at which channels form vary thedebrisflowsubsequentlydepositsasadamthat among different catchments and within a catchment cangiverisetoaggradation,terraces,andoutburst through time in response to differences in surface floods (Lancaster et al., 2003; Comiti et al., 2008; and subsurface downslope pathways of water and Rigon et al., 2008). Mountainous headwater chan- sediment. nels subject to debris flows can display downstream A variety of morphometric indices have been coarsening of median bed surface grain size (Brum- used to characterize the spatial distribution of mer and Montgomery, 2003), in contrast to the more channel networks and topography within catch- typical downstream fining in fluvial channels (Sec- mentsasameansoffacilitatingcomparisonamong tion 5.2.2). catchments. Stream order, drainage density, and relief ratio are particularly widely used. Investi- gators have also used mathematical frameworks 2.6 Summary suchasoptimalityandfractalstosearchforuni- versal physical laws that underlie and can provide Every process discussed in this chapter is charac- insight into observed patterns in river networks. terized by diverse levels of spatial and temporal This approach remains controversial at least in part variability. Beginning with inputs and downslope because it is unclear to what degree actual river movement of water, the details of how precipitation networks, with their departures from ideal math- is produced and how water moves downslope via ematicalformsasaresultofspatialvariationin surface and subsurface pathways vary significantly substrate resistance, sediment inputs, and so forth, across even a small catchment and through rela- canbeeffectivelydescribedbyoptimalityorfrac- tively brief intervals (hours to days) of time. Water tals.Atamoregenerallevel,mostcatchmentscan movement, in particular, is characterized by thresh- be usefully viewed as being spatially differentiated olds. Once precipitation or glacier or snow melt into: begins, the downslope pathways and rates of water movement change substantially as different compo- r headwaters that are characterized by relatively nentsofahillslopeturnon,suchaswhenasub- efficient movement of water, solutes, and sedi- surface pipe network becomes active or infiltration ment from uplands into first- and second-order gives way to surface flow. Similarly, sediment pro- channels; 44 Rivers in the Landscape r mid-basin zones characterized not only by down- r lower basin depositional zones characterized by stream transfer of water, solutes, and sediment, but more spatially extensive and longer-term storage also by some storage of these materials in valley- of water, solutes, and sediment in well-developed bottom environments such as alluvial fans and floodplains and deltas. floodplains; and Channel processes I

Chapter 3 Water dynamics

This chapter introduces the basic physical proper- 3.1 Hydraulics tiesofwaterflowwithinachannel.Thediscussion of hydraulics starts by explaining how flow is clas- Water flowing down a channel converts potential sified based on velocity, the ratio of viscous to iner- energy to kinetic energy and dissipates energy. The tial forces, and the ratio of inertial to gravity forces. rate and manner in which energy are expended The next section describes measures of the energy depend on the configuration of the channel, of water flowing in a channel and the implications including the frictional resistance of the channel of energy level for stability of the flow and adjust- boundaries, and the amount of sediment being ments of the channel boundaries. Energy and stabil- transported. Velocity is one of the most commonly ityarecloselyconnectedtosourcesofflowresistance measured hydraulic variables, and is particularly and equations used to quantify resistance, as well as variable in time and space because of its sensitivity velocity and turbulence that result from the interac- to frictional resistance. The basic flow continuity tions between available flow energy and resistance. equation introduced in Chapter 1, Q = wdv,isquite The final section on hydraulics returns to the idea mathematically simple. Any of the dependent vari- of energy and introduces variables used to express ables can be very difficult to predict in natural chan- the energy exerted against the channel boundaries. nels,however,evenifQ is known, because each is Understanding the physical properties of flowing influenced by other properties of the flow and chan- water is necessary to everything that follows in this nel boundaries. Width, for example, reflects Q,and volume because the quantity of energy available and alsotheabilityofflowtoerodethechannelbanks. the resistance created by the channel boundaries Depth reflects Q,aswellastheerodibilityofthebed. govern the movement of sediment and adjustment Velocity reflects Q, and also frictional resistance of of channel geometry. the channel boundaries. Width, depth, and velocity The second part of the chapter first addresses alsoinfluenceeachother.Increasingflowdepth methodsusedtoestimatevolumeofflowoverdif- changes the resistance to flow and hence the velocity fering time periods. This is followed by a discus- associated with a particular bed grain-size distribu- sion of the effects on stream discharge of surface– tion, for example, as the grains protrude into a pro- subsurface water exchanges, and the effects on dis- gressively smaller proportion of the total flow depth. charge of flow regulation and channel engineering.

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

47 JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

48 Rivers in the Landscape

(a) 1 2

hf 2 v1 Energy line, slope = S 2g f 2 v2 2g

d1 θ Water surface, slop v1 e = Sw 90° v2 d2 Stream bed, slope = S θ o

z1 z2 datum line

(b) Energy line

Water surface

Stream bed

Figure 3.1 Parameters of open-channel flow; variables are defined in the text. Longitudinal view of a short channel segment, with flow from left to right. (a) Uniform flow; assumed parallel flow with a uniform velocity distribution and small channel slope. (b) Gradually varied flow, with velocity increasing and depth decreasing downstream.

Knowledge of basic hydraulic properties is neces- McDonald (1978), a classic fluid mechanics text- sary to understand the complex interactions among book, and from Robert (2003), which contains a flowing water, channel geometry, and sediment thorough and comprehensive treatment of water transport, and to understand the assumptions that and sediment dynamics in rivers. underlie many of the equations applied to processes Waterflowinginaconduitcanbeeitheropen- in natural channels. Consequently, the first por- channelfloworpipeflow.Open-channel flow has a tion of this chapter introduces hydraulic param- free surface at the boundary between the water and eters and equations commonly applied to natural theatmosphere,andthisfreesurfaceissubjectto channels. atmospheric pressure (Chow, 1959). Figure 3.1 illus- trates several important parameters used to charac- terize open-channel flow. For simplicity, individual 3.1.1 Flow classification flow lines are parallel and moving at the same veloc- ity, and the slope of the channel is small. The total Much of the following discussion of basic hydraulics energy in the flow of the section with reference to in open-channel flow comes from Chow (1959), a datum line is the sum of the elevation z,theflow a classic hydraulics textbook, from Fox and depth y,andthevelocityheadv2/2g,wherev is the JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 49

mean velocity of flow and g is gravitational acceler- Table 3.1 Types of open-channel flow. ation (9.8 m/s2). The energy is represented by the Type of flow Criterion energy grade line or energy line,andthelossofenergy resulting from water flowing from section 1 to sec- Uniform/varied Velocity is constant with position/ velocity is variable with position tion 2 is hf (Chow, 1959). The depth of flow, the dis- charge, and the slopes of the channel bottom and the Steady/unsteady Velocity is constant with time/velocity water surface are interdependent. The flow parame- is varied with time Laminar/turbulent R < 500/ R > 2500 (transitional flow ters in Figure 3.1 form the basis for procedures such e e between) as step-backwater calculations that are used to esti- Subcritical/critical/ F < 1/F = 1/F > 1 matedischargefrompaleostageindicators(Supple- supercritical mental Section 3.2.1). Water flowing in an open channel is subject to gravityandfriction.Externalfrictionresultsfrom Considering only the vertical dimension within a the channel boundaries. The portion of the flow channel, eddy viscosity can be expressed as closest to the boundaries typically moves more slowly than portions of the flow further away from 𝜀 2 the boundaries as a result of external friction. Exter- = l dv∕dz (3.1) nal friction can vary widely in natural channels because of the presence of an irregular surface over where l is the mixing length—the characteristic individual grains, the presence of bedforms, bends, distance traveled by a particle of fluid before its changes in channel width, and other factors, as dis- momentum is altered by the new environment cussed in more detail in Section 3.1.3. (Chanson, 1999), dv is change in velocity, and dz is Internal friction results from eddy viscosity change in height above the bed (Robert, 2003). The (Chow, 1959). Viscosity represents the resistance of a mixing length represents the degree of penetration fluid to deformation. The molecular or dynamic vis- of vortices within the flow and depends on distance cosity, 𝜇, is the internal friction of a fluid that resists from the boundary. (A vortex is the rotating motion forces tending to cause flow (Robert, 2003). The of water particles around a common center (Lugt, greater the dynamic viscosity the smaller the defor- 1983).) The mixing length is assumed to be mation within the fluid for a given applied force and the lower the degree of mixing and turbulence. Kine- l = 𝜅z (3.2) matic viscosity, 𝜐,istheratioofmolecularviscos- ity to fluid density (𝜐 = 𝜇/𝜌; typically 𝜐 ∼ 1 × 10−6 where 𝜅 is the von Karman constant, which is 0.41 m2/s). Both types of viscosity decrease significantly in clear water flowing over a static bed (Robert, as water temperature increases (Fox and McDon- 2003). Understanding of eddy viscosity is important ald, 1978). The effect of changing density on kine- because eddies allow dissolved or particulate mate- matic viscosity is such that very high suspended sed- rial carried in the water to spread throughout the iment concentrations can increase kinematic viscos- flow field. ity (Colby, 1964). Open-channel flow is classified into types based Eddy viscosity is friction within the flow that on four criteria. The classifications segregate based results from the vertical and horizontal circulation on properties important to natural processes (Table of turbulent eddies. Eddy viscosity expresses the 3.1). Uniform flow occurs when the depth of flow vertical and horizontal transfer of momentum, or isthesameateverysectionofthechannel(Fig- exchange between slower and faster moving parcels ures 3.1a, 3.2), and is very rare in natural channels. of water, and varies with position above the bed. The Flow is varied if the depth changes along the length coefficient of eddy viscosity, 𝜀, represents momen- of the channel, and can be either rapidly varied tum exchange or turbulent mixing (Robert, 2003). (Figure 3.2) if the depth changes abruptly over a JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

50 Rivers in the Landscape

Energy line

d1 d1 = d2 d

2 Discharge

Uniform Months Energy line Energy line Unsteady – gradual variation Hydraulic jump – series of steady flows d1 d1 ≠ d2 d2 Discharge

Gradually varied Rapidly varied Months (discontinuous) Figure 3.2 Illustration of differences between uni- Unsteady form, gradually, and rapidly varied flow. Longitudinal – rapid variation view of a short channel segment, with flow from left to Figure 3.3 Idealized hydrographs for different types of right. d is flow depth. Figure courtesy of David Dust. unsteady flow. Figure courtesy of David Dust.

comparatively short distance, or gradually varied at the cross section), 𝜌 is mass density, and 𝜇 is (Figures 3.1b, 3.2). dynamic viscosity. Uniform flow can be steady or unsteady. In steady The Reynolds number can be used to differentiate flow, the local depth does not change or can at least laminar, turbulent, and transitional flow. In laminar be assumed to be constant during the time interval flow,theviscousforcesaresufficientlystrongrelative under consideration. Like uniform flow, steady flow to the inertial forces that viscosity exerts a signifi- is rare in natural channels. Unsteady flow can occur cant influence on flow behavior. Each fluid element as gradual variation among a series of steady flows, movesalongaspecificpathwithuniformveloc- or as rapid variation (Figure 3.3). ity and no significant mixing between adjacent lay- Most natural flows are varied and unsteady, ers. When the velocity or hydraulic radius exceeds a althoughthevariationsintimeandspacecanbe critical value, the flow becomes turbulent.Thevis- gradual. Although uniform flow was traditionally cousforcesareweakrelativetotheinertialforces. assumed for simplicity when calculating hydraulic The fluid elements follow irregular paths and mixing parameters in natural channels (Chow, 1959), many occurs, involving transfer of momentum by large- applications now assume gradually varied flow. scale eddies. The state of open-channel flow is governed by the The transition from laminar to turbulent depends effects of viscosity and gravity relative to the inertial on the energy dissipation within the fluid: high vis- forces of the flow. The effect of viscosity relative to cosityequatestomoreenergynecessarytomixthe inertia can be represented by the Reynolds number flow. A smooth or glassy water surface in a channel does not mean that the flow is laminar. The flow in vR𝜌 most channels is turbulent, but laminar motion can R = (3.3) e 𝜇 persist in a very thin layer (typically <1mmthick) next to the channel boundary that is known as the where v is velocity, R is hydraulic radius (cross- laminar sublayer (Figure 3.4). sectional area divided by the wetted perimeter; wet- The Reynolds number provides a useful indicator ted perimeter is the wetted length of bed and banks of the magnitude of mixing within a natural channel. JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 51 3.1.2 Energy, flow state, and hydraulic jumps Laminar sublayer k s As noted earlier, water flowing downhill within a Figure 3.4 Side view of vertical velocity distribution channel is converting potential energy (PE = mgh, in a channel (flow from left to right). Arrows indicate for water mass m and height h above a given datum) 1 relative velocity and k indicates roughness height. Fig- to kinetic energy (KE = mv2), some of which is s 2 ure courtesy of David Dust. dissipated at the microscale of turbulence and the molecular scale of heat. A streamline is a visual rep- resentation of the flow field in the form of a line > drawnintheflowfieldsothat,atagiveninstantof The flow is typically turbulent when Re 2300, and < time, the streamline is tangent to the velocity vector laminar when Re 2300, although there is no single at every point in the flow field (Fox and McDonald, value of Re at which the flow changes from laminar to turbulent (Fox and McDonald, 1978). 1978). In other words, streamlines indicate the paths The effect of gravity on the state of flow is repre- offlow,andtherecanbenoflowacrossastreamline. sented by the ratio of inertial forces to gravity forces, The total energy in any streamline passing through as given by the Froude number, F a channel section can be expressed as the total head H,whichisequaltothesumoftheelevationabove a datum, the product of the depth below the water v F = √ (3.4) surface d and the cosine of the bed angle 𝜃,andthe gd velocity head (Figure 3.1)

where d is the hydraulic depth, defined as the cross- 𝛼v2 H = z + dcos𝜃 + (3.5) sectionalareaofthewaternormaltothedirectionof 2g flow in the channel divided by the width of the free surface (Chow, 1959).√ For channels of small slope, 𝜃 ∼ 0, and the cos 𝜃 ∼ When F = 1, v = gd,andtheflowisinacrit-√ 1. Because velocity is not uniformly distributed over < < ical state. When F 1, v √gd,andtheflowis a channel section, the velocity head of open-channel subcritical.WhenF > 1, v > gd,andtheflowis flow is typically greater than the value of v2/2g.The supercritical. Critical flow can occur when the flow true velocity head is expressed as 𝛼v2/2g,where𝛼 is constricted or subject to a substantial increase is the energy coefficient and typically varies from in bed slope and passes√ through a state of mini- about one to two for natural channels (Chow, 1959). mum specific energy (v/ gdc= 1, where dc is crit- The value is higher for small channels and lower ical hydraulic depth), as it funnels through or drops for large, deep channels. Every streamline passing over the channel contraction or slope break (Section through a cross section will have a different veloc- 3.1.2). ity head because of the nonuniform velocity distri- Subcritical flow has lower velocity and greater bution in real channels, but the velocity heads for depth for a given channel configuration and dis- allpointsinthecrosssectioncanbeassumedtobe charge, and is stable and persistent. In a sense, sub- equal in gradually varied flow (Chow, 1959). critical flow is the usual flow state. Supercritical flow The line representing the elevation of the total is inherently unstable and changes in channel bed head of flow is the energy line,andtheslopeof

gradient or cross-sectional area can cause the flow the energy line is known as the energy gradient Sf to become subcritical, with implications for erosion (Chow, 1959). The water-surface slope is Sw and the 𝜃 𝜃 and deposition along the channel, as discussed in the slopeofthechannelbedisSo = sin ,where is next section. theslopeangleofthechannelbed.Inuniformflow, JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

52 Rivers in the Landscape

𝜃 Sf = Sw = So = sin (Chow, 1959). Based on the prin- ciple of the conservation of energy, the total energy head at upstream section 1 (Figure 3.1) should equal the total energy head at downstream section 2 plus SubcriticalCritical flow staterange the loss of energy hf betweenthetwosectionsinpar- d allel or gradually varied flow

𝛼 2 𝛼 2 Supercritical flow range 1v1 2v2 z + d + = z + d + + h (3.6) Flow depth, dc 1 1 2g 2 2 2g f

𝛼 This is known as the energy equation.When 1 = 𝛼 2 = 1andhf = 0, this becomes the Bernoulli equa- tion, which holds in friction-free flow Specific energy, E Figure 3.5 Specific energy curve; dc is critical flow v2 v2 depth. z + d + 1 = z + d + 2 = constant (3.7) 1 1 2g 2 2 2g cal. Specific energy changes as discharge changes, as Specific energy in a channel section is the energy showninFigure3.5.Thecriterionforcriticalflow perkilogramofwateratanysectionofachan- is that the velocity head equals half the hydraulic 2 nel measured with respect to the channel bottom depth (v /2g = d/2), assuming that flow is parallel or (Chow, 1959). Using Equation 3.5, with z = 0, spe- gradually varied, the channel slope is small, and the cific energy E is energy coefficient equals one (Chow, 1959). Other- wise, 𝛼v2/2g = (d cos 𝜃)/2. 𝛼v 2 Aflowatornearthecriticalstateisunsta- E = dcos𝜃 + (3.8) 2g ble because a minor change in specific energy will cause a major change in depth (Chow, 1959). Con- For a channel of small slope and 𝛼 = 1, sequently, extended areas of critical or supercritical flow are relatively uncommon in natural channels. v2 The flow is likely to return to a subcritical state via a E = d + (3.9) 2g hydraulic jump. When a rapid change in the flow depth from so that specific energy is equal to the sum of the flow low stage to high stage occurs, the water surface depth and the velocity head. Because v = Q/A, E = rises abruptly in a hydraulic jump (Figure 3.6). This d + Q2/2gA2. Therefore, for a given channel section is common where a steep channel slope abruptly and discharge, E is a function of only the depth of decreases. An undular jump forms when the change flow (Chow, 1959). A specific energy curve is a plot in depth is small and the water passes from low to of E and d for a given channel section and discharge high stage through a series of undulations that grad- (Figure 3.5). ually diminish in size (Chow, 1959). A direct jump For a given specific energy, two possible depths occurs at a large change in depth and involves a rel- exist,thelowstageandthehighstage,eachofwhich atively large amount of energy loss through dissipa- is an alternate depth for the other. At point C on the tion in the turbulent flow in the jump.

curve, known as critical depth dc, the specific energy Thegeomorphicsignificanceofahydraulicjump is a minimum, which corresponds to the critical state is that the jump creates intense turbulence and large of flow. At greater depths, the velocity is less than kinetic energy losses. Flow velocity decreases sub- thecriticalvelocityforagivendischargeandflowis stantially across the jump and sediment can be subcritical. At lower depths, the flow is supercriti- deposited downstream, stabilizing the position of JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 53

Critical flow and depth, F = 1 Water flowing down an open channel encoun- Subcritical F < 1 Subcritical F < 1 ters resistance that is counteracted by gravity act- Supercritical Hydraulic jump ing on the water in the direction of motion. Uni- Energy line F > 1 form flow occurs when the resistance is balanced by gravity forces. Velocity is related to flow resistance or boundary roughness using one of the three com- monly applied uniform flow equations, the Chezy,´ Darcy–Weisbach, and Manning equations. Because steady uniform flow is rare in natural channels,

Hydraulic jump the results obtained by applying these equations are Gradually varied Rapidly varied Gradually varied approximate and general (Chow, 1959). The equa- tions continue to be widely used because they are relatively simple to apply and provide satisfactory Figure 3.6 Longitudinal view of a hydraulic jump and approximations for many natural channels. the associated velocity distributions and flow conditions The French engineer Antoine de Chezy´ devel- in channel segments immediately upstream and down- oped the first uniform flow formula stream from the jump. Flow from left to right. F is Froude number. Length of arrows indicates relative velocity. The √ boulder at the change in slope can localize the jump, v = C RS (3.10) but is not necessary for the jump to occur. The abrupt change in slope can be sufficient to create a hydraulic where v is mean velocity, R is hydraulic radius, S is jump. Figure courtesy of David Dust. theslopeoftheenergylineSf,commonlyapproxi- mated as water-surface slope Sw or bed slope So,and C is a factor of flow resistance that represents the the jump (Carling, 1995). If the channel boundaries ratio between the driving force (RS)andthevelocity areadjustableunderagivenflow,sitesofsupercrit- sustained in the presence of frictional resistance to ical or critical flow are likely to be preferentially flow. Secondary equations developed for calculating eroded so that the channel cross section enlarges or the value of C typicallyrelyonanempiricallydeter- gradient declines until subcritical flow occurs (Kief- mined coefficient of roughness. fer, 1989; Grant 1997). Natural channels thus exhibit Henry Darcy, Julius Weisbach, and other feedbacks between flow energy and channel geom- nineteenth-century engineers developed the etry that tend to maintain subcritical flow (Grant, approach now known as the Darcy–Weisbach 1997). equation

( ) 1 8gRS 2 v = (3.11) 3.1.3 Uniform flow equations f and flow resistance The Darcy–Weisbach equation was developed Both velocity distribution within a channel and aver- primarily for flow in pipes, but the form of the equa- age velocity are sensitive to boundary roughness that tion stated earlier can be applied to uniform and retards flow. Velocity can be directly measured with nearly uniform flows in open channels. The variable a velocity meter, and the measured value can be used f is sometimes preferred over other friction factors to calculate flow resistance. Or flow resistance coef- becauseitisnondimensionalandhasamoresound ficientscanbeusedtoestimatevelocityundercondi- theoretical basis. tions in which direct velocity measurements are not Naturalchannelshaveirregularboundariesthat feasible. can be more difficult to characterize in terms JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

54 Rivers in the Landscape

Bedforms Pool Riffle Bend

Bridge piers Bank vegetation

Irregularities in Logjam wetted perimeter Sediment suspended in the flow

Bed vegetation Sediment moving along the bed

Bed grain size and shape Figure 3.7 Schematic illustration of some of the sources of resistance (named in italicized text) that are lumped together in resistance coefficients such as C, f, or n. Sources of resistance in natural channels include boundary irregularities, vegetation along the channel boundaries, sediment forming the bed and banks, and sediment in transport, obstructions, and channel bends.

of resistance than flow in pipes. The Gauckler– Determining an n value—or a value for C or f—is Manning empirical formula was developed in the the most difficult part of applying these equations, late nineteenth century by French engineer Philippe in part because n incorporates so many forms of Gauckler, and subsequently modified by Irish engi- roughness and flow resistance. Each of the resistance neer Robert Manning, to describe the relation coefficients listed earlier includes multiple sources among velocity and flow resistance in natural chan- of resistance (Chow, 1959) (Figure 3.7). Values of nels. The formula was subsequently modified to its n are commonly quite variable in space and in present form time because n value tends to decrease as stage and discharge increase, although the rate of change in R2∕3S1∕2 n values can be nonlinear because of interactions v = (3.12) n between the flow and the channel boundaries. The most common approach is to visually estimate the where n is the coefficient of resistance, typically n value using standard tables (Chow, 1959) or com- known as Manning’s n.Thisisthemostwidelyused parisons to field-measured values from diverse sites formula for uniform flow. (Barnes, 1967). Numerous empirical equations also The flow resistance factors C and n relate to one exist for estimating n values for particular types of another as channels (Supplemental Section 3.1.3). A key consideration in applying these uniform ( ) 1 flowequationsisthatmanynaturalchannelshave 1 D 6 C = H (3.13) local accelerations and decelerations associated with n 4 protruding grains or bedforms, so that bed, water surface, and energy slopes may not be equal and uni-

where DH is the hydraulic diameter, which is four form flow assumptions are not met. Under these cir- times the hydraulic radius R (Chanson, 1999). cumstances, a uniform flow resistance equation can JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 55

δ δ δ k k ks s s Smooth boundary Transition Rough boundary 훿 Figure 3.8 Hydraulically smooth and rough boundaries, showing ks roughness height and laminar sublayer .In these beds of uniform grain size, ks is effectively the grain diameter. In beds with a range of grain sizes, ks is better approximated by some multiplier of a characteristic grain size (e.g., 3.5D84), as explained in the text. (From Julien 1998, Figure 6.2.)

be used to relate reach-averaged velocity to reach- penetrates closer to the bed (Richards, 1982; Robert, averaged depth and bed or water-surface slope if 2003) but, as noted earlier, the sublayer is typically slopeismeasuredovermultipleobstaclesandbed- <1mmthick. forms, starting and finishing at sections with the The thicknesses of the laminar and turbulent same depth and velocity (Ferguson, 2012). Other- portions of the boundary layer partly depend on wise, S mustbetheenergyline.Theexistenceofpro- the characteristics of the boundary roughness. The truding grains and bedforms also makes estimation effective height of the irregularities forming the

of R very difficult. The most common approxima- rough boundary is the roughness height ks (Fig- tion is to use a cross-sectionally averaged or reach- ure 3.8), and the ratio of the roughness height to

averaged value of R in the flow resistance equations. the hydraulic radius, ks/R,istherelative roughness The Chezy,´ Darcy–Weisbach, and Manning (Chow, 1959). If ks is only a small fraction of the equations are based on the recognition that the thickness of the laminar sublayer, the surface is velocity of water flowing in a channel is influenced hydraulically smooth. If the effects of the rough- by the channel boundary roughness. Within the ness elements extend beyond the laminar sublayer, boundary layer,velocityvariesaccordingtodistance the surface is hydraulically rough and the velocity from the channel surface. Outside the boundary distribution depends on the form and size of the layer, the velocity distribution is practically uniform roughness projections. Because the laminar sublayer (Chow, 1959). In many natural channels, the bound- is commonly extremely thin, and most natural chan- ary layer extends to the water surface. As noted in nels have individual grains and other features that Section 3.1.1, a laminar sublayer can be present at extend beyond this distance into the flow, natural thebaseoftheboundarylayer.Thetopsurfaceof channels have hydraulically rough surfaces. the laminar sublayer corresponds to the transitional Theturbulentboundarylayerisofmostinterest zonefromlaminartoturbulentflowandgives for river processes because it is within this layer that way to the turbulent boundary layer, in which velocity is measured, shear stress is estimated, and the vertical velocity distribution is approximately sedimenttransportislinkedtohydraulicparame- logarithmic. The thickness of the laminar sublayer, ters (Robert, 2003). The law of the wall describes the 훿,isdefinedby variation of velocity with height above the bed sur- face within the turbulent boundary layer, and is used 훿 = 11.6휐∕v∗ (3.14) to derive shear stress and roughness height from where 휐 is kinematic viscosity and v∗ is shear veloc- measured velocity profiles (Robert, 2003) ity, which has the dimensions of velocity and is v = 2.5v∗ ln(z∕z ) (3.16) determined from shear stress 휏 z 0 √ 0 ∗ 휏 휌 where vz is the mean one-dimensional (1D) velocity v = ( 0 ) (3.15) at a height z above the bed, z0 is the projected height The thickness of the laminar sublayer decreases abovethebedatwhichvelocityiszero,andthecon- with an increase in shear stress as turbulence stant 2.5 is equal to 1/휅. JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

56 Rivers in the Landscape

(a)

ks

λ

(b)

ks

λ

(c)

ks

λ Figure 3.9 Longitudinal view of differing types of bed roughness configurations. (a) Widely spaced roughness elements create isolated-roughness flow in which the wake and vortex at each element develop and dissipate before the next element is reached. (b) More closely spaced roughness creates wake-interference flow in which the wake and vortex from each element interfere with those developed at the next roughness element downstream. This creates intense turbulence and complex vorticity. (c) Roughness elements sufficiently close together to create flow that skims the crest of each element is quasi-smooth flow. Each depression between elements has extremely low velocity and a stable eddy. (From Chow, 1959, Figure 8-4, p. 197.)

Thelawofthewallcanbeintegratedfromz = roughness dimensions can be used to represent sur- z0 to z = d,whered is flow depth, to calculate mean facesofvariableroughness,asinchannelswitha 1D velocity at a vertical section via the Keulegan wide range of grain sizes, although much effort has equation been devoted to determining how to most accurately characterize an average. ∗ . v∕v = 2 5 ln(d∕ks) + 6 (3.17) The loss of energy in turbulent flow over a rough boundary results from the formation of wakes where v is the average velocity at a vertical, v∗ is behind each roughness element, so the longitudi- 휆 shear velocity, and ks is roughness height (Robert, nal spacing of the roughness elements is par- 2003). The primary challenge is the determination ticularly important. Isolated-roughness flow occurs

of ks. With densely packed grains of uniform size where individual roughness elements are sufficiently and shape and a flat, immobile bed, ks is approx- far apart that the wake and vortex at each element imately the median diameter of the bed sediment are completely developed and dissipated before the (Robert, 2003). With poorly sorted, heterogeneous flow reaches the next element (Figure 3.9). Rough-

bed sediment and bedforms, ks can vary widely. ness results from form drag on the roughness ele- Commonly used, empirically based approximations ments and friction drag on the wall surface between

include 3.5D84,6.8D50,and2D90,whereDx is the elements. Wake-interference flow occurs where the grain size for which x percentage of the cumulative roughness elements are sufficiently closely spaced grain size distribution is finer. Average values of the for the wake and vortex at each element to interfere JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 57

with the wake and vortex developed at the next ele- ments that cause the water surface to shift to varied ment. This results in intense and complete vorticity flow. (In engineering hydraulics, roughness refers to and turbulence mixing, as well as the highest val- a coefficient that represents the effects of boundary ues of flow resistance. The transition from isolated- shear stress in uniform flow equations, resistance is roughness to wake-interference flow occurs when aforcewithavector(directionalroughness),and the ratio of downstream spacing between obstacles energy dissipation refers to everything that retards (L)toheightoftheobstacles(H)isapproximately flow, including flow transitions such as hydraulic 9–10 (Wohl and Ikeda, 1998). Skimming flow occurs jumps, eddies, and grain and form resistance.) wheretheroughnesselementsaresocloselyspaced Einstein and Banks (1950) used flume experi-

that the flow skims the crests of the elements. Large ments to demonstrate that grain, fg,andform,fb, roughness projections are absent and the surface resistance could be additive acts hydraulically smooth (Chow, 1959). Isolated-

roughness elements can inhibit entrainment of sur- f = fg + fb (3.18) rounding grains and affect bedload motion, and these effects become more pronounced as roughness Grain resistance is negligible in channels formed elements are more closely spaced on the bed. in sand-sized and finer sediment. Individual parti- As noted earlier, resistance to flow along the cles do not protrude into the flow above the rest bedinmostnaturalchannelsresultsfromdiverse of the bed and, except at very shallow flow depths, features. These features are commonly divided into individual particles influence only a small portion three categories (Griffiths, 1987; Robert, 2003). of the flow. Grain resistance can become quite sub- The friction created by individual grains within stantial in channels with larger grains, particularly heterogeneousbedsedimentsisdesignatedgrain where the grains are poorly sorted and are large rel- resistance (orskinorgrainfrictionorgrainrough- ativetoflowdepth.Poorlysortedsubstrateswitha ness). Friction from individual grains organized wide range of grain sizes allow large grains to pro- into bedforms such as , pools and riffles, or trude into the flow above the rest of the bed. Indi- steps and pools is known as form resistance (or form vidual grains may create the greatest resistance in roughness or form drag). Friction also results from shallow, steep, bouldery headwaters, whereas bed- pressure and viscous drag on sediment in transport forms such as step–pool and pool–riffle sequences above the bed surface (Griffiths, 1987). Sometimes (Section 4.3.2) formed in gravel dominate flow resis- a distinction is also drawn between relatively small tance in the middle segments of drainage networks bedforms and large bed undulations that create spill (Prestegaard, 1983). Other sources of flow resistance resistance because of a vertical drop in the water become more important in the deeper, low-gradient surface (Leopold et al., 1960). segments of drainages. An important consideration is that roughness is The importance of grain resistance can be

relatedtouniformflowequations.Anythingthat expressed using R/ks when ks is a function of causesthewatersurfacetoshifttograduallyor,espe- grain size. This ratio is known as the relative grain

cially, rapidly varied flow (such as a large submergence,commonlyexpressedasR/D84,and or a logjam) is an obstruction that does not formally is the inverse of relative roughness. Relative grain fall under the category of form roughness as used by submergence can be used to distinguish large-scale < < hydraulic engineers. In practice, however, any large roughness (0 R/D84 1) in which individual obstacle to flow is lumped under the category of grains protrude a substantial distance into the flow, < < formroughnessinmostgeomorphicdiscussionsof intermediate-scale roughness (1 R/D84 4) and > flow resistance in natural channels, and roughness small-scale roughness (R/D84 4) in which grains is used generically to refer to various forms of flow protrude above the bed only a small proportion of resistance. In the discussion that follows here, “resis- flow depth (Bathurst, 1985) (Figure 3.10). Relative tance” is used to include roughness elements and ele- form submergence, R/H, where H is bedform JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

58 Rivers in the Landscape

(a) (b) 10 100

5 10 1/√ff = 0.82 log (4.35 R/D84)

1 1

0.5

Friction factor 0.1

0.1 Darcy-Weisbach friction factor 0.01 0.05 0.001 0.1 1 10 100 0.5 1 5 10 50 100 Relative grain submergence (R/D84) R/D84

(c)

1 log ff = 1.18 – 0.87 log Q + 0.22 (log Q)2

Friction factor 0.5

1 510 Discharge (m3/s)

Figure 3.10 The relationship between Darcy–Weisbach friction factor and relative grain submergence (R/D84) differs between (a) low-gradient alluvial channels with sand bedforms in which friction factor decreases as grain roughness is drowned out (from Knighton, 1998, Figure 4.2A) and (b) high-gradient alluvial channels with cobble- to boulder- beds in which the range of values for friction factor is relatively constant (from Wohl and Merritt, 2008, Figure 9). Step–pool channels are solid circles, pool–riffle channels are open circles, and plane-bed channels are solid triangles. (c) Friction factor tends to decrease with discharge, as shown in this plot of data from the River Bollin, UK (from Knighton, 1998, Figure 2B), although the rate of decrease varies among channels as a function of the sources of roughness.

amplitude, can be similarly used to express the tanceresultsfromviscousshearoverthegrain importance of form roughness (Wohl and Merritt, surfaces and local pressure gradients around the 2008). grains. Large grains that protrude into the flow cre- The phrases “grain resistance” and “form resis- ateformresistanceaslarger-scaleflowacceleration tance” can be misleading in that individual large and deceleration generate large turbulent eddies in grains actually create form resistance. Grain resis- which energy is dissipated by viscous forces (Fergu- tance is the cumulative effect of many individual son, 2013). Form resistance thus includes the effects grains that retard water flowing over the bed in a of drag on large roughness elements such as pro- uniform manner, without accelerations or deceler- truding boulders or logs, the spill losses over steps ations and with parallel streamlines and the same or riffles, standing waves around protruding obsta- mean depth and velocity (Ferguson, 2013). Resis- cles, and transverse accelerations caused by sharp JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

Chapter 3 Water dynamics 59

changes in flow direction (Ferguson, 2013). Grain tant influence on total resistance and the distribu- and form resistance can be very difficult to distin- tion of hydraulic forces (Wohl et al., 1999; Kean guish in many natural channels. and Smith, 2006a, 2006b). Bank resistance results A widely used approach (Robert, 2003) for par- from: individual grains; bank vegetation; irregular- titioning resistance in natural channels has been to ities such as those produced by slumping of bank

select an appropriate value of ks for grain resistance material or the presence of vegetation; and repetitive andusethisvaluetocalculatefg via the equation variations in the channel planform such as bends. √ Examining Lost Creek, a small (4 m wide), gravel- . . bed channel with rough banks (±20 cm ampli- 1∕ fg = 2 11 + 2 03 log10(d∕ks) (3.19) tude), Kean and Smith (2006a) found that additional Grainresistanceisthensubtractedfromtotal flow resistance created by drag on bank topographic resistance, f,toestimateformresistance.How- features substantially reduced near-bank velocity ever, flume studies that combined grain resistance and shear stress. Neglecting these effects in Lost and form resistance with resistance from individ- Creek resulted in a 56% overestimate of discharge. ual instream wood pieces and jams indicate the exis- Enhanced bank resistance caused by vegetation can tenceofsubstantialinteractioneffectsbetweenresis- alterthespatialdistributionofvelocity,shearstress, tance components. Consequently, a simple additive and sediment deposition, and thus channel geom- approach can be inaccurate (Wilcox et al., 2006). etry (Griffin et al., 2005; Gorrick and Rodr´ıguez, An alternative to assuming that any portion of f 2012). Friction associated with woody riparian veg- etation on the lateral boundaries of the sand-bed not included in fg is fb involves directly calculating RioPuercoinNewMexico,USA,(averagechannel fb width ∼11 m) reduced perimeter-averaged bound- . 휌 2 fb = 0 5Cdh v (3.20) ary shear stress by almost 40% and boundary shear stress in the channel center by 20% (Griffin et al.,

where Cd isadragcoefficientthatvarieswithobsta- 2005). The effects of bank resistance may be more cle shape and h istheheightofbedundulationsasso- important along channels formed in finer sediment, ciated with the bedforms (Robert, 2003). Quanti- suchassand,thaninmorecoarse-grainedcobble-

fying Cd is difficult, however, because form drag is to boulder-bed channels in which bed resistance is influenced by the pressure field, and thus the veloc- very large (Yochum et al., 2012). ityfield,abovethebedforms.Thepressure field is Momentum equations define the hydrodynamic the instantaneous water pressure exerted on a given forces, such as drag, exerted by flow. Houjou et al. surface, such as the bed, within a control volume of (1990) proposed that the velocity field in flows for fluid, expressed as a function of three spatial coordi- whichlateralshearisanimportantfactorcanbecal- nates and time. The velocity field is the instantaneous culatedusingamomentumequationintheformof velocity of the center of gravity of a volume of fluid ( ) ( ) instantaneously surrounding a point in the flow (Fox 휕 휕u 휕 휕u − 휌gS = 휅 + 휅 (3.21) and McDonald, 1978). In other words, Cd for a given 휕z 휕z 휕x 휕x obstacle is not constant, but varies with discharge and with the configuration of the channel geome- where 휌 is fluid density, g is acceleration due to grav- try and bed resistance in the immediate vicinity of ity, S is the downstream water-surface gradient, 휅 is the obstacle, as these influence local velocity and kinematic eddy viscosity, z is the coordinate normal pressure distributions. Quantifying and partitioning to the bed, x is the transverse coordinate, and u is flow resistance, particularly in coarse-grained chan- the downstream velocity component. Houjou et al. nels,continuestobeanelusivegoal. (1990) explained that, in a channel with a rectan- Although more attention has been devoted to bed gular cross section, flow structure reflects width– resistance, bank resistance can also exert an impor- depthratioandtheratioofwallandbedresistance. JWST426-c03 JWST426-Wohl Printer: Yet to Come February 7, 2014 11:12 246mm×189mm

60 Rivers in the Landscape

Equation 3.21 permits computation of velocity and Supplemental Section 3.1.3 provides more dis- stress fields affected by both the channel bed and cussion of resistance estimates and partitioning banks. resistance. Water flowing through a bend experiences a transverse force proportional to v2/r,wherer is the 3.1.4 Velocity and turbulence radius of curvature. This adds to the total flow resis- tance in sinuous channels (Ferguson, 2013). Total Velocity is a vector quantity, with magnitude and head loss along a reach scales with R/r2 (Chang, direction, and one of the most sensitive and vari- 1978). Radius of curvature increases more rapidly able flow properties. Because of the presence of a than v or R as river size increases, so bend losses free surface and friction along the channel bound- decrease in importance with river size and are aries, velocity is not uniformly distributed within insignificant in large rivers for which banks are a channel. Velocity varies with distance from the readily erodible and in which regular can bed, across the stream, downstream, and with time develop (Ferguson, 2013). (Figure 3.11). At a cross section, the maximum

(a) (b)0.8 (c) 0.7 0.6 0.6 d 0.5 d W v 0.4 0.3 Flow depth (m) Logarithmic vertical 0.2 Horizontal velocity velocity profile 0.1 profile 0.0 0.4 0.8 1.2 1.6 2.0 Velocity (m/s)

(d) (e)

v v d1 d1 ≠ d2

d2

Minutes

Downstream velocity Turbulent flow variation velocity variation Figure 3.11 Different characteristics of velocity through time and space. (a) Idealized logarithmic vertical velocity profile (flow from left to right, length of arrows indicates relative velocity); d is flow depth, v is mean velocity. (b) Vertical velocity profile from a boulder-bed channel. Abrupt increase in velocity between 0.3 and 0.4mflow depth reflects top of boulders and transition from highly turbulent flow with low average velocity to rapidflowin upper profile. (c) Horizontal velocity profile looking down on a channel ofwidth w. (d) Longitudinal view illustrating downstream variations in velocity, as reflected in differences within successive vertical velocity profiles. (e) Temporal variations in velocity at short time intervals associated with turbulence. Parts (a), (c), (d), and (e) courtesy of David Dust. Chapter 3 Water dynamics 61

1D flow occurs where velocity is represented by onlyonedimension,suchasvelocityalongthechan- nelcenterlineofalong,straightchannelwithcon- stant cross section, where velocity can be primar- ily influenced by distance above the channel bed. 1D representations have cross-sectionally averaged velocity along the hydraulic axis that follows the . From the continuity equation, mean 1D velocity is discharge divided by cross-sectional area Figure 3.12 Isovelocity lines in differently shaped = channels. In each of these channels, the view is look- (v Q/A). ing downstream within the channel. These are idealized 2D flow occurs when the velocity field is predom- distributions: actual distributions are likely to be less inantly influenced by two of the space coordinates, symmetrical with respect to cross-sectional boundaries such as distance above the bed and from the bank. because of upstream effects. 2D representations of flow have vertically integrated velocity with only the cross- and downstream com- ponents. velocity typically occurs just below the free sur- Thecomplexityofanalysisincreasessubstantially face. The location of the maximum velocity with withthenumberofdimensionsoftheflowfield. respect to distance between the two channel banks Consequently, the simplest case of 1D flow is widely dependsonthesymmetryorasymmetryofthe used to provide approximate solutions for evaluat- channel (Figure 3.12). Velocity tends to remain con- ingfeaturessuchasthalwegvelocityprofile(Foxand stant or increase slightly downstream as resistance McDonald, 1978). from the channel boundaries affects a smaller pro- Flow along channels with low gradients and rel- portion of the total flow volume. Velocity varies at atively uniform bed material is commonly approx- short time intervals of a few seconds as a result imated by a semilogarithmic velocity profile in of flow turbulence, and at longer time intervals which velocity varies with distance from the bed as a result of roughness of the channel bound- (Leopold et al., 1964): this is an example of assum- aries and fluctuations in discharge during unsteady ing 1D flow. The profile includes a laminar sub- flow. layer, a turbulent boundary layer with logarithmic As noted earlier, velocity can be characterized in profile, and an outer layer that deviates slightly terms of the velocity field. At a given instant in time, from logarithmic (Ferguson, 2007). For channels the velocity field is a function of the space coordi- with rough boundaries or irregular bed material, nates x or w (cross stream),yor u (downstream), velocity is proportional to roughness-scaled dis- and z or v (vertical with respect to the bed), so that a tance from the bed, rather than just distance from complete representation of the velocity field is given the bed. In effect, the velocity profile becomes seg- by V = V(x, y, z) (Fox and McDonald, 1978). The mented, with low velocity flow between larger boul- usual convention is for velocity terms u, v, w to cor- ders or below the average surface elevation of the respond to spatial coordinates x, y, z:hencetheuse bed, and an abrupt transition to higher velocity of z for depth or vertical position in some of the flow above the level of the boulders (Figure 3.11b). equations in this chapter. A flow field described in This two-part velocity profile has been described this manner is three dimensional because the veloc- as s-shaped, although the profile varies substan- ityatanypointintheflowfielddependsonthethree tially in relation to channel boundary roughness. coordinates required to locate the point in space. A Thelawofthewallisusedtodescribethevari- flow can be classified as one, two, or three dimen- ation of velocity with height above the bed sur- sional, depending on the number of space coordi- face in the bottom 20% of fully turbulent flows nates required to specify the velocity field. (Robert, 2003). 62 Rivers in the Landscape

Table 3.2 Types of current meters (Clifford and French, 1993b).

Type of meter Characteristics

Mechanical impellor Inexpensive; durable; poor frequency response (<1 Hz), 1D data; requires maintenance and recalibration Electromagnetic Intermediate cost; robust; good frequency response (5–20 Hz); tolerate particle and other contamination in the flow; provide 1D, 2D, or 3D measurements; frequency response affected by head design Ultrasonic Expensive; fragile; sensitive to particle and air bubble contamination; provide 2D and 3D data; excellent frequency response (up to 30 Hz) Laser velocimeters Highest cost; sensitive to suspended sediment; produce very high frequency response; 3D data; do not perturb flow being measured; do not function adequately near the streambed

Supplemental Section 3.1.4 provides more infor- As water particles move in irregular paths, mation about velocity profiles. momentum is exchanged between different por- Velocity in natural channels can be measured tions of the water (Chanson, 1999). Turbulence using current meters or dilution tracers. Point mea- occurs when water parcels flow past a solid sur- surements use various types of current meters, face or past an adjacent water parcel with a different including mechanical impellers, electromagnetic velocity (Clifford and French, 1993a, 1993b). Tur- and ultrasound meters, and laser velocimeters. Table bulence appears as irregular velocity fluctuations. 3.2 summarizes the characteristics of each type of Time series of velocity in natural channels com- current meter. One of the limitations of point mea- monly reveal intriguing variations in the magnitude surementsisthenumberandspatialdensityofmea- of even 1D measurements (Figure S3.1), let alone surements needed to characterize mean velocity. three-dimensional (3D) flow, and a great deal of Acoustic Doppler current profilers can obtain humanenergyisdevotedtocharacterizingthepat- spatially dense point measurements in a very short ternsofvelocityfluctuationsinordertorevealthe period of time, but are very expensive and can be underlying processes and thus improve understand- of limited usefulness in very shallow or highly aer- ing of the relations among resistance, hydraulics, and ated flow. Doppler current meters use the Doppler channel adjustment. Understanding of velocity and effect of sound waves scattered back from parti- turbulence is important because of the feedbacks cles within the water column. The meter generates amongturbulence,channelmorphology,andsedi- and receives sound signals. The traveling time of ment transport (Figure 3.13) (Leeder, 1983; Buffin- sound waves can be used to estimate distance, and Belanger´ et al., 2013). the change in sound wavelength can be converted to Channel morphology creates resistance and con- velocity. tributes to the variability in the intensity of turbu- Dilution tracer techniques using a fluorescent dye lent exchanges with the water column. The structure or a chemical such as NaCl can also be used to of turbulent flows consists of the mean structure, characterize the average velocity for a channel seg- as reflected in the 1D vertical velocity profile, and ment (Wohl, 2010). This approach is inexpensive thetemporalfluctuationsofthestructureassociated and fast, but provides no information on the dis- with turbulence. Both of these elements influence tribution of velocity within the measured channel sediment mobility. Mobile sediments can alter the segment. flow structure and contribute to bedforms, which Supplemental Section 3.1.4 provides more infor- then in turn influence flow resistance and turbu- mation about velocity measurement techniques. lence. Chapter 3 Water dynamics 63

Lower regime Transition Upper regime

f value

Dunes and , superimposed Washed out Antidunes, Ripples Dunes breaking waves Chute and ripples dunes Plane bed standing waves

Increasing roughness Increasing pool

0.052 < f < 0.13 0.042 < f < 0.16 0.02 < f < 0.03 0.02 < f < 0.035 0.03 < f < 0.07 0.07 < f < 0.09

Increasing flow intensity Figure 3.13 Sequence of bedforms that develop with increasing discharge or velocity in mobile-bed channels, and associated flow resistance. Small drawings at the base of the figure are longitudinal views of a segment of bed, indicating the bed profile (gray shading) and the water surface configurations (upper black line). Note that resistance, as indicated by the Darcy–Weisbach friction factor f, generally increases going from ripples to dunes in the lower regime, declines dramatically when the dunes wash out to a plane bed, and then increases less in the progression toward chutes and pools within the upper regime, as indicated by the undulating line partway up the y-axis in this figure. (From Simons and Richardson, 1966.)

Turbulence is quantified based on measurements near-bed region. The greater the value of RMS of velocity fluctuations at a point in the flow. The the greater the turbulence and the greater the turbulence intensity of any of the three velocity potential for sediment entrainment and transport. components (x, y, z) is the average magnitude of The turbulent kinetic energy of the flow, TKE, the deviation from the mean for a given velocity is then series (Robert, 2003). Variability around the mean for a normal or near-normal distribution is rep- . 휌 2 2 2 TKE = 0 5 (RMSx + RMSy + RMSz) (3.23) resented by the standard deviation of the velocity distribution,knownastheroot-mean-squarevalue, where 휌 is water density (Clifford and French, 1993a; RMS, which is used to represent turbulence intensity Robert, 2003). TKE is the mean kinetic energy per (Robert et al., 1996; Robert, 2003). unit mass associated with eddies in turbulent flow, √ and can be conceptualized as the energy extracted √ √ (∑ ) √⎛ ⎞ from the mean flow by turbulent eddies (Bradshaw, √⎜ v2 ⎟ √ y 1985; Robert, 2003). The downstream component RMSy = √⎜ ⎟ (3.22) ⎜ N ⎟ of TKE is typically dominant in natural channels ⎝ ⎠ (Robert, 2003). Turbulencecanalsobestudiedbyfollowingthe where vy isthedownstreamcomponentofveloc- trajectories of tracers traveling with the flow. This ity and N is the total number of observations in a flow visualization approach uses hydrogen bubbles, given series. The RMS value can be computed for colored fluids, or fine particles of neutral submerged anyofthethreedimensionsofvelocityandthethree density (Buffin-Belanger´ et al., 2013). Flow visu- expressions can be combined into an index of total alization does not provide the quantitative infor- turbulence intensity (Robert, 2003). mation of Equation 3.22 or 3.23, but does provide Flow structures in the vicinity of large obsta- insight into the location, size, and structure of tur- cles increase the local turbulence intensity in the bulence. 64 Rivers in the Landscape

Reynolds stresses quantify the degree of momen- moving sand grains and influencing the size of bed- tum exchange at a point in the flow (Robert, 2003). forms (Best, 1992, 2005) (Section 4.3.1). Six Reynolds stresses can be calculated as the prod- Bursts and sweeps occur only a small fraction uctofthenegativevalueofwaterdensity(−휌) of the time, but cause most of the momentum and the product of the average fluctuations of exchange and can strongly influence sediment trans- 2 either a single velocity component (e.g., vx )ortwo port (Robert, 2003). This is the key point regard- different velocity components (e.g., vxvy) (Buffin- ing turbulence in natural channels, and one of the Belanger´ et al., 2013). Like TKE,Reynoldsstresses reasonssomucheffortisfocusedonquantifying are important descriptors of the intensity of turbu- turbulence. Numerous studies of sediment mobility lent exchanges and are linked to sediment transport and associated channel stability indicate that turbu- (Buffin-Belanger´ et al., 2013). Values of RMS, TKE, lent fluctuations exert a stronger influence on sed- and Reynolds stresses are each widely used to quan- iment dynamics, particularly initiation of motion, tify the turbulence characteristics of flow in natural than does mean velocity, especially in channels with channels. beds composed of sediment coarser than sand size. Quadrant analysis is commonly used to under- Turbulent boundary layers include various types stand and describe the structure of turbulence in two of vortices and irregular, but spatially and tem- dimensions. The instantaneous horizontal and ver- porally repetitive, flow patterns known as coher- tical velocities are divided into four quadrants based ent flow structures. Coherent flow structures are on their deviation from the mean (Robert, 2003). self-perpetuating (Smith, 1996; Robert, 2003). Low- Quadrant I represents positive deviation in the hor- speed streaks alternate with zones of high-velocity izontal and vertical. Quadrant II represents nega- water and give rise to coherent flow structures such tivedeviationinthehorizontalandpositivedevia- as horseshoe vortices. Horseshoe vortices,whichare tion in the vertical. Quadrant III is negative devi- namedfortheirshape(Figure3.14),areadvected ationinbothquadrants.QuadrantIVispositive upward from the bed. deviation in the horizontal and negative deviation A second major group of vortices and coherent in the vertical. Quadrant II events have slower than flow structures are present over coarse-grained sur- average downstream flow velocity and positive ver- faces in which vortices are induced by individual tical flows away from the channel boundary and are large particles or bedforms (Roy et al., 1999; Robert, knownasejectioneventsorbursts because water 2003) (Figure 3.15). Eddies are shed from the down- isejectedfromthebedupwardintotheouterflow stream end of obstacles and from the shear layer (Robert, 2003). Quadrant IV events have greater along the flow separation zone downstream from than average downstream flow velocity and nega- obstacles. Standing vortices form upstream of indi- tive vertical flows toward the bed and are known vidual obstacles (Robert, 2003). Horseshoe vortices as sweeps. Alternate zones of low- and high-velocity form from the flow separation zone downstream water near the bed create low-speed streaks that are from isolated obstacles. relatively narrow zones of low-velocity water near The frequency of eddy shedding downstream the bed (Robert, 2003). The spacing of low-speed from isolated obstacles is expressed by the dimen- streaksreflectstheshearvelocityandfluidkinematic sionless Strouhal number, Str viscosity. Low-speed streaks culminate in ejections ( ) upward from the bed (bursts). To preserve continu- feD Str = (3.24) ity, bursts are followed by high-velocity outer layer v flow penetrating the near-bed flow (sweeps). Sweeps lose momentum as they impact the bed and diffuse where fe isthefrequencyatwhicheddiesareshed laterally (Best, 1993; Robert, 2003). This sequence from obstacles, D is obstacle size, and v is mean of bursts and sweeps exerts an important control flow velocity (Robert, 2003). Large Strouhal num- on sand-bed bedforms such as ripples and dunes by bers (∼1) indicate that viscosity dominates fluid Chapter 3 Water dynamics 65

Bursting Low-speed Ejection streak Lift-up

Inrush Sweep re-establishment

Y X Low-speed streak Z Figure 3.14 Schematic representation of turbulent bursting, according to Allen (1985). Flow from left to right, with length of arrows representing relative velocity. A developing horseshoe vortex lifts, stretches, and bursts. Associated velocity profiles shown at rear (from Bridge, 2003, Figure 2.12a, p.29). flow. Low Strouhal numbers (≤10−4)indicatethat the turbulent boundary layer reattaches to the bed the high speed, quasi steady-state portion of the or bank downstream from the obstacle, and the area fluid flow dominates. Intermediate Strouhal num- between the separation and reattachment points is bers reflect the formation and rapid shedding of vor- the zone of recirculating flow (Buffin-Belanger´ et al., tices (Sobey, 1982). 2013). Zones of recirculating flow are typically sites Eddy shedding may be the dominant mecha- where finer sediment is deposited as a result of lower nism of energy dissipation in coarse-grained chan- velocity.Recirculatingflowcanalsoformimpor- nels (Robert, 2003), and may cause the greater-than- tant aquatic habitat by concentrating organic mat- expected flow resistance of coarse-grained channels ter in transport and providing a low-velocity resting (Clifford et al., 1992). Pseudo-periodic oscillations area for organisms such as fish. Strong gradients in in time series of velocity fluctuations may reflect the waterpropertiessuchasvelocityordensitywhere periodicity at which eddies are shed (Robert, 2003). tworiversjoincanalsopromotelateralmomen- Very large roughness elements or an abrupt tum exchanges and large and intense shear layers change in the channel boundary or orientation cre- between the flows (Buffin-Belanger´ et al., 2013). The ate separated flows, or portions of the channel in energy dissipated at these boundaries can facilitate which there is no downstream flow (Robert, 2003). sediment deposition. Examples of sites inducing flow separation include channel bends, channel expansions, pools, bed- 3.1.5 Measures of energy exerted forms, and very large grains. The boundary between against the channel boundaries the lower velocity separation zone and the higher velocity main flow is a zone of rapid change in veloc- The force driving a flow, Fd, along a unit length of ity, intense mixing, and high turbulence intensity channel can be expressed as known as a shear layer (Figure 3.16). The accelera- = sin 휃 tion of flow around the obstacle can allow the turbu- Fd W (3.25) lentboundarylayertodetachorseparatefromthe where W is the weight of the water and 휃 is bed or bank. The reattachment point occurs where the channel bed slope. Weight equals mass times 66 Rivers in the Landscape

Typical eddies Flow (a)

Large-scale motion Large-scale motion

Superburst

(b) Flow

High-speed High-speed region region Low-speed Low-speed region region

High-low High-low interface interface

Flow (c) Fast Low Fast

U’ FastSlow Slow Fast Fast V’ Up Up Down Down Up Quadrant 1 2 3 4 1 Figure 3.15 Longitudinal views of turbulence. (a) Large-scale flow motions with the presence of typical eddies at their boundaries. (b) High- and low-speed regions within the flow. (c) A sequence of turbulent flows associated with the passage of high- and low-speed wedges. (From Buffin-Belanger et al., 2000, Figure 1.)

휃 gravitational acceleration, g,andW sin represents where R is hydraulic radius. For small slopes, sin 휃 is the downslope components of mass acted on by g. approximately equal to tan 휃, which is approximated 휌 Mass equals mass density, ,andthevolumeofwater, by slope, S,sothat V; (mass = 휌V), so 휏 휌 휌 휃 0 = gRS (3.28) Fd = gV sin (3.26) This is the basic equation for average bed shear Stress is force per unit area (A), so the average stress in natural channels (Robert, 2003). Although shearstressexertedbytheflowonthebed,휏 ,under 0 steep channels violate the assumption that bed slope conditions of steady uniform flow is approximates sin 휃, Equation 3.28 is commonly 휏 휌 휃 0 = gR sin (3.27) applied to steep channels. Chapter 3 Water dynamics 67

(a) Spatial variations in stream power are addressed in a model developed by Knighton (1999). Depend- S ingonthedownstreamratesofchangeforQ and R S, the model predicts that total power peaks at an intermediate distance between the drainage divide (b) and the river outlet. Specific stream power is more sensitive to the rate of change in S and peaks closer R to the headwaters. The model accurately predicts S observed conditions along lowland alluvial rivers (Knighton, 1999), but limited tests for mountain rivers reveal substantial deviations from expected Figure 3.16 Examples of flow separation (S is separa- patterns because of local factors that influence Q, S, tion point, R is reattachment point) as indicated by time- and w (Fonstad, 2003). averaged streamline patterns at (a) a downward trans- Allofthemeasuresofstreampowernotedear- verse step, as seen in side view, and (b) a sharp bend in lier are essentially instantaneous values at a point or an open channel, as seen in planform (from Allen, 1994, cross section, and do not account for variations in Figure 2.24). flow, and thus stream power, through time. Tempo- ral variations in stream power relate mainly to varia- Stream power is also commonly used to express tionsindischarge.InonestudyalongtheTapiRiver, energy exerted against the channel boundaries. India, the annual peak flood contributed up to 34% Stream power is the rate of doing work in trans- of the total energy expended during a monsoon sea- porting water and sediment. Stream power can be son (Kale and Hire, 2007). Integrating variations in expressed as total power, Ω stream power through time and along a channel is necessary to effectively understand the influence of Ω=훾QS (3.29) energy expenditure on channel form and process. where 훾 is the specific weight of water (assumed to be 9800 N/m3), Q is discharge, and S is channel gra- 3.2 Hydrology dient. Stream power can also be expressed as stream 3.2.1 Measuring, indirectly 휔 power per unit area, estimating, and modeling discharge 휔 = 휏 0v (3.30) The basic data on river discharge come from gag- ing stations, typically operated by either government 휏 where 0 is bed shear stress and v is velocity. agencies or private companies that require knowl- Finally, stream power can be expressed as specific edge of water supply. In the United States, for exam- 휔 stream power, ple,themajorityofgagesareoperatedbytheU.S. Geological Survey, although state agencies also oper- 휔 =Ω∕휔 (3.31) atemanygages.TheEnvironmentAgencyoper- ates most stream gages in England and Wales. In where w is channel width. these countries, discharge data are publicly acces- Each measure of stream power quantifies the sibleonlineatnocharge.Governmentagencies energy available to perform geomorphic work in other countries, such as India, consider stream against the channel boundaries in the form of records sensitive information and will not readily entrainment and erosion. release the data. 68 Rivers in the Landscape

Although the continuity equation (Q = wdv) stations, but many tributaries have few or no gages. suggestsasimplemeanstomeasuredischargeat Under these circumstances, a number of techniques agage,w, d,andv are not measured continuously. canbeusedtoindirectlyestimatethedischargeofa Instead, they are initially measured over a range discreteevent,suchasaflood,ortoestimatestatis- of flows and used to establish a rating curve for tical characteristics of discharge through time, such the cross section. This curve relates Q to stage as mean annual flow. (water-surface elevation), which is in turn related In regions with reasonably complete spatial and to d, the most readily measured parameter. A rating temporal coverage of discharge gages, the exist- curve is typically recalibrated periodically to ensure ing records can be composited to create regional that gradual, progressive changes in cross-sectional discharge–drainage area relations.Theserelations geometry do not affect the accuracy of discharge can then be applied to estimate flows at ungaged sites calculations. This method of calculating discharge based on the drainage area of the ungaged site (San- is most accurate where cross-sectional geometry born and Bledsoe, 2006; http://streamstats.usgs.gov/ is relatively simple, without split flow or extensive index.html). Related to this is the practice of storm backwaters, and stable through time. Cross sections transposition,inwhichaparticularlyextremestorm with bedrock exposures or banks or bed stabilized for which records exist at one gage site is used to by concrete or other artificial materials are chosen estimate the discharge that would result at another for gaging sites where possible. site if such a storm occurred (Changnon, 2002). The simplest technique for measuring flow depth This type of extrapolation can be inaccurate where is to install a calibrated vertical scale along one bank discharge–drainage area relations are nonlinear as a and read depth from this scale. This technique was result of changes in hydroclimatology with elevation originallydevelopedforusealongtheNileRiver or changes in rainfall–runoff relations with geology circa 620 AD. Most gages now use a stilling well or land cover (Pitlick, 1994). in which the stage equals height of the water sur- The most extreme precipitation input physically face in the channel. A pressure sensor in the well likely at a given site, the probable maximum precip- is connected to instruments that record fluctuations itation, is defined as the greatest depth of precipita- in the stage through time, using something as sim- tion for a given storm duration. This value is used pleasapaperstripchartandastylus,orsend- to estimate the resulting probable maximum flood, ing the electronically transduced measurements to which is mandated for design of structures such as a central receiving and recording station via satellite dams where failure of the structure would result in transmission. Discharge can be measured at regular loss of human life. intervals, the length of which typically reflects some In the absence of direct measurements at dis- tradeoff between accuracy and cost. Various statisti- charge gages, flow can be indirectly estimated using cal measures such as mean daily, mean monthly, and several approaches. Most of these are based on a mean annual discharge can be computed from these rearrangement of the Manning equation as presented data. earlier Supplemental Section 3.2.1 provides further 1 2 1 information on gaging techniques and illustrations Q = AR 3 S 2 (3.32) n of different types of gages. Direct discharge measurements are relatively lim- where A is cross-sectional area, R is hydraulic radius, ited. Some large rivers have continuous records S is channel gradient, and n is the Manning rough- extending back more than a century. More com- ness coefficient. The Manning equation is based monly, records are of short duration or with time on the assumption of steady, uniform flow so that gaps,andincludeonlypartialspatialcoverageofa slope, discharge, and velocity are constant with time river network such that the main channel and a lim- and space along a segment of channel. This assump- ited number of tributaries have one or more gaging tion may not apply very well to flow in natural Chapter 3 Water dynamics 69 channels, particularly during high flows. Nonethe- TheManningequationcanalsobeusedtoesti- less, the commonly used indirect methods of mate the magnitude of discharges that occurred slope-area and step-backwater computations in the distant past, if cross-sectional geometry and assume steady, uniform flow and use 1D hydraulic water-surface profiles are preserved. Information theory as their basis (Webb and Jarrett, 2002). on water-surface elevation can come from histor- Both the slope-area and step-backwater meth- ical records created by people, including physical ods use the conservation of mass, conservation of marks on buildings or bedrock channel walls, or energy, and Manning equations to calculate flow. diaries, journals, damage or insurance reports (Gur- Slope-area uses known water-surface elevations to nell et al., 2003). compute discharge (Dalrymple and Benson, 1967), Information on water-surface elevation can also whereas step-backwater uses discharge to com- come from botanical and paleostage indicators. pute stage (O’Connor and Webb, 1988). Both of Botanical records offlowstagetaketheformof these approaches, and use of the Manning equa- damage to riparian trees caused by flood-borne tion, assume that a water-surface profile can be used debristhatleavesascaronthetree(Huppand with surveyed cross-sectional geometry to calculate Bornette, 2003). Paleostage indicators in the form R as an approximation of flow depth. The accuracy of erosional or depositional records of maximum of this calculation depends on the representative- stage can be used to reconstruct a water-surface ness of the surveyed cross-sectional geometry and profileatsitessuchasbedrockcanyonswherethe water-surface profile. Geometry and water-surface cross-sectional geometry changes relatively slowly elevationcanvaryalongariverreach,andcanvary (Wohl and Enzel, 1995; Jarrett and England, 2002; through time, even during a single flood, because Benito and O’Connor, 2013) (Table 3.3). Using of scour and fill of the bed, rapid changes in flow, each of these types of information, A, R, and S substantial sediment transport, and flow transitions are directly measured from existing channel geom- between subcritical and supercritical (Jarrett, 1987; etry and high-water marks, and n is estimated, Sieben, 1997). allowing at least an approximate value of Q to be Variations of parameters through time are par- calculated. ticularly problematic when the Manning equation Other indirect methods can be used to estimate is used to calculate flood discharge. The accuracy unrecorded discharge. Regime-based reconstruction of indirect flood discharge estimations based on the uses sedimentary features to reconstruct the cross- Manning equation typically declines in at least four sectional geometry of relict channels preserved on scenarios. The first scenario is rivers with very short afloodplainorincutbanksofthecontemporary duration and high magnitude discharges that do not channel (Jacobson et al., 2003). Channel geometry have steady, uniform flow. This scenario is charac- is then related to a relatively frequent flow, such teristic of smaller catchments, particularly in arid as mean annual flood, using an approach such as regions and the tropics. The second scenario is rivers the Manning equation. Competence estimates use the with seasonal ice in which much of the runoff dur- averageclastsize,bedforms,orbeddingstructure ing the season when the ice is melting occurs on top to estimate past mean flow, or use the largest clasts of the ice (Priesnitz and Schunke, 2002). A third sce- likely to have been transported by fluvial processes nario that limits accuracy of indirect flood discharge to estimate the shear stress, stream power, or veloc- estimations is on small, steep rivers with very rough ity of the associated peak flow (Wohl and Enzel, boundaries (Bathurst, 1990). Finally, indirect dis- 1995). In addition to the scars mentioned earlier, charge is difficult to estimate on rivers such as sand- several other characteristics of riparian vegetation bed channels that can undergo substantial changes canbeusedtoinferpastflowandchanneldynamics in cross-sectional geometry during a flood as a result (Table 3.3) (Yanosky and Jarrett, 2002). oferosionofthebedontherisinglimbanddeposi- Supplemental Section 3.2.1 provides more tion on the bed on the falling limb. information on the use and relative accuracy of 70 Rivers in the Landscape

Table 3.3 Types of paleoflood and paleoflow boundary conditions (Sieben, 1997). Deterministic indicators. approaches include a variety of models, such as the Category Types and information U.S.ArmyCorpsofEngineers’HEC-1modelthat estimates discharge resulting from rainfall. Historical Written records of date and extent and/or depth Hydrologic models of rainfall–runoff processes of flood at the watershed scale depend on input data. Input Photographs of flood data can include topography of the catchment sur- Maps of flood extent face and parameters that are spatially distributed Physical marks recording peak stage over (land cover, land use) and beneath (soil tex- Botanical Vegetation structured by age or type Impact scars record flood stage and date ture and moisture) the surface. Input data can also Adventitious stems or split-base sprouts record include the spatial characteristics of the river chan- flood date nel network (e.g., drainage density) and channel Adventitious roots indicate burial by overbank geometry.Manyofthesedataareacquiredusing sedimentation remote sensing and manipulated using GIS soft- Exposed roots indicate date of bank erosion ware. This scale of hydrologic modeling presents Variations in tree-ring width and symmetry numerous challenges because catchment-scale mod- indicate high and low flows els integrate data derived from a wide range of Geologic Regime based: channel dimensions related to sources, at a range of spatial scales and resolution, mean flow and over a period of time (Downs and Priestnall, Competence: average clast size, bedforms, 2003). Understanding the sources and magnitude of bedding structure related to mean flow; uncertainty in the diverse input data is particularly maximum clast size related to peak flow Paleohydraulic: depositional or erosional important for evaluating the accuracy and represen- indicators of hydraulics (e.g., lateral gravel tativeness of these models. berms, berms at downstream end of plunge Data on the spatial distribution of basic hydro- pools, lateral potholes) logical variables such as rainfall and runoff can be Stage indicators: scour lines, lichen limits, combined with digital terrain analysis to estimate truncation of landforms impinging on channel discharge at varying points in a channel network (e.g., alluvial fans), silt lines, organic debris, and to create spatially explicit water budgets (Mont- boulder bars, slackwater sediments gomery et al., 1998). Marks and Bates (2000) pro- vided an example of combining different types of models and data sets for modeling flood parameters. different techniques for indirectly estimating Working on the River Stour in the United Kingdom, ungaged discharge. they used a two-dimensional (2D) model of flow in Different measures of discharge, from flood peak the channel and across the floodplain with LiDAR- flow to average annual discharge, can also be derived topographic data to simulate the extent of numericallysimulated.Theuseofexistingdischarge flood inundation, as well as local values of hydraulic records to estimate flow at ungaged sites is based on parameters such as flow depth and velocity. statistical analyses and the assumption that repre- Another approach to numerically modeling sented trends are consistent and continuous and can floods involves combining gaged and ungaged therefore be extrapolated across space and time, as streamflows with physical and statistical data well as extrapolated to the very large flows that may to develop probability models for all uncertain notbepresentinthegagerecords. parameters (Campbell, 2005). Hydrologists accept An alternative to the statistical approach of esti- the fact that hydrologic model prediction is not mating discharge is a deterministic approach based deterministic, but rather must explicitly include on a mechanistic concept used to simulate precipi- an estimate of uncertainty. Uncertainty in model tation inputs and discharge outputs under imposed predictions reflects measurement errors for input Chapter 3 Water dynamics 71 and output parameters, model structural errors 350 St. Vrain Creek at Lyons, Colorado from the aggregation of spatially distributed real- 300 (drainage area 600 km2) world processes into a mathematical model, and 250 errors of parameter estimation. Bayesian model /s) 3 200 selection uses the rules of probability theory to select among different possibilities, automatically 150 choosing simpler, more constrained models. This 100 approach can be applied to rainfall–runoff models Discharge (m 50 (Sharma et al., 2005) or to models of streamflow, 0 among other things. Both types of models can also be run using a Generalized Likelihood Uncertainty 0.1 1 10 100 1000 Estimation scheme or GLUE (Beven and Binley, Recurrence interval (years) 1992). The GLUE framework was one of the first Figure 3.17 Sample flood–frequency curve: here, for attempts to represent prediction uncertainty. This St. Vrain Creek in Colorado, USA (drainage area 600 km2). framework conditions the parameter distributions The typical annual peak flow results from snowmelt, but and generates prediction uncertainty envelopes that at much longer recurrence intervals, convective storms incorporate parameter uncertainties (Wyatt and can generate very high magnitude flash floods. The two Franks, 2006). A great deal of attention is given to outliers in the upper right of this plot are rainfall flash floods that occurred in 1941 (largest flood) and 1919. spatial variability in parameters and processes and to uncertainty in parameter estimation (Sivapalan et al., 2006). Although direct measurements of flow stage at duration series that includes all floods above a spec- gaging stations include errors and uncertainty, these ified discharge. The partial duration series is likely data remain the standard against which indirect to be more important geomorphically, particularly if estimates, statistical extrapolations, and numerical the specified discharge has some physical relevance simulations are validated. Changing priorities in to sediment transport or channel morphology. government expenditures in countries such as the Thesimplestwaytocalculaterecurrenceinter- UnitedStates,however,continuetoresultindeclin- val T (years) for either annual maximum or partial ing numbers of active gages. Estimation of discharge duration series is via the Weibull formula using satellite-based remote imagery may help to fill this widening gap in discharge data, but such T = (n + 1)∕N (3.33) approaches are still being developed (Legleiter et al., 2009; Flener et al., 2012). where n is the number of years of record and N is the rank of a particular flood when floods are ranked 3.2.2 Flood frequency analysis from largest to smallest (Dalrymple, 1960). T can then be plotted against Q to create a flood–frequency The frequency with which a flood of a given mag- curve (Figure3.17).Inanyoneyear,theprobability nitude recurs is critical to understanding the phys- of the largest flow exceeding a flood with a recur- ical and ecological significance of floods, as well as rence interval of 10 years is 1/10. A flood with an mitigating hazards associated with flooding. Flood- averagerecurrenceintervalof10yearsiscommonly frequency analysis uses different techniques to relate referred to as a 10-year flood, but this type of abbre- flood magnitude to recurrence interval. The record viated description can be very misleading. There of floods at a site can be considered in terms of is no physical reason that a 100-year flood cannot an annual maximum series, which includes only the occurduringtwoconsecutiveyears;itisimproba- largest discharge in each year of record, or a partial ble, but not impossible. 72 Rivers in the Landscape

There are no strong theoretical justifications for ingfromconvectivethunderstormsandsnowmelt applying any particular statistical distribution to floods, or flash floods and longer duration floods hydrologic data. Log-Pearson Type III is commonly caused by dissipating hurricanes. Analyses that sep- used in the United States because it is a skew dis- arate populations with different flood-generating tribution bounded on the left and therefore of the mechanisms result in improved parameter estimates general shape of most hydrologic distributions. The of the component distributions and a better under- recommended procedure for using the log-Pearson standing of the physical basis for extreme floods distribution is to convert the data to logarithms and (Hirschboeck, 1987). computethevalueofafloodofspecifiedprobability, The presence of nonstationarity is a primary X, for any probability level as concern for any form of flood-frequency analysis. Stationarity is the assumption that any statistical property of the flow record, such as mean annual log X = log X + K휎 (3.34) x flow or mean annual flood peak, has a probability density function that does not vary through time. where X is the mean of the flood series, K is obtained This assumption facilitates extrapolation of trends 휎 from a standard table, and x is the standard devia- in hydrologic data forward in time and prediction of tion of the series. The Gumbel extreme value distri- the magnitude of future 100-year floods, for exam- bution is commonly used in the United Kingdom. ple, based on the magnitude of 100-year floods in the Regardless of the distribution used, treating the past, or in a situation where the record length is less highest known discharges for a site is particularly than 100 years. problematicbecausethesedischargestendtobe Assumptions of stationarity can be incorrect if outliers that do not follow trends present in the the particular period of record represents some remainderofthedata.Historicalandpaleoflood deviation from longer-term averages. Precipitation information can be combined with systematic gage can exhibit long-period variability such that more measurements of discharge to extend the length zonal or meridional circulation patterns charac- of the flow record at a site. The historical and terize periods up to several decades in length at paleoflood data must be treated differently than regional to continental scales (Hirschboeck, 1988). the systematic data because they represent what is More meridional circulation tends to produce more known as a censored record: only floods above a mag- severe flooding, so using 20 years of record from nitude threshold are recorded, rather than all flows a period of meridional circulation to predict future being recorded, as at a stream gage. Early work used flood frequency might overestimate the recurrence threshold-exceedance maximum likelihood esti- intervalofaparticularfloodmagnitude.Othercli- mators, which employed the number of floods that matic circulation patterns can produce fluctuations exceeded a known threshold, without differentiating in flooding at timescales of decades (Webb and the magnitude of each flood (Stedinger and Baker, Betancourt, 1990) to centuries and millennia (Ely 1987). Subsequent work relies on moments-based et al., 1993; Benito et al., 1996; Redmond et al., 2002) parameter estimation procedures (Cohn et al., 2001; across broad regions. Human alterations of rainfall– England et al., 2003) or simply uses field evidence of runoff relations as a result of changing land cover or paleoflood magnitude to quantify nonexceedance flow conveyance in river networks can also compro- bounds for discharge over a time interval defined mise the assumption of stationarity. Finally, ongoing using various geochronologic techniques (England global climate change associated with CO2-induced et al., 2010). warming is rendering the assumption of stationar- Another difficult issue for flood-frequency analy- ity inappropriate in many regions. Efforts are under- sis is the existence of mixed distributions.Manysites way to develop nonstationarity probabilistic models include floods produced by more than one hydrocli- (Milly et al., 2008), but no widely used method is yet maticmechanism—forexample,flashfloodsresult- available. Chapter 3 Water dynamics 73

3.2.3 Hydrographs evenly over the drainage basin) (Rodriguez-Iturbe and Valdes, 1979). A geomorphologic unit hydro- A hydrograph is a curve of discharge plotted against graph relates the unit hydrograph to the morpho- time. An event or flood hydrograph represents a sin- logic parameters of the river network and defines the gle flood. An annual hydrograph represents flow travel time distribution of water particles to the out- overthecourseofanaverageyearasderivedfrom let of the basin (Rodriguez-Iturbe and Valdes, 1979). averages of some flow interval (typically, minutes Anyhydrographincludesbaseflowanddirect to 1 day) over the period of gage records. A unit runoff, a rising limb, a peak, and a falling limb (Fig- hydrograph is based on runoff volume adjusted to ure 3.18). Base flow results primarily from ground- a unit value (e.g., 100 mm of precipitation spread water inputs to a channel and is the stable, low

(a) 8

Snowmelt Rainfall

/s) 6 3

4

2

Average daily discharge (m 0

0 100 200 300 Days since 1 January (b) 8

Peak (6.94 m3/s) 6 /s) 3

Rising Recessional 4 limb limb

Discharge (m Runoff 2

Base flow

0 0 50 100 150 200 250 300 350 Cumulative time (hours) Figure 3.18 Sample hydrographs. (a) Average annual hydrographs based on average daily discharge over the period of record for a snowmelt-dominated stream (North St. Vrain Creek, Colorado, USA; drainage area 91 km2)anda rainfall-dominated stream (Cedar Creek, South Carolina, USA; drainage area 161 km2). (b) Flood hydrograph for a single rainfall-generated flood on Gills Creek, South Carolina; drainage area2 167km ). 74 Rivers in the Landscape flow to which a river returns following precipita- regime. Ignoring other influences such as drainage tion inputs. Direct runoff results from the combined basin size, snowmelt runoff tends to be spread inputs of overland flow and flow in the unsaturated over a longer time and to produce a peak flow zone (Section 2.2.1). The rising limb is the portion of longer duration and lower peak than rainfall of the hydrograph where discharge increases as a runoff. Different types of rainfall produce differently result of direct runoff inputs, the peak represents the shaped hydrographs. Convective rainfall, because of maximum discharge within a particular timespan, its greater intensity and short duration, tends to and the falling or recession limb reflects progressive produce a more peaked hydrograph than cyclonic declines in direct runoff with time until discharge rainfall. Rain-on-snow, as the name implies, occurs returns to base flow. when rain falls on a snowpack. The resulting runoff The term flashy is commonly applied to hydro- depends on the magnitude of the rain and the graphs with short duration peaks and rapid rise and water equivalent and spatial extent of the snow- recession. Baker et al. (2004) proposed a flashiness pack, and can be quite large but of shorter dura- index tion than snowmelt (McCabe et al., 2007). Because ∑ n | | of these characteristic differences in event and |q − q | i=1∑ i i−1 annual hydrographs, a river’s flow regime is typically Richards–Baker Flashiness Index = n i=1 qi described as being snowmelt dominated or rainfall (3.35) dominated.Thesecategoriesarenotmutuallyexclu- sive because different types of flooding can occur where q is mean daily flow, qi is flow on a particu- during different seasons or over a period of many lar, and qi−1 is flow on the following day. Although years in the same region. For example, a snowmelt- many drainages are described as being flashy, the cri- dominated hydrograph can have secondary peaks teria used for this designation vary widely and are associated with late summer convective storms, or not necessarily comparable between studies. a rainfall-dominated hydrograph can have predomi- The lag time or basin lag indicates the delay nantly frontal cyclonic precipitation with occasional between the center of mass of precipitation inputs hurricane rainfall. and the center of mass of streamflow. Lag time In addition to describing flow regime in terms reflects the manner in which precipitation is trans- of the dominant type of runoff, flow regime can be mitted from hillslopes into the channel, as well as characterized by its spatial and temporal continuity. thesizeofthecatchment,spatialarrangementof Perennial flow is continuous through time and channels in the river network, and geometry of val- space: some level of flow is always present through- leys and channels. Lag time for a given type of pre- out the river network so designated. Intermittent cipitation input increases with basin size because flow is spatially discontinuous such that some direct flow must travel longer distances along hill- portions of a river contain flow while other portions slope paths before reaching a channel, and longer are dry (Figure S3.20). This situation can arise distances within the river network before reach- where a channel crosses into a different climate, as ing the basin outlet. Lag time is lower for equant- when perennial river segments flowing from moun- shaped basins than for linear basins. Equant-shaped tainous highlands with large annual precipitation basins tend to produce larger, shorter peak dis- enter drier lowlands and the surface flow evaporates charges because of efficient concentration at the or infiltrates. Intermittent flow can also result from basin outlet (Strahler, 1964). Lag time also increases longitudinal contrasts in subsurface permeability, as where broad valley bottoms and floodplains allow when flow in a channel crossing a thick, permeable direct runoff to move relatively slowly downstream alluvial layer infiltrates and moves downstream in in overbank areas or multiple channels. the subsurface, returning to the surface channel Hydrograph characteristics can also be strongly where an impermeable layer close to the surface influenced by the hydroclimatology,orprecipitation forces the water upward once more. Ephemeral flow Chapter 3 Water dynamics 75 is temporally discontinuous, with periods of surface charge during floods as a result of transmission flow shortly after precipitation inputs to the basin losses from infiltration into unconsolidated alluvial interspersed with periods of no flow. channel beds, and evaporation and transpiration, as Intermittent and ephemeral flows are most likely well as a common absence of appreciable tributary to occur in dry environments such as arid and semi- inflows in the lower parts of many dryland river arid regions of the polar, temperate, and tropical networks (Tooth, 2013). Maximum flood peaks and latitudes, and in very small catchments that may discharge per unit drainage area can be quite large have limited base flow inputs. Intermittent flow in dryland rivers, particularly where small, steep canalsooccurinkarst terrains where underground catchments and abundant low-permeability surfaces drainage networks developed through chemical dis- associated with bedrock or crusted soils are present. solution of carbonate rocks capture surface flow This situation can result in highly skewed flood– in some portions of a network. Surface–subsurface frequency distributions and steep flood–frequency exchanges in karst environments can produce blind curves that reflect a high ratio of large to small flows valleys that end suddenly where a stream disappears (Knighton and Nanson, 2002; Tooth, 2013). underground and exsurgence where an underground stream with no surface headwaters reaches the sur- face. 3.2.4 Other parameters used to The spatial and temporal extent and magnitude characterize discharge of flow also influence the connectivity of river net- works. Ecologists distinguish riverine connectivity, Discharge measurements that are continuous during which indicates spatial linkages within rivers, and a year and over several years can be used to estab- hydrologic connectivity,whichisthewater-mediated lish a flow–duration curve in which mean discharges transport of matter, energy, and organisms within over a specified time interval (e.g., daily mean) are orbetweenelementsofthehydrologiccycle(Free- grouped into selected classes of discharge magnitude man et al., 2007). Both forms of connectivity are vital and plotted against the percentage of time that class to maintaining the ecological integrity of riverine is equaled or exceeded (Figure 3.19). The shape of ecosystems, where ecological integrity is the undi- the flow–duration curve reflects the drainage basin’s minished ability of an ecosystem to continue its nat- response to precipitation inputs. The steeper the ural path of evolution, its normal transition over flow–duration curve, the more quickly storm runoff time, and its successional recovery from perturba- enters the channel. tions (Westra et al., 2000). Atsiteswithmeasurementsofsuspendedsedi- The typical scenario is that discharge and occur- ment concentration at varying discharges that allow rence of perennial flow increase downstream within construction of a sediment rating curve,theflow– a river network as contributing drainage area duration curve can be used to develop a cumulative increases.Iftherateofincreaseindischargewith sediment transport curve indicating how flows of dif- drainage area is known, the more easily measured fering magnitude and recurrence interval contribute drainage area can be used as a surrogate for dis- to sediment transport during an average year. charge.Thismaynotbethecase,however,indry- Ecologists developed the flood pulse concept for lands (semiarid and arid regions). Low annual pre- large, lowland rivers in recognition of the impor- cipitationandhighevaporationindrylandsresult tance of the magnitude, timing, duration, and in low annual runoff, and the interannual variabil- ratesofriseandfallofannualorseasonalfloods ityofrunoffincreasesinthedriestregions.Spa- to aquatic and riparian ecosystems. Natural, pre- tial variability of precipitation and runoff limit the dictable floods—the flood pulse—that inundate at useofdrainageareaasasurrogatefordischarge least a portion of the floodplain are associated with in drylands (Tooth, 2013). Instead, most dryland much higher biological productivity.The flood pulse river networks exhibit downstream decreases in dis- provides clear, shallow water in overbank areas 76 Rivers in the Landscape

10 North St. Vrain Creek, Colorado (91 km2) /s) 3

1 Discharge (m

0.1 0 20 40 60 80 100 Percent of time flow equaled or exceeded discharge

10 Cedar Creek, South Carolina (186 km2) /s) 3

1 Discharge (m

0.1 0 20 40 60 80 100 Percent of time flow equaled or exceeded discharge Figure 3.19 Sample flow–duration curves for North St. Vrain Creek, Colorado (snowmelt flow regime) andCedar Creek, South Carolina (rainfall flow regime). for primary, photosynthetic production. The flood of a channel (Tockner et al., 2000). Although these pulse also creates fish nursery habitat and feeding, fluctuations do not cause overbank flow, they change as well as more abundant and diverse habitat for a theextentofflowinsecondarychannelsandinareas variety of other organisms. Inundation of overbank of flow separation along a single, confined chan- areas and then subsequent recession of streamflow nel and provide some of the same biogeochemical into channels during the flood pulse facilitates the exchanges and habitat abundance and diversity as exchange of nutrients, organic matter, and organ- overbank flows. isms between the channel and floodplain (Junk et al., A third key concept in riverine ecology is the 1989; Bayley, 1991). natural flow regime (Poff et al., 1997). This phrase Ecologists also describe flow pulses that are fluc- refers to the magnitude, frequency, duration, timing, tuations in surface waters below the bankfull level andrateofchangeofflowconditionsintheabsence Chapter 3 Water dynamics 77 of human interference through processes such as greater precipitation inputs equate to higher river flow regulation. Channel geometry and aquatic and stageandgreaterdischarge.Thisisnotnecessar- riparian communities are typically adjusted to the ily the case in high-latitude rivers with strongly natural flow regime so that altering this flow regime developed seasonal ice cover. A river-ice season is likely to cause corresponding alterations in the that lasts more than 100 days between autumn physical and ecological characteristics of the river. freeze-over and spring breakup characterizes many Management of rivers with flow regulation increas- rivers of high latitudes, including some high- ingly emphasizes protecting or restoring at least elevation or interior rivers as far south as 42◦ Nin some approximation of the natural flow regime, as North America and 30◦ N in Eurasia (Prowse and discussed in Chapter 7. Beltaos, 2002). The hydraulic resistance added by an ice cover elevates river water levels, particularly 3.2.5 Hyporheic exchange when the ice cover is hydraulically rough during and hydrology freeze-over and breakup. River ice can modify the quantity of flow in achannelviathreemechanisms(Prowseand The hyporheic zone is the portion of unconfined, Beltaos, 2002; Ettema and Kempema, 2012). The near-stream aquifers where stream water is present. first involves reduction of contributing area, when Hydrologists define this zone as a flow-through sub- anchor ice frozen to the river bed cuts off ground- surface region containing flow paths that originate water inflow. The second involves direct storage of and terminate at the stream. Gooseff (2010) pro- water in river ice, which typically involves slower posed that the hyporheic zone be defined based abstraction during freeze-over and rapid re-supply on the timescale of flow, analogous to definitions during breakup. The third mechanism, which can of floodplains based on inundation frequency, such produce the most significant effect, is hydraulic stor- that there are 2–24-hour hyporheic zones. The age in the channel. This involves increased water hyporheic zone can extend up to 2 km laterally levelsbecauseoftheincreasedhydraulicresistance fromtheactivechannelalongriverswithbroad, caused by an ice cover. The latter effect can be gravel floodplains (Stanford and Ward, 1988) (Fig- particularly significant when ice jams form during ure S3.21). Hyporheic flow may constitute only breakup. a minor portion of stream discharge (<1%) in Ice-jam frequency and the resulting floods are steep, headwater channels with limited alluvium highly variable from year to year (Boucher et al., (Wondzell and Swanson, 1996), but can be 15% or 2012) as a result of the strong influence of interan- more of surface discharge in larger, lowland alluvial nual fluctuations in air temperature. The reduction rivers (Laenen and Risley, 1997). In the latter case, in flow created by the three mechanisms described ignoring hyporheic flow can lead to errors in dis- earlier can be equivalent to nearly 30% of the flow charge estimation, particularly given that the char- that would otherwise occur at a particular site acteristics of the hyporheic zone vary widely in space (Prowse and Carter, 2002). Release of this water dur- and time. ing breakup can account for nearly 20% of the spring Supplemental Section 3.2.5 provides more infor- peak flow (Prowse and Carter, 2002). Because of mation about the processes of, and controls on, these effects, early winter freeze-over can create the hyporheic exchange. lowest discharge of the year, even though runoff is lowest during late winter, and breakup frequently 3.2.6 River hydrology in establishes the annual maximum water levels, even cold regions though maximum discharge is more likely to result from spring snowmelt or summer rainfall later in the Hyporheic and groundwater exchange can certainly year (Prowse and Ferrick, 2002). One implication of complicate the relations between precipitation the presence of ice is that drainages with ice cover inputs and outputs of river flow, but in most basins, may not exhibit the direct relationship between flow 78 Rivers in the Landscape and stage that exists in most rivers (Prowse and and other industrialized countries have equally high Beltaos, 2002). rates of flow alteration by dams. Flow regimes in high-latitude rivers can be dis- Dams vary in their design and operation. Dams tinguished as nival, proglacial, wetland, and prola- that store water for flood control or water supply, for custrine (Woo and Thorne, 2003). Snowmelt and example, release the stored water in large, sustained river ice breakup generate the largest flows in nival pulses. Dams designed for hydroelectric power gen- regimes.Highflowsareprolongedintothesummer eration typically have substantial daily fluctuations by glacial melt water in proglacial regimes;someof in water releases to maximize power production the highest flows occur in mid to late summer when during periods of high demand. The typical effect of glacier melt is most pronounced. Rivers below wet- any type of dam, however, is to homogenize regional lands also have prominent snowmelt peaks because river flow regimes (Poff et al., 2007) by reducing the wetlands have low storage capacity while frozen, peak flows and reducing the variability in lower but once the ground thaws, the wetlands can retain flows. water and retard summer flows in wetland regimes. Flow diversion can either reduce flow in the The influence of large lakes creates a fairly even source stream or augment flow in the receiving runoff throughout the year in prolacustrine regimes, stream. In extreme cases, all water is removed from although the outflow channel of an arctic lake is thesourcestreamduringatleastaportionofthe more likely to be blocked during freeze-over and year, or the peak flow is more than doubled in the inflow to the lake can be very small, minimizing receiving stream. Although flow diversion has his- winter outflow. toricallybeenmostcommononsmall-tomedium- size rivers, massive diversion projects on very large rivers are now underway or being planned in China 3.2.7 Human influences and parts of Africa and the Middle East. on hydrology Direct alterations of channel form that change flow conveyance and at least some characteristics Human activities can alter flow regimes in chan- of the hydrograph take several forms. Construction nels indirectly through changes in climate, topog- of levees prevents or limits overbank flows, increas- raphy, and land cover that influence the amount ing the magnitude and velocity of water conveyed and downslope pathways of water entering channels within the channel during higher discharges. Chan- from uplands (Supplemental Section 2.2.1). Direct nel engineering in the form of bank stabilization, human alterations of streamflow involve changes in dredging, or straightening (channelization) typically thevolumeandtimingofflowasaresultofflow increases cross-sectional areas, reduces boundary regulation,andchangesintheabilityofchannelsto resistance, and limits exchanges between sediment convey flow downstream. Flow regulation includes in transport and sediment in the channel bed and dams and diversions. banks. All of these alterations typically result in Dams can be small barrages or weirs that are fre- larger magnitude but shorter duration flood peaks. quently overtopped during moderate to high dis- Removal of beavers (Castor fiber in Europe, Cas- charges, but serve to pond water upstream during tor canadensis inNorthAmerica)andtheirsmall lowdischarges.Attheoppositeendofthesizescale, dams typically reduces overbank flooding and asso- largedamsarethosetallerthan15mandmajor ciated floodplain wetlands, and increases flood peak dams are taller than 150 m. More than 45,000 large magnitude and velocity by confining flood dis- dams and more than 300 major dams currently exist. charges to a main channel. Alteration of riparian More than half (172 of 292) of the world’s large river vegetation occurs either through removal of veg- systems are affected by dams (Nilsson et al., 2005). etation or through introduction of exotic species Only about 2% of the total river kilometers in the with substantially different characteristics than United States are not affected by dams (Graf, 2001), thenativespecies.Thesealterationschangebank Chapter 3 Water dynamics 79

Table 3.4 Types of direct human influences on streamflow.

Type of influence Effects Sample references

Flow regulation Dams typically reduce the mean and coefficient of variation of annual Graf (2006) peak flow, increase minimum flow, shift the seasonal variability, and increase diurnal fluctuations; diversions typically reduce base flows Ryan (1997), Wohl and and flood peaks in the source stream and increase base and peak Dust (2012) flows in the receiving stream Channel engineering Levees prevent or limit overbank flow and thus typically increase the Kondolf (2001) magnitude and decrease the duration of peak flows Bank stabilization can decrease channel complexity and increase conveyance Channelization (straightening and dredging) increases channel Shankman and Pugh (1992) conveyance and typically increases the flashiness of streamflow Alterations of Removal of riparian vegetation decreases near-bank and overbank Anderson et al. (2006a) riverine biota roughness, can decrease bank stability, decreases transpiration Introduction of exotic riparian vegetation can change vegetation density Dean and Schmidt (2011) and function, typically with increased density and roughness that can facilitate channel narrowing Removal of beaver decreases channel–floodplain connectivity, increases Wohl (2001) flow velocity and peak flow magnitude, decreases elevation of alluvial water table

stability, near-bank and overbank roughness, and Our activities alter the water reaching Earth’s sur- transpiration rates (Wohl, 2000a). Table 3.4 summa- face, alter how that water moves downslope and rizes the effects of different types of direct human into channels, and alter how the water moves down- influences on streamflow. stream through river networks. These cumulative Human effects on nearly all aspects of river form alterations are extremely intensive and widespread and function are now nearly ubiquitous. Sediment globally. fluxes are very difficult to quantify accurately as a Supplemental Section 3.2.7 provides more details result of spatial and temporal fluctuations within and example case studies of indirect and direct each drainage basin, as well as incomplete systematic human alterations of streamflow. measurements of sediment flux in most basins, but global-scale estimates have been published. People have simultaneously increased the sediment trans- 3.3 Summary port by global rivers through soil erosion (Hooke, 2000) by more than 2 billion metric tons per year Equations used in basic engineering hydraulics start (of a total estimated 14 billion tons per year prior from assumptions that flow in a channel is uniform to human alteration), but reduced the flux of sedi- and steady, and that the channel has relatively ment reaching the world’scoasts by 1.4 billion metric smooth boundaries and limited turbulence. Natural tonsperyearbecauseofretentionwithinreservoirs channels seldom meet these assumptions, and the behind dams (Syvitski et al., 2005). Global manip- degree to which flow is varied and unsteady deter- ulations result in a doubling to tripling of the resi- mines how far actual hydraulic characteristics devi- dence time of continental runoff and a 600%–700% atefromexpectedconditions.Theinstabilitiesasso- increase in fresh water stored in channels within ciated with hydraulically rough boundaries, flow reservoirs (Vor¨ osmarty¨ et al., 2004). transitions, and spatial and temporal variations in 80 Rivers in the Landscape velocity and turbulence are inherent in natural chan- channels—are typically quite poorly characterized nels, as are the interactions between flowing water using standard assumptions from engineering and the channel boundaries that result in sediment hydraulics. movement and channel adjustment. A substantial The energy of flow quantified in hydraulic equa- amount of research is focused on quantifying and tions inherently reflects the volume of water mov- predicting hydraulic variables such as resistance ing down a channel, so a substantial amount of (n or f), velocity, turbulence intensity, shear stress, research also focuses on quantifying and predicting and stream power because these variables can be discharge. Systematic, gage-based measurements of used to predict sediment movement and channel discharge seldom provide sufficient length and spa- geometry. Our ability to accurately characterize tial density of record to infer the characteristics of hydraulics in natural channels is greatest in those very large, infrequent flows. A variety of indirect channels that best approximate the assumptions methods are used to estimate the magnitude and underlying engineering hydraulics—sand-bed recurrence interval of a wide range of flows, from channels with readily deformable boundaries and rare floods to the shape of an average annual hydro- simple geometry. Natural channels with more graph. Human activities have directly and indirectly hydraulically rough boundaries, more complex and altered river discharge for centuries in much of the spatially variable geometry, and greater erosional world, making it challenging to infer a natural flow thresholds—gravel-bed, boulder-bed, and bedrock regime in the absence of human manipulation. Channel processes II

Chapter 4 Fluvial sediment dynamics

This chapter examines the movement of particulate Some rivers, particularly sand-bed channels, have and dissolved sediment within river channels. The relatively mobile beds in which sediment moves fre- chapter is divided into four main sections. The first quently and sometimes at relatively high rates, at section describes the channel bed and initiation of least in terms of grain numbers. These channels are sediment motion from the bed. Within this part of also referred to as live-bed or regime channels. In the chapter, the first subsection addresses a station- contrast, rivers with a bed dominated by gravel and ary streambed and the methods used to character- larger particles are threshold channels in which bed ize bed sediment. The ability to characterize the bed sediment moves relatively rarely and at low rates. substrate is important to understanding one source This section of the chapter ends with a description of of sediment supply when sediment is in transport, as bedforms in cohesive sediment, and processes and well as sources of flow resistance that influence the forms of sediment deposition within the channel distribution of hydraulic forces moving sediment. boundaries. The next two subsections discuss the initiation of The third section of the chapter discusses fac- motion of individual particles on a non-cohesive tors that control bank stability and processes of bank bed, and the removal of aggregates of particles from erosion. Bank erosion gets a great deal of atten- acohesivebed. tionbecauseofthepotentialforpropertylossand The second section of the chapter focuses on damage to infrastructure as a result of bank failure. the processes of sediment transport and deposition. Erosion of banks is distinct from that of streambeds Sediment can be transported downstream in solu- for at least two reasons. First, bank erosion com- tion, as particles suspended in the flow, or as par- monly occurs via mass failures, as well as grain- ticles that remain in contact with the bed. Each by-grain detachment from the banks. Second, bank mode of transport is influenced by similar hydraulic stability and erosion can be strongly influenced by forces, but is described by a distinct suite of pro- vegetation. cesses and equations. Sediment moving in contact The final section examines sediment budgets, with the bed is differentiated between readily mobile which use information on sediment input and stor- and infrequently mobile bedforms. This division age to quantitatively estimate fluxes of sediment recognizes a fundamental distinction among rivers. past a given point in a watershed. Sediment yield is

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

83 84 Rivers in the Landscape characterized by great spatial and temporal variabil- Volumetric samples of finer sediment can also be ity, with much of the sediment coming from a lim- combined with in situ measurements of coarser sedi- itedportionofadrainagebasinduringarelatively ment in a bed with mixed grain sizes (Bunte and Abt, short period of time. Globally, a disproportionate 2001). amount of sediment originates in small, mountain- In situ methods for characterizing bed sediment ous drainages. canemploydirectmeasurementsofclastsizealong a grid or a random walk, as originally proposed by Wolman (1954). In situ approaches can also employ systematic subsampling, in which all par- 4.1 The channel bed and ticlesexposedatthesurfacewithinadefinedarea are measured, or are removed from an area of the initiation of motion bed using an adhesive such as clay or wax: removal only works where particles are small enough to be 4.1.1 Bed sediment lifted from the bed (and subsequently sieved) using characterization an adhesive layer. In situ methods in medium to large rivers increasingly use Fourier or other spec- Knowledge of the size distribution, packing, sorting, tral analysis of particle outlines on digitized pho- shape of particles, and stratigraphy of bed sediments tographs (Supplemental Section 4.1.1). is important for understanding the initiation of Much effort has gone into determining the mini- motion of particles on the bed, sediment transport, mum sample size necessary to accurately character- and channel change. Accurately measuring the char- ize bed sediment using either bulk or in situ sam- acteristics of the bed sediments becomes more diffi- pling. The default method remains the 100 clasts cult as these sediments become more coarse grained originally proposed by Wolman (1954), which is and heterogeneous, not least because progressively usually sufficient for measures of the central ten- larger samples are necessary to accurately character- dency such as mean grain size. Fripp and Diplas ize the grain-size distribution. Consequently, a great (1993), Rice and Church (1996), and Bunte and Abt deal of effort has been devoted to measuring the (2001) described how to determine the minimum surface and subsurface bed sediments in gravel-bed sample size based on the size fraction of interest rivers. and the level of precision required. The desired pre- The grain-size distribution of bed sediments can cision depends on the intended use of the grain- be measured by taking a bulk sample and sieving size distribution, but commonly results in a sample the sediments in a laboratory. This works well for size closer to 300 clasts. Measurements repeated at a pebble-size or finer sediments, but quickly becomes site through time with the intent of monitoring the impractical with coarser sediments because of the percentage of finer sediment on the surface of a large sample size needed. As the intermediate axis gravel-bed stream as an indicator of fish or macroin- of the largest clast approaches 64 mm, sample sizes vertebrate habitat, for example, may require greater required to accurately characterize the distribution precision than a reach-averaged estimate of bed exceed 400 kg (Church et al., 1987). Samples in roughness based on grain size. which the weight of the largest clast is <1% of the Once the grain-size data are obtained, they can total weight can provide unbiased estimates of mean be used to develop a cumulative frequency distribu- grain size (Mosley and Tinsdale, 1985), although this tion of frequency by weight or frequency by number makes sampling coarse sediments almost entirely (Supplemental Section 4.1.1). Either type of data can impractical. In situ measurements of clast size can be then be used to estimate Dx,wherex is the percent- more practical in terms of avoiding sediment collec- age of the grain-size distribution that is finer than tion, but such methods may not consistently sample the reference size (e.g., D50 is the grain size for which clasts <15 mm in diameter (Fripp and Diplas, 1993). 50% of the distribution is finer). Chapter 4 Fluvial sediment dynamics 85

In addition to size distribution, bed sediment can coarse particles touch one another in an interlocked be characterized in terms of density, shape, sorting, bed,whichisframework supported,whereascoarse and packing. Grain density is the mass per unit vol- particles are surrounded by finer sediment in a umeofthesediment,whichisusuallyassumedto matrix-supported bed. Open-framework gravels lack be 2.65 g/cm3—the density of quartz. (Bulk density finer sediment in the pore spaces between individual of aggregates of grains is more typically in the range gravel particles. The distinction between framework of 1.4–1.8 g/cm3.) Grain density affects the fall or supported and matrix supported can be important settling velocity, the velocity at which a grain moves in contexts such as aquatic habitat. A framework- down a water column. Specific gravity is the density supported or open-framework bed allows water, of the sediment relative to the density of water (2.65 dissolved oxygen, and nutrients to circulate more gcm−3/L g/cm3 = 2.65). readily around fish eggs laid in the streambed, for Grain shape includes form and surface texture of example, than does a matrix-supported bed. a sediment grain. Shape affects fall velocity: flatter Additional information about characterizing bed particles fall more slowly than spherical particles of sediment is included in Supplemental Section 4.1.1. similar weight and density. Shapes can be described as spheres, blades, discs, and rods, and with respect to roundness and sphericity (Folk, 1980). Shape fac- tors such as grain elongation (the ratio of the long 4.1.2 Entrainment of non-cohesive axis to the intermediate axis) can also be quantified sediment (Robert, 2003). The shape of a grain influences the surfaceareaexposedtotheflowandtheeasewith Most grains of sand size and coarser (Table 4.1) on which the grain can be rolled along the bed. a streambed are non-cohesive. Entrainment refers to Sorting describestherangeofparticlesizes the initiation of motion of sediment. Similarly, incip- present. Well-sorted sediments have a narrow range ient motion describes the threshold condition when of particle sizes. Various measures of sorting exist, the hydrodynamic moment of forces acting on a par- including the widely used inclusive graphic standard ticle balances the resisting moment of force (Shields, deviation (Folk, 1980): 1936; Julien, 1998). Being able to predict when sed- iment will be entrained is critical to understanding 𝜙84 − 𝜙16 𝜙95 − 𝜙5 bed erosion, movement of bedforms, maintenance 𝜎 = + (4.1) I 4 6.6

𝜙 where x is the phi size for which x%ofthedistri- Table 4.1 Grain-size categories. 𝜙 bution is finer grained: =−log2x,wherex is grain 𝜎 < 𝜙 Category Size range (mm) Phi unitsa (𝜙) size in millimeter. Values of I 0.35 are very well sorted, 0.35–0.5𝜙 are well sorted, 0.5–0.71𝜙 mod- Boulder ≥256 −8to–12 erately well sorted, 0.71–1𝜙 moderately sorted, 1– Cobble 64–256 −6to−8 2𝜙 poorly sorted, 2–4𝜙 very poorly sorted, and >4𝜙 Gravelb 2–64 −1to−6 extremely poorly sorted. Sandc 0.062–2 4 to −1 Packing describes the organization of particles on Silt 0.004–0.062 8–4 thebed.Packingcanbequalitativelycategorized Clay ≤0.004 ≤8 as imbricated, interlocked, and open (Figure S4.3). a 𝜙 Phi units are calculated from =−log2x,wherex is grain size in mm. Imbricated clasts lie with their long axes parallel bGravel is sometimes divided into pebble (4–64 mm, −2to−6phi) to flow and dipping slightly up-current. Imbrica- and granule (2–4 mm, −1to−2phi). c tion makes it more difficult to entrain clasts. Imbri- Sand is sometimes divided into very coarse sand (1–2 mm, 0 to −1 phi), coarse sand (0.5–1 mm, 1–0 phi), medium sand cated clasts can provide useful paleocurrent indi- (0.25–0.50 mm, 2–1 phi), fine sand (0.125–0.25 mm, 3–2 phi), and cators when preserved in stratigraphy. Individual very fine sand (0.0625–0.125 mm, 4–3 phi). 86 Rivers in the Landscape andstabilityofaquatichabitat,sedimenttransport, size distribution and flow roughness in diverse chan- and hydraulic roughness. nels produce a range of values for the Shields num- Studies of entrainment initially focused on well- ber. (3) The use of reach-averaged conditions results sorted, sand-bed channels, and these remain the in a greater range of values. channels for which entrainment can be predicted It is also important to distinguish between the most accurately. As cohesion between particles apparent Shields number and the actual Shields increases in grains finer than sand, or in bedrock, number. The apparent Shields number is calculated bedsedimentislesslikelytodetachasindividual using Equation 4.2. The apparent Shields num- particles. Under these conditions, substrate char- beriswidelyusedbecauseitisderivedfromrela- acteristicssuchasporosityandpermeabilityor tively simple measurements of the reach-averaged bedrock jointing influence entrainment. As the shear stress and grain-size distribution. The appar- grain-size distribution becomes wider in coarse- ent Shields number, however, is relatively simplis- grained channels, effects such as packing, sorting, tic because entrainment is typically not spatially or shielding of finer particles, protrusion of coarser temporally uniform and thus is not accurately repre- particles, and particle shape influence entrainment. sented by a single shear stress for a particular grain, Modern studies of entrainment in sand-bed or a single grain size for a streambed (Yager and channels start with Shields (1936), who conducted Schott, 2013). The actual Shields number of a par- flume experiments on the initiation of motion in ticular grain reflects the force balance acting on the channels with relatively uniformly sized sand grains. grain, which depends on the local particle arrange- Shields quantified entrainment using dimensionless ment and flow turbulence. Any given grain size thus 𝜏∗ critical shear stress, c , calculated by dividing the hasarangeofpossiblecriticalshearstressvalues critical shear stress by an approximation of the (Figure 4.1). weight of an immersed grain, Research on entrainment in beds of mixed grain sizes focuses on the force balances acting on grains, 𝜏 ∗ c the grain properties, and the turbulence param- 𝜏 = (4.2) c 𝜌 𝜌 ( s − )gD eters that influence motion. Biotic processes that influence entrainment have also recently received 𝜏∗ 𝜌 where c is the dimensionless critical shear stress, s increasing attention. is the sediment density, 𝜌 is the water density, g is the acceleration due to gravity, and D is the grain size.

D50 is typically used for D when the grain-size distri- bution is relatively narrow. This approach, although Potential developed for sand-bed channels, is now also the motion ) starting point for studying entrainment across a 2 Motion 1000 much wider range of bed grain sizes. No movement Dimensionless critical shear stress is commonly Movement known as the Shields number. This empirically 100 derived number typically varies between 0.03 and 0.08 for channels with S < 0.03, but can include No values between 0.01 and 0.2 (Buffington and motion Bed shear stress (N/m 10 Montgomery, 1997). The wide variation in Shields number results from at least three factors (Fergu- son, 2013; Yager and Schott, 2013). (1) Diverse stud- 10 100 1000 ies use different criteria for defining the initiation Grain size (mm) of grain motion and different methods of measuring Figure 4.1 Bed shear stress versus grain size, showing the initiation of motion. (2) Differences in the grain- range of entrainment (from Williams, 1983, Figure 1). Chapter 4 Fluvial sediment dynamics 87

Velocity Lift force, FL

Buoyancy, FB

Drag force, FD

Δv Velocity difference between top and bottom of grain creates a vertical pressure gradient Resisting force, Fr Weight, Fg Figure 4.2 Schematic illustration of the forces acting on a non-cohesive grain under steady uniform flow on a nearly horizontal surface.

Forces acting on a grain Reynolds number describes the ratio of kinetic to viscousforcesappliedonamovingparticle),shape, The balance of forces acting on a grain is schemat- size, orientation, and relative submergence; 𝜌 is the ically illustrated in Figure 4.2. The downstream water density; is the square of the fluid and upward driving forces acting on the grain are velocity averaged over the grain; 𝜏 is the reach- the downstream component of particle weight F , b g averaged boundary shear stress; f(z/z ) is a func- buoyancy F ,liftF ,anddragF .Thefrictional o B L D tionthatdeterminestheformofthevelocitypro- resisting forces acting on the grain are F ,thebed- r filearoundthegrain;andA is the cross-sectional perpendicular component of particle weight, and grain area, which varies with relative submergence the friction angle. Entrainment occurs when the and protrusion. driving forces exceed the resisting forces. Each of Lift force results from differences in velocity and these forces reflects the local turbulence and particle pressure at the top and bottom of a grain and is characteristics, including diameter, drag coefficient, friction angle (resisting pocket angle of a grain), ( ) 1 2 and protrusion of the grain into the flow (Yager and F = 𝜌AC v2 − v2 L 2 L T B Schott, 2013). ⟨ ( ) ( )⟩ 1 z z For spatially and temporally averaged flow condi- = AC 𝜏 f 2 T − f 2 B (4.4) 2 L b z z tions, the forces acting on a grain can be derived as 0 0 follows (Wiberg and Smith, 1987; Yager and Schott, 2013). Drag force results from flow of water past a where zT and zB are the heights of the top and bot- graininthedownstreamdirectionandis tom of the grain, respectively; CL is the lift coeffi- cient; and v and v are the flow velocities of the ⟨ ( )⟩ T B top and bottom of the grain, respectively. The larger 1 ⟨ ⟩ 1 z F = 𝜌AC v2(z) = AC 𝜏 f 2 the velocity and pressure gradients between the top D 2 D 2 D b z 0 and base of grains, the more important the lift force (4.3) becomes. The force exerted by particle weight is where CD is the drag coefficient and is a func- ( ) 𝜌 𝜌 tionoftheparticleReynoldsnumber(theparticle Fg = s − gV (4.5) 88 Rivers in the Landscape where g is the gravitational acceleration and V is Grain properties the particle volume. This force is the difference ( ) The grain properties that influence entrainment between the grain weight 𝜌 gV and buoyancy ( ) s include grain size, shape, orientation, packing, 𝜌 gV ,whichistheamountoffluidweightdis- w and imbrication. As noted previously, the original placed by the grain of interest. The net gravita- Shields curve of dimensionless shear stress and sed- tional force is divided between the down slope (driv- iment transport rate was empirically derived for rel- ing force) and bed-perpendicular (resisting force) atively uniformly sized sediment. Larger grains have components. ahighercriticalshearstressbecauseoftheirgreater Thedrivingforcescanthenbewrittenas weight, and weight-driven transport represents one end-member for entrainment from a bed of mixed 𝜙 𝛽 FD + FL tan + Fg sin (4.6) grain sizes. Equal mobility transport represents the other end-member. where 𝜙 is the friction angle and 𝛽 is the channel Equal mobility canbedefinedasthesamecrit- 𝜏 slope. The friction angle is the angle through which ical shear stress ( c)forallgrainsizes(Parker 𝜏∗ agrainwillpivotwhenitisdislodgedinthedown- et al., 1982). Under this condition, c decreases with stream direction. increasing grain size because coarser particles pro- The resisting force is given by trude into the flow and are easier to move than on a bed of uniform grain size, whereas finer particles arehiddenandthereforehardertomovethanona 𝛽 𝜙 Fg cos tan (4.7) bed of uniform grain size. Equal mobility can also be defined as occurring when the grain-size distri- For entrainment, Equations 4.6 and 4.7 can be bution of the bed load equals that of the channel bed equated, Equations 4.3 and 4.4 can be substituted (Parker and Toro-Escobar, 2002). for FD and FL,respectively,andtheequationcanbe Under conditions of intermediate mobility solvedforthecriticalshearstresstoyield between the two end-members of weight-driven and equal mobility, fine sediment is likely to be more 𝜏 2V (tan𝜙cos𝛽 −sin𝛽) c = ⟨ ( )⟩ ( [ ] ) mobile than coarse sediment, but less mobile than (𝜌 − 𝜌)gD F s 2 z L 𝜙 itwouldbeonabedofuniformfineparticles.This CDAD f 1 + tan z0 FD is known as selective entrainment (Parker, 2008). (4.8) The competing effects of grain weight and hid- ing can be represented by hiding functions that esti- mate the influence of mixed sediment sizes and flow This equation requires an assumed velocity pro- conditions on entrainment, as in this example from file. The logarithmic profile is commonly used. Parker et al. (1982) Other representations of the force balance use dif- ferent forces and alter the parameterization of each ( ) 𝜏∗ D b force (Yager and Schott, 2013). Bagnold (1977), for ci = i 𝜏∗ (4.9) example, proposed using stream power per unit area c50 D50 rather than shear stress to evaluate entrainment and 𝜏∗ 𝜏∗ transport. where ci and c50 are the dimensionless critical Although Equation 4.8 addresses the basic forces shear stresses for the ith grain size (Di)andthe acting on a grain, it is in practice cumbersome to median grain size (D50), respectively, and b indicates apply to real channels with a range of grain sizes. the relative importance of grain weight and hiding This reflects the substantial scatter in actual, site- effects. When grain weight dominates, b = 0. When specific values of individual forces associated with hiding effects dominate and create equal mobility, grain properties and flow turbulence. b =−1 (Yager and Schott, 2013). Chapter 4 Fluvial sediment dynamics 89

Because hiding functions are derived from site- Spatial and temporal variability in bed texture specific empirical measurements, they are not trans- and hydraulic forces make it effectively impossible ferable between locations (McEwan et al., 2004). to predict exactly when a particular grain will be Accounting for hiding effects when trying to quanti- entrained. Entrainment of a particular grain or a tatively predict entrainment becomes progressively population of mixed grain sizes is instead commonly more important, however, as the grain-size distri- described in terms of a range of critical values for bution grows wider and entrainment of the smallest velocity, shear stress, or some other parameter. grains is more influenced by hiding. In addition to grain size, grain protrusion, p, Biotic processes grain exposure, e,andfrictionangle,𝜙,influence entrainment. Protrusion is the distance a particle Biotic processes can also influence entrainment protrudes above the mean bed elevation. Exposure (Riggsbee et al., 2013). Instream wood creates large- is the distance a grain extends above or below scale hiding effects and alters the distribution of the neighboring grains. Friction angle,alsoknown pressure and velocity near the bed. Microbial mats as pivoting angle, would be about 30—32 degrees and macroinvertebrates such as net-spinning cad- for uniform sediments (Robert, 2003). In hetero- disfly larvae can increase the cohesion of sand- geneous sediments, the angle through which a size and finer sediments, and thus increase the grain must pivot to be entrained varies significantly, shear stress necessary for entrainment (Statzner depending on the size distribution of the surround- et al., 1996). Algae and microorganisms can facil- ing grains. Protrusion, exposure, and friction angle itate deposition of calcium carbonate, which lim- can vary with relative grain size, shape, and packing, its sediment mobility and enhances the formation and are typically empirically estimated. of quasi-permanent steps and pools (Marks et al., Beds with increasingly wide distributions of grain 2006). Bioturbation by invertebrates and the build- size have increasingly larger effects from protrusion, ingofspawningmoundsordepressionsbyfishcan exposure,andhiding.Large,protruding,immobile alter surface grain-size distribution and topography grains can be particularly important in creating (Statzner and Sagnes, 2008). higher values of dimensionless critical shear stress The magnitude of these effects varies in part in steep, shallow flows. The additional resistance with the abundance of the biota. Instream wood created by these grains and by immobile bedforms was historically widespread and abundant along reduces the shear stress available for grain entrain- rivers in forested regions and undoubtedly strongly ment, so that the Shields number increases with influenced sediment dynamics, but centuries of slope (Ferguson, 2013). removing wood from channels has diminished these effects. Bed sediment dynamics are substantially influenced by spawning activities in rivers that still Turbulence have abundant salmon populations (Hassan et al., Near-bed turbulent bursts, sweeps, and inward and 2008), although such rivers are now rare. outward interactions can influence entrainment by inducing fluctuations in the pressure and velocity fields around a particle and thus influencing drag 4.1.3 Erosion of cohesive beds and lift forces (Nelson et al., 1995). The relationship between local turbulent forces and grain entrain- Channels formed in cohesive material can have a ment is quite complex and sediment motion can thin, continuous or discontinuous veneer of uncon- be dominated by drag, lift, or some combination of solidated sediment along the bed. The cohesive both forces (Schmeeckle et al., 2007). Both magni- material underlying this alluvium limits the rate tude and duration of turbulent fluctuations influ- and manner of bed and bank erosion and exerts ence entrainment (Diplas et al., 2008). a strong influence on cross-sectional geometry 90 Rivers in the Landscape and hydraulics during large discharges because the Quarrying refers to the entrainment of blocks boundary is not readily erodible. Cohesive mate- at least partially detached from the surrounding rial in this context includes channel boundaries with bedrock, typically along joints. Lift force is par- sufficient silt and clay to create strong interparticle ticularly important in quarrying, and the large lift cohesion and limit the detachment of individual silt forces generated in shallow, swift flow can effec- or clay particles, and bedrock. Cohesive boundaries tivelyquarryblocksmuchgreaterthan1macross are not necessarily more resistant to erosion than (Wohl, 1998; Whipple et al., 2000; Chatanantavet alluvial boundaries. A very soft siltstone or sand- and Parker, 2009; Dubinski and Wohl, 2012). Quar- stone may have a lower erosional threshold than rying can be the dominant mechanism of bedrock an alluvial channel formed in boulders. In general, channel erosion in strongly jointed bedrock. however, cohesive material has a higher erosional Abrasion is abrasive erosion of bedrock by clasts threshold than non-cohesive alluvium. carried in the flow, and macroabrasion is sometimes Because individual particles are not readily used to refer to the combined effects of particle detached from cohesive material, sediment entrain- impacts that fracture bedrock into fragments that mentisnottheonlyprocessbywhichchannels can be quarried (Chatanantavet and Parker, 2009). in cohesive beds erode, although sediment entrain- The effectiveness of abrasion depends on (i) the ment can be an important component of other ero- relative hardness of the clasts in transport and the siveprocesses.Theprimaryerosiveprocessesin cohesive boundaries, (ii) the amount of sediment indurated, cohesive material (bedrock) are corro- in transport, and (iii) the frequency and duration sion, cavitation, abrasion, and quarrying (Whipple of flows during which clasts able to create abra- et al., 2013). sionareintransport.Abrasionhasanonlinearrela- Corrosion is the chemical dissolution of bedrock. tion with the amount of sediment available. Very Most chemical weathering occurs outside the chan- low sediment supply limits the tools for abrasion, nel, where water moves more slowly through or past but high sediment supply can cover and protect the rock matrix. Chemical weathering can be effec- the bed from abrasion (Sklar and Dietrich, 2004) tive in eroding carbonate rocks along channels, how- (Figure S4.4). ever, and in weakening other rock types, particularly Bed erosion in cohesive materials typically does those with carbonate cement (Springer et al., 2003). not occur evenly across a cross section or through- Chemical weathering can thus make the bedrock out a reach, and bedrock channels seldom have a more susceptible to processes of physical erosion planar bed. Instead, erosion tends to occur most (Hancock et al., 2011). actively over a fraction of the bed width that Cavitation involves shock waves generated by the varieswithbedloadsupplyandtransportcapacity collapse of vapor bubbles in a flow with rapidly fluc- (Finnegan et al., 2007). This results in features such tuatingvelocityandpressure.Theminuteirregular- as inner channels and variable bed gradients (Baker, ities along channel boundaries caused by joints or 1978; Wohl, 1998). crystal or grain boundaries in the bedrock can facil- Just as alluvial bedforms such as ripples and itate the formation and collapse of vapor bubbles. dunesbothreflectandinfluenceboundaryrough- Although this effect might sound insignificant, mil- ness,velocity,andturbulence,sofeedbacksamong lions of bubbles imploding and sending out shock boundary roughness, rate of erosion, velocity and waves during a flood can very effectively weaken turbulence, and sediment mobility influence the and erode bedrock, a phenomenon common on the location and style of erosion in bedrock channels concrete spillways of dams (Barnes, 1956; Eckley (Goode and Wohl, 2010b). A distinctive result of and Hinchliff, 1986; Wohl, 1998). The importance of localized erosion is the creation of sculpted forms cavitation relative to the processes of quarrying and such as potholes, grooves, and undulating walls abrasion remains largely unquantified, however, for (Richardson and Carling, 2005). The location and natural channels. dimensions of some sculpted forms can be used Chapter 4 Fluvial sediment dynamics 91 to infer hydraulics during formative flows (Wohl, mobilization of soluble material that accumulates 2010). prior to the flood (Walling and Webb, 1986). Rivers Equations for processes of bedrock erosion, and with strongly seasonal flow regimes typically have further information on processes of bedrock ero- an annual flush of high dissolved loads, as in the sion, are in Supplemental Section 4.1.3. rising limb of a snowmelt-dominated river. Erosive processes in cohesive but non-indurated A global survey of 370 rivers indicates that solute materials, such as silt and clay, are similar to concentration, C, varies with discharge, Q,as those in bedrock channels, but are also influenced by upward-directed seepage and matric suction. C = aQb (4.10) Upward-directed seepage is seepage directed verti- cally upward. This leads to static liquefaction, which where b is mostly 0 to −0.4, with a mean value of creates an additional driving force. Matric suction −0.17 (Walling and Webb, 1986). occurs when negative pore water pressure above the As noted in Chapter 2, the primary constituents − water table increases the apparent cohesion of a soil; ofdissolvedloadarethedissolvedionsHCO3 , 2+ 2− − + 2+ + this is also called matrix suction. Matric suction cre- Ca ,SO4 ,H4SiO4,Cl ,Na ,Mg ,andK ;dis- ates an additional force of resistance to bed erosion solved nutrients N and P; dissolved organic matter;

(Simon and Collison, 2001). dissolved gases N2,CO2,andO2;andtracemetals Much of the research on erosion of cohesive, non- (Berner and Berner, 1987; Allan, 1995). The sum indurated materials focuses on stream banks rather of the concentrations of the dissolved major ions is than the streambed, and is discussed in Section 4.3. known as the total dissolved solids (TDS). An average natural value for rivers is 100 mg/L, and pollution 2+ − adds on average another 10 mg/L. Ca and HCO3 from limestone weathering tend to dominate TDS 4.2 Sediment transport (Berner and Berner, 1987). Because dissolved fluxes in rivers partly reflect rates of continental weather- 4.2.1 Dissolved load ing, which in turn reflect rates of tectonic uplift, the largest contemporary fluxes occur in rivers drain- The sediment carried in solution within a river is ing the Himalayan and Andean mountains and the sometimes ignored as a component of sediment Tibetan Plateau (Raymo et al., 1988). transport, but can be substantial in some rivers. A Reach-scale dissolved load can also strongly survey of the world’s large rivers indicates anywhere reflect interactions between surface and hyporheic from 2% of the total load carried in solution in the water. Downwelling into the hyporheic zone trans- Huang He River of China to 93% in Canada’s St. fers solutes and surface water rich in dissolved oxy- Lawrence River (Knighton, 1998). gen, whereas upwelling flow can transfer nutrients to Solute concentration is highest in water entering a streams (Tonina and Buffington, 2009). Hyporheic channel through subsurface pathways in which gen- exchange exposes solutes in stream water, includ- erally slower rates of flow provide longer reaction ing nutrients, to alternating anoxic and oxic zones in times with the surrounding matrix. Solute concen- the bed that are composed of geochemically reactive tration typically declines with increasing discharge sediments and microbial communities (Lautz and as water moving more rapidly and along surface flow Siegel, 2007). Anoxic zones that indicate the pres- pathsentersariver.Totaldissolvedloadcontinuesto ence of sulfate, iron, and manganese reduction typ- increase with discharge, but much of the dissolved ically occur upstream of streambed structures, as in load is carried by relatively frequent flows. low-velocity pools. Oxic zones that reflect the pro- Soluteconcentrationalsodisplayshysteresis duction of nitrate are typically downstream of bed effects,withhighersoluteconcentrationsonthe structures, as in turbulent riffles (Lautz and Fanelli, rising limb than on the falling limb as a result of 2008). These zones can enhance biogeochemical 92 Rivers in the Landscape reactions and increase nutrient utilization, and may shallow overbank flow across floodplains—that pro- be critical to maintaining stream water quality (Han- motes flow separation and retention of organic mat- cock et al., 2005). ter can substantially increase nitrogen uptake. By Patrick (1995) reviewed the importance of river retaining organic matter, these sites provide biota an chemistry to aquatic organisms. Dissolved organic opportunitytoaccessandingesttheorganicmat- matter provides an important energy and nutrient ter (Naiman et al., 1986; Fanelli and Lautz, 2008). source.Ca,Mg,oxidizedS,N,andphosphates,along Partly for this reason, river management increas- with small amounts of Si, Mg, and Fe, are desirable ingly emphasizes protection or restoration of phys- formanyspecies.ThepHofthewateraffectsthesol- ical channel complexity, as well as protection or ubility and availability of elements. restoration of riparian corridors that can “buffer” Nitrogen and carbon are important nutrients in channels from excess nitrogen. river water. Rivers are major conduits for nitrogen Human activities, including use of inorganic andcarbontransport,butriverscanalsoremoveand fertilizer and emissions from fossil-fuel combus- transform dissolved nitrogen and carbon in trans- tion, have dramatically increased nitrogen inputs to port (Hall and Tank, 2003). The details of nitrogen watersheds (Boyer et al., 2006). Simultaneously, river and carbon transport versus removal provide a nice engineering that limits channel boundary complex- example of how physical, chemical, and biochemi- ity and retention, as well as channel–floodplain con- cal processes interact in rivers to govern transport nectivity, has reduced the ability of many rivers to of solutes. process nitrogen inputs. The result is excess nitro- Of the total nitrogen loss from the land, only genloadsandeutrophicationincoastalareas(Boyer about 18%–20% is carried to the oceans by rivers et al., 2006). Many large river systems are reach- (Van Breemen et al., 2002) because of removal and ing their limits for processing excess nitrogen. Pre- transformation en route. Rates of dissolved nitro- industrial nitrogen fluxes were greatest from the gen uptake from river water are especially high in largest rivers, such as the Amazon. Post-industrial shallow streams with algae and microbes in attached fluxes are greatest from rivers in the industrialized biofilms (Hall and Tank, 2003), but riverine process- zones of North America, Europe, and southern Asia ing of nitrogen is heterogeneous in time and space. (Green et al., 2004). Examples of biogeochemical hot spots in which Organic matter can be present in river water in accelerated chemical reactions occur include anoxic particulate and dissolved forms (Berner and Berner, zones beneath riparian environments (Lowrance 1987). Particulate organic matter enters rivers as et al., 1984) or the convergence of ground and sur- fossil carbon from sedimentary rocks and as car- face waters in hyporheic zones (Harvey and Fuller, bon from contemporary soils and vegetation. Car- 1998). Examples of biogeochemical hot moments bon inputs can be gradual and continuous. Carbon include rare high flows that occupy secondary chan- inputs can also be episodic, as in active mountain nels in arid environments (McGinness et al., 2002) beltswheresedimentyieldtoriversisdominatedby or snowmelt that enhances leaching of dissolved landslides triggered by tropical cyclones, which also nutrients in high-elevation catchments (Boyer et al., generatefloodsthatresultinlargeriverinefluxes 2000). of carbon to the ocean (Hilton et al., 2008a, 2008b, Much of the nitrogen removal within rivers 2011a, 2011b). occurs in stream sediments (Nihlgard et al., 1994) Dissolved organic matter includes different and riparian zones (McClain et al., 1999) through classesoforganiccompoundsthatdifferinreactiv- theactivitiesofbenthic(bottom-dwelling)organ- ityandecologicalrole,aswellasvaryinginquantity ismssuchasbacteria,algae,andinsects,andriparian with time and space in response to seasonal vari- soil organisms and plants that take up nitrogen. Any ation and precipitation inputs. Dissolved organic source of fluvial complexity—logjams, beaver dams, matter is usually expressed as dissolved organic channel margin irregularities, hyporheic zones, and carbon (DOC). Values in river waters are typically Chapter 4 Fluvial sediment dynamics 93

2–15 mg/L but can reach 60 mg/L in rivers draining derived from rock weathering, or from human activ- wetlands (Drever, 1988). DOC varies with the size ities that include mining, burning fuels, smelting of the river, the climate, and vegetation (Thurman, ores, or disposing of waste products (Drever, 1988). 1985). DOC values are typically high for temperate Sediments contaminated with trace metals associ- and tropical rainforests and taiga, for example, and ated with human activities are commonly referred low for arid and semiarid environments. to as legacy sediments (as are other types of deposits DOC is highly reactive and influences river- thatreflectpasthumanactivities).Thesecontam- ine ecosystems by controlling microbial food webs inated sediments create a long-term toxic legacy. (Kaiser et al., 2004). The efficiency with which rivers Trace metals receive most attention as contaminants retain and oxidize organic carbon depends on the occurring at concentrations above background lev- presence of microbial communities in response to els or hazards occurring at levels potentially harm- riverine features that increase the residence time of fultoorganisms.Examplesoftracemetalsthat organic molecules in transport, as explained above commonly create pollution are arsenic, cadmium, for nitrogen (Battin et al., 2008, 2009). Simply put, chromium,copper,lead,mercury,nickel,selenium, storage of organic matter supports microbial com- and zinc, each of which is very toxic and biochem- munities, which retain and oxidize organic carbon. ically accessible to living organisms. Because many Substantial differences in carbon dynamics are trace metals are adsorbed to fine sediment, the present between rivers with well-developed flood- mobility and storage of the fine sediment strongly plains, which exchange more than half of the recent influence the spread of trace metals through a river biogenic carbon in transport with floodplain car- network and the bioavailability of the metals. Metals bon, and small mountainous rivers that transport adsorbed to silt and clay traveling in suspension or carbon directly to the ocean (Galy et al., 2008). resting on the surface of the streambed, for example, Even small mountainous rivers can include seg- are more readily ingested by aquatic organisms than ments with substantial carbon storage, however, if are metals adsorbed to fine sediment deeply buried physical channel complexity creates flow obstruc- on a floodplain. tions and enhanced channel–floodplain connectiv- As with other aspects of hydrology, cold-region ity (Wohl et al., 2012c). Consequently, river manage- riversandriversofthehumidtropicshavesomedis- ment that facilitates maintenance or restoration of tinctive chemical characteristics. Cold-region rivers channel complexity has numerous benefits, includ- in alpine or permafrost terrains with high levels ing carbon and nitrogen retention and increased of impermeable surface area may be more influ- diversity of aquatic habitat. enced by precipitation chemistry than other types DOC is an important component of the global of rivers with greater infiltration and deeper, slower carbon cycle. An estimated 2.7 petagrams (Pg) of downslope pathways of water (Meixner et al., 2000). carbon are eroded from terrestrial sources and deliv- Selective weathering in glaciated catchments can ered to freshwater environments each year (Auf- allow more reactive minerals to contribute dispro- denkampe et al., 2011). Of this quantity, 0.6 Pg is portionately to dissolved load relative to their abun- buried in sedimentary reservoirs such as floodplains dance in the local rock. Chemical denudation rates or deltas. Another 1.2 Pg is released to the atmo- can be higher for glaciated areas than for adjacent sphere as a result of metabolism and respiration of non-glaciated catchments. Both seasonal snowmelt riverine organisms. The remaining 0.9 Pg is carried and glacial melt can create large temporal and spa- byriverstotheocean.Riverdynamicsstronglyinflu- tial variations in stream water chemistry as the ence the partitioning between sedimentary reser- source and contact time of melt water with sed- voirs, release to the atmosphere, and transport to the iment and rock change through the melt season ocean. (Taylor et al., 2001; Anderson et al., 2003). Total Trace metals typically occur at concentrations less organic carbon export from cold-region rivers is than 1 mg/L in river waters. Trace metals can be likely to increase under global warming because 94 Rivers in the Landscape of accelerated decomposition of soil organic mat- Interparticle forces are greater than the gravity ter as the seasonal duration of frozen soil decreases force among portions of the wash load finer than (Trumbore and Czimczik, 2008). 0.004 mm in diameter, and these very fine sediments Rivers of the humid tropics have warmer water, are cohesive. Cohesive sediments typically travel as higher annual exports of dissolved constituents aggregated or flocculated material, rather than as (Lyons et al., 2002), and lower seasonal variability single grains (Kuhnle, 2013). Unlike coarser sedi- in water temperature and chemistry than rivers at ments, for which transport may be limited by stream higher latitudes (Scatena and Gupta, 2013). Carbon energy (and which are thus known as transport lim- exportsoccurprimarilyasDOC,andbothDOC ited), fine cohesive sediments are more likely to be and particulate carbon exports can increase sig- supply limited, such that the amount in transport is nificantly when tropical storms create widespread limited by the amount supplied to the flow. Under defoliation and/or landsliding (Hilton et al., 2008a, supply-limited conditions, the actual transport rate 2008b). Consequently, humid tropical rivers can cre- is likely to be less than the transport capacity,which ate important global sinks of carbon by transporting reflects what the flow is capable of transporting. carbon to oceanic burial (Goldsmith et al., 2008). Bed-material load includes grains typically coarser than 0.062 mm. These grains move either in contact with the bed by rolling, sliding, or saltating 4.2.2 Suspended load as bed load, or in suspension just above the bed, with concentration declining upward from the The particulate sediment transported by rivers can bed. Bed-material load can also be subdivided into be subdivided in different ways. One distinction is suspended load and bed load. If sediment is divided betweenwashloadandbed-materialload,anotheris into only suspended load and bed load, then the between suspended load and bed load. The distinc- suspended load includes wash load. tion between wash and bed-material load reflects Non-cohesive sediment will remain in suspen- different mechanics of transport, but the coarsest sion if the strongest vertical velocity fluctuations grain size moving as wash load is not easy to measure exceed the particle fall velocity. Close to the bed, directly, blurring this distinction. The distinction therootmeansquareoftheverticalvelocityfluc- between suspended and bed load primarily reflects tuationsreachesamaximumthatisapproximately measurement technology, with some forms of bed equal to the magnitude of the shear velocity, v∗, load transport (saltation) being transitional between which provides an approximate criterion for suspen- the two end-members. sion (Kuhnle, 2013) Wash load is the finest size fraction of the total √ sediment load, typically grains with intermediate 휏 v = 0 (4.11) diameter ≤0.062 mm. Washload consists of particles ∗ 휌 typically not found in large quantities on the bed sur- 휏 휌 face. These particles have settling velocities so small where 0 is the bed shear stress and is water den- that they move at approximately the same velocity sity. Non-cohesive suspended sediment concentra- as the flow and only settle from suspension when tion typically decreases with distance away from the velocity declines substantially. Wash load is verti- bed in a manner described using the Rouse equation cally mixed by turbulent flows, so that concentra- tion varies little with flow depth. Because relatively ( ) Z little energy is needed to transport wash load, trans- C (a)(d − 훾) = 훾 (4.12) port rates typically reflect sediment supply rather Ca ( )(d − a) than flow energy. Much of the wash load comes from bank erosion and surface erosion across the drainage where C is the concentration of sediment at a dis- 훾 basin. tance above the mean bed elevation, Ca is the Chapter 4 Fluvial sediment dynamics 95 concentration of sediment at the reference level a Large suspended sediment concentrations com- above the bed, d is flow depth, and the exponent, Z, monly occur during high discharge. Using an equa- known as the Rouse number,isdefinedas tion similar to Equation 4.10, the exponent b is gen- 휔 erally in the range of 1 to 2 (Walling and Webb, s 1986). Wolman and Miller (1960) first suggested Z = 휅 (4.13) u∗ that most suspended sediment is carried by rela- tively frequent flows with a recurrence interval of 1– 휔 휅 where s istheparticlefallvelocity, is the von 2years,althoughthemagnitudeofflowcarryingthe Karman constant, and u∗ is the shear velocity greatest proportion of annual suspended sediment (Kuhnle, 2013). load−also known as the effective discharge−varies The Rouse number is basically the ratio of the widely among catchments. grain fall velocity to the strength of the flow. The Effective discharge is usually a relatively fre- relation between the Rouse number and the distri- quentflowinlargerbasins.WorkinginCanada’s bution of suspended sediment through the water Saskatchewan River basin, for example, Ashmore column (Figure 4.3) reflects the fact that, as the value and Day (1988) found that small headwaters basins of the flow strength (u∗) increases relative to the fall tended to have effective discharge during the most velocity of the sediment grains, the resulting gra- extreme events with annual flow duration much less dient of sediment concentration through the flow than 1%, whereas larger rivers had most suspended depth grows less steep (Kuhnle, 2013). sediment transported during frequent events with durations often greater than 10%.

1 Supplemental Section 4.2.2 includes more infor- 0.03 mation about suspended sediment concentrations in

0.8 Z values 0.06 relation to flow magnitude and recurrence interval. Most rivers display some form of hysteresis for 0.6 0.12 suspended sediment, but the details can vary widely. 0.25 Williams (1989) identified three forms of hystere- 0.4 sis (Figure 4.4). One form is a clockwise loop in 0.5 which the peak sediment concentration precedes 1.0 elevation ( y – a )/( h ) 0.2 the peak discharge because available sediment is 2.0 depleted before runoff peaks, a situation more com- Relative distance from reference Relative 4.0 0 moninsmallbasins.Asecondformofhysteresisis 0 0.2 0.4 0.6 0.8 1 an anticlockwise loop in which peak sediment con- Relative sediment concentration (C/Ca) centration lags the peak discharge, a situation more Figure 4.3 Diagram illustrating the effect of a range of common in large basins. The third form of hys- Rouse numbers (Z values) on the prediction of suspended teresis is a figure-of-eight loop in which peak sed- sediment concentration profiles. The variables on thex iment concentration precedes peak discharge, but and y axes are as follows: a is a reference level above the theshapeofthesedimentoutputisskewedrelative bed, h is depth of flow, C is the concentration of sedi- to the flood peak. This can occur where suspended ment at a distance y above the mean bed elevation, and sediment increases more rapidly than discharge, and C is the concentration of sediment at the reference level a thus peaks first, but post-peak sediment availabil- a. Z values reflect the ratio of grain fall velocity tothe strength of the flow. Greater values of flow strength rel- ity and transport are sufficiently high to create sedi- ative to the fall velocity (smaller Z values) of the grains ment concentrations that decrease more slowly than result in a less steep gradient of sediment concentration water discharge. with flow depth—that is, greater concentrations of sus- Sources of suspended sediment are typically pendedsedimenthighintheflowthanatlargeZ values not uniformly distributed across a drainage basin, (from Kuhnle, 2013, Figure 2). particularly in larger basins. This is dramatically 96 Rivers in the Landscape

C increases erosion from upland surfaces via processes Q such as rainsplash, sheetwash, and rilling. An exam- I ple comes from the 539 km2 Lago Lo´ıza basin of Puerto Rico. Completion of a dam in 1953 created thelake(LagoLo´ıza). Gellis et al. (2006) used his-

Q toric records of suspended sediment load on incom- C II ing rivers, repeat bathymetric surveys of the lake, sediment cores from the lake, precipitation records, and land use maps to evaluate potential controls on suspended sediment yield to the reservoir. They

C Q foundthatuplandclearingforagricultureresultedin III C increased sheetwash, rilling, and downstream sedi- ment yields, as did subsequent urbanization during the construction phase. Some of the sediment mobi- Time Q lized from uplands during agricultural development Figure 4.4 Three types of hysteresis in suspended was stored along channels and floodplains for sev- sediment—discharge relations when C (concentration) eral decades. This sediment was remobilized when and Q (discharge) are not synchronous (from Williams, second-growthforesttookovercroplands,sothat 1989, Figures 3, 5, and 7). downstream sediment yields remained high despite upland reforestation. Potential effects of land use on suspended sed- illustrated by Bob Meade’s “sedigraphs” for drainage iment are covered in more detail in Supplemental basins such as the Mississippi River in North Amer- Section 4.2.2. ica (Meade, 1996) (Figure 4.5). The majority of dis- Suspended load is the dominant component of charge in this drainage comes from the humid tem- sediment transport on many rivers, particularly perate Ohio River basin and the headwaters of the rivers draining areas larger than about 5000 km2 Mississippi, whereas the majority of suspended sedi- (Kuhnle, 2013). Suspended load is also the most ment comes from the semiarid Missouri River basin. commonly measured component of total sediment Therelativeimportanceofdifferentsub-basinsas load. Globally, suspended sediment accounts for sources of suspended sediment reflects sediment about 70% of total fluvial sediment transport to yieldsasinfluencedbyfactorssuchasrocktype, the oceans. Small, mountainous basins produce topography, climate, and land use. Sediment budgets muchgreateryieldperunitarea(tons/km2), but fortheAmazonRiverbasinshowsimilarspatialdis- large drainages produce the greatest total sediment paritiesasthosefortheMississippi,withtributary output. basins draining highlands contributing substantial Suspended sediment can influence hydraulics, suspended sediment, whereas other tributaries con- total sediment transport, and channel form in tribute negligible amounts. diverse ways. Large concentrations of suspended Suspended sediment dynamics along the Ama- sedimentcanincreaseflowviscosity,reducesettling zon also indicate the importance of the floodplain velocities, and thus decrease turbulence and facil- and channel banks. Much of the suspended sedi- itate the transport of coarser grains than in clear ment in transport during the rising and peak stages water (Simons et al., 1963). Suspended sediment of the annual flood is deposited in the floodplain, can also create turbidity that limits photosynthe- where the sediment remains until returned to fluvial sis for aquatic plants. Infiltration of fine suspended transport via bank erosion (Dunne et al., 1998). sediment into an alluvial streambed can increase Suspended sediment can also be strongly influ- bed cohesion and reduce porosity and permeabil- enced by land use that disrupts land cover and ityofthesedimentmatrix.Reducedporosityand Chapter 4 Fluvial sediment dynamics 97

Circa 1700 1980–1990

Mis souri (a)

Misso Ohio uri Ohio

Arkansas Arkansas

Red Red Icá Japurá Negro (b) Suspended 1000 sediment 0 200 400 Gulf of Gulf of Mexico Suspended sediment Mexico in millions discharge, in millions 500 of metric tons per year of tonnes Juruá per year 0

Missou Mississippi Madeira

ri Negro Trombetas Jarí Japurá Icá 6000 Discharge

Illinois 3000 in cubic hio O kilometers Jutaí per year Juruá Purús 0

Arkansa MadeiraTapajós Xingu s e esse Tenn Red 0 500

Atchafalaya Mean water discharge, 3000 2000 1000 0 0 330 km in cubic kilometers per year River distance from Atlantic Ocean in kilometers

Figure 4.5 (a) Schematic illustration of historical and contemporary contributions to suspended sediment discharge (above) and water discharge (below) of the Mississippi River (from Meade, 1996, Figures 5 and 6). The majority of the suspended sediment carried in the Mississippi continues to come from the Missouri River, although this value has declined substantially through time as successively more large dams have been built along the Missouri, resulting in substantial sediment storage in reservoirs. The Ohio River is the single largest contributor of discharge to the Mississippi. (b) Schematic representation of downstream changes in suspended sediment and discharge in the Amazon River (from Meade, 2007, Figure 4.6, p. 53). The Rio Negro and Rio Madeira contribute similar magnitudes of discharge to the Amazon, but the Rio Madeira contributes substantially more suspended sediment. permeabilitycaninturnreducehyporheicexchange deposition of fine sediments can strongly influence and the ability of stream organisms from aquatic thedistributionofcontaminantswithintheriver insectstofishembryostosurvivewithinthebed corridor. Nineteenth century silver and gold min- matrix. High suspended sediment concentrations ing using a mercury-amalgamation procedure led can correlate with high rates of overbank deposi- to release of thousands of kilograms of mercury- tion and both lateral and vertical accretion of flood- contaminated tailings into the Carson River in plains. High suspended loads can also correlate with Nevada,USA.Mercury,whichisassociatedwith stream banks with a greater percentage of silt and fine-grained sediment fractions (<63 휇m), has sub- clay, greater bank cohesion, and relatively narrow, sequently been dispersed downstream for tens of deep cross sections (Schumm, 1960). kilometers via overbank deposition during floods A variety of pollutants, including heavy metals, (Miller et al., 1999). pesticides, and excess nutrients, can travel adsorbed Supplemental Section 4.2.2 discusses measure- to cohesive fine sediments, so that the transport and ment and modeling of suspended load. 98 Rivers in the Landscape

4.2.3 Bed load a granular-fluid flow at high concentrations. Con- sequently, the mechanics of bed load motion vary Althoughthecontributionofbedloadtototalsed- because fluid–grain and grain–grain interactions iment transport is relatively small except in moun- change with concentration (Haschenburger, 2013). tainous river networks, bed load dynamics strongly As in the discussion of bed sediment characteri- influence river process and form. The movement zation,itisusefultodistinguishbedloadinpredom- of bed load influences the stability of the bed and inantly sand-bed channels from bed load in chan- disturbance regime of aquatic habitat. Moving bed nels with coarser substrate and a wider range of grain material influences flow resistance and river mor- sizes. For both types of channels, bed load move- phology, as well as lateral channel movement and ment is typically episodic and spatially heteroge- floodplain formation. In tectonically active regions, neous across the channel and downstream. the rates of channel incision and landscape evolution In sand-bed channels, much of the bed load fundamentally reflect the rate at which bed load is moves as migrating bedforms such as ripples, dunes, moved out of the channel network. and antidunes, which are treated in the next sec- Grainmotionwithinbedloadoccursviarolling tion. Transport rate reflects the energy of the flow, and sliding at lower flows, saltation at intermediate because the individual sand grains are relatively flows, and sheetflow during high flows, although the mobile and transport is not typically limited by sed- three types of motion can coexist (Haschenburger, iment supply. Sand-bed channels commonly exhibit 2013). Grains rolling or sliding along the bed can bedscourduringtherisingstagesofflow,butthen make microleaps across distances smaller than the fill to approximately pre-flood levels on the falling grain diameter, or begin to saltate. Saltation,from stage (Leopold et al., 1964). the Latin for “leap,” describes grains launched into In channels with substrate coarser than sand, bed the water column that follow a predictable, ballis- load movement is more episodic in time and space, tic trajectory before once more contacting the bed. and is typically limited by sediment supply. Supply The maximum saltation height, which can be up limitationsontransport,aswellaslargespatialand to ten times the grain diameter, defines the upper temporal variations in local flow field and bound- extent of the bed load layer. Sheetflow occurs when ary configuration, make it difficult to predict bed grains move in multiple granular sheets with a grain load transport using foundational equations of bulk concentration approaching that of the underlying sediment transport developed for uniformly sized stationary bed. (Sheetflow is also used to refer to grains (typically sand sizes) in flumes. Consequently, sediment movement under very shallow flows on research has shifted toward using two-phase equa- hillslopes.) tions that differentiate sand and gravel. Gravel-bed The ratio of dimensionless shear stress to crit- channels can also exhibit scour during the rising 휏 휏 ical dimensionless shear stress ( ∗/ ∗c)correlates limb of a flood and fill during the falling limb. with type of sediment transport. Bed load typically 휏 휏 dominates where ∗/ ∗c is ∼1–3, bed load and sus- 휏 휏 Bed load in channels with coarse-grained pended load occur where ∗/ ∗c is between 3 and 33, 휏 휏 substrate: coarse surface layer and suspended load dominates where ∗/ ∗c exceeds ∼33 (Church, 2005). In streams with coarser bed material, flow energy is Most bed load comes from the streambed and important, but the finer portion of the grain-size dis- lower banks. Bed load flux depends on grain veloc- tribution can also be supply limited by the existence ity and concentration. At low concentrations, bed of a coarse surface layer. Many channels with rela- load particles collide primarily with stationary bed tively coarse, poorly sorted grain-size distributions grains. Collisions between mobile grains become have at least slightly finer-grained sediments imme- more frequent as bed load concentrations increase. diately below the coarse surface layer (Figure 4.6). Bed load grains are supported by grain collisions in Thecoarsergrainsizesatthesurfaceincreasethe Chapter 4 Fluvial sediment dynamics 99

(a) (b) Figure 4.6 Coarse surface layer (a) and underlying finer sediment (b) on an alternate along the Poudre River, Colorado, USA. Scale with inches (right) and centimeters (left) at upper left in (b). threshold of hydraulic force that must be exceeded have less developed coarse surface layers than peren- before the bed is mobilized, as well as increasing nial channels (Hassan et al., 2006). Coarse surface the hydraulic roughness of the bed. The coarse sur- layers can also be spatially variable at the reach scale face layer is characteristically one grain diameter in as a result of variations in shear stress and sediment thickness and is both more coarse grained on aver- dynamics associated with alternate bar topography, age and better sorted than the underlying sediment. hydraulic resistance caused by instream wood, or The underlying sediment has a grain-size distribu- variations in sediment supply. tionsimilartothatofthebedloadinmanygravel- Several studies suggest that the coarse portion of bed rivers. a surface layer moves down a river at nearly the same The coarse surface layer can be called pavement, rate as the finer sediment in the layer. Coarse grains armor, or a censored layer. Each of these terms are intrinsically less mobile than finer grains, but is used differently by different investigators. Pave- the coarse surface layer might provide an equalizing ment is most commonly used for a coarse surface mechanism by exposing proportionally more coarse layer that is rarely disrupted. Armor most commonly grains to the flow. The coarse surface layer can thus refers to a coarse lag layer developed at waning flows beamobile-bedphenomenonpresentunderarange that is regularly disrupted. Censored layer commonly of flows, with only a few clasts of diverse sizes being refers to a layer that forms as finer matrix material is entrained at any instant by even the peak flows removed from around the surface framework parti- (Andrews and Erman, 1986; Clayton and Pitlick, cles as flow increases (Carling and Reader, 1982). 2008). Themechanismbywhichthecoarsesurfacelayer forms, and the stability and frequency of mobiliza- Bed load in channels with coarse-grained tion of the layer, vary among channels and with time substrate: characteristics of grain movements for a given channel. Coarse surface layers typically are well developed when local bed load supply from In coarse-bed streams, grains move as individual upstream is less than the ability of the flow to trans- particles or in discontinuous pulses. Pulses have portthebedload.Thissituationiscommondown- been described as sheets in which migrating slugs stream from dams (Dietrich et al., 1989), and may of bed load 1 to 2 grain-diameters thick alternate develop with time following a pulse sediment input between fine and coarse particles. Pulses of bed load such as that from a volcanic eruption (Gran and canalsotaketheformoftraction carpets,whichare Montgomery, 2005). Ephemeral channels typically highly concentrated bed load layers, or streets,which 100 Rivers in the Landscape are longitudinally continuous tongues of bed load petent flow (Haschenburger, 2013). Local flow and that do not span the full width of the channel. Pulses bed conditions strongly influence the start and end can also occur as waves,whicharegroupsoflow- of step lengths and path lengths. Longer duration amplitude bars migrating downstream in braided flow can create longer path lengths when grains rivers,withplanebedtransportacrosstheupper move over relatively smooth beds. Bed morphol- bar surface and avalanching on downstream faces. ogysuchasbarsorpoolscanlimitpathlengthsfor Each of these forms of transport reflects the fact that even longer duration flows, such that the modal path coarsebedloadisunlikelytobeequallydistributed length correlates to spacing of pool—riffle–bar or in time and space along a reach of channel. step–pool units (Pyrce and Ashmore, 2003). In channels with mixed grain sizes, rolling and The number and size of grains entrained from sliding grains move episodically. Field observations the surface increase with flow and, when the entire suggest that rolling and sliding grains move less than surface is mobile, grains entrain from the subsur- 20% of the time during which flow conditions are face. As the bed transitions from being immo- sufficient for entrainment (Haschenburger, 2013), bile to partial mobility and then full mobility, the although periods of transport increase as rates of upper limit to partial mobility is equal transport grain entrainment increase. Turbulent bursts cause mobility, where all grain sizes are transported in much of the grain motion (Nelson et al., 1995). proportion to their presence on the bed (Parker Step length is the travel distance of a particle dur- and Toro-Escobar, 2002). This condition, which is ing continuous motion. The mean step length of relatively uncommon in coarse-grained channels, grain motion appears to exceed 100 grain diam- occurs when all surface grains are mobilized at flows eters, although step length increases with flow exceeding entrainment thresholds by about four and decreases with increasing grain size and lower times (Wilcock and McArdell, 1997). sphericity (Schmidt and Ergenzinger, 1992). In par- The distinction between partial and full mobil- ticular, step length varies only moderately for grains ity is also known as phase I and phase II trans- smaller than D50, whereas step length of larger grains port (Andrews and Smith, 1992). Phase I transport correlates inversely with grain size because the step starts when flow energy is just sufficient to rotate length is limited by the energy required for transport someofthegravel-sizedparticlesoutoftheirrest- of these larger grains (Church and Hassan, 1992). ing place, sending the particles rolling or bouncing Grain collisions or areas of lower hydraulic force or downstream. As shear stress increases, more parti- rougher surface topography initiate periods of rest. cles become mobile. Where a well-developed coarse The distributions of step lengths and rest periods surface layer is present, phase I involves transport of have been described using gamma and exponential a finer fraction over a stable coarse bed. Sometimes, density functions, as well as a Poisson distribution. three phases of bed transport are distinguished. In Virtual velocity of grains quantifies the rate of this case, phase I is transport of a finer fraction over downstream movement, including both step and astablecoarsebedandphaseIIispartialandusu- rest periods during flows when entrainment can ally size-selective local entrainment (Ashworth and occur. This is known as grain velocity.Grainveloc- Ferguson, 1989; Warburton, 1992). Most of the bed ity, which is not particularly sensitive to flow veloc- is mobile during phase II transport (or phase III in a ity but does reflect grain characteristics such as den- three-phase conceptualization). The two phases can sity, can also be described with a gamma density be quantified using dimensionless shear stress, 휏∗, 휏∗ 휏 휌 휌 function. Mean virtual velocity increases with flow, where = /g( s − )D: phase I occurs in the range and has been documented at up to 100 m/h (Hassan 0.020 <휏∗ < 0.060, and phase II when 휏∗ > 0.060. et al., 1992). The nonlinear effects of sand on the trans- The path length of moving bed load is the port of coarse particles in a bed with mixed grain total downstream displacement of particles, typi- sizes illustrate the complexities of bed load move- cally through multiple steps, during a period of com- ment (Wilcock and Kenworthy, 2002). Sand can be Chapter 4 Fluvial sediment dynamics 101 preferentially transported at low discharges that do to the streambed. Anchor ice can form when large- notmovecoarsesediment.Astheamountofsand scale turbulence mixes suspended ice crystals and present increases, local sandy areas decrease the frazil ice across the full depth of flow, allowing some protrusion of large grains. Moderate amounts of of the ice to adhere to the bed or individual boul- sand can thus decrease the entrainment of large ders. This is particularly common in riffles and can grains.Oncealargegrainisentrained,though,it result in anchor ice forming in flows as deep as can move faster over the relatively smooth sand 20 m (Ettema and Daly, 2004). Under sufficiently bed and may move farther because sand fills the cold temperatures and substantial flow turbulence, pore spaces between large grains and reduces resting extensive areas of the streambed can become cov- places. ered by anchor ice. Larger amounts of anchor ice In addition to moving downstream, grains on the typically form on coarser beds, which facilitate heat bed surface are also vertically exchanged with those flux from a sub-bed zone that is 1–2 degrees above in the subsurface through local scour and fill. Under freezing to supercooled flow over the full depth someconditions,thebedisinactiveandmobile above this zone. Anchor ice is less likely to form on grains pass over the bed. Finer grains can prefer- fine, non-cohesive sediment, which is readily lifted entially infiltrate, particularly if framework gravels by ice and therefore cannot hold a large accumula- on the bed dilate prior to movement (Allan and tion of ice (Ettema and Daly, 2004). Frostick, 1999). (Dilation refers to a phenomenon in Both frazil and anchor ice can be at least briefly which grains lift and move apart from one another attached to the bed and then float up with sedi- just before entrainment (Allan and Frostick, 1999).) mentfrozenintotheice.Diurnalformationofthese More active beds facilitate grain exchange and sort- types of ice can result in repeated ice-rafting of sedi- ing, although finer particles are still exchanged into mentalongariverduringthecoldseason.Although the subsurface more quickly than coarser particles. observations indicate that sediment is entrained and Increased flow magnitude or duration increases the rafted downstream by these forms of ice, there is lit- frequency and depth of vertical exchange (Wong tle quantitative information on exactly how impor- et al., 2007), with exchange depths reaching up to tant this process is relative to other forms of sedi-

10 times the diameter of D90,butmoretypically2D90 ment transport (Ettema and Daly, 2004). Much of (Haschenburger, 2013). the sediment entrained by the ice is stored in a sea- Entrainment and transport of alluvium along sonal ice cover until the cover breaks up during bedrock rivers likely varies along a continuum from warmer weather. an end-member with continuous alluvial cover, in which processes operate as described in Section Bed load in channels with coarse-grained 4.1.2, to a bedrock end-member in which sediment substrate: controls on bed load dynamics entrainment and transport are independent of sizeandmorelikelytoreflectthelocationoflocal Spatial and temporal discontinuities in bed load roughness and hydraulics (Goode and Wohl, 2010a) movement have been explained in relation to sev- that stabilize alluvial sediment patches (Hodge et al., eral potential controls, each of which may dominate 2011). in different channels or as conditions within a chan- Cold-region rivers have unique forms of sedi- nel change. Multiple controls can also influence bed ment transport related to the formation and move- load movement at a particular time and place. ment of frazil and anchor ice. Frazil ice is a collection Temporal variations in bed load can be distin- of loose, randomly oriented, needle-shaped ice crys- guished as occurring at three scales. tals that resembles slush. Frazil ice forms sporadi- cally in open, turbulent, supercooled water, typically 1. Long-term variations reflect the rate of on clear nights when the air temperature drops to sediment supply to the channel (Vericat et al., −6◦C or lower. Anchor ice is submerged ice attached 2006). Altered sediment supply from outside the 102 Rivers in the Landscape

channel can occur as discrete sediment inputs of The frequency of the effective discharge for bed largevolume,suchasalandslide,thattranslate load decreases as the size of the bed sediment downstream as waves or disperse (Lisle, 2008; increases. In other words, coarser beds are mobile Sklar et al., 2009). Increased sediment can also less frequently than finer beds. Rates of bed load enter the channel in prolonged, widespread transport increase rapidly and nonlinearly within a inputs that produce responses such as an increase channel once sand becomes mobile, and can span up in gravel bars with a broad-scale wave-like form to seven orders of magnitude (Milhous, 1973). that move downstream over periods of tens to The migration of readily mobilized bedforms hundreds of years (Jacobson and Gran, 1999). can account for the majority of bed load move- 2. Short-term variations in bed load reflect tempo- ment in coarse-grained channels, as in sand-bed rary changes in sediment supply and the move- channels. Bedforms can also strongly influence the ment of bedforms, such as dunes (Gomez et al., hydraulics that control sediment transport. In pool– 1989). Bedforms may not span the entire channel riffle sequences, the pools have higher sediment cross section, and time gaps can occur between transport during high flows, whereas riffles and passage of successive bedforms. During a single alternate bars are sites of sediment deposition dur- flood lasting hours to weeks, bed load can dis- ing high flows (Thompson et al., 1996). The steep- play hysteresis analogous to that of suspended ened water-surface gradient over riffles and bars sediment, with greater bed load transport rates during low flows can promote dissection of the during the rising stage. Discontinuities in bed riffles and bars and the removal of finer parti- load movement have also been attributed to cles that are then stored in pools until the next longitudinal sediment sorting, which results in high flow. Clasts in pools of step–pool sequences pulses of coarser or finer material moving down- are also preferentially entrained during the rising stream (Whiting et al., 1988). Lateral shifting limb of a flood and deposited during the falling ofbedloadstreets,whichtypicallydonotspan limb. theentirechannelwidth,cancreatespatialdis- Instream wood can be a major source of cross- continuities in bed load and apparent tempo- sectional to reach-scale heterogeneity of bed struc- ral discontinuities at a point (Ergenzinger and ture and relief, and can exert a particularly strong de Jong, 2002). Sediment storage in secondary influenceonbedloadtransportinchannelswith channels can fluctuate nonlinearly with changes mixed grain sizes. By creating form roughness and in stage, resulting in discontinuities in supply obstructions, wood increases flow resistance and to the main channel. Bed load discontinuities flow separation, leading to enhanced entrainment can reflect cross-sectional to reach-scale hetero- of material from portions of the bed and banks geneity of bed structure and relief that influence subject to scour because of current deflected by both sediment supply and transport. In a channel wood. Wood appears to be most effective, how- with mixed grain sizes, for example, protruding ever, in promoting localized sediment storage (Fig- clastscancreateleestorageoffinersedimentuntil ure S4.10). Many studies document preferential stor- fluctuatingdischargealtersthelocalhydraulics age of bed material near individual wood pieces or and the finer sediment is entrained (Figure S4.9) logjams and greater overall storage, particularly of (Thompson, 2008). finer sediment, in channel segments with abundant 3. Instantaneous variations in bed load reflect the wood (Piegay´ and Gurnell, 1997; Buffington and inherently stochastic and probabilistic nature Montgomery, 1999; Montgomery et al., 1996, 2003b; of the processes governing entrainment, as Skalak and Pizzuto, 2010). Wood that breaks or is examined in Section 4.1.2. Typically, instanta- mobilized releases a pulse of bed load, and long-term neous variations increase with channel boundary removalofwoodcanchangeadepositionalchan- roughness and increasing spread of bed grain nel reach to a net sediment source (Brooks et al., sizes. 2003). Chapter 4 Fluvial sediment dynamics 103

In addition to the physical processes acting to ments, from enhanced bank erosion to bed coars- move bed load, aquatic organisms can displace or ening and incision. Globally, over 100 billion metric sort bed material and alter the susceptibility of tons of sediment are now stored in dams, and sed- bed sediment to entrainment and transport. Vari- iment transport to the oceans has been reduced by ous species of fish create redds or nests for their an estimated 1.4 billion metric tons per year (out of eggs or larval fish. Where large numbers of fish 14 billion metric tons per year total) (Syvitski et al., such as salmon concentrate in a channel during the 2005). spawning season, their disruption of the bed can result in nearly half of the annual bed load yield Estimating bed load flux (Hassan et al., 2008, 2011). Crayfish also disrupt the bed sediments in the course of moving about Bed load can be examined using a Lagrangian and burrowing into the bed, and these activities approach, which tracks grains along a river network, can significantly influence clast entrainment and or an Eulerian approach, which describes grain pas- transport during low flows (Statzner and Peltret, sage at a fixed location such as a limited segment of 2006). channel (Doyle and Ensign, 2009). Using an Eule- Themagnitude,frequency,anddurationofbed rian perspective, bed load flux can be estimated load transport also influence aquatic biota by limit- using three methods (Haschenburger, 2013). ingthestabilityofthechannelsubstrateandphysi- cally disturbing plants and animals via abrasion, dis- 1. Flux can be estimated from grain kinematics— lodging, and displacing or burial. Even if the coarse the movement of grains as described by their dis- surface layer of a mixed grain-size bed remains placement and velocity. This approach relies on mostly stable, sand and finer gravel moving across coupling grain entrainment rate from a given bed the coarse surface can disturb benthic or bottom- area with the associated displacement length of dwellingorganisms.Otherfactorsbeingequal,sand- the mobilized grains. Entrainment rate and dis- bed channels typically have lower abundance and placement length are then extended to explicitly diversity of benthic organisms than less mobile sub- account for grain-size fractions (Wilcock, 1997; strates,andlessmobilefeaturessuchasinstream McEwan et al., 2004). Bed load transport rates wood attached to the bank have a disproportionately can also be derived from the spatial concentra- large concentration of organisms in sand-bed chan- tionofbedloadtravelingatagivengrainveloc- nels (Minshall, 1984). ity. Or, transport rates based on the displacement The greatest biological influence on bed load of grains over some fixed time, together with the transport is undoubtedly human alterations of portionofthebedthatisregularlymobilized,can rivers. Changes to channel form through activities be used to estimate flux. such as dredging, channelization, bank stabiliza- 2. Flux can be estimated from changes in river tion, and straightening influence the erodibility of morphology that reflect bedform migration bed and bank sediments and thus sediment sup- or bed aggradation and degradation (Trayler ply. Sediment supply to the channel is also strongly and Wohl, 2000; Haschenburger, 2006). This influenced by human activities outside of the chan- approach depends on being able to directly access nel, as discussed in Supplemental Section 4.4 and thebedusingsurveyinstruments,orindirectly Section 8.1.2. image the bed with fathometers or other remote Likely, the greatest alteration of suspended and sensing, at time intervals and spatial resolution bed load transport in many rivers is associated with relevant to bed load flux. flow regulation. Dams effectively trap all bed load 3. Bed load transport rates can be derived from and much of the suspended load moving down a (semi) empirical equations, typically based on river, leading to dramatically reduced sediment sup- laboratory experiments. Starting with Shields’ plies downstream and a range of channel adjust- flume experiments in sand-bed channels 104 Rivers in the Landscape

휏 (Section 4.1.2), investigators initially developed where is bed shear stress], Fi is proportion of frac- bulk models of bed load flux with equations tion i on the bed surface, and the subscript i repre- based on one representative grain size. Most of sents either the sand or gravel fraction. thesemodelswerevariationsonShields’formula More recent research focuses on grain–grain and assumed that transport is proportional interactions during transport (Frey and Church, to a bulk flow parameter such as shear stress 2012). Experiments with sand and gravel tracers or stream power. An example of a bed load reveal diffusion of the tracers, statistically defined transportequationthatfollowsthisapproach as the spreading of particles downstream. Diffusion and remains widely used is the Meyer-Peter and reflectsthefactthatindividualgrainsmoveatdiffer- Mueller (1948) equation, which can be written in ent rates as a result of the complexity of grain–grain dimensionless form as and grain–fluid interactions (Martin et al., 2012). The progression of differing approaches through

( ) 3 time reflects the application of increasingly sophis- gb = 휃′ − 휃 2 [( ) ] 1 8 c (4.14) ticated physics to the problem of understanding and 휌 − 휌 2 s gD3 predicting bed load flux, particularly in channels 휌 of mixed grain size. Changes in flow and bed load transport rate typically do not correspond well in 휌 휌 where is water density, s is sediment density, g spaceortimeinthesechannels,forthemanyrea- is acceleration due to gravity, D is the size of the sons outlined previously, making prediction of bed bed material, gb is the transport rate (volumetric load transport extremely difficult. transport rate per unit width, in m2/s), 휃′ is the Whatever their form, most bed load transport 휃 dimensionless grain stress, and c is the critical equations do not accurately predict fluxes in chan- dimensionless shear stress (McLean et al., 1996). In nels with coarse or mixed grain sizes. The discrep- 휃 theoriginalformoftheequation, c was set equal ancy reflects at least three factors (Haschenburger, to 0.047 (Robert, 2003). Actual transport rate is 2013). First, the equations require a specific chan- assumed to equal theoretical transport capacity in nel response based on defined hydraulic conditions. thisequation,althoughrateislikelytobelowerthan In other words, the equations require equilibrium capacity in coarse-grained channels because of lim- transport conditions, which do not always occur in ited sediment supply. coarse-grained channels. Second, the equations do Starting in the 1980s, size-specific formulae were not integrate the spatial and temporal complexities developedforbedloadflux.Thisapproachassumes that influence entrainment and transport, such as that different grain sizes move independently of each spatial and temporal heterogeneities in bed rough- other, although movement or stationarity by grains ness, which can be substantial in coarse-grained of a given size can influence the movement of other channels. Finally, the equations rely on input data grains via effects such as shielding. An example of that are typically difficult to characterize, such as this approach is the two-fraction transport model spatial heterogeneity of the bed. proposed by Wilcock and Kenworthy (2002). The Prediction of bed load flux for specific rivers can two fractions of sand and gravel are described by the be improved by collecting bed load data. Supple- ∗ dimensionless transport rate Wi via mental Section 4.2.3 provides details on measuring bed load. (s − 1) gq W∗ = bi (4.15) i F u3 i ∗ 4.3 Bedforms where s is the ratio of sediment to water density, g is gravity, qbi is volumetric transport rate per unit Bedforms arebedundulationsthatresultfromsedi- 휏 휌 0.5 width of size i, u∗ is shear velocity [u∗ = ( / ) , ment transport. Bedforms in sand-bed channels are Chapter 4 Fluvial sediment dynamics 105 smaller than channel-scale bar forms. Sand bed- quickly changing bedforms. At the other extreme forms are sometimes distinguished as: arecohesive-bedchannelsinwhichthebeddeforms only by erosion. Between these end-members are r microforms such as ripples for which the occur- coarse-grained channels in which a larger threshold rence and geometry are controlled by boundary- of flow energy (relative to sand-bed channels) must layer characteristics including bed grain size; be exceeded before the bed material begins to move r mesoforms such as dunes that are controlled by andbedformsarecreatedormobilized. boundary layer thickness or flow depth; and r macroforms such as bars that have lengths of the same order as channel width and heights com- 4.3.1 Readily mobile bedforms parable to the flow producing the form (Bridge, 2003). Readily mobile bedforms are primarily those in sand- bedchannels,whichoccurinthebroadclassesof Bridge (2003) also distinguished between a small-scale ripples and larger-scale dunes, both of bedform—a single geometrical element such as a which have their long axes perpendicular to down- ripple, and a bed configuration—the assemblage of streamflow.Othertypesofbedformsthatdevelop bedforms of a given type that occurs in a particu- under specific hydraulic conditions and are rarer larbedareaatagiventime.Bedformswillbeused includeupper-stageplanebedsandantidunes(Fig- here in a more general sense that includes individual ure 3.13). bedforms and the bed configuration. Understanding of bedforms in sand-bed chan- Explaining bedforms involves considering both nels is particularly important for at least three rea- sediment transport dynamics and channel geome- sons.First,bedformsarethemostimportantsource try. Migration of bedforms creates bed load flux, and of flow resistance at the local scale, creating two to alters the conditions for mobility of sediment not three times the resistance caused by grains. Second, within a bedform. In the presence of bedforms, a sig- bedform initiation, growth, and migration dominate nificant proportion of the total stress is caused by sediment dynamics in sand-bed channels, both by form drag on the bedform and this component of servingasabasictransportprocessandbyaffecting the total stress is not effective in moving sediment. the near-bed flow field and distribution of hydraulic Bedforms therefore reduce bed load transport rates forceandturbulence.Finally,bedformsleavechar- relativetoaflowwiththesamemeanvelocitythat acteristic depositional records, the interpretation of does not include bedforms. which is fundamental to understanding past river The presence and characteristics of bedforms are environments (Venditti, 2013). interconnected with channel geometry to the extent A typical sequence of bedforms evolves as veloc- thattheoccurrenceofcertaintypesofbedforms ityincreases(Figure3.13).Atverylowvelocities, reflects not only channel geometry variables such sand grains on a bed will not move. As velocity as flow width and depth or bed gradient, but also increases, the lift and drag forces exceed the sub- influences these variables. Step–pool sequences, for merged grain weight and other resisting forces. The example, form within a restricted range of bed gra- bed begins to develop millimeter-scale grooves par- dient. Steps and pools also alter flow path length by allel to the flow and spaced at the scale of the high creating a stepped rather than uniform longitudinal and low speed streaks that are ubiquitous in shear profile. flow.Thisisalower-regime plane bed,whichcan As noted in the introduction to this chapter, the persist for long periods of time or develop into subdivisions within this section recognize an impor- bedforms as a result of either defects or instanta- tant difference in bed mobility. At one extreme are neous initiation (Venditti et al., 2005). Defects such sand-bed channels in which readily mobile bed sub- as a small depression or protrusion into the flow strate results in relatively high bed load fluxes and create flow separation and local shear stresses able 106 Rivers in the Landscape to move sediment while the average shear stress The planimetric morphology of lower-regime remains below the threshold of motion. At higher bedforms can be distinguished as being two or shear stress and general sediment transport, bed- threedimensional.Two-dimensional(2D)bed- formsappearovertheentirebednearlyinstanta- forms have fairly regular spacing, heights and neously. This presumably reflects instability in the lengths, and straight or slightly sinuous crestlines form of turbulence and velocity fluctuations at oriented perpendicular to the mean flow lines. the water–sediment interface that rapidly increase Three-dimensional (3D) bedforms have irregular the extent of entrainment and grain–grain interac- spacing, heights and lengths with highly sinuous or tions, allowing grains to start collecting into bed- discontinuous crestlines (Venditti, 2013; Figures 4.7 forms. and S4.11). As with symmetry, differences in the ThestagesinFigure3.13upthroughdunesare mechanics of formation of 2D versus 3D bedforms lower flow regime bedforms.Thesetypicallyexistin remain under discussion, although 3D forms may subcritical flow and have relatively small bed load reflect persistent, unidirectional flows, whereas 2D transport rates. The size of lower-regime bedforms bedforms reflect variable flows. Lower-regime bed- varies widely between diverse rivers, from a few cen- forms can also be superimposed on one another, as timeters to several meters in height and a few tens of when ripples and dunes are superimposed on bars, centimeters to as much as a kilometer in length. Bed- or ripples are superimposed on dunes. form height, H,andlength,L,areempiricallyrelated As the Froude number approaches 1, dunes are (Flemming, 1988) as increasingly likely to wash out to a plane bed, which is the start of the upper flow regime bedforms (plane . H = 0.677L0 8098 (4.16) beds can be present at Froude numbers well below 1). These bedforms exist in supercritical flow and The existence of two subpopulations in a plot have relatively large bed load fluxes and small flow of H versus L is used as evidence that ripples and resistance. Upper-stage plane beds have low-relief dunes are distinct forms (Ashley, 1990). Ripples do bed waves only a few millimeters in height. These not affect the water surface and form only in sedi- bed waves migrate downstream beneath a water sur- ment where D < ∼0.7 mm. When dunes form, waves face with waves that are out-of-phase with the bed develop on the water surface that are out of phase waves. At the high velocity end of upper flow regime with the troughs and crests. bedforms, antidunes occur in all grain sizes and the Dunesforminsedimentfromfinesandtogravel water surface waves are in-phase with the bed sedi- size. The ability to form dunes thus reflects flow ment waves. Antidunes are upstream-moving undu- magnitude, rather than grain size, as evidenced by lations that typically grow to an unstable steepness, dunes formed of boulder- to gravel-sized parti- break, and then reform (Southard, 1991). Antidunes cles during late Pleistocene outburst floods (Baker, canformfromdunesorupper-stageplanebedsas 1978). Dunes are typically less steep than ripples, velocity increases (Venditti, 2013). although slope angles for the two types of bedforms Bedforms can also be described using phase dia- overlap. grams. Figure 4.7 illustrates the distribution of bed- Lower-regime bedforms can be low-angle sym- formtypesinrelationtovarioushydraulicand metric (downstream slope ∼10 degrees) or angle of sediment variables, including fall velocity, stream repose asymmetric with an upstream slope averag- power, Froude number, and dimensionless excess ing 2–6 degrees and downstream slope ∼30 degrees. shear stress (Venditti, 2013). Examining bedforms Symmetric forms are common in large rivers in the context of a phase diagram emphasizes the and estuaries, whereas asymmetric forms seem to relation of individual types of bedforms to a contin- be more common in small channels, although it uum defined by hydraulic and sedimentary variables remains unclear how mechanics of formation differ (Southward and Boguchwal, 1990; Southard, 1991) between the two types. (Figure S4.12). Chapter 4 Fluvial sediment dynamics 107

(a) (b)

Figure 4.7 (a) Schematic illustration of the planimetric morphology of two-dimensional (2D) sand bedforms with planar cross-stratification and three-dimensional (3D) sand bedforms with trough cross-stratification (from Venditti, 2013, Figure 7). (b) Plan view photographs of 2D (upper photo) and 3D (lower photo) ripples, both in the shallows along the edge of the Niobrara River of Nebraska, USA. Sedge stem in upper photo gives some indication of scale: both sets of ripples are a few centimeters tall. Primary flow direction is indicated by a white arrow in each photograph. Dark material in lower photograph is fine-grained organic matter temporarily deposited in the lee of each ripple.

Most examinations of flow over sand bedforms Flow have focused on asymmetric, 2D dunes. Flow accel- (a) erates and converges over the upstream side of these Acceleration Deceleration Acceleration dunes, then separates at the dune crest. The zone Outer flow layer of flow separation is associated with a turbulent Shear layer wake and shear layer originating at the dune crest Wake and extending and expanding downstream. Flow Separation reattaches at a distance downstream about four to zone X Reattachment point six times the bedform height. An internal bound- Flow ary layer grows from the reattachment point down- (b) stream beneath the wake toward the crest. An outer, Deceleration Acceleration overlying wake region is present above the level of Acceleration flow separation and reattachment (Venditti, 2013) Outer flow layer (Figure 4.8a). Shear layer Wake Flow over symmetric dunes includes a well- Separation defined region of decelerated flow in the lee of the zone X Reattachment point bedform, a shear layer between the main flow and Figure 4.8 Hydraulics over sand bedforms. (a) Two- the decelerated flow, and turbulent eddies generated dimensional asymmetric dunes at the angle of repose and along this shear layer that dominate the macrotur- (b) low-angle asymmetric dunes. X is the reattachment bulent flow structure (Venditti, 2013) (Figure 4.8b). length (from Venditti, 2013, Figure 9a and 9b). Symmetric dunes have less intense mixing along the shear layer in the dune lee, and only intermittent flow separation. 108 Rivers in the Landscape

All forms of coherent flow structures and turbu- 4.3.2 Infrequently mobile lence over asymmetric and symmetric dunes effec- bedforms tively increase the resistance of a channel with such bedforms and influence sediment entrainment and Infrequently mobile bedforms are primarily those mobility. Symmetric dunes have lower flow resis- occurring in gravel-bed channels. The cobbles and tance, however, than asymmetric dunes. coarser sediment in these channels have a higher The structure of turbulence over ripples and entrainment threshold that may be exceeded much dunes controls the flow resistance. The flow struc- less than half of the time during an average flow year, ture over dunes is dominated by depth-scale turbu- although minor adjustments associated with indi- lent eddies, known as macroturbulence or kolks when vidual grain movements or the mobility of the finer they are in the water column and boils when they fraction of the bed sediment can occur over a wide disturb the water surface (Venditti, 2013). Kolks and range of flows. Lower-regime plane beds and dunes, boils likely originate from the combined effects of as well as upper-stage plane beds and antidunes, can boundary layer bursts, instabilities along the shear occur in gravel-bed channels (Bridge, 2003). layer, and vortex shedding. Aswiththemorereadilymobilizedsand-bed Individual bedforms continue to change their bedforms, infrequently mobilized bedforms in geometry under constant flow conditions, as well coarser-grained channels both respond to and as under changing flow. Bedform changes typically strongly influence flow resistance, the distribution lag changes in flow, and thus display bedform hys- of velocity, turbulence, and other hydraulic vari- teresis that can be minutes to hours for ripples, and ables, and sediment transport. The interactions as much as days to months for dunes (Venditti, among bedforms, flow, and sediment in transport 2013). within gravel-bed channels typically occur over In addition to changing their geometry, bedforms larger spatial scales and longer timescales than anal- migrate with time, thus contributing to sediment ogous interactions in sand-bed channels, but exhibit transport. Asymmetric bedforms migrate by ero- similar levels of complexity. The primary bedforms sion on the upstream side and deposition on the treated here are particle clusters, transverse ribs, downstream side via avalanching down a steep slip- pool–riffle sequences, step–pool sequences, and bars. face. Over symmetric dunes, migration can also involve large amounts of sediment entrained into suspension on the upstream side of the dune and Particle clusters deposited from suspension on the downstream side. Particle clusters are closely nested groups of clasts Migration creates inclined packets of sediment aligned parallel to flow, typically 0.1 to 0.2 m in that can be preserved as cross-strata in the sed- length in the downstream direction, and about twice imentary or rock record. Ripples typically create as long as they are wide (Brayshaw, 1984) (Figure cross-strata millimeters thick, and dunes can cre- S4.14). These features are also known as pebble clus- ate cross-strata millimeters to centimeters thick. ters, imbricate clusters, cluster bedforms, and micro- Planar cross-strata formed by 2D bedforms (Figure forms. Particle clusters can take a variety of shapes, S4.13) are parallel to the bedform slipface. These including comet shaped, triangular, rhomboid, or beds form over a planar erosional surface scoured diamond. Typically, an obstacle clast anchors an by a laterally continuous recirculation cell as the upstream-side accumulation of imbricated particles bedform trough passes downstream. Trough cross- with a tail on the downstream side (Papanicolaou strata formed by 3D bedforms (Figure S4.13) accu- et al., 2003). mulate over a curved erosional surface resulting Particle clusters are most common in gravel-bed from flow separation in the lee of a saddle-shaped channels with low rates of bed load transport and a bedform crest (Bridge, 2003). stable coarse surface layer, and usually occur in riffle, Chapter 4 Fluvial sediment dynamics 109 alternatebar,orplane-bedsectionsofthechannel. the rib crest (Robert, 2003). Spacing between ribs is Although more than one mechanism of formation typically of the order of decimeters to meters, and mayexist,clustersappeartoformduringthereces- heights are one to two clast diameters (Bridge, 2003). sional limb of a flood as bed load is deposited around Ribs appear to be most common on bars in gravel- large, protruding clasts (Brayshaw, 1984). Clusters bed rivers. create abrupt roughness transitions along the bed Transverse ribs likely represent the crests of and increase average boundary roughness (Hassan antidunes (Koster, 1978). Water-surface waves com- and Reid, 1990). This leads to localized flow sepa- monly break upstream of crests, creat- ration and turbulence (Lacey et al., 2007). By creat- ing temporary hydraulic jumps that could promote ing relatively stable points on the streambed, clusters coarse sediment deposition. As with other bed- also delay incipient motion and limit the availability forms,thepresenceoftransverseribsinfluences of bed material for transport (Brayshaw, 1984). boundary resistance and sediment entrainment and transport. Transverse ribs Steep alluvial channel bedforms Transverse ribs, like particle clusters, are microto- pographic features that can be difficult to see when The classification of reach-scale channel morphol- looking at a streambed. Ribs are formed by a series ogy proposed by Montgomery and Buffington of regularly spaced pebble, cobble, or boulder ridges (1997) for steep, relatively laterally confined gravel- perpendicular to the flow. The spacing between ribs bedchannelsislargelybasedondominantbedforms is proportional to the size of the largest clasts in in these channels (Figure 4.9). Channels with the

Debris flows Initiation Scour Deposition

Hillslope Wood Hollow Largely immobile, Largely mobile, traps sediment acts as sediment Colluvial

Cascade

Step–pool Plane–bed Pool–riffle Dune–ripple Transport reaches Response reaches Diffusion Debris flow Fluvial dominated dominated Figure 4.9 The Montgomery and Buffington (1997) channel classification based on dominant bedforms. This schematic illustration of downstream trends includes changes in reach-scale gradient (light gray shading), chan- nel type (cascade/step–pool/plane-bed/pool–riffle/dune–ripple), indication of transport versus response reaches, dominant sediment transport process (hillslope diffusion/debris flow/fluvial), and the behavior of debris flows and instream wood (from Montgomery and Buffington, 1997, Figure 4). 110 Rivers in the Landscape highest bed gradients are typically cascade channels The progression with decreasing reach-scale gra- with disorganized, very coarse bed material and dient from cascade to dune–ripple channels cor- small pools that partially span the channel and are responds to decreasing average bed grain size and spaced less than a channel width apart. In these increasing bed mobility. This represents a gen- channel segments, flow may not be competent to eral downstream trend that can be interrupted or organize the very coarse-grained bed material into reversed. Mountain rivers are commonly longitu- bedforms. dinally segmented and alternate repeatedly down- At gradients of 0.03–0.10 m/m, step–pool chan- stream between high and low gradient reaches nels become particularly common. Although sub- (Wohl, 2010), so that cascade segments can be strate mobility typically remains limited in these downstream from pool–riffle segments, and vice channel segments, flow is competent to create reg- versa. Step–pool and pool–riffle channels are par- ularly spaced bedforms and bed variability in ele- ticularly common and well-studied channel types vation and grain size, primarily in the downstream within this progression. direction. Plane-bed channels lack well-defined, rhythmi- Step–pool channels cally occurring bedforms and are most common at gradients of 0.01 to 0.03. The relatively planar The dominant bedforms in step–pool channel seg- bedformedincoarseclastsorbedrockcanpersist ments are channel-spanning steps of clast, wood, or through conditions of bed stability and mobility. bedrock that alternate downstream with a plunge Despite the similar name, coarse-grained plane-bed pool at the base of each step (Chin and Wohl, 2005; channels are much less mobile than plane-bed sand Zimmermann, 2013) (Figures 4.10 and S4.18). Step– channels. pool sequences are most common where rela- Alternating pools and riffles characterize chan- tively immobile large clasts or wood trap sedi- nels with gradients less than 0.015 m/m, which ment in a wedge tapering upstream, with flow exhibit consistent patterns of bed variability both plunging over the immobile obstacle to scour a downstream and across the channel. Pool–riffle in the bed downstream. Although steps sequences transition at lower gradients to dune– likely form via more than one mechanism, those in ripple channels with sand beds and readily mobile mobile sediments (rather than bedrock) appear to bedforms. result from the interlocking of large, keystone clasts

Interstitial flow

EGL Weir flow

H Oscillating flow

Step Pool Figure 4.10 Longitudinal view of the components and hydraulics associated with step–pool bedforms. H is bedform amplitude: here, the step height. EGL is energy grade line. Interstitial, weir, and oscillating flows represent flow regimes under progressively increasing discharge. Flow is from left to right. (From Dust and Wohl, 2012a, Figure 5b; Dust and Wohl, 2012b, Figure 2.) Chapter 4 Fluvial sediment dynamics 111

(Zimmermann and Church, 2001) under conditions jump disappears and an oscillating or skimming of limited sediment supply, leading to a jammed state flow regime develops in which the water flows as (Zimmermann et al., 2010). a coherent stream and recirculating vortices occur Individual steps can be destroyed by erosion at the base of each step (Chanson, 1996; Dust of the keystones and subsequent mobilization of and Wohl, 2012a) (Figure 4.10). Flow resistance the step (Lenzi, 2001). Step–pool sequences can decreases significantly when flow transitions from give way to braided or plane-bed channels dur- nappe to skimming conditions (Comiti et al., 2009). ingextremelylargedischargesthatmobilizethe Values of resistance coefficients such as Man- entire bed and greatly enhance the sediment sup- ning’s n are very difficult to estimate in step–pool ply. Such mobilization occurs anywhere from annu- channels, which have non-uniform flow. An alter- ally in some channels, to intervals of 50 years or native approach to estimating discharge is to apply longer in other channels (Wohl, 2010). As the mobi- the equation developed for broad-crested weirs, lizing flood recedes and sediment supply once again under the assumption that a step approximates such declines, steps and pools gradually reform during a weir subsequent, smaller flows. These smaller flows can . mobilize the finer sediments exposed at the sur- Q = C∗g0 5Wh3∕2 (4.17) face,butnotnecessarilythekeystoneclasts.An example comes from the 5 km2 RioCordoncatch- where C∗ is a dimensionless discharge coefficient, ment in Italy, where floods with recurrence inter- W is the crest width, g is the acceleration of grav- vals of 30–50 years destroy step–pool sequences ity and h is the upstream flow depth above the step and move individual clasts approximately twice the crest (Dust and Wohl, 2012a). downstream distance of movement during smaller Bed load transport is extremely spatially and tem- floods with recurrence intervals of 1–5 years. Selec- porally variable and very difficult to quantify or tive entrainment and transport during the smaller predict in step–pool channels. Yager et al. (2012) floods gradually reform step–pool sequences (Lenzi, proposed a modified version of the Parker (1990) 2001, 2004). bed load equation to include the resistance asso- Steps and pools create extremely large values of ciated with steps and selective transport of rel- flow resistance relative to lower gradient channels ativelymobilesedimentusingarangeofhiding (Curran and Wohl, 2003). Steps are typically defined functions. by height, H, and downstream spacing, L,andmany Particles in pools are preferentially entrained and step–pool sequences have a ratio of H/L/S (S is transported longer distances. After placing tracer bedgradient)between1and2.Bedformswiththis clasts in a small step–pool stream in Germany, ratiohavebeeninterpretedtomaximizeflowresis- Schmidt and Ergenzinger (1992) found that all tance and channel stability (Abrahams et al., 1995). the tracers placed in pools were entrained during Wood incorporated into clast steps can increase step lower discharges than those in steps. All of the height, backwater effects, and dissipation of flow pool particles were transported, whereas 57%–76% energy (Wilcox et al., 2011). The large grain and of the particles placed in other sites on the bed form resistance of step–pool channels also creates were transported. Pool particles also had greater high turbulence intensities, particularly in pools and transport lengths. Particles finer than or equal to at high discharges (Wilcox and Wohl, 2007). D40 of the bed surface in step–pool channels do Lower discharges in step–pool sequences can not exhibit substantial differences in travel dis- create interstitial flow, with water primarily pass- tance during floods, whereas particles larger than ing through gaps between the step-forming clasts. the bed surface D84 have very limited mobility With increasing discharge, free falls over steps create (Lenzi, 2004). nappefloworweirflow,withahydraulicjumpbelow Step–pool channels are known as transport the step. At the highest discharges, the hydraulic reaches (Montgomery and Buffington, 1997) 112 Rivers in the Landscape because they are relatively insensitive to changes in Melhorn, 1978). Pools are commonly described as water and sediment supply. Field studies support being spaced at approximately five to seven times the idea that, when water and sediment supplied the average channel width. Despite this commonly to a river network change, step–pool bedforms are cited rule, the relation between pool spacing and less likely to alter their dimensions than bedforms channel width is in reality quite variable. Pool spac- present at lower channel gradients (Ryan, 1997; ing varies with substrate, such that pools tend to be Wohl and Dust, 2012). The resistance to change more widely spaced in more resistant substrate such likely reflects the combined effects of high values as bedrock (Roy and Abrahams, 1980; Wohl and of boundary irregularity and flow resistance that Legleiter, 2003). Pool spacing also varies with val- effectively dissipate flow energy, and very coarse, leywidthandassociatedflowconvergence.Working relatively interlocked clasts in steps. on the Yuba River of California, USA, White et al. (2010) found that riffles persisted in locally wide areas of the valley and pools were associated with Pool–riffle channels long valley constrictions over a period of 20 years, Pool–riffle bedform sequences are analogous to despite rapid channel incision and planform change meanderingintheverticaldimension(Kellerand in response to frequent floods. Melhorn, 1978) in that regularly spaced deeps with Pool spacing also varies with the longitu- typically finer sediment (pools) and coarser-grained dinal spacing of obstructions such as wood shallows (riffles) create an undulating longitudinal (Montgomery et al., 1995; Buffington et al., 2002) or profile at the reach scale (Figures 4.11 and S4.19). large, immobile boulders (Lisle, 1986; Thompson, Riffles are usually wider and shallower at all stages 2001), which typically create localized scour and of flow than pools. pool formation. Pool–riffle sequences associ- Pool and riffle bedforms occur in alluvium and ated with local scour and deposition around an in bedrock, and are likely formed and maintained obstruction are sometimes referred to as forced bydiversemechanisms.Theamplitudeandwave- pool–riffle sequences (Montgomery et al., 1995). lengthofthesebedformsistypicallydescribedin Obstructions pond water upstream and create terms of average downstream spacing of pools, flow convergence, higher water-surface slopes, which has been related to channel width (Keller and andhigherflowvelocitythroughtheconstricted zone. This creates localized scour that leads to pools. As the flow passes downstream beyond the EGL obstruction, flow divergence and deceleration along the downstream end of the pool center facilitate sediment deposition, leading to formation of a riffle (Lisle, 1986; Thompson et al., 1998; Thompson, 2013). H Pools and riffles formed in relatively well-sorted alluvium with few obstructions are dominated by

Riffle Pool feedbacks between hydraulics and sediment trans- port,andaremorelikelytoexhibitpoolspacingat Figure 4.11 Longitudinal view of the components and five to seven times channel width. Pools and rif- water surface associated with pool–riffle bedforms. H is bedform amplitude: here, the elevation difference fles formed in poorly sorted, coarse-grained allu- between the riffle crest and the next pool thalweg. EGL vium with abundant obstructions are also strongly is energy grade line, shown here for relatively low flow influenced by hydraulic–sediment interactions, but conditions. During high flow, the EGL over riffles flattens, local variations in channel boundary configura- whereas that over pools grows steeper. Flow is from left tion exert a much stronger influence on bedform to right. (From Dust and Wohl, 2012b, Figure 2.) characteristics. Chapter 4 Fluvial sediment dynamics 113

The locations of freely formed (rather than charges. Cross-sectionally averaged velocity does forced) pools and riffles have been explained in not exhibit velocity reversal, but the presence of termsofatleasttwomechanismsofhydraulicsand strong eddy flow along the margins of pools during bed load movement. In some channels, kinematic high discharges permits the formation of a central jet waves of bed load appear to create bars and rif- of high velocity flow that does exhibit velocity rever- fles where the wave stops moving during the falling sal with respect to riffle velocity (Thompson et al., limb or where transport capacity locally declines 1998, 1999) (Figure S4.20). (Langbein and Leopold, 1968). Accelerating flow Bed load in pool–riffle channels tends to move converging downstream from the bar then scours during competent flows in spatially discrete steps apool.Anotherscenarioinvolvespersistentpoint that are strongly influenced by pool and riffle spac- sources of coarse sediment, such as tributary junc- ing (Thompson et al., 1996; Pyrce and Ashmore, tions, that create riffles (Webb et al., 1987). The rif- 2003). Clasts entrained from a pool during high flow fles create accelerated flow that leads to pool scour are deposited on the next riffle downstream. Clasts downstream (Dolan et al., 1978). In each of these on a riffle can be moved into the next pool down- explanations, creation of an initial bed irregular- stream during waning or low flows as progressively ity in the form of either a pool or a riffle then steeper water-surface profiles over riffles at lower perturbs hydraulics and bed load transport suffi- stages result in dissection of the riffle (Harvey et al., ciently to initiate downstream bed undulations that 1993). lead to successive pools and riffles (Clifford, 1993). In contrast to step–pool sequences, pool–riffle Average longitudinal pool spacing likely reflects a channel segments are response reaches (Montgomery minimum distance related to the backwater and and Buffington, 1997). Because pool–riffle channels turbulent conditions needed for pool formation aremorelikelytobetransportlimitedwithrespect (Thompson, 2012). to sediment, bedform dimensions and bed grain- The velocity-reversal hypothesis has been used size distributions of pool–riffle channels are more to explain the maintenance of pools and riffles likely to change in response to altered water and sed- (Gilbert, 1914; Keller, 1971). Field observations iment discharge than in supply-limited step–pool indicate that the near-bed velocity increases more channels. Increased sediment load typically causes rapidly with discharge in a pool than in a riffle, preferential pool filling. Filling pools reduces form sothatflowinpoolsismorecompetentathigh roughnessandcreatesamoreuniformreach-scale stagethanflowoverriffles.Thisexplainsthecom- bed gradient and flow depth that enhance the abil- mon observation that pools scour at high flow and ity of moderate flows to transport bed load (Lisle, fill at low flow, whereas riffles are depositional 1982). Decreased sediment load and/or increased sites at high flow. The velocity reversal has been discharge can lead to enhanced pool scour and difficult to explain, however, given the tendency bank erosion (Wohl and Dust, 2012). The bedform of pools to have greater cross-sectional area (and amplitude and wavelength of pool–riffle sequences therefore presumably lower velocity) than riffles. canthusadjusttovariationsinflowandsediment Hydraulic modeling of the River Severn in Eng- supply. land indicated that channels in which pools are An important implication of the distinction hydraulically rougher than riffles during high flow, betweentransportandresponsereachesisthat and riffles are substantially wider than pools, are diverse types of infrequently mobile bedforms are morelikelytoexhibitvelocityreversal(Carlingand not likely to respond uniformly to changes in water Wood, 1994). and sediment yield associated with changes in cli- Numerous field observations and flume exper- mate,landcover,orotherexternalcontrols.Iftimber iments illustrate the difficulty of demonstrating a harvest within a catchment increases sediment yield, velocity reversal, at least in part because of the dif- for example, bedform dimensions in the lower gradi- ficultyofmeasuringbedvelocityduringhighdis- ent response reaches will change disproportionately, 114 Rivers in the Landscape while those in the higher gradient transport reaches the height and lateral thickness of a single cobble or may remain relatively unaffected. boulder, or can be more than a meter tall and wide. River reaches with different bedforms exhibit dif- Berms tend to be persistent features that form dur- ferences in sensitivity (the ability to react to a stim- ing high discharges and remain during subsequent ulus) and resilience (the ability of a river to return lower discharges. to initial conditions following a disturbance such Alternate bars occur on alternating sides of the as a flood or debris flow) (Brunsden and Thornes, channel in a downstream progression. Alternate 1979). These reach-scale differences within a river bars can form in mixed grain-size channels when network have important implications for chan- concentrations of coarse clasts at the downstream nel stability, aquatic and riparian habitat, water end of bed load pulses aggrade along portions of quality, and river management (Montgomery and the channel with high flow roughness (Lisle et al., Buffington, 1997; Wohl et al., 2007). Sensitive chan- 1991). Alternate bars are typically asymmetrical nel segments may be less stable, for example, longitudinally and generally migrate in the down- whereas resilient segments may be more stable. stream direction (Bridge, 2003) as particles being By influencing flow velocity and depth, sub- transported along the top of the bar reach the strate size and stability, dissolved oxygen, hyporheic downstream edge and cascade down the bar front. exchange, and solute fluxes, infrequently mobile ThesebarsarediscussedmoreinSection5.2as bedforms also influence habitat and the distribu- part of the treatment of meandering and braided tion of aquatic biota. In mountainous headwater channels. channels, for example, pool–riffle channel segments Point bars form along the inside edge of mean- typically have greater pool volume and fish habi- der bends, where fine grains are swept inward over tat than step–pool or cascade channels (Moir et al., the and coarse grains are routed out- 2004). ward toward the pool (Clayton and Pitlick, 2007). Depending on the channel substrate and sedi- ment supply, point bars can be formed of sand to Bars boulder-size grains, and are scaled to channel width. Diverse types of bars occur in rivers, including Pointbarsareessentiallyalternatebarsinsinuous berms, alternate bars in straight channels, point bars channels. in sinuous channels, transverse bars that form rif- Point bars migrate and adjust in size and fles or rapids, braid bars or mid-channel bars in geometry as sediment supply and the distribution braided channels, and bars at tributary–main chan- of hydraulic forces fluctuate with discharge, and nel junctions. Bars have lengths comparable to chan- as meander geometry changes through time. An nel width. Bars form when flow energy is sufficient extreme example comes from the highly sinuous to transport bed material and simultaneously scour Fall River in Colorado, USA, which received a huge other portions of the channel bed, such as pools. As influx of sediment following a damburst flood in the with other bedforms, bars are both influenced by headwaters. The flood waters eroded large volumes boundary resistance, hydraulic forces, and sediment of glacial outwash sediment from the valley bottom dynamics, but also influence roughness, hydraulics, en route to the meandering reach, where the result- and sediment dynamics. ing sediment supply increased by about 1000 relative Berms are accretionary features that form at sites to normal sediment influx (Anthony and Harvey, of energy loss associated with hydraulic jumps or the 1991). During the 4 years following the damburst, shear zone of flow separation (Carling, 1995) (Fig- point bars aggraded to the bankfull water surface ureS3.16).Bermscanformdownstreamfromlateral during higher flows, then eroded laterally and ver- channel constrictions or at the downstream end of a tically during periods of lower flow when sediment plunge pool below a vertical step such as a water- supply from upstream decreased. fall. Berms can be relatively small features that are Point bars are also discussed more in Section 5.2. Chapter 4 Fluvial sediment dynamics 115

4.3.3 Bedforms in cohesive 10000 sediments Erosion 1000 Characteristic repetitive bed undulations occur in Erosion velocity streambeds with sufficient silt- and clay-sized par- 100 ticles to act as cohesive materials. These bed- Transport Deposition Velocity (m/s) forms are strictly erosional and display a sequence 10 with increasing flow strength analogous to that of lower- and upper-regime bedforms for sand-bed 1 0.01 0.1 1.0 10 100 1000 channels. In cohesive streambeds, straight longitudinal Clay Silt Sand Gravel grooves and ridges form initially (Allen, 1982). Grain diameter (mm) These features are parallel to mean flow direction Figure 4.12 Hjulstrom’s¨ diagram of erosion, transport and reflect streaks in the viscous sublayer. Flute and deposition thresholds for varying grain sizes as a marksappearasvelocityincreases.Flutemarksare function of velocity. (From Hjulstrom,¨ 1935.) shaped like the depression in a spoon and reflect a site where locally enhanced near-bed turbulence vegetation, or with irregularities in the channel mar- results in differential bed erosion and flow separa- gins such as tributary junctions or bends. Local tion (Bridge, 2003). Further velocity increase creates deposition alters boundary roughness, grain-size transverse ridge marks shaped like ripples, which distributionsatthechannelsurface,andsubsequent reflect supercritical flow (Bridge, 2003). particle entrainment. Local deposition that occurs in As with depositional bedforms in non-cohesive isolated but numerous locations across the channel sediments, hydraulics can be inferred from the type produces features such as particle clusters, ripples and dimensions of erosional bedforms in cohe- and dunes, and various types of bars. sive sediments. Richardson and Carling (2005) pre- The size and spacing of bars, in particular, can sented a catalog of cohesive bedforms, sometimes change with sediment supply and flow energy. An known as sculpted features,characteristicofbedrock example comes from a headwater tributary of the channels (Figure S4.21). River Tay in Scotland, where a sediment wave gener- ated by lateral channel movement caused bar aggra- 4.4 In-channel dation in some reaches. Bar aggradation resulted in flow confinement and erosion from adjacent and depositional processes downstreamportionsofthechannel,aswellastrim- ming of bar margins, thus limiting the growth of the Each of the bedforms discussed in Sections 4.3.1 and bars (Wathen and Hoey, 1998). 4.3.2 is also a depositional feature, particularly when Deposition across the entire channel is facilitated the bedform stops moving. Deposition occurs when by channel-spanning obstructions such as logjams the flow or shear velocity falls below the settling and beaver dams. These obstructions can create suf- velocity of a particle. This is a lower threshold than ficient backwater to accumulate a wedge of sedi- that required for entrainment (Figure 4.12). Deposi- ment as thick at the downstream end as the obstruc- tion can be highly localized and limited to a zone of tion is tall, with the wedge tapering upstream in lower velocity and flow energy, or more widespread a manner dictated by channel gradient and sedi- across and down a length of channel. ment supply (Figure S4.22). Sediment accumulating Local deposition occurs in association with upstream from the obstruction is likely to be finer isolatedroughnessfeaturessuchasprotruding grained than sediment elsewhere on the streambed, boulders, instream wood, or aquatic or riparian enhancing the diversity of aquatic habitats and 116 Rivers in the Landscape the storage of fine organic material. The pressure this process as vegetation-mediated landform devel- gradient associated with backwater at the obsta- opment. cle also enhances hyporheic exchange within the Changes in vegetation density can reflect natu- deposited sediment (Hester and Doyle, 2008). A ral fluctuations in flood magnitude and frequency. very large obstruction such as a sediment dam asso- Working on ephemeral and intermittent channels in ciated with a rockfall, landslide, or debris flow cre- thesemiaridsteppeoftheUSGreatPlains,Friedman ates a correspondingly large deposit, although natu- and Lee (2002) documented repeated alternation in ral dams caused by slope instability typically breach channel geometry and riparian vegetation through within a year (Costa and Schuster, 1988). Some time. Channel widening and removal of vegetation types of large-scale natural obstructions, such as occurred over a period of hours during infrequent lava dams, can persist for 102–103 years, but many floods caused by convective storms. Post-flood nar- smaller-scale obstructions such as logjams or land- rowing associated with regrowth of riparian forests slides may be present for less than a year or at most occurred over decades between floods. The channels afewyears. alternated through time between broad, braided, Someportionsofthechannelthatserveasdepo- unvegetated rivers and narrow, sinuous, forested sitional sites at lower discharge can be scoured as rivers. dischargeincreases.Poolstypicallyaccumulatefiner In many cases, vegetation density changes as a sediment during low flow, but converging flow off a result of invasive exotic species (Graf, 1978) or a riffleoracentraljetdevelopinginalaterallycon- reduction in flood peaks associated with flow reg- stricted pool results in removal of sediment dur- ulation that allows germinating seedlings to suc- ing high flows. Similarly, lee deposits accumulating cessfully colonize lower portions of channel mar- around protruding clasts during lower discharges gins. As a result of flow regulation, larger, perennial can be mobilized as hydraulic forces around the rivers in the US Great Plains metamorphosed from clasts change with increasing discharge (Thompson, braided, unvegetated rivers present in the early nine- 2008). teenth century, to sinuous or anastomosing, densely Riparian vegetation can be especially effective in forested rivers by the mid-twentieth century (Nadler facilitating deposition within the channel and along and Schumm, 1981). the channel margins, and increases in the density Cross-channel, longitudinally continuous depo- of vegetation can facilitate sediment accumulation sition typically represents a reduction in transport and channel narrowing (Tal et al., 2004). An exam- capacity or an increase in sediment supply. Trans- ple comes from the braided, gravel-bed Tagliamento port capacity can decline as a result of decreased dis- River of Italy, where Gurnell and Petts (2006) docu- charge or decreased gradient. Sediment supply can mented how deposition of large pieces of driftwood increase following climate change, hillslope instabil- or entire trees that are capable of regrowth influences ity, volcanic eruptions, or changes in land cover or sediment deposition. These trees are deposited dur- land use. The 1991 eruption of Mount Pinatubo, for ing waning flood stage, grow quickly and, where suf- example, covered a third of the Pasig-Potrero River ficient sand and finer sediment is available, facilitate watershed in the Philippines with sand-sized pyro- island and floodplain development by trapping sedi- clastic flow deposits. This abrupt change in sediment ment and allowing more vegetation to become estab- supply caused the gravel-bedded river to become lished. a planar, highly mobile, sand-bedded channel that Establishment of riparian vegetation can initiate a gradually transitioned back toward a gravel channel self-enhancing feedback as plants trap and stabilize over the succeeding 15 years (Gran et al., 2006). sediments and organic matter than can provide ger- Cross-channel, longitudinally continuous depo- mination sites for other plants, which in turn rein- sition can occur as a relatively minor or tempo- force the development of floodplains, islands, and rary effect that results in finer sediment being other landforms. Gurnell et al. (2012) referred to depositedonthesurfaceofacoarse-grainedchannel Chapter 4 Fluvial sediment dynamics 117

(Figure S4.23). Although the effects on cross- Given the tendency to build immediately adjacent to sectional geometry can be minor, fining of the bed channels and then attempt to prevent bank erosion, sediment can alter hyporheic exchange and habi- mucheffortisdevotedtoanalyzingandmodeling tat suitability for benthic organisms such as algae, bank stability. aquatic insects, or larval fish. As with bed erosion, the forces driving sediment More substantial and sustained deposition removal from banks must exceed the resisting forces throughout a channel is known as aggradation for banks to erode. This becomes complicated to and can result in sufficient loss of channel cross- assess in part because stream banks are commonly sectional area to cause enhanced overbank flooding layered, with coarser sediment at the bottom as a and lateral channel movement or the formation of result of deposition within the channel and finer secondary channels. A common cause of aggra- sedimenttowardthetopofthebankasaresultof dation leading to loss of cross-sectional area is overbank deposition (Figure S4.24). Bank sediment the clearance of native vegetation and start of typically grows progressively finer grained down- agriculture within a drainage basin, as discussed in stream or along channel segments of lower gradient, more detail in Supplemental Section 4.4. butbankscanbeveryspatiallyheterogeneous.The finer sediments in the upper bank add cohesion, and the roots of riparian vegetation can further enhance 4.5 Bank stability the resistance of the bank sediment to erosion (Fig- ure S4.25). and erosion The strength of stream banks reflects the fric- tional properties of the bank sediment, effective The ability of flow to erode the banks of a channel, normal stress, and effective cohesion (Simon et al., and the processes and rates of bank erosion, strongly 1999). Effective normal stress, 휎′, is the difference influence sediment supply, channel geometry, ripar- between normal stress (the perpendicular com- ian vegetation, and rates of lateral channel migra- ponent of total stress, 휎) and pore pressure, u: tion. Bank erosion thus represents, like bedforms, a in other words, 휎′ = 휎 − u.Effectivenormal means of adjusting channel geometry, external resis- stress results from static friction that keeps parti- tancetoflow,andsedimentsupply. cles together in a stream bank minus the effect of Bank erosion can result in property loss and dam- pore pressure that keeps particles separate. Effec- age to infrastructure such as buildings and bridges. tive cohesion results from true cohesive forces, from Human alterations of water and sediment supply to matric suction within the unsaturated portion of the the channel commonly result in altered rates of bank bank, and/or from root reinforcement via vegetation erosion. Even unaltered, “natural” bank erosion is (Eaton, 2006). Acting against these resisting forces inmanycasesregardedasanindicatorofchan- are the driving forces of gravity and hydraulic force nel instability and a sign that mitigation is needed expressed as shear stress. to limit bank erosion. This represents a fundamen- Pore water contentandpressureisoneofthemore tal misunderstanding of river processes. A stable importantinfluencesonbankstability(Rinaldi channel transporting bed material must entrain an and Darby, 2008), and involves multiple effects. amount of sediment equivalent to that deposited. Pore water reduces shear strength by increasing Most in-channel deposition occurs in bars, and most lubrication between sediment particles. Pore water “compensating” erosion comes from banks as the increases the unit weight of the bank material, mak- channel expands to maintain conveyance. Conse- ing unsupported material more susceptible to fail- quently, local lateral instability in the form of local ure. Pore water destabilizes banks by facilitating the bank erosion is a natural feature of stable rivers presence of water in tension cracks. Pore water also transporting bed material, but is commonly mis- creates seepage forces that can stabilize or destabilize interpreted as a sign of general channel instability. the banks. 118 Rivers in the Landscape

Pore water pressure results from the pressure of water filling the voids between particles. Negative pore water pressures above the water table reflect the surface tension of pore water in voids, which creates a suction effect on surrounding particles and stabi- lizes the banks. Positive pore water pressures below the water table, in contrast, help to force particles apart and destabilize banks. Pore water pressures are extremely transient in response to changes in precip- itation and stream flow, but they typically promote bank failure during the falling limb of the hydro- graph when the bank sediment is at or near satura- tion and the confining pressure of the river water is removed. Figure 4.13 Scalloped bank along a small creek in Con- Vegetation also influences stream banks in necticut, USA, caused by locally increased bank resis- diverse ways (Merritt, 2013). Plants increase the tance as a result of tree roots. mass of stream banks and can thus facilitate bank failure. However, plants near the water’s edge create flow resistance that reduces near-bank velocity and are gradually undercut and topple into the chan- shear stress and increases bank stability (Griffin nel, resulting in a larger volume of bank erosion et al., 2005; Gorrick and Rodr´ıguez, 2012). Plant and a new round of the cycle (Pizzuto et al., 2010) roots increase the resistance of the sediment to (Figure 4.13). shearing (Pollen and Simon, 2005). And, plants alter The primary processes eroding stream banks bank pore water pressure by affecting infiltration, involve hydraulic action or mass failure. Fluvial evaporation, and transpiration (Griffin et al., 2005; detachment via hydraulic action is likely to be Pollen-Bankhead and Simon, 2010). Vegetation more important in coarse-grained or smaller rivers. exerts the greatest influence on bank stability along Mass failure becomes progressively more important low, shallow banks in weakly cohesive sediments downstream as bank heights increase and bank sed- and along small channels (Eaton and Millar, 2004). iments become finer grained (Lawler, 1992). Not all vegetation is created equal with regard to Hydraulic action results from shear stress exerted bank stability. Channels with forested banks tend against a bank and is related to the near-bank veloc- to be wider than channels in grasslands for at least ity. Individual particles can be detached from the tworeasons.First,woodrecruitedtothechannel bank face in non-cohesive or weakly cohesive sed- from the riparian zone promotes channel widening. imentandthebaseofthebankcanbeprefer- Second, grass grows readily on point bars, facilitat- entially eroded (Rinaldi and Darby, 2008). Even ing more rapid deposition of suspended sediment cohesive sediments can be made more susceptible and narrower channels (Allmendinger et al., 2005). to hydraulic action by processes that weaken and Among woody vegetation, densely growing species detach sediment, such as shrink–swell or freeze– such as willows (Salix spp.) can limit bank erosion thaw cycles (Wynn et al., 2008). Conversely, flow and channel widening more effectively than species resistance associated with riparian vegetation can in which individual plants are more widely dispersed reduce bank erosion through hydraulic action. (David et al., 2009) (Figure S4.26). Hydraulic action is typically quantified using an Cyclic bank erosion can occur over decades when excess shear stress equation similar to Equation 4.14, small volumes of sediment are removed between with fluvial bank erosion rate per unit time substi- large trees on the bank, creating a scalloped bank tutedforbedloadtransportrate(RinaldiandDarby, morphology buttressed by large trees. The trees 2008). Chapter 4 Fluvial sediment dynamics 119

Processes that weaken and detach bank sediment also promote mass failure through slumping or top- pling of a slab. Seepage also reduces the cohesion of bank sediment by removing clay and, in some cases, promoting piping in the bank. Seepage can be particularly effective where rapid reductions in flow stage leave saturated banks without lateral sup- port, a situation common in rivers regulated for hydropower generation. Trampling by large num- bers of grazing animals congregating along streams can further enhance bank erosion by breaking down the banks and creating hydraulic roughness (Trimble and Mendel, 1995). Mass failure of banks takes several forms (Osman Figure 4.14 Fine-grained bank collapse block that will and Thorne, 1988; Knighton, 1998). Shallow slips act as an isolated roughness element until the vegeta- in which relatively thin segments of bank detach tion holding the sediment together decays. This block, andthendisintegratedominateinnon-cohesivesed- indicated by the white arrow parallel to flow, is about iment. Slab-type failures in which a vertical slab 40 cm in diameter, and is much larger than the average detaches from the bank and then topples are most size of the bed substrate. common in weakly cohesive sediment and near- vertical banks. Deep-seated rotational slips domi- nate in cohesive sediment. In these failures, a mass of sediment slides down the bank along a curved fail- with the floodplain surface (Van De Wiel and Darby, uresurface,rotatingsothatthetoeofthefailurepro- 2007). trudes into the river. Cold-region rivers with a seasonal ice cover can All forms of mass failure are enhanced by scour have a distinctive bank morphology in the form of at the base of the bank that oversteepens the bank. a two-level bank structure that reflects ice scouring In strongly layered banks, removal of the coarser, (Boucher et al., 2012). Elevation of the ice surface non-cohesivesedimentinthelowerbankcancre- during freeze-over and jamming of ice blocks dur- ate overhangs that eventually collapse as blocks into ing break-up lead to abrasion of the banks by ice and thechannel.Suchblockscaneitherquicklybreak the formation of a steep segment of bank above the up and disperse, or persist for more than a year, in bankfull stage. Ice gouging and overbank sedimen- part because of the dense, fine plant roots within tation during ice-jam flooding can create an elevated the block. Persistent blocks act similarly to boul- ridge or bench along some rivers, which is referred ders and can protect the lower bank from further to as a bechevnik from the Russian word for tow rope, erosion or deflect the current toward the bank toe because these benches formed convenient paths for and enhance basal scouring (Figures 4.14 and S4.27). towing boats upstream along Siberian rivers (Ettema Bank collapse via overhanging blocks is particularly and Kempema, 2012). common in permafrost regions where slow thawing Severalmodelsquantifyandpredictbankstabil- of the soil ice facilitates the persistence of collapsed ity, as reviewed by Pizzuto (2003) and Rinaldi and blocks (Walker et al., 1987), and in wet meadows Darby (2008). Models can be differentiated as mech- where bank erosion creates peat blocks (Warburton anistic models based on the physics of particular ero- and Evans, 2011). Root reinforcement of bank sedi- sional processes, and parametric models that relate ment by riparian vegetation can limit mass failure, bank erosion to potential controlling parameters— particularly where the vegetation grows along the typically near-bank velocity or shear stress—using bank toe or at the intersection of the failure plane empirical coefficients (Pizzuto, 2003). 120 Rivers in the Landscape

Bank erodibility parameters are modeled using stress, uw istheporewaterpressure,(ua − uw)is methods similar to those for entrainment of bed sed- the matric suction, and 휙b is the angle expressing iments, with modifications to account for the effect therateofincreaseinstrengthrelativetothematric of bank angle on the downslope component of par- suction. ticle weight and for partly packed and cemented Parameter uncertainties in bank stability mod- sediments (Rinaldi and Darby, 2008). Near-bank elsaretypicallysolargeasaresultofnaturalvari- hydraulic force is either directly measured or esti- ability in the parameters that the likelihood of gen- mated using hydraulic models. Models developed erating unreliable predictions exceeds 80% (Samadi for non-cohesive banks include the mechanistic et al., 2009). In addition, processes that weaken model of Kovacs and Parker (1994), which uses and strengthen banks interact in sometimes unpre- a bed load transport model to compute bed load dictable manners. Consequently, rather than use a transport on steeply sloping banks coupled with deterministic model with a single value for each the near-bank velocity. Because vegetation, mois- bank material property, a probabilistic representa- ture, and even small amounts of fine sediment add tion of effective bank material strength parameters cohesion to banks, most models focus on cohesive may be the most appropriate approach (Parker et al., banks. 2008). Mechanistic models for cohesive banks are gen- Supplemental Section 4.5 discusses measuring erally 2D models that evaluate bank stability in bank stability and erosion. terms of the soil strength and bank geometry (Pizzuto, 2003). For example, Simon et al. (2000) developed a bank stability algorithm and the mecha- 4.6 Sediment budgets nistic bank stability and toe erosion (BSTEM) model for layered, cohesive stream banks. They combined A sediment budget quantifies fluxes of sediment past the Coulomb equation for saturated banks with a given point in a watershed. Sediment budgets are the Fredlund et al. (1978) equation for unsaturated basedontheverysimpleformulaofinputs(I)minus banks. The Coulomb equation is thechangeinmassorvolumeofsediment(ΔS) stored in the channel reach during a specified time 휎′ 휙 St = c + tan (4.18) interval,

휙 휑 where St is shear strength, c is cohesion, is the I −ΔS = (4.20) angle of internal friction, and 휎′ is the effective nor- mal stress. Total normal stress, which tends to hold where 휑 isthemassorvolumeofsedimentoutput sediment together, is the sum of effective normal from the channel reach during the specified time stress and pore pressure. Under saturated condi- interval. 휑 is also known as sediment yield. Sediment tions, pore pressure is positive and effective normal inputs represent those coming from upstream and stress is lower. In partially saturated soils, effective subsurface sources on the main channel and trib- normal stress is increased. Consequently, bank sed- utaries, as well as from sources beyond the chan- iment is more susceptible to mass failure under sat- nel, including hillslopes, glaciers, terraces and other urated conditions (Robert, 2003). valley-bottom deposits, and eolian inputs. Storage The failure criterion of Fredlund et al. (1978) is occurs within the channel bed, banks, bedforms, and inoverbankareassuchasthefloodplain.Outputis 휏 ′ 휎 휙′ 휙b = c + ( − ua) tan + (ua − uw) tan (4.19) transported within the channel as dissolved, wash, suspended, and bed load (Figure 4.15). where 휏 is the shear strength, c′ is the effective cohe- The mathematical simplicity of Equation 4.21 is 휎 sion, is the normal stress, ua is the pore air pres- deceptive because each of the three primary vari- sure, 휙′ is the friction angle in terms of effective ables can fluctuate greatly through time, across Chapter 4 Fluvial sediment dynamics 121

I – ΔS = φ

Ieolian

Iglacier Iuplands

Itributary

Iupstream

Iterrace, fan, flooplain Sfloodplain φ dissolved φ suspended φ bed load

Sbed, banks, bedforms

Figure 4.15 Schematic illustration of the basic components of a sediment budget. Inputs I to a river segment can come from uplands and glaciers, eolian deposition, mainstem and tributary transport, or erosion of terraces, alluvial fans, floodplains, and other valley-bottom deposits. Storage S of sediment can occur in the floodplain or in various portions of the channel. Sediment outputs 휑 occur via downstream transport as dissolved, suspended, and bed load. a drainage basin and between basins, making it In large river basins as diverse as the Amazon extremely difficult to accurately quantify input, stor- and the Mississippi, one portion of the catchment age and output even over relatively short time inter- supplies the great majority of total sediment output vals. Sediment inputs can be gradual, as in slope (Figure 4.5), as noted earlier. In the case of the Ama- wash, average tributary sediment transport, or soil zon, the Andes of Peru and Bolivia supply more than creep.Or,sedimentinputscanbeabrupt,asindebris 80% of the sediment load but constitute only ∼10% flows or tributary flash floods. Inputs can also be of the basin area (Meade et al., 1985; Meade, 2007). seasonally driven, aperiodic, or variable over time Therelativeimportanceofdiversesediment spans of thousands of years because of fluctuations sources varies greatly among basins. High-relief in base level, climate, or land use. basins tend to be dominated by hillslope sources. Sediment inputs tend to be temporally and spa- Laterally mobile rivers with extensive floodplains tially heterogeneous in a wide variety of basins. can obtain sediment predominantly from bank ero- Research in mountainous basins provides exam- sion. High-latitude basins can be dominated by ples: 75% of the long-term sediment flux in Taiwan glacial sources (Gurnell, 1995). occurs during typhoon-generated floods occupying Therelativeimportanceofdiversesediment <1% of the flow duration curve (Kao and Milliman, sources also varies among solute, wash, suspended, 2008). Over a period of 70 years, half of the sedi- andbedload.Inmostcatchments,solutescomepri- ment load from a river in California, USA, was deliv- marily from groundwater inputs. Bank sediments ered in less than 5 weeks (Farnsworth and Milliman, can contribute nearly 40% of the total suspended 2003). Sediment yield in drainages with periodic dis- sediment, even in relativelylow-energycatchments turbances such as volcanic eruptions and wildfires is (Walling et al., 1999), with up to 80% of the total sus- likely to be dominated by episodic inputs from both pendedsedimentyieldcomingfrombanksources disturbed hillslopes and continuing channel insta- in some highly unstable, incised channel networks bility (Gran et al., 2011). (Simon and Darby, 2002). Bed load can come 122 Rivers in the Landscape primarily from mobilization of streambed sedi- because such catchments have proportionally less ments, inputs from smaller, steeper tributary catch- sedimentstorageinfeaturessuchasfloodplains. ments or, particularly in high-relief catchments, All rivers with large sediment loads originate in from upland sources such as hillslope mass move- mountains, and the majority of sediment load within ments. most river basins comes from the mountainous por- The predominant source of sediment can also tion of the basin (Milliman and Syvitski, 1992). The vary through time. This is illustrated by the Upper correlation between sediment load and small, steep MississippiRiver,USA,wherethedominantsource catchments reflects the fact that such catchments of suspended sediment shifted from agricultural soil typically have higher within-hillslope connectivity, erosion in the mid-twentieth century to accelerated hillslope–channel connectivity, and within-channel erosion of stream banks since 1980 as discharge has connectivity for surface water and sediment. Steep increased from climate change and channelization hillslopes, numerous first-order channels and steep, (Belmont et al., 2011). narrow valley bottoms limit sediment storage in The existence of sediment storage is reflected colluvial (hillslope) and alluvial sites, creating effi- in the sediment delivery ratio.Thisratiorepresents cient delivery of sediment into and through river thedifferencebetweenvolumeofsedimentgener- networks. Magnitudes and timescales of sediment ated and volume of sediment stored or transported storageincreaseinlargerbasinswherewideval- from the basin, typically as a function of drainage ley bottoms buffer channels from hillslope inputs area, A and provide larger storage areas with terraces, flood- plains, and alluvial fans. 훾 = 훼A휑 (4.21) The influence of climate on sediment inputs appears in plots of sediment yield versus a variable where 훾 is sediment delivery ratio and 훼 and 휑 are such as effective precipitation. (Effective precipita- empirical parameters. Sediment delivery ratio is typ- tion is the precipitation that actually contributes to ically calculated from observed sediment yields (y) runoff.) Such plots originated with Langbein and andmeasuredorestimatedgrosserosionrates(e): Schumm (1958), who used sedimentation records from the United States to demonstrate that sedi- y = 훾e (4.22) ment yield peaks at ∼300 mm effective precipita- tion. Lesser values of precipitation produce insuffi- Combining Equations 4.22 and 4.23 results in cient runoff to mobilize large quantities of sediment. Continuous vegetation cover effectively limits ero- 휃 y = bA (4.23) sion at higher precipitation values. Subsequent work indicates a secondary sedimentation peak in the sea- where y is areal average sediment yield and 휃 is an sonal tropics. Despite the mostly continuous vegeta- empirical parameter that varies from −0.52 to 0.12 tion cover in the tropics, the great intensity of pre- (Milliman and Meade, 1983; Lu et al., 2005). cipitation may limit the ability of vegetation to retard Sediment delivery ratio is typically between 0 and runoff and surface erosion or, in steep terrains, mass 1, indicating that only a fraction of the total sedi- movements such as landslides. ment detached from eroding sources actually leaves The influence of lithology on sediment yield a catchment. Although sediment delivery ratio is reflects relative rates of bedrock weathering and ero- commonly related to drainage area, other factors sion,aswellasthecharacteristicsofweathering such as topography, land cover, land use, lithology, products. Other factors being equal, more erodi- andcyclicchannelprocessessuchasgullyerosion ble and/or clastic sedimentary rocks tend to yield and filling can strongly influence sediment deliv- greater sediment volumes than rocks more resis- ery ratio (Wohl, 2010). In general, smaller, steeper tant to weathering and erosion. Chemical sedimen- catchments have a greater sediment delivery ratio tary rocks such as limestone also typically have low Chapter 4 Fluvial sediment dynamics 123 yields of granular sediment because most weather- portingthesediment.Oncetheentrainmentthresh- ingproductsmoveasdissolvedload. old in a channel is exceeded, sediment transport The dominant influence on sediment yield in tends to increase with flow duration. The magnitude many catchments is land use. Hooke (2000) esti- of the entrainment threshold thus exerts an impor- mated that humans move more sediment glob- tant control on sediment output. The majority of ally than any other geomorphic agent, and sedi- thesedimentoutputinsand-bedchannelswithrel- ment yields from areas where natural vegetation and atively low entrainment thresholds can occur dur- topography have been altered are among the highest ing relatively frequent flows (Wolman and Miller, intheworld.SupplementalSection4.6discussesin 1960). In contrast, the majority of sediment out- more detail how diverse human activities influence put in very coarse-grained or otherwise resistant- sediment yield and how sediment yield is measured boundary channels may occur during infrequent, or estimated. high magnitude flows (Wohl, 1992). The volume and residence time of sediment Casestudiesillustratesomeofthespatialand storage in depositional features can also vary temporal complexities of sediment budgets. greatly. Residence time of floodplain sediments, for example, reflects the balance between rates of r Trimble (1983) examined historical changes in the sediment accumulation through vertical and lateral sediment budget of Coon Creek in Wisconsin, accretion, rates of floodplain erosion through bank USA, using: measured suspended and bed load migration, and the size of the floodplain (Section yields from Coon Creek, reservoir sedimentation 6.1). The mainstem Amazon in Brazil is bordered data from a nearby analogous basin, estimates of by ∼90,000 km2 of floodplain, and floodplain upland sheet and rill erosion based on the Univer- width increases downstream. Mertes et al. (1996) sal Soil Loss Equation, sediment cores from valley estimated mean residence times for floodplain bottoms,andmappedhistoricalaccumulationsof sediments in this portion of the river as being sediment from aerial photographs. Examining the ∼1000–2000 years, and reaching periods as long as period 1853–1975, he estimated that about half of 10,000 years. the upland sediment brought into the river net- Residencetimeofsedimentinfloodplains work as a result of agricultural development went and along river networks typically increases with into floodplain storage, with less than 7% of the drainage area. At larger drainage areas, flows of sediment transported out of the basin. sufficient magnitude to “turn-over” (i.e., erode) r Similarly, Fryirs and Brierley (2001) found that theentirefloodplainandthemajorityofstorage lowland portions of the 450 km2 Bega River catch- sites along the channel are very rare (Harvey, 2002), ment in southeastern Australia have stored more in part because precipitation capable of creating than 80% of the sediment mobilized from the a high percentage of contributing area within a uplands since European settlement of the catch- drainagebecomesinfrequentatverylargedrainage ment in the mid-nineteenth century. Fryirs and areas. A relatively small convective cell can produce Brierley used historical ground and aerial pho- heavyrainfalloverallofasmalldrainagebasin, tographs, stratigraphy and sedimentology of val- increasingthechancesthattheentirebasinwillbe ley bottoms assessed from geomorphic mapping contributing and that widespread erosion will occur. and sediment cores, and sediment delivery to Such basin-wide precipitation is more common in theestuarytodevelopanalluvialsedimentbud- small basins, creating shorter residence times for get through time for sand and fine gravel in the sediment and higher delivery ratios. catchment. Sediment output varies not only with the occur- renceofprecipitationandflowscapableofmobiliz- These examples illustrate how strongly deposi- ing sediment in colluvial and alluvial storage areas, tional features can influence sediment yields and but also with the duration of flows capable of trans- sediment budgets. In this context, depositional 124 Rivers in the Landscape features such as floodplains reduce upland-channel portrelationbecauseofeffectssuchasshield- and river network sediment connectivity by storing ing and protrusion and limited sediment supply. substantial amounts of sediment. Mobilization of particles from beds of mixed grain Supplemental Section 4.6 discusses methods of sizesoccursinphases.Equationsusingtwo-fraction measuring and modeling sediment budgets and transport or grain–grain interactions better capture outputs. the processes operating in these channels. The distinction between readily mobile and infre- quently mobile bedforms reflects basic differences in thresholds of entrainment and mobility between 4.7 Summary sand-bed and other channel types. The readily mobile bedforms of ripples, dunes, and antidunes Natural channels differ from engineered channels respond to a range of flow magnitude. Infrequently in being able to adjust boundary roughness, channel mobile bedforms, including particle clusters, trans- geometry, and sediment transport in response to verse ribs, step–pool and pool–riffle sequences, changes in water and sediment supplied to the change most actively under relatively high flow mag- channel. Most of these adjustments involve sedi- nitudes. ment dynamics—the entrainment, transport, and Sedimentcanalsobeerodedfromordeposited deposition of individual particles and aggregates of along stream banks. As with streambeds, bank ero- particles. sion reflects the frictional properties of the sedi- Historically, quantification of sediment dynam- ment. Bank erosion is also strongly influenced by ics began with the simple scenario of a channel bed bank stratigraphy, pore water pressure, and riparian with relatively uniform, sand-sized grains that could vegetation, and is more likely than bed erosion to be readily collected and sieved to quantify the grain- taketheformofmassfailure.Aswithbedsediment, size distribution. Sand-bed channels have a relatively efforts to quantitatively model and predict bank low entrainment threshold and can be adequately erosion are limited by large spatial and temporal represented by bulk sediment transport equations. variability in sediment properties and in the forces The entrainment and transport equations developed acting on the sediment. forsand-bedchannelsarebasedonestimationof A mass balance of sediment in the form of a sed- the magnitude of shear stress above a critical value iment budget reflects the inputs, storage, and out- necessary to mobilize sediment, as formulated by puts of sediment across a reference area that can Shields. These equations remain the starting point vary from a small subcatchment to the Amazon. for characterizing sediment dynamics. Regardless of spatial scale or geographic setting, sed- Quantitative studies of sediment dynamics have iment budgets reveal that sediment comes dispro- expanded to include a progressively broader range portionately from limited areas of a catchment, and of natural channels, from fine-grained cohesive moves disproportionately during limited intervals of beds, to bedrock beds, and gravel or boulder beds. time. Sediment export includes solute, wash, sus- Entrainment and transport in these types of chan- pended, and bed load, but the majority of sediment nels are not adequately approximated with an excess moved from most river catchments is transported in shear stress relation for entrainment or a bulk trans- suspension. Chapter 5 Channel forms

Channel form can be examined at many levels, as work has been devoted to identifying parameters reflected in diverse classifications for rivers. The that can usefully describe the mean state and the predominant bedforms present in a river can be variations in process and form. In the preceding used to distinguish step–pool from pool–riffle chan- chapter, channel form was characterized in terms nels (Section 4.3.2), for example, or the focus on of the dominant bed material (e.g., sand-bed chan- channel form can be at the cross-sectional scale, nel, gravel-bed channels) or the predominant bed- based on descriptors such as width/depth ratio or forms (e.g., pool–riffle channel, step–pool channel). exponents of at-a-station hydraulic geometry rela- Channel form can also be described with respect tions, both of which are introduced in this chap- to cross-sectional geometry, with a focus on mean ter. Channel planform is commonly used to distin- cross-sectional characteristics. guish straight, meandering, braided, and - Cross-sectional form parameters such as width, ing channels with differing degrees of and depth, cross-sectional area, wetted perimeter, and single- or multi-thread channels. Channel form can hydraulic radius (Figure 5.1) are typically measured also include the vertical dimension over the entire at bankfull stage, under the assumption that bankfull length of a river, as when longitudinal profiles are stage defines a channel area that is filled by flow at categorized in terms of concavity or stream gradient leastonceevery2years,ormorefrequently.Bankfull indices. Just as river classifications attempt to iden- flowisaproblematicconceptthatisusedquitedif- tify consistent thresholds that distinguish channel ferently in different studies, but nonetheless forms forms existing within a continuum, the distinction the standard for measuring cross-sectional geom- of cross-sectional, planform, and longitudinal chan- etry.Cross-sectionalparameterscanalsobemea- nel forms in this chapter represents an arbitrary divi- sured for specific flow magnitudes, such as the mean sion of aspects of that intergrade annual flood or base flow. with and influence one another. 5.1.1 Bankfull, dominant, and 5.1 Cross-sectional effective discharge geometry Bankfull discharge is one of the most widely used Process and form in natural channels typically reference discharges in fluvial geomorphology, yet exhibit large spatial and temporal variability. Much the definition and implications of this flow remain

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

125 126 Rivers in the Landscape

Width can be top width or weighted average channel width logical definitions using various criteria proposed Depth can be maximum depth or weighted average Cross-sectional area is the product of width and depth in individual studies—top of bank, bank inflection, Wetted perimeter is the length of the channel boundary ratioofchannelwidthtomeandepth,levelofsignif- beneath the water surface (gray shading) Hydraulic radius is cross-sectional area divided by wetted perimeter icantchangeintherelationbetweenwettedareaand top channel width, and first maximum local bank Water Ground surface slope—indicate that, on average, discharge estimates surface Width varybyafactorofthreeatagivensite(Radecki- Depth Pawlik, 2002; Navratil et al., 2006). Another source of uncertainty is that bankfull discharge defined in relation to a channel mor- Wetted perimeter phologic feature can have very different recur- Figure 5.1 Downstream view in a channel, indicating rence intervals among different sites. In hydrocli- cross-sectional geometry form parameters. matic regions with extreme annual and interan- nual flow variability, such as arid and semiarid controversial. Wolman and Leopold (1957) defined regions, the morphologically defined bankfull flow bankfulldischargeasthestagejustbeforeflow may have a recurrence interval much longer than 1– begins to overtop the banks. Bankfull discharge has 2years.Recurrenceintervalsassociatedwithacon- subsequently assumed various implications because sistent morphological feature can vary by a factor it has been equated to a specific recurrence inter- of two between channel segments within regions as val and geomorphic function. Numerous studies diverse as Puerto Rico (Pike and Scatena, 2010) and indicate that a discharge that nearly overtops the snowmelt rivers in the US Rocky Mountains (Segura banksrecursapproximatelyevery1–2yearsonmany and Pitlick, 2010). channels (Leopold et al., 1964; Castro and Jackson, Finally, whether defined from channel morphol- 2001). Consequently, bankfull discharge is some- ogy or recurrence interval, bankfull discharge does times defined based on recurrence interval rather not necessarily transport the majority of suspended than stage with respect to channel morphology. sediment or exert a greater influence on channel Because a discharge that recurs every 1–2 years form than other magnitudes of flow (Wohl, 2010). transports the majority of suspended sediment in Analyses of suspended and bedload transport on a many rivers (Simon et al., 2004), bankfull discharge variety of rivers indicate that discharge with a recur- has been interpreted as the most important flow renceintervalof1–2yearstransportsmuchofthe magnitude for controlling channel process and form sediment moved in many rivers, but there are excep- (Wolman and Miller, 1960; Dunne and Leopold, tions to this generalization, and channel morpho- 1978). logic features are likely to reflect multiple recurrent Problems arise in that bankfull discharge can be discharge magnitudes (Pickup and Warner, 1976; difficult to define based on channel morphology. O’Connor et al., 1986; Turowski and Rickenmann, Designating bankfull stage is relatively straightfor- 2009). ward where a clearly defined, regular top of bank As a broad generalization, large magnitude, infre- separates the channel from overbank areas. Bank- quent flows are more likely to strongly influence full stage can be difficult to define morphologically sediment transport and channel morphology in if a river has inset channels. The channel may be rivers with higher thresholds for mobilizing bed so deeply incised, for example, that the top of the and bank materials, and greater temporal varia- bank has little relevance to flow volume. The top tion in discharge magnitude. In other words, sedi- of the bank may be at different levels on each side ment transport and channel morphology are more ofthechannel.Thechannelcanalsodisplaymulti- likely to reflect larger, infrequent flows in gravel-bed, ple convexities along its side slopes that reflect dif- boulder-bed, and bedrock rivers, and rivers in very ferent flow magnitudes. Comparisons of morpho- dry or seasonal tropical climatic regimes. Chapter 5 Channel forms 127

The concept of a dominant discharge,whichis network. Widening a channel requires less sedi- sometimes equated with bankfull discharge, reflects ment transport capacity than deepening a channel. theideaofasingleflowmagnitudewhich,ifmain- Bank erosion simply requires eroding the banks, but tained, will maintain the same average dimensions theresultingsedimentcanthenbestoredinthe and channel morphology as those that result from channel. Bed erosion requires both entraining the astablestream’sentirehydrologicregime(Crowder bed sediment and transporting the sediment down- and Knapp, 2005). When used in numerical mod- stream. Banks are also more likely to become unsta- eling, dominant discharge is the flow that moves the bleastheygrowtaller,unlesstheyareinverycohe- same amount of bed material, with the same size dis- sive material, so that deepening a channel typically tribution, as the actual flow regime. The assumption results in associated widening as over-steepened is that channel morphology under a dominant dis- bankscollapse.Inaddition,thechannelbedtypically charge will also remain similar to the morphology consists of coarser sediment than the banks and thus under the actual flow regime. requires more flow energy to entrain and remove. Effective discharge is the discharge that trans- Relative erodibility of bed and bank materials can portsthelargestamountofsediment(Schmidtand reflect differences in properties such as cohesion or Morche, 2006). This magnitude of discharge may or grain size between the bed and banks that influence may not equate to bankfull stage. the ability of stream flow to remove boundary mate- Bankfull discharge, effective discharge, mean rial. As banks become more erodible, w/d increases: annualflood,andthe1.5-yearflowhaveallbeen braided channels typically have greater w/d ratios proposed as constituting dominant discharge (Ros- than single-thread channels. Low bank erodibility gen and Silvey, 1996; Griffiths and Carson, 2000). can reflect high percentages of bedrock or cohesive The idea that a single magnitude of flow strongly sediment in the bank, or effective bank stabilization dominates river process and form is an oversim- by vegetation. plification of the complexity of interactions among If other factors are equal, channels with forested flow, sediment, and channel geometry through time. banks tend to be wider and have lower rates of bank This oversimplification can have important conse- erosion and channel migration than channels with quences when used in river restoration as an index grassy banks (Allmendinger et al., 2005). As noted in value for scaling channel dimensions and designing Section 4.5, the degree to which forest or other vege- stable channels, particularly if excessive emphasis on tation influences bank stability and w/d ratio varies. dominant discharge results in neglecting the geo- Riparian trees with dense root networks and greater morphic and ecological importance of less frequent, stem density that increases overbank roughness tend largerflowsorofbaseflows. to result in narrower channels than channels in more open woodlands. This is illustrated by plots of bank- full width versus discharge for gravel-bed channels 5.1.2 Width to depth ratio in Colorado, USA, (Andrews, 1984) and the United Kingdom (1986). In both cases, rivers with banks Theratioofchanneltopwidthtomeanflowdepth heavily vegetated with trees and thick brush had nar- (w/d) is one of the most commonly used parameters rower channels for a given unit discharge than did of cross-sectional geometry. The w/d ratio can be rivers with banks primarily covered in grasses and used to infer the limit strength and relative erodibil- brush (Figure S5.1). ity of bed and bank materials, relative base level sta- Theeffectofdenseriparianvegetationwasalso bility, and consistency of water and sediment inputs. demonstrated in a comparison of channel responses Channels of diverse size flowing on similar bed to increased peak flows associated with snowmaking materials become wider more rapidly than they at commercial ski resorts in Colorado, USA. Chan- become deeper as discharge increases with time at nels lined with dense willow communities showed acrosssectionorproceedingdownstreaminariver much less erosion and widening following increased 128 Rivers in the Landscape

10 North Fork Poudre River, Colorado, USA North Fork Poudre River, Colorado, USA

Run Riffle Run

Pool Riffle Riffle 1 1 Riffle Pool

Flow depth (m) Run Average velocity (m/s)

0.1 1 10 100 1 10 100 Discharge (m3/s) Discharge (m3/s) Figure 5.2 At-a-station hydraulic geometry curves for depth and velocity on the North Fork Poudre River in Colorado, USA (drainage area 980 km2). Individual curves within each plot are for different channel units along this pool–riffle channel. flows than channels lined with open conifer forests 5.1.3 Hydraulic geometry (David et al., 2009) (Figure S4.21). Grazing animals can also increase channel width At-a-station hydraulic geometry and w/d ratios indirectly by selectively grazing on and removing riparian vegetation, and directly by At-a-station hydraulic geometry is used to describe trampling stream banks and enhancing bank ero- changes in cross-sectional parameters at a site with sion. Such effects have been documented for domes- changes in flow, and to compare rates of change ticated animals (Trimble and Mendel, 1995) and for in these parameters between sites. At-a-station wild animals such as bison (Butler, 2006). hydraulic geometry characterizes how changing dis- Channel w/d ratio can also reflect base level con- charge alters width, depth, and velocity at a cross straints. Rivers do not incise below base level, so section (Figure 5.2). Starting with the continuity an increase in discharge when base level remains equation, and assuming that discharge is the pri- constant is more likely to result in channel widen- mary influence on hydraulic variables, Leopold and ing or some planform change such as formation of Maddock (1953) proposed that secondary channels, than in deepening of the river. w = aQb (5.1) Conversely, channels develop lower w/d ratios (i.e., d = cQf (5.2) deepen faster than they widen) under relative base m level fall. v = kQ (5.3) Channel cross-sectional geometry can also reflect where w is width, d is mean flow depth, v is mean changesorrelativeconsistencyinwaterandsedi- flow velocity, Q is discharge, and a, c, k, b, f, and m are ment inputs. An increase in sediment yield is likely numerical constants. Based on the continuity equa- to cause bed aggradation and channel widening, tion, the product of a, c,andk is one, and the sum of leading to a larger w/d ratio. A decrease in sedi- b, f,andm is one. ment yield can cause bed erosion, but is also likely The rates of change of w, d,andv reflect to result in bank erosion, leading to less predictable changes in w/d ratio. Four conceptual models com- r the shape and relative erodibility of the channel— monlyusedtodescribeandpredictchangesincross- channels with nearly vertical, cohesive banks, for sectional geometry are discussed in the next four example, have a very low rate of change in width sections. (Knighton, 1974); Chapter 5 Channel forms 129 r sediment transport—channels carrying large resistance elements or the initiation of overbank amounts of bedload are typically wide and shal- flow. Hydraulic geometry curves can be divided into low, and depth increases little with discharge, three portions (Knighton, 1998): whereasvelocityincreasesrapidly(Wilcock, 1971); r one portion for low discharges below the threshold r the slope of the water surface; and of sediment movement, when flow characteristics r the roughness of the wetted perimeter—velocity reflect a cross-sectional geometry left from earlier increases rapidly in a channel in which increasing high flows; flow depth quickly makes the protrusion height r a middle portion when sediment is being ofroughnesselementsasmallproportionofflow entrained from the bed; and depth (Ferguson, 1986; Eaton, 2013). r an upper portion when overbank flow occurs and flow width expands rapidly, whereas flow depth Studies of steep, coarse-grained channels suggest increases relatively slowly. that channel segments in which velocity increases more rapidly than w and d are dominated by Hydraulic geometry is also used to examine grain resistance, because relative grain submergence downstream changes in the relations between dis- increases quickly as discharge increases, and associ- charge and channel form. At-a-station hydraulic ated flow resistance declines. Channel segments in geometry exponents are typically smaller and more which width and depth increase more rapidly are variable than those for downstream hydraulic geom- dominated by form resistance (David et al., 2010). etry (DHG) relations, indicating lower rates of Ratesofchangeindepthandvelocityinrelation change with discharge at a cross section than down to resistance may not be linear, however, if differ- a river. ent sources of resistance occur with changing stage. Although velocity may increase more rapidly once Downstream hydraulic geometry grain resistance becomes negligible, for example, form resistance associated with a mobile bed can DHG relations take the same basic form as at- become more important at higher flows and slow the a-station relations, with power functions relating rate of velocity increase at the largest discharges. width, depth, and velocity to discharge. DHG rela- Althoughthecomplexityofat-a-stationhydraulic tions describe changes in w, d,andv as a discharge geometry relations limits generalizations, the width ofthesamefrequencyvariesinmagnitudedown- exponent primarily reflects channel geometry and stream (Leopold and Maddock, 1953). Mean annual boundary composition. The depth and velocity or bankfull discharge is typically used for DHG anal- exponents reflect cross-sectional form as well as yses (Park, 1977), with the assumption that chan- hydraulic resistance and sediment transport, which nel dimensions primarily reflect the forces exerted tend to be more variable than form parameters by a flow with this recurrence interval. DHG rela- (Knighton, 1998). tionsareessentiallyscalingfunctionsforchanges Average values of at-a-station hydraulic geom- in cross-sectional geometry resulting from down- etry exponents are b = 0.23, f = 0.42, m = 0.35 stream changes in discharge magnitude. (Park, 1977). These values indicate that depth typ- Typical exponent values for the relations between ically increases more rapidly with discharge than Q and w, d,andv, respectively, are 0.4–0.5, 0.3–0.4, does width or velocity, although the exponents vary and 0.1–0.2 (Park, 1977). DHG is used for engineer- widely among channels and among different cross ingstablechannelsundertheassumptionthatw ∼ sections on a single channel (Reid et al., 2010). Q0.5 when Q has a recurrence interval of 1–2 years. Inflection points that mark variations in the rate As noted in earlier discussions of bankfull and dom- of increase in w, d, or v with increasing Q can inant discharge, this can be a reasonable assump- reflect increasing submergence of grain or form tion for channels with limited hydrologic variability, 130 Rivers in the Landscape but can be an inappropriate assumption for channels channel response to changing discharge is difficult with greater hydrologic variability in which channel for at least two reasons: width reflects a discharge of longer recurrence inter- r val.DHGisalsousedtopredicttheeffectsofflow interactions among w, d, v and other channel form regulation, and to understand the geomorphic role variables (e.g., bedform amplitude) not included in hydraulic geometry relations; and of floods or infer the magnitude of past floods based r on channel dimensions (Ferguson, 1986) (supple- the influence of sediment supply, which is also not mental Section 3.2.1). included in hydraulic geometry relations. Strong correlations between Q and w, d,andv At-a-station and downstream hydraulic geom- exist where channel boundaries have relatively low etry are nonetheless useful in understanding the erosional thresholds, and where substantial varia- manner in which channel form adjusts to fluctuat- tions in sediment supply or imposed gradient are not ing discharge. present. Numerous site-specific case studies, how- ever, document weak or poorly developed DHG relations in mountainous regions with persistent 5.1.4 Lane’s balance geomorphic effects from glaciation or strong down- stream contrasts in substrate erodibility or grain-size Lane’s balance refers to a conceptual model of chan- distribution of sediment inputs (Wohl, 2010). nel adjustment that includes several cross-sectional Wohl (2004b) proposed an empirical thresh- parametersandaccountsforbothwaterandsedi- old for well-developed DHG relations in high- ment discharge. Lane (1955) conceptualized equilib- relief drainage basins as a function of the ratio of rium within a channel segment as reflecting a bal- stream power (hydraulic driving forces) to sedi- ance among discharge (Qw), channel gradient (S), ment size (substrate resistance). Higher values of sediment load (Qs), and sediment size (Ds) stream power associated with large discharge per Q S ∝ Q D (5.4) unit drainage area, as in the tropics, can result w s s in well-developed DHG relations even in channels This relation is referred to as Lane’s balance and with strong colluvial or bedrock influences (Pike has been widely depicted using a drawing of a bal- et al., 2010). DHG relations for bedrock channels ance (Figure 5.3) (Grant et al., 2013). This drawing are within the range of those in alluvial channels, is intuitively appealing and easy to understand. An although bedrock channels tend to be narrower and ↑ increase in sediment load ( Qs), for example, will deeper (Montgomery and Gran, 2001; Wohl and ↑ cause aggradation, coarsening ( Ds), and steepening David, 2008). (↑S)ofthestream. Bothformsofhydraulicgeometryarescaling TheabilityofLane’sbalancetodescriberiver relations between discharge and the other vari- adjustments is inherently limited, however, because ables of the continuity equation. The exponents of theexpressiondoesnotaccountforchangesincross- the hydraulic geometry relations reflect the relative sectional, planform, and bedform geometry that are magnitude of response among w, d,andv to changes commonly associated with channel adjustments to in Q,andcanbeusedtoinferthenatureofchannel changes in discharge and sediment load. Dust and adjustment. The DHG average width exponent of Wohl (2012b) expanded Equation 5.4 to include ∼0.5, for example, indicates that approximately half these terms of the channel adjustment to downstream increases ( ) ( ) in discharge occurs via channel widening, whereas Δz w Qw ̄ ∝ QsDs (5.5) the at-a-station average depth exponent of 0.42 indi- PHa d cates that adjustment to greater flows at most cross sections occurs primarily via increased flow depth where S is proportional to total change in elevation rather than increased width. Precisely predicting along a channel (Δz) and inversely proportional Chapter 5 Channel forms 131

Coarse Fine Flat Steep S DS

Sediment Size Stream Slope

QS + Qw +

Degradation0 Aggradation

Original (a)

Coarse Fine Flat Steep S DS

Slope Sediment Size Stream

QS + Qw +

+ Increase 0 Decrease – Width-to-Depth Ratio (W/d)

Expanded (b)

Figure 5.3 An illustration of Lane’s relation, which is widely used to conceptualize channel adjustment in response to changes in water or sediment yield. (a) The original illustration, based on a drawing by Whitney Borland of Colorado State University. (From Dust and Wohl, 2012b, Figure 1.) (b) An illustration of an expanded version of Lane’s relation, including adjustments to channel width-to-depth ratio. (From Dust and Wohl, 2012b, Figure 9.)

̄ ↓ to sinuosity (P) and bedform amplitude (Ha), and to depth ratio ( w/d), as well as coarsening and (w/d) is width/depth ratio. Equation 5.5 suggests steepening. ↑ that an increase in sediment load ( Qs)willcause Either version of Lane’s balance is useful primar- aggradation (↑z), decreased sinuosity (↓P)and ily as a conceptualization of the various potential ↓ ̄ bedform amplitude ( Ha), and decreased width ways in which channel geometry can adjust to the 132 Rivers in the Landscape

(a)

(b) Terrace

Braided (c) channel

(d) Terrace

Figure 5.4 Complex response of a channel to a single fall in base level: perspective is upstream within the valley (gray shading) at a point midway along the channel between the mouth and the headwaters. (a) Stream alluvium (black) deposited before base level lowering. (b) After base level lowering, a knickpoint migrates upstream, causing the channel to incise into alluvium and bedrock of the valley and leaving the alluvium as a terrace. Following incision, bank erosion widens channel and partially destroys the terrace. (c) Continued upstream migration of the knickpoint causes increased sediment discharge to mid-portion of river. Inset alluvial fill is deposited as sediment discharge from upstream increases. A broad, shallow, braided channel (lighter gray shading) develops. (d) Upstream migration of the knickpoint ceases and the sediment supply to downstream channel segments declines. The channel becomes deep and narrow and incises slightly, creating a second terrace. (From Schumm and Parker, 1973, Figure 1.)

controlling variables of Qw and Qs.Ananalogous its furthest upstream point, causing downstream method to quantitatively predict channel adjust- reachestoincise,andsometimescreatingasecond ment remains elusive because of the numerous inter- head cut that migrated upstream. Consequently, the dependent variables. upstream portion of a river can be incising while the downstream portion is aggrading. Any given point 5.1.5 Complex response within the channel alternates between incision and aggradation with time, sometimes going through Schumm (1973) and Schumm and Parker (1973) multiple cycles of incision/aggradation in response introduced the phrase complex response to describe to the initial base level fall before the channel once asynchronous, discontinuous river response to a more stabilizes. single external perturbation. During flume exper- Subsequent studies have applied the idea of com- iments, Schumm and Parker (1973) observed that plex response to a variety of spatial and temporal base level fall initiated a head cut that migrated scales and types of rivers (Marston et al., 2005; Har- upstream. Upstream migration increased sediment vey, 2007). In general, although a river can incise or load to channel segments downstream from the aggrade throughout its length, the more common head cut, which then began to aggrade (Figure 5.4). scenario is that aggradation in some part of the Sediment supply declined once the head cut reached river’s length is likely associated with enhanced Chapter 5 Channel forms 133 erosion of the channel boundaries elsewhere. modified to six or more stages (Simon and Rinaldi, Among the implications of complex response is that 2013) that reflect shifts in the dominant adjustments r and associated rates of sediment transport, bank multiple terraces along a river may not represent stability, and cross-sectional geometry (Figure 5.5). multiple external changes (Womack and Schumm, Thetimenecessarytodevelopeachstageofevo- 1977) (Section 6.2); r lution and the relative magnitudes of vertical and incision or aggradation within a segment of chan- lateral adjustments vary widely between different nel can be a transient response to upstream or streams and stream segments (Simon and Rinaldi, downstream changes in channel geometry, base 2006), and the entire sequence can require 102–103 level and sediment supply, with the time span of years (Simon and Castro, 2003) to complete. Like the transience dependent on factors such as the Lane’s balance and complex response, channel evo- size of the channel and the erodibility of the bed lution models are useful conceptualizations rather and banks; and r than quantitative predictions. consequently, restoration or management that Incised channels of the type described in channel seeks to limit incision or aggradation may actu- evolution models can form anywhere, but are partic- ally enhance undesirable channel change if imple- ularly common in arid and semiarid regions. Chan- mented just as the direction of channel response nels in these regions undergo repeated episodes of (i.e., incision vs. aggradation) is changing. alternating incision and aggradation in response to internal thresholds, as described in Section 1.5.1 5.1.6 Channel evolution models (Schumm and Hadley, 1957) or to external changes in flood magnitude and frequency (Webb and Baker, The ideas underlying Lane’s balance and com- 1987), other aspects of climate (Leopold, 1976), or plex response, in particular, are incorporated into land uses such as grazing and groundwater with- channelevolutionmodels.Alluvialchannelscan drawal (Cooke and Reeves, 1976). Artificially chan- develop very small w/d ratios when undergoing nelized streams, which are described in Chapter 7, rapid incision, but this is typically a transient con- also typically go through the stages described in dition. Alluvial channels in diverse environments channel evolution models following completion of that incise in response to base level fall or changes channelization. in water and sediment supplied to the channel pass Adjustments to cross-sectional geometry repre- through a characteristic sequence of channel geom- sent an intermediate level of channel response to etry with time that is summarized in channel evo- changing water and sediment dynamics. Bed grain- lution models. The channel adjustment described size distribution and bedform configuration are in these models typically begins with a relatively more likely to change first as water and sediment narrow,deepchannel.Thechannelsubsequently supplies fluctuate, as described in the preceding widens, aggrades, and eventually stabilizes. Schumm chapter. Adjustments to channel w/d ratio typically et al. (1984) proposed empirical relations (e.g., top represent the next level, in response to larger mag- width = 46.77 (drainage area)0.39) between top width nitude or more prolonged fluctuations in water and and drainage area, derived from observations on sediment. Changes in channel planform represent a stabilized channels, which can be used to identify third level of channel adjustment. whether incising channels are close to a stability threshold at which widening stops. A specific form of this relation can be useful within a limited geo- 5.2 Channel planform graphic region, but the coefficient and exponent vary between regions. Several classification schemes have been proposed Schumm et al. (1984) initially proposed a five- to describe the wide variety of forms assumed by stage model of channel evolution. This has been rivers when viewed on a two-dimensional planar 134 Rivers in the Landscape

Stage I. Sinuous, prior to modification

hc Critical bank height h Direction of bed or bank change

h < hc

Stage II. Constructed Stage IV. Degradation Stage III. Degradation (channelized) and widening

Terrace h h h h < h c h < hc h < hc

Slumped material Stage V. Aggradation and widening Stage VI. Quasi-equilibrium

Terrace Terrace

h Bankfull h

h > hc h < hc

Aggraded material Aggraded material Slumped material

Primary Stages I,II knickpoint

Stages III Stages IV Precursor Top of bankStages V knickpoint Stages VI Aggraded Oversteepened reach Secondary Aggradation zone material knickpoints Figure 5.5 Illustration of a six-stage channel evolution model. View is upstream or downstream in the first six boxes. The lower box contains a longitudinal profile illustrating different stages occurring simultaneously along different segments of a channel. Light gray shading is valley material (sediment or bedrock), darker gray is recent alluvium. (From Simon and Castro, 2003, Figure 11.11, and Wohl, 2010, Figure 4.14.) surface such as a map (Buffington and Montgomery, additional categories such as wandering gravel-bed 2013). An obvious distinction is between rivers with rivers (Carson, 1984), anabranching (Brice et al., single channels and those with multiple channels. 1978), and compound rivers that regularly alternate Single channels are typically differentiated on the between two or more planforms over time. basis of sinuosity,theratiodefinedbyactualflow Any classification imposes more or less arbitrary path downstream to straight line distance between boundaries on a continuum of channel form. Ideally, two points. Multiple channels are typically differ- the boundaries reflect thresholds in the processes entiated based on the lateral mobility of secondary that create channel form. As with substrate, bed- channels. forms, and cross-sectional geometry, channel plan- Leopold and Wolman (1957) proposed a tri- form reflects an adjustment of the rate of energy partite classification of straight, meandering, and expenditure in relation to water and sediment sup- braided channels. Although this classification is still plied to the channel and erosional resistance of used, there are many channel planforms that do not the channel boundaries. Any channel planform that fit well within these categories. Schumm (1985) pro- deviates from a single, straight channel—as many posed a broader range of 14 channel types (Fig- planforms do—represents an adjustment of resis- ure 5.6), and other investigators have described tance to flow. A channel with meanders, for example, Chapter 5 Channel forms 135

High Relative stability Low Low Bed load/total load ratio High Small Sediment load Large Small Sediment size Large

Channel type Suspended load Mixed load Bed load tagtMadrn Anastomosing Meandering Straight Straight Straight eneigtawgBraided Meandering thalweg Meandering braided Island

Figure 5.6 Classification of alluvial (from Schumm, 1981). Gray shading indicates depositional areas in the form of islands, bars or riffles. Arrows indicate flow paths. effectively has a longer flow path for a given down- Straight alluvial channels without mobile bed- stream length and gradient, and thus greater bound- formshavemostlysuspendedsediment,alowgradi- ary resistance than an otherwise equivalent straight ent, and a deep, narrow cross section. Such channels channel. The most commonly occurring planforms are rare (Schumm, 1985). Straight alluvial channels are described in more detail next. with mobile bedforms are typically sand-bed chan- nels in which the sequence of immobile bed, ripples, 5.2.1 Straight channels dunes, mobile plane bed, and antidunes described in Section 4.3.1 occurs as flow changes. Straight channels have a single channel with sinu- Themajorityofstraightalluvialchannelsin osity less than 1.5. Straight channels can be straight mixed substrates or substrates coarser than sand becausetheyarecloselyconfinedbysteepvalley havepool–rifflesequencesofgreaterorlessermobil- walls or other uplands such as terraces, or they can ity, depending on the substrate and sediment sup- occur in erodible, alluvial boundaries. Even straight ply. The downstream alternation between pools and channels in erodible alluvial boundaries, however, rifflesisassociatedwithcross-sectionalasymmetry remain straight because the banks have sufficient and undulations in bed elevation, so that a regu- erosional resistance relative to available flow energy lar channel planform does not equate to regular- to limit bank erosion (Paola, 2001). Straight chan- ity in other dimensions, or to uniformity of bound- nels can be further distinguished as those with or ary roughness. Pools typically occur on alternate without mobile bedforms, those with alternating sides of the channel in a downstream direction, with bars and a sinuous thalweg, and slightly sinuous intervening depositional features such as riffles and channels with point bars. alternate bars (Figure 5.7). The thalweg—the line of 136 Rivers in the Landscape

(a) (b) face to develop a rotational component of flow as a result of greater hydraulic resistance along the chan- Meandering Riffle–pool nel boundaries than in the center of the channel. thalweg units This rotational component, described as helical flow, creates downstream alternations in the location of Pool Pool greatest velocity that are expressed in differences in

Scour Scour boundary erosion, sediment transport and deposi- tion, and channel geometry. Sinuosity in flow and Crossing Riffle channel planform scales consistently across features asdiverseaschannelslessthanameterwideon Deposition Scour glacial ice and channels more than a kilometer wide in very large rivers (Leopold, 1994). Primary flow direction

Pool Pool

5.2.2 Meandering channels Crossing Riffle Single channels with a sinuosity greater than 1.5 appear to be the most widespread and common type Pool Pool of channel planform (Leopold, 1994). Meandering channels can be differentiated based on the regular- ity of bends (Figure S5.2) (Kellerhals et al., 1976) into

Crossing Riffle r irregular meanders with a weakly repeated down- stream pattern; r Figure 5.7 Plan view showing a sinuous thalweg and regular meanders with a repeated pattern and a alternate bars within a straight channel, with associ- maximum deviation angle between the channel ated helical flow and cross-sectional asymmetry shown and down-valley axis <90 degrees; and by small channel cross-sectional views at right. (a) Model r tortuous meanders with a less clearly repeated pat- of twin, periodically reversing, surface-convergent heli- tern and a maximum deviation angle >90 degrees. cal cells based on work by Einstein and Shen (1964). (b) Model of surface-convergent flow produced by inter- actions between the flow and a mobile bed, creating This differentiation reflects the fact that naturally pool–riffle units of alternate asymmetry based on work formed meanders are seldom perfectly regular, but by Thompson (1986). Black lines indicate currents and instead include randomness that reflects local con- progressively darker gray shading indicates progressively trols on the erodibility of the channel boundaries. deeper portions of the channel (from Knighton, 1998, fig In any given meander, the outer banks are com- 5.15, p. 199.) monly steep and eroding and a pool is present at the bend apex. Cross-sectional bed topography slopes downward from the point bar on the opposite inner deepestflow—isthussinuouseventhoughthechan- bank. Riffles are present in the inflection regions of nel boundaries are straight. The flow alternately con- the bend, in the straight limbs between bend apices, verges passing through pools and diverges passing where cross-sectional and bank geometry are more over depositional zones. symmetrical (Hooke, 2013). The ubiquity of either a sinuous thalweg or a sinu- Meander geometry is typically characterized ous channel planform presumably reflects an inher- using meander wavelength, 𝜆,andradiusofcur- 𝜃 ent tendency in water flowing over a rough sur- vature, rc, for individual bends. Path direction, , Chapter 5 Channel forms 137

demonstrated that, for a given discharge, meander λ * wavelength also varies with boundary cohesion and w gradient. Wavelength decreases as boundary cohe- rc sion and gradient increase. λ The thalweg of a meandering river does not

Inflection maintain a central location along the channel in point a downstream progression. Instead, the thalweg migrates across the channel through each bend from λ meander wavelength near the inner bank at the bend entrance to near the λ path wavelength * outer bank at the bend exit (Figure 5.9). The strongly rc radius of curvature w channel width helical flow in a meandering channel is expressed as Figure 5.8 Components of meander geometry. superelevation of the water surface against the outer bank of each bend and a transverse current directed toward the outer bank at the surface and toward the and change of direction, Δ휃, characterize meander inner bank at the bed. Helical flow facilitates prefer- geometry for multiple bends (Figure 5.8). ential erosion of the outer bank and deposition on a Langbein and Leopold (1966) introduced the idea point bar along the inner bank. of modeling meander geometry using the equation The details of strength and location of the trans- for a sine-generated curve verse,secondarycurrentsreflectflowstage,mean- der geometry, and channel cross-sectional geometry. 휃 = 휔 sin kx (5.6) Secondary currents weaken with increasing stage and the flow follows a straighter path. Tighter bends where 휃 is channel direction expressed as a sinu- withalowerratioofradiusofcurvaturetochannel 휔 soidal function of distance x, with parameters for width (rc/w)havestrongersecondarycirculation. the maximum angle between a channel segment and A large width/depth ratio is associated with more themeandown-valleyaxisandk = 2휋/휆.Asine- extensive development of point bars, and point bars generated curve represents a path in which the sum also reflect grain-size distribution (Hooke, 2013). ofsquaresofchangesinchanneldirectionperunit Although models, in particular, typically assume length is minimized, which effectively distributes relatively symmetrical and regular meander bend stress uniformly along the curve. A sine-generated geometry, natural meanders can have substantial curve is a good approximation of the geometry deviations in form and flow resistance, with asso- of regular, symmetrical meanders (Leopold, 1994). ciated deviations from an idealized distribution of Many—perhaps the majority of—natural meanders, hydraulic variables. The presence of instream wood however, are not symmetrical (Carson and Lapointe, along a meander, for example, strongly modifies the 1983). Measurements of meander geometry across a three-dimensional (3D) flow structure in a manner rangeofenvironmentssuggestthat휆 is typically 10– highly dependent on the arrangement, density, and

14 times the channel width and rc is typically 2–3 mobility of the wood (Daniels and Rhoads, 2004). times the channel width (Knighton, 1998). In some bends, wood can create sufficient resis- Just as channel width is approximately propor- tance to reduce flow velocity and constrain the high- tional to Q0.5, meander wavelength also typically velocity flow and helical motion to the channel cen- varies as ∼Q0.5,whetherQ is mean annual flood, ter, whereas in other bends the wood can strongly mean annual discharge, mean monthly maximum deflect the flow away from the banks (Daniels and discharge, or some other measure of flow that Rhoads, 2004). is likely to transport most of the sediment and Hypotheses regarding the initiation of meander- thus strongly influence channel form parameters ing tend to focus either on inherent properties of (Dury, 1965; Knighton, 1998). Schumm (1963, 1967) flow, or on interactions between the flow and a 138 Rivers in the Landscape

(a) i ii High High τ * * h , max Near-surface velocity τ Near-bed velocity , max * h , max * τ h Low , max Low τ h Low * Bar exposed * at low flow High Thalweg * relative water surface elevation Break in break in bed slope bed slope

(b) Outer Inner Mid-channel bank bank region region region Super-elevated water surface

Outward shoaling flow across point bar

path lines of secondary flow Figure 5.9 Sinuosity of the thalweg and associated velocity and shear stress in meandering channels. (a) (i) 휏 Location of maximum boundary shear stress ( b) and (ii) flow field in a bend with a well-developed bar (from Dietrich, 1987). (b) Secondary flow at a bend apex showing the outer bank cell and shoaling-induced outward flow over the point bar (from Markham and Thorne, 1992). mobile channel boundary. The argument for flow (2012), helicoidal, curvature-driven secondary flow is properties rests on the fact that water flowing down generated by superelevation of the water surface at a rough, inclined surface—even a relatively smooth the outside of a bend. This creates substantial cross- plate in a flume—in the absence of sediment, devel- streamvariationinvelocity,whichredistributes ops sinuosity at the base of the flow where frictional downstream momentum, leading to a decrease in resistance is greatest. Progressively more of the flow bed shear stress and deposition of a point bar along depthisinvolvedinsinuousflowasdischargeand theinnerbend,aswellasadownstreamincreasein velocity increase. The forces involved in this type of bed shear stress and erosion along the outer bend. flowdonotnecessarilyscalewelltonaturalchan- The point bar deflects flow laterally toward the cut nels (Hooke, 2013), however, which typically have at bank, creating topographically driven secondary flow. least partly mobile beds. Arguments for interactions Seminara (2006), Guneralp¨ and Marston (2012), and between the flow and the boundary as the underly- Hooke (2013) reviewed theoretical and mechanical ing cause of meandering emphasize deformation of explanations for meandering in more detail. the bed leading to the development of alternate bars Attention has also focused on why meanders that initiate meandering by deflecting flow toward form.Extremalhypothesesfocusontheinfluence the opposite banks (Seminara and Tubino, 1989). of meanders on energy expenditure. The reason- Sinuous channels develop in materials without bars, ing is that meanders either (i) create the most uni- however, such as those formed atop glacial ice or in form rate of energy expenditure along a channel by bedrock. minimizing variance in hydraulic variables (Lang- Most likely, meanders are not the outcome of a bein and Leopold, 1966), or (ii) establish a minimum single cause. As explained by Guneralp¨ and Marston channel slope for given input conditions (Chang, Chapter 5 Channel forms 139

1988). Subsequent research, however, emphasizes and increase habitat diversity for aquatic and ripar- the absence of equilibrium (Hooke, 2013), focus- ian organisms. Secondary channels gradually fill ing instead on continuous evolution and instabil- with sediment settling from suspension during over- ity (Eaton et al., 2006). Discharge, sediment size bank flows. Because this sediment is commonly and supply, bank resistance, and gradient are still finer grained than adjacent depositional areas of the acknowledged to influence meander migration and floodplain, even secondary channels that completely morphology, implying that any change in these vari- filled with sediment decades or centuries earlier cre- ables results in a response in meander form and pro- ate persistent differences in groundwater, soil mois- cess (Hooke, 2013). ture, and plant communities. Meanders can migrate very slowly, but even Supplemental Section 5.2.2 discusses measure- deeply incised bedrock channels that are sinuous ment and modeling of meandering rivers. show evidence of meander migration during inci- sion (Harden, 1990). Migration of individual mean- ders (Daniel, 1971) can occur through 5.2.3 Wandering channels r translation, which is downstream shifting of the Wandering gravel-bed rivers have rapid bend migra- bend without alteration in basic shape; tion, numerous bars, and frequent dissection of r extension, during which the bend increases its point bars. Wandering rivers are probably transi- amplitude by migrating across the valley; tional between meandering and braided planforms r rotation, in which the bend axis changes orienta- (Carson, 1984) (Figure S5.4). Wandering channels tion; and have also been subdivided into those with single r lobingandcompoundgrowth,duringwhichthe channels, high channel migration rates, and fre- bend becomes less regular and symmetrical. quentdissectionofpointbars,andthosewithmul- tiple channels, a large supply of bed sediment, and Individual bends along a river can have differ- low to moderate bank erodibility (Carson, 1984). ent styles and rates of migration, and bends typically Wandering channels are most commonly described deform and become asymmetrical with migration. for mountainous regions with headwater glaciers

The fastest migration tends to occur when rc/w is or downstream from large terraces, although this between 2 and 3 (Hickin and Nanson, 1984). Ratios channel planform can occur in any environment outside of this range lead to preferential migration in (Church, 1983; Burge, 2005). Although the num- theupstreamordownstreamlimbofthebend,with berofpapersusingthecategoryofwanderingchan- associated changes in curvature rather than rapid nels is limited, the designation of this channel type migration of the entire bend (Knighton, 1998). does reflect the fact that distinguishing meandering, Meander migration that increases the amplitude braided, and anabranching channels is not always and tightness of bends can exceed a stability thresh- straightforward (Nanson and Knighton, 1996). old and trigger a chute cutoff that creates a shorter channel across the inside of the point bar or a neck cutoff atthebaseofthebend(Constantineetal., 5.2.4 Braided channels 2010). Because a cutoff increases channel gradient and local transport energy, the cutoff becomes the Braided rivers are multi-thread channels in which main channel and the longer flow path becomes flow is separated by bars within a defined chan- a secondary or overflow channel that accumulates nel(FigureS5.5).Someofthebarscanbesub- sediment during high flows. Secondary channels can merged at high flows, but all are typically exposed persist for decades to centuries, depending on the at low flows. The degree to which bars are sta- rates of sediment filling (Figure S5.3). While present, bilized by vegetation varies, but the usual dis- secondary channels can form floodplain wetlands tinction between multi-thread braided rivers and 140 Rivers in the Landscape multi-thread anabranching rivers is that bars in (a) i braided rivers have less vegetation, are narrow rel- ative to the width of the channel, and are relatively Flow direction mobile compared to anabranching rivers. When Pool Bar front bars and islands are distinguished, mid-channel bars Exposed bar are unvegetated and submerged at bankfull stage, Cut off whereas islands are vegetated and emergent at bank- ii full stage (Brice, 1964). Bars can be: r longitudinal and formed of crudely bedded gravel (b) sheets; r linguoidbarsthatarelobateinshape,ofsandor gravel deposited by downstream avalanche-face progradation; and r point or side bars formed by coalescence of smaller bedforms such as dunes and linguoid bars at sites of lower energy (Miall, 1977) (Figure S5.6).

Flume experiments with initially straight chan- nels indicate that braiding can develop from the for- mation of single alternate bars, single mid-channel Figure 5.10 (a) Schematic illustration of the develop- bars, or multiple mid-channel bars (Ashmore, 2013). ment of braiding from initial (i) alternate bars by channel widening and chute cutoff, or (ii) central or higher mode These bars are low-amplitude bedforms occupy- bars. (From Ashmore, 2013, Fig 3.) (b) Photograph of a ing most of the channel width and connected to wide (∼1 km) in northern Alaska. Flow is upstream scour pools. The bars deflect flow and from left to right. startthedevelopmentofchannelsinuosity.Braid- ing develops either by cutoff of single bars at a criti- cal bend amplitude, or bifurcation (splitting of flow) experiments indicate a strong, positive relationship around mid-channel bars (Ashmore, 2013). Braid- between sediment supply and frequency of chan- ing is maintained by repetition of these initial bar- nelavulsion(lateralmovementandformationofa scale processes within the individual channels of new channel), such that increased sediment sup- the braided network, but the initially simple, well- ply causes greater channel mobility, bifurcation, and defined pool-bar units (pool head to downstream (Ashworth et al., 2007). Bars more strongly barmargin)ofasinglechannelarereplacedbythe influence bifurcation processes as the bar height confluence-bar/bifurcation units (the distance over above the bed (bar amplitude) increases (Bertoldi which two channels join and then split again down- et al., 2009). Adjustments to braided channel geom- stream) that form a basic morphological element of etry alter flow resistance, and the degree of braiding braided rivers (Bridge, 2003; Ashmore, 2013) (Fig- tends to increase with slope (Parker, 1976). ure 5.10). The degree of braiding has been quantified Repeated division and joining of channels, and using various braiding indices. The most common associated divergence and convergence of flow, cor- approach is to count the mean number of active respond to rapid shifts in channel position and the channels or braid bars per transect across the chan- size and number of bars, particularly during floods. nel belt (Bridge, 2003; Egozi and Ashmore, 2008). Braiding is produced by processes active at higher Braided channels are much less common flows, rather than resulting solely from dissection than meandering channels. However, both the of bars during low flows (Ashmore, 2013). Flume rock record prior to the evolution of land plants Chapter 5 Channel forms 141

(Montgomery et al., 2003a; Davies and Gibling, of sediment supplied (Griffiths, 1979). Locally 2011) and flume experiments suggest that braided reduced transport capacity can facilitate sedi- rivers are the default channel planform in rivers ment deposition that allows a bar to form and lacking sufficient riparian vegetation or cohesive grow, deflecting the current toward the adjacent bank sediments to substantially increase bank banks, creating local bank erosion, and introduc- resistance to erosion (Paola, 2001). Where woody ingmoresedimentintotheflow.Thisisthecen- riparian vegetation expands, commonly as a result tral bar mechanism that Leopold and Wolman of changes in flow regime, braided channels can (1957) invoked to explain the initiation of braid- metamorphose to a meandering or anabranching ing. Related to this is the transverse bar conversion planform (Nadler and Schumm, 1981; Piegay´ and mechanism of initiating braiding, in which flow Salvador, 1997). Field studies indicate that the rate convergence through a pool scours the bed and and locations of growth of woody riparian vegeta- provides sufficient bedload for deposition down- tion strongly influence the formation and erosional stream from the pool where the flow diverges, resistance of islands (Gurnell and Petts, 2006), as eventually causing flow to be deflected around an well as the stability of outer banks along a braided elevated bar (Ashmore, 1991). Increased bedload channel. Instream wood can also strongly influence supply causes aggradation and an increase in the thelocationandstabilityofbars,aswellasthe degree of braiding, whereas decreased supply has establishment of riparian forests on the bars, when theoppositeeffect(GermanoskiandSchumm, sediment is deposited around a bar-apex logjam 1993; Thomas et al., 2007). (Abbe and Montgomery, 2003) (Figure S5.7). The 2. Erodible banks facilitate continued channel tendency toward braiding may be influenced by a widening and the development of multiple river’s ability to turn over its bed within the char- bars in wide, shallow flow with heterogeneous acteristic time for riparian vegetation to establish transport capacity. This corresponds to the and grow to a mature, scour-resistant state (Paola, argument presented earlier: single-thread or 2001). This dimensionless time-scale parameter can anabranching channels tend to form where predictwhetherachannelwillbraid(Hicksetal., cohesivebanksresultfromsiltandclayorfrom 2008). riparian vegetation. Braided channels occur in diverse environments 3. Rapid fluctuations in discharge contribute to and across a broad range of scales, and are par- bank erosion and heterogeneous bedload move- ticularly common in arid and semiarid regions, ment,andlargefloodscaninitiatebraiding,in downstream from glaciers, and in mountainous part by removing stabilizing riparian vegetation environments with abundant coarse sediment and dramatically increasing bank erosion and supply and limited riparian vegetation (Ashmore, channel width (Burkham, 1972; Friedman and 2013) (Figure S5.8). Proglacial braided channels are Lee, 2002; Jaquette et al., 2005) (Figure S5.9). known by the Icelandic word sandurs at the point Some types of channels alternate repeatedly where the channel system expands freely, and valley between braided (immediately after a large flood) sandurs where development of the channel network and meandering (developing gradually during is confined by valley walls (Krigstrom, 1962). lower discharges over years to decades following Braiding tends to be associated with four con- a large flood). ditions, although no single one of these conditions 4.Steepvalleygradientsappeartopromotebraid- is either sufficient or necessary to create a braided ing, although high stream power (QS)maybea channel (Knighton, 1998). better measure than simply gradient (Knighton, 1998). Different threshold values have been pro- 1. Abundant bedload can cause braiding if the chan- posed, but braided rivers tend to have higher val- nel lacks capacity to transport the volume of sed- ues of stream power than meandering rivers of iment supplied, or competence to move the size similar grain size (Ashmore, 2013). The physical 142 Rivers in the Landscape

explanation for this observed correlation remains ing length of flow path for a given vertical drop. under debate. Mid-channel bars exhibit similar width/length ratios over a wide range of spatial scales (Kelly, 2006). This In addition to mechanisms of initiating braid- suggests that braiding is a self-organized, emergent ing through primarily depositional processes, as property of the interaction between flow and a non- described earlier, braiding may also initiate in cohesive sediment bed that must be constrained in response to erosion of bars. Flume experiments indi- some way if morphology other than braiding is to cate that dissection or the formation of cutoff chan- develop (Paola, 2001; Ashmore, 2013). nels across various types of bars occurs when flow Supplemental Section 5.2.4 discusses measure- follows a steeper route across a bar, causing head- ment and modeling of braided rivers. ward incision across the bar, which increases the braiding index (Ashmore, 1991; Germanoski and Schumm, 1993). The presence of bars is still basic 5.2.5 Anabranching channels to developing a braided channel, indicating that the fundamental underlying process is local deposition Anabranching channels are multi-thread channels in response to loss of transport capacity. in which individual channels are separated by veg- Once braiding starts, all of the erosional and etatedorotherwisestablebarsandislandsthat depositional processes described earlier may be arebroadandlongrelativetothewidthofthe operating to maintain braiding, as long as short- channels, and that divide flows at discharges up term transience in bedload transport is maintained. to bankfull (Nanson, 2013). The islands persist Such transience is closely tied to longitudinally alter- for decades to centuries and are similar in eleva- nating convergent flow zones at confluences, in tion to the floodplain (Knighton, 1998). Individ- which sediment from upstream is transported and ual anabranching channels can be straight, mean- bed scour occurs, and divergent flow zones at bifur- dering, or braided but, unlike distributary networks, cations, in which sediment is deposited (Ashmore, the channels in an anabranching network eventually 2013). rejoin. Anabranching channels are relatively uncom- Confluence geometry reflects relative discharge, mon, but occur in very diverse environments (Fig- junction angle, and orientation of the confluent ure S5.10). channels. Maximum depth of scour increases as Anabranching occurs in bedrock channels and the junction angle increases and as discharges in the characteristics of individual channels appear to the confluent channels approach equal magnitude be influenced by joint geometry in the bedrock. (Mosley, 1976; Best and Rhoads, 2008). Avulsion Anabranching reaches are particularly common frequency is the frequency at which a new chan- immediately upstream of waterfalls (Kale et al., 1996; nel forms that carries ≥50% of the discharge of Tooth and McCarthy, 2004), but can also occur in the old channel. Avulsion frequency scales with the lower gradient reaches without knickpoints (Her- time necessary for bed sedimentation to produce a itage et al., 2001). deposit equal to one channel depth, at which time Anabranching is also particularly common in thechannelislikelytoavulse.Thisisreflectedina very large alluvial rivers (Jansen and Nanson, 2004; dimensionless mobility number based on the rela- Latrubesse, 2008), although many of the historically tive rates of bank erosion and channel sedimentation anabranching segments of these rivers have been (Jerolmack and Mohrig, 2007). channelized to single-thread channels (Piˇsut,´ 2002). As in the case of sinuosity in meandering chan- Anabranching is the dominant channel pattern of nels, braiding represents a morphological adjust- the Amazon, Congo, Orinoco, Parana, and Brahma- ment of channel gradient to a stable configuration: putra, among other large rivers, and anabranch- in this case, by progressive channel subdivision, ing was historically present along rivers such as the which effectively alters the channel gradient by alter- Danube and Rhine (Latrubesse, 2008). Chapter 5 Channel forms 143

Alluvial can develop as erosional in Australia to very large rivers in South America channels scour into the floodplain during channel and Asia. avulsion. Avulsion can also be triggered by sedi- 4. Sand-dominated, ridge-forming anabranching ment accumulation within a channel, particularly rivers have long, narrow, parallel ridges stabilized where an obstruction such as a channel-spanning by riparian vegetation and straight anabranches logjam creates a backwater effect (Abbe and Mont- with steep banks. This form of anabranching also gomery, 2003; O’Connor et al., 2003; Wohl, 2011c). has moderate stream power and has mostly been Rapid aggradation can produce frequent avulsions described in Australia. and a network of channels in various stages of for- 5. Gravel-dominated, laterally active anabranching mation and abandonment that can be wandering rivers are wandering channels with higher stream or anabranching (Nanson, 2013). Anabranches can power. Although the basal coarse sediment is also develop from mid-channel bars that become overlain by finer sediment on the islands, vege- islands dividing the flow in previously wider chan- tation provides most of the bank stability. These nels,particularlyinverylowgradientchannels. channels, in particular, can be facilitated by Finally, anabranches can develop from delta progra- obstructions such as logjams. dation and modification of the distributary network 6. Gravel-dominated, stable anabranching rivers (Nanson, 2013). In this situation, anabranching also exhibit higher stream power. Like type 5 may reflect an efficient means of redistributing channels,riparianvegetationiscriticaltoenhanc- and storing excess sediment across a wide val- ing bank stability in these channels and obstruc- ley, because the lower w/d ratios within indi- tions appear to facilitate anabranching. The vidual anabranches enhance bed shear stress and steeper, confined valleys in which type 6 channels sediment transport (Huang and Nanson, 2000). occur limit lateral mobility. Conversely, anabranching increases boundary resis- tance and may effectively consume surplus energy along rivers with very low sediment loads (Nanson and Huang, 2008). 5.2.6 Compound channels Although anabranching and anastomosing are sometimes used synonymously, Nanson Compound channel canbeusedtodescribeany and Knighton (1996) distinguished six types of channel that has distinct low- and high-flow por- anabranching rivers, including anastomosing. tions, including channels that simply overflow onto a floodplain. Compound channel can also refer specif- 1. Anastomosing rivers form in cohesive sediment ically to low- and high-flow portions with distinctly and low gradients, and have very low stream different planforms. Examples of the latter include power. This type of river has been described proglacial streams (Fahnestock, 1963) or tropical in diverse environments including the Rocky rivers (Gupta and Dutt, 1989) that switch seasonally Mountains of Canada, arid Central Australia, between meandering and braided as discharge fluc- Africa, and Wyoming, USA, and tropical South tuates. America. An intriguing example of a compound stream is 2. Sand-dominated, island-forming anabranching Cooper Creek in central Australia. This creek has a rivers rely on riparian vegetation to provide bank clay-bed anastomosing planform at low flow, then cohesion.Suchchannelshavebeendescribedfor dries out completely, allowing the clay particles to Australia, and have relatively low values of stream aggregateintosand-sizedpellets.Atthestartofthe power. wetseason,thepelletsaretransportedasbedloadin 3. Mixed load, laterally active anabranching rivers braided rivers, but the pellets disaggregate as flow have moderate values of stream power. This type continues, transitioning to the anastomosing clay of anabranching river ranges from small channels rivers as flows recede (Nanson et al., 1986). 144 Rivers in the Landscape 5.2.7 Karst channels 5.2.8 Continuum concept

Channels developed in karst terrains form a unique As noted earlier, classifications impose bound- subset of rivers in that they exhibit the proper- aries on continuous variations in channel planform. ties of open-channel flow, but exist in subterranean Leopold and Wolman (1957) recognized this explic- environments. As noted earlier, karst processes and itly, arguing that channel pattern reflects interac- forms are associated with rocks that are readily sol- tions among continuous variables such as sediment uble under surface or near-surface conditions, typ- grain size and volume, gradient, boundary erodibil- ically carbonates and evaporates. Rivers in karst ity, and flow energy. They proposed that a contin- terrains can flow underground for substantial dis- uum of channel patterns exists, with each pattern tances and then abruptly resurface as a spring where being defined by a combination of control variables. the water table and the karst aquifer intersect the Efforts to identify the most important control surface,orasariverinapocketvalleywherean variables and consistent correlations with channel impermeable substrate forces flow to the surface. pattern include, but are not limited to Conversely, surface streams can abruptly disappear underground in a blind valley.Or,asurfacestreamin r slope versus discharge (Leopold and Wolman, a dry valley can contain flow only during large runoff 1957) (Figure 5.11), under the assumption that inputs, despite existing in a wet climate, because discharge is an independent variable and slope of only large runoff inputs effectively fill subsurface alluvial channels reflects roughness, particle size, conduits and force flow to the surface (Ritter et al., and drainage area; 2011). r the percentage of silt and clay in the channel Subterranean rivers in cave systems can behave boundaries (Schumm, 1963), because the presence similarly to surface rivers in the sense of transport- of cohesive sediment strongly influences channel ing sediment from clay to boulder size clasts and stability, shape, and sinuosity; developing channel geometry that reflects adjust- r the ratios of depth to width (d/w)andslopeto ments between available energy and substrate resis- Froude number (S/Fr) (Parker, 1976), as a means tance (Springer et al., 2003). Karst rivers differ from of including channel form and flow energy param- most surface streams formed in bedrock in that eters; and a large portion of the sediment load can be car- r unit stream power versus median bed grain size ried in solution, and erosion of resistant channel (D50) (Van den Berg, 1995), as two boundary con- boundaries can produce distinctive sculpted forms ditions that are nearly independent of channel such as scallops and pockets via both abrasion and pattern. solution (Springer and Wohl, 2002). Along surface channels, a floodplain can attenuate the greater dis- Brotherton (1979) and subsequent investigators charge and energy present during a flood. Subter- have argued for the importance of the relative ease of raneankarstriversmayrespondtoenhanceddis- eroding the channel banks versus transporting bank charge with much greater flow depths and veloci- material. Where transport dominates, the channel ties, if the cave passage occupied by river channel remainsstraight,whereaserodiblebanksfacilitate is sufficiently large, and may also develop condi- braiding, with meandering as an intermediate sce- tions of pipe flow (Springer, 2004). (Pipe flow does nario (Knighton, 1998). not have the free surface found in open-channel In general, the progression from laterally stable flow.) In many respects, subterranean karst chan- (straight) channels through meandering to braided nels are analogous to laterally confined bedrock channels tends to correlate with increasing stream channelsatthesurface,withdownstreamchanges power, increasing w/d ratio (and hence increas- in substrate erodibility strongly influencing channel ing bank erodibility), and increasing volume and geometry. grain size of bedload (Ferguson, 1987). Although Chapter 5 Channel forms 145

(a) (b) Straight 0.1 Straight Meandering Braided Anastomosing 0.05 10–1 2–3 1–2braids braids Meandering 3–10 braids Braided s = Fd/w 0.01 Braided > 10 braids 0.005 10–2 s = 0.012 Q–0.44

0.001 0.0005 10–3 Channel slope (m/m) Slope/froude number Meandering 0.0001 BraidedTransition Meandering Transition Straight 0.00005 10–4 550 500 5000 50,000 10–5 10–4 10–3 10–2 10–1 100 10 100 1000 10,000 100,000 (Width:Depth ratio)–1 (d/w) Bankfull discharge (m3/s)

(c) +

Braided

Island braided Multi– channel Landform instability – Catastrophic change (bifurcation) – Single channel Ve (straight) Single channel g d e (meandering) yn ta a t m io i n cs + – Hydrogeomorphic + disturbance Figure 5.11 Different approaches to distinguishing channel planforms. (a) Braided and meandering with respect to slope versus discharge (from Leopold and Wolman, 1957). (b) Multiple channel types distinguished via slope/Froude number versus width/depth (from Parker, 1976). (c) Schematic representation of the continuity among different channel planforms, shown here in terms of a cusp catastrophe model (from Thom, 1975). This type of representation highlights how channels close to a threshold can have very different planforms (e.g., island braided vs. meandering), a condition in which a small change in the controlling variables, such as hydrogeomorphic disturbance or vegetation dynamics, can produce an abrupt change in channel planform (from Francis et al., 2009b, Figure 5). plotssuchasFigure5.11over-simplifythecom- change, are particularly well represented using a plexity of channel planforms, these plots do pro- cusp catastrophe model such as that in Figure 5.11c. vide useful insights regarding a given channel seg- By representing channel boundary resistance on the ment’s proximity to a planform threshold and its x and z axes, and magnitude, frequency, and dura- likely response to relatively small changes in exter- tion of flow energy on the y axis, this model incor- nal variables (Schumm, 1985; Knighton, 1998). porates the balance between hydraulic driving forces Transitions between different planforms, and the and substrate resistance in a manner that facilitates ability for channels to repeatedly cross thresholds as understanding of the relative importance of individ- conditions of boundary resistance and flow energy ual variables influencing channel planform. 146 Rivers in the Landscape

(a) (b) 450 m

(c) (d)

100 m

Figure 5.12 Schematic plan view illustration of river metamorphosis on the Great Plains of the United States. (a) Early 1800s: highly seasonal discharge maintains a broad, shallow braided channel with unvegetated bars. (b) Late 1800s: flow regulation creates more consistent flows with higher base flow and lower flood peaks, allowing riparian vegetation (gray shading) to establish along the channel banks and on some of the braid bars. (c) Early 1900s: droughts and flow regulation allow vegetation to establish below mean annual high water level, bars become islands, and single thalweg is dominant. Dashed lines indicate vestiges of historic channels that remain on floodplain. (d) Contemporary channel: bars and islands have become vegetated and attached to the floodplain, creating a wide, forested riparian corridor with a single narrow channel. (From Nadler and Schumm, 1981, Figure 10.)

5.2.9 River metamorphosis narrowing, deepening, and alteration of braided channels to meandering channels (Birken and The idea of continuity between individual chan- Cooper, 2006; Reynolds et al., 2012). r nelpatternsisalsoreflectedintherecognitionthat Similarly, river metamorphosis occurred along pattern can change abruptly in space and time. rivers of the US Great Plains in which flow reg- Schumm (1969) used the phrase river metamorpho- ulation caused a decrease in peak flows and an sis to describe the “almost complete transforma- increase in base flows. Changes in flow regime tion of river morphology” in response to natural or allowed native riparian vegetation to grow more human-induced changes. River metamorphosis typ- densely along the channel banks, and the com- ically refers to a rapid change in channel planform. bined effects of loss of peak flows and increased Numerous examples have been described: bank resistance caused braided rivers to narrow to a meandering planform (Williams, 1978b; Fig- ure 5.12). r r Klimek (1987) described metamorphosis of a Pol- A final example of a biotic external change that ish river when native vegetation was cleared across triggers river metamorphosis involves the pres- muchofthedrainagebasinanddramaticincreases ence or absence of beaver. As noted earlier, in sediment yield caused a meandering river to when present in a river, beaver build dams that become braided. promote channel–floodplain connectivity, over- r Flow regulation, along with introduction of exotic bank sedimentation and floodplain wetlands, and riparian vegetation that grows more densely along an anabranching planform. Beaver have been river banks, resulted in increased bank stability removed by trapping throughout much of their and sediment trapping that caused river metamor- historic range. In parts of the western United phosis along rivers in the southwestern United States, beaver have also been outcompeted by States. This metamorphosis took the form of large herbivores such as elk that have dramatically Chapter 5 Channel forms 147

increased in population density following twen- commonly smaller than the sum of the areas of the tieth century removal of predators. Elk eat the contributing channels. The presence of a flow sepa- same riparian woody species that support beaver, ration zone downstream from the junction can fur- and the combined loss of beaver dams and heavy ther reduce effective cross-sectional area. The higher elk grazing of riparian vegetation can result in velocitycanbeassociatedwithascourzoneinthe metamorphosis within two to three decades of an bed. The size and location of the scour zone reflect anabranching channel to a single incised channel thejunctionangleattheconfluenceandtheratioof with unstable banks (Beschta and Ripple, 2012; discharge in the confluent channels: dimensionless Polvi and Wohl, 2012). Where the effects of elk scour depth increases as the junction angle increases aresomehowreduced(e.g.,ripariangrazingexclo- and as the ratio of the minor tributary discharge to sures or reintroduction of predators) and beaver the mainstem discharge increases, as noted earlier return to a site, the channel can again metamor- for braided channels. A shear layer develops at the phose to multi-thread within a decade (John and margin of the flow separation zone because of the Klein, 2004). low velocity in this zone relative to the main flow. A shear layer with significant vorticity also devel- River metamorphosis is not so much a model of ops near the middle of the channel where tribu- how rivers adjust to changing external controls, as tary and mainstem flows merge (Figure 5.13). Tur- the concept that pronounced channel change can bulence associated with this shear layer may signif- occurveryrapidlyandinresponsetolimitedexter- icantly influence the bed scour zone. Strong sec- nal change—such as loss of flood peaks or change ondary circulation and helical flow also character- in biota—which then triggers numerous, nonlinear ize confluences. Velocity typically decreases beyond responses in river form. the confluence zone as flow merges into the single A specific aspect of changing channel planform receiving channel. is the presence, geometry, and erosional and depo- Bars commonly form at confluences. The loca- sitional processes occurring where two segments tion and type of bar varies in relation to plan- of channelized flow meet at confluences. Conflu- form geometry. Symmetrical confluences with sim- ences are inherent in multi-thread planforms such ilar angles between each tributary and the receiving as braided or anabranching rivers, but can also be channel commonly have a mid-channel bar down- very important where tributary channels join single stream from the confluence that reflects deposition or multi-thread main channels, or where secondary of sediment eroded from the scour zone (Best, 1986). floodplain channels rejoin the main channel. Asymmetrical confluences typically have bars asso- ciated with the zone of flow separation, in which low velocity and recirculation promote sediment depo- 5.3 Confluences sition(Figure5.13).Amongthemostwell-studied examples are the sand bars that form downstream Confluences between tributary and larger chan- from tributary junctions along the Colorado River nelsareubiquitousindrainagenetworks,andcon- in Grand Canyon, USA. fluences between subchannels are widespread in TheColoradoRiverinGrandCanyonexem- braided and anabranching channels. Confluences plifies a river within a laterally confined bedrock have received increasing attention in recent years canyon. Smaller, steeper tributaries entering the because these highly turbulent locations strongly Colorado deposit bars and fans, commonly with influence processes as diverse as contaminant dis- episodic deposition during flash floods and debris persal and navigation (Gaudet and Roy, 1995). flows on the tributaries. Coarse-grained tributary Average and maximum values of flow velocity deposits create rapids on the Colorado, and con- typicallyincreaseatconfluencesbecausethecross- strict the mainstem, leading to development of sectional area downstream from the confluence is a shear layer, strong secondary circulation, and 148 Rivers in the Landscape

(a)Uniform (b) flow

Q V t t Scour Stagnation w Flow separation zone 2 zone Flow directions at bed Avalanche face edge Zones of sediment deposition QmVm Flow w3 Qtot w 1 recovery Approximate cross-channel Uniform flow sediment transport rates Boundary of flow separation zones Helicoidal flow cells Shear layer Maximum Flow velocity Flow separation zone deflection A’ zone ayer Shear l

A

Helicoidal flow cells A Shear layer A’

Figure 5.13 Hydraulics, cross-sectional geometry, and depositional features associated with channel confluences. (a) Conceptual model of flow dynamics at channel confluences, as derived from laboratory observations foran asymmetrical, 90 degrees confluence with main channel and tributary of equal width and depth. Q is discharge, V is average velocity, and w is channel width: Subscript m indicates the main channel and subscript t indicates the tributary. (From Best, 1987, Figure 4.) (b) Features of flow dynamics, bed morphology, and sediment transport ata symmetrical planform confluence. (From Mosley, 1976, Figure 3, and Best, 1987, Figure 5.) finer-grained separation and reattachment deposits strongly influenced by mass movements coming (Figure 5.14). The separation and reattachment down the steeper tributaries. Low-order tributaries deposits are typically sand bars that form impor- prone to debris flows can introduce abundant coarse tant riparian habitat and recreational sites for thou- sediment and wood into the main channel, result- sandsofpeoplewhofloatdowntheColorado ing in more heterogeneous channel morphology and through Grand Canyon each year, and the eddy greater habitat diversity in the main channel (Benda return-current channel creates a backwater that pro- et al., 2003a). Tributary inputs can have minimal vides important habitat for endangered native fish effectswheretransportcapacityofthemainstemis (Schmidt and Graf, 1990). Flume experiments and sufficient to rapidly redistribute tributary sediments. hydraulic modeling indicate that the length of the The likelihood that tributary inputs will significantly separation zone reflects hydraulics and topography influencechannelmorphologyonthemainstem ofthechannelbeddownstreamfromtheconfluence. increases with the size of the tributary relative to the Aggradation within the separation zone effectively mainstem (Benda et al., 2004) decreases the length of this zone (Schmidt et al., Confluences can be characterized as concordant 1993). The great majority of marginal deposition in a when two channels of equal depth join, or the channel configuration such as that in Grand Canyon more common discordant confluences when chan- occurs in recirculation zones (Wiele et al., 1996). nels of different depth join (Robert, 2003). (Dis- Confluences between tributary and main chan- cordant confluences reflect differences in conflu- nels in mountainous environments can also be ent channel dimensions as a result of differences in Chapter 5 Channel forms 149

(a) Debris fan

Recirculation zone Reattachment point Secondary eddy Low-velocity area

Return current

Separation Primary eddy point surface Separation

Constriction Expansion

Primary-eddy return-current (b) channel Separation Reattachment deposit deposit

Upper-pool deposit Debris fan

Main channel Channel-margin deposit

Figure 5.14 Separation and reattachment deposits in the Grand Canyon of the Colorado River. (a) Flow patterns. (b) Configuration of bed deposits. Gray shading indicates sand deposition. (From Schmidt and Graf, 1990, Figure3, p. 5.) discharge and boundary resistance; i.e., differences most extreme case, this can result in a hanging valley in hydraulic geometry.) The vertical mixing layer (Section 5.4.3), with a substantial waterfall between present in concordant confluences becomes more the tributary and mainstem. complicated at a discordant confluence, where the suddendropinbedelevationfromthetributaryto the mainstem creates complex 3D flow and vertical 5.4 River gradient upwelling (Best and Roy, 1991). The greater the bed discordance, the shorter the mixing length for the The final level of adjustment in channel form occurs converging flows (Gaudet and Roy, 1995). via changes in gradient. These can be changes in In regions with active tectonic uplift or relative reach-scale gradient that occur over lengths tens base level fall, tributaries may not be able to incise to hundreds of times the average channel width. as rapidly as the mainstem, leading to a pronounced Gradient can also change over all or much of the lon- drop at the tributary–mainstem confluence. In the gitudinal profile of a river. 150 Rivers in the Landscape

Like channel planform, gradient can be largely val. Rivers in the northwest Indian Himalaya, for imposed on a river by external constraints such as example, display increased concavity downstream changing base level or erosionally resistant substrate, from glacially modified reaches and require more or gradient can be a response to existing water and than 500,000 years to recover a graded condition sediment supply. In the latter case, gradient repre- (Hobley et al., 2010). sents another dimension of channel adjustment and Discontinuities in process and form along a chan- exhibits changes in response to varying water and nel can occur in association with tributary junctions. sediment supply to the channel. The idea of gradient These junctions can represent a substantial increase as a reflection of external inputs to the river is for- in discharge and/or an increase in the volume or mally expressed in the concept of a graded stream. average grain size of sediment (Section 5.6.1) sup- The most widely cited definition of a graded plied to the channel. stream is that proposed by Mackin (1948, p. 64): “A Discontinuitiesinchannelprocessandformcan graded stream is one in which, over a period of years, also occur at transitions between bedrock, coarse- slope is delicately adjusted to provide, with avail- grained alluvial, and fine-grained alluvial substrates able discharge and the prevailing channel character- (Howard, 1980). The gradient of alluvial channels istics, just the velocity required for transportation of is commonly interpreted as reflecting hydraulic alloftheloadsuppliedfromabove.”Mackintraced regime (Moshe et al., 2008). This condition is this idea back to Gilbert (1877) and Davis (1902a), referred to as transport limited (Howard, 1994) although Davis viewed a graded stream as a condi- because the transport of sediment is limited primar- tion developed only over very long periods of geo- ilybyflowenergyratherthansedimentsupply.In logic time. Mackin, following Gilbert’s perception, contrast, sediment transport in supply-limited chan- identified a graded stream as reflecting equilibrium, nels is constrained primarily by the availability of and applied this idea to the entire longitudinal pro- sediment. Bed material grain size in a supply-limited fileofariver,aswellastomorelimitedsegments. channel adjusts by coarsening until the remaining The earlier definition of a graded stream implies that exposed sediment is moved only in proportion to stream geometry is relatively constant over the time supply of that grain size. If sufficient supply is not period of interest, although fluctuations about a con- available, bedrock is exposed along the channel, cre- sistent mean can occur. ating a subset of supply-limited conditions known A graded stream can develop relatively quickly— as detachment limited. In detachment-limited chan- in some cases, over a matter of hours to a few days— nels, gradient may be less dependent on hydraulic in channels with readily erodible substrate. A graded regime because weathering must precede erosion condition typically requires longer to develop over (Howard, 1994, 1998). an entire longitudinal profile or in more erosionally Alluvial channel gradient shows a weak inverse resistant substrate. Developing a graded condition correlation with discharge. Similarly, the relation- over the length of a river requires that all portions ship between gradient and median bed material size of the river adjust to base level and that bed sed- is not simple, but can be significant when sites of iment along the river is distributed in adjustment similardrainageareaordischargearecompared with discharge. Before these adjustments can be (Hack, 1957). This indicates that gradient reflects accomplished along substantial portions of a river, both discharge and sediment supply (Knighton, however, boundary conditions (base level, sediment 1998). This is logical, given that bed material size and water yield, substrate resistance) are likely to and discharge influence particle mobility, channel change. Developing a graded condition in more ero- boundary roughness, rates of energy expenditure, sionally resistant substrate requires a sufficient fre- and thus ability to adjust gradient. In channels quency and duration of flows of high magnitude that with mixed grain sizes, questions arise as to which exceed the threshold of boundary erodibility, and grain-size fraction exerts the strongest influence on such flows are likely to be of longer recurrence inter- ability to adjust gradient. Most investigators use Chapter 5 Channel forms 151 some grain-size fraction coarser than D50 as better and a reference portion of a river changes. This can representing the influence of grain size on gradient occur when tectonic uplift raises the drainage basin, in gravel-bed streams. for example, even if the ultimate base level of sea Athresholdgrainsizeof10mmseparatessand- level remains constant during the period of uplift. bed and gravel-bed streams (Howard, 1987). Gra- Sealevelhasofcoursefluctuateddramaticallydur- dient reflects quantity and grain size of load below ing the Quaternary in association with the advance this threshold, whereas gradient on channels with and retreat of continental ice sheets, causing ulti- bed material coarser than 10 mm reflects thresh- mate and relative base level to change repeatedly. old of motion of large grains rather than quantity In the description of longitudinal profiles, as in of sediment. Sand-bed channels are referred to as many other areas of fluvial geomorphology, G.K. live-bed or regime channels because sediment trans- Gilbert led the way with his three laws of land sculp- portoccursatallbutthelowestflows.Gravel-bed ture, which included the law of divides.Thissim- channels are known as threshold or stable channels ply states that the gradient of a river steepens with because sediment moves only near bankfull dis- proximity to the drainage divide (Gilbert, 1877). charge or during extreme flows (Howard, 1980). Although longitudinal profiles can be straight or Thesechannelscommonlyhavegradientsnearthe convex (Figure 5.15), overall gradient of medium- thresholdofmotionforthecoarsegrainfraction to large-sized rivers typically decreases downstream, (Howard et al., 1994). creating a concave longitudinal profile that Hack (1957) described as 5.4.1 Longitudinal profile S = kLn (5.7) where S is gradient, k incorporates mean bed parti- Longitudinal profile typically refers to gradient along cle size, L is distance downstream from the drainage theentirelengthofariverfromthechannelheadto divide, and the exponent n is an index of profile con- the base level at which a river enters a larger river or cavity. S in this equation is a tangent to the curve a body of standing water. The concept of base level is that defines the relation between fall H and length particularly important in understanding adjustment L along a river (i.e., a longitudinal profile). of river longitudinal profile. First articulated by J.W. Based on field data from the eastern United Powell (1875, 1876), base level can be conceptual- States, Hack (1957) proposed an empirical form of ized as a lever arm that influences channel bed ele- Equation 5.7 using mean bed particle size (D50,in vation and gradient upstream from the base level. mm) and distance downstream from the drainage If base level increases in elevation, upstream gra- divide (L,inmiles) dient declines, transport energy declines, and sedi- ( ) mentistypicallydeposited,causingthechannelto D0.6 S = 25 50 (5.8) aggradetothenewbaselevel.Ifbaselevelfalls,gra- L dient increases, and the channel incises to the new base level starting at the point of base level drop and For channel segments with constant D50, integra- propagating upstream. tion of Equation 5.7 produces Local base level can occur partway along a river, where the river enters a lake that has river outflow H = kInL+ c (5.9) downstream. Local base level also refers to an ero- where c and k are empirically derived constants. sionally resistant layer that limits upstream trans- missionofultimatebaselevelfall.Sealevelformsthe (The integration is ultimate base level for rivers. S = kLn Relative base level change refers to a scenario in H = kLndL which the elevation difference between base level = klogeL + c,wheren equals − 1) 152 Rivers in the Landscape

(a) Ouzel Creek Cony Creek Hunters Creek (area 14 km ) (area 20 km ) (area 12.5 km ) Elevation (m) Elevation Elevation (m) Elevation (m) Elevation

Distance downstream (m) Distance downstream (m) Distance downstream (m)

3300 North St. Vrain Creek 3200 (drainage area 90 km2)

3100

3000

2900 Ouzel Elevation (m) 2800 Cony 2700 Hunters 2600

2500 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Distance downstream (m) (b) 4000

3000 AMAZON RIVER (drainage area 6,150,000 km2) RHINE RIVER 2000 2000 (drainage area 252,000 km2)

1000 Elevation (m) 1000 Elevation (m)

0 0

0 1000 2000 3000 4000 0 200 400 600 800 1000 Distance from source (km) Distance from source (km) Figure 5.15 Sample longitudinal profiles for diverse rivers. (a) A mountainous river network in Colorado, USA. Ouzel, Cony, and Hunters Creeks are each tributary to North St. Vrain Creek at the points indicated on the longitudinal profile of North St. Vrain Creek. Each of these longitudinal profiles is relatively straight or convex, and the tributaries steepen as they enter the main valley of North St. Vrain Creek, which was glaciated during the Pleistocene. (b) The Amazon and Rhine Rivers, which display much more concave profiles. In each of these large rivers, the majority of elevation loss occurs in the mountainous headwaters.

In the much more common scenario in which Snow and Slingerland (1987) used numerical particle size changes systematically downstream, simulations to demonstrate that the specific form (i.e., n does not equal −1), the equation takes the of mathematical function that best approximates a form of ( ) river’s longitudinal profile depends on the relative k rates of change in water and sediment discharge and H = L(n+1) + c (5.10) n + 1 grain size downstream. Exponential, logarithmic, Chapter 5 Channel forms 153 and power functions all create smooth, concave- inputs of sediment, tectonic activity, or continuing upward longitudinal profiles, but the profiles of response to base level fall. Although irregularities real rivers typically contain irregularities (Knighton, generated by any of these mechanisms can persist in 1998). alluvial channels, they are more likely to persist in The index of profile concavity (exponent n in bedrock channels. Alluvial channels respond more Equation 5.7), as suggested by the correlations quickly than bedrock channels to local perturba- between reach-scale gradient and discharge and tions, and alluvial channels are more likely to adjust grain size, reflects rates of downstream change in channel parameters such as cross-sectional geome- discharge and substrate resistance. Profiles tend to tryandplanform,aswellasgradient(Schummetal., be more concave where discharge increases rapidly 1987). downstream or grain size decreases rapidly. Increas- Irregularities in the longitudinal profile have the ing discharge translates to ability to transport the potential to provide important insight into river his- same bed-material load over progressively lower toryandadjustment.Consequently,theyhavebeen slopes. Similarly, decreasing grain size implies the the subject of much attention, including how to ability to transport the available load over progres- quantify the irregularities using SL or DS indices. sively lower slopes. Adjustments to increasing dis- charge or decreasing grain size can also occur in terms of bed roughness, cross-sectional geometry, 5.4.2 Stream gradient index and planform (Dust and Wohl, 2012b). The longitudinal profile of a river can also be A semi-log plot of the longitudinal profile of chan- 휃 characterized in terms of the concavity index nel segments with a constant value of D50 should be a straight line, the slope of which (k) Hack (1957) −휃 S = ksA (5.11) referred to as the stream gradient, or SL, index.SL index can be calculated as the product of gradient where S is channel gradient, ks is the steepness index, (S) and total stream length (L) from the divide. This and A is drainage area (Flint, 1974). This relation is the most common theoretical form of an equilib- holds downstream of a critical drainage area. Crit- rium longitudinal profile against which actual pro- ical drainage area varies, but is typically in the range files are assessed (Goldrick and Bishop, 2007). of 0.2–0.9 × 105 m2 (Whipple, 2004). Values of 휃 for Hack (1973) noted that abrupt spatial changes bedrock channel segments vary from 0.3 to 1.2, and in stream gradient, as indicated by discontinu- can be negative over short reaches. Concavity val- ities in the longitudinal profile of a river, can be ues <0.4 are associated with short, steep drainages an equilibrium response to substrate variations, or strongly influenced by debris flows or with down- adisequilibriumresponsecausedbylocaldefor- stream increases in incision rate or rock strength mation or upstream propagation of a knickpoint that result in knickpoints. Moderate values of 0.4– caused by relative base level fall. An equilibrium 0.7 correlate with active uplift and homogeneous response to spatial variations in substrate erodibil- substrate, and high values of 0.7–1 correlate with ity should persist because steeper channel segments downstream decreases in rock uplift rate or rock form in more resistant substrate. A disequilibrium strength (Whipple, 2004). The concavity index indi- response should be transient because reach-scale cates the rate of decline in channel bed gradient with gradient decreases after the knickpoint migrates increasing drainage area: the higher the index, the upstream. There is no inherent mechanism, how- faster the rate of decline in channel bed gradient. ever, for differentiating equilibrium steepening asso- As noted previously, even predominantly concave ciated with greater substrate resistance from dise- longitudinal profiles tend to exhibit irregularities in quilibrium steepening associated with relative base theformoflocalsteepening.Theseirregularities level fall on an SL plot. Consequently, Goldrick and can result from more resistant substrate, large point Bishop (2007) proposed the DS approach, derived 154 Rivers in the Landscape from the power relationship between discharge and Consequently, bedrock longitudinal profiles have downstream distance, and the dependence of stream beenusedasanindexofrockupliftrateinsteady- incision on stream power state landscapes that maintain statistically invariant ( ) topography and constant denudation rate (Whip- L1−휆 ple, 2001), and in landscapes experiencing transient H = H − k (5.12) o 1 − 휆 increases in erosion rates (Snyder et al., 2000; Whit- taker et al., 2008; Roberts and White, 2010). The where Ho is an estimate of the theoretical elevation of assumption is that steeper portions of the profile thedrainagedivideifhydraulicprocesseswereactive reflect greater uplift. The ability of a river to main- right to the drainage head, k reflects factors such tain profile concavity in response to uplift, greater as equilibrium incision rate and stream hydraulic substrate erosional resistance, large sediment inputs, geometry as well as lithology, and 휆 is the exponent orbaselevelfallcanbeafunctionofdischarge,with of the relationship between discharge and down- larger rivers capable of responding more quickly and stream distance. Equation 5.12 is essentially a repa- maintaining smoother, more concave profiles (Mer- rameterized version of Equation 5.10, with –k for k ritts and Vincent, 1989). Knickpoints, in particular, and –휆 for n.TheDS formmaybemoreappropriate reflect an inability to maintain longitudinally con- for tectonically quiet areas than the SL form, because tinuous profile concavity. the DS form differentiates equilibrium steepening, A knickpoint is a step-like discontinuity in a river’s which appears as a parallel shift in a DS plot, from longitudinal profile. A knickzone is a river segment disequilibrium steepening, which appears as disor- steeper than upstream and downstream segments, dered outliers on a DS plot (Goldrick and Bishop, but which does not have a pronounced vertical 2007) (Figure S5.11). discontinuity such as a waterfall. Knickpoints can occurinweaklyconsolidatedalluvium,buttheyare best developed in cohesive alluvium or bedrock. 5.4.3 Knickpoints Knickpoints can be stepped, buttressed, or under- cut, with headward erosion via parallel retreat or As noted previously, irregularities are more likely to rotation of the knickpoint face (Figure 5.16). Knick- persist in bedrock channels than in alluvial channels. points can migrate upstream from a site of base

(a) (b) Figure 5.16 Persistent and ephemeral knickpoints along channels. (a) This bedrock knickpoint where a tributary enters the Wulik River in Alaska, USA, represents the inability of the tributary to incise through the bedrock as rapidly as the mainstem river (The Wulik River flows right to left in the foreground of this view). (b) This knickpoint along an ephemeral channel tributary to the South Fork Poudre River in Colorado, USA formed in response to enhanced water yield after a wildfire that burned the catchment 4 months before the photograph was taken. Chapter 5 Channel forms 155 level fall (Crosby and Whipple, 2006), or as a result oftheincreaseindischargerelativetosediment supply. Knickpoints can form at a local base level created by more resistant substrate, in which case they are likely to disappear once the river incises an inner channel through the resistant material (Wohl et al., 1994). Knickpoints can also form where alargepointsourceofsedimentsuchasaland- slide overwhelms fluvial transport capacity (Korup et al., 2006). Rates of knickpoint retreat can be two orders of magnitude greater than erosion rates elsewhere along the channel (Seidl et al., 1997). Knickpoints are the geomorphic hot spots of incision along a river’s longitudinal profile. Rates of knickpoint retreat can correlate with drainage area and thus fol- low the stream power law (Equation S4.1) (Bishop et al., 2005). Rates of retreat also reflect spatial vari- ations in rock erodibility (Harbor et al., 2005) and sediment flux passing over the knickpoint lip (Lamb et al., 2007). Cosmogenic 10Be dating of strath ter- races downstream from headward-retreating knick- points in western Scotland, for example, indicates that knickpoint retreat rates have declined since the mid-Holocene. Jansen et al. (2011) attributed this decline to a depletion of paraglacial sediment sup- Figure 5.17 An example of a tributary hanging valley ply and a deficiency of abrasive tools to erode the in a glaciated river network, here in Yosemite National bedrock knickpoints. Park, California, USA. The mainstem river flows left to At the network scale, knickpoints at tributary right in the foreground. junctions can create hanging valleys in which the tributary junction is perched well above the main valley bottom. Hanging tributary valleys are com- which large tributaries can keep pace with base level mon in regions with alpine glaciation. Smaller trib- fall and mainstem incision. utary glaciers do not erode the valley as effectively Crosby et al. (2007) predicted the maximum as larger glaciers, leaving the tributary valley hang- drainage area at which a tributary hanging valley, ing when the glacial ice recedes (Figure 5.17). Hang- Atemp,canform ing valleys can also form in tectonically active and ( ) 1 incising landscapes with no history of glaciations. b 1−bc kwk I Under these conditions, rapid mainstem incision q max Atemp = 훽 (5.13) over-steepens tributary junctions beyond a thresh- KGA Uinitial oldslope,orlowtributarysedimentfluxduring mainstem incision limits the tributary’s ability to where kw is a coefficient for the relation between inciseasrapidlyasthemainstem(Crosbyetal., channel width and discharge, kq is a coefficient for 2007). The amount of over-steepening needed to the relation between discharge and drainage area, formahangingvalleyincreasesastributarydrainage KGA is a dimensional constant equal to (r/Ls), with area increases up to some maximum value, above r asthefractionofthevolumedetachedoffthebed 156 Rivers in the Landscape

5 10 Climate, tectonics 104 Profile Land use, gradient river engineering 3 Profile 10 concavity Reach Profile gradient form 2 10 Meander Wavelength Plan form Bed configuration: 101 Gravel-bed streams

Increasing length scale, m Channel width Channel 100 depth Cross-sectional Bed configuration: form Sand-bed streams 10–1 10–1 100 101 102 103 104 Increasing time scale, years Figure 5.18 Temporal and spatial scales of river adjustment. (Modified from Knighton, 1998, fig 5.3, p158).

훽 with each collision and Ls as saltation hop length, of the material covered thus far in this chapter and isthepercentageoferodedmaterialtransportedas in preceding chapters. The magnitude and duration bedload, Imax is maximum incision rate, and Uinitial of a flow hydrograph influence the energy avail- is background rate of base level fall. able to perform work in the channel and thus influ- This equation describes a transient instability in ence channel form, for example, but also respond a river with a profile that reflects incision dependent to channel form as form influences travel time of on sediment flux. Crosby et al. (2007) assumed that flood pulses. Hydraulic forces influence sediment changes in sediment flux lag behind profile adjust- mobility and bedforms, but also respond to changes ment because the hillslope response that determines in boundary roughness caused by mobile sediment sediment flux depends on the transmission of base and bedforms. Rate of erosion governs whether a level fall through the network. Temporary hanging longitudinal profile can maintain concavity during valleys form after base level fall if the mainstem tectonic uplift, but profile concavity also influences transient incision rate exceeds the tributary maxi- energy available for erosion. mum incision rate associated with the initial sedi- Channel forms are typically described at the spa- ment flux. The terms in Equation 5.13 describe how tial scales of cross-sectional geometry, reach-scale the over-steepened reach Atemp adjusts as the main planform and gradient, and basin-scale longitudinal stem incision rate returns to just balancing the back- profile. Adjustment of channel form can similarly ground rate of base level fall (Uinitial)whilethetrib- occur at various spatial scales, including changes in utary continues to incise at the rate Imax,whichis w/d ratio or grain size and quantity of bed material, dependent on abrasion reflected in the variables of changes in sinuosity or number of subparallel chan- 훽 KGA and . nels (if the channel is not laterally confined), and changes in gradient and shape of the longitudinal 5.5 Adjustment of profile. Figure 5.18 illustrates the temporal and spatial channel form scales over which various form components of rivers adjust to changing inputs and processes. Feedbacks between channel processes and various This figure provides a very useful conceptual aspectsofchannelgeometryareimplicitinmuch frameworkforthinkingaboutfeedbacksand Chapter 5 Channel forms 157 adjustment in rivers because the diagram illustrates mizes or minimizes the value of a specific parameter that (Darby and Van De Wiel, 2003). Examples include minimization of rate of energy dissipation (Yang, r smaller spatial-scale components of a river net- 1976) or stream power (Chang, 1988), and the max- work (e.g., bed configuration) can change rela- imization of sediment transport rate (White et al., tively quickly compared to larger scale compo- 1982) or friction factor (Davies and Sutherland, nents such as planform; 1983; Abrahams et al., 1995). This approach stems r for a given spatial scale or type of component, such from Langbein and Leopold’s (1964) conceptualiza- as bedforms, the rate of adjustment increases as tion of river form as the most probable state between the erosional resistance of the material increases— the opposing tendencies of minimum total rate of in other words, live-bed channels and sand bed- work and uniform distribution of energy expendi- forms adjust more quickly than threshold chan- ture. nels and boulder-bed step–pool sequences; Extremal hypotheses have been criticized as r riversincludemultiplecomponentsthatcanadjust being teleological and lacking explanatory power to changes in boundary conditions (water and sed- (Ferguson, 1986), but extremal hypotheses do iment input, substrate erodibility, base level), mak- explain a wide range of observations (Darby and ing it difficult to precisely predict the nature of Van De Wiel, 2003). The foundational assumption river adjustment; and in extremal hypotheses is typically that a high rate r space and timescales over which adjustments of energy expenditure at a specified point in a chan- occur in diverse river features overlap, implying nelwilleventuallyresultinboundarydeformation that multiple aspects of a river are likely to respond until the rate of energy expenditure declines or the to a change in boundary conditions, as implied by boundary is no longer adjustable by the energy avail- the expanded Lane’s balance (Section 5.1.4). able. An example for steep, mobile-bed channels is the hypothesis that interactions between hydraulics and Channel adjustment fundamentally represents bed configuration prevent the Froude number from some alteration of channel form in an attempt to exceeding 1 for more than short distances or periods reach equilibrium, as reflected in channel geome- of time (Grant, 1997). Observed cyclical patterns of try and rate of flow energy expenditure, and as con- creation and destruction of bedforms (particularly trolled by independent variables such as sediment antidunes) effectively maintain critical flow in these and water yield, substrate resistance, and base level. channels. Extremal hypotheses have also been used The final two sections of this chapter examine (i) to explain the development of bank roughness in conceptual models developed to explain the physi- extremely deep, narrow bedrock slot canyons (Wohl cal processes that limit channel adjustment, and (ii) et al., 1999), cross-sectional geometry (Eaton et al., the high flows that are responsible for a substantial 2004), and anabranching channel planform (Huang proportion of channel adjustment in many channels, and Nanson, 2007), among other characteristics. respectively.

5.5.2 Geomorphic effects of floods 5.5.1 Extremal hypotheses of channel adjustment Thegeomorphicimportanceoffloodsofvarying magnitude differs widely among rivers. Along rivers Channel adjustment is sometimes conceptualized with relatively low hydrologic variability through using extremal hypotheses,whicharemodelsbased time, the largest floods are unlikely to have per- on the assumption that equilibrium channel mor- sistent effects on channel and valley morphology phology corresponds to the morphology that maxi- because subsequent smaller flows quickly modify 158 Rivers in the Landscape erosional and depositional features created during r longitudinal grooves—elongate linear grooves large floods. Large floods are more likely to cre- parallel or subparallel to the local direction of ate persistent effects as erosional resistance of chan- flood flow, tens to hundreds of meters long, nel boundaries increases and as hydrologic vari- form in groups, individual grooves spaced 0.5–3 ability increases (Kochel, 1988), because subsequent mapart,widthanddepthvaluesrangefrom smaller flows are less able to modify the geomorphic centimeters to greater than a meter; effects of large floods. r channel widening and incision; Within a region, large floods are typically more r stripped floodplains, which occur where general important in highland rivers than in lowland rivers scouring that is not restricted to a well-defined (Froehlich and Starkel, 1987; Patton, 1988; Eaton scourmarkorchannelremovesvegetationand et al., 2003). Highland rivers have steep gradi- fine-grained alluvium to depths of up to 1.5 m; ents, narrow valley bottoms, coarse bedload, rela- r anabranching erosion channels, which reflect tively flashy hydrographs, and less erodible chan- incomplete channel widening that creates rem- nel boundaries, all of which magnify the geomor- nant islands in expanded channels; phic effects of floods (Kochel, 1988). Lowland rivers r cutoffchutesintheformofwell-definedchannels have larger buffering capacity associated with well- that are typically several hundred meters long; and developed floodplains and greater drainage area, as r erosion of tributary fans impinging on the flood- well as finer bedload and channel boundaries that plain and main channel (Figure 5.19). can be mobilized by smaller discharges. Floods can thus be of varying geomorphic effectiveness—defined Depositional flood features include as the ability of a flow to modify channel morphol- ogy (Wolman and Gerson, 1978)—across different r gravel bars within and along the margins of the segments of a river network. channel; The geomorphic effectiveness of a particular r wake deposits in the lee of a large obstacle to flow; flood typically varies spatially along a river and r slackwater deposits of sediment deposited from between neighboring rivers in response to spatial suspension in areas of flow separation; variations in flood hydraulics, sediment supply, and r terrace-like boulder berms; and erodibility of the channel boundaries (Miller, 1995; r aggradationwithinchannels. Cenderelli and Wohl, 2003; Procter et al., 2010). Antecedent conditions and the magnitude of a flood Depositional features in overbank areas include relative to earlier floods also influence geomor- gravel splays and gravel and sand sheets (Miller and phic effectiveness (Eaton and Lapointe, 2001). The Parkinson, 1993). Gravel splays are lobate features greater the difference in magnitude between a par- in planform, with a convex profile. They are asso- ticular flood and previous flows, the more likely ciated with severe channel or floodplain erosion, that particular flood is to create substantial channel and are deposited where confined flow becomes change. unconfined, such as main-channel flow breaching Erosional features created by large floods over a levee and entering the floodplain. Gravel and cohesive substrates (Baker, 1988) include sand sheets are typically broader and thinner than splays. Floods can also drive substantial changes in r sculpted features (Section 4.3.3; Supplemental channel planform, typically from single channel to Section 4.1.3); braided (Scott and Gravlee, 1968; Cenderelli and r inner channels (Section 4.1.3); and Cluer, 1998). r knickpoints (Section 5.4.3). Themagnitudeofchannelchangeduringa flood can be difficult to predict, but the longitu- Flood erosional features in unconsolidated mate- dinal distribution of predominantly erosional and rials (Miller and Parkinson, 1993) include depositional flood features correlates strongly with Chapter 5 Channel forms 159

(a) (b) Figure 5.19 Examples of flood erosional features along the Big Thompson River in Colorado, USA during an extreme flood in 1976. (a) Channel widening along the upper, alluvial reaches of the river. (b) Channel widening in abedrock- constrained canyon resulted in destruction of a bridge. Both photographs courtesy of Stanley A. Schumm. valley and channel geometry and is thus predictable. Steep, narrow reaches are likely to experience, pre- A B Bedrock erosion threshold )

dominantly, erosion during floods, whereas rela- 2 tively low-gradient, wide reaches become sites of Energy available for deposition (Shroba et al., 1979). Despite the site- geomorphic change specific nature of erosion and deposition, various thresholds have been proposed for significant chan- nel modification during floods, including a mini- Alluvial erosion threshold 2 Stream power (W/m mumstreampowerperunitareaof300W/m for minimal erosion low-gradient, alluvial channels (Magilligan, 1992) C and an empirical power relation between stream power per unit area and drainage area (휔 = 21 A0.36) (Wohl et al., 2001). Costa and O’Connor (1995) pro- Time posed distinct thresholds for different types of sub- Figure 5.20 Schematic illustration of how the energy strates and channel modification during floods (Fig- available for geomorphic change during a flood varies in ure 5.20). If such thresholds can be quantified for a relation to flood magnitude and duration. Hydrograph A channel reach, and a measure of flow energy avail- represents a peaked, short duration flood of the type that results from a convective storm or damburst. Hydrograph able to perform geomorphic work (e.g., excess shear B represents a longer duration, high-magnitude flood, stress or stream power) can be quantified over some such as might result from a cyclone. Hydrograph C repre- time interval based on flood magnitude, duration, sents a sustained, low-magnitude flood of the type that and frequency, then quantitatively estimating flood could occur during snowmelt. (From Costa and O’Connor, effectiveness should be possible. 1995, Figure 11, p. 54). 160 Rivers in the Landscape 5.6 Downstream trends Abrasion is commonly quantified using some form of Sternberg’s (1875) law Some components of channel form, such as cross- = −훼x sectional geometry, typically exhibit consistent D D0e (5.14) downstream trends as drainage area and, usually, discharge increase. Other components, such as bed- where initial grain size D0 wears down to D at form configuration or bank stability, may or may distance x from the origin at a rate given by the 훼 not exhibit consistent downstream trends (Wohl, rock erodibility parameter , which sets the abra- 훼 2008b). The absence of consistent longitudinal sion length scale, 1/ (Sklar et al., 2006). Values of 훼 canvarybyatleasttwoordersofmagnitude,from trends typically results from spatial variation in −3 −5 other controls, such as tectonic regime, lithology, 10 /m to 10 /m, but tend to decrease as the bed or climate, that override the effects of increasing material grows finer. In this general form of the 훼 discharge. Among those components of channel Sternberg law, incorporates both selective sort- formthatexhibitsomedownstreamtrends,butcan ing and abrasion (Knighton, 1998). Fracturing of also exhibit substantial longitudinal variability, are clasts is typically lumped under abrasion, although bed grain size and instream wood. Chatanantavet et al. (2010) suggested distinguishing these processes (Figure S5.12). Downstream fining can be interrupted by 5.6.1 Grain size coarse sediment from hillslope or tributary inputs (Attal and Lave,´ 2006) or from knickzone erosion Average bed grain size decreases downstream in (Deroanne and Petit, 1999). Mountainous catch- most rivers at scales of tens to hundreds of ments with close coupling of hillslopes and channels kilometers—a pattern known as downstream fining. may have downstream grain-size distributions that Downstream fining has been attributed to selective closely mirror the size distribution of sediment sorting (entrainment, transport, and deposition), supplied from hillslopes because local resupply abrasion in place or during transport, or some com- fromhillslopesoffsetstheinfluenceofchannel bination of these processes (Powell, 1998). The rela- processesthatcreatedownstreamfining(Sklar tive importance of selective sorting versus abrasion et al., 2006). reflectsfactorssuchasclasterodibility.Abrasionis Mountainriverscanalsohavedownstreamcoars- more important where clasts abrade readily (Parker, ening. Grain size coarsens to a maximum value 1991). atthelocationalongthechannelwheretransport In other settings, gravel-to-sand transitions occur dominatedbydebrisflowsgiveswaytotransport over distances too short for substantial abrasion. dominated by fluvial processes. Downstream fining Ferguson (2003) attributed these transitions to non- occurs below this transition (Brummer and Mont- linearities in bedload transport in which small gomery, 2003). increases in sand content lead to large increases in Significant lateral sediment sources that influ- the mobility of sand and gravel. As sand deposi- ence downstream fining can be identified using tion decreases the effective bed roughness, near-bed drainage basin area, network magnitude, and the velocities increase, shear stress and turbulent kinetic basin area–slope product to define individual chan- energy decrease, burst and sweep events become less nel links within which downstream fining occurs frequent,andthesandfractionremainsmoremobile (Rice, 1998). The travel distance required to abrade than coarser particles. All of these factors contribute coarser tributary sediments to the size of mainstem to creating a sand-bed channel downstream (Sam- inputs from upstream can be predicted as brook Smith and Nicholas, 2005), as do decreases ( ) in transport capacity (Ohmori, 1991; Parker et al., 1 D L∗ = ln t (5.15) ΔD 훼 2008). Dm Chapter 5 Channel forms 161

∗ where L ΔD is the distance over which the grain size the absence of wood (MacFarlane and Wohl, 2003; perturbation decays, D is mean grain size from the Mao et al., 2008); t r tributary, and Dm is mean grain size in the main- creating forced alluvial reaches by creating back- stem, and 훼 is the rock erodibility parameter (Sklar water areas in which bedload is at least temporarily et al., 2006). deposited—in the absence of the obstructions cre- Changes in cross-sectional geometry and trans- ated by instream wood, the bedload would remain port capacity (Constantine et al., 2003; Rengers and in transport (Montgomery et al., 2003b); Wohl, 2007) and river engineering (Surian, 2002) r enhancing overbank flooding and floodplain het- can also create substantial variation in downstream erogeneity by creating in-channel obstructions grain-sizetrends.Eachofthesestudiesindiverse that result in local aggradation and decreased flow environments indicates minimal downstream fin- conveyance (Jeffries et al., 2003); ing, or disruption or reversal of system-wide fining r promoting bar growth and lateral channel move- trends, along reaches in which the channel is later- ment (O’Connor et al., 2003); and ally confined (and, typically, steeper). r initiating multi-thread morphology by enhanc- In addition to decreases in average grain size ing magnitude and duration of overbank flows downstream, bed material also becomes better that can then erode secondary channels across the sorted and individual particles become more floodplain (Wohl, 2011c; Collins et al., 2012). rounded with distance from sediment source areas (Knighton, 1998). As with downstream fining, the Themagnitudeandrelativeimportanceofthese distance over which these changes occur reflects diverse effects vary downstream, as do the mech- clast erodibility and transport. anisms of wood recruitment, transport, and stor- age (Gurnell, 2013). This can be illustrated using the relations proposed by Benda and Sias (2003) 5.6.2 Instream wood for wood dynamics within a channel segment of length x As noted in previous sections, wood within a chan- [ ] nel has numerous geomorphic and ecological effects Q Q ΔS = L − L + i − 0 − D Δt (5.16) (Wohl, 2010; Gurnell, 2013) (Figure S5.13), includ- c i 0 Δx Δx ing: where ΔS is change in storage within the reach over r c increasing flow resistance and decreasing bank time interval t; Li is lateral wood recruitment into erodibility (Curran and Wohl, 2003; Manners the channel; L0 is loss of wood to overbank deposi- et al., 2007); tion during floods and abandonment of jams; Qi is r deflecting flow toward the channel boundaries fluvial transport of wood into the reach and Q0 is flu- and thus accentuating local boundary erosion vial transport out of the reach; and D is in situ decay. (Hassan and Woodsmith, 2004); Lateral wood recruitment reflects multiple forms of r increasing channel width by deflecting flow supply toward the banks (Nakamura and Swanson, 1993); r facilitating localized storage of sediment and Li = Im + If + Ibe + Is + Ie (5.17) organic matter by creating obstacles to flow and

associated zones of flow separation (Andreoli where Im is chronic forest mortality, If is toppling et al., 2007); of trees following mass mortality such as caused by r changing the dimensions of bedforms such as fire or windstorms, Ibe is inputs from bank erosion, steps and pools—large instream wood pieces help Is is wood from hillslope mass movements, and Ie to stabilize steps, for example, creating taller and is exhumation of buried wood from the floodplain more widely spaced steps than otherwise form in (Benda and Sias, 2003) (Figure S5.14). Equation 5.16 162 Rivers in the Landscape is a wood budget, analogous to a sediment budget Most pieces move by floating, although water- (Section 4.6). logged or otherwise particularly dense pieces can

In most regions, Sc decreases downstream. Wood move in contact with the streambed. Transport dis- load, typically quantified as volume of wood per tance increases with flow depth and duration above a unit area or unit length of channel, decreases down- floating threshold (Haga et al., 2002). Transport dis- stream as transport capacity increases (Lienkaem- tance also increases as the spatial density of obstacles per and Swanson, 1987; Bilby and Ward, 1989; such as large boulders or stationary wood declines Hassan et al., 2005). Simple ratios of wood piece (Bocchiola et al., 2008). Wood rating curves that length to average channel width and wood diame- relate wood transport to discharge can be developed, tertoflowdepthindicaterelativetransportcapac- analogous to sediment rating curves (Section 3.2.4), ity.Woodmobilityincreasesastheaveragepiece if records of wood transport can be developed using becomes shorter than channel width and as wood techniques such as video monitoring (MacVicar and diameter becomes a progressively smaller fraction of Piegay,´ 2012). flow depth (Lienkaemper and Swanson, 1987; Mar- Borrowing the concept of alternative stable states tin and Benda, 2001). from ecology, Wohl and Beckman (2014), proposed Therelativeimportanceofdifferentmecha- that rivers in forested environments tend to be wood nisms of lateral inputs also varies. Individual trees rich.Althoughrecruitmentofwoodtotheriverfluc- falling into the channel as a result of chronic for- tuates through time in response to disturbances such est mortality can be important in small channels, as forest fires, debris flows, or blowdowns, wood but become progressively less important as chan- decay is sufficiently slow in most rivers of the tem- nel size increases. Bank erosion and floodplain perate zone that at least some wood is always present exhumation become more important downstream in the stream. This wood helps to trap and retain as bank erodibility and the extent of floodplains newly recruited wood. If instream wood is removed increase. Hillslope mass movements add progres- or recruitment is eliminated for a period of time by sively less wood downstream as channels become clear-cutting, the river can enter a wood-poor condi- buffered from hillslopes by wide valley bottoms tion in which the absence of instream wood reduces and as topographic relief decreases. Lateral losses of the likelihood that any subsequently recruited wood wood also increase downstream as overbank pro- will be retained in the river rather than transported cesses and lateral channel movement grow more downstream. common. Ecologists use alternative stable states to describe Drag and buoyancy are the main forces mobi- a scenario in which an ecosystem can exist under lizing wood (Alonso, 2004). Wood entrainment is multiplestates,orsetsofuniquephysicalandbio- influenced by factors such as logical conditions. Alternative states are stable over ecologically relevant time spans, but ecosystems can r piece angle relative to flow direction—pieces par- transition from one stable state to another when per- allel to flow tend to be more stable (Braudrick and turbed sufficiently to cross a threshold. Collins et al. Grant, 2000); (2012) applied the concept of alternative stable states r presence of a root wad, which tends to decrease to floodplains along forested rivers. entrainment; Riversarelikelytotransitionfrombeingtrans- r wood density, which partly governs whether a log port limited with respect to wood in headwaters, rolls, slides, or floats; to supply limited in larger channels (Marcus et al., r piece length and diameter (Braudrick and Grant, 2002). Jams resulting from local decreases in wood 2000); and transport increase downstream, whereas jams pri- r inclusion in a jam, which tends to stabilize indi- marily forming around stationary wood pieces are vidual pieces and decrease mobility (Wohl and moreimportantinheadwaters(AbbeandMont- Goode, 2008). gomery, 2003; Wohl, 2013b) (Figure S5.15). The Chapter 5 Channel forms 163 size and number of logjams are likely to be great- 5.7 Summary est in middle portions of river networks (Wohl and Jaeger, 2009). At the mid-basin transition from Water and sediment are the drivers of channel transport- to supply-limited conditions for wood, form. We study hydraulics and sediment dynamics sufficient transport capacity exists to mobilize sub- to understand channel form—to predict the chan- stantial volumes of wood, but local decreases in nel form likely to result from specific engineering transport capacity associated with features such as manipulations such as altered flow regime down- mid-channel bars, zones of flow expansion, or chan- stream from a dam or stabilization of banks through nel bends (Braudrick and Grant, 2001) can trap large an urban area, and to interpret the environment that volumes of wood. has resulted in an existing natural channel form. The geomorphic effects of individual wood pieces The fundamental questions in understanding chan- decrease downstream as each piece interacts with nel form through time and space are (1) What mag- a progressively smaller proportion of the channel nitude and frequency of flow are most important boundaries and cross-sectional flow, but aggregate in creating and maintaining channel form? and, effects of wood can be substantial in large chan- (2) Through what processes, and how frequently, nels. Historical accounts exist from diverse forested do changes in channel form occur? The relations environments of enormous volumes of wood con- between water, sediment, and channel form are not centrated in portions of large alluvial rivers (Triska, a simple, one-way process. Channel form responds 1984). These log rafts and congested transport—when to discharges of water and sediment, but form also numerous logs move together as a single mass, influences the distribution of hydraulic variables and observed during flume studies (Braudrick et al., the dynamics of sediment in the channel. 1997)—significantly influenced channel and flood- Channel forms—from cross-sectional geometry plain dynamics along many forested rivers prior to to longitudinal profiles—represent the intersection extensive modification by human activities such as of physics and history. The physical balance between timber harvest and snagging (removal of instream hydraulic driving forces and substrate erodibility wood) (Wohl, 2013b). The volume of instream governs the creation and maintenance of river form. wood within most rivers flowing through forested Butthehistoryofpastchanneladjustmentsandof drainage basins has decreased so substantially as a larger factors as diverse as tectonics, climate, and result of human activities that our perceptions of land use, which influence driving forces and resis- “natural” wood loads, and of the geomorphic impor- tance, constrain the manner in which the balance is tanceofinstreamwood,aredistorted(Chinetal., expressed. A river is a physical system with a history. 2008).

Chapter 6 Extra-channel environments

Extra-channel environments refer to fluvial erosional designate floodplains based on average recurrence and depositional processes and the resulting fluvial interval of flooding, as in the 10-year floodplain or landforms created or now found outside of active 100-year floodplain. Nanson and Croke (1992) refer channels. These processes and landforms are con- to such surfaces as the hydraulic floodplain,because trasted with those within the boundaries of the the designation does not imply anything about geo- active channel (Chapters 3, 4, and 5). This is an morphic history or fluvial influence. Geomorpholo- arbitrary distinction because of the close coupling gists historically referred to floodplains as those sur- between active channels and marginal areas of flu- facesthatarefloodedatleastonceevery2years, vial erosion and deposition, but the distinction with the assumption that such surfaces are com- servestoemphasizesomeoftheuniqueaspectsof posed largely of fluvial sediments deposited under floodplains, terraces, alluvial fans, deltas, and estu- the current flow regime, rather than relict sediments aries.Allofthesefeaturescanbelaterallyandlon- deposited under very different conditions. Nanson gitudinally extensive and persistent landforms. As and Croke (1992) describe this as a genetic flood- terrestrial–aquatic interfaces, they can also be chem- plain. Even a genetic floodplain can be flooded less ically complex and biologically diverse and produc- frequently than approximately once every 2 years, tive environments that are strongly influenced by however, depending on the hydroclimatic regime. biotic activities. Globally, these riverine environ- Analyzing data from 28 gaged sites on rivers in the ments are also heavily altered by human activities, western United States, Williams (1978a) found that with consequences for hazards and for ecological the most frequent recurrence interval for floodplain sustainability. inundation was about 1.5 years, but recurrence inter- valsvariedfrom1to32years.Floodplainscanform along river courses in valley bottoms, on alluvial 6.1 Floodplains fans, and on deltas (Bridge, 2003). Floodplainscanbeexaminedinmanycontexts. Floodplains arelow-reliefsedimentarysurfacesadja- They exert an important influence on lateral and centtotheactivechannelthatareconstructedbyflu- longitudinal connectivity within a drainage net- vial processes and inundated frequently. Floodplain work,bufferinghillslopeinputs,andcommonly relief is typically ∼0.1–0.5 times the bankfull depth slowing the downstream movement of water, of the river (Dunne and Aalto, 2013). Engineers sediment, nutrients, contaminants, and organisms.

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

165 166 Rivers in the Landscape

Floodplains thus serve as reservoirs, in the sense of arrives, and by clear water that appears distinctly at least temporarily storing diverse materials. Stor- different from the turbid flood waters of the main agetimecanrangefromhourstomonthsforflood channel. waters, and up to thousands of years for sediment. An example of the influence of non-river water on Floodplains act as a safety valve during large flooding comes from the Altamaha River of Georgia, floods. As flood water spills over the channel USA, where measurements during several floods banks and across the floodplain, the water typically revealed that the floodplain was inundated by water encounters greater flow resistance and moves more carrying virtually no sediment while the river was slowly, facilitating sediment deposition. Slower still rising but below bankfull level (Mertes, 1997). overbank flow attenuates the flood, creating a peak Deeppondingonthefloodplainpriortooverbank discharge of lower magnitude and longer duration. flow limited the ability of overbank flow to spread Turbulence and flow separation between the over- laterally across the floodplain once the river over- bank and channelized flow dissipate energy and topped its banks. On this sinuous river, overbank reduce channel-bed and bank erosion. flow entered the floodplain at the inflection point Floodplains typically enhance the storage of sed- of river bends, flowed obliquely onto the floodplain, iment and solutes over diverse time spans that and retained a significant down-valley component. depend on factors such as frequency of overbank The river water moved down the floodplain in a rel- flow and width of the floodplain relative to rate atively straight line between river bends and covered of lateral channel migration. Sediment and solutes onlyanarrowzone(∼2–4 km) on a floodplain up to entering the floodplain can come from upland envi- 10 km wide (Mertes, 1997). ronments, upstream or tributary portions of the The perirheic zone referstothezoneofmix- river network or, particularly in the case of solutes, ing surrounding the flowing river water (Mertes, from groundwater sources. 2000). Mixing between flooding river water and Floodplains also tend to be biologically rich habi- locally derived water is governed by characteris- tats. Many aquatic organisms use the floodplain dur- tics such as the spatial density of floodplain chan- ing periods of inundation for breeding, nursery, and nels: high density limits mixing because the flood- feeding areas. The flush of nutrients that returns plain channels contain flood waters leaving the main to the channel as flood waters recede helps to sup- channel. port aquatic communities in the channel (see Sec- The Amazon River in some respects exempli- tion 3.2.4). The diversity of floodplain habitat and fies floodplain dynamics. As the largest river in the moisture levels commonly supports more diverse world,theAmazonhasenormousfloodplains.Asa plant communities than are found in the adjacent largely unregulated river without human-built lev- uplands. ees, the Amazon still experiences a natural flooding The water inundating a floodplain does not nec- regime and high channel–floodplain connectivity. essarily come exclusively from overbank flow (Fig- Limited mixing occurs in the upper reaches of the ure S6.1). Flooding can occur before natural lev- Amazon River, which have a high density of flood- ees overtop because flood waters can enter a flood- plain channels, whereas the middle reaches have plainfromrisinggroundwater,overlandflowfrom enhanced mixing from overbank flows across lev- adjacent slopes, and tributaries, as well as overbank eesalongthemainchannel.Thelateralextentofthe flow from the main channel. Mertes (2000) dis- incursion of river water into the floodplain varies tinguished inundation patterns on dry floodplains, from about 5 km in upstream reaches of the Ama- where most water overflows channel banks, from zon to 5–15 km in middle reaches. Mixing is also saturated floodplains in which rising groundwa- limited in the lower reaches of the Amazon, where ter can contribute significantly. Flooding associated large floodplain lakes have sufficient hydraulic head with rising groundwater is evidenced by increas- to prevent river water from entering large portions ing depth of standing water before the flood wave of the floodplain (Mertes, 2000). Chapter 6 Extra-channel environments 167

Floodplains along large, lowland rivers such ees, or channelization fail for some reason. In addi- as the Amazon can be extremely extensive tion to the obvious hazard of flooding, lateral chan- (Figure S6.2). Floodplains extend for approximately nel migration and avulsion can create substantial 4000 km along the Amazon or the Nile River, and change, creating hazards for people and structures. canextendupto100kmacross.DunneandAalto Relatively few large rivers with largely unaltered (2013) distinguished the modern, most recently flood regimes and floodplains remain. Among these active channel belt landforms on a floodplain from are the Amazon and the Congo, the two largest the wider floodplain that can be of Holocene or riversintheworld,andthegreatnorthernrivers even Pleistocene age. The channel belt along a the Lena and Mackenzie, which are rated as mod- meandering or braided river can be up to several erately affected by dams, as well as the Yukon, which times the width of the channel bends, and such is unaffected (Nilsson et al., 2005). The commonal- bends can be up to 100 times the bankfull channel ityamongtheseriversisthattheyallhavelowaver- width. Channel belts on multi-thread large rivers age population density within the drainage basin. can be even wider. In contrast, large rivers such as the Mississippi, Active floodplains were historically much more Danube, Nile, Murray–Darling, and Yenisey, which extensive. People have systematically reduced the historically had large floods and extensive flood- extent and duration of overbank flow by plains,havelostfloodpeaksasaresultofflowreg- ulation, and lost floodplains to disconnection from r regulating the flow and removing flood peaks to the river and human settlement and land use within the extent possible; the floodplain. r constructing levees or transportation corridors (Blanton and Marcus, 2009) that physically block access to the floodplain; and 6.1.1 Depositional processes and r enlarging channels so that higher flows remain floodplain stratigraphy within the channel. Floodplain form reflects a history of both erosion Floodplains have been particularly targeted by and deposition, but the majority of floodplains are these activities because extensive and prolonged predominantly depositional environments that store overbank flooding is regarded as a nuisance or a haz- large volumes of sediment for varying lengths of ard that limits human access to the flooded areas time. Floodplain deposits can be categorized as ver- or travel across these areas, and because floodplains tical accretion deposits that form when sediment have long been regarded as highly desirable loca- settles from suspension in the lower-energy envi- tions for human communities as a result of fertile ronment of the floodplain or as lateral accretion soils and easy access to water supply and water- deposits. Vertical accretion deposits include levee basedtransport.Manyoftheworld’sgreatcitiesare and backswamp or flood basin sediments. largelyorwhollywithinfloodplains,includingLon- Sedimentation rates typically decrease across the don (capital of the United Kingdom), Seoul (capi- floodplain with distance from the channel (Swanson talofSouthKorea),Paris(capitalofFrance),Tokyo et al., 2008), and active sedimentation extends no (capital of Japan), Dhaka (capital of Bangladesh), farther than a few kilometers from the channel even Caracas (capital of Venezuela), and Kinshasa (capital on very large floodplains (Dunne and Aalto, 2013). of the Democratic Republic of Congo). Obviously, Rates of vertical accretion reported for diverse rivers governments in every continent think they can con- range from 0 to nearly 6000 mm per year (Knighton, trol flooding sufficiently to continue to base them- 1998), but are typically at most a few centimeters per selves within a floodplain. The location of major year. Sedimentation is commonly extremely slow cities within floodplains can result in tremendous at distances of more than a kilometer from the damageandlossoflifewhenflowregulation,lev- main channel, and these lower elevation portions of 168 Rivers in the Landscape

(a)

Sand Terrace Silt-loamy sand Point bar Silty clay

Sand bar Ridge Chute Channel fill Swale Splay

Levee Meander-loopallounit Direction of point- bar migra Overbank tion deposit

Lateral accretion surface

(b)

Figure 6.1 (a) Idealized diagram of a large, sinuous river floodplain, showing types of lateral and vertical accretion deposits discussed in text. (From Holbrook et al., 2006, Figure 2, and Dunne and Aalto, 2013, Figure 21.) (b) Levee formed along a distributary channel of the Slave River in northern Canada. The Slave River forms an extensive delta as it enters Great Slave Lake and this photograph was taken several kilometers upstream from the distributary channel’s entry into the lake. Arrow indicates flow direction in channel and dashed line indicates approximate shape of levee, which has been truncated in the foreground in order to create a parking area and boat access to the channel. the floodplain can be covered by lakes or swamps that form where the coarsest suspended sediment (Dunne and Aalto, 2013). In regions wet enough to is deposited as velocity drops in the mixing zone support at least partial vegetation cover, floodplain between channelized and overbank flows (Fig- deposits commonly include organic matter trans- ure 6.1). Sediment accumulates faster along the levee ported by flood waters and varying in size from fine than it accumulates as flow spreads farther from particulate organic matter (∼0.5–1 mm in diame- the channel. This results in a continuous berm that ter) to large tree trunks. Floodplain sediments are gradually declines in elevation with distance from also typically bioturbated by plant roots and animals the channel (Dunne and Aalto, 2013) and can be (Bridge, 2003). up to four channel widths across (Bridge, 2003). Natural levees, as distinct from levees constructed Levee height commonly decreases downstream as by humans, are discontinuous, linear features imme- the grain size of sediments in suspension decreases diately adjacent and parallel to the channel banks (Dunne and Aalto, 2013). Chapter 6 Extra-channel environments 169

Backswamp or flood basin deposits are typically bulkoffloodplainsedimentingravel-bedriversthat finer-grained sediment—wash load—settling from are laterally mobile. suspension across the portions of the floodplain Well-developed bars can concentrate flow along between the levees and the valley walls or the the outer margins of bends and accelerate bank ero- outer limit of flooding. Flood basins are typically sion and bend growth (Legleiter et al., 2011). Bend much longer than wide, and segmented by allu- migration results in lower shear stress and sediment vial ridges and crevasse splays (Bridge, 2003). Sed- transport across the bar, which facilitates growth of iment deposited in this environment can fill aban- vegetation and accumulation of finer sediment from doned or secondary channels (known as sloughs, suspension, raising the bar surface and blocking billabongs, and other names in different parts of the minor channels that might have developed across world), or can be deposited more uniformly across the bar (Braudrick et al., 2009). As bars are incorpo- the floodplain flats. Very large depressions on the rated into the floodplain, sequences of arcuate bars flood basin can contain lacustrine (lake) deposits with intervening, poorly drained depressions create and deltas. Ephemeral lakes in arid regions can scroll-bar topography that can persist for centuries as contain evaporite minerals. Backswamp deposits differences in soil grain size and moisture support are typically rich in organic matter, which can be different plant communities (Figure S6.3). autochthonous (derived from plants growing at the The lateral accretion of bars in a braided river site)orallochthonous(transportedfromuplandor creates a braid plain of channel-margin sediments. upstream sites and deposited in the backswamp This plain, which is analogous to the channel belt environment). Flood basins form the lower elevation of a sinuous river, has an irregular surface and thin, areas of floodplains. discontinuous, vertically accreted sedimentary cov- Alluvial ridges are slightly higher areas formed of ers in depressions and former channels (Dunne and active and abandoned channels and bars. The ridges Aalto, 2013). include accretionary features, levees, and crevasse The juxtaposition of different types of deposits channels and splays (Bridge, 2003). creates vertical relief on floodplains. A floodplain Splay or crevasse splay deposits occur when a levee created primarily by lateral accretion is some- is breached by piping or overtopping. The breach- times known as a flat floodplain (Butzer, 1976). ing process creates a narrow cut through the levee A flat floodplain consists mainly of bed-material that contains higher velocity flow able to transport deposits—either bedload in gravel-bed rivers, or coarser suspended sediment into the backswamp suspendible sandy or silty bed material in large, area. Sand splays formed in this manner can be lowland rivers (Dunne and Aalto, 2013). This type quite extensive. The 1993 flood on the Mississippi of floodplain contrasts with what Butzer (1976) River created sand splay deposits greater than 60 cm called a convex floodplain along rivers dominated thick that extended hundreds of meters in length by vertical accretion. In a convex floodplain, the and breadth (Jacobson and Oberg, 1997). thickestandcoarsestdepositsclosetothechannel The other major category of floodplain deposits create a slightly higher floodplain elevation adja- is associated with lateral accretion. Lateral accre- cent to the channel than at the margins of the tion deposits result from sediment deposition along floodplain. themarginsoftheactivechannel:thissediment Floodplainscanalsoincludenon-fluvialsedi- can become part of the floodplain as the channel ments intermixed with the fluvial material. Eolian migrates laterally. Lateral accretion deposits include sediments can be interbedded with fluvial flood- islandsandbars,aswellaslagdepositsofthe plain deposits in arid regions as a result of eolian channel-bed sediment. The classic example is point input from adjacent regions, or reworking of unveg- bar sediment incorporated into the floodplain as a etated floodplain sediments. Colluvial sediments meander bend migrates toward the outer, can be deposited on the outer margins of a flood- of the bend. Lateral accretion deposits can form the plain with steep valley walls. 170 Rivers in the Landscape

Floodplains thus form from a combination of translation times for passage of a flood wave along within-channel and overbank deposits, and the rela- large rivers create gradually rising and falling hydro- tive importance of these two basic depositional envi- graphs. Overbank flow can last from many weeks to ronments varies in relation to factors such as rate as long as 6 months (Richey et al., 1989). The com- of lateral channel movement and frequency, extent, bination of long duration overbank flows and abun- duration, and sediment concentration of overbank dant fine-grained particles with low settling veloc- flows. In an influential paper on floodplain deposi- ity can result in large transfers of sediment into the tion, Wolman and Leopold (1957) proposed that lat- floodplain (Dunne and Aalto, 2013). eral accretion deposits dominated most floodplains, Recent work emphasizes biotic influences on accounting for up to 90% of the total sediment floodplain form and process. Riparian vegetation deposited. Subsequent research, however, suggests growing on bars can help to stabilize and enlarge the that the situation is more complicated and varies bars, facilitating lateral accretion to the floodplain greatly among different rivers. (Gurnell and Petts, 2006). Formation of channel- The timescales of vertical and lateral floodplain spanning obstructions such as beaver dams and log- accretion can differ, and rapid lateral channel migra- jamscanleadtolocalaggradationandlossofcon- tion can remove gradually accumulated vertical veyance that enhances overbank flow. This can in accretion deposits. The relative importance of lat- turn lead to the formation of multi-thread chan- eral and vertical accretion can alter through time in nels, as well as enhancing vertical accretion (Collins response to and Montgomery, 2002; Sear et al., 2010; Westbrook et al., 2011; Wohl, 2011c; Polvi and Wohl, 2012). r changes in the volume and grain size of sedi- Collins et al. (2012) conceptualized these influences ment available during floods—finer-grained sed- in terms of a floodplain large-wood cycle in which iment carried in suspension is more likely associ- logjams create alluvial patches and protect them ated with vertical accretion, and coarser material from erosion, providing sites for trees to mature movingincontactwiththebedismorelikelyasso- over hundreds of years in valley bottoms where ciated with lateral accretion; floodplain turnover is typically much faster. These r transport energy of the flood waters—higher patchesofold-growthforestinturnprovidelarge transport energy can result in overbank velocities wood to rivers that can form persistent logjams, thus too high to permit settling from suspension and perpetuating the cycle. Floodplains influenced by vertical accretion; and thisprocesshavemulti-threadchannelsseparatedby r geometry of depositional sites—floodplains with surfaces of different age and elevation, as well as a numerous depressions that enhance locally low mosaic of forest stands of varying age (Figure 6.2). velocityandsettlingfromsuspensioncanpromote Episodically forming channel-spanning logjams vertical accretion. canalsofacilitategreaterdepthanddurationofover- bank flow and floodplain aggradation. Oswald and Therelativeimportanceofverticalandlateral Wohl (2008) described how jokulhlaups¨ (outburst accretion can also vary downstream. Along the floods)fromaheadwaterglacierintheWindRiver Brazilian portion of the Amazon River, floodplain MountainsofWyoming,USA,createpeakflows deposition in the upstream reaches results from much greater than the typical snowmelt peak flow. lateral accretion in floodplain channels. Overbank Higher peak flows enhance bank erosion and wood deposition dominates downstream portions of the recruitment, creating numerous logjams. Jokulh-¨ floodplain (Mertes et al., 1996). laups occurring every few decades create logjams As a relatively well-studied and still-natural river, thatbreakupoveraperiodofadecade,butthe the Amazon River provides insight into some of the overbankaggradationthatoccurswhilethelogjams aspects of floodplain sedimentation characteristic of are present can persist over time periods of 102–103 verylargerivers.Thehugedrainageareaandlong years (Figure S6.4). Chapter 6 Extra-channel environments 171

Meandering Abandoned Although the conceptual model of a flood- Forest channel/ chronosequence oxbow plain large-wood cycle was developed specifically Channel for moderate-sized rivers in the Pacific Northwest region of the United States, it is important to remem- Channel migration ber that forested floodplains were once much more common. In his classic 1837 textbook, Principles of Buried Geology,CharlesLyellcommentedontheenormous Anastomosing wood Mature forest jam log rafts characteristic of large American rivers in In-channel patches “hard Secondary thelowerMississippiandAtchafalayaofthesouth- wood jam point” channel eastern United States. Historical descriptions indi- cate that these log rafts occurred on rivers across the United States, including the northeastern conifer Avulsion Migration forests, the temperate deciduous forests of the north- around between hardpoints hardpoints central region, and the humid temperate rainforest of the Pacific Northwest (Wohl, 2013b). Similar log Braided rafts may have occurred historically on continents Ephemeral patches of pioneer vegetation besides North America. Lyell specifically mentioned how log rafts in the southeastern United States, which could stretch for >200 km along a river (Phillips and Park, 2009), Frequently shifting channels enhanced overbank flooding. A sedimentary facies Figure 6.2 Cross-sectional views of three idealized suite attributed to low-energy, organic-rich rivers models of floodplain landforms and forest development with multiple, anabranching or anastomosing chan- for river valleys of moderate gradient in the US Pacific nels and stable alluvial islands first appeared during Northwest. The meandering river model is dominated the Carboniferous, when tree-like plants could cre- by lateral migration of meanders and associated mean- ate channel blockage by logjams (Davies and Gib- der cutoffs. These processes create scroll-bar topography, ling, 2011). In other words, instream wood has been oxbow ponds, and sloughs. Plant germination on newly accreted sediments, and continued lateral migration and influencing flow resistance, channel conveyance, forest succession, create chronosequences of surface and and overbank flows for a very long time, and these forest ages. Maximum forest age is limited by the rate at influences extended across the forested regions of which the river meanders across the floodplain. In the the world. Wohl (2013b) proposed that widespread anastomosing river model, stable alluvial deposits asso- removal of instream wood during the past two cen- ciated with wood jams can resist lateral channel erosion turies resulted in a fundamental change in flood- for hundreds of years, providing sites for tree germina- plain dynamics because of the associated decrease tion. Maximum forest age is limited by the stability of in overbank flows and sedimentation, as well as these deposits. Channels avulse around stable patches decreased avulsion and formation of multi-thread and migrate as in the meandering river model, creating planform (Figure S6.5). a mosaic of multiple channels and floodplain elevations, Beavers were also once extremely abundant and and greater diversity in forest patch age. Braided rivers ubiquitous in forested environments of North Amer- are dominated by multiple, frequently shifting channels. Mature forest vegetation is limited to the channel mar- ica (Castor canadensis)andEurope(Castor fiber). gins, although ephemeral patches of pioneer vegetation The channel-spanning dams that these animals built can grow on braid bars within the channel network. (From on smaller rivers certainly influenced magnitude Collins et al., 2012, Figure 3.) and duration of overbank flows, and thus floodplain dynamics. The balance through time among vertical accre- tion, lateral accretion, and lateral channel migration 172 Rivers in the Landscape is reflected in floodplain stratigraphy. Sinuous chan- blockage of a channel by wood or ice, for example, nels typically create a characteristic fining-upward can create sufficiently high water levels to promote sequence (Figure S6.6). As a meander bend migrates avulsion in the absence of flooding. The new chan- laterally, a geographic location that starts as a deeper nel belt follows the zone of maximum floodplain portion of the channel near the outside of the bend slope while migrating to the lowest part of the flood- transitions to a point bar environment and even- plain. This process can occur over a period of years tually to a floodplain. The stratigraphic sequence to centuries (Bridge, 2003), or during a single very at that location thus fines upward from relatively large flood. coarse channel-lag deposits, to slightly finer-grained The frequency of avulsion increases with increas- point bar deposits, which are in turn capped by even ing deposition rate. Deposition enhances topo- finer overbank deposits. Floodplains created by sin- graphic relief and cross-valley slope relative to uous channels also commonly have predictable lat- down-valley slope of the channel belt. Frequency of eral stratigraphy associated with the migration of avulsion also increases with base level rise, which numerous meander bends. causes aggradation and growth of alluvial ridges Braided channels or sinuous channels subject to (Bridge, 2003). Numerical simulations, and the age frequent, large avulsions are likely to have much distribution of floodplain vegetation, suggest that less predictable stratigraphy. Floodplains along these mid-sized sinuous alluvial channels with sand and rivers can be described as a three-dimensional gravel beds preferentially reoccupy floodplain loca- (3D) mosaic in which different grain sizes repre- tions that were recently abandoned, resulting in a senting diverse depositional environments can be decreasing probability of floodplain reoccupation abruptly juxtaposed laterally and vertically (Fig- with time since abandonment (Konrad, 2012). ure S6.7). Arid-region braided channels, for exam- ple, repeatedly avulse across a broad floodplain. This led Graf (2001) to suggest that historical infor- 6.1.2 Erosional processes and mation about channel location—including flood- floodplain turnover times plain stratigraphy—be used to develop a locational probability map that provides a statistical view Floodplain sediment can be removed by localized of the river’s most likely location at any point erosion during extreme floods. Examples of local- in time. ized erosion include the longitudinal grooves, scour Lateralaccretionofbarsdominatesfloodplain marks, stripped floodplains, chutes, and anasto- stratigraphy of braided rivers, creating an irregular mosing erosion channels described by Miller and surface capped by a thin, discontinuous cover of Parkinson (1993) after a 1985 flood in West Vir- finer, vertical accretion sediment accumulated in ginia, USA. Longitudinal grooves are elongate linear former channels (Dunne and Aalto, 2013). Levees grooves parallel or subparallel to the local direction along braided rivers tend to be less tall as a result of flood flow that extend down the valley floor tens of lateral channel movement. These levees are also to hundreds of meters, with depths and widths from easily eroded, so that sand splays are an important a few centimeters to greater than 1 m. Miller and mechanism of floodplain sedimentation. Relatively Parkinson (1993) interpreted these features as the steep gradients and sparse vegetation facilitate early stages of floodplain erosion, because the rest of chute cutoffs and avulsions that can be rapidly filled thevalleybottomwaslargelyunaltered.Scourmarks if suspended sediment concentrations are high varyinshapefromsmallcircularorellipticalpitsto (Dunne and Aalto, 2013). elongate parabolic or spindle-shaped and irregular Although avulsion is particularly characteristic of marksthatappeartobeassociatedwithflowsepara- braided rivers, this process can occur in any type tion and formation of vortices. Individual marks can of river. Avulsion referstotheshiftingofachannel belessthan0.3mdeepand1mindiameter,butthe or channel belt across the floodplain. Avulsion typ- largest scour marks reach tens to hundreds of meters ically occurs during floods, although downstream long and tens of meters wide. Stripped floodplains Chapter 6 Extra-channel environments 173 occurwherefloodwatersremoveaveneeroffiner Along the Brazilian portion of the Amazon River, overbank deposits to reveal underlying cobble and sediment exchanges between the channel and flood- boulder sediments across most or all of the flood- plain in each direction exceed the annual flux of sed- plain. Concentrated flow on the floodplain produces iment out of the river at Obidos´ (∼1200 Mt/year) a well-defined chute channel comparable in dimen- in the lower portion of the drainage (Dunne et al., sions to the pre-flood river channel, rather than 1998). The annual sediment supply entering the stripping the entire floodplain surface. Anastomos- channel from bank erosion was estimated at 1.3× the ing erosion channels across the floodplain are asso- Obidos´ flux (1570 Mt/year). The annual volume of ciated with incomplete channel widening that leaves sediment transferred to the floodplain via channel- remnant islands in the expanded channel (Miller ized flow (460 Mt/year) and diffuse overbank flow and Parkinson, 1993). Removal of floodplain veg- (1230 Mt/year) totaled 1.7× the Obidos´ flux. These etation by the flood strongly influences the ability estimates indicate a net accumulation of 200 Mt/year of flood waters to create such localized erosional on the floodplain (Dunne et al., 1998). Quantitative features. studies of floodplain dynamics along the Strickland Gradual change in floodplain morphology can River of Papua New Guinea similarly indicate that facilitate abrupt, localized erosion during floods, ∼50% of the total sediment load is recycled between a phenomenon Nanson (1986) described as catas- the channel and floodplain via cut bank erosion and trophic stripping. Studying partly confined flood- point bar deposition (Aalto et al., 2008). In gravel- plains along high-energy coastal rivers in New South bed channels such as the Fraser River of Canada, Wales, Australia, Nanson described gradual over- sediment inputs from bank erosion can exceed net bank deposition over periods of 102–103 years that downstream sediment flux by an order of magni- builds a floodplain of fine-textured alluvium with tude, partly because bank sediment is largely trans- large levees. Growth of the levees and adjacent flood- ferred to adjacent bars, creating short sediment step plain surfaces progressively restricts peak flows to lengths (Ham, 2005). the main channel and floodplain back-channels, The continual exchange between sediment within concentrating erosional energy until widespread the channel and sediment stored in the floodplain scour of the channel boundaries and floodplain results in floodplain turnover,orerosionofexist- occurs. This catastrophic stripping is discontinuous ing floodplain sediment and deposition of new along a channel (Nanson, 1986). sediment (Figure S6.8). Turnover times typically Longitudinal variation in valley geometry increase downstream for two reasons. First, increas- andchannelgradientcanalsoresultinlocalized ing floodplain width downstream means that longer floodplain erosion at the reach scale. Comparing time spans are needed for lateral channel migra- theeffectsofwidespreadfloodinginthenorth- tion to completely cross the floodplain. Second, even eastern United States, Patton (1988) found that large floods are less likely to completely erode the sedimentation during large floods is important in floodplainthaninsmaller,laterallyconfinedpor- constructing floodplains within highland drainages. tions of the floodplain in upstream reaches (Patton, In these drainages, reaches of lateral confinement 1988). Mertes et al. (1996) estimated turnover times and steep gradient experienced erosion during of ∼1000 years in upstream portions of the Brazil- flooding. Less confined, lower-gradient reaches ian Amazon, and >4000 years in the downstream within the highland drainages experienced primar- portions of the river near Obidos.´ Based on cosmo- ily deposition. Lowland drainages, in contrast, were genic 26Al and 10Be, Wittmann et al. (2011) inferred notasstronglyinfluencedbytheflood,because that some of the sediment in transport within the magnitudes of flood erosion and deposition did Amazonchannelnetworkhasbeenstoredindistal, not exceed magnitudes during a succession of more deeply buried portions of floodplains for as long as moderate flows. 5millionyears. More continuous floodplain erosion occurs as a Because of the slow turnover times of sediment result of channel lateral migration and bank erosion. in many large floodplains, floodplain stratigraphy 174 Rivers in the Landscape can provide a variety of paleoenvironmental infor- erodibility, glacial history, sediment supply, network mation. The dimensions and stratigraphy of pale- structure, and biota (Wohl, 2010). Changes in lithol- ochannelscanbeusedtoinferpastflowregimes ogy and jointing can influence bedrock erodibility, (see Section 3.2.1). Pollen and fossils such as gas- so that a river alternates between wider, lower gradi- tropods in the sediments filling abandoned flood- entsegmentswherebedrockismorereadilyeroded plain lakes and channels can be used to infer past and a floodplain develops, and steep, narrow gorges climates. Archeological sites are also common in largely lacking floodplains where bedrock is more floodplains and can provide useful information on resistant (Ehlen and Wohl, 2002). Portions of a valley floodplain chronology and on local plants and ani- immediately upstream from a glacial moraine can be mals if food remains are present. widerandoflowergradient,allowingmoreexten- One of the implications of floodplain turnover sive floodplain development. Local point sources of is the potential for storage and subsequent release coarse sediment, such as debris flows or landslides, of contaminants. Several studies document prefer- can laterally constrict floodplains, as well as creat- ential floodplain storage of mining-related contam- ingatleastatemporarylocalbaselevelthatalters inants. Floodplain storage results from dispersal of channel form (Korup, 2013). Network structure, tailings during large floods caused by tailings dam as expressed in the arrangement and relative size failures (Marcus et al., 2001). Floodplain storage of tributaries entering a main channel, influences alsoreflectsthetendencyofmining-relatedcontam- water, sediment, and other fluxes (e.g., instream inants such as mercury to travel adsorbed to the fine wood) into the mainstem, and affects the form and sediment that accumulates in floodplains via verti- processes of the mainstem floodplain (Benda et al., cal accretion (Miller et al., 1999). Once deposited on 2004). Biota, including beavers that build dams and the floodplain, the contaminants can slowly leak into forests that allow sufficient wood recruitment to the river via bank erosion over a period of centuries support channel-spanning logjams, can create chan- (Macklin et al., 1997). nel segments of lower gradient and greater overbank flows (Westbrook et al., 2011; Polvi and Wohl, 2012). Exceptions to the general downstream trend of 6.1.3 Downstream trends in increasing floodplain width can also result from floodplain form and process structural controls or changes in substrate resis- tance. The Amazon River is entrenched where it The general pattern of floodplain form is a progres- crosses each of four structural arches, resulting in sive downstream increase in the longitudinal conti- restricted channel movement and narrower flood- nuity and the lateral extent of floodplains. Excep- plains through these segments (Mertes et al., 1996). tions can occur in low-relief terrains where even Similar effects have been described along the Mis- headwater channels have little lateral confinement, sissippi (Schumm et al., 1994) and other large allu- allowing wider, longitudinally continuous flood- vial rivers (Bridge, 2003). The very low gradients of plains to develop. Exceptions also occur in rivers large alluvial rivers allow these rivers to be modi- that cross mountain ranges. The Danube River of fied by even slow, small tectonic movements (Dunne Europe, for example, heads in mountainous regions, and Aalto, 2013), resulting in downstream changes but the middle and lower sections of the river’s in gradient, sediment transport, channel planform, course alternate between wide alluvial basins with and floodplain width and processes. multi-thread channels and extensive floodplains, Regional-tocontinental-scaletectonicsetting and narrowly confined bedrock gorges with little canalsoinfluencefloodplainextentonverylarge floodplain development (Wohl, 2011a). rivers. The largest rivers with long, low-gradient Valley and floodplain width within regions of reaches that facilitate extensive sedimentation highreliefcanalsoexhibitsubstantiallongitudinal and floodplain development drain the tectonically variation as a result of local changes in bedrock passive margins of continents (Potter, 1978). Rivers Chapter 6 Extra-channel environments 175 that head in orogens (belts of deformed rocks, eral channel migration, so vertical accretion and typically associated with higher topographic relief) infrequent channel avulsion dominate floodplain have greater sediment supplies than those that drain deposition. cratons (tectonically stable continental interiors, typically with lower topographic relief) (Milliman Because floodplains do not necessarily progres- and Meade, 1983). If these rivers flow away from sively increase in width downstream, and because theorogen,liketheAmazon,theyareabletocreate channel and floodplain characteristics are so closely extensive floodplains, whereas rivers that flow related, some river classifications emphasize the parallel to the orogen and receive sediment from degree of valley confinement and floodplain devel- tributaries along much of their length, as does the opment. The River Styles framework of Brierley and Orinoco, tend to have narrower and asymmetrical Fryirs (2005), for example, begins by differentiat- floodplains where tributary fans of coarse sediment ing valley setting as being confined, partly confined shift the main river away from the orogen (Dunne or unconfined. Similarly, designations of river geo- and Aalto, 2013). morphic spatial differentiation in mountainous set- tings emphasize the degree of valley confinement. Polvi et al. (2011) found that valley confinement 6.1.4 Classification of floodplains and elevation explained nearly 90% of the variabil- ity in field-delineated width of the riparian zone, as Nanson and Croke (1992) suggested a three-part defined based on plant species present, in headwater classification for floodplains, based on specific rivers of the Colorado Rocky Mountains. stream power and sediment texture as an influence Afloodplainsurfacecanremainactiveforthou- on substrate resistance. The intent of this classifica- sands of years along a river with stable base level tion is to distinguish floodplains with respect to pro- and relatively consistent water and sediment inputs. cesses of origin and resulting morphology. Along many rivers, however, the longevity of a par- 1. High-energy, non-cohesive floodplains with spe- ticular floodplain surface is limited because inci- cific stream power >300 W/m2 at bankfull are sion or aggradation reduces the lateral connectiv- disequilibrium landforms that partly or com- ity between the channel and the floodplain surface. pletely erode during infrequent extreme floods. When this occurs, a terrace can form. These floodplains typify steep upland portions of a drainage, in which bedrock or very coarse allu- 6.2 Terraces vium limits lateral channel migration. Relatively coarse vertical accretion deposits dominate these Terraces are relict channel–floodplain features that floodplains. have been isolated from contemporary river pro- 2. Medium-energy non-cohesive floodplains with cesses. Each terrace consists of a tread,theflatpor- bankfull specific stream power of 10–300 W/m2 tion that represents the former floodplain surface, are in dynamic equilibrium with annual to and a riser, the steep portion that separates the ter- decadal flow regime. Geomorphic change is lim- race from adjacent surfaces. ited during extreme floods because energy is dis- Terraces can be created by aggradation of the sipated by overbank flow. Lateral point bar accre- floodplaintoalevelnolongeraccessedbyrelatively tionandbraid-channelaccretiontendtodomi- frequent flows, or aggradation followed by incision. nate deposition on these floodplains. Terracescanalsobecreatedbyincisionofthechan- 3. Low-energy cohesive floodplains with bankfull nel. These various combinations can result from specific stream power <10 W/m2 form along lat- altered sediment supply or water flow caused by erally stable channels with low gradients. Energy changes in climate, land cover, or tectonics. Terraces is dissipated by overbank flow during extreme canalsoresultfromintrinsicprocessesassociated floods and cohesive bank sediments limit lat- withsedimenttransportthatisdiscontinuousin 176 Rivers in the Landscape time and space (Section 1.4). A change in base strath and fill terraces (Figure 6.3). Strath terraces level or water or sediment discharge can lead to have low-relief terrace treads formed in bedrock or aggradation or degradation and terrace formation. other cohesive materials such as glacial till, overlain Terraces are not necessarily present along a river, byaveneerofalluviumthinenoughtobemobilized but where they do exist, these landforms can be used throughoutitsentiredepthbytheriver.Existence to infer past longitudinal profiles of rivers, as well of a strath terrace implies a period of vertical sta- as paleoenvironmental conditions. Where terraces bility during which a river forms a relatively planar are present in a drainage basin, they typically do not bedrockvalleybottombylateralerosion,followed extend fully into the headwaters because of lack of by a period of vertical incision as transport capac- formation or subsequent erosion. ity increases beyond sediment supply. Strath terraces Terraces have been widely used by people because arelesslikelytoformasupliftandrateofincision they provide flat, in many cases stable surfaces, increase, because the river is cutting downward too which are close to rivers and suitable for agricul- fast to form a strath tread (Merritts et al., 1994). ture and cities. Terraces containing coarse-grained Strath terraces tend to be more extensive where channel-lag deposits are also mined for placer met- rivers flow over bedrock less resistant to weather- als and construction aggregate. ing and erosion as a result of lithology or jointing, and more poorly developed over more resistant rock 6.2.1 Terrace classifications (Montgomery, 2004; Wohl, 2008a). Fill terraces are alluvial sequences too thick to Numerous classifications are used in connection be mobilized throughout their depth by the river. withterraces.Someofthesearesimpleanddescrip- Because fill terraces can form more rapidly than tive. Paired terraces have equivalent surfaces on both straths, they do not necessarily imply a period of ver- sides of a river valley. Unpaired terraces do not match ticalstability.Theydoimplyaperiodofaggrada- acrossthevalleyandoccurinsituationssuchascon- tion, followed by incision. Alluvium in fill terraces is tinued incision while a sinuous river migrates later- commonly topped by 0.1–1 m of fine overbank sed- ally across the valley. iments and eolian deposits (Pazzaglia, 2013). Other terrace classifications focus on the domi- A relatively common scenario among fill terraces nant process creating the terrace. The names of ero- associated with high-latitude climatic fluctuations sional terraces and depositional terraces, for exam- during the Quaternary is illustrated by the exam- ple, are self-explanatory, although in practice it can pleoftheMaasRiverofnorthernEurope.Terrace be difficult to assign a primary cause of terrace for- alluvium aggraded along the Maas during periods of mation because both erosion and deposition are cold climate and glacial advance, whereas the river involved in the formation of many terraces. Simi- incised during periods of warm climate and glacial larly, tectonic terraces and climatic terraces can be dif- retreat (van den Berg and van Hoof, 2001). Alternat- ficult to distinguish in a region with active tectonic ing glacial and interglacial periods during the Qua- uplift or relative base level fall that has also under- ternary produced 21 paired and unpaired fill ter- gone climatically driven changes in water and sedi- races and ∼100 m of incision along the Maas (van ment yield over the period of terrace formation. den Berg and van Hoof, 2001). One terrace classification distinguishes the dura- tion of terrace formation. This classification dif- ferentiates event terraces associated with a single 6.2.2 Mechanisms of terrace extreme perturbation such as a very large flood or formation and preservation debris flow, and sustained terraces that represent a persistent change in water and sediment yield. Terraces can be interpreted as deviations from a One of the more straightforward classifications graded condition (Section 5.4). A graded river can focuses on terrace composition and differentiates aggradeinresponsetobaselevelriseorincisein Chapter 6 Extra-channel environments 177

(a) (c)

(b) (d)

(d)

Figure 6.3 Examples of river terraces. (a) Two small fill terraces (top outer edge of each terrace indicated bydashed white line) are visible in this view of a small creek in northern California, USA. Successive waves of aggradation and incision along this creek created the higher fill terrace just visible at the base of the forest, then the younger fill terrace about 1 m in elevation below the upper terrace, and finally resulted in lowering and widening of the channel, killing the riparian trees that left the stumps visible here. Flow in the channel is from left to right. (b) Aerial view of numerous fill terraces along a river in Nepal. (c) A strath terrace along the Mattole River in California, USA.The white material exposed in the lower portion of the cut bank is bedrock. An upward fining alluvial sequence of cobbles to loamy soil overlies the bedrock strath. Flow is from the foreground toward the rear in this view. The cut bank is approximately 6 m tall. (d) Distant view of very tall strath terrace along the Duke River in northwestern Canada. Strath surface is approximately 100 m above the active channel. Lower view shows close-up of bedrock-fill contact, which is indicated by dashed line in both photographs. Arrow indicates flow direction on main channel. 178 Rivers in the Landscape response to base level fall, as long as the channel Fill terraces differ from strath terraces in that remains fixed to the changing base level such that the fill terraces can be event-based features result- river maintains a steady-state profile with uniform ingfromasingleflood,debrisflow,orlandslide valley and channel geometry and constant concav- that overwhelms sediment transport capacity and ity and steepness (Gilbert, 1877; Mackin, 1937; Paz- causesfloodplainaggradationbeforeriverflowsub- zaglia, 2013). Creation and preservation of a terrace sequently incises the deposit (Miller and Benda, reflects unsteadiness in the profile because the chan- 2000; Strecker et al., 2003). Fill terraces are also more nel aggrades and/or incises, and geometry, concav- likely to be reduced in planform size during lateral ity, and steepness can be altered. This unsteadiness erosion after the terrace forms (Moody and Meade, is a transient response to changes in water and sedi- 2008). ment input or base level. The mechanisms by which terraces form, along Strath terraces have been attributed to a wide with terrace geochronology, have been the subject variety of changes in river dynamics (Wohl, 2010; of much research because a terrace can record some Pazzaglia, 2013), including change in external control variables, such as water r and sediment yield or base level, and thus pro- periods of balanced sediment supply; vide paleoenvironmental information. Terrace stud- r altered sediment supply; r ies began in the nineteenth century with the con- glacial–interglacial transitions; ceptual model that river terraces reflected an exter- r tectonically induced changes in rock uplift; r nally imposed change in river dynamics associated falling local base level, eustatic base level fall with glaciation, uplift, or changing base level (Davis, (global sea level change); and r 1902b). Terraces continued to be studied under autocyclic oscillations in erosion rate in laterally this assumption until Schumm’s work on semi- migrating channels. arid alluvial channels (Schumm and Hadley, 1957) Numerical modeling suggests that formation of and physical experiments (Schumm, 1973) demon- strath terraces depends on input variability that cre- strated that terraces could result from crossing inter- ates a changing ratio of vertical to lateral erosion nal thresholds while external inputs to the river net- rates (Hancock and Anderson, 2002). A wide variety work remained unchanging. Schumm and Hadley of external and internal factors can create input vari- (1957) described a for ephemeral ability, which explains how strath terraces can result riversinsemiaridregionsthatinvolveslongitudi- from different causes in diverse river catchments or nally discontinuous incision. Infiltration into the within a catchment through time. streambed during brief periods of flow causes down- Similarly, fill terraces have been attributed to stream decreases in discharge, local aggradation, diverse changes (Wohl, 2010; Pazzaglia, 2013), steepened channel gradients, and incision, all with- including out any change in climate or land cover that alters water and sediment inputs to the river network. r fluctuating water and sediment discharge during This represented a major shift in thinking about glacial cycles, volcanic eruptions, climatic change, howsometypesofterracecouldform,andledto and changing land use; the idea that event-based terraces could result from r repetitive lateral shifting and stillstands during temporaryfluctuationsinwaterandsedimentsup- continuous downcutting; ply without major shifts in climate, tectonics, or r fluctuatingbaselevel; land use. r complex response to base level change; and Supplemental Section 6.2.2 reviews physical r sediment waves migrating down catchments over experiments and numerical simulations used to periodsoftenstohundredsofyearsinresponseto understand mechanisms of terrace formation and hillslope mass movements or land use. preservation. Chapter 6 Extra-channel environments 179

6.2.3 Terraces as paleoprofiles and ate strath surfaces along segments of more resistant paleoenvironmental indicators bedrock (Wohl, 2008a). Third, event-based terraces that result from local disturbances such as landslides create fill terraces Grain size and stratigraphy of terraces can be used of limited longitudinal extent (Montgomery and to infer sediment supply and flow conditions (Sec- Abbe, 2006; Wohl et al., 2009). Even more sustained tion 3.2.1). The longitudinal profile of a terrace changessuchasbaselevelriseorfallmayonlycause canapproximatethepaleoprofileoftheriver.Age terraces in the lower portion of a river network of the terrace can be used to constrain the tim- (Merritts et al., 1994). Strath terraces, in particular, ingofeventsthatresultedinterraceformation. may require some minimum drainage area to form However, inferring paleoenvironmental conditions (Garcia, 2006) because of the stream energy required from terrace characteristics is complicated by several to create a strath surface in bedrock. The forma- factors. tion and preservation of straths can also vary with First, a given change in climate that alters water differences in bedrock erosional resistance (Mont- and sediment yield to a river network can cre- gomery, 2004; Wohl, 2008a). Terraces may be thick- ate very different responses within the river net- est, widest, and best preserved near tributary junc- work, depending on the climatic conditions present tions where tributary sediment widens the main- prior to the change. This is illustrated by Bull’s stem and helps to build a terrace sufficiently large (1991) work in the Charwell River basin of central to be preserved (Pazzaglia, 2013). New Zealand. Different portions of the river basin Fourth, a single perturbation such as base level responded differently when regional precipitation fall can result in multiple terraces if complex increased slightly and temperature increased more response occurs (Section 6.1.1; Schumm, 1973). substantially. The lower portion of the drainage, Each terrace along a river therefore cannot necessar- from the basin mouth to the contemporary tree line, ilybecorrelatedwithanexternalchange. changed from periglacial conditions to greater vege- Fifth, the duration of time represented in terrace tation density and soil thickness. This reduced sedi- alluvium can vary widely. The alluvium capping a ment yield, increased peak discharge, and resulted strath terrace is synchronous with formation of a in stream incision. The upper portion of the strath, for example, whereas the alluvium of a fill drainage above timber line experienced an increase terraceisyoungerthantheunderlyingsurface,and in periglacial processes and increased sediment can accumulate over long periods of time (Pazzaglia, yield, so that channels in this zone did not incise 2013). (Bull, 1991). Sixth, the timing of terrace formation can sub- Second, a change in base level, such as that stantially lag the timing of any external perturba- caused by tectonic uplift, can result in different tioninwaterandsedimentyieldorbaselevel(Paz- responses across a river network. For example, zaglia, 2013). Base level on the Drac River in the larger rivers in the vicinity of a triple junction French Alps dropped 800 m following removal of in northern California, USA, can maintain uni- an ice-dam during glacial retreat. As a knickpoint form incision into bedrock during relatively rapid moved upstream through the basin in response to uplift, creating sequences of strath terraces along base level lowering, terrace formation lagged behind the rivers (Merritts and Vincent, 1989; Merritts base level drop by 2000–5000 years, depending on et al., 1994). Smaller, headwater rivers cannot incise location within the drainage (Brocard et al., 2003). rapidly enough to keep pace with uplift, and instead These complicating factors do not by any means become steeper with time rather than creating strath render terraces useless as paleoenvironmental indi- terraces. Studies elsewhere indicate that, even on cators. The presence of spatial and temporal dif- larger rivers, terrace formation can be longitudi- ferencesandlagtimesinterraceformationwithin nally discontinuous if a river lacks the power to cre- 180 Rivers in the Landscape arivernetworkonlymeansthatterracesmustbe tive decay—the energy stored is a function of time interpreted carefully and in a broader context than thatthesedimenthasbeenburied,aswellasback- a single terrace exposure, as with any other river ground decay rate); feature. An example of paleoenvironmental inter- r U-series dating of pedogenic carbonate in terrace pretation comes from Merritts et al. (1994), who sediments (Candy et al., 2004); interpreted Quaternary strath and fill terraces along r cosmogenic isotope dating of terrace sediments or three rivers in northern California in the context of bedrock in strath terraces (Riihimaki et al., 2006); river response to tectonic uplift and fluctuating base r paleomagnetic dating of terrace sediments (Pan level associated with changes in sea level. Fluctuat- et al., 2003); ing sea level strongly influenced terraces along the r tephrochronology of volcanic ashes incorporated lower portion of each river (tens of kilometers long). in terrace sediments (Dethier, 2001); and Aggradation led to formation of fill terraces dur- r radiometric ages of bounding lithologic units such ing sea level high stands. Incision occurred during as basalt flows (Maddy et al., 2005). low stands. Long-term uplift created unpaired strath terraces in the middle and the upper portion of Relative methods infer differences in age between each river. terraces by characterizing parameters that change Strath and fill terraces can be used to recon- through time, although the rate of change in a structriverlongitudinalprofileatthetimeofstrath parameter is not necessarily linear. Relative methods or floodplain formation (Paola and Mohrig, 1996), applied to terrace chronology include although creation of straths, in particular, can lag by several thousand years the input changes that cause r development of weathering rinds on coarse clasts their formation (Merritts and Vincent, 1989). Both in terrace sediments (Pazzaglia and Brandon, types of terraces also can subsequently be tectoni- 2001); cally deformed (Finnegan, 2013). The terrace sur- r soil characteristics (thickness, accumulation of face is the easiest proxy to use for paleoprofile. How- translocated clays or soluble salts, color, and iron- ever, this surface can be deformed by weathering and oxide speciation) (Tsai et al., 2007); inputs of colluvium from adjacent hillslopes, so the r amino acid racemization of organic compounds more difficult-to-discern underlying bedrock strath such as shells (Penkman et al., 2007); and provides a better indicator of original longitudinal r lichenometry (the diameter of lichen colonies, profile (Pazzaglia, 2013). which increases with time) (Baumgart-Kotarba Any paleoenvironmental inferences drawn from et al., 2003). terraces rely on establishing a chronology of ter- race formation (Wohl, 2010; Pazzaglia, 2013). Ter- Many studies utilize multiple techniques (Gural- race chronology can be based on absolute methods. nik et al., 2010). The actual event or period of time Absolute methods provide a numerically precise age, dated can be either the phase of tread construction although the accuracy of this age is not necessarily (erosionofastrathordepositionofafill)orthe greater than that achieved using relative methods. incision that produces a terrace scarp (Ritter et al., Absolute methods applied to terrace ages include 2011). The phase of tread construction is time trans- gressiveandoflongerduration,whereasincisioncan r 14C dating of organic matter in terrace sediments be widespread and rapid (Pazzaglia, 2013). (Merritts et al., 1994); As numerous detailed case studies accumulate in r dendrochronology of terrace vegetation (Pierson, the scientific literature on fluvial geomorphology, 2007); diverse groups have begun to establish digital cat- r luminescence techniques (Pederson et al., 2006) alogs and databases. The Fluvial Archives Group (luminescence measures energy stored in sedi- (FLAG) is actively documenting the global record ment as a result of natural, background radioac- of terrace stratigraphy (Vandenberghe and Maddy, Chapter 6 Extra-channel environments 181

2000). FLAG creates databases for Pleistocene and ment. Paraglacial fans are alluvial fans composed Holocene sedimentary archives. mainly of reworked glacial deposits (Ryder, 1971). As with other types of fluvial deposits, LiDAR Fan deltas formatthemarginsoftheocean,where and DEMs facilitate designation and mapping of ter- rivers flow from a mountainous region across a nar- races (Jones et al., 2007). Shallow geophysical tech- row coastal zone, and can grade into deltas that form niques such as ground penetrating radar and electri- in freshwater or marine subaqueous environments cal resistivity (Froese et al., 2005) are used to image (Section 6.4). Small alluvial fans form at crevasse terrace subsurface geometry. These images can sub- splays below levees on floodplains. stantially enhance information available from lim- Alluvial fans can form in any climatic region. ited surface exposures. However, fans in arid and semiarid environments are by far the most well studied, not least because the lack of continuous vegetation cover results in 6.3 Alluvial Fans good exposure of fan features (Figures 6.4, S6.9, and S6.10). Alluvial fans are primarily depositional environ- Lower transport capacity than upstream chan- ments that are shaped by the combined effects of nel reaches promotes aggradation, overbank flows, erosion and deposition, similar to the other features andfrequentavulsiononalluvialfans,whichcre- addressed in this chapter. Channels exiting high gra- ates numerous hazards for structures built on these dient, laterally confined canyons in regions with surfaces. Fans are also heterogeneous with respect to high relief commonly create a fan-shaped deposit process, form, grain size and stratigraphy, and soils known generically as an alluvial fan, although collu- and vegetation. Fan sediments tend to be relatively vial processes can contribute to sediment accumu- porous and permeable, which facilitates infiltration lation. Water flow, hyperconcentrated flow, debris and subsurface flow, allowing fans to hydrologically flow, rockfall, landslide, and snow avalanches can buffer a catchment (Vivoni et al., 2006). Herron and all contribute sediment to alluvial fans. Fans can be Wilson (2001), for example, described how an allu- differentiated based on primary depositional pro- vial fan in the subhumid temperate climate of south- cess and morphology, resulting in debris cones dom- eastern Australia significantly buffered hydrology in inated by rockfalls, debris fans dominated by debris a 26 ha catchment by storing 20%–100% of surface flow, and so forth. Fans classified in this way can be runoffdeliveredtothefan.Wetlandscanoccurat differentiated as those dominated by inertial trans- the downstream end of fans in humid climates as port (debris flows, rockfalls) versus those dominated lower gradients promote deposition of finer, less per- by traction transport (hyperconcentrated and fluvial meable sediments and subsurface flow rises back flows) (Stock, 2013). toward the surface (Woods et al., 2006). Where parallel channels drain from a mountain Fanscanalsobuffersedimentdeliverybyatleast range into a basin, adjacent alluvial fans can coalesce temporarily storing sediment. Fans thus lower the to form an alluvial apron or bajada. Fluvial megafans connectivity of sediment movement and the cou- are unusually large alluvial fans that mostly occur pling between hillslopes and channels or between between 15◦ and 35◦ north and south of the equator tributaries and larger rivers. The volume of sediment where rivers that undergo moderate to extreme dis- stored can fluctuate substantially through time, as charge fluctuations enter actively aggrading basins processes of fan erosion—incision of the main or (Leier et al., 2005). Australians use the term flood- distributary channels, or erosion of the toe of a out to describe sites where a marked reduction in tributary fan by a larger river—alternate with pro- channel capacity creates greater overbank flows and cesses of deposition. In the Grand Canyon of Ari- associated deposition (Tooth, 1999). Floodouts bear zona,USA,forexample,flashfloodsanddebris some similarities to fans, but are not necessarily flows along small, steep, ephemeral channels create associated with a change in lateral channel confine- tributary fans that laterally constrict the mainstem 182 Rivers in the Landscape

(a) (b)

Figure 6.4 Examples of alluvial fans. (a) A fan created at the junction of a tributary with the Colorado River in the Grand Canyon, USA. The older fan deposits are outlined with the dashed white line, the most recent fan deposits with the dotted white line at far right center. Vegetated deposits in the foreground are the floodplain. Mainstem flow is left to right. (b) An event-based alluvial fan created by a damburst flood down the Roaring RiverinColorado, USA. The Roaring River enters the Fall River valley: the Fall River flows from lower left to upper right in this view. The outburst flood occurred in 1982, and the resulting alluvial fan appears lighter in color (and outlined indashed white line) in this photograph, taken nearly 30 years later, because of slow growth of vegetation on the fan surface. See also Figure S6.4 for additional, color photographs of fans.

Colorado River. Fans continue to grow until the the apex of many alluvial fans, with active deposi- associated lateral constrictions create supercritical tion concentrated farther down the fan (Blair and flow during large floods on the mainstem, when the McPherson, 2009). Channels also frequently avulse fans are eroded sufficiently to allow mainstem flow or are abandoned through stream capture. to become subcritical (Kieffer, 1989). Downstreamfromthezoneofmostactiveinci- sion, channels are commonly braided and can decrease in size downstream as a result of discharge 6.3.1 Erosional and depositional infiltration, so that channelized flow can give way to processes sheetflow (Bridge, 2003). Although deposition can produce relatively smooth surfaces, secondary chan- Alluvial fans are geomorphically complex and nel networks can dissect these surfaces once deposi- dynamic environments. Channelized water flows, tion shifts elsewhere on the fan. hyperconcentrated flows, and debris flows alternate Blair and McPherson (2009) distinguished pri- spatially and temporally with unconfined sheet flow mary and secondary depositional processes. Pri- or with rapid infiltration that results in sieve deposits mary processes transport sediment from the catch- and subsurface flow (Griffiths et al., 2006). Deposits ment onto an alluvial fan and cause fan construction. onmanyfansarecoarsegrainedandpoorlysortedas Secondary processes modify previously deposited a result of relatively short transport distances, inputs sediment via overland flow, wind erosion or depo- by debris flows and flash floods, and rapid loss of sition, bioturbation, soil development, weathering, flowcapacityasaresultofinfiltrationorovertopping and toe erosion. Secondary processes dominate fan of shallow channels (Blair and McPherson, 2009). surfaces away from the usually limited areas of Episodic aggradation at the fanhead causes over- recent deposition. steepening that initiates incision. Consequently, an Blair and McPherson (2009) also distinguished incised channel, or fanhead trench,ispresentnear 3classesand13differenttypesofalluvialfans Chapter 6 Extra-channel environments 183

(Table S6.1) based on dominant primary processes example comes from the Southern Alps of New and textures. The three classes (bedrock, cohe- Zealand, where Korup et al. (2004) compiled case sive colluvium, non-cohesive colluvium) reflect the studies documenting a series of alluvial fans dom- catchment slope material from which primary pro- inated by sediment inputs from large (>106 m3) cesses are triggered. landslides. As Stock (2013) noted, very few direct observa- Supplemental Section 6.3 describes approaches tions of transport and depositional processes on flu- used to measure and model processes on alluvial vially dominated fans have been published. These fans. observations consistently indicate Complex and intermittent erosion and deposi- tion on alluvial fans create special challenges for haz- r fluctuations in water and sediment supply during ard zoning and mitigation. Fans are commonly the asinglestorm, only gently sloping surfaces available in high-relief r formationofcoarsebarsbyparticleaccretion environments, and hazards associated with erosion and subsequent diversion of flow and sediment and deposition are intermittent in time. Conse- around these bedforms on gravel-bed surfaces— quently, urbanization concentrates on fans (Zorn these processes of channel bifurcation and avul- et al., 2006). Approaches used to mitigate hazards sion are analogous to those in braided rivers, (Kellerhals and Church, 1990; Pelletier et al., 2005; r shallow, hydraulically rough flow, and unstable Wohl, 2010; Stock, 2013) include flows, r critical flow with standing waves and presumably r channelization and bank stabilization to limit antidunes in sand- to cobble-bed surfaces, and r overbank flows, sheetwash, and channel avulsion; large concentrations of suspended and bedload r debris interception barriers and detention basins transport. to contain sediment entering the fan, rather than allowingthesedimenttodisperseandburyinfras- Processes of sediment accumulation seem to be tructure or fill active channels; consistently associated with bifurcation, avulsion, r hazard mapping to guide development of infras- and surges of water and sediment on alluvial fans. tructure; and Alluvial fans are typically built over time by con- r warning devices to facilitate rapid evacuation in tinuing deposition, but individual fans can result case of an erosional or depositional event. from a single extreme event, such as massive erosion associated with a damburst flood. The Roaring River alluvial fan in Colorado, USA, (Figure 6.4b) was cre- 6.3.2 Fan geometry and ated by a dam failure on a headwater lake in 1982. stratigraphy The resulting outburst flood flowed more than 7 km down the steep, narrow Roaring River, eroding sub- An alluvial fan approximates a segment of a cone stantial quantities of sediment from the Pleistocene- radiating downslope from a point where a channel age glacial tills that underlie the river course. Where flows out from a high-relief catchment (Blair and the Roaring River enters the wider, lower gradient McPherson, 2009). In plan view, a fan is arcuate, Fall River valley, the flood created an alluvial fan creating a 180 degree semicircle unless the toe of the covering 0.25 km2 and up to 14 m thick (Blair, 2001), fan is truncated by erosional processes, or the sides whichpersistsmorethan30yearslater. of the fan are restricted by adjacent fans. Longitu- Fanscanalsobedominatedbyaspecificand dinal fan profiles within and away from the channel sometimes repetitive type of disturbance, such as are typically segmented, with the steepest slopes tributary glaciation, volcanic eruption, wildfire- along the flanks (Hooke and Rohrer, 1979). Seg- induced mass movements (Figure S6.11), and land- mentation can reflect intermittent uplift, channel slides (Korup et al., 2004; Torres et al., 2004). An incision, and episodic deposition of various types 184 Rivers in the Landscape

(Bull, 1964; Arboleya et al., 2008). Channel gradient regions such as piedmonts (Stock, 2013). These typically decreases downstream from 0.10–0.04 areasfavorlargefansbecauseofthelackofcon- m/mto0.01m/monfluviallydominatedfans(Stock finement on depositional space. Extensional regions et al., 2007). This decrease in gradient reflects pro- such as the US Basin and Range, Tibetan Plateau, gressive deposition and declining sediment loads and Andean back-arc are characterized by moun- and grain size down-fan. Radial fan profiles tend tain ranges and intervening subsiding basins along to be concave or planar, and cross-fan profiles the margins of which alluvial fans form. These convex (Bull, 1977; Blair and McPherson, 2009). regionstendtobearidbecauseofrain-shadow Fans typically extend 0.5–10 km from a mountain effects from bounding mountain ranges and current front, but large fans can extend 20 km (Blair and global circulation patterns, which facilitates identi- McPherson, 2009). fication and study of their alluvial fans. Some of the The volume of sediment available for deposition largestfans,suchastheKosifan,forminconver- on a fan and the processes of deposition exert first- gent tectonic settings such as along the Himalayan order controls on fan geometry. Climate, lithology, or Taiwan thrust fronts (Stock, 2013). In this setting, tectonics, and drainage area all influence sediment active uplift provides abundant sources of sediment, delivery to a fan. Climate influences weathering and but the depositional area can be less constrained caliber of sediment supplied to the fan, as well as than in an extensional setting with numerous moun- flow available to transport sediment. Oguchi and tain ranges. Other examples of formerly tectoni- Ohmori (1994) compared alluvial fans formed in the cally active settings conductive to fan formation are wet climate of Japan against fans formed in the arid retreating escarpments such as South Africa’s Drak- American Southwest, across a range of lithologies. ensberg or the southeastern Australia passive mar- Theyfoundthat,althoughfanareaineachregionis gin, which juxtapose sediment sources in higher- proportional to the catchment sediment output per relief areas with relatively flat piedmonts. unit time, sediment yield per unit area is higher in Insomeoftheearliestquantitativeworkonallu- theJapanesebasinsforthesamebasinslope. vial fans, Bull (1962) found that drainage area corre-

Lithologiesthatweathertoverycoarsesediment sponds to fan size (Af)andfanslope(Sf)as deposited by mass movements typically create the n steepest fans. Lithologies that weather rapidly and Af = aAd (6.1) release very fine sediments result in moderately d Sf = cAd (6.2) steepfansoflargervolumes,whereaslithologiesthat weather to sand-sized sediment create low-gradient with a varying slope of 0.7–1.1 for the regression fans of smaller volume per unit drainage area (Bull, line (n), which averages 0.9 for fans in the United 1964; Blair, 1999). States. Coefficient a varies by more than an order Tectonics influence rates of sediment production of magnitude as a result of lithology, climate, and and delivery to rivers in upstream catchments, and the space available for deposition on the fan (Vis- the accommodation space in depositional sites. A eras et al., 2003). Although these equations con- basin that is subsiding relative to the source area can tinue to be widely applied, Blair and McPherson develop a thicker fan of more limited spatial extent, (2009) noted limitations. These include the com- whereas a tectonically stable basin can develop a parison of plan-view areas for 3D features (Equa- thinner but more areally extensive fan (Whipple and tion 6.1), which are not of constant thickness and Trayler, 1996). for which thickness is commonly unknown, and the Tectonics can dominate fan formation in regions lack of clear guidelines for defining area, which may with active extension (divergence of plates), com- not be straightforward (Stock, 2013). pression (convergence of plates), or transtension Various morphometric indices, including relief (lateral plate movement) in which confined, steep and ruggedness, as well as particle size and round- valleys are juxtaposed with unconfined depositional ness, and fabric of deposits, can be used to Chapter 6 Extra-channel environments 185 distinguish spatial differences in depositional pro- past 2000 years. Alluvial fan sediments provided a cesses on fans (de Scally and Owens, 2004; Stock, longer record (8000 year) of fire than tree rings, were 2013) and to distinguish dominant depositional pro- ubiquitous in this mountainous environment, and cesses on individual fans (Al-Farraj and Harvey, recorded stand-replacing fires that limited tree-ring 2005). Debris flows with high sediment concentra- records. Although alluvial fan deposition is episodic tioncancreatesteepandruggedtopographyon in time and space, discontinuities can be offset by the upper fan, for example, whereas low sediment compiling records from numerous individual strati- concentration can create smoother topography on graphic sections (Pierce and Meyer, 2008). the lower fan surface by filling channels and other depressions (Whipple and Dunne, 1992). Spatial and temporal heterogeneity in deposi- 6.4 Deltas tionalprocessesonalluvialfansmaketheresult- ing stratigraphic records challenging to interpret. A delta is analogous to an alluvial fan in that it is pri- Nonetheless, an extensive literature, summarized in marily a depositional feature that results from a loss Wohl (2010), documents how changes in the volume oftransportcapacity:inthecaseofadelta,wherea and style of fan deposition through time have been river enters a body of standing water. Just as an allu- used to infer, for example, vial fan is roughly triangular in shape, the word delta was first applied to a fluvial deposit 2500 years ago r climatically driven changes in sediment supply bythehistorianHerodotusbecauseoftheNiledelta’s to alluvial fans in Japan (Oguchi and Oguchi, resemblance to the Greek letter delta, Δ (Ritter et al., 2004; Waters et al., 2010), and changes in sedi- 2011). ment supply associated with alpine glacier retreat A delta can protrude well beyond the adjacent in Switzerland (Hornung et al., 2010); coastline when a river carries large volumes of sed- r human-induced and tectonically driven changes iment, and currents in the receiving body have in sediment supply in the Solway Firth– limited ability to rework that sediment. Examples Morecambe Bay region of Great Britain (Chiver- of deltas that protrude well beyond the adjacent rell et al., 2007) or deposition in the Tagliamento shoreincludetheMississippiRiverdeltainthe River region of Italy (Spaliviero, 2003); Gulf of Mexico, the Danube in the Black Sea, the r the magnitude and frequency of event-based sedi- Nile in the Mediterranean Sea, Italy’s Po–Adige mentation associated with storms or wildfires in delta in the Adriatic Sea, Russia’s Volga the US Northern Rocky Mountains (Pierce and intheCaspianSea,theGanges–Brahmaputraand Meyer, 2008); and Irrawaddy in the Bay of Bengal, and Africa’s Niger r to construct short- and long-term sediment bud- intheGulfofGuinea. gets for the upstream catchment of the Storutla Or the delta deposits can be primarily upstream River in southern Norway (McEwen et al., 2002). from the adjacent coastline, so that the delta does not protrude into the ocean as a result of limited An example comes from the work by Pierce particulate sediment transport from the river basin and Meyer (2008) in Yellowstone National Park or thorough reworking and transport of fluvial sedi- in Wyoming, USA. Much of the park is covered ment by currents in the receiving water body. Exam- in conifer forests that burn at regular intervals, ples include the Amazon and the Rio de la Plata mostly as a result of lightning strikes during peri- in South America, Australia’s Murray–Darling, the odsofdrought.PierceandMeyerusedtree-ring Congo, and the Columbia in North America. records of fire and drought, pollen and charcoal in Regardless of the shape, deltas are biologically lake sediments, and charcoal and other sediments diverse and productive environments that typically in numerous alluvial fans across the park to infer support commercial fisheries. Deltas also host veg- changes in fire regime and sedimentation over the etative communities that can help to reduce the 186 Rivers in the Landscape velocity and associated damage of incoming oceanic in Greece, for example, are predicted to retreat storm surges. Delta features such as area of islands 700–2000 m inland under projected sea level rises and bars and type and extent of vegetation influence of 0.5–1 m as a result of combined effects of inun- the tidal prism—the volume of water exchanged in a dation during sea level rise and associated coastal tidalcycle—aswellasvaluesofbedshearstressand erosion (Poulos et al., 2009); thus sediment dynamics (Canestrelli et al., 2010). r loss of water supply, transport, or waste disposal if As river flow entering the body of standing water river flow shifts location across the delta; and expands and decelerates, sediment is deposited. Var- r salinization of delta soils and groundwater if river ioustypesofcurrentsinthereceivingwaterbody flow decreases, sea level rises, or groundwater is reworktheriversediment,sometimeswiththeresult pumped at a rate faster than natural recharge—in that no delta forms. The morphology and stratig- southeastern Spain, for example, increasing water raphyofthedeltareflectthebalancebetweenriver demand for crop irrigation and reduced irriga- inputs and reworking of sediment by the receiv- tion return flow linked to new irrigation tech- ing water body. In general, the coarsest sediment is niques have reduced recharge of some coastal deposited closest to the and finer sedi- aquifers, which have become more saline since ment in suspension is carried farther into the receiv- 1975 (Rodr´ıguez-Rodr´ıguez et al., 2011). ingbody.Largedatabasescontainingdatafrom numerous deltas indicate that the area of a delta is Like alluvial fans, deltas exemplify terra non- best predicted from average discharge, total sedi- firma. ment load entering the delta, and offshore accom- modation space (Syvitski and Saito, 2007). Many of the world’s great cities are located on 6.4.1 Processes of erosion deltasalongmarineshorelines.Thisistypicallya and deposition mixed blessing for cities such as Shanghai, New Orleans, Calcutta, Rotterdam, Marseille, and Lis- Deposition on deltas, as on alluvial fans, constantly bon. On the plus side, deltas provide ease of trans- shifts location across or down the delta in response port and navigation of goods from the terrestrial to changes in fluvial water and sediment discharge, interior and across the ocean, fertile soils and rich as well as changes in delta morphology result- fisheries, and reserves of oil and natural gas in delta ing from deposition, subsidence, and reworking by sediments. These benefits are offset by some of the waves and tides. The classic model of delta deposi- same hazards common on alluvial fans, such as over- tion involves deposition of a longitudinal bar where bank flooding and channel avulsion. Deltas also river transport capacity declines because of mixing include hazards associated with between river water and the water of the receiving basin. The initial bar forces the river to bifurcate, r subsidence of delta sediments—an example comes and the process continues, developing a distributary from the Pearl River delta of China, where channel network. Distributary channels formed at measurements based on interferometric synthetic least in part by avulsion are lined with natural lev- aperture radar data indicate average subsidence ees. With continued deposition, the distributary net- of 2.5 mm/year within 500 m of the coast, and work progrades into the receiving body. maximum values of 6 mm/year associated with Slingerland and Smith (1998) developed a model rapid urban development, in a region less than 2 m predicting avulsion stability based on the ratio of above current mean sea level (Wang et al., 2012); thewatersurfaceslopesofthetwobranchesofa r flooding of low-lying portions of the delta during bifurcation. Ratios < 1(wherethesecondarychan- marine storm surges; nel slope is the numerator and the main channel r marine erosion of the delta if river sediment sup- slope is the denominator) result in a failed avulsion ply decreases or sea level rises—deltaic coastlines in which the incipient avulsion fills with sediment. Chapter 6 Extra-channel environments 187

Ratios > 5 create avulsions in which the river aban- chemical processes; lake flooding; and even surface dons its original course in favor of the steeper albedo as floodwaters hasten the melting of snow channel. and ice, causing a slight increase in air tempera- Jerolmack and Mohrig (2007) proposed that avul- tures at a critical time for the growth of terres- sion frequency scales with the time needed for sedi- trial and aquatic plants (Goulding et al., 2009a). Ice mentationonthestreambedtocreateadepositequal transported by the floodwaters into overbank areas to one channel depth. They used relative rates of enhances localized erosion. Decreases in the sever- bank erosion and channel sedimentation to derive a ity of river-ice break-up during recent decades have dimensionless mobility number to predict the con- lessenedthefloodingofsomedeltalakes,which ditionsunderwhichdistributarychannelsform. mayleadtolossofthesewaterbodies,aswellas An active delta includes subdeltas created when changes in the biogeochemical interactions between sediment diverted through breached levees accumu- river water and the floodplain ecosystem, and the latesascrevassesplays.Channelsonthesubdelta processes described earlier (Prowse et al., 2011). also bifurcate and prograde until they develop gradi- As a delta front progrades further into the receiv- entssimilartothoseofthemainchannelanddepo- ing basin, shorter routes for water flow into the sition shifts to a new subdelta. On the Mississippi basin become available along the sides of the delta. River, the process of forming and abandoning sub- These shorter routes can also access topographi- deltas occurs over a few decades (Morgan, 1970). cally lower portions of the coastline than the delta, Physical experiments suggest that the timescale for which has been progressively built up as well as out. these processes on river-dominated deltas reflects River flows commonly access such shorter routes via the time over which a river that is pro- crevasses developed in levees well upstream from grading basinward reaches a critical size and stops the delta, causing delta switching,arelativelyabrupt prograding (Edmonds et al., 2009). This triggers shift in river course and delta location. Over time, a wave of bed aggradation that moves upstream, delta switching results in distinct depositional lobes increasing overbank flows and shear stresses exerted within the general depositional complex (Aslan and on the levees, triggering avulsion and the growth of Autin, 1999). Delta switching occurs approximately subdeltas. every 1000–2000 years along the Mississippi River Inhigh-latituderivers,iceandice-jamfloodscan (Aslan and Autin, 1999), for example, and along the exert an important control on processes of delta ero- Ebro River in Spain (Somoza et al., 1998). sion and deposition. Canada’s Mackenzie River pro- vides a well-studied example. The Mackenzie, which 6.4.2 Delta morphology drains 1.8 million km2, enters the Arctic Ocean at 70◦ N. The Mackenzie Delta stretches across and stratigraphy 12,000 km2, including a 200 km maze of mean- dering channels and 45,000 riparian lakes (Gould- Deltas reflect the changing balance through time ing et al. 2009b). High spring flows from snowmelt between at least four factors (Bridge, 2003). dominate Mackenzie River runoff. The northerly flow direction of the river results in the spring 1.Thevolumeandgrain-sizedistributionofriver flow peak progressing downstream with the seasonal sediments: The characteristics of the river sedi- advance of warm weather. The flood wave com- ments supplied to the delta integrate everything monly encounters an intact and resistant ice cover in the river catchment that influences sediment downstream, causing the formation of ice jams and supply—lithology, relief, tectonics, climate, land enhanced overbank flooding (Prowse and Beltaos, cover, and land use—as well as river discharge and 2002). The spatial extent and duration of overbank transport capacity. The greater the sediment load flooding influence erosion and deposition; habi- coming down the river, the more likely the river tatabundance,diversity,andconnectivity;biogeo- is to build a delta. Conversely, the stronger the 188 Rivers in the Landscape

currents in the receiving water body, the more gradient and transport capacity, the location of likely these are to rework the river sediments and the shoreline, and the energy of currents in the dominatethesizeandformofthedelta,orto receivingbasin.Eitherriseorfallofrelativebase completely remove the river sediments. level can shift the locations of active erosion and 2. The relative density of the river water and the deposition across a delta. An example comes from receivingwaterbody:Therelativedensitiesof a Pleistocene-age lake delta in northwestern Ger- river and receiving waters reflect salinity, temper- many,inwhichperiodsoflake-levelrisecorre- ature, and concentration of suspended sediment. spond to vertically stacked delta systems as depo- When the densities are similar (homopycnal flow), sitional centers shifted upslope, and lake-level fall the river and basin water mix in three dimensions corresponds to development of a single incised closer to the shoreline and much of the sediment valley with coarse-grained delta lobes in front of is deposited close to the river mouth (Bridge, the valley (Winsemann et al., 2011). 2003). This is most likely to occur where a river enters a freshwater lake. River water denser than One of the more commonly used classifications basin water (hyperpycnal)—such as very cold or of deltas emphasizes the relative importance of flu- turbid water entering a freshwater lake—will flow vial processes or reworking by the receiving body as a bottom current, mixing only along its lat- (Figure 6.5) (Ritter et al., 2011). High-constructive eral margins, carrying sediment farther into the deltas reflect predominantly fluvial deposition; these basin.Hyperpycnalflowscanalsooccurwherea are sometimes known as river-dominated deltas riverentersanocean,asontheHuangHe(Yellow (Fagherazzi, 2008). River-dominated deltas can be River) of China. River water of lower density than elongate, with greater length (in the downstream the basin water (hypopycnal)—such as freshwater direction) than width, like the contemporary delta of entering saline water—can flow over the denser the Mississippi River. Elongate deltas have more silt water as a buoyant plume, mixing along the bot- and clay and subside rapidly once deposition ceases tom and sides, and carrying sediment further into or shifts elsewhere on the delta. River-dominated thebasinthaninthecaseofhomopycnalflow. deltascanalsobelobate, with greater width than This type of flow is common on the delta of the length, and slightly coarser sediment that subsides Mississippi River. more slowly upon abandonment, allowing time for 3. The currents of the receiving water body: wavesandtidestoreworkthesandysediment.The Lakes and enclosed seas typically experience now-abandoned Holocene deltas of the Mississippi lower-energy waves and tidal currents, although River are lobate. storm-generated winds can temporarily enhance High-destructive deltas are shaped by currents in wave energy. Deltas in these lower-energy the receiving water body. Sediment is moved along- environments are likely to reflect predomi- shore to form beach ridges and arcuate sand barriers nantly fluvial processes. River mouths entering in wave-dominated deltas such as those of the Rhone long, narrow embayments such as the Bay of or Nile River. Tide-dominated deltas have strongly FundyonCanada’ssoutheasterncoastline,or developed tidal channels, bars, and tidal flats, with semi-enclosed seas, such as the North Sea in the sand deposits that radiate linearly from the river Atlantic or the Sea of Cortez off northwestern mouth, such as those in the Gulf of Papua (Ritter Mexico, can have very strong tidal currents that et al., 2011). causetidalborestotravelupthebayandintothe Anytypeofdeltacanbesubdividedintodelta entering river, strongly influencing delta shape plains and a delta front. The majority of most and size. deltas consists of delta plains,extensiveareasoflow 4. Subsidence and water-level change in the receiv- slope crossed by distributary channels. Delta plains ing body, which influence relative change in base include subaerial portions of floodplains, levees, level:Changesinrelativebaselevelinfluenceriver and crevasse splays, and subaqueous portions Chapter 6 Extra-channel environments 189

Lobate Elongate 17 km (Lafourche (modern Mississippi) Mississippi)

Prodelta 17 km

Distributary channel, Delta plain (marsh, Delta front (channel mouth levee, crevasse splay swamp, lake) bar and sheet sands) High-constructive deltas

High-destructive deltas Tide dominated Wave dominated (Gulf of Papua) (Rhone River)

17 km

17 km Prodelta Shelf

Channel Delta plain Tidal Channel and Delta plain (flood basin tidal flat channel-shelf meander belts and marine coastal basin) Delta plain Tidal Tidal Coastal barrier– Channel mouth bar (non-tidal) sand bar channel deeps strandplain Figure 6.5 Basic delta types. (Adapted from Ritter et al., 2011, Figure 7.35.) of marshes, lakes, interdistributary bays, tidal China), rather than woody vegetation. Regardless drainage channels, and tidal flats (Bridge, 2003). of the specific plant communities, vegetation cover The gradient of the delta plain correlates with the strongly increases the resistance of delta sediments ratio of sediment supply to sediment retention on to erosion. Removal of vegetation typically results in the delta, sediment concentration (a proxy for delta much faster and more severe delta erosion, as well as plain sedimentation), and mean water discharge lessefficienttrappingandstorageofincomingsedi- (Syvitski and Saito, 2007). The delta plain gradient ment.Analysisofhistoricalaerialphotographsofthe increases as sediment accumulates on the plain, but Skagit River delta in Washington, USA, for example, the gradient decreases as river discharge increases. suggests that sand bars formed within tidal chan- Delta plains can have extensive vegetation cover nels are colonized by marsh vegetation, which forms that varies with climatic setting, rates of deposition, a marsh island that narrows the gap between the and—down the delta—salinity gradients. Some of island and the mainland marsh. This constricts the the world’s great deltas are covered in dense subtrop- tidal channel and eventually leads to coalescence of ical forest and mangrove swamps (e.g., the Sundar- the island with the adjacent mainland (Hood, 2010). bans at the mouth of the Ganges River). Other deltas Delta plains are more affected by river currents have primarily sedges, rushes, and marsh plants than by wave and tidal currents, and thus resem- (e.g., Mississippi River delta, Pearl River delta in ble inland floodplains (Bridge, 2003). Deposition 190 Rivers in the Landscape along the lower portions of tributary channels can be enhanced during periods of high tides or storm surges, leading to increased frequency of avulsion (Bridge, 2003). The delta front is the steeper, basinward por- Delta plain tion of a delta, where sediment-carrying river water Topset enters open water. High rates of deposition and lim- ited reworking by currents in the receiving basin can result in steepening of the delta front beyond the Bottomset Foreset Foreset angleofreposeofthesediment,causingasediment Bottomset gravity flow that can develop into a turbidity cur- rent. A turbidity current is analogous to a subaque- Younger delta Older delta ousdebrisflow,inthatitisarelativelydense,flowing Figure 6.6 Schematic block diagram of a Gilbert-type slurry of sediment that follows bottom topography or foreset-dominated delta, showing the three primary and can travel farther into the receiving basin than types of beds, the delta plain, bars, and distributary suspended sediment. Oceanographers first docu- channels. River flow enters a body of standing water from mented turbidity currents when slumps initiated by the right. the 1929 Grand Banks were transformed into a turbidity current that flowed down the slope of of water fluxes and bed shear stresses to varying sea the Atlantic Ocean basin, breaking submarine tele- levelandtheareacoveredbydeltaislands.Model graph cables en route (Heezen et al., 1954). results suggest that a decrease in island area or an Deposition on a delta front is episodic, as is increase in sea level results in increased tidal prism deposition on the basin floor beyond the delta. and bed shear stresses. Stratigraphy in a prograding delta that is progres- Fagherazzi and Overeem (2007) provided a thor- sively building into the receiving basin is charac- ough review of numerical models of delta dynamics. terized by bottomsets—fine-grained, low inclination Numerical models include: deposits of the basin floor; foresets—deposits of the r delta front at the angle of repose; and topsets— 2D cross-shelf models that focus on the evolution relatively flat-lying and coarse deposits of the delta of the continental shelf profile and ignore along- shelf variations in morphology; plain (Bridge, 2003) (Figure 6.6). G.K. Gilbert (1885) r introduced these terms, so deltas with this type 2D along-shelf models, which focus on the evolu- of stratigraphy are known as Gilbert-type deltas. tion of the along-shelf morphology and consider Edmonds et al. (2011) refer to Gilbert-type deltas either a constant cross-shelf profile or integrate across the shelf; as foreset-dominated deltas, and distinguish them r from topset-dominated deltas in which the distribu- pseudo-3D shelf area models based on depth- tary channels incise into pre-delta sediment. averaged formations of hydrodynamics and sed- iment transport; and As with other fluvial processes, delta dynam- r ics can be studied using physical experiments fully 3D models that reproduce the full structure (Edmonds et al., 2009) and numerical simulations of the flow field coupled with sediment transport. (Geleynse et al., 2011). Canestrelli et al. (2010) pro- vided an example of a numerical simulation used 6.4.3 Paleoenvironmental records to describe the propagation of the tidal wave within the delta and along the channel of the Fly River in The depositional environment that is a delta inte- Papua New Guinea. Using a high-resolution bathy- grates upstream changes in yields of water, sedi- metricmapofthedeltaandatwo-dimensional(2D) ment, and other materials. Consequently, delta mor- finite element model, they examined the sensitivity phology and stratigraphy are commonly examined Chapter 6 Extra-channel environments 191 as a source of information on catchment paleoenvi- of direct and indirect human alterations of water ronments. The length of paleoenvironmental record and sediment yield to the delta, and of the delta dis- that can be accessed in this manner partly depends tributary network. As noted earlier, many deltas are on the age and erosional history of the delta (in the densely populated and heavily farmed—close to half sense of erosion removing portions of the deposi- a billion people live on or near deltas, and an esti- tional record), and partly on the techniques used to mated 25% of the world’s population lives on deltaic examine deltas. Changes in sedimentary facies, min- lowlands (Tamura et al., 2012). These people are eralogy, and isotopic ratios, as well as changes in increasingly vulnerable to changes in delta morphol- location of delta sublobes, are used to infer changes ogy and dynamics. in depositional processes driven by either upstream Oneoftheprimarychangesissedimentcom- changes in water and sediment yield or changes in paction and delta subsidence, so that deltas are sink- coastal wave and tide energy (Hori et al., 2004; Hori ing more rapidly than sea level is rising (Syvitski and Saito, 2007; Liu et al., 2010). Chronologies of et al., 2009). Delta subsidence results from several delta deposition can be constrained using lumines- factors. Sediment consolidates with time as pore cence and radiocarbon techniques (Hori and Saito, water is expelled. Trapping of sediments and river 2007; Tamura et al., 2012). flow in reservoirs upstream from the delta reduces The Nile River delta provides a nice example ongoing inputs that would otherwise offset progres- (Krom et al. 2002). One of the Nile’s major tribu- sive compaction and subsidence. Removal of oil, gas, taries, the White Nile, drains predominantly crys- and water from delta sediments reduces pore pres- talline rocks. Other major tributaries—the Blue Nile sure and facilitates compaction. Floodplain engi- and the Atbara—drain the Tertiary volcanic rocks neering on the delta, including channelization, lim- of the Ethiopian Highlands. The different rock types its overbank deposition and conveys sediment more have different ratios of 87Sr/86SR and Ti/Al. Changes efficiently beyond the delta and into the ocean. And, in these ratios through time, combined with known rising sea level subjects deltas to more intense ero- changes in river discharge, indicate that periods of sion by wave and tidal currents, as well as storm higher river flow during the past 7000 years corre- surges. spond to decreased input of total sediment and sed- In China’s Changjiang (Yangtze) delta, for exam- iment derived from the Blue Nile watershed because ple, construction of approximately 50,000 dams of higher rainfall and increased vegetative cover in throughout the contributing watershed has resulted the Ethiopian Highlands (Krom et al., 2002). in erosion of the delta (Yang et al., 2011). Numerous Over shorter time periods, changes in delta mor- other examples of reduced sediment yield to deltas phologyandprocesscanbestudiedusingaerial and associated erosion have been documented (Jun photographs, satellite images, and repeat ground- et al., 2010; Nageswara Rao et al., 2010). Diversion based surveys. Tamura et al. (2010), for example, ofwaterandsedimentflowseitherawayfromthe used topographic maps and satellite images to study delta or to a different portion of the coastline exacer- changes in the shoreline of the Mekong River delta bates delta shifting, causing the abandoned delta to since 1936 (Tamura et al., 2010). They found that the compact, subside, and erode (Jabaloy-Sanchez et al., delta has changed asymmetrically over this period 2010; Restrepo and Kettner, 2012). in response to net southwest-ward transport of sed- Examining 33 deltas around the world, Syvitski iment associated with winter monsoons. et al. (2009) found that 85% of the deltas experi- enced severe flooding during the previous decade, causing the temporary submergence of 260,000 km2. 6.4.4 Deltas in the Anthropocene They also conservatively estimated that delta surface area vulnerable to flooding could increase by 50% As with other aspects of river networks, many deltas under current projections of sea level rise during the have recently undergone major changes as a result twenty-first century. 192 Rivers in the Landscape

Coastalwetlandsassociatedwithdeltasarealso In addition to changes in morphology caused by being lost to submergence or erosion. A 2005 sur- human alterations, delta sediments can be locally vey of 42 deltas across the globe indicated the loss of contaminated by toxic synthetic chemicals such as nearly 16,000 km2 of wetlands during the preceding pesticides (Feo et al., 2010) or petroleum byprod- 20 years, with an average rate of loss of 95 km2 per ucts (Wang et al., 2011). Any portion of a river year (Coleman et al., 2005). network can be affected by such contaminants, but Mining of construction aggregate, land reclama- deltas are particularly vulnerable because they inte- tion, and changes in delta morphology can also grate diverse upstream sources of contamination. influence the pathways of water and sediment across Contaminants can affect the delta environment and, the delta, as well as delta submergence and the abil- when remobilized by wave or tidal erosion, can be ity of waves and tides to rework delta sediments. spread into the nearshore environment. Naturally Large-scalesandexcavationfromChina’sPearlRiver occurring contaminants, such as arsenic associated delta since the mid-1980s, for example, has resulted with highly reducing deltaic sediments in the Bengal in increased water depth and lowered streambed and Mekong River deltas, can also strongly influence elevation (Zhang et al., 2010). This has led to an delta environments (Berg et al., 2007). increased tidal prism and upstream movement of the tidallimit,whichhavefacilitatedincreasedsaltwater intrusion into the . 6.5 Estuaries Some deltas changed little during the twentieth An estuary represents an incised river valley along century and their aggradation rate remains in bal- the coastline of an ocean (Figure S6.12). The river ance with, or exceeds, subsidence or relative sea- valley incised during a period of lower sea level that level rise. Examples include deltas on the Amazon caused base level lowering for the river, typically (Brazil), Congo (Democratic Republic of Congo), during a period of continental ice sheet formation. Orinoco (Venezuela), Fly (Papua New Guinea), and Subsequent sea level rise during ice sheet melting Mahaka (Borneo) Rivers (Syvitski et al., 2009). and retreat backflooded the estuary. The typical sce- These deltas typically occur on rivers with lower narioisthattheestuarygrowsprogressivelyshal- density of human settlement in the upstream lower and less subject to marine influence as ter- drainage basin and minimal flow and sediment reg- restrial and marine sediments accumulate. An estu- ulation. Deltas with minimal human manipulation ary that is a sediment sink is also referred to as a provide a control against which to evaluate the microtidal estuary (Cooper, 2001). Usually, estuar- effects of direct human alteration of deltas within the ine filling occurs during and/or shortly after back- context of changing climate and sea level. flooding on sediment-rich rivers. Estuary filling can Conversely, the Pearl Delta in China and the require time periods longer than the Holocene on Mekong Delta in Vietnam are among the most at- rivers with very low sediment loads, in settings in risk deltas. Each of these deltas is inhabited by mil- which river floods periodically scour the estuary, lions of people, is exposed to typhoons, and has lim- or in settings in which sediment is primarily trans- ited coastal barrier protection. Each of these deltas ported through the estuary to the marine environ- also has compaction and/or reduced aggradation, so ment (Cooper, 2001; Cooper et al., 2012). that much of the delta surface area is below mean sea Numerous factors influence the form and pro- level (Syvitski et al., 2009). cesses of estuaries (Dalrymple and Choi, 2007). Human-induced changes are also exacerbated Among the most fundamental are: where very little or no river flow now reaches the delta as a result of consumptive water uses upstream. thebathymetryoftheestuary,whichrangesfrom Deltas subject to this severe change in flow regime a relatively shallow-water, channelized environ- includethoseontheIndusRiverandtheColorado mentlandwardofthecoast,todeeper,unconfined River in the United States and Mexico. settings on the shelf, Chapter 6 Extra-channel environments 193 the source of the energy that moves sediment, from Dalrymple et al. (1992) distinguished simply predominantly river currents to tidal, wave and wave- and tide-dominated estuaries. Wave- oceanic shelf currents, dominated estuaries typically have three well- the frequency, rate and direction of sediment move- definedportions:amarinesandbodycomposedof ment, which vary from barrier, washover, tidal inlet and tidal delta deposits; r unidirectional, continuous movement to sea- a fine-grained central basin; and a bay-head delta sonal or episodic when driven by rivers, that experiences tidal and/or salt-water influences r reversing in tidal settings, (Figure 6.7). Examples of wave-dominated estu- r episodic and coast-parallel when driven by aries include the Hawkesbury Estuary and Lake waves, and Macquarie in Australia, eastern shore estuaries in r onshore–offshore in tide-dominated shelf envi- Nova Scotia and the Miramichi River in Canada, ronments, and and San Antonio Bay and Lavaca Bay in the United water salinity, which ranges from fresh, through States (Dalrymple et al., 1992). Tide-dominated brackish, to fully marine. estuaries have a marine sand body composed of elongate sand bars and broad sand flats; a Given this breadth of river and coastal influences, central zone of tight meanders where bedload estuaries exhibit significant diversity. Most estuaries, is transported by flood-tidal and river currents; however, possess a three-part structure composed and an inner, river-dominated zone with a single, of (i) an outer, marine-dominated portion where low-sinuosity channel. Examples of tide-dominated the net bedload transport is headward, (ii) a rela- estuaries include the Adelaide and Ord Rivers in tively low-energy central zone with net bedload con- Australia, the Cumberland Basin and Avon River vergence, and (iii) an inner, river-dominated (but in Canada, and Alaska’s Cook Inlet in the United still marine-influenced) portion where net sediment States (Dalrymple et al., 1992). transport is seaward (Dalrymple et al., 1992). Each Estuaries share some characteristics with deltas zone is developed to differing degrees among diverse in that they include depositional surfaces on which estuariesasareflectionofsite-specificsediment tidally influenced channels form. Numerous stud- availability, gradient of the coastal zone, and the ies demonstrate that these tidal channels are fun- stage of estuary development. nel or trumpet shaped, such that channel width Various classifications exist for estuaries. Exam- tapers upstream in an approximately exponential ining estuaries in South Africa, for example, Cooper fashion (Chappell and Woodroffe, 1994; Fagherazzi (2001) distinguished five types of estuaries based and Furbish, 2001; Savenije, 2005). The shape of the on contemporary morphodynamics. Normally open exponential width profile relates to the mouth width estuaries maintain a semi-permanent connection and river output. Channels and estuaries with larger with the open sea. The three primary categories mouths have a more pronounced funnel shape than of open estuaries are barrier-inlet systems main- systems with narrower mouths, and higher river tained by river discharge (river-dominated estuar- water and sediment discharge create a longer “fun- ies) and tidal discharge (tide-dominated estuaries), nel” (Davies and Woodroffe, 2010). and open estuaries that lack a supratidal barrier Asinthecaseofdeltas,humancoastalsettle- becauseofinadequateavailabilityofmarinesedi- ment and resource use concentrate around estuar- ment. Closed estuaries are separated from the sea ies. Estuaries play an important role in the cycling for long periods by a continuous supratidal bar- of carbon and other nutrients through the exchange rier. Perched closed estuaries develop behind high and modification of terrestrial organic matter trans- berms and maintain a water level above high tide ported by rivers to the ocean (Canuel et al., 2012). level, whereas non-perched closed estuaries develop Because estuaries typically retain organic matter for behind barriers of lower elevation with wide beach some period of time, these environments have high profiles. rates of primary production and support abundant 194 Rivers in the Landscape

(b) (a) Tidal fluvial Tidal sand UFR channel bars sand flat

Bay-head Tidal delta limit Straight Sinuous Straight Tidal Shallow Mudfalt limit marine Central

Barrier basin Flood- Alluvial tidal valley Alluvial Inlet Salt marsh delta valley

Mean sea level Mean high tide Bay-head delta Barrier/ Flood-tidal Shore- Central basin Alluvial tidal inlet delta face Tidal sand Tidal sediments Alluvial bars Tidal fluvial sediments UFR meanders sand flat

UFR = upper flow regime Figure 6.7 Distribution of morphological elements in plan view and sedimentary facies in longitudinal section within an idealized (a) wave-dominated and (b) tide-dominated estuary. For (a), the barrier/sand plug is shown as attached to the headland, but may not be connected on low-gradient coasts, where it can be separated from the mainland by a lagoon. The longitudinal section represents the onset of estuary filling after a period of sea level rise. For (b), the longitudinal section is along the axis of the channel and does not show all sedimentary facies present. (From Dalrymple et al., 1992, Figures 4 and 7.) and diverse wetlands and fisheries. Dramatically 6.6 Summary increasedfluxesofnitrogentoestuariesasaresultof livestock wastes, industrial-scale application of agri- The fluvial landforms discussed in this chapter— cultural fertilizers, and other human activities have floodplains, terraces, alluvial fans, deltas, estuaries— caused many estuaries to experience eutrophication are inherently dynamic and rapidly changing envi- (Green et al., 2004). Eutrophication is an ecosystem ronments. Ever-shifting balances among different response to increased nutrients. This response typi- processes of deposition and erosion can reflect cally involves a substantial increase in phytoplank- changes external to the landform such as variation ton and associated decreases in dissolved oxygen, in climate or base level. Shifting erosion and depo- which together cause hypoxia, or the so-called dead sition can also reflect internal adjustments within zones. At least 375 hypoxic coastal zones have been the environment of the landform while the external identified around the world, primarily around west- setting remains relatively stable, as when continued ern Europe, the eastern and southern coasts of the development of levee crevasses results in delta shift- United States, and Asia (CENR, 2000). ing. This continual variability can make fluvial land- Many of the same techniques used to study deltas forms difficult to interpret, but floodplains, terraces, are applied to the study of process and form in estu- fans, and deltas, in particular, are nonetheless valu- aries. Bathymetric mapping of submerged features is able repositories of information on river dynam- particularly useful in studying estuaries, and naviga- ics over time spans of tens to tens of thousands tional charts for individual estuaries can date from of years. the early nineteenth century (van der Wal and Pye, Humanactivitiesincreasinglyalterprocessand 2003). form in these extra-channel environments. Indirect Chapter 6 Extra-channel environments 195 effects associated with climate change and rising sea biological productivity and ecological sustainability, level combine with direct effects including flow reg- and increased hazards to human populations and ulation, levee construction, and channelization to infrastructure as a result of eroding landforms. The change fluxes of water, sediment, solutes, and organ- greatchallengeofthefutureistounderstandhow isms between channels and extra-channel environ- past human actions have changed river networks, ments. Altered fluxes—altered connectivity—have and how undesirable changes can be mitigated or in many cases led to reduced water quality, lower reversed as we go forward.

Chapter 7 Humans and rivers

Each of the preceding sections of this book has networks goes back more than a century (Marsh, included some mention of how people alter drainage 1864; James and Marcus, 2006). basin and channel process and form. Because human Direct human impacts result from activities influence is now so ubiquitous and, in many cases, within the river network that directly alter channel intensive, this chapter focuses specifically on sev- form or process. Flow regulation, and specifically eral aspects of human activities that exert an impor- the construction of dams, has received the most tant influence on river process and form. One of the scientific and public scrutiny, in part because indi- definitions for “impact” in the Oxford English dic- vidual dams can be extremely large and can alter tionary is the “strong effect of one thing, person or physical, chemical, and biological characteristics action,onanother;aninfluence.”Thisisthesense of entire watersheds. Other, typically smaller-scale in which this chapter discusses human impacts on alterations of channel form and connectivity— rivers. associated with flow diversion, construction of Human impacts can be indirect or direct. Indi- levees, dredging and channelization, bank stabi- rect human impacts on rivers result from activi- lization, and instream mining of placer metals and ties that occur outside of the river network and do sediment for construction aggregate—can have not directly alter channel form or process. Human- cumulative effects equally substantial to those induced changes in atmospheric chemistry, and the associated with large dams. Most studies of these resulting changes in atmospheric and oceanic cir- impacts, however, focus on individual or regional culation, and temperature and precipitation pat- examples, and there is not yet an effective analysis terns, are one category of indirect impacts that have of the cumulative, global effects of these activities. received increasing attention within the past two Thethirdportionofthechapterexaminesriver decades. These changes have been underway for engineering designed specifically to reverse or miti- more than a century, but are receiving widespread gate the negative consequences of earlier river engi- attention only now. Land cover is strictly defined as neering. River restoration or rehabilitation attempts the type and density of vegetation, although alter- to alter river process and form in a manner per- ations in vegetation are commonly accompanied by ceived as beneficial for fisheries, recreation, esthet- alterations in topography and soil characteristics. ics, or some other characteristic not considered in Changes in land cover have been occurring for thou- prior engineering. Instream flows, channel mainte- sandsofyears,andrecognitionoftheeffectsonriver nance flows, and environmental flows are a form of

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

197 198 Rivers in the Landscape river engineering that focuses primarily on the flow polar ice sheets, record increasing atmospheric con- regime as a means of protecting or restoring river centrations of CO2 and other compounds that per- process and form. mit a given volume of air to absorb more infrared

The chapter concludes with a discussion of river radiation. Increased CO2 levels have caused mea- health, as defined from ecological and geomorphic surable increases in air temperature since the start perspectives. River health is an intuitively appeal- of systematic measurements, and particularly since ing concept that evaluates the cumulative effects the 1950s. International scientific panels estimate of numerous human-induced alterations in river that continued increases in CO2 will cause a rise ecosystems and provides a conceptual framework in average global temperature during the twenty- for undertaking river restoration. first century of anywhere from 1◦ to 5◦C (IPCC, Humans impact landscapes through the effects 2008). Increases in atmospheric temperature cas- of economic activities, including management cade through atmospheric, oceanic, freshwater, and intended to mitigate natural hazards, but land- terrestrial systems as changes in numerous charac- scapes also impact humans through patterns of teristicsofthesesystems(Table7.1;Figure7.1).Even resource availability and natural hazards. Werner a brief consideration of the extent and complexity of and McNamara (2007) modeled these interactions local to regional changes in river process and form as hierarchical complex systems that provide associated with global changes in air temperature insight into how human manipulation of landscape reveals the difficulty of predicting and adapting to processes both influence and respond to landscape these changes on more than a relatively crude, basic process and form. level. The concept of landscape sensitivity (Brunsden For river networks, changes in precipitation and Thornes, 1979) is also important in thinking associated with changing atmospheric temperature about humans and rivers. A sensitive landscape is are the most immediate manifestation of climate likelytorespondtoachangeinexternalcontrols.A change. Most studies of the effects of climate change resilient landscape is likely to return to initial con- on rivers consequently start with an examination of ditions following a disturbance. Whether examin- how altered precipitation characteristics will likely ing how rivers respond to climate change, deforesta- influence a river’s flow regime. Predictions of future tion, or river restoration, the idea that some river precipitation commonly rely on general circulation networks or some portions of a river network are models (GCMs). At present, these models typically more sensitive or resilient (Section 4.3.2) is critical haveaspatialresolutionbasedongridcellsassmall to understanding and predicting river adjustment as 12,100 km2. They can thus be used for examining (James and Lecce, 2013). River segments with very large river basins, but smaller spatial scales require high sediment transport capacity may adjust rela- either statistical down-scaling to relate local climate tively quickly and easily to increased sediment yield, variables to large-scale meteorological predictions for example, whereas portions of the network that (Gyalistras et al., 1997; Andreasson et al., 2003), or are transport limited may exhibit substantial change a more detailed regional model that is nested within in channel geometry in response to a similar magni- the GCM (Giorgi et al., 1994; Marinucci et al., 1995). tude of increased sediment yield. Inferred precipitation characteristics must be cou- pled with simulations of soil moisture, land cover, infiltration, and runoff, as discussed in Chapter 2. 7.1 Indirect impacts Examples of the complexity of such modeling efforts come from studies in mountainous regions, 7.1.1 Climate change where even a small increase in air temperature can significantly influence the distribution, volume, Systematic sampling of atmospheric gases during snow water equivalent, and snowmelt timing of thepasthalfcentury,aswellasairbubblestrappedin mountain snowpacks (Lopez-Moreno´ et al., 2009; Chapter 7 Humans and rivers 199

Table 7.1 Examples of how changing global climate will impact rivers.

Description Reference

Because temperature and precipitation thresholds influence the magnitude and frequency of mass Lu et al. (2010) movements, climate change is increasing sediment yields to mountain rivers in the Himalayan and Tibetan Plateau In the semiarid/arid Western United States, changes in air temperature, moisture, and Hauer et al., 1997 thunderstorms alter the frequency and severity of wildfires, which causes changes in water, Westerling et al. sediment and nutrient yield to streams, as well as changes in forest structure and wood (2006) recruitment to streams; predictions include loss of floodplain forests Rood et al. (2008) In Alaska, USA, warming increases glacial runoff, which changes flow regime, water temperature Oswood et al. and sediment discharge, causing changes in channel substrate, bedforms, channel stability, leaf (1992) litter quantity and quality, and habitat complexity Along arid-region rivers, changes in flood magnitude and frequency can cause channel incision Grimm and Fisher that removes the deep hyporheic sediments that support microbial communities: changes in (1992) precipitation and runoff can also alter the availability of nitrogen, which is a limiting nutrient in these rivers As the proportion of precipitation coming in the form of rain (rather than snow) increases at high Eckhardt and elevations, winter rains increase flood hazards and decrease groundwater and summer Ulbrich (2003) streamflow by up to 50% in the mountains of central Europe Droughts in the United Kingdom during the late twentieth century highlighted the sensitivity of Wilby (1995) water resources to climatic fluctuations; predictions suggest that rainfall is likely to become more intense and less frequent in future In Australia, changes in rainfall will create the greatest runoff changes in arid catchments; some Chiew and McMahon regions will receive more runoff (more intense and frequent rainfall in eastern Australia), others (2002) will receive less; frequency and severity of droughts will also increase Changes in temperature and hydrology that promote disturbances to vegetation such as wildfire, Goode et al. (2012) insect outbreaks, and drought-related die off are predicted to increase sediment yield to rivers in the northern US Rocky Mountains; increased sediment yields will affect aquatic habitat and downstream water-storage reservoirs Increasing air temperatures in the monsoonal Mahanadi River basin of India during the twentieth Rao (1995) century correlate with declining river flows

Gillan et al., 2010). Nonlinear responses compli- ica, India, and Pakistan, disappearance of mountain cate attempts at modeling and prediction. Snow- glaciers will substantially reduce water supply and pack on a glacier, for example, reduces absorbed likely cause social upheaval (Hasnain, 2002; Barnett solar radiation and melt rate (Oerlemans and Klok, et al., 2005; Singh et al., 2006). 2004). Removal of the snowpack can increase daily Changes measured during the past few decades discharge amplitude by more than 1000% (Willis canprovideinsightintosuchmulti-facetedalter- et al., 2002). Upward retreat of rainfall–snow ele- ations. Shifts in large-scale atmospheric circula- vation limits may thus substantially increase rates tion have resulted in altered flood frequency across of glacial melting. Changes in glacial mass balance Switzerland (Schmocker-Fackel and Naef, 2010). are particularly important in mountainous regions Earlier snowmelt, reduced snow accumulation, and because even relatively small alpine glaciers pro- altered streamflow are particularly well documented vide large reserves of freshwater that sustain sum- for the European Alps, Himalaya, and western North mer peak flows and autumn base flows. In populous, America (Kundzewicz et al., 2007). Regional-scale relatively dry lowlands such as parts of South Amer- studies document how differences in characteristics 200 Rivers in the Landscape

Increasing atmospheric temperature

Atmospheric

Freshwater Terrestrial

Systems Oceanic

Changing

Atmospheric Ocean currents and circulation exchanges of heat and moisture systems between ocean and atmosphere

Magnitude, Temperature, Rainfall– Type and timing and moisture and runoff extent of type of precipitation infiltration capacity relations vegetation of soils communities

Magnitude, Volume and Ground water Sediment Biological timing and spatial extent and hyporheic transport productivity duration of surbsurface water exchanges in channels of rivers of streamflow along rivers

Figure 7.1 Schematic illustration of changes affecting rivers as a result of increasing atmospheric temperature. such as subsurface drainage and seasonal distribu- higher latitudes (Nijssen et al., 2001). Understanding tion of precipitation influence between-watershed and preparing for the implications of these changes differences in sensitivity of streamflow to climate at the scale of individual drainage basins requires warming.IntheOregonCascadeRangeofthe a great deal more effort, but it is worth emphasiz- western United States, for example, differences in ing that hydroclimatic changes that influence river bedrock geology between watersheds correspond to ecosystems are not hypothetical. Many regions of the differences in volume and seasonal flux of subsur- world already exhibit statistically significant differ- face water (Tague et al., 2008; Tague, 2009). Water- ences between the latter half of the twentieth cen- sheds dominated by extensive, low-relief basaltic tury and earlier periods (Marchenko et al., 2007; Van lava flows have deeper groundwater flow and are Der Schrier et al., 2007; Rood et al., 2008). As of predicted to show greater absolute reduction in 2012, successive records for the smallest June snow summer streamflow under expected temperature cover extent have been set each year in Eurasia since increases than are watersheds dominated by shallow 2008, and in 3 of the past 5 years in North America subsurface flow, where stream flow is already flashy (Derksen and Brown, 2012). and highly seasonal. Atthebroadestscale,modelingsuggeststhatpre- 7.1.2 Altered land cover cipitation totals will increase in high latitudes and in mid-latitudes during winter. Precipitation variabil- Alterations in land cover began thousands of years ity and possibly the frequency of extreme precipita- agowhenpeoplebegantogrowcrops,domesticate tion will increase in the tropics. Snow cover dura- grazing animals, or even use fire to alter vegeta- tion will decrease and mid-latitude soil moisture tion communities in favor of plants preferred by may decrease in summer (Kattenberg et al., 1996; animals that people hunted. Primary types of land- Mote et al., 2005). The largest hydrological changes cover alterations include deforestation, afforesta- are predicted for snow-dominated basins of mid-to- tion, grazing, crops, urbanization, mining, wetland Chapter 7 Humans and rivers 201 drainage,and,morelocally,commercialrecreational transpiration and precipitation (Costa, 2005; Wohl development. et al., 2012a). At present, the three main humid tropical forest regions of South America, Africa, and southeast Asia have particularly accelerated and Deforestation widespread deforestation (Drigo, 2005). People have reduced global forest cover to about Sediment yields increase as mineral soils exposed half of its maximum extent during the Holocene and compacted during deforestation become more and have eliminated most old-growth forests (Mont- susceptible to erosion via overland flow, rilling, and gomery et al., 2003a). Forestry (wood harvest, with landslides and debris flows. Bank erosion, headward anticipated regrowth of the forest) is now prac- expansion of channels, and windthrow of remain- ticed in many industrialized countries, but histori- ing trees can also increase sediment yields. Roads cal deforestation was seldom conducted in a manner decrease hillslope stability by redistributing weight, conducive to forest regrowth, and minimal attention changing surface angles, reducing infiltration, alter- is currently paid to regrowth in many areas experi- ing conveyance via culverts under roads, and initiat- encing deforestation in developing countries. Con- ing gullies (Figure 7.2). Unpaved roads continue to sequently, the discussion that follows here applies contribute increased fine sediment yields to chan- primarily to clearcutting, or complete removal of nels long after harvested vegetation has regrown forest cover over areas of varying spatial extent, (Jones et al., 2000). Water yields increase as removal rather than to selective cutting of a limited number of vegetation reduces interception and transpiration, of trees within a forest. and compaction decreases infiltration. Changes in An extensive literature documents the effects of sediment and water yield cause changes in stream- deforestation on rivers from mountains to lowlands flow, stream chemistry, and channel morphology. and from the boreal regions to the tropics. Thorough Dependingonthemagnitudeandtimingofalter- reviews are provided by Foley et al. (2005), Scanlon ations in sediment and water yields, numerous chan- et al. (2007), Douglas (2009), and Wohl (2010). Cut- nel changes can result indirectly from timber har- ting of trees and the associated building of roads typ- vest and road building (Figure 7.3). Changes in ically greatly increase hillslope sediment yield over channelmorphologycanpersistforatleastsev- a period of a decade or less, and increase water eral decades (Madej and Ozaki, 2009). Associated yield over periods of multiple decades until vege- processes such as recruitment of instream wood tation recovers. Deforestation in the tropics, how- may require two centuries to return to pre-cutting ever, can decrease water yield because of reduced levels because trees must grow to maturity and

(a) (b)

Figure 7.2 Two of the ways in which roads influence sediment yields. (a) In this view of a paved road, the steep face cut into the hillslope above the road bed has destabilized the slope and caused mass movement. (b) Small gullies forming in this unpaved road enhance sediment yield to adjacent streams. 202 Rivers in the Landscape r Increased flooding the connectivity between deforested areas and channels—this can be strongly influenced by the Increased solute fluxes extent and placement of roads, which are major Increased water temperatures sources of sediment and can provide effective con- from loss riparian shading and subsurface flow inputs duits for water and sediment from deforested areas toadjacentchannels,asshownbygullydevelop- Altered Bank erosion ment from road outlets to streams in southeastern water and Loss of pool habitat and instream Australia (Croke and Mockler, 2001). sediment wood yields Aggradation Clearcutting an entire drainage causes major changes in water and sediment yield and rivers. Increased potential for overbank flow and channel change during floods Selective cutting, with riparian buffer strips of unal- tered vegetation left in place, minimizes change. Increased substrate mobility

Bed scour Afforestation

Figure 7.3 Potential channel responses to timber har- Afforestation, or the regrowth of forest cover, vest and associated road building as a result of changes whether natural or human induced, can reverse in water and sediment yield entering the river network. some of the trends created by deforestation. Doc- umented effects of afforestation include increased then be recruited into and retained within channels infiltration and base flow, decreased runoff and (Bragg et al., 2000). sediment yield, and channel narrowing and deep- Thedegreetowhichwaterandsedimentyields, ening. The magnitude and timing of these changes and thus rivers, are altered depends on can be complicated by continuing remobilization of sediment eroded during deforestation and stored in colluvial and alluvial features within the drainage r the portion and spatial distribution of defor- basin (Larsen and Roman,´ 2001). If afforestation ested areas within a drainage basin—logging at occursasaresultofdeclinesinpopulationinregions lower elevations in catchments of mountainous, that have been extensively deforested for centuries, snowmelt-dominated catchments in southwestern as is now occurring in portions of the European Canada caused little or no change in peak flow Alps, channels that had become stable under former because of relatively small low-elevation snow- land use patterns can become unstable under newly packs, for example, whereas logging in the upper changing water and sediment yields (Latocha and elevations of these catchments resulted in substan- Migon,´ 2006). tial increases in peak flow (Whitaker et al., 2002); r the methods used to harvest trees—clearcutting Grazing typically results in substantially larger effects on water and sediment yields than selective logging, Upland grazing can have many of the same effects as as shown, for example, by Kasran (1988) in Penin- timber harvest. These include reduced and altered sular Malaysia; vegetation cover, soil compaction, reduced infil- r site characteristics including climate, topogra- tration, increased runoff and sediment yield, and phy, and soils—subcatchments most responsive associated changes in streamflow, channel morphol- to changes in water and sediment yield within ogy, and channel stability. The more intense and the Dromeˆ River basin in France, for example, widespread the grazing, the more severe are these were high-energy environments of high elevation, effects. Upland grazing can be practiced with min- high relief ratio, and abundant sediment sources imal effect if the density of grazing animals is kept (Liebault´ et al., 2002); and below a level that negatively impacts vegetation type Chapter 7 Humans and rivers 203 and density, and does not compact soils. Over- magnitude and timing (Chin, 2006; Gurnell et al., grazing characterizes mountain livestock husbandry 2007; Chin et al., 2013). worldwide (Hamilton and Bruijnzeel, 1997), how- Removal of vegetation and leveling or artifi- ever, leading to land degradation and altered river cially contouring the land surface during the ini- process and form. tial phase of construction dramatically elevates sed- iment yields and causes aggradation and planform Crops changes in receiving channels. Completion of con- struction stabilizes ground surfaces beneath roads, As with deforestation and upland grazing, the plant- buildings, and lawns, causing sediment yield to ing of crops alters soil infiltration capacity, soil decline to a negligible value, and triggering bed exposure and erosion, and thus water and sedi- coarsening, incision, and bank erosion in receiv- ment yields to channels. Compacted footpaths and ing channels. Water yield increases as impervi- roads around crop lands exacerbate these effects. ous surface area within the contributing basin Land drainage associated with crops can increase or increases. Storm sewers rapidly drain urban areas, decrease runoff. further decreasing infiltration and conveyance time The magnitude of alterations in water and sedi- of water, creating flood peaks of shorter duration ment yield depends on the type and extent of crops, and higher magnitude for at least small to mod- as well as topographic and soil characteristics at the erately sized precipitation inputs, and exacerbating site. Not all crops are created equal. Replacement of channelerosion.(Wherestormsewerdrainageisfed grains with potatoes in nineteenth-century Poland, into the sanitary sewer system, or into buffer areas for example, increased sediment yields and flood planted with vegetation, these effects can be mini- peakstothepointthatmeanderingchannelsbecame mized.) The magnitude, extent, and speed of change braided (Klimek, 1987). in channel networks reflect factors such as the per- The most common effect of planting crops is centageofthecatchmentthatbecomesimpervious, increased sediment yield and associated changes the connectivity and conveyance of impervious sur- in channel pattern and stability. This effect can be faces, and the characteristics of the receiving chan- documented in sedimentary records of prehistoric nels (Bledsoe and Watson, 2001). land use (Mei-e and Xianmo, 1994). Compilation Urbanizationtypicallyalsoinvolvesdirectalter- of numerous case studies indicates that soil erosion ation of channels via channelization, bank protec- under conventional agriculture exceeds rates of soil tion, and other engineering, as well as substantial production and geological erosion by up to several changes in water chemistry via contaminants mov- orders of magnitude, a situation that is clearly unsus- ingwithrunoffandsediment.Headwaterchannels tainable with respect to soil fertility (Montgomery, are commonly paved or otherwise filled, or diverted 2007). Crops can also increase nutrient fluxes to into sewer systems, effectively eliminating this por- channels as soil nutrients are lost and excess fertiliz- tion of a river network. ers are mobilized from uplands (CENR, 2000; Wang et al., 2004; Boyer et al., 2006). Upland mining Urbanization Upland mining that occurs within a watershed but Wolman’s (1967a) classic study of the effects outside of the channel and floodplain can take the of urbanization on sediment yield and channel form of hard-rock mining of metals or surface or response delineated a sequence of changes within subsurface excavation of coal, construction aggre- a small watershed in the Piedmont region of gate, or building stone. Extensive surface disrup- Maryland, USA (Figure S4.25). Subsequent studies tion or creation of large tailings piles can increase fromthetropicstothehighlatitudesdocument sediment yield to channels (Harden, 2006). Alter- similar sequences that differ only in the details of ationofrunoffpathwayscanincreasepeakrunoff 204 Rivers in the Landscape

(a) (b) (c)

Figure 7.4 Headwater channels in West Virginia, USA, a region with extensive coal mining and mountain-top removal. (a) A stream in an unmined catchment typically has intermittent or ephemeral flow. (b) A channel down- stream from a mining area with valley fill is more likely to have perennial flow, and the increased hydraulic forces result in channel widening, coarsening, and steepening. The channel boundaries become less complex, habitat diver- sity declines, and disturbances in the form of large flows and bed mobility become more common. (c) Constructed, terraced slopes and “channels” lined with riprap in a mined headwater area. Photographs courtesy of Kristin Jaeger.

(McCormick et al., 2009). Treatment of ores or Justasmostpeopletodaycannotreallyimaginehow concentration of naturally occurring toxic contam- much wood was historically in rivers and how many inants can substantially alter the chemistry of sur- rivershadamulti-threadplanform,wehavetrou- face and subsurface waters and fluvial sediments ble understanding how extensive various types of (Stoughton and Marcus, 2000). Extreme examples wetlands—marshes, swamps, fens, and mires—once of upland mining impacts come from the practice were, and how much their drainage and loss has of mountain-top removal in coal-bearing regions of altered water and sediment dynamics across large the eastern United States (Palmer et al., 2010a), in regions (Vileisis, 1997). which material overlying coal seams is removed to Field drainage systems are designed to control vertical thicknesses up to 300 m and dumped into surface water and the water table via surface features adjacent headwater valleys, obliterating surface flow such as bedded systems used on flat lands grow- and valley topography (Figure 7.4). ing rice or graded systems used in sloping lands forothercrops.Fielddrainagecanalsousesub- Land drainage surfacefeaturessuchashorizontalorslightlyslop- ing channels or trenches, wells with pumps, and Land drainage has been undertaken for centuries in buried pipe drains. In many cases, an extensive net- many parts of the world in order to decrease the work of collector and main drains exists to trans- extent of standing water at the surface, lower the fer water to a gravity outlet structure or a pumping water table, and improve access to low-lying lands station. Drainage can result in soil compaction and for agriculture and settlement. The English, work- increased runoff, soil erosion, and suspended sedi- ing with Dutch engineers, had drained more than ment transport, although the changes are not neces- 38,000 ha in the eastern counties of England (known sarily substantial (Walling et al., 2003). as the Fens) by 1649 (Simco et al., 2010). George Washington’s library included a copy of “Practical Commercial recreational property development Treatise on Draining Bogs and Swampy Ground,” first printed in 1775 (Stephens and Stephens, 2006). Commercial ski resorts and other spatially extensive Extensive wetland drainage occurred in southern changes in land cover undertaken for recre- Europe during the seventeenth and eighteenth cen- ational purposes, such as golf courses, can involve turies in an attempt to control malaria, which was deforestation, road construction, alteration of then believed to result from “miasma,” or the moist topography, and transfer and application of large air coming from marshes and swamps (Davis, 2006). volumes of water and/or pesticides. Such activities Chapter 7 Humans and rivers 205 can alter water, sediment, and nutrient and contam- lation. Dams have been built since circa 2800 BC inant yields to nearby channels, resulting in channel (Smith, 1971) for diverse purposes, including water change (Keller et al., 2004; David et al., 2009). Very supply, flood control, navigation, and hydroelectric few studies, however, have systematically evaluated power generation. The construction of dams larger the effects of commercial recreational property than 15 m tall has accelerated substantially since development on river networks. the 1950s (Goldsmith and Hildyard, 1984; Nilsson et al., 2005). Although relatively few large dams are An important consideration in reviewing the effects now built in Europe and North America, numer- of changes in land cover on river process and form ous large dams are being built and are proposed for is that very few locations have had only one change rivers in Africa, Asia, and South America. Multiple in land cover through time. Most river networks dams are under consideration in the few large river have experienced multiple changes that overlapped basins not yet extensively regulated, including the in time and space. This complicated history, com- Amazon and the Congo (Wohl, 2011a). Although bined with factors such as equifinality, thresholds, dams are sometimes promoted as an environmen- lag times, and complex response, can make it very tally benign, “clean” source of hydroelectric power, difficult to decipher cause and effect in relation to flow regulation nearly always substantially disrupts any particular past land-cover alteration, or to pre- physical process and form, water chemistry, and eco- dict responses to ongoing or likely future alterations logical communities along rivers. in land cover. As reviewed by Petts and Gurnell (2005), geomor- Acommonresearchtechniqueistocompareoth- phologists began to pay increasing attention to the erwise similar rivers with and without a particular effects of dams in the late 1960s and into the 1970s land-cover change in order to detect the influence (Wolman, 1967b; Gregory and Park, 1974; Petts, of the land-cover change on river process and form. 1979). Coupled with advances in measurement tech- The river without human impacts reflects reference niques and process-based studies, this led to a series conditions (Section 7.3.1). This paired watershed of influential papers during the 1980s on the geo- approach is effective to the extent that other poten- morphic effects of dams on river networks (Petts, tial control variables for river process and form can 1984; Williams and Wolman, 1984; Carling, 1988). be held constant. Where the absence of unaltered The specific effects of flow regulation depend on drainage basins makes such comparisons infeasible, the character of the alterations in water and sed- paleoenvironmental records of past changes (Sup- iment discharge and on the characteristics of the plemental Section 3.2.1) or numerical modeling of channel (Grant et al., 2003; Salant et al., 2006; Do river response to altered water and sediment yield Carmo, 2007; Magilligan et al., 2013). The presence canbeusedtoinfercauseandeffect.Understand- of a dam typically ing of process domain or river style (Section 8.2) can also be used to infer reference conditions and the r reduces the mean and the coefficient of variation evolutionary trajectory of a river (Fryirs et al., 2012) of annual peak flow—analysis of 29 dams in the in the absence of a single, highly similar, reference central and western United States, for example, watershed. revealed decreases in average annual peak flows downstream from the dam that ranged from 3% to 90% (Williams and Wolman, 1984); 7.2 Direct impacts r increases minimum flows, as illustrated by por- tions of the Tennessee and Columbia Rivers in the 7.2.1 Flow regulation United States, where annual 7-day low flows have nearlydoubleddownstreamfromdamswithsub- As discussed briefly in Chapter 3, most of the world’s stantial amounts of reservoir storage (Hirsch et al., rivers are affected to some degree by flow regu- 1990; 206 Rivers in the Landscape r shifts the seasonal flow variability—the Colorado High (>>1) Textural shifts at confluences River below Glen Canyon and Hoover Dams in Island and bar construction Vegetation encroachment Arizona, USA, for example, shifted from a pre- Channel aggradation regulation May and June snowmelt peak flow to a broad peak from April to September that corre- sponds to the period of maximum irrigation and municipal water-supply demands (Hirsch et al., Effects subtle – supply above dam S*, ratio of sediment supply below dam to Bed scour 1990); and depend on channel history r Armored channel can greatly increase diurnal flow fluctuations if Bar and island erosion the dam is operated for hydroelectric power gen- Low (<<1) Channel degradation, narrowing 0 T*, fractional change in frequency of High (>1) eration, as illustrated by the Connecticut River sediment-transporting flows below Wilder Dam in the northeastern United Figure 7.5 Response domain for channel adjustments States, where the number of hydrograph reversals predicted in response to the presence of a dam. Adjust- peryearhaveincreasedby30%(Magilliganand ments are conceptualized in relation to the fractional Nislow, 2001). change in frequency of sediment transporting flows (T∗) and the ratio of sediment supply downstream Numerous studies document the response of the from the dam to supply upstream from the dam (S∗). downstream channel to these changes in hydrol- End-member textural and morphologic adjustments are ogy,andtothechangesinsedimentsupplyasthe shown. Response of rivers plotting within the shaded majority of bedload and, in some cases suspended diagonal region is likely to be strongly influenced by load,istrappedupstreamfromthedam.Geomor- geological factors, including the history of floods and phic response reflects the ratio of sediment supply landslides that leave legacies of large volumes of coarse below the dam to that above the dam, as well as the material and bedrock incision in the valley and channel bottom. (From Grant et al., 2003, Figure 3, p. 209.) fractional change in frequency of flows transport- ing sediment (Grant et al., 2003) (Figure 7.5). Where can be difficult. Methods used to quantify hydro- sediment supply decreases substantially and flow logic changes include: competence remains sufficiently high, bed coars- ening, channel incision, and bank erosion occur, r the range of variability, also known as the indices and these effects can extend hundreds of kilome- of hydrologic alteration (IHA) (Supplemental Sec- ters downstream from a large dam, as shown for tion 3.2.7), in which the interannual variability of several large dams in North America (Galay, 1983) 67 streamflow parameters that reflect magnitude, and for the Changjiang (Yangtze) in China (Xu, timing, frequency, duration, and rate of change of 1996). Alternatively, the reduction in peak flows and discharge are compared pre- and post-regulation flood scouring can result in bed fining, aggradation, (Richter et al., 1996); and narrowing, especially if large sediment sources r wavelet transform, a mathematical tool used to such as tributary inputs are present just downstream extract dominant modes of variation from statis- from the dam or encroaching riparian vegetation tically nonstationary signals (Zolezzi et al., 2009); helps to stabilize sediment, as shown for the West and River in Vermont, USA (Curtis et al., 2010). Channel r comparison of the seasonal probability density adjustment varies with distance downstream from function of natural and regulated streamflows a dam, but many rivers have sequential dams that (Botter et al., 2010). cause substantial cumulative disruption of water and Of these techniques, IHA is the most com- sediment fluxes. monly used. Where dams have been present for many decades and discharge measurements are sparse, quantifying In addition to altering water and sediment flux the changes in flow regime associated with a dam and resultant changes in channel form, dams alter Chapter 7 Humans and rivers 207 water temperature and chemistry, and the move- Flow ment of nutrients, plant propagules, instream wood, regulation and organisms. Numerous studies highlight the effects of dams on fish populations (Brooker, 1981; Ligon et al., 1995; Bunn and Arthington, 2002). Hydrology Sediment supply Theseeffectsincludephysicalblockageofmigration. Fish lose floodplain spawning and nursery habitat whenreducedpeakflowslimitoverbankflooding. Hydraulics Reductions in minimum flow can prevent spawning Substrate mobility by restricting access to spawning areas. Fine sedi- Channel form ment deposition on gravels used for spawning lim- its survival of eggs and embryos. Rapid submer- gence and exposure through dewatering of the varial Biota zone—the shallow borders of channels—can expose and kill fish eggs and embryos. And, disruption of Figure 7.6 Schematic illustration of the changes in a seasonal thermal cues that regulate timing of lifecy- river as a result of flow regulation. cles can stress or eliminate fish populations. Riparian communities also experience numer- ousdisruptionswheredamsarepresent(Nilsson flow regulation (Dynesius and Nilsson, 1994; Moyle et al., 1997; Nilsson and Berggren, 2000; Katz et al., and Mount, 2007; Poff et al., 2007; Braatne et al., 2005; Merritt and Wohl, 2006). Altered channel 2008). A common theme among diverse case studies morphology or lateral disconnection of the chan- of regulated rivers is the loss of physical and ecologi- nel and floodplain as a result of reduced peak cal complexity (Surian, 1999; Graf et al., 2002). This flows decreases or eliminates freshly scoured sur- can be conceptualized with changes in hydrology faces necessary for seedling establishment. Seedlings and sediment supply as first-order effects; changes in can be killed by prolonged submersion because of hydraulics, substrate mobility, and channel form as increased base flows. Changes in timing and mag- second-order effects; and changes in biota as third- nitude of flow relative to release of riparian seeds order effects (Burke et al., 2009) (Figure 7.6). can disrupt hydrochory—the downstream transport Damsbegantoberemovedinsomeportionsof of seeds and other plant propagules to germination the United States at the end of the twentieth century. sites by streamflow. Changing grain-size distribu- These were mostly relatively small, nineteenth- cen- tion and moisture content along channel banks and tury dams built for a purpose that no longer existed, bars can limit seedling establishment. Changes in such as powering a local mill that was long gone. A the riparian water table can cause soil salinization primary concern with dam removal is remobiliza- that kills riparian vegetation. The integrated effects tion of sediment stored behind the dam, particu- of these diverse changes are illustrated by a compar- larly if this sediment contains toxic contaminants or ison of four free-flowing and four regulated rivers in high nutrient concentrations (Hart et al., 2002; Piz- northern Sweden (Jansson et al., 2000). The number zuto, 2002). Issues of increased mobility for aquatic of plant species and their cover per unit area were organisms and loss of lentic (still water) habitat can lower along regulated rivers, and wind-dispersed also be important (Stanley and Doyle, 2003). (as opposed to water-dispersed) species were more Removal of large dams is typically not econom- common along regulated rivers. ically or politically feasible, but disruptions to river Global and regional syntheses indicate that ecosystems can be reduced by modifying the opera- aquatic and riparian species are becoming increas- tions of the dam (Graf, 2001). The most widely stud- ingly homogeneous—a few hardy generalist species ied example involves experimental flood releases tend to dominate many communities—as a result of from Glen Canyon Dam on the Colorado River just 208 Rivers in the Landscape upstream from Grand Canyon National Park in the munities in both source and receiving streams. The United States (Collier et al., 1997; Wright et al., magnitude of these disruptions depends on the mag- 2008). These releases are timed in order to achieve nitude of hydrologic changes and the characteristics mobilization and downstream redistribution of sed- of the affected channels (Ryan, 1997; Parker et al., iment inputs from tributaries just downstream from 2003; Wohl and Dust, 2012) (Figure S7.1). Glen Canyon Dam. The objective of redistribut- Among the most insidious ecological effects of ing sediment is to enhance channel-margin sand flow diversions are the introductions of new species deposits and associated camping sites and channel- from one river network to another. Millions of margin fish habitat. dollars are now being spent in the United States, The complexity of most river ecosystems, and the for example, to construct barriers using concrete, complications introduced by other factors such as wire mesh, and electrical fields designed to pre- exotic species, makes it extraordinarily difficult to vent four introduced species of Asian carp from predict and create targeted effects on these ecosys- migrating through a canal connecting the head- tems by modifying dam operations, particularly waters of the Illinois River system, where the fish when each experimental flow release is expensive are established, into the Great Lakes ecosystem via and comes under intense public scrutiny. Nonethe- a canal between the Illinois and Lake Michigan less, the adaptive management being used on the (Sandiford, 2009). In southeastern Australia, water Colorado River, in which hypotheses regarding river impounded and diverted from the headwaters of the response to experimental flow releases are tested, eastward-draining Snowy River into the headwaters and modified hypotheses are used to design the of the westward-draining Murray River catchment next experimental release (Wohl et al., 2008), is now since 1967 has been accompanied by pioneering being applied to other regulated rivers (Bednarek fish species, including climbing galaxias (Galaxias and Hart, 2005; Jorde et al., 2008). brevipinnis), which may compete with many of the Flow regulation via diversion of flow from a already endangered native species of the Murray– source to a receiving channel has occurred primar- Darling basin (Waters et al., 2002). ily in dry regions and on relatively small channels, although diversions are present in a range of envi- ronments from northern Canada (Kellerhals et al., 7.2.2 Altered channel form and 1979) to southern Australia. Massive flow diver- sions are now being proposed or built for some connectivity of the world’s largest rivers, however, including the South–North Transfer Water Project to trans- In addition to regulating flow, humans have been fer water from China’s Huang He (Yellow River) directly altering channel form for centuries. The north to the Changjiang (Yangtze River); the Jonglei intent behind these alterations is as diverse as the Canal Project that would divert waters more than alterations: 300 km from the Sudd wetlands along the Nile; and the Solomon pipeline that would divert water 1000 r increasing channel conveyance by dredging the km south from the mouth of the Congo River to channel or building levees to reduce overbank Namibia (Wohl, 2011a). flooding; Diversion has most commonly been undertaken r straightening sinuous channels and stabilizing for water supply or flood control. Diversion tends banks to reduce channel mobility and bank ero- toreducebaseflowandfloodpeaksonthesource sion; channel, and increase these parameters in the receiv- r removing instream wood or beaver dams to ing channel. As with dams, changes in flow regime increase conveyance and downstream water sup- associated with diversions disrupt sediment dynam- plies, to reduce overbank flooding, or to enhance ics, channel morphology, and aquatic riparian com- fish passage; Chapter 7 Humans and rivers 209 r building check dams and weirs to limit bed inci- at least since 2600 BC, when levees were built in sion and/or downstream movement of coarse the Indus River valley. Levees date to more than bedload; 3000 years ago in Egypt along the Nile, and extensive r extending channel networks via canals that can levees were built by ancient Mesopotamian civiliza- connect the headwaters of two networks that share tionsandtheChinese.Atpresent,particularlyexten- a common drainage divide, or can shift water sive permanent levees line the Danube, Po, Rhine, across tens to hundreds of kilometers to a differ- Meuse, and Rhone Rivers of Europe, the Mississippi ent river network; and Sacramento River systems in the United States, r burying or laterally shifting channels perceived as andmostofthelargeriversofChina.TheMissis- inconveniently placed with respect to urban areas, sippi levee system is one of the world’s largest. Begun transportation corridors, or other land uses; by French settlers in Louisiana during the eighteenth r removing sediment from channels to mine placer century, it now includes more than 5600 km of lev- metals or construction aggregate; and ees along the middle and lower portions of the river, r reconfiguring a channel to a form considered and most major tributaries such as the Illinois, Ohio, more esthetically pleasing. and Missouri Rivers also have extensive levees. Levees reduce or eliminate all of the channel– Early, well-documented examples include Li floodplain exchanges discussed in Chapter 6. Levees Ping’s extensive system of irrigation canals and facilitate higher magnitude, shorter duration floods, flood-control structures built more than 2100 years and exacerbate flooding in downstream areas with- ago in the Szechuan region of China (Yao, 1943). out levees. As levees cause flood peak discharge and Levees were constructed along the Yodo River in shear stress to increase, the river bed between the Japan during the fourth century AD. Check dams levees coarsens and may incise (Frings et al., 2009). were built along mountain rivers in Japan at least Levees severely reduce lateral connectivity of water, as early as 806 AD. (Japanese Ministry of Con- sediment, organic matter, and organisms, leading to struction, 1993). Wood has been cleared from chan- loss of habitat, animal abundance, and biodiversity nels since the twelfth century in France (Piegay´ and in channel and floodplain environments (Hohensin- Gurnell, 1997). Flood embankments were built ner et al., 2004). The floodplain can be a major sedi- along Italy’s Po River during the fourteenth century ment source as the result of bank erosion along sin- (Braga and Gervasoni, 1989). In North and South uousrivers(thisisanetsedimentsourceonlyifthe America, Australia, New Zealand, and other areas eroded sediment added to the channel is not com- colonized by Europeans, alterations to channel form pensated by deposition on the opposing bar surface), typically start soon after European settlement. but levees and associated bank protection truncate As with flow regulation, an extensive literature this sediment exchange, commonly leading to chan- documentsnumerouscasestudiesofalteredchan- nel incision (Kesel, 2003). nel form and associated changes in longitudinal, lat- Along rivers with extremely high suspended sed- eral, and vertical connectivity. This section provides iment loads, including the Huang He in China, lev- a brief review of some of the most widespread types ees limit lateral channel movement and facilitate of direct channel alteration. sediment deposition within the channel. With time, Levees are as ubiquitous along lowland (as thesechannelshavebecomeelevatedabovethesur- opposed to mountain) rivers as are dams along all rounding floodplain, a situation that the Chinese rivers.And,likedams,leveeshaveaverylonghis- describe as “hanging rivers” (Xu, 2004). Levees have tory. Linear mounds constructed along rivers to been built along the lower Huang He since 475 BC, limit overbank flooding are known as levees, dikes and the bed of the river is now mostly 3–5 m above or dykes, embankments, and floodbanks, among thefloodplainbeyondthelevees.Insomeplacesthe other things (Petroski, 2006). Levees can be perma- riverbedis10mabovethesurroundings.Water nent or temporary, but they have been constructed seeps into the ground from this elevated river at rates 210 Rivers in the Landscape

Figure 7.7 A highly stabilized urban river: the Vienna River in Vienna, Austria. of 1.29 m3 per meter of river length per day during heavy machinery. Bank stabilization (also known as high flow and 0.59 m3 during low flow, reducing dis- bank hardening) involves increasing the erosional charge by an estimated 83 million m3 each year (Xu, resistance of the streambanks using methods that 2004). Along with other changes in the drainage and rangefromplantingriparianvegetation,tocover- with consumptive water use, this seepage loss caused ing the banks in large boulders (riprap) or concrete thelowerHuangHetobecomeephemeralstarting (Figure 7.7). in 1972, a condition for which there was no earlier Collectively, dredging, channelization, and bank evidence in the historical record. With time, periods stabilization are so ubiquitous and of such antiquity of no flow in this large river have become more fre- in many regions that most people have no idea what quent and the point of no flow has moved farther altered rivers looked like historically. Channels at upstream (Xu, 2004). opposite ends of the drainage basin in the Danube Dredging, channelization (sometimes known as River of Europe provide an example. canalization in Europe), and bank stabilization com- Alpine headwater tributaries of the Danube monly occur together and in association with lev- have been extensively “trained” since the sixteenth ees, the intent being to make the channel more century. Training refers to engineering channel physically uniform in dimensions, including flow form and mobility, and includes bank stabilization, depth, and to limit overbank flooding, bank erosion, check dams, and confinement of braided channels and lateral channel movements. Dredging involves into a single, straightened, and stabilized channel physically removing sediment from the channel in (Figure 7.8). Larger, mid-basin tributaries have been order to increase cross-sectional area and flow con- similarly engineered since the nineteenth century to veyance, usually for purposes of navigation or flood alter braided channels to single channels. Sediment control. Channelization involves making the chan- introduced to channels from hillslopes and formerly nel straighter and larger in cross-sectional area, typ- transported downstream has been stored in check ically via dredging and other forms of sediment dams and sediment detention basins in the headwa- removalsuchasdiggingoutthestreambankswith ters. Sediment temporarily stored in floodplains and Chapter 7 Humans and rivers 211

(a) (b)

Figure 7.8 Examples of trained alpine rivers. (a) In this view upstream in the town of St. Jakob, Austria, the river is channelized, with stabilized banks and regularly spaced concrete steps. (b) This river in the Italian Dolomites flows between stabilized banks and over pronounced artificial steps. alluvial fans along mid-network valley bottoms has (Piˇsut,´ 2002). As in the headwater tributaries, these been stabilized to limit mobilization during floods. alterations have resulted in diverse problems from The net effect of these alterations has been to sub- bed and bank erosion that undermine engineering stantially reduce sediment supply to downstream infrastructure along the stabilized channel to dra- river segments, exacerbating incision and bank ero- matic losses in abundance and diversity of aquatic sion (Habersack and Piegay,´ 2008). and riparian species (Bloesch, 2003). Floodplain habitat has also been reduced. The Channelization of lowland alluvial channels, typ- Upper Danube drainage, for example, has lost 95% ically undertaken to increase conveyance and reduce of historically present floodplain habitat (Bloesch, overbankflooding,wasparticularlywidespreadin 2003). Lateral connectivity between the main chan- portions of the United States until the 1970s (Wohl, nel and secondary channels has declined, along with 2004a). By making channels straighter, steeper, and associated river complexity, aquatic habitat and pro- less complex, channelization triggered responses ductivity (Hohensinner et al., 2004), and abundance that varied from incision and widening in the and diversity of fish (Aarts et al., 2004). channelized reaches and upstream segments of the Downstream from the Alps, the Danube alter- drainage to aggradation and exacerbated overbank nately flows through narrow canyons and broad flooding in downstream reaches (Schoof, 1980; alluvial basins that now contain cities such as Simon, 1994; Wyzga,´ 2001). Channel evolution Vienna, Austria and Bratislava, Slovakia. Histori- models (Section 5.1.6) were developed to describe cally, the Danube had a braided or anastomosing the successive responses through time of channel- planform in these basins, with multiple channels ized streams. separated by side arms, backwaters, and forested Check dams are a type of channel alteration com- floodplain. River training began along these seg- mon in steep, mountain channels (Lenzi, 2002), and ments of the river in the eighteenth century, and are primarily designed to retain sediment. Check accelerated during the nineteenth century, resulting damssharesomesimilaritieswithweirs.Weirs are in a single, straight, highly stabilized channel rel- designed to provide grade control, or local base ative to historical, multi-thread channel planforms level, in lowland alluvial channels that are actively 212 Rivers in the Landscape incising, sometimes because they have been arti- et al., 2006; Burchsted et al., 2010). (Conversely, ficially straightened or channelized. Check dams beavers introduced to Chile in 1946 have locally arealsousedtocreatealocalbaselevelthatlim- eliminated riparian forests and greatly expanded its upstream migration of headcuts in steep chan- meadow environments (Anderson et al., 2006b).) nels.Bystoringsediment,checkdamsaredesigned Wood has been removed from channels for cen- to limit downstream deposition, overbank flood- turies to reduce flooding, to enhance river navi- ing, and channel avulsion, and to enhance profile gationandfishpassage,andbecauseitisconsid- irregularity and dissipation of flow energy. Check ered esthetically unattractive (Section 5.6.2). Wood dams are particularly abundant in steep channels removal has gone on so long that most peo- of Europe and Asia (Castillo et al., 2007; Shieh ple have no understanding of how abundant and et al., 2007). widespread instream wood was historically, and Closed check dams built of concrete or rock are consequently harbor negative attitudes toward wood designed to trap all sediments until the structure (Chin et al., 2008). Records of wood removal indi- fills, at which point the sediment can be removed or cate that individual pieces, logjams and, in very large another structure can be built (Figure S7.2). Open rivers, log rafts extending many kilometers along check dams builtofsometypeofverycoarsemesh a river, were present from the smallest headwa- can pass finer sediment downstream (Figure S7.3). ter channels to the largest alluvial rivers anywhere Check dams are typically built in series, rather than in the boreal and temperate zones where forest singly, along a channel. was historically present (Sedell and Froggatt, 1984; Although check dams typically do not strongly Piegay´ and Gurnell, 1997; Phillips and Park, 2009). affect flow regime, they greatly disrupt sediment (Limited studies of wood in tropical environments fluxes. Disruption of sediment fluxes commonly indicate that, although instream wood performs causes exacerbated channel erosion downstream important geomorphic and ecological functions in (Wyzga,´ 1991), as well as altering channel geom- tropicalstreams,thewoodtendstobemuchmore etry, substrate characteristics, and aquatic and transient as a result of greater discharge per unit riparian communities (Bombino et al., 2009). drainage area and faster decay rates (Wohl et al., New approaches to check dams include designing 2012b).) structures that mimic naturally occurring step–pool Instream wood provides hydraulic resistance, bedforms both in appearance and function (Lenzi, enhances storage of sediment and organic matter, 2002). increases substrate diversity and aquatic habitat, Although people have built hundreds of thou- and enhances overbank flooding and channel– sands of dams and check dams along rivers through- floodplain connectivity (Section 5.6.2). Active out the world, they have also actively removed removalofinstreamwood,andpassivelossofsuch naturally occurring channel-spanning obstructions, wood because of reduced recruitment from riparian including instream wood and beaver dams (Sup- zones and hillslopes, has caused reduced stability plemental Section 3.2.6). Beavers, once extremely and complexity along a wide variety of channels. abundant and widespread throughout Europe Attempts to actively reintroduce instream wood and North America, are now highly restricted in intheformofengineeredlogjams(Southerland distribution and abundance, and are effectively and Reckendorf, 2010; Abbe and Brooks, 2011) absent from much of their former range. Loss of or individual pieces remain limited by lack of beaver populations results in loss of beaver dams. detailed, quantitative understanding of how wood This causes increases in conveyance, flow velocity, characteristics (abundance, size, spatial distribution sediment erosion and transport, as well as decreases along a channel or river network) relate to channel in overbank flooding, riparian water tables, nutrient form and process. storage, and aquatic and riparian habitat diversity Another form of channel alteration involves and abundance (Naiman et al., 1988; Westbrook actively or passively altering riparian vegetation. Chapter 7 Humans and rivers 213

Active alteration includes removal and replanting. ing substrate grain-size distribution and abundance, Passive alteration includes changes associated with all forms of instream mining disrupt sediment riparian grazing or the spread of invasive, exotic dynamics and cause a variety of channel adjust- species. Removing or replanting riparian vegeta- ments. Large, localized excavations can initiate a tion alters bank erodibility, as well as near- and knickpoint (Wishart et al., 2008). Bed and bank overbank flow resistance and sediment deposition erosion downstream from mining can be exacer- (Section 4.5). bated by reduced sediment supply (Lagasse et al., Ripariangrazingaffectsthedensity,type,and 1980). Sediment deficit at the mining site can spatial extent of riparian vegetation, and also result in bed coarsening and loss of aquatic habitat exposes the bank substrate and directly erodes (Kondolf, 1997). Disruption of a coarse surface the banks via trampling (Trimble and Mendel, layer can increase sediment mobility and down- 1995). Grazing animals can create ramps along stream turbidity, aggradation, overbank deposition, stream banks and trails along the floodplain that and lateral channel mobility (James, 1997; Parker enhance localized erosion (Cooke and Reeves, et al., 1997) (Figure 7.9). These changes can be 1976). Enhanced bank and overbank erosion can so severe that they cause a meandering channel result in deposition of fine sediments on the bed, to become braided, as in the case of the Middle reducingpoolvolume,spawningandmacroinverte- Fork of the South Platte River in Colorado, USA brate habitat, and hyporheic exchange (Myers and (Hilmes and Wohl, 1995). Knighton (1989) docu- Swanson, 1996). Increased water temperature and mented increases in channel width of up to 300% fol- reduced dissolved oxygen can result from loss of lowing alluvial tin mining on the Ringarooma River shading by overhanging and riparian vegetation and in Australia, and the development of a braided plan- contamination by fecal material of grazing animals. form that later reverted to a single channel when Several studies indicate that these effects can be mining ceased and sediment supply declined. rapidly reduced or eliminated if short-duration graz- Mining-related changes in channel form and pro- ing or grazing exclosures replace continuous grazing cess can significantly disrupt aquatic communities, of the riparian corridor (Magilligan and McDowell, from primary production by algae to survival of 1997; Magner et al., 2008). fish. This was documented for the headwaters of the Invasive, exotic riparian species with different Chatanika River in Alaska by Van Nieuwenhuyse characteristics than native riparian species can alter and LaPerriere (1986). Comparing otherwise anal- the density of riparian plants, and thus influence ogous mined and unmined headwater tributaries, streambank resistance to erosion and near- and they found that the watersheds with mining had tur- overbank sedimentation (Graf, 1978; Allred and bidity values up to two orders of magnitude higher, Schmidt, 1999). Exotic plants can also change pat- which resulted in 50% reduction in primary pro- terns of water uptake and transpiration, thus alter- ductivityinamoderatelyminedstream,andno ing streamflow and riparian water tables, as well detectable primary production in a heavily mined as nutrient cycling and riparian habitat for other stream. species (Hultine and Bush, 2011). Invasive, exotic Diversion of flow to process placer deposits fur- riparian species are particularly widespread and well ther disrupts river form and process, and toxic documented in southeastern Australia and the west- material such as mercury associated with placer ern United States. mining can contaminate aquatic and riparian sys- Although instream mining for sand and gravel tems for decades to centuries after mining ceases used in construction and for placer deposits of pre- (Taylor and Kesterton, 2002; Macklin et al., 2006). ciousmetalshasoccurredformillenniainsome Metal contamination from sulfide-ore mining in regions, systematic studies of the physical and eco- the headwaters of Soda Butte Creek continues to logical effects of these activities began only dur- contaminate riparian vegetation and aquatic insects ing the twentieth century (Gilbert, 1917). By alter- along the creek in Yellowstone National Park, USA, 214 Rivers in the Landscape

Figure 7.9 Historical photograph of placer mining pit along the Middle Fork of the American River, Monte Rio Mining Company, 1903. (Photograph courtesy of the Bancroft Library, University of California, Berkeley) for example, following a tailings dam failure in 1950 the past.” In the context of rivers, legacy refers to (Marcus et al., 2001). Mining and associated river sediments or channel form resulting from histori- contamination in parts of Europe, in particular, cal land uses. An example comes from a study of began centuries ago (Macklin et al., 2006). streams in the mid-Atlantic Piedmont of the east- In summary, it is important to once more empha- ern United States, where streams that historically size that the cumulative, combined effects of direct had small, anabranching channels with extensive and indirect human alteration of rivers are ubiq- vegetated wetlands and low sediment accumulation uitous (Gregory, 2006; James and Lecce, 2013). were transformed by the construction of tens of River networks in even seemingly remote areas of thousands of milldams during the seventeenth to polar regions, the tropics, vast inland deserts, and nineteenth centuries (Walter and Merritts, 2008). sparsely settled mountains have been directly and After dam construction, 1–5 m of fine, slackwater indirectly altered by a variety of human activities sediment buried the original channel network and (Wohl, 2006, 2011a; Comiti, 2012). In many cases, wetlands upstream from each dam. As the dams these activities and the resulting changes in river were abandoned, they fell into disrepair and were systems occurred so long ago that there is little or breached, causing channel incision down to Pleis- no historical record, let alone collective awareness, tocene basal gravels. The resulting incised, mean- of them. dering gravel-bed streams were assumed to repre- Increasing scientific attention to prehistoric or sent a channel form with little human influence until historical activities that altered rivers has led to Walter and Merritts (2008) undertook detailed useofthephraseslegacy effects or legacy sediments historical and stratigraphic reconstructions of the (Walter and Merritts, 2008). A legacy can be defined regionandrevealedthelegacyofhistoricalalter- as “something received from a predecessor or from ation. These insights have proved controversial Chapter 7 Humans and rivers 215 becauseofdifferingviewsofhowtoprotectand 7.3.1 Reference conditions restore river corridors in the region. At present there is a dichotomy between those who seek to pro- Any river restoration is undertaken to achieve some tect single-thread channels with riparian forests and desired end result of river process and form. Restora- thosewhoseektorestoremulti-threadchannels tion is undertaken for a variety of reasons, including with floodplain wetlands. Similarly dense networks those related to recreation, water quality, esthetics, of mill dams were present throughout Britain and protection of aquatic and riparian species, bank sta- much of northern Europe. bilization, fish passage, flow modification and dam removal, and creation of a more natural environ- ment (Table 7.2) (Bernhardt et al., 2005). The latter 7.3 River management in point is perhaps the most difficult, because achiev- ing a more natural environment entails addressing an environmental context at least one fundamental question: what is natural? (Graf, 1996; Wohl, 2011d). Thisportionofthechapterfocusesonriverman- Natural is commonly assumed to imply minimal agement undertaken specifically to restore rivers human alteration, although natural is increasingly in an environmental context. People have been being defined in terms of geomorphic integrity (Graf, managing—or at least attempting to manage—rivers 2001), or the ability of the river to adjust to exist- for millennia. Past management actions were typi- ing conditions (Fryirs and Brierley, 2009). Humans cally undertaken with the intent of making rivers or have been manipulating natural landscapes for many associated resources more conveniently accessible to thousands of years by using fire to alter land cover, humans: damming rivers to ensure water supply, for selectively hunting some animals to extinction, example, or building levees and channelizing rivers domesticating plants and animals and then altering to enhance agricultural use or settlement on flood- ecosystems to favor domesticated species, and cut- plains. Although river restoration and rehabilitation ting trees for fuel and building materials. So at what are sometimes viewed as being fundamentally dif- point in history do we consider a given ecosystem to ferent than past river management, they are the latest have last been natural: prior to agriculture, prior to iteration of trying to reconfigure rivers to conform the Industrial Revolution, or prior to some arbitrary to human expectations: in the case of restoration, human population density? expectations of more natural or esthetically attrac- Whatever (pre)historical period is chosen, the tive rivers. characteristics of rivers during that period are typ- Restoration activities such as planting riparian ically known as reference conditions,whichcanalso vegetation to stabilize streambanks date to the sev- be defined as the best available conditions that could enteenth century in Europe (Evette et al., 2009), but be expected at a site (Norris and Thoms, 1999). The projects designed to restore rivers have increased latter definition can be highly problematic, how- dramatically in number and scope since the 1990s ever, because great disagreement or uncertainty can (Bernhardt et al., 2005), particularly in the United arise as to what constitutes “best available.” Refer- States,WesternEurope,Australia,andNewZealand. ence conditions encompass all of the aspects of river As with any form of river management, some process and form discussed to this point, including projects have largely achieved their original pur- flow regime; sediment regime; water chemistry; sub- pose, whereas others have been thorough failures. strate;bedforms;channelmorphology,planform, The rationales that underpin river restoration, and and gradient; and longitudinal, lateral, and vertical the factors that result in success or failure, are worth connectivity. examiningbecausetheabilitytorestoreriverspro- In a region where some river basins have under- vides an effective measure of our understanding of gone minimal human alteration, contemporary river process and form. rivers can provide reference conditions for altered 216 Rivers in the Landscape

Table 7.2 NRRSS working group list of goal categories and operational definitions for river restoration projects.

Category Description

Esthetics/recreation/education Activities that increase community value: use, appearance, access, safety, and knowledge Bank stabilization Practices designed to reduce or eliminate erosion of banks Channel reconfiguration Alteration of channel geometry: includes restoration of meanders and in-channel structures that alter river thalweg Dam removal/retrofit Removal of dams and weirs or modifications to existing dams to reduce negative ecological impacts (excludes dam modifications solely for improving fish passage) Fish passage Removal of barriers to longitudinal migration of fishes: includes physical removal of barriers, construction of alternative pathways, and construction of barriers to prevent access by undesirable species Floodplain reconnection Practices that increase overbank flows and flux of organisms and materials between channel and floodplain Flow modification Practices that alter the timing and delivery of water quantity (does not include stormwater management) Instream habitat improvement Altering structural complexity (bedforms, cross-sectional geometry, substrate, hydraulics) to increase habitat availability and diversity for target organisms and provide breeding habitat and refugia from disturbance and predation Instream species management Practices that directly alter aquatic native species distribution and abundance through the addition (stocking) or translocation of plant and animal species and/or removal of exotic species Land acquisition Practices that obtain lease, title, or easements for streamside land for the explicit purpose of preservation or removal of impacting agents and/or to facilitate future restoration projects Riparian management Revegetation of riparian zone and/or removal of exotic species of plants and animals Stormwater management Special case of flow modification that includes the construction and management of structures (ponds, wetlands, flow regulators) in urban areas to modify the release of storm runoff Water quality management Practices that protect existing water quality or change the chemical composition and/or suspended load, including remediation of acid mine drainage

Source: From Bernhardt et al. (2007), Table 1. river basins. This approach must be used with cau- conditions (Hughes et al., 2005; Newson and Large, tion, however, because contemporary conditions 2006). Consequently, reference conditions may be constitutea“snapshot”intimethatreflectsonlya most appropriate as an ideal rather than as a goal for single state or a limited portion of the fluctuations restoration (Osborne et al., 1993). that naturally occur in rivers (SER, 2002). A tremendous amount of effort may be neces- In many regions of the world there are no rela- sary to characterize reference conditions, not least tively unaltered rivers, so reference conditions must because river process and form can vary substan- be inferred from historical, botanical, and geologic tially across even a relatively small drainage basin, records (Supplemental Section 3.2.1) (Morgan et al., and because rivers are never static in time. Even in 1994; Nonaka and Spies, 2005; Stoddard et al., 2006; the absence of human manipulation, rivers undergo Wohl, 2011d). Lack of information on reference con- fluctuations in process and form associated with ditions, as well as continuing change in catchment natural events such as floods or droughts, land- parameters, can limit the usefulness of reference slides, wildfires, tectonic uplift or subsidence, and Chapter 7 Humans and rivers 217 continuing adjustment of the river to long-term river corridor coincident with the ditch had in fact changes in climate or tectonics. Consequently, a key exceeded the HRV. component of understanding reference conditions is Ongoing climate change, abundant and beingabletoquantifythenatural or historical range widespread invasive species, and human population of variability (HRV)foragivenparameterorsetof growthandresourceusecausesomescientistsand parameters (Morgan et al., 1994; Nonaka and Spies, managers to question the relevance of HRV (Safford 2005; Wohl, 2011d). et al., 2008). If the world already looks fundamen- AnexampleofquantifyingHRVcomesfroma tally different than prior to human manipulation, segment of the Upper Colorado River in Rocky and will grow increasingly different in the future, Mountain National Park in the United States. The what do past river process and form matter? Other national park was designated in 1915. Prior to that, a scientists and managers contend that, even if a river flow diversion known as the Grand Ditch was exca- cannot be restored to HRV, detailed, quantitative vated across steep slopes uphill from, and parallel to, understanding of prior and existing river character- theColoradoRiver.TheGrandDitchcutsacrossa istics can inform management by constraining the series of small tributaries flowing into the Colorado range of potential river process and form. River from the west, diverting their waters into Knowledge of HRV provides insight into the con- theditchandthenintotheadjacentPoudreRiver ditions that native riverine species or communities drainage. The unlined ditch periodically overtops might require for survival, as well as the thresholds during high snowmelt flows, triggering debris flows or minimum values of process or form that must be that enter the Colorado River valley and the national maintained in order to sustain biological communi- park. A particularly large debris flow in 2003 buried ties (e.g., flood threshold for overbank flooding that much of a large wetland along the river and resulted provides fish access to the floodplain for spawning in deposition of ∼36,000 m3 of sediment along the andnurseryhabitat).KnowledgeofHRValso river. In a subsequent lawsuit between the national facilitates delineation of the spatial distribution of park service and the private company operating the different suites of geomorphic processes, such as irrigation ditch, two of the primary questions were portionsofamountainousheadwatercatchment whether the 2003 debris flow was beyond the natu- dominatedbydebrisflowsversusportionsdomi- ral range of variability for that site, and how best to nated by fluvial processes. This facilitates evaluating restore the river and wetland altered by the debris the location, relative rarity, and connectivity of flow. The question regarding range of variability sensitive stream segments that are likely to respond was not so obvious, because the bedrock along to alterations or that contain biologically unique both sides of the Upper Colorado River exhibits communities (McDonald et al., 2004; Brierley and hydrothermal alteration that creates faster weather- Fryirs, 2005; Wohl et al., 2007). Insight into HRV ing and erosion. Consequently, the upper river expe- provides information on the relative magnitude of riences more frequent debris flows than other, adja- variationinspecificriverattributesamongprocess cent regions, including the eastern side of the val- domains (Wohl, 2011d). Knowledge of the HRV ley,whichisnotaffectedbyGrandDitch.Examina- thus underpins our understanding of process and tion of the stratigraphy in the Upper Colorado River form in any river network. valley, using ground-penetrating radar and augering to obtain sediment samples and material for radio- carbon dating, produced an ∼4300-year record of 7.3.2 Restoration sedimentation (Rubin et al., 2012). This record indi- catedamarkedincreaseinthevolumeofdebris-flow Restoration is undertaken for many reasons, and deposition on the valley bottom subsequent to con- at many scales, from a single river segment only struction of Grand Ditch, suggesting that hills- a few hundred meters in length, to entire large lope instability and associated sediment yield to the drainage basins. The National River Restoration 218 Rivers in the Landscape

Science Synthesis database includes more than and recognizes uncertainty in the resulting river 37,000 restoration projects in the United States. responses. Riffle–pool sequences might be the key- Most of the projects in this database were imple- stones of river form and process in a project mented on river segments less than 1 km in length designed to restore fish habitat, so that restoration (Bernhardt et al., 2005), although the most high- focuses on the parameters necessary to create and profilerestorationprojectsarethoseinvolvinglarge maintain riffles and pools. segments of river basins, such as the Grand Canyon River restoration as currently practiced is not of the Colorado River in Arizona (Melis, 2011) or commonly scientific because hypotheses regarding the Kissimmee River in Florida (Toth et al., 1993; river response to a given restoration action are Warne et al., 2000; Wohl et al., 2008), both in not posed and tested. A large-scale survey of river the United States, or the Danube River in west- restoration in the United States found that fewer ern Europe (Tockner et al., 1998, 1999; Bloesch and than 10% of projects included any form of moni- Sieber, 2003). Restoration in the context of this chap- toringorassessment,althoughprojectswithhigher ter is used to include both restoration—a return to a costs were more likely to be monitored (Bernhardt close approximation of the river condition prior to et al., 2005). Less than half of all projects set mea- disturbance (however disturbance may be defined), surable objectives for the project, although nearly and rehabilitation—improvements of a visual nature, two-thirds of project managers felt that projects were sometimes described as “putting the channel back “completely successful” (Bernhardt et al., 2007). This into good condition” (however good condition may reflects the fact that perceived ecological degrada- be defined) (National Academy, 1992). tion typically motivated the projects. Post-project Three basic approaches to river restoration appearance and positive public opinion were the are currently employed (Palmer et al., 1997; most commonly used metrics of success (Bernhardt McDonald et al., 2004). The “field of dreams” et al., 2007), however, rather than more objective approach uses traditional engineering techniques to metrics or metrics grounded in scientific under- modify river form to a desired condition with the standing of river process and form. expectation that this will create processes necessary The lack of monitoring for restoration effec- to maintain that form. This is the most widely used tiveness is highlighted by a consideration of the approach for restoration on relatively short river historyofriverrestoration.Thedesignofinstream segments. This approach is named in reference to structures such as rock and log dams and deflectors an American movie about baseball that became used for habitat improvement goes back to at least famous for the phrase “build it and they will come.” the 1880s in the United States and even earlier Inthecaseofriverrestoration,theimplicationis in Europe (Thompson and Stull, 2002). Many of that restoring river form will also restore function, these structures are still used today with very little sothatorganismssuchasfishwillreturnandthrive. modification of initial designs, although systematic The “system function” approach identifies and examination indicates that such structures do alters the initial conditions required to achieve not necessarily ensure demonstrable benefits for restoration goals. This could involve modifying fine fish communities (Thompson, 2006), and may sediment inputs to a targeted segment of river by in fact decrease habitat abundance and diversity establishing riparian buffer strips, for example, or over a period of many decades (Thompson, 2002) adding instream wood to increase flow resistance (Figure7.10).Inotherwords,becausewetypically and sediment retention. The underlying rationale is do not objectively and systematically evaluate the that modifying initial conditions such as water or success of restoration projects over periods of sediment inputs will cause river process and form to severalyearsfollowingprojectcompletion,weare adjust in a desired manner. notlearningfromourmistakes. The “keystone” method identifies and incorpo- Small- to medium-scale river restoration has rates crucial components of process and form, become an industry with designs developed by those Chapter 7 Humans and rivers 219

(a)

(b)

Figure 7.10 Instream structures used for restoration. (a) Vortex weir along a river in Charlotte, North Carolina, USA. A vortex weir is u- or v-shaped with the apex pointing downstream. The structure is designed to deflect toward the channel center and promote bed scour that forms a pool. (b) Collapsed lunker in the Catskills region of New York State, USA. Lunkers are designed to stabilize stream banks and promote edge cover for fish. Both photographs courtesy of Douglas M. Thompson.

with relatively little knowledge of river process and First, restoration should be designed with explicit form that are implemented by consulting firms most recognition of complexity and uncertainty regard- likely to have a background in civil engineering ing river process and form, including the his- worksratherthaninriverscience.Underthesecir- torical context of variations in process and form cumstances, the scientific community has become through time. A well-documented example of failed increasingly vocal in criticizing restoration practices restoration imposed a stabilized single-thread chan- (Pasternack, 2013). Numerous papers emphasize nel on a river segment that had repeatedly alter- thatriverrestorationmustbebasedonorinclude nated between multi- and single-thread planforms five factors (Kondolf and Larson, 1995; Hughes et al., during previous decades in response to fluctua- 2001; Kondolf et al., 2001; Ward et al., 2001; Hilder- tions in flood magnitude and frequency (Kondolf brand et al., 2005; Wohl et al., 2005; Kondolf et al., et al., 2001). The “restored” channel was com- 2006; Sear et al., 2008; Brierley and Fryirs, 2009; pletely altered by a flood within 3 months of project Hester and Gooseff, 2010). completion. 220 Rivers in the Landscape

Second, restoration should emphasize processes objectives of restoration will continue to be met over that create and sustain river form, rather than impo- a period of many years. sition of rigid forms that are unlikely to be sustain- Numerous papers also examine how numerical able under existing water and sediment regimes. Per- simulations can be used to predict restoration out- haps the most egregious and common example is comes prior to project implementation (Brooks and braided river segments that are “restored” to sin- Brierley, 2004; Singer and Dunne, 2006). However, uous, single-thread rivers without addressing the as Bernhardt et al. (2007) emphasized in their sur- water and sediment yields that produced a braided vey of river practitioners, publishing more scientific planform, and without any consideration of mean- studies of river restoration will not by itself change der dynamics (the constructed bends are commonly the existing situation. River restoration can only stabilized to prevent migration). improve through direct, collaborative involvement Third, monitoring of projects after comple- among scientists, managers, and practitioners. tion must be included, using the set of variables Such collaborations appear to be an obvious most effective for evaluating achievement of project next step, but can be very difficult to achieve. objectives,andatthecorrectscalesofmeasurement. Interdisciplinary scientific teams provide significant Iftheprimaryobjectiveofrestorationistoenhance challenges because of differences in terminology, biodiversity, for example, then monitoring habitat conceptual models, qualitative versus quantitative heterogeneity (under the assumption that habitat knowledge,andtemporalandspatialscalesofinter- heterogeneity always correlates with biodiversity) is est between disciplines (Benda et al., 2002). These notasappropriateasdirectlymonitoringmetricsof challenges multiply when the pool of participants biodiversity (Palmer et al., 2010b). is broadened beyond the scientific community. Fourth, consideration of the watershed context, There is no question, however, that river restoration rather than an isolated segment of river, is crucial requires an interdisciplinary approach. As reviewed because of the influences of physical, chemical, and by Pasternack (2013), wetland rehabilitation pro- biological connectivity on alterations undertaken vides a model. Wetland rehabilitation is facilitated by for river restoration. The Carmel River in California, a certification program hosted by the Society of Wet- USA, provides an example where restoration using land Scientists, which includes research scientists, native riparian vegetation was not initially success- governmental regulators, and practitioners. There is ful because decades of groundwater withdrawal had no equivalent for river restoration, partly because lowered the water table below a depth that could theriversciencecommunityisdiverseandoriented be accessed by the vegetation (Kondolf and Curry, toward specific academic disciplines, as well as being 1986). Consequently, the native plants had to be strongly divided between research scientists and artificially irrigated to ensure their survival. practitioners, and poorly organized (Castro, 2008; Fifth, accommodation of the heterogeneity and Pasternack, 2013). Although regulators and practi- spatial and temporal variations inherent in rivers is tioners would like to establish a universal restoration necessary (Brierley and Fryirs, 2009). Rivers con- approach that would standardize methods, research tinually adjust parameters such as bedform con- scientists remain highly skeptical that such a “cook- figuration, bed grain-size distribution and chan- book” technique can be effective. There presently is nelwidth/depthratioinresponsetofluctuations no scientific consensus about the scientific founda- in water and sediment yield to the channel. These tions for restoration, what the practice should entail, adjustments are commonly not synchronous or or who should be allowed to undertake restoration of exactly the same magnitude between distinct (Darby and Sear, 2008; Pasternack, 2013). reaches of the river. Allowing the channel some free- An important consideration in river restoration dom to adjust to changes imposed during restora- is that restoration is not an “all or nothing” pro- tion,aswellaschangesthatwillinevitablyoccur cess. A river does not have to be—and typically after restoration, increases the likelihood that the cannot be—restored to some completely natural Chapter 7 Humans and rivers 221 condition that existed prior to intensive resource improve and protect water quality, with a set dead- use. Partial restoration within the constraints exist- lineforachieving“goodstatus”forallsurfacewaters ing in a watershed can nonetheless restore a great within member nations by 2015. This status is to deal of physical and ecological form and function. be achieved by meeting requirements for ecologi- Small watersheds in the central United States that cal protection and minimum water quality standards are effectively completely devoted to agriculture, for analogoustothoseenforcedbytheUSEnvironmen- example, have highly channelized streams. These tal Protection Agency. These goals, although laud- streams cannot be restored in the traditional sense. able, are difficult to implement: 95% of England’s River engineering is ubiquitous and occurred prior rivers,forexample,areatriskoffailinglegislated to scientific study of these streams, so that little is environmental objectives (Green and Fernandez-´ known of reference conditions. In addition, land Bilbao, 2006; Pasternack, 2013). cover has been altered throughout the watershed, Supplemental Section 7.3.2 discusses examples of so that water and sediment yields are completely river restoration from diverse river basins, including altered (Rhoads et al., 1999). Physical and ecological the Danube. function of these streams can be improved with nat- uralization,however,whichdefinesaviablemanage- 7.3.3 Instream, channel ment goal for watersheds within landscapes inten- sively modified by humans. A naturalized channel maintenance, and environmental has sustainable hydraulic and morphologic diversity flows thatsupportsgreaterbiodiversitythananunrestored channel, even though the naturalized channel may A vital aspect of river restoration at many sites is still be completely altered relative to its natural state. preserving or restoring a natural, as opposed to A naturalized channel might be deepened but not regulated, flow regime. A widely cited paper, “The straightened, for example, allowing the development Natural Flow Regime” (Poff et al., 1997), outlines of limited sinuosity and associated physical diver- the importance of magnitude, frequency, duration, sity in hydraulics, substrate, and channel geometry timing, and rate of change of flow in river net- (Rhoads et al., 1999). works for physical process and form and for eco- River restoration is also commonly spatially con- logical communities. When flow regime is altered strained by existing land ownership and use. Sci- directly by flow regulation or indirectly by changes entists undertaking basin-scale restoration in the in land cover that alter water yield to a river, chan- Missouri–Mississippi River drainage of the United nel characteristics and aquatic biota are affected. States refer to a “string of beads” approach in which Subsequent papers have documented that regulation land acquisition and restoration activities focus on tends to homogenize river flow regimes, with the key floodplain habitats such as flood-prone areas consequence that riverine physical characteristics near tributary confluences or remnant backwaters and biotic communities also become more homo- that form beads along the string of the otherwise geneous with time (Moyle and Mount, 2007; Poff altered river corridor (Galat et al., 1998). This spa- et al., 2007). Growing awareness of the importance tially discontinuous approach to river restoration of all aspects of a river’s flow regime led to the cur- yields demonstrated improvement in water quality, rent emphasis on environmental flows. flood hazard mitigation, and biodiversity. Initial efforts to protect river flow focused on The European Water Framework Directive pro- minimum flows. In arid and semiarid regions such vides an example of a new governmental approach as the western United States, dams and diversions that may help to organize efforts toward com- designed to manipulate water for consumptive uses, mon objectives and enhance river restoration at a including agricultural irrigation, can result in river national or transnational scale (European Commis- segments that are completely dewatered for some or sion, 2000). The directive is designed primarily to alloftheyear.Astwofishbiologistswroteinoneof 222 Rivers in the Landscape their papers about rivers in such regions, “it is obvi- alimitedobjective,suchaspoolscour,ortheycan ous that without water, there can be no fish” (Fausch incorporate a broader range of flow magnitudes and Bestgen, 1997). designed to maintain a physically diverse chan- The concept of instream flows developed as a nel (Andrews and Nankervis, 1995). Analogous to means to preserve some minimum flow level within instream flows, the intent behind channel mainte- thechannel.Anearlyversionofthiswasbasedon nance flows is to quantify and then legally establish the instream flow incremental methodology (IFIM) or protect the magnitude and frequency of flow nec- (Bovee and Milhous, 1978). IFIM uses a biologi- essary to maintain specific components of river pro- cal model that describes the habitat preference of cess and form. individual fish species in terms of depth, velocity, Thelatestiterationinthisprogressivelyexpand- and substrate, and a hydraulic model that estimates ing view of the components of the flow regime nec- how habitat availability varies with discharge. The essary to preserve physical and ecological integrity intent behind this method is to be able to specify in a river is environmental flows. The concept of the minimum flows below which individual species environmental flows grew out of experimental flow likely cannot be sustained in a river segment, as well releases from dams, such as those on the Colorado astheinferredgainsinhabitatandpotentiallyin River through the Grand Canyon in 1996, 2004, fishbiomassasflowincreases.Althoughthemethod and 2008 (Melis, 2011). An experimental flood is has been criticized for being overly simplistic—the tied to a qualitative or quantitative model of a river biological model does not account for interactions ecosystem that predicts some beneficial effect from such as competition or predation, for example— the flood. In the case of the Colorado River, the IFIM remains widely used in evaluating alternative floods are designed primarily to deposit finer sed- water management options (Stalnaker et al., 1995). iments (silt and sand) along the channel margins in Minimum flows can allow river organisms to sur- order to restore riparian and backwater habitat that vive for a period of time, but a river that in essence has been lost through progressive erosion of sand hasonlycontinualbaseflowswilleventuallylose bars and backwater habitats since construction of its capacity to support a diverse aquatic community. Glen Canyon Dam in 1963. Ideally, the experimental Periodic high flows are indirectly necessary to biotic flow release is conducted in an adaptive management communities because higher flows maintain habi- context in which the results of the flood are system- tatbyperformingfunctionssuchasscouringpools, atically assessed, and the underlying model of the winnowing fine sediments from the bed, and lim- river ecosystem is modified as needed (Melis, 2011). iting channel narrowing through encroachment of Experimental releases from dams are now docu- riparian vegetation. High flows are also directly nec- mented for several rivers in diverse settings (Murle¨ essary because they provide a window of opportu- et al., 2003; Konrad et al., 2011). nity during which organisms can disperse longitudi- Environmental flows now refer both to such nally and laterally, accessing new habitat for breed- experimental releases, which are typically limited in ing and feeding. duration, and to an annual hydrograph that spec- Recognition of the importance of higher flows ifies magnitude, frequency, timing, duration, and first led to the concept of channel maintenance flows, rate of change in flow, commonly for diverse con- typicallydefinedasthecomponentsofariver’sflow ditions such as wet years and dry years (Rathburn regime necessary to maintain specific physical chan- et al., 2009) (Figure S7.8). As with other forms of nel characteristics, such as sediment transport or river restoration, developing the guidelines for envi- flood conveyance. Channel maintenance flows are ronmental flows is time-consuming and challeng- an applied equivalent of the concepts of bankfull, ing because this process forces geomorphologists effective, or dominant discharge. In this applied con- and riverine ecologists to specify flow thresholds text, channel maintenance flows can specify a par- related to targets such as winnowing fine sediment, ticular magnitude and frequency of flow to achieve mobilizing the entire bed, creating overbank flows Chapter 7 Humans and rivers 223 for channel–floodplain connectivity, or maintain- and transport sediment present along the river cor- ing diversity of species and individual ages within ridor. A body of scientific literature on sediment a riparian forest. The complexity and uncertainty dynamics in the context of environmental flows is associated with river process and form, as discussed just beginning to appear (Rubin et al., 1998; Wiele atlengthinChapters3,4and5,meanthatthistaskis et al., 2007). sometimes straightforward, but more often includes a great deal of uncertainty. Uncertainty that can be acceptable in scientific research becomes more chal- 7.4 River health lenging in a management context in which every cubic meter of water released from a hydroelec- River restoration in any form is driven by the per- tric dam during an experimental flood, for example, ceptionthatariveristosomeextentunhealthy equates to a loss of revenue for the entity operating and can be improved. The concept of river health the dam. is intuitively appealing to many people and easy The procedure of assessing environmental flow to communicate at a general level to non-scientists requirements has gradually assumed characteristic (Karr, 1999). Scientists debate whether such a steps of (i) quantifying natural and altered stream conceptualization is useful or appropriate, however, flowsandthechangesintheriverflowregimein as well as how to quantify river health (Boulton, terms of relevant hydrologic metrics (Richter et al., 1999; Fairweather, 1999; Harris and Silveira, 1999). 1996; Gao et al., 2009) and (ii) quantifying relation- Much of this debate occurs in the biological liter- ships between hydrologic metrics and physical and ature, partly because river health is an example of biological river attributes, which essentially involves ecosystem health (Norris and Thoms, 1999). coupling physical and biological models (Sanderson Riverhealthisrelatedtoecological integrity, et al., 2011). Once recommendations are developed which is the ability of an ecosystem to support and for environmental flows, the process moves into the maintain a community of organisms with species policyarenainwhichacommunitymuchbroader composition, diversity, and functional organization than scientists typically weighs in on water avail- similartothosewithinnaturalhabitatsinthesame ability and use. As Arthington and Pusey (2003) region (Parrish et al., 2003). This definition of eco- describedtheprocessinanAustraliancontext,the logical integrity emphasizes biota, but implicitly two vital questions are: “How much water does a includes physical and chemical processes that sus- riverneed?andHowcanthiswaterbeclawedback tain the biota. from other users?” (p.377). An extensive literature Biologists typically explicitly include physical and has come into being during the past decade that chemical aspects of rivers in the consideration of describes environmental flow assessments and rec- river health, as exemplified by defining river health ommendations,aswellasavarietyofcasestud- as the degree to which a river’s energy source, ies (Tharme, 2003; Arthington et al., 2006; Shafroth water quality, and flow regime, plus the river’s biota et al., 2010). and their habitats, match the natural condition at Environmental flows are in many cases driven all scales (Karr, 1991; Harris and Silveira, 1999). by the need to preserve endangered species, and Numerous qualitative and quantitative metrics of ecologists tend to focus on flow regime. Geomor- river health have been developed, typically focused phologists increasingly emphasize the equal impor- on biological metrics (Harris and Silveira, 1999; tance of sediment dynamics in maintaining chan- Karr, 1999), although cumulative metrics sometimes nel complexity, habitat heterogeneity, and nutrient include measures of water quality (Bunn et al., 1999), cycling (Pitlick and Wilcox, 2001). Altered flow habitat (Maddock, 1999; Norris and Thoms, 1999), regimes are commonly accompanied by altered sed- or flow regime (Richter et al., 1996). iment dynamics as a result of sediment trapping Many of the biological metrics used to character- behind dams or changed ability of flows to entrain izeriverhealthfocusonsomeaspectofbiodiversity. 224 Rivers in the Landscape

Biodiversity is typically defined in terms of number what constitutes a natural or healthy river. Some ofspecieswithinagivenecosystem,butcanbe rivers are naturally depauperate in species, for exam- quantified in a wide variety of ways, each of which ple, because of harsh physical or chemical condi- provides specific information about the ecosystem tions or a history of geographic isolation. Some under consideration. Biodiversity reflects biological rivers receive large sediment inputs and exhibit sub- influences such as competition and predation, stantial channel instability because of natural fac- as well as physical influences such as the diver- tors such as semiarid climate or erodible lithology in sity, abundance and stability of habitat, and the the watershed. Recognition of complexity and diver- connectivity of habitat (Gaston and Spicer, 2004). sity as inherent properties within a river network Geomorphologists have been slower to develop and between river networks remains crucial for both metrics of physical river condition to facilitate quan- research and management of rivers. This takes us tification of difference between contemporary and back to viewing rivers in the context of the greater reference conditions for a river, as well as evalua- landscape. tion of river health. Geomorphic conceptualizations of rivers emphasize diversity of form and process through space and through time (McDonald et al., 7.5 Summary 2004; Brierley and Fryirs, 2005). Diversity of form and process reflects the hydrology, sediment sup- A diverse array of human activities have indi- ply, hydraulics, substrate, geomorphic history, and rectly and directly altered the supply of water, sed- biota at the site. A desirable level of physical diversity iment, nutrients, and contaminants to rivers, as well can constitute physical integrity, which Graf (2001) as altering channel geometry, fluxes of materials definesasasetofactivefluvialprocessesandland- and organisms, and the six degrees of connectiv- forms such that the river maintains dynamic equi- ity between a river and the greater landscape. Even librium, with adjustments not exceeding limits of regions that have never experienced dense human change defined by societal values. In other words, populations or intensive resource extraction, such a river has physical integrity when river process as some high-latitude or high-altitude regions, are and form are actively connected under the current now affected by atmospheric warming and associ- hydrologic and sediment regime. ated changes in hydrologic balance. An appropriate A geomorphic perspective on river health would way to recognize that we live during the Anthro- characterize a healthy river as having two basic pocene is to assume by default that any particular characteristics. First, the ability to adjust form and environment has been altered by human activities process in response to changes in water and sedi- (Wohl, 2013c). ment yield, whether these changes occur over many A major challenge for fluvial geomorphologists decades to centuries (e.g., climate variability) or rel- is to effectively integrate our understanding of atively short time periods (e.g., a large flood or river process and form into contemporary efforts landslide).Second,ahealthyriverhasspatialand to restore rivers and to assess and protect river temporal ranges of water and sediment inputs and health. Among the unique contributions that geo- river geometry similar to those present under nat- morphologists can make to river restoration are (i) ural conditions. Discussions of how to evaluate recognition of the historical context of landscapes, river health have taken on increased importance as including historical human alterations of riverine broad governmental regulations such as the Euro- ecosystems, (ii) knowledge of connectivity and pean Union’s Water Framework and Habitats Direc- thresholds that influence river response to natural tive mandate delineation of diverse aspects of river and human-induced alterations, and (iii) ability to health (Newson and Large, 2006). quantify thresholds, alternative stable states of a Onedangerinvolvedinthistypeofnationalor river, landscape resilience, and physical integrity of transnational evaluation is that of over-simplifying the critical zone (Wohl, 2013c). Chapter 8 Rivers in the landscape

This final chapter looks at interactions between cific geographic location and the history of that rivers and landscapes across varying time and space location. scales, starting with interactions over geological timescales of 103–107 years and entire continents. Thesignificanceofspatialcontextandthepoten- 8.1 Rivers and topography tial for relatively abrupt spatial transitions in pro- cess and form within a drainage basin are exam- Having examined the details of how water and sed- ined in the section on geomorphic process domains. iment move down hillslopes and into channels, and The third section of the chapter returns to the then move through a river network, it is useful to idea of connectivity, which was introduced in the step back and consider the larger scale distribution first chapter, and explores the implications of con- of rivers across continents. Important insights into nectivity for the diverse river processes and forms interactions between rivers and topography can be discussedthroughoutthisvolume.Thenextsec- gained by examining river configuration. tion of the chapter explores distinctive climatic The manner in which rivers both respond to signatures of river process and form associated and shape surrounding topography has been inves- with high latitudes, low latitudes, and warm dry- tigated systematically for more than a century. Early lands,respectively.Fiveregionalexamplesarethen work explored why some rivers cut through moun- used to illustrate how details of climate, geology, tain ranges rather than simply flowing downward and human resource use through time influence from high points in the landscape. Significant ques- river process and form and create a context that tions at the time of this late-nineteenth-century must be considered if river management is to be research included, What is the role of rivers (as successful. opposed to glaciers or other processes) in cutting Preceding chapters have introduced the basic deep river canyons?, and How do large-scale struc- processes of water, solute, and sediment movement tures (mountains, deep canyons) relate to move- into and through channel networks, and the result- ments of Earth’s crust? Subsequent research has ing river forms and adjustments through time. This emphasized (i) how redistribution of mass at the sur- final chapter returns to the idea that river pro- face by river erosion can influence redistribution of cess and form reflect not only physics but also dis- subsurface mass via movements of molten material tinctive processes and forms associated with a spe- in the crust and tectonic movements and (ii) how

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

225 226 Rivers in the Landscape river gradient and channel width can be used as indi- nel bed and enlarge the channel cross section relative cators of spatial variations in rock uplift. to alluvial segments. The most obvious topographic influences on segments can be interpreted as geo- river process and form occur in mountainous envi- logically transient features that are not confined to ronments where rivers cut deep, narrow canyons headwaters. The middle and lower Danube River of as they flow down toward adjacent lowlands. For Europe alternately flows across large alluvial basins more than a century, however, investigators have andthencutsthroughmountainranges.Thelower recognized that rivers do not always follow topog- MississippiRiver,commonlythoughtofasafully raphy. Observing rivers that cut across mountain , is better described as a mixed bedrock- ranges in the western United States, Powell (1875, alluvial channel because of the presence of a cohe- 1876) distinguished antecedent drainage networks in sive, Pleistocene-age clay unit that influences river which pre-existing channels maintained their spa- process and form in a manner analogous to bedrock tial arrangement while the underlying landmass was (Schumm et al., 1994; Nittrouer et al., 2011). deformed and uplifted, and superimposed channels The relative lack of erosive ability that produces which incised downward to a buried structure. In a bedrock river segment can reflect greater ero- either case, the river flowed through or across the sional resistance where the river crosses a differ- mountain range, rather than being a consequence of ent lithology (Wohl, 2000b), or changes in relative the topography. Today these two types of drainages base level associated with base level fall or with are commonly referred to as transverse drainages uplift of the drainage (Howard, 1980; Seidl et al., (Figure S8.1) that cut across bedrock topographic 1994). Where the presence of bedrock river seg- highssuchasanticlinesorupwarps(Douglassand mentsreflectsenhancedincisioninresponsetorel- Schmeeckle, 2007). Transverse drainage subsumes ativebaselevelfall,riverincisionistheprimary antecedent and superimposed drainages, overflow, non-glacial mechanism of transmitting base level and drainage piracy. In drainage piracy,adrainage changes across the landscape (Hancock et al., 1998). network on one side of a bedrock high erodes head- Bedrock river incision steepens adjacent hillslopes, ward sufficiently to lower or breach the drainage increases topographic relief between summits and divide and divert flow from a network on the other valleybottoms,removesmassfromthelandscape, side of the divide. Continuing research has devel- and ultimately sets the rate at which the entire land- opedtoolsthatilluminatetheinfluencesoftecton- scape evolves (Howard, 1994; Burbank et al., 1996; ics on the spatial arrangement of rivers and the Hancock et al., 1998) (Figure 8.1). geometry of individual river channels, as well as The idea that hillslopes and rivers mutually adjust the influences of river incision on tectonics and was first expressed by G.K. Gilbert (1877), and sub- topography. sequently formally stated in the context of dynamic Thegradientandwidthofriversincisedinto equilibrium by Hack (1960) as a condition in which bedrock are the geometric parameters most com- “ ... every slope and every channel in an erosional monly used to infer the spatial distribution and rel- system is adjusted to every other. When the topogra- ative magnitude of tectonic forces. A river incised phy is in equilibrium and erosional energy remains into bedrock, rather than alluvium, implies that the the same all elements of the topography are down- channel’s capacity to transport sediment exceeds wasting at the same rate” (Hack, 1960, p. 80). the sediment supply (Howard, 1980). Regardless of where they occur along a river’s course, bedrock river segments typically have smaller width/depth 8.1.1 Tectonic influences on ratios and steeper gradients than alluvial segments river geometry (Montgomery and Gran, 2001; Wohl and David, 2008). These differences in geometry effectively Increased availability of topographic data in the enhance the flow’s limited ability to incise the chan- form of electronic digital elevation models (DEMs) Chapter 8 Rivers in the landscape 227

(a) River and hillslopes stable

(b) Base level fall causes river to incise into bedrock

(c) River incision oversteepens adjacent hillslopes, triggering landslides that cover river bed with sediment, limiting further incision until sediment is removed Figure 8.1 Illustration of the effects of bedrock channel incision on landscapes. In these schematic views looking upstream within a river valley, base level change triggers river incision that affects the stability of adjacent hillslopes and tributaries. Gray shading represents hillslope regolith and valley alluvium. greatly enhanced the ability to detect irregularities tial uplift (Whittaker et al., 2007a, 2007b; Attal et al., in river longitudinal profile starting in the 1990s. 2008; Yanites et al., 2010) as well as changes in rock Profile irregularities (Figure S8.2) can reflect down- erodibility (Wohl and Merritt, 2001). Typically, seg- stream variations in lithology and erodibility (Valla ments of higher uplift or more resistant rock have et al., 2010), glacial history (Hobley et al., 2010), sed- deeper, narrow cross-sectional geometry. iment inputs (Cowie et al., 2008), and rock uplift Adjustments of gradient and width in response (Snyder et al., 2000), so interpreting the signifi- to increasing substrate resistance or uplift are typ- cance of irregularities typically requires knowledge ically tightly coupled (Whipple, 2004; Stark, 2006). of other characteristics of the river environment. Themostcommonlyusedapproachistointerpret Where investigators have independent evidence of downstream variations in scaling relations among uplift rate, as in the central Apennines of Italy (Whit- channel width w,drainageareaA, gradient S,and taker et al., 2008) or the Santa Ynez Mountains of discharge Q—in other words, downstream hydraulic California (Duvall et al., 2004), steeper gradients geometry—as reflecting changes in rock erodibility strongly correlate with greater rates of rock uplift or uplift rate (Duvall et al., 2004; Cowie et al., 2006; (Whipple et al., 2013). Jansen, 2006). Scaling laws change along bedrock Spatial variations in channel width/depth ratio channels crossing an active in the central Ital- are not as readily detected using remote information ian Apennines, for example (Whittaker et al., 2007a), as are changes in river gradient, but variations in the and along rivers crossing growing folds in New width of bedrock channels can also reflect differen- Zealand (Amos and Burbank, 2007). 228 Rivers in the Landscape

8.1.2 Effects of river incision Erosion on tectonics Greater precipitation Removes surface mass Decades of research indicate that rivers can convey Greater river erosion Rise of molten rock tremendous volumes of sediment from high-relief landscapes.Riversremoveuptofivetimesmore Uplift sediment per unit area from mountainous basins Precipitation than from lowland basins (Corbel, 1959; Willen- bring et al., 2013). Milliman and Syvitski (1992) esti- Enhanced relief mated that 96% of the approximately 7819 million Orographic effects tons of sediment delivered to the oceans by rivers Figure 8.2 Schematic illustration of the interactions each year originates in mountainous settings. among tectonics, topography, and climate, as illustrated By the 1990s, investigators realized that one by research in the Himalaya. implication of this ability to remove mass from mountainous regions is that river incision can affect et al., 2002). This is demonstrated in the crustal structure in mountain belts by changing the (Montgomery and Stolar, 2006), where erosion distribution of stress in the crust (Molnar and Eng- along major rivers causes focused rock uplift. The land, 1990; Hoffman and Grotzinger, 1993; Beau- uplift creates anticlines and the anticlines correlate mont and Quinlan, 1994; Small and Anderson, withlocalrainfallmaximabecausemonsoonprecip- 1998). Local rheological variations arise in a deform- itation is advected up the river valleys. The greater ing orogen as a result of deep and rapid inci- rainfall in turn increases the erosive capability of the sion by glaciers or rivers (Zeitler et al., 2001). The rivers (Figure 8.2). crust weakens as the strong upper crust is locally stripped from above by erosion. This causes the local geotherm (or rate of change in temperature 8.1.3 Indicators of relations with depth below the surface) to steepen from below as a focused, rapid uplift of hot rock occurs. In between rivers and landscape other words, incision by glaciers or rivers removes evolution enough mass that molten material rises preferen- tially into the eroding area from Earth’s interior. Davis (1899) first attempted to relate river geom- If efficient erosion continues, material continues to etry to landscape evolution in a conceptual model flow into the weakened zone, maintaining local ele- known as the cycle of erosion.Thismodelposited vation and relief (Koons et al., 2002; Booth et al., high topographic relief and rivers with steep gra- 2009a, 2009b). This conceptualization of the inter- dients in geologically young landscapes. As ero- actions between river erosion, uplift, and topogra- sion gradually transferred mass to lower elevations, phy is known as the tectonic aneurysm model (Zeitler topographic relief and river gradients progressively et al., 2001). decreased from mature to old landscapes. A river’s ability to incise depends partly on dis- The cycle of erosion, despite the name, assumed charge and the climate that supplies runoff, and a highly linear landscape evolution with time and the picture now emerging emphasizes strong cou- implied that an observer could readily interpret the pling among climate, erosion, and tectonics. Gra- relative geologic age of distinct landscapes based dients in climate (Bookhagen and Burbank, 2010) on their topography. This conceptualization is typ- and tectonic forcing influence erosional intensity, ically contrasted with the ideas of Davis’ contem- which governs the development of topography, porary, G.K. Gilbert. Gilbert emphasized nonlin- whichinturninfluencesclimateandtectonics(Roe ear change with periods of little net change, or Chapter 8 Rivers in the landscape 229 equilibrium, as a result of feedbacks such as those 1 subsequently recognized as tectonic aneurysms or isostatic rebound (Gilbert, 1877). Isostatic rebound is delayedupwardflexureofEarth’scrustinresponse to removal of mass such as a continental ice sheet Front Range, Colorado, USA that had depressed the crust when the ice sheet Great Smokey Mountains, North Carolina, USA was present. The implication of tectonic aneurysms or isostatic rebound is that elevation or relief may change relatively little over time spans of 103–104 Blue Ridge Mountains, years, despite continued erosion and transfer of mass Virginia, USA to lower elevations.

Many subsequent investigators have demon- basin height of total Proportion strated that rates of landscape change fluctuate sub- stantially through time and space (Bierman and Nichols, 2004). Conceptual models now tend to 0 0 1 emphasizethatmostlandscapechangesoccurdur- Proportion of basin area ing relatively short periods of time and are concen- Figure 8.3 Hypsometric curves for three small drainage trated in relatively small portions of a drainage basin, basins in mountainous regions of the Colorado Front although net change in elevation or relief may be Range, the Blue Ridge Mountains of Virginia, and the minor. Great Smokey Mountains of North Carolina, all in the Topographic metrics are still used to infer rela- United States. The river in the Colorado Front Range tiveratesorstagesoflandscapeevolution.Among begins on a broad, low-relief upland and then descends these metrics are hypsometric curves (Strahler, 1952), steeply through a narrow canyon. The catchment in the which illustrate the distribution of mass within a Great Smokey Mountains is well dissected, with no allu- basin by plotting proportion of total basin height vial fans, no floodplains, and minimal deposition in the against proportion of total basin area (Figure 8.3). valleys. The catchment in the Blue Ridge Mountains has Strahler (1952) proposed that these curves could be alluvial fans at the base of the ridges and extensive val- used to distinguish relative basin age as a function ley infill, so that more of the mass is at lower elevations in this catchment. Figure courtesy of Gregory S. Springer. of decreasing hypsometric integral—the area under (From Wohl, 2010, Figure 2.30.) the hypsometric curve—with increasing age. Sub- sequentresearchsuggeststhathypsometriccurves canbeusedtoinferthehistoryandprocessesof 8.1.4 Tectonics, topography, basin development. The distribution of mass within and large rivers a basin reflects uplift rates and variations in erodi- bility of different lithologic units (Walcott and Sum- Although much of the research summarized in the merfield, 2008; Perez-Pe´ na˜ et al., 2009), as well preceding sections has focused on rivers in moun- as differences in diffusive (hillslope) versus fluvial tainous terrain, investigators have also examined sediment transport (Willgoose and Hancock, 1998) the spatial arrangement of very large rivers in the and glacial versus fluvial erosion (Sternai et al., context of tectonic history and topography. Potter 2011). Hypsometric curves are likely to be concave- (1978) described how the location and configura- down everywhere, for example, within landscapes tion of most of Earth’s largest river basins reflect the dominated by diffusive transport (Willgoose and tectonic assembly and deformation of continental Hancock, 1998), and glacial valleys are more likely land masses. He noted that the basic configuration to have concave-up curves than are fluvial valleys of some major rivers has been extremely persistent. (Sternai et al., 2011). The Mississippi drainage, for example, has existed 230 Rivers in the Landscape about 250 million years, or about 1/16 of Earth’s his- that deeper crustal structures and geologic processes tory (Potter, 1978). This persistence partly reflects occurring over millions of years can strongly influ- the control of deeper crustal structures. ence river characteristics. These influences may be The 28 largest rivers discharge across trailing- most readily detected in mountainous portions of edge coasts without compressional deformation, a river network, but can also influence the world’s which reflects the continental asymmetry of many largest lowland rivers, such as the Amazon or the watersheds (Inman and Nordstrom, 1971). The Mississippi. mouths of many large rivers occupy grabens or crustal downwarps. The alignment of several major rivers, including some in South America (Potter, 8.2 Geomorphic process 1978), the Nile (Schumm and Galay, 1994), and the domains Rio Grande in North America, correlate with large- scale crustal fracture patterns. Turning to smaller spatial scales at the subwatershed The Amazon River provides an example of the orreachlevel,distinctsuitesofgeologicandclimatic three scales at which tectonics can influence large- processes can create spatial differentiation within scale features of a drainage network (Mertes and river networks. Schumm (1977) conceptualized Dunne, 2007; Dunne and Aalto, 2013). At the conti- drainage basins as consisting of an upstream zone nental scale (5 × 103 km), the assembly of orogen of production from which water and sediment (mountains), foreland basin, cratons, and grabens arederived,acentralzoneoftransferinwhich influences production of runoff, sediment supply, inputs can equal outputs in a stable river, and a and accommodation space. Specifically, the Ama- downstream zone of deposition (Figure S2.20). zon heads on the Andean arc at the leading edge of Although acknowledging that production, transfer, theSouthAmericanPlate.TheAndesprovidethe and deposition occur continuously throughout a majority of the sediment that the Amazon carries drainage basin, this organization recognizes the across a very broad lowland before entering a graben existence of spatial zonation in dominant processes that localizes the mouth of the river on the trailing within a catchment. edge of the continent. Subsequent conceptual frameworks have also At an intermediate scale (102–103 km), the spac- emphasized spatial zonation. Montgomery and ing of crustal warping transverse to the river course Buffington (1997), for example, distinguished influences gradient, valley width, channel sinuos- source, transport, and response segments in their ity, accommodation space, and sediment distribu- reach-scale classification of mountain channel tion across the floodplain. Four major structural morphology (Figure 4.9). Sklar and Dietrich (1998) arches lie transverse to the main course of the Ama- hypothesized consistent changes in dominant inci- zon between the Peru–Brazil border and the Atlantic sion mechanism (debris flow, fluvial) and substrate Ocean (Mertes and Dunne, 2007). As the Ama- type (coarse bed alluvial, fine bed alluvial) at thresh- zon crosses each of these arches, channel gradient old slopes, regardless of drainage area (Figure 2.11). increases, the valley grows narrower, and the chan- Montgomery (1999) built on this work in describ- nel becomes less sinuous. ing process domains,whichhedefinedasspatially At the local scale (101–102 km), brittle crustal identifiable areas of a landscape or drainage basin fracturing influences channel orientation and gra- characterized by distinct suites of geomorphic pro- dient. Channel alignment follows fractures because cesses (Figure 2.8). The existence of process domains thesecreatelocalizedzonesofmorereadilyeroded implies that river networks can be divided into dis- bedrock (Latrubesse and Franzinelli, 2002). crete regions in which ecological community struc- These examples illustrate that, when examining ture and dynamics respond to distinctly different river process and form, it is important to remember physical disturbance regimes (Montgomery, 1999). Chapter 8 Rivers in the landscape 231

As noted earlier, some river geomorphic parame- Otherconceptualframeworksthatemphasize ters exhibit progressive downstream trends, whereas spatial zonation include the River Styles framework others exhibit so much local variation that any sys- (Brierley and Fryirs, 2005) structured around the tematiclongitudinaltrendsthatmightbepresentare five spatial scales of the obscured (Wohl, 2010). Local variation that over- whelms progressive trends is particularly character- r catchment, istic of mountainous terrain, where spatially abrupt r landscape units of relatively homogeneous topog- longitudinal zonation in substrate resistance, gra- raphy within the catchment, dient, valley geometry, and sediment sources can r reaches with consistent channel planform, assem- create substantial variability in river process and blage of channel and floodplain geomorphic units, form. Under these conditions, characterizing river and bed-material texture, dynamics at reach scales can be more accurate than r channel and floodplain geomorphic units such as assuming that parameters will change progressively pools, and downstream. Examples of geomorphic parameters r hydraulic units of homogeneous flow and sub- for which spatial variation is better explained by pro- strate characteristics. cess domain classifications than by drainage area or discharge in mountainous drainage basins include River style is classified at the reach scale. The riparian zone width (Polvi et al., 2011), floodplain catchment-wide distribution of river styles can be volume and carbon storage (Wohl et al., 2012c), usedtounderstandspatialdistributionofcontrols and instream wood load (Wohl and Cadol, 2011) on river process and form, relative abundance of (Figure S8.3). different river styles, and potential sensitivity and Process domains can also apply to very large resilience of different portions of the river network, drainage basins that have distinct spatial differences analogous to the application of process domains. associated with topography or climate. The wet, Stream biologists also emphasize zonation at dif- high-relief, sediment-producing Ethiopian High- fering spatial scales. This is exemplified by the lands portion of the Blue Nile, for example, is dis- widely used hierarchy of stream system (103 m tinctly different than the dry, low-relief mainstem length), stream segment (102 m), reach (101 m), Nile lower in the drainage. Similarly, the steep, nar- pool/riffle (100 m), and microhabitat (10−1 m) of rowly confined segments of the Danube that cut Frissell et al. (1986). through mountainous terrain are distinctly different Aquatic and riparian ecologists have also exam- than the intervening anastomosing or braided seg- ined the relative importance of progressive trends ments in broad alluvial basins. versus local controls. One of the primary conceptual Process domains can provide a useful organiza- models of aquatic ecology, the river continuum con- tional framework for delineating the spatial distri- cept,emphasizescontinuouslongitudinalgradients bution and relative abundance of different valley in the structure and function of ecological commu- and channel types (Wohl et al., 2007), and this can nities along a river network (Vannote et al., 1980). In facilitate identification of sensitive or rare areas and contrast, the serial discontinuity concept focuses on formulation of different management strategies for how dams disrupt longitudinal gradients along river distinct physical settings. The concept of process courses (Ward and Stanford, 1983, 1995). Hierarchi- domains can be readily applied at the reach scale cal patch dynamics (Pringle et al., 1988; Poole, 2002) atwhichmostrivermanagementoccurs.Witheven emphasizes the existence of relatively homogeneous minimal field calibration, process domains can also unitsfromthescaleofmicrohabitatuptochannel provide a framework for remotely predicting at least reaches, with distinct changes in process and form relative variations in numerous valley-bottom char- between patches. Reach-scale patches, like process acteristics. domains or river styles, can result from changes in 232 Rivers in the Landscape physical processes such as glaciation or differential vialfanscanlimittherateatwhichsedimentis rock erodibility (Ehlen and Wohl, 2002) and from introduced from hillslopes to channels. Extensive biotic influences such as beaver dams (Burchsted floodplains can increase the time necessary for et al., 2010). sediment entering a river network to move com- The appropriateness of conceptual models that pletely through the network. As described in Sec- emphasize progressive downstream changes versus tion 6.1.2, sediment can reside on the floodplains of patchiness varies depending on the river character- the Amazon River for thousands to millions of years. istic being described and the specific drainage basin. Erosionally resistant portions of a river network Recognition that a variety of river processes and can influence landscape connectivity. Segments of forms can exhibit abrupt spatial transitions, how- bedrock river commonly act as local base levels, ever,illustratestheimportanceofconsideringland- forexample,andtheupstreamtransmissionofbase scape context when examining river process and levelchangeislimitedtotherateatwhichtheriver form. The presence and characteristics of connec- canincisethebedrocksegment.Largewaterfallsor tivity exert an important influence on both spatially portions of an ephemeral or intermittent river net- continuous and discontinuous processes and forms. work that are dry can limit migration of organisms and thus limit biological connectivity (Jaeger and Olden, 2012). 8.3 Connectivity Features that limit connectivity can be conceptu- alized as reservoirs that store materials, as exempli- Although the first chapter of this volume introduced fied by alluvial fans storing sediment or floodplains theideaofdiverseformsofconnectivity,itisworth storing peak flows during a flood. Features that limit returning to this concept and exploring the implica- connectivity can also be conceptualized as barriers, tions of connectivity for several of the processes and asinthecaseofalocalbaselevelthatlimitspro- forms discussed to this point, including file adjustment or a dry stream segment that lim- its fish dispersal. Whether a reservoir or a barrier, r interactions between hillslope, floodplain, chan- these aspects of river networks exert critical con- nel, and hyporheic environments; trolsonfluxesofmaterialandorganismsandmust r sources, transport, and residence time of water, be included when understanding or quantifying all sediment, and solutes; and aspects of river networks, from production of water, r spatial zonation within a drainage basin. solutes, and sediment, to movement of these mate- rials downslope into channels and through channel Figure 8.4 illustrates how various forms of con- networks. nectivity change throughout a drainage basin. Among the challenges in managing rivers are High connectivity implies that matter and organ- those of quantifying connectivity and under- isms move rapidly and easily within a river net- standing how human activities have increased or work. Landscapes typically include some charac- decreased connectivity within a landscape (Kondolf teristics that create at least temporary storage and et al., 2006). As discussed in the introductory chap- limit connectivity. Subsurface units of low perme- ter, most human activities decrease hydrological, ability can limit the downslope transmission of sediment, biological, and landscape connectivity water from hillslopes to channels, or limit hyporheic within a river network, although a few alterations and groundwater exchanges along channels. Lakes, such as flow diversions and removal of naturally broad floodplains with extensive wetlands, and occurring obstructions such as beaver dams can numerous channel-spanning obstructions such as increase connectivity. beaver dams or logjams can substantially decrease Unanticipated side effects of altered connectiv- therateatwhichfloodsmovethrougharivernet- ity can require many decades to become apparent. work. Similarly, depositional features such as allu- Before the 1970 completion of the Aswan High Dam Chapter 8 Rivers in the landscape 233

Downstream changes Headwaters Characteristics and in connectivity connectivity Source zone Slopes and channels coupled Tributaries and mainstem coupled Limited floodplain development Longitudinal sediment transfer is efficient Organic matter input direct from uplands and riparian zone; dominated by CPOM Midcatchment Limited wood mobility; individual pieces important Limited hyporheic exchange Transfer zone Irregular slope–channel connectivity Floodplain width irregular, longitudinally discontinuous Moderate channel–floodplain connectivity

Increasing sediment storage Longer residence times of sediment Decreasing sediment delivery Decreasing slope–channel connectivity Increasing channel–floodplain connectivity Increased wood transport Longitudinal sediment transfer is efficient Organic matter breakdown to FPOM Wood mobile; jams important Floodplain Floodplain Parafluvial hyporheic hyporheic zone Greater hyporheic exchange zone zone Lowland plain Depositional zone Slopes and channels decoupled Floodplains broad and longitudinally continuous High channel–floodplain connectivity Longitudinal sediment transfer inefficient Organic matter dominantly FPOM Wood highly mobile; limited in-channel storage Greatest hyporheic exchange

Floodplain Parafluvial Floodplain hyporheic Hyporheic zone hyporheic zone zone zone Lateral connectivity Longitudinal connectivity (hillslope–channel) Vertical connectivity Lateral connectivity (channel–floodplain) Figure 8.4 A schematic illustration of changes in connectivity with distance downstream along a river with high- relief headwaters. Moving downstream, the river flows through headwater valleys with relatively thin, narrow alluvial veneers over bedrock, and then through progressively wider and deeper alluvial valleys with greater floodplain devel- opment and hyporheic exchange. The presence of a floodplain buffers the mainstem river from hillslope and tributary inputs by creating depositional zones along the river, and progressively more extensive floodplains typically equate to greater average residence time of sediment, surface flow during overbank floods, and subsurface flow. CPOMis coarse particulate organic matter (>1 mm in diameter) and FPOM is fine particulate organic matter (0.45 μm–1 mm). (From Brierley and Fryirs, 2005, Fig 2.10, p. 44.) on the Nile River in Egypt, the river annually carried ical, and biological characteristics of the lower river more than a hundred million tons of silt to the Nile and the nearshore zone (Wohl, 2011a). Delta. The dam now traps much of this sediment, Along Siberia’s Ob River, the Novosibirsk Dam causingsubsidenceanderosioninthedelta:former reduces seasonal peak flows along more than a thou- delta villages are now 2 km out to sea. Plankton for- sand kilometers of river. The Ob historically pro- merly nourished by nutrients in the river flow have vided an important commercial fishery, but many dramatically decreased in abundance, contributing species of fish require access to the floodplains for to a collapse of sardine populations that fed on the spawning and nursery habitat during the spring plankton. Reduction of sediment connectivity along peak flows. Fish in the Ob River need at least 20 this major river has thus altered the physical, chem- days of flooding in order to spawn, hatch, and grow. 234 Rivers in the Landscape

The Novosibirsk Dam has reduced floodplain habi- occurrence of very cold temperatures that main- tatbyhalfduringyearsofaverageflow,andelimi- tain seasonal ice cover on rivers (Figure S8.4). As natedthishabitatduringdryyears.Thisdam,along discussed in Section 3.2.6, the formation, presence, with several others in the Ob catchment, also lim- and break-up of ice create distinct hydrologic and its longitudinal movements by fish, creating genetic hydraulic effects on stage and discharge. The pres- isolation and constraining the ability of each fish ence of ice can also strongly influence channel and population to find appropriate habitat during dry floodplain erosion, deposition, and connectivity. periods. The commercial fishery along the Ob has Ice cover and ice jams that form during break- largely collapsed since construction of the dams up create backwater effects that enhance overbank (Wohl, 2011a). flooding. When ice jams break, the resulting The effects of altered connectivity along the surge can enhance overbank flooding downstream. Nile and the Ob might appear obvious in retro- Changing hydraulic forces during ice jams and spect, but the challenge of anticipating not only break-up, along with the mechanical effects of large the type, but the magnitude, of altered connectivity chunks of moving ice, can enhance bank erosion in connection with river engineering remains sub- and the overbank flooding and scouring that struc- stantial. Numerical simulation can be particularly ture riparian vegetation (Beltaos, 2002). Avulsion useful in this context by facilitating the ability to induced by ice jams has been described for both evaluate alternative scenarios, but any model used meandering and braided channels, and channel ice must be accurately parameterized and underpinned can facilitate bank erosion by gouging the banks, by a solid understanding of the sources and char- increasing the bank loading, and reducing vegeta- acteristics of connectivity within a drainage basin. tion growth along the banks (Ettema and Kempema, Again, knowledge of landscape context is critical to 2012). Ice cover can alter lateral variations in flow effective understanding and management of river depth and boundary shear stress within a channel process and form. (Ettema and Kempema, 2012). Ice–backwater effects can alter the direction of flow within a channel and the connectivity between the main channel, sec- 8.4 Climatic signatures ondary channels, and the floodplain (Prowse and Beltaos, 2002). Ice that retards flow can decrease Many of the preceding chapters have briefly men- bed-material transport, but ice can also increase sed- tioned differences in river process and form in rela- iment transport by directly moving the bed sed- tion to climate. Climatically induced diversity in iment via ice rafting (transport of sediment by processandformhasbeenrelativelyneglectedin floating ice), ice gouging, or ice push (Ettema and scientific river studies, however, because the great Kempema, 2012). majority of studies have been conducted in the tem- Icejamsandthesurgesthatresultfromtheir perate latitudes where many scientists live. This release can cause habitat degradation or loss, species imbalance and relative neglect of high latitudes, low stress or mortality, and deposition of fines and latitudes, and drylands has been changing in recent deterioration of spawning grounds. The highest decades, and this portion of the chapter reviews the suspended sediment concentrations occur during implications of climatic differences for interactions freeze-over and break-up, however, and ice jams and between process and form. surges also replenish adjacent floodplains with sed- iment and nutrients (Beltaos, 2002). Althoughthereisnouniquetypeofcold-region 8.4.1 High latitudes river geometry, sandur (a valley segment undergo- ing rapid aggradation, with a downstream decrease Thesalientfeatureofhighlatitudeswithrespectto in particle size), braided channels, meandering interactions between river process and form is the channels, and anastomosing channels in wetland Chapter 8 Rivers in the landscape 235 environments are particularly common in cold Extremely intense rainfall and preferential shal- regions (Vandenberghe and Woo, 2002). For all of low flow paths such as macropores (Section 2.2.1) these channel types, the periods of dynamic change result in large hydrologic inputs to tropical chan- in seasonal ice cover—freeze-over and break-up— nels, which tend to have a very flashy flow regime tend to be periods of the greatest channel and in smaller drainages and a prolonged peak flood floodplaingeomorphicchangeincold-regionrivers (exceeding 3 months) in very large basins such as the (Prowse and Beltaos, 2002), even though these Amazon and Congo. Smaller catchments are more periods do not coincide with the greatest sea- likely to have basin-wide intense storms than are sonal discharge. The enhanced erosion associated equivalently sized catchments at higher latitudes. with ice also increases the frequency of changes Monsoons, hurricanes, and ITCZ-related storms in channel cross-sectional geometry, planform, tend to cover sufficiently large areas that even larger and channel–floodplain connectivity (Ettema and catchments receive geomorphically significant rain- Kempema, 2012). falls at the same time (Scatena and Gupta, 2013). Widespread intense rainfall results in high percent- ages of contributing area and channel-modifying 8.4.2 Low latitudes discharges that occur simultaneously throughout the catchment. These characteristics contrast with Thesalientfeatureoflowlatitudesisthemagnitude the more spatially restricted precipitation inputs and and speed with which various fluxes occur (Wohl channel modifications of equivalently sized catch- et al., 2012a). Low latitudes here are synonymous ments at higher latitudes. with the tropics, the area of surplus radiative energy Continual high air temperatures and abundant bounded by anticyclonic circulations near 30◦ N vegetation combine with high values of precipita- and 30◦ S (Scatena and Gupta, 2013). This region tion to create rapid weathering of rock and soil includes the humid tropics, where average annual and of organic inputs such as wood. Large inputs rainfall is greater than potential evapotranspiration of material to channels occur in high-relief tropi- and precipitation is sufficient to support evergreen cal environments when cyclones or hurricanes trig- or semi-deciduous forests, and the dry tropics. The ger widespread landslides that strip hillslopes of dry tropics are sufficiently similar to temperate dry weathered rock and vegetation (Scatena and Lugo, regions to be treated in the next section. 1995; Goldsmith et al., 2008; Hilton et al., 2011b; The humid tropics can be further distinguished Wohl et al., 2012b) (Figure S8.5). Instream wood as seasonal and aseasonal (Gupta, 1995). The sea- does not persist in low-latitude rivers. Although sonal humid tropics have a marked seasonal con- individual pieces of wood and large jams can cre- centration of rainfall and runoff, typically reflecting ate important geomorphic and ecological effects the intertropical convergence zone (ITCZ) or mon- (Wohl et al., 2009, 2012b), the transience of wood soonal circulation patterns, and 80% of the annual asaresultofcombinedrapiddecayandhigh runoff can occur in 4 or 5 months of the year transport rates is particularly noticeable (Spencer (Scatena and Gupta, 2013). This can result in sub- et al., 1990; Cadol and Wohl, 2011; Soldner et al., stantial variations in river process and form between 2004). wet and dry seasons, including distinctly different Although the hydrology of low-latitude rivers has wet- and dry-season channel geometry (Wohl, 1992; distinctive characteristics and process–form inter- Gupta, 1995). The aseasonal humid tropics typically actions occur more rapidly and frequently than have mean annual rainfall between 2000 and 4000 in higher latitudes, a recent review concluded mm/year, whereas mean annual rainfall in the sea- that tropical rivers do not have diagnostic land- sonal humid tropics varies between 1000 and 6000 forms that can be solely attributed to their low- mm/year and interannual variability in runoff can latitude location (Scatena and Gupta, 2013). A key be large (Scatena and Gupta, 2013). distinction relative to rivers in higher latitudes is the 236 Rivers in the Landscape high frequency of geomorphically significant flows correspondingly substantial erosion and deposition. and channel changes. Larger river networks are likely to have more sus- tained flows. Downstream transmission losses in drylandchannels(Section3.2),alongwithlimited 8.4.3 Warm drylands contributing area where small convective storms affect only a limited portion of a large drainage The salient feature of interactions between process basin, create large spatial variability in discharge and and form within rivers in warm dryland regions temporal variability between storms (Tooth, 2013). is spatial and temporal discontinuities. As with Although ephemeral dryland channels can have tropical rivers, a recent review of dryland chan- less dense riparian vegetation than perennial rivers, nels concluded that there are no features unique dryland vegetation can be concentrated in and along to this subset of rivers, although certain charac- channels and can generate substantial flow resis- teristics are more common in drylands than else- tanceduringflows.Resistancecanincreasewith where (Tooth, 2013). Warm dryland here includes stage as woody vegetation becomes immersed, par- hyperarid, arid, semiarid, and dry subhumid envi- ticularly along the tops of banks and bars (Knighton ronments, but excludes cold, high-latitude or high- and Nanson, 2002). Resistance associated with ripar- altitude regions. Warm dryland regions are particu- ian vegetation can lead to a scenario of streambed larly prevalent within the subtropical, high-pressure scour at lower stages during a flood, when vege- belts of the northern and the southern hemisphere, tation is not inundated and thus contributes little and much of the scientific research on rivers in resistance, and fill during the higher stages when these regions has been conducted in the western bank flow resistance is strongly affected by vegeta- United States, Australia, the Middle East, and south- tion (Merritt and Wohl, 2003). ern Africa. Dryland rivers typically transport large quanti- Warm drylands share high, but variable, degrees ties of suspended and bed load and display strong of aridity that reflect low precipitation and high hysteresis of bed scour during the rising limb and potential evapotranspiration. Long periods with fill during the falling limb. Scour is facilitated by little or no precipitation and stream flow are the absence or poor development of coarse surface interrupted by intense rainfall and runoff that can layers.Poorlydevelopedcoarsesurfacelayers generate short-duration flash floods. Irregular pre- may reflect abundant upland sediment supplies, cipitation and low water tables keep many dryland enhanced sediment mixing during scour and fill, channels ephemeral or intermittent, although rivers high rates of bedload transport, and short-duration that originate in wetter mountainous highlands flows that minimize winnowing of fine particles before flowing into dry regions can be perennial from the bed (Tooth, 2013). Lack of coarse surface (Tooth, 2013). Short periods of high water and layersalsofacilitateshighratesofbedloadtransport sediment connectivity between uplands and chan- that increase more consistently with increasing nels can result from overland flow and limited flow because particles across a large size range are upland vegetation. Limited riparian vegetation and available for entrainment at the start of flow and the sediment cohesion can create unstable banks that bed becomes highly mobile at even modest flows promote a braided planform, and floodplains are (Tooth, 2013). limited in extent. Muchoftheresearchonsmall-tomedium- Flashfloodswithinchannelsoracrosspiedmonts size dryland channels emphasizes abrupt transitions are particularly characteristic of catchments less between incising and aggrading conditions across a than 100 km2 in size (Tooth, 2013). Flash floods channel network and through time. This emphasis is have steep rising and recessional limbs and may exemplified by the extensive literature on arroyos in last only minutes to hours, but can generate very the American Southwest (Graf, 1983, 1988; Harvey, high values of discharge per unit drainage area and 2008). Chapter 8 Rivers in the landscape 237

Repeated large floods appear to dominate pro- nal inputs of water, solutes, and sediment and ongo- cess and form in many dryland channels, as evi- ing adjustments within the river. Under these cir- denced by numerous studies of flood-related chan- cumstances, considerable insight can also be gained nel changes and recovery times of decades or longer by asking why, under variable external forcing, rivers (Tooth, 2013). Dryland channels may be particularly do not change even more. susceptibletochangeduringfloodsbecauseoflim- The second context for adapting Asimov’s phrase ited bank cohesion in the absence of dense riparian to rivers is that, although any river or river segment vegetation and abundant silt and clay, and because followsthebasiclawsofphysicsandchemistry,char- ofthelackofinterveningsmallerflowsthatcould acterizing feedbacks between channel process and modify flood-generated erosional and depositional form using a numerical equation or qualitative con- features (Tooth, 2013). ceptual model that applies to all rivers is limited In contrast to the emphasis on equilibrium condi- by the place-specific effects of lithology, tectonics, tionsalongperennialrivers,theabilityofrarefloods weathering regime, landscape history, river history, to cause substantial change along dryland rivers has theseasonalpresenceoficecoverorcyclones,and led some investigators to describe dryland chan- so forth. The only constant within a river network is nels as non- or disequilibrium systems (Graf, 1988). changes through time and space. The only constant Tooth’s (2013) review of dryland rivers, however, among river networks is river-specific changes in emphasized the global diversity of dryland river pro- the interactions between process and form. Recogni- cess and form as a result of varying degrees of aridity, tion of these characteristics further emphasizes the tectonic activity, and structural and lithological set- importance of understanding landscape context for tings (Tooth, 2000; Nanson et al., 2002). In addition, any particular river network or river segment. the identification of equilibrium or disequilibrium is highly dependent on temporal and spatial frames of reference, so that individual dryland rivers can 8.5 Rivers with a history exhibit both conditions (Tooth and Nanson, 2000). As explained in Chapter 5, a river is a physical sys- Isaac Asimov once wrote—perhaps in a moment tem with a history. The influence of previous cli- of frustration—that the only constant is change. maticandtectonicregimesonriverformandpro- This phrase can be adapted to rivers in at least cess can extend back in time beyond the Quaternary two contexts. First, natural rivers are continually because of the slow response of some aspects of river adjusting process and form—spatial distribution of networks to change. Among those characteristics of hydraulic forces; water chemistry; sediment entrain- rivers and drainage basins that can be particularly ment, transport and deposition; instream wood; resistant to changes are: bed configuration; channel cross-sectional geome- try; channel planform; reach gradient and longitudi- r topography, nalprofile—inresponsetochanginginputsofwater, r the spatial arrangement of river channels within a sediment,andwood,orinresponsetocontinuing network, development of the river. Indeed, one definition of a r relief ratio, natural, as opposed to a completely engineered, river r drainage density for river segments larger than isthatanaturalriverpossessesgeomorphicorphysi- first- and second-order channels, cal integrity. Graf (2001) defined physical integrity as r longitudinal profiles of rivers, and a set of active fluvial processes and landforms such r valley geometry. that the river maintains dynamic equilibrium, with adjustments not exceeding limits of change defined Consequently, inherited characteristics of these by societal values. River process and form reflect featurescancontinuetoinfluencecontemporary some balance between continually changing exter- process and form. 238 Rivers in the Landscape

Amazon

Amazon Ganges Mekong Mississippi Paraná Magdalena

Longshore transport Major depositional basin

Mountains marginal to Marginal to large craton folded mountains Along strike of Mississippi folded mountains Columbia Congo Danube

Superimposed over Danube Large craton mountain chains Figure 8.5 Types of big river settings. Lighter gray shading indicates basement bedrock; darker gray shading indicates major depositional areas. (Adapted from Potter 1978, Figure 9.)

An example comes from continental-scale con- be subdivided into reaches that differ in gradient trols on large rivers. The location of large rivers pre- and sinuosity where the river crosses more erosion- dominantly follows structural lows such as deep- ally resistant Tertiary-age sediments and fault zones, seated rifts, aulacogens (the failed arm of a triple even though many of these faults show relatively junction in a crustal rift zone), and major fracture little recent activity (Schumm et al., 2000). Rivers systems. Potter (1978) discussed how the continen- with drainage areas larger than 10 km2 in the Cen- talsettingsoflargeriversfallpredominantlyintoone tral Apennines of Italy have long-profile convex- of the five settings (Figure 8.5). Large rivers drain- itieswheretheycrossfaultsthathaveundergone ing mountains marginal to a large craton include the an increase in displacement during the past million Amazon and Mississippi Rivers. Large rivers flow- years, whereas those crossing faults with constant ing parallel to folded mountains include the Ganges displacement rates lack such convexities (Whittaker and Parana.´ The Mekong and the Magdalena are et al., 2008). large rivers flowing along the strike of folded moun- Landscape configurations or persistent erosional tains. Large rivers can be superimposed over moun- and depositional features relict from Quaternary tain chains, as in the case of the Columbia and the glaciation provide another example of how past Danube, or large rivers can flow across a large cra- events continue to influence river form and process. ton, as in the case of the Congo. Glaciated mountains can have distinctly different At smaller spatial scales, a single river channel valley geometry above and below the elevation lim- flowing across different lithologies, structural fea- its of Pleistocene valley glaciers, with greater cross- tures, or tectonic zones can exhibit striking differ- sectional area and steeper valley walls in glacial val- ences in river and valley geometry and rate of inci- leys relative to fluvial valleys (Montgomery, 2002; sion in response to these geological controls that Amerson et al., 2008). Glaciated and fluvial portions formed millions to hundreds of millions of years ofamountainrangecanalsobeerodingatdiffer- ago. Examples include the lower Mississippi River ent rates in response to the effects of differing sedi- in the eastern United States. This sinuous river can ment supply and base level controls (Anderson et al., Chapter 8 Rivers in the landscape 239

2006c). Tributary glacial valleys that eroded to a base megafloods, most were associated with ice-marginal level defined by the upper level of the main valley lakes or water released from within the ice sheet. glacier can persist as hanging valleys with large verti- Among the megafloods documented thus far (Baker, caldropsbetweenthetributaryvalleymouthandthe 2013) are those of: mainvalleyfloorforthousandsofyearsafterglacial ice retreats. r the Channeled Scabland in Washington, USA; Recessional moraines are persistent depositional r theLaurentideIceSheetice-marginallakesin features perpendicular to valley orientation that north-central North America, which flowed down create local base levels and valley segments with theMississippi,St.Lawrence,Mackenzie,and lower river gradient and finer substrate than seg- Hudson Rivers; mentsimmediatelyup-anddownstream,evenlong r the Patagonian Ice Sheet of southern Argentina afterthemoraineisincisedbyariver.Reces- and Chile; sional moraines that fill with meltwater as valley r Icelandic jokulhlaups;¨ glaciers retreat can fail catastrophically. Although r southwarddrainagefromtheFennoscandianIce individual moraines fail only once, failure of suc- SheetthatinfluencedtheEnglishChannelandthe cessive moraines along a valley can create numer- North Sea; and ousoutburstfloodsalongarivernetworkover r the northern mountain areas of central Asia, periodsofdecadestocenturiesasmorainessuc- including Kirgizstan, Mongolia, and Siberia. cessively up-valley are abruptly drained. The dis- charge and stream power of these outburst floods These exceptionally large floods created erosional typically greatly exceed discharge and stream power and depositional features of such large magnitude generated during “normal” floods induced by rain- andextentthatsubsequentgeomorphicprocesses fall or snowmelt (Cenderelli, 2000; O’Connor et al., duringtheHolocenehaveonlypartially—orinsome 2013), and normal floods may be largely incapable cases little—modified megaflood terrains. of modifying the outburst-flood erosional and depo- More recent, Holocene history can also continue sitional features (Cenderelli and Wohl, 2001). Por- to influence river process and form. River response tions of a valley subject to repeated outburst floods to disturbance partly depends on time elapsed since canalsobecomelessresponsiveinthatearlierfloods the last disturbance of a similar magnitude. Investi- have already modified valley morphology to convey gating erosion along mountainous headwater catch- exceptionally large discharges (Cenderelli and Wohl, ments in which wildfires and subsequent rainfall 2003). induced debris flows, Wohl and Pearthree (1991) The continental-scale ice sheets that covered found that a particular debris flow strongly influ- portions of North America, northern Europe, and enced channel morphology only if a minimum northern Asia also left enduring signatures on river periodhadpassedsincethelastdebrisflow.This networks, diverting existing channels, altering water minimum time interval allowed sufficient sediment and sediment supply to channels beyond the ice to accumulate in the channel and then be eroded by margins, and changing local river gradients via the next debris flow. isostatic flexure of the crust. One of the most Human use of resources can also continue to spectacular categories of ice sheet effects on river influence river process and form long after the par- networks is the occurrence of megafloods dur- ticular human activity has ceased. Earlier chap- ing periods of glacial retreat. Megafloods are rel- ters provided numerous examples of such influ- atively short-duration flows that constitute the ences, including milldams along rivers in the largest known freshwater floods, with discharges eastern United States (Walter and Merritts, 2008). that generally exceed 1 million m3/s (Baker, 2013). Sediment accumulated behind these milldams is Although other mechanisms, such as failure of rock now influencing river form and process, as well dams or caldera lake impoundments, can generate as downstream sediment and nutrient loads, more 240 Rivers in the Landscape than a century after the milldams ceased to be erated by compressional tectonics in the late Cre- used and maintained. A 1998 study in the south- taceous and early Tertiary (Erslev, 2001), and has ern Appalachian Mountains of the United States been relatively quiescent for the past 40 million years found that whole watershed land use in the 1950s (Anderson et al., 2006c). Despite contemporary tec- was the best predictor of contemporary inverte- tonic stability, the Precambrian-age crystalline rocks brate and fish diversity because of persistent effects that constitute most of the Front Range underwent on aquatic habitat (Harding et al., 1998). Study- repeated episodes of uplift and erosion following ing streams down which cut logs had been floated the Precambrian, and are now densely jointed. Pre- for railroad ties in the Medicine Bow National For- cambrian shear zones that were reactivated during est of Wyoming, USA, Young et al. (1994) found themostrecentorogenycreateparticularlydensely that, a century after the log floating had ceased, jointed bedrock that correlates with the location of these streams had less instream wood, lower densi- wider valley segments with lower downstream gra- ties of large riparian trees, lower channel complexity, dients, and the creation and preservation of strath a greater proportion of riffles, and fewer pools than terraces (Ehlen and Wohl, 2002; Wohl, 2008a). did otherwise analogous streams that were not used Exhumation of the adjacent, sediment-filled for log floating. As noted earlier, 200 years may be basins beyond the mountain front began sometime required before instream wood volumes completely during the late Cenozoic, and many of the basins recover following timber harvest (Bragg et al., 2000). havenowbeenerodedtodepthsof1kmorgreater The following sections provide a few further, in- (Leonard, 2002). Basin exhumation lowers base level depth examples of how the history of geology (lithol- for the rivers draining the Front Range, creating ogy, structure, tectonics), climate, biota, and human knickpointsorknickzonesthatcurrentlylieatvary- resource use within a catchment influence contem- ing distances upstream from the mountain front, porary river process and form. These place-specific partly as a function of differing drainage area, dis- examples are largely drawn from my own research charge, and incisional capabilities on the rivers trib- experience because I can write most effectively of utary to the South Platte (Anderson et al., 2006c). places with which I am familiar. None of the details The upper elevations of most valleys were also are unique, however, to these case studies. Many influenced by Pleistocene valley glaciers (Madole rivers of the semiarid western United States share et al., 1998). In addition to eroding bedrock valley someaspectsofthehistoryoftheUpperSouthPlatte boundaries and creating knickpoints along tributary River,forexample,justasotherriversinthehumid valleys, the glaciers deposited large moraines that tropics share some components of the history of the continue to create local base levels for upstream val- Rio Chagres. Together, these case studies illustrate ley segments (Wohl et al., 2004). Wide, low gradi- theimportanceofbeingawareofthegreaterland- ent valley segments immediately upstream from the scape context when interpreting and managing river terminal moraine are likely to host beaver mead- process and form. ows where historically abundant beaver populations created numerous dams, multi-thread channel net- works, and thick sequences of organic-rich fine sed- 8.5.1 Upper South Platte River iments (Westbrook et al., 2006, 2011; Kramer et al., drainage, Colorado, USA 2012; Polvi and Wohl, 2012; Wohl et al., 2012c). The lower limit of Pleistocene glaciation, TheUpperSouthPlatteRiverdrainageincludes ∼2300 m elevation, corresponds to a transition in numerous tributaries that drain east from the crest modern hydroclimatology. Above this elevation, of the Colorado Front Range toward the adjacent snowmelt dominates the annual streamflow regime lowlands of the Great Plains and the Missouri– and discharge per unit drainage area seldom exceeds Mississippi River (Figure S8.6). The Front Range, 1.1 m3/s/km2 (Jarrett, 1989). Below this elevation, like other mountain ranges in the region, was gen- snowmelt continues to be the primary source of Chapter 8 Rivers in the landscape 241

annual peak flow, but late summer convective 2005). As at lower elevations, these fires can trigger storms can produce localized flash flooding that hillslope instability. generatesasmuchas40m3/s/km2 of discharge People have lived in the Front Range for at (Jarrett, 1989). These flash floods generate suf- least 12,000 years (Benedict, 1992), but the hunter– ficiently high values of shear stress and stream gatherer cultures of the early inhabitants do not power to create substantial and persistent erosion appear to have altered river process and form. in deep, narrow valley segments and deposition in Human impacts to rivers started with the advent wide, lower-gradient segments (Shroba et al., 1979). of beaver trapping for furs during the first decades Floods with a recurrence interval of hundreds of of the nineteenth century, and accelerated with the years thus create the valley-bottom template at 1859 discovery of placer gold in several tributaries of middle to lower elevations in the Front Range. the Upper South Platte River (Figure S8.7), and the Hillslope disturbance regimes also influence associated deforestation, construction of roads and valley-bottom process and form. Primary sources of railroads (Figure S8.8), floating of cut logs down- hillslope instability are intense precipitation at eleva- stream to processing areas (Figure S8.9), flow reg- tions below 2300 m, or disruption of the forest cover ulation for agricultural water supply (Figure S8.10), by wildfire at all elevations. Steppe vegetation at the and urbanization (Wohl, 2001). Although a few rel- base of the mountains gives way to montane forests atively unimpacted, reference rivers remain in the (1830–2740 m elevation), with open ponderosa pine Front Range, process and form have changed sub- (Pinus ponderosa) woodlands in the lower mon- stantiallyduringthepasttwocenturiesalongthe tane (1830–2350 m) and slightly more dense mixed majority of rivers. This can be illustrated by consid- conifer forests that include Douglas fir (Pseudot- ering the effects of instream wood. suga menziesii) in the upper montane (2440–2740 River segments flowing through old-growth for- m).Frequent,low-severityfiresthatburnmainlythe est, and which experienced minimal historical dis- ground surface over areas of approximately 100 ha ruption from activities such as placer mining, tend recur at intervals of 5–30 years in the lower montane to have greater volumes of instream wood per unit zone (Veblen and Donnegan, 2005). Because the channel area, and larger and more closely spaced canopy and the root structure of the forest remain channel-spanning logjams (Beckman, 2013; Wohl intact, these fires are unlikely to trigger widespread and Beckman, 2014). Larger instream wood loads hillslope instability, although destruction of the sur- correspond to more physical diversity of channel face organic layer can create local slopewash and planform along the length of a river (Wohl, 2011c; rilling. Polvi and Wohl, 2013), greater storage of nutrients A complex pattern of mixed low- and high- (Wohl et al., 2012c), and greater abundance and severity fires that burn areas of approximately 100 diversity of riparian and aquatic habitat (Richmond ha and recur at intervals of 40–100 years occurs and Fausch, 1995). In the absence of wood, chan- inthemiddleandtheuppermontanezone.High- nels tend to be much more uniform with respect severity fires can trigger widespread hillslope insta- to cross-sectional geometry (e.g., fewer pools, and bility when unconsolidated hillslope sediment is a more continuous riffle-run morphology), plan- exposedtosnowmelt,rainfall,anddryravel. form (almost entirely single-thread channels, some- The subalpine forest (2740–3400 m) includes times with lower sinuosity), and aquatic and ripar- Engelmann spruce (Picea engelmannii), subalpine fir ian habitat (Figure S8.11). Water, dissolved and (Abies lasiocarpa), lodgepole pine (Pinus contorta), particulate nutrients, and sediment all move more aspen (Populus tremuloides) and other species. The rapidly downstream in the absence of obstructions subalpine forest is characterized by infrequent, high- and in-channel storage associated with instream severity fires that kill all canopy trees over areas wood, and channel–floodplain connectivity and of hundreds to thousands of hectares at inter- hyporheic exchange decrease. Pronounced longitu- vals greater than 100 years (Veblen and Donnegan, dinal variations in valley geometry associated with 242 Rivers in the Landscape

Precambrian shear zones and Pleistocene glaciations Consequently, the Upper Rio Chagres has received strongly influence wood transport and storage at more scientific attention than many of the water- channel lengths of 101–102 m(WohlandCadol, sheds in the neotropics of Central America. 2011), and wildfire and other forest disturbances Panama is an east-west-oriented isthmus between influence wood recruitment to rivers (Wohl, 2011b), NorthandSouthAmericacomposedofadiverse but the historical legacy of nineteenth and twenti- suite of geological units created and assembled since eth century human activities along these rivers dom- the late Cretaceous (Harmon, 2005b). The igneous inates instream wood loads and, consequently, river rocksthatcomprisemostofPanamaformedduring process and form. the Tertiary as an oceanic plateau and volcanic Effective management and restoration of rivers in islandarccomplex.Thisarcbegantocollidewith theUpperSouthPlattedrainagemustrecognizeat northwestern South America approximately 10 mil- least four salient factors: lion years ago and the land bridge between the two continents was in place around 3 million years ago r 1 2 substantial reach-scale (10 –10 m) variation in (Harmon, 2005b). At present, the Pacific side of process and form imposed by geologic factors Panama is a geologically active margin with a (Figure S8.2), which means that many river char- deep oceanic trench, narrow marine shelf, active acteristics are best described using the conceptual subduction, volcanic activity, and , model of process domains; r whereas the Atlantic side is a passive, stable margin episodic natural disturbances, primarily in the (Harmon, 2005b). The bedrock of the Upper Rio form of debris flows and floods associated with Chagres basin is mainly a mixture of volcanic and wildfires and/or convective storms, which create intrusiverocksthatrangefromfelsictoultramafic, persistent erosional and depositional features that most of which are strongly deformed and chemically shape valley and channel geometry at time spans altered (Worner¨ et al., 2005). of 101–103 years; r The Rio Chagres basin was divided into an upper simplification and homogenization of most rivers and a lower basin by completion of the Panama as a result of nineteenth and twentieth century Canal and associated dams in 1914, which flooded human activities, several of which (e.g., flow regu- a significant portion of the original Chagres basin lation) continue today; and r (Harmon, 2005a) (Figure S8.12). The 1936 construc- ongoing climate change toward a warmer, drier tion of Madden Dam formed Lake Alhajuela: the climate with smaller annual snowpack, earlier Upper Rio Chagres drains 414 km2 above this lake. snowmelt, changes in forest composition associ- The river network of the Upper Chagres is deeply ated with changing climate and disturbance (wild- incisedintoanextremelysteepterrain.Hillslopes fire, insects, blowdowns), and increased human exceed 45 degrees in over 90% of basin, most of consumptive demands for surface waters. which is covered by intact primary tropical and sea- sonally tropical rainforest. Trees can attain a height Under these circumstances, maintaining and of 30 m and a diameter of 2.2 m (Wohl, 2005). Mean restoring the physical and ecological integrity of annual precipitation at the basin outlet is 2590 mm, rivers in the Upper South Platte River drainage 90% of which falls during May to December. Rain- presents a substantial challenge. fallintheupperwatershedisunmeasured,butlikely closer to 5000 mm based on vegetation type. 8.5.2 Upper Rio Chagres, Panama Soil surfaces are relatively dry and water repellent at the start of the wet season, and soil tension cracks The Upper Rio Chagres is the largest headwater have developed. Landslides and tree falls create per- basin within the 2982 km2 watershed of the Panama sistent, uneven hillslope topography and numer- Canal, and supplies almost half of the water needed ous small surface depressions. These depressions for the operation of the canal (Harmon, 2005a). accumulate water relatively quickly during rains at Chapter 8 Rivers in the landscape 243

(a) (b)

Figure 8.6 Views of diverse segments of Upper Chagres River formed in (a) mafic bedrock (gradient 0.014, channel width 30 m), and (b) felsic rocks (gradient 0.001, channel width 35 m). the start of the wet season, and these rains produce Chagres basin, and is likely never more than 2–3 m anomalously high runoff volumes. Water repellent below the channel. Mafic units are the most resistant soils limit infiltration from the depressions, which, to weathering and erosion, and correspond to steep, along with the soil cracks, feed a well-developed narrow valley segments and knickpoints (Wohl and network of soil pipes that rapidly deliver runoff to Springer, 2005). Granitic rocks are more readily stream channels. As the wet season continues and weathered and correspond to straighter, wider val- the soil surface becomes moist, water repellency and ley segments (Worner¨ et al., 2005) (Figure 8.6). soil cracks disappear, infiltration into the soil matrix Downstream hydraulic geometry relations are well fills the soil water storage reservoir, less runoff developed despite these local variations, with expo- moves downslope via soil pipes, and runoff volume nents similar to average values for rivers worldwide per precipitation input decreases. Runoff volume (Wohl, 2005) (Figure S8.13). This suggests that high again increases later in the wet season, however, as values of discharge per unit drainage area gener- saturated soils create saturated overland flow (Hen- ate sufficient erosional force to reduce variations drickx et al., 2005). Runoff efficiency of a storm— in channel geometry associated with local differ- defined as the ratio of the volume of runoff to the ences in rock erodibility, although poor correlations volume of rainfall—varies from 6% to 59% during between grain size or reach gradient and drainage the wet season in the Upper Chagres, with an aver- area (Rengers and Wohl, 2007) suggest that land- age of 40% (Niedzialek and Ogden, 2005). slidesandbedrocklithologystronglyinfluencesome Sequential changes in hillslope hydrology dur- aspects of river process and form. ing the wet season are reflected in streamflow. Landslides are frequent and widespread through- Theincreaseinsaturationoverlandflowlaterin out the Upper Chagres (Figure S8.14), and are the wet season, for example, appears as increas- an important mechanism of instream wood ingly higher quasi-stable base flows (Niedzialek and recruitment. Wood recruitment and retention are Ogden, 2005). Unit discharges in the headwaters extremely spatially and temporally episodic along portionofthebasincanbequitelarge:heavyrainfall the Upper Chagres. High decay rates of wood and in 2007 resulted in a peak discharge of 41 m3/s/km2 large fluvial transport capacity combine to create in a 20.6 km2 catchment (Wohl and Ogden, 2013). extremely low “background” levels of instream Bedrock is discontinuously exposed along the wood. When intense rainfall generates landslides, bed and banks of channels throughout the Upper however, large volumes of wood can be abruptly 244 Rivers in the Landscape introduced to the river network. Widespread r high rainfall, runoff, and discharge per unit area, landsliding and wood recruitment can result in r widespread, frequent landslides and introduction the formation of channel-spanning logjams and of sediment, wood, and organic carbon to rivers, upstream wedges of sediment accumulation in r high transport capacity within rivers and atten- tributary channels (Figure S8.15) and at sites of dant rapid changes in erosional and depositional reducedtransportinthemainchannel,suchas features along channels, and bends or expansions (Wohl et al., 2009). These jams r arivernetworkthatislikelytobehighlyrespon- persist for only 1–2 years, although longitudinally sive to changes in land cover and climate. discontinuous fill terraces 1–2 m above the active channel may persist when a jam breaks up and inci- sion removes sediment stored above the jam along 8.5.3 Mackenzie River the channel axis (Wohl et al., 2009). Infrequent drainage, Canada storms that generate greater rainfall over much of the upper basin trigger even more abundant The Mackenzie River drains 1.8 million km2 in landslides and wood recruitment, but sustained CanadaandisthefourthlargestArcticriverwith high flows limit formation of logjams and instead respect to water discharge and the largest in terms flush the newly recruited wood all the way through of sediment discharge (Holmes et al., 2002; Rachold the Upper Chagres river network and into Lake et al., 2004). The mainstem Mackenzie flows north Alhajuela (Wohl and Ogden, 2013). This represents to the Arctic Ocean from Great Slave Lake, but pri- a substantial flux of organic carbon from the basin. marytributariesextendmuchfarthersouth(Fig- Wohl and Ogden (2013) estimated values of 24 Mg ure S8.16). Permafrost underlies two-thirds of the C/km2,whichisanorderofmagnitudehigherthan Mackenzie drainage (Heginbottom, 2000), and the background rates of wood-based carbon export presence of this permanently frozen ground strongly from other catchments. influences hydrology and geomorphology within In situ-produced cosmogenic 10Be analyses indi- the region. The drainage is unique in containing 8 of cate sediment generation rates averaging ∼270 tons/ the 15 ecozones identified in Canada (de Rham et al., km/year in the Upper Chagres, a value compara- 2008a), but much of the drainage basin (79%) is cov- bletootherregionswithhighrainfallandrunoff ered by forest, and the river has long been recognized (Nichols et al., 2005). Adjacent drainages with exten- as a primary source of wood export to the Arctic sive land clearance have twofold to threefold greater Ocean (Eggertsson, 1994). Like other high-latitude, sediment yields per unit drainage area (Nichols forested catchments, the Mackenzie also stores sub- et al., 2005). Land clearance represents the great- stantial carbon in forest and wetland ecosystems and est immediate potential source of changes in sed- associated peat deposits (Dixon et al., 1994). iment and water yield, water quality, and carbon The western Mackenzie drainage includes indi- stocks within the Upper Chagres (Dale et al., 2005). vidual ranges of the Canadian Rockies, whereas the Although much of the watershed lies within the Par- central and eastern portions of the basin are on the que Nacional Chagres, illegal clearing and settle- low-relief Canadian Shield and Interior Plains (Woo ment within the basin have increased during the and Rouse, 2008). The western mountainous terrain past decade as human population increases rapidly and the central and eastern plains have distinctly within the region. different pathways for water and sediment moving As in the Upper South Platte drainage, effec- from uplands and into the river network. tive management and restoration of rivers in the The lowlands of the Mackenzie drainage are an Upper Chagres drainage basin requires recogni- arid to semiarid region that receives relatively lit- tion of fundamental river processes governed by tle precipitation, but nonetheless contains tens of the geologic and climatic setting. This setting thousands of small lakes (Figure 8.7). This apparent creates paradox reflects low evaporation rates and, for very Chapter 8 Rivers in the landscape 245

The mountainous uplands also have extensive permafrost, but the steeper terrain likely contributes to a runoff pattern more like the source area con- cept (Section 2.2.1). On a per unit area basis, the mountainous areas contribute more runoff than the lowlands (Woo and Thorne, 2003). The uplands also contribute substantial sediment to the mainstem, and much of this sediment enters the river network via landslides (Figure S8.17). A 1990s inventory of 3400 landslides within the Mackenzie drainage found that 69% of the landslides occurred in uncon- solidated sediments. Landslides are concentrated along the banks of the Mackenzie and its tributary channels where the rivers have eroded Quaternary sedimentsandalongsteepslopesinthemountains west of the Mackenzie River (Aylsworth et al., 2000). Runoff in the Mackenzie drainage is dominated by high spring flows, but the presence of seasonal ice creates unique stage-discharge relations (Hicks and Beltaos, 2008). The additional hydraulic resis- tance of a stable ice cover elevates water levels in a channel (Prowse and Beltaos, 2002), and this effect is greatest when the ice cover is hydraulically most Figure 8.7 An aerial view of widely scattered small rough, such as during freeze-over and break-up. lakes, bedrock outcrops, and patches of forest in the low- Pronounced seasonal changes in hydrographs also lands of the eastern Mackenzie River drainage basin. reflect storage of significant quantities of water as ice during freeze-over and subsequent release dur- ing break-up (Prowse and Ferrick, 2002). Early win- shallow lakes, impermeable permafrost that pre- ter freeze-over can create the lowest flow of the vents infiltration of precipitation or meltwater when year,eventhoughactualrunoffproductionreachesa soil ice thaws during summer. minimum later in the winter. This apparent discon- The element threshold concept described in Sec- nect results from three factors: reduced contributing tion 2.2.1 was developed for the lowlands of the area because anchor ice frozen to the river bed cuts Mackenzie drainage. The central tenet of this con- off groundwater inflows; storage of water within ice; cept is that spatially and hydrologically distinct areas and backwater storage upstream from ice-covered such as bedrock uplands, soil-filled valleys, and lakes river segments in which the ice reduces the chan- create a mosaic on the landscape. Each area, or ele- nel cross-sectional area (Prowse and Beltaos, 2002). ment, within the mosaic can store, contribute, or Releaseofstoredwaterandmeltoftheicecovercan transmit water, and the properties of each element significantly contribute to spring peak flows (Prowse are not necessarily synchronized with those of adja- and Carter, 2002). Peak water levels can be enhanced cent elements, so that a bedrock upland can be con- byicejamming,whichcanalsocauseaccelerated tributingwaterwhileanadjacentlakeisstoring channel and bank erosion (Goulding et al., 2009a; water, and an upslope element may contribute over- Ettema and Kempema, 2012). land flow before a downslope element (Spence and Peak flows can create extensive overbank flood- Woo, 2006). The lowlands contribute sediment to ing in portions of the drainage such as the Macken- the river network primarily through bank erosion. zie Delta, where overbank flows strongly influence 246 Rivers in the Landscape hydrologic and biogeochemical exchanges between up, lower peak discharge, slower rate of rise in dis- channels and delta lakes (Goulding et al., 2009b; charge, and lesser ice thickness in the Mackenzie Prowse et al., 2011). The delta covers 12,000 km2 Delta (Goulding et al., 2009a). Analysis of break- and contains more than 25,000 lakes and a sinu- up during the period 1913–2002 also indicates sig- ousmazeofdistributarychannels(Gouldingetal., nificantly earlier trends in the timing of break-up 2009a, 2009b). Mean annual flow through the delta in upstream portions of major tributaries of the is 8300 m3/s, but peak flows can reach 14,000 m3/s Mackenzie drainage (de Rham et al., 2008b). during May and June (Mackay, 1963). Hydraulic Despite the low population density and remote forces, along with transport of ice blocks (Gould- location of most of the Mackenzie drainage, we have ing et al., 2009a, 2009b) and wood, facilitate con- a relatively extensive knowledge base for this river stant bank erosion (Figure S8.18) and sedimenta- network partly because of the history of proposed tion. An estimated 128 million tons of sediment and actual resource extraction. Starting with a pro- reach the delta annually from the Mackenzie River posed natural gas pipeline during the 1970s that and the tributary Peel River (Carson et al., 1998). led to a range of baseline environmental studies, Approximately 103 million tons are at least tem- and continuing with additional proposals for oil and porarily deposited in the delta, of which 43 mil- gas extraction and transmission corridors at present, lion tons remain: the rest moves from the delta into we know more about climate, hydrology, biogeo- Mackenzie Bay and the Beaufort Sea (Carson et al., chemistry, forest ecology, and geomorphology in the 1998). The main distributary channels of the delta Mackenzie drainage than in other Arctic drainages have nonetheless remained relatively stable since such as the Yukon in North America and the major they were mapped by the Franklin expedition in rivers of Siberia. Effective management of rivers in 1826, which may reflect the stabilizing presence of the Mackenzie drainage requires recognition of dis- permafrost. As the permafrost underlying the region tinctive factors associated with the geologic and cli- melts, rates of channel avulsion, bank erosion, and matic setting, including wood recruitment have the potential to increase dra- matically. r permafrost that limits infiltration while intact, TheMackenziedrainageisattheforefront r seasonal ice cover that alters flow resistance, water of environmental changes associated with global storage, downstream transmission of runoff dur- warming, as are other rivers draining to the Arctic ing snowmelt, and timing and magnitude of peak (Guo et al., 2007). Increases in average air tempera- flows, ture will cascade through arctic ecosystems in com- r differences in runoff and sediment yields from plex ways, altering the distribution of permafrost, mountainous versus lowland portions of the surface water hydrology (especially frequency and drainage, and magnitude of ice jam floods), vegetation, and bio- r rapidchangesinallcomponentsofthehydrologic geochemical fluxes (Francis et al., 2009a; Frey and cycle as global air temperatures rise. McClelland, 2009). As permafrost recedes, thaw- ing of frozen organic soils and stream banks will releasecarbonintoriversviagroundwaterpath- 8.5.4 Oregon Coast Range, USA ways, surface erosion, bank instability, and land- slides (Dyke et al., 1997; Holmes et al., 2008). Warm- TheOregonCoastRangeisarugged,deeplydis- ing air temperatures will not only influence precip- sected mountain range in the northwestern United itation and runoff, but will also alter the dynamics States (Figure S8.19). The range sits over the Cas- of river ice by delaying freeze-over and changing cadia subduction zone and has experienced vari- the timing of break-up (Prowse and Beltaos, 2002). able uplift rates during the past 20–30 million years. Records during the period 1974–2006 indicate a Rock uplift rates average 30–300 m per million trend toward longer melt interval prior to break- years, basin denudation rates average 50–80 m per Chapter 8 Rivers in the landscape 247 millionyears,andbedrockloweringaverages70m within this zone are also steeper than in adjacent per million years (Heimsath et al., 2001). The range areas and may experience more frequent landslides is underlain by a thick sequence of arkosic sandstone (Kobor and Roering, 2004). and siltstone, with minor areas of mafic volcanic and Landslides and debris flows strongly influence intrusive rocks (Personius et al., 1993). Summits in river form and process in the Oregon Coast Range. the Oregon Coast Range lie below 1250 m elevation, Terrain prone to large landslides has lower values of and the range was not glaciated during the Pleis- drainage density (Roering et al., 2005). Debris-flow tocene. Peaks and ridges are soil mantled, with a recurrence intervals vary from ∼100 to 400 years in typical profile of colluvium overlying saprolite, frac- headwater basins (May and Gresswell, 2004). Once tured bedrock, and then intact bedrock that can be debris flows enter the river network, they typically 1.5 m thick (Heimsath et al., 2001). strip the steeper channel segments to bedrock, evac- The combination of active uplift, readily erodi- uating stored sediment and instream wood. These ble sedimentary rocks, and wet climate (mean channel segments can remain bedrock channels for annual precipitation averages 200 cm) creates abun- substantial periods of time, but wood that accumu- dant landslides and debris flows. The Coast Range lates in the channel can effectively trap sediment, is densely forested, and landslides occur in areas and low-order channels with wood can form one of reduced root strength associated with diverse of the more important sediment storage reservoirs vegetation (Roering et al., 2003). Landslides also in this steep landscape (May and Gresswell, 2003). correlate with greater amounts of siltstone rel- Tributary basins with larger drainage areas have ative to sandstone, and occur preferentially on more potential landslide source area and a greater slopes on which the downslope aspect corresponds frequency of scouring debris flows than smaller to the direction of bedrock dip (Roering et al., basins, facilitating the formation of larger, more per- 2005). Wildfires are also important in destabilizing sistentdebrisfansatchannelconfluences(Mayand slopes. Post-fire erosion rates exceed long-term rates Gresswell, 2004). (∼0.1 mm/year) by a factor of six, and numerical Debris flows commonly dam headwater val- simulations suggest that fire-related processes may leys with deposits of sediment and wood, causing account for ∼50% of sediment production on steep upstream alluviation that covers bedrock channel slopes (Roering and Gerber, 2005). segments (Lancaster and Grant, 2006). These tem- Rivers in the Oregon Coast Range flow mostly on porary sediment storage sites are evacuated at rates bedrock or thin alluvial beds bordered by bedrock, that reflect process transitions. Channel segments and bedrock straths and discontinuous terraces are higher in the river network that experience fre- widespread (Personius et al., 1993). An episode of quent debris-flow erosion have shorter transit time regional aggradation at the start of the Holocene for sediment than segments lower in the network has been interpreted to reflect climate-induced that experience only fluvial erosion (Lancaster and changesinthefrequencyofevacuationofcollu- Casebeer, 2007). Although most sediment in both vium from hillslope hollows (Personius et al., 1993) portions of the network has relatively short tran- and increased fire frequency and sediment yield sit times (<600 years), significant volumes of sedi- (Roering and Gerber, 2005). mentcanremainforthousandsofyears,sothatthese Spatial variations in the gradient and longitudi- valley-bottom sediment reservoirs buffer down- nal profile of rivers in the Oregon Coast Range have stream aquatic habitat from hillslope disturbance been used to infer the location of synclinal tilting (Lancaster and Casebeer, 2007). and uplift (Rhea, 1993). A zone approximately 20 km The Oregon Coast Range has also been an area of wide in which channel gradients are about twice the intensive timber harvest for more than a century. As regional average coincides with the strike of N-S- noted in Chapter 7, timber harvest typically results trending folds, and likely reflects differential rock in substantial changes to water and sediment yield, uplift (Kobor and Roering, 2004). Colluvial hollows as well as instream wood recruitment. In the Coast 248 Rivers in the Landscape

Range, instream wood forms about half of the pools in old-growth forests, but fewer pools in indus- 8.5.5 Yuma Wash, Arizona, USA trial forests, which have substantially fewer jams per length of channel (Montgomery et al., 2003b). Log- Yuma Wash is a mixed sand- and gravel-bed chan- jams that do form within river networks in industrial nel that drains 186 km2 in southwestern Arizona, an forests tend to be associated with debris flows and extremely arid region (Figure S8.20). The channel to form at river confluences, in contrast to the more is formed primarily in alluvium, but bedrock out- widely distributed logjams of old-growth forests. crops discontinuously along the upper reaches of the The widely distributed jams in old-growth river net- channel network (Figure S8.21). Headwater reaches works facilitate storage of channel alluvium and thus are relatively narrow, single-thread channels incised dampen bedrock incision rates. Jams in old-growth into bedrock or cohesive alluvium cemented by sec- areas also contribute to the presence of three pro- ondary carbonate. Further downstream, Yuma Wash cess domains (Montgomery et al., 2003b). (1) An grows progressively wider, until valley width exceeds alluvial zone exists in the lower portions of main- 450mandthechannelbecomesbraided.Theouter stream channels where channel and valley gradients edges of the braided channel pattern are defined by are sufficiently low to permit a persistent sediment Holocene and Pleistocene terraces with steep risers layer over the bedrock. (2) A forced alluvial zone is of indurated alluvium. present in the middle reaches of the river network, Vegetation cover in the surrounding uplands is where instream wood increases flow resistance and only 1%–5%, but averages 31% within the mid- creates backwater effects that allow sediment cover dle and lower portions of the channel network to persist along valley segments that would other- (Merritt and Wohl, 2003). Dominant vegetation wise have exposed bedrock in the streambed. (3) A includes woody xeric species, succulents and grasses zone dominated by debris flows occurs in headwa- (Merritt and Wohl, 2003). Clumps of vegetation ter channels with drainage areas less than ∼1kmand form linear or lemniscate patterns on depositional channel gradients >0.2. surfaces in the valley bottoms. Woody species grow- In the Oregon Coast Range, identifying the fac- ing on bars and adjacent to active channels are resis- tors that correlate with greater landslide activity tant to removal by scour and are adapted to survive facilitates identification of the areas of greatest land- periodic flooding (Figure S8.22). scapechange,aswellasareaslikelytobemostsen- The mainstem and all of the tributaries are sitive to climate variability and land use. Effective ephemeral, and flows occur less frequently than management and restoration of rivers in the Coast once a year. The region receives a total annual aver- Range requires recognition of age rainfall of 93 mm (NOAA, 1998) from convec- tive thunderstorms, frontal systems, and dissipating r the importance of tectonics, lithology, and climate tropical storms. Most of the summer rainfall comes in promoting hillslope instability, from isolated, fast-moving convective storms that r the resulting spatial and temporal variability in cause localized flash flooding. Frontal storms dur- sediment supplied to rivers, ing autumn and winter create more widespread but r patterns of sediment storage and evacuation typically low-intensity rainfall that generates little within river networks that reflect process stream flow. Floods with the longest duration (typ- domains, debris flow inputs, and instream wood ically hours to a few days) and largest spatial extent dynamics, and result from dissipating tropical storms during the r the influence of land use, specifically timber har- autumn months. vest, on slope stability and wood recruitment to Examination of channel change resulting from channels, and the associated differences in river 1997 Hurricane Nora indicated that wider, braided form and process between undisturbed and indus- reaches aggraded substantially during the 1997 trial forests. flood, whereas narrower reaches incised. Patterns Chapter 8 Rivers in the landscape 249 of aggradation and incision reflected a threshold ical losses and reduced channel stability as a result of relationship between flow depth and flow resis- military activities in the channel network. tance associated with vegetated bars. Degradation Ongoing research designed to identify the most occurred as long as flow was confined within a chan- ecologically diverse and resilient channel segments nel or subchannel. At flow depths sufficiently high to focuses on a six-part channel classification (Sutfin, overtop vegetated bars, greater flow resistance asso- 2013) of ciated with the vegetation facilitated lower veloci- ties, deposition, and aggradation (Merritt and Wohl, r montane bedrock channels entirely confined by 2003). exposedbedrockandlackingpersistentalluvium; Downslope from the steep, mountainous terrain r upper piedmont bedrock with alluvium channels that characterizes the headwaters of Yuma Wash, that are at least partly confined by bedrock but channels are incised into more gently sloping alluvial contain enough alluvium to create bedforms that piedmont surfaces of alluvial fans and pediments persist through time; (Figure S8.23). Eolian sedimentation and accumula- r incised alluvium channels bounded only by tion of silt facilitate the formation of extensive desert unconsolidated alluvium into which the channel pavement (Bacon et al., 2010). Desert pavement con- segment is incised; sists of a coarse upper layer of pebble to cobble size r depositional braided washes with multi-thread clasts that are typically covered in a secondary iron– channel planform regardless of the degree and manganese weathering rind known as desert var- composition of confining material; nish. Many of the upland surfaces host cryptobiotic r piedmont headwater channels that are first- or soil crusts that help to stabilize finer sediment near second-order streams confined only by unconsol- the surface. Desert varnish, desert pavement, and idated alluvium and which initiate as secondary cryptobiotic soil crusts take many years to form, but channels on piedmont surfaces; and these surfaces are disrupted by military activities in r flood-outs, a name originally applied in Australia the vicinity of Yuma Wash. (Tooth, 1999) to channels in arid regions in which YumaWashlieswithintheUSArmy’sYuma all signs of defined channels and surface flow dis- Proving Ground (YPG), which was established in appear as a result of transmission losses (evapora- 1943 as a hot desert training center. YPG is used tion and infiltration). primarily for tactical ground exercises, testing of heavyartillery,andaerialreleaseofheavyequip- Statistical analyses indicate that these six stream ment.StartingduringtheeraofWorldWarII, types differ significantly in terms of stream gradi- tactical ground exercises focused on the piedmont ent, width/depth ratio, entrenchment ratio (ratio of uplands. Disturbance of desert pavement and cryp- valleywidthattwotimesthereferencestage(ref- tobiotic soils in association with these activities has erence stage here is stage of a relatively common proven to be persistent: the outlines of roads and tent discharge analogous to bankfull)), shear stress, and camps from that era still appear clearly in contem- stream power per unit area (Sutfin, 2013). porary aerial imagery. Consequently, military exer- Effective channel management and restoration in ciseshavemorerecentlyfocusedinthechannelnet- Yuma Wash requires recognizing some of the unique work.Mostactiveroadsthroughthearea,forexam- aspects of this hot desert environment, including ple, use dry washes. Although dry washes might be thoughttohavelesserbiologicalvaluethanintermit- r the temporally and spatially discontinuous nature tent or perennial channels, the vegetative abundance of channel change—individual channel segments and diversity of the channel network at Yuma Wash, within the river network commonly do not change as well as the importance of woody riparian vegeta- in the same direction (i.e., some reaches are incis- tion in creating flow resistance and limiting channel ingwhileothersareaggrading),atthesametime, erosion, suggest the potential for substantial ecolog- or at the same rate, 250 Rivers in the Landscape r not all rainfall produces runoff or stream flow, accelerated globally since the mid-twentieth cen- and only the most extensive and sustained rain- tury, abundant evidence has appeared that reflects fallaboveaminimumintensitythresholdpro- riverine influences on the entire critical zone. duces discharge through substantial portions of Changes in coastal environments provide a vivid the channel network—this typically does not hap- example. pen every year, Human activities now dominate nitrogen bud- r discharge can increase downstream to some gets in regions such as Asia, Europe, and North extent, but at some point in the network discharge America (Boyer et al., 2006). Substantial increases in during a single flood decreases downstream as a nitrogen yields to rivers result from industrial-scale result of infiltration and evaporation losses, agriculture, feedlots, and septic systems. Riverine r alluvial upland surfaces are highly sensitive to corridors have been simplified via channelization, physical disruptions of desert pavement and cryp- removal of riparian vegetation, levees and flow regu- tobiotic soil crusts, and disruption of these fea- lation, all of which reduce channel–floodplain con- tures likely alters rainfall–runoff relations, and nectivity. This simplification reduces nitrogen reten- r woody riparian vegetation strongly influences tion and processing by rivers. Greater inputs and flow resistance and channel adjustment during less storage create a “one-two punch,” resulting in floods. substantially increased nitrogen fluxes down rivers to coastal areas. This has created eutrophication of 8.6 The greater context estuaries and other nearshore environments. For the most part, we cannot yet predict in detail The nineteenth-century American conservationist how diverse changes in rivers resulting directly John Muir famously wrote in 1911, “When we try and indirectly from human activities will affect topickoutanythingbyitself,wefindithitched the greater landscape. Synthesizing studies in the to everything else in the Universe.” Many other northern high latitudes, for example, Woo (2010) thinkers have stated the same concept in different explained how freshwater discharge during spring words, recognizing the interconnectedness of nat- snowmelt influences numerous and diverse pro- ural systems and people. Rivers are no exception cesses in the nearshore zone and greater Arctic tothisrule,asIhaveemphasizedfromtheopen- Ocean (Figure 8.8). Freshwater discharge influences ing pages in this book. As the intensity and extent thedynamicsofcoastalseaice,terrestrialsedi- of human alteration of river form and process have ment and organic matter plumes into the ocean, and

Freshwater discharge during spring snowmelt

Dynamics of coastal sea ice: Exports organic carbon river discharges enhance and nutrients that ice breakup in enhance biological nearshore zone productivity of Arctic Ocean

Creates sediment plumes Thermohaline circulation that can extend in polar seas: hundreds of kilometers freshwater layer forms into the ocean over saline ocean water or floods the landfast ice

Figure 8.8 Schematic diagram of the effects of freshwater discharge during spring snowmelt from rivers draining to the Arctic Ocean. Chapter 8 Rivers in the landscape 251 thermohalinecirculationinpolarseas.Thecom- r More than 100 million tons of silt historically car- plex effects of global warming on these interactions ried by the Nile River each year to the Mediter- remain largely unknown. Less sea ice may mod- ranean Sea nourished marine plankton commu- ify existing energy and moisture fluxes, for exam- nities at the river’s mouth. Sardines fed on the ple, and thus alter coastal storm patterns and inland plankton. Completion of the High Aswan Dam water balances (Woo, 2010). in 1970 dramatically decreased water and sedi- Unfortunately, we too commonly recognize the ment yield from the Nile drainage basin to the importance of landscape context and connectivity Mediterranean, and contributed (along with pol- within and between river networks once our activ- lution) to the collapse of the commercial sardine ities have altered connectivity and caused unin- fishery (Collins, 2002). Diminished sediment sup- tended negative consequences. Examples span spa- ply has also resulted in subsidence and erosion tial scales from headwaters to the world’s largest oftheNiledelta,whereformerdeltavillagesare rivers, climatic zones from hot and cold deserts to now 2 km out to sea (Stanley et al., 2004). Here, rainforests, and tectonically active and passive ter- reduced river longitudinal connectivity alters delta rains with varying lithology and structure. The fol- and nearshore environments. lowing list expands on some of the examples men- r The Novosibirsk Dam has altered flow along tioned briefly in Chapter 1 and introduces others. more than 1000 km of the upper and middle Ob River in Siberia since the dam’s completion in r Nineteenth- and twentieth-century wetland 1957 (Bityukov, 1990). The Ob historically pro- drainage and channelization along thousands of vided one of Russia’s most important fisheries, kilometers of small, headwater streams in the partly because of the abundant floodplain habi- Illinois River basin of the central United States tatalongtheriver.Annualoverbankfloodsduring reduced channel–floodplain connectivity and snowmelt provided fish spawning habitat in flood- caused precipitous declines in fish and water- plain meadows during May, and young fish could bird communities in this once extraordinarily grow rapidly in the warm, slow-moving, shallow productive watershed (Mattingly et al., 1993). waters during June and early July. A minimum Here, reduced lateral connectivity alters biotic of 20 days of flooding each year are required for communities. fish spawning, hatching, and growth along the Ob, r The Los Angeles River drains 2200 km2 of south- but flow regulation associated with the Novosi- ern California, USA, flowing <60 km from the birsk Dam reduces floodplain spawning and feed- SantaSusanaMountainstothePacificOcean.The ing areas by half in years of average flow, and fire-prone chaparral vegetation in these moun- completely during years of drought (Wohl, 2011a). tains and the dry climate historically contributed Catches of commercially valuable species such as substantial amounts of sediment to the nearshore Siberian sturgeon (Acipenser baerii)droppedto zone via the river: on the order of 600,000 m3 per nearly nothing by the mid-1990s (Ruban, 1997). year (Flick, 1993). The combined effects of sedi- In this example, reduced lateral connectivity alters ment detention basins in the high-relief headwa- riverine biota. ters and extensive channel stabilization along the r North America’s Mississippi River integrates river’s course (tributaries are confined by concrete land uses and human alterations of river process within1kmofthestartofsurfaceflow,andmore and form across 3.5 million km2.Amongthe than 700 km of tributaries are within concrete), many substantial, human-induced changes in have reduced sediment output from the river to the drainage basin during the past century have 200,000 m3 per year. This has severely exacerbated been the loss of naturally vegetated floodplains coastal erosion and loss of beaches (Flick, 1993). In and riparian zones, and the construction of this example, reduced river longitudinal connec- hundreds of dams within the watershed. Loss tivity affects coastal environments. of valley-bottom forests and wetlands and their 252 Rivers in the Landscape

biogeochemical buffering effects, channelization plankton communities to shift from silica-using and loss of the river-form complexity that pro- diatoms to coccolithophores and flagellates that motes retention and biological processing of nutri- do not require as much silica. This has caused ents, and the spread of industrial agriculture and algal blooms that destabilize the Black Sea ecosys- associated fertilizer and pesticide use during the tem (Humborg et al., 1997). Eutrophication, algal latter half of the twentieth century, together cre- blooms, and contaminants carried by river water ated enormous increases in nutrients and contam- have contributed to the collapse of once com- inants transported from the river catchment into mercially valuable Black Sea sprat and anchovy the Gulf of Mexico in the Atlantic Ocean (Goolsby fisheries. As in previous examples, increased lat- et al., 1999). This has caused persistent hypoxia in eral connectivity of nutrients and contaminants, theGulf,anda“deadzone”thatfluctuatesbetween and decreased longitudinal connectivity of sedi- 13,000 and 20,000 km2 in size (Mitsch et al., 2001). ment, have dramatically altered delta, nearshore, The Missouri River drains the western half of the and coastal environments. Mississippi River basin, contributing relatively r The Fly River drains 75,000 km2 in Papua New little water (12% of the Mississippi’s discharge) but Guinea. The Ok Tedi, one of the Fly’s major trib- large amounts of sediment (>50% of the Missis- utaries, originates in the steep, tectonically active sippi’s load) (Meade et al., 1990). Numerous large Southern Mountains, and includes the Ok dams built along the Missouri since 1950 now trap Tedi mine. The open-pit copper–gold porphyry about 80% of the sediment load. Along with other mine is the second largest copper-producing mine changes in land use, this reduced sediment dis- in the world. Production began in 1984. No tail- charge to the Gulf contributes to rapid delta and ings dam was constructed, and ∼80,000 tons of coastal erosion (69 km2 of coastal wetlands lost waste tailings and 121,000 tons of mined waste each year, for an estimated total of 3900–5300 km2 rock are dumped directly into the Ok Tedi River since the 1930s) (Turner, 1997). In this example, each day, resulting in ∼66 million tons of extra increased upland-river connectivity in the form sediment per year (Hettler et al., 1997). Metals of increased nutrient fluxes and decreased pro- travel downstream in dissolved and particulate cessing, and reduced longitudinal connectivity of form (Yaru and Buckney, 2000). Elevated levels sediment fluxes, substantially alter delta, coastal, of copper have been detected in floodplain lakes and nearshore environments and biota. (Nicholson et al., 1993) and in the Gulf of Papua r The Danube River drains 816,000 km2 of Europe beyond the mouth of the Fly River (Apte and and,liketheNileandtheMississippi,changes Day, 1998). The bed of the Ok Tedi aggraded in land use and flow regulation have negatively over6minplacesduringthedecadefollowing impacted delta and nearshore geomorphology, the start of mining (Swales et al., 1998). Min- water quality, and biota. Compared to the 1960s, ing sediment is also dispersed across the flood- nitrogen loads entering the Black Sea from the plain. Although direct overbank flooding and ver- Danube have increased fivefold, phosphorus loads tical accretion are limited to less than 1 km on have doubled, and silica has decreased by about either side of the main channel, sediment is dis- two-thirds. The increased nitrogen and phospho- persed across tens of kilometers of floodplain by rus reflect commercial agriculture and loss of backflooding of the mainstem up tributaries and delta lakes and floodplains that promote biogeo- secondary channels connecting the tributaries to chemical processing of nutrients. The result has floodplain wetlands (Dietrich et al., 1999, 2007). been eutrophication of the lower river, delta, and In this example, disposal of mining wastes arti- nearshore areas. Decreased silica likely reflects the ficially increased hillslope–channel connectivity, 70% drop in sediment yields to the delta as a resulting in widespread downstream dispersal of result of widespread dam construction upstream. sediment and metals, and altered river, floodplain, Declines in silica have caused Black Sea phyto- and nearshore environments. Chapter 8 Rivers in the landscape 253

The point of these examples of altered connectiv- integral to human communities than is any other ity is not to induce despair as we contemplate past single landscape component. The challenge of devel- failures to account for the importance of connectiv- oping ways to co-exist with healthy, functional rivers ity, but rather to highlight the importance of being is integral to both a scientific understanding of rivers fully cognizant of various forms of connectivity as and to applying that understanding through river we move forward with river management. Rivers management. This challenge can only be met by are physical systems with a history, and rivers exist treating rivers as part of the greater landscape. If we within a global landscape that includes the atmo- do not view rivers in this integrative, holistic con- sphere,landmasses,oceans,groundwater,andall text,wemakethesamemistakesoverandoverand, the wondrously diverse organisms that live on our ultimately, risk nothing less than the survival of our planet. Rivers are at the heart of nearly every land- own societies. scape on Earth, and river form and process are more

References

Aalto, R., J.W. Lauer, W.E. Dietrich. 2008. Spatial and phology, Sedimentology, Dynamics. Geological Society temporal dynamics of sediment accumulation and Special Publication 251, pp. 85–94. exchange along Strickland River floodplains (Papua Allan, J.D. 1995. Stream Ecology: Structure and Func- New Guinea) over decadal-to-centennial timescales. tion of Running Waters. Chapman & Hall, London, Journal of Geophysical Research 113, F01S04. doi: 388 p. 10.1029/2006JF000627 Allan, A.F., L. Frostick. 1999. Framework dilation, win- Aarts, B.G.W., F.W.B. van den Brink, P.H. Nienhuis. 2004. nowing, and matrix particle size: the behavior of some Habitatlossasthemaincauseoftheslowrecov- sand-gravel mixtures in a laboratory flume. Journal of eryoffishfaunasofregulatedlargeriversinEurope: Sedimentary Research 69, 21–26. the transversal floodplain gradient. River Research and Allen, J.R.L. 1982. Sedimentary Structures: Their Character Applications 20, 3–23. and Physical Basis, Volumes 1and 30A (Developments in Abbe, T., A. Brooks. 2011. Geomorphic, engineering, and Sedimentology). Elsevier, Amsterdam. ecological considerations when using wood in river Allen, J.R.L. 1985. Principles of Physical Sedimentology. restoration. In: A. Simon, S.J. Bennett, J.M. Castro, Allen and Unwin, London. eds., Stream Restoration in Dynamic Fluvial Systems: Allen, J.R.L. 1994. Fundamental properties of fluids and Scientific Approaches, Analyses, and Tools.American their relation to sediment transport processes. In: Geophysical Union Press, Washington, DC, pp. 419– K. Pye, ed., Sediment Transport and Depositional 451. Processes. Blackwell Scientific Publications, Oxford, Abbe, T.B., D.R. Montgomery. 2003. Patterns and pro- pp. 51–60. cesses of wood debris accumulation in the Queets River Alley, W.M., R.W. Healy, J.W. LaBaugh, T.E. Reilly. 2002. basin, Washington. Geomorphology 51, 81–107. Flow and storage in groundwater systems. Science 296, Abrahams, A.D., G. Li, J.F. Atkinson. 1995. Step-pool 1985–1990. streams: adjustment to maximum flow resistance. Allmendinger, N.E., J.E. Pizzuto, N. Potter, T.E. Johnson, Water Resources Research 31, 2593–2602. W.C. Hession. 2005. The influence of riparian vegeta- Adams, R.K., J.A. Spotila. 2005. The form and function of tion on stream width, eastern Pennsylvania, USA. Geo- headwater streams based on field and modeling investi- logical Society of America Bulletin 117, 229–243. gations in the southern Appalachian Mountains. Earth Allred, T.M., J.C. Schmidt. 1999. Channel narrowing by Surface Processes and Landforms 30, 1521–1546. vertical accretion along the Green River near Green Aizen, V., E. Aizen, J. Melack. 1995. Characteristics of River, Utah. Geological Society of America Bulletin 111, runoff formation at the Kirgizskiy Alatoo, Tien Shan. 1757–1772. In:K.A.Tonnessen,M.W.Williams,M.Tranter,eds., Alonso, C.V. 2004. Transport mechanics of stream-borne Biogeochemistry of Seasonally Snow-Covered Catch- logs. In: S.J. Bennett, A. Simon, eds., Riparian Vegeta- ments. IAHS Publication no. 228, pp. 413–430. tion and Fluvial Geomorphology.AmericanGeophysi- Al-Farraj, A., A.M. Harvey. 2005. Morphometry and cal Union Press, Washington, DC, pp. 59–69. depositionalstyleoflatePleistocenealluvialfans:Wadi Amerson, B.E., D.R. Montgomery, G. Meyer. 2008. Rela- Al-Bih,northernUAEandOman.In:A.M.Harvey, tive size of fluvial and glaciated valleys in central Idaho. A.E. Mather, M. Stokes, eds., Alluvial Fans: Geomor- Geomorphology 93, 537–547.

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

255 256 References

Amos, C.B., D.W.Burbank. 2007. Channel width response Wilcock, eds., Natural and Anthropogenic Influences in to differential uplift. Journal of Geophysical Research Fluvial Geomorphology. American Geophysical Union, 112, F02010. doi:10.1029/2006JF000672 Washington, DC, pp. 151–164. Anderson, R.S., S.P.Anderson. 2010. Geomorphology: The Andrews, E.D., J.D. Smith. 1992. A theoretical model for Mechanics and Chemistry of Landscapes.Cambridge calculating marginal bed load transport rates of gravel. University Press, Cambridge, 637 p. In:P.Billi,R.D.Hey,C.R.Thorne,P.Tacconi,eds., Anderson, S.P.,W.E.Dietrich. 2001. Chemical weathering Dynamics of Gravel-bed Rivers.JohnWiley&Sons, and runoff chemistry in a steep headwater catchment. Chichester, pp. 41–52. Hydrological Processes 15, 1791–1815. Anthony, D.J., M.D. Harvey. 1991. Stage-dependent cross- Anderson, C.B., C.R. Griffith, A.D. Rosemond, R. Rozzi, section adjustments in a meandering reach of Fall River, O. Dollenz. 2006b. The effects of invasive North Amer- Colorado. Geomorphology 4, 187–203. ican beavers on riparian plant communities in Cape Apte, S.C., G.M. Day. 1998. Dissolved metal concentra- Horn, Chile: do exotic beavers engineer differently in tions in the Torres Strait and Gulf of Papua. Marine sub-Antarctic ecosystems? Biological Conservation 128, Pollution Bulletin 36, 298–304. 467–474. Arboleya, M.-L., J. Babault, L.A. Owen, A. Teixell, R.C. Anderson, S.P., S.A. Longacre, E.R. Kraal. 2003. Patterns Finkel. 2008. Timing and nature of Quaternary fluvial of water chemistry and discharge in the glacier-fed incision in the Ouarzazate foreland basin, Morocco. Kennicott River, Alaska: evidence for subglacial water JournaloftheGeologicalSocietyofLondon165, 1059– storage cycles. Chemical Geology 202, 297–312. 1073. Anderson, R.S., C.A. Riihimaki, E.B. Safran, K.R. Mac- Arthington, A.H., S.E. Bunn, N.L. Poff, R.J. Naiman. 2006. Gregor. 2006c. Facing reality: late Cenozoic evolution of The challenge of providing environmental flow rules smooth peaks, glacially ornamented valleys, and deep to sustain river ecosystems. Ecological Applications 16, rivergorgesofColorado’sFrontRange.In:S.D.Willett, 1311–1318. N.Hovius,M.T.Brandon,D.M.Fisher,eds.,Tectonics, Arthington, A.H., B.J. Pusey. 2003. Flow restoration and Climate, and Landscape Evolution. Geological Society protection in Australian rivers. River Research and of America Special Paper 398, Boulder, CO, pp. 397– Applications 19, 377–395. 418. Ashley, G.M. 1990. Classification of large-scale subaqeous Anderson, B.G., I.D. Rutherfurd, A.W. Western. 2006a. bedforms: a new look at an old problem. Journal of Sed- An analysis of the influence of riparian vegetation on imentary Petrology 60, 160–172. the propagation of flood waves. Environmental Mod- Ashmore, P.E. 1991. How do gravel-bed rivers braid? elling and Software 21, 1290–1296. Canadian Journal of Earth Sciences 28. 326–341. Andreoli, A., F. Comiti, M.A. Lenzi. 2007. Characteristics, Ashmore, P. 2013. Morphology and dynamics of braided distribution and geomorphic role of large woody debris rivers. In: E. Wohl, ed., Treatise on Fluvial Geomorphol- inamountainstreamoftheChileanAndes.Earth Sur- ogy. Elsevier, Amsterdam, pp. 290–312. face Processes and Landforms 32, 1675–1692. Ashmore, P.E., T.J. Day. 1988. Effective discharge for Andrews, E.D. 1980. Entrainment of gravel from naturally suspended sediment transport in streams of the sorted riverbed material. Geological Society of America Saskatchewan River basin. Water Resources Research 24, Bulletin 94, 1225–1231. 864–870. Andrews, E.D. 1984. Bed-material entrainment and Ashworth, P.J., J.L. Best, M.A. Jones. 2007. The relation- hydraulic geometry of gravel-bed rivers in Col- ship between channel avulsion, flow occupancy and orado. Geological Society of America Bulletin 95, 371– aggradation in braided rivers: insights from an exper- 378. imental model. Sedimentology 54, 497–513. Andrews, E.D., D.C. Erman. 1986. Persistence in the size Ashworth, P.J., R.I. Ferguson. 1989. Size-selective entrain- distribution of surficial bed material during an extreme mentofbedloadingravelbedstreams.Water Resources snowmelt flood. Water Resources Research 22, 191– Research 25, 627–634. 197. Aslan, A., W.J. Autin. 1999. Evolution of the Holocene Andrews, E.D., J.M. Nankervis. 1995. Effective discharge Mississippi River floodplain, Ferriday, Louisiana: and the design of channel maintenance flows for gravel- insights on the origin of fine-grained floodplains. bed rivers. In: J.E. CostaA.J. Miller, K.W. Potter, P.R. Journal of Sedimentary Research 69, 800–815. References 257

Attal, M., J. Lave.´ 2006. Changes of bedload characteris- Barnes, H.L. 1956. Cavitation as a geological agent. Amer- tics along the Marsyangdi River (central Nepal): impli- ican Journal of Science 254, 493–505. cations for understanding hillslope sediment supply, Barnes, H.H. 1967. Roughness Characteristics of Natural sediment load evolution along fluvial networks, and Channels. U.S. Geological Survey Water-supply Paper denudation in active orogenic belts. In: S.D. Willett, 1849, Washington, DC, 213 p. N. Hovius, M.T. Brandon, D.M. Fisher, eds., Tectonics, Barry, R.G., R.J. Chorley. 1987. Atmosphere, weather and Climate, and Landscape Evolution. Geological Society of climate, 5th edn. Methuen, London. America Special Paper 398, pp. 143–171. Bathurst, J.C. 1985. Flow resistance estimation in moun- Attal, M., G.E. Tucker, A.C. Whittaker, P.A. Cowie, G.P. tain rivers. Journal Hydraulic Engineering 111, 625–641. Roberts. 2008. Modeling fluvial incision and tran- Bathurst, J.C. 1990. Tests of three discharge gauging tech- sient landscape evolution: influence of dynamic chan- niques in mountain rivers. In: L. Molnar, ed., Hydrol- nel adjustment. Journal of Geophysical Research 113, ogy of Mountainous Areas. IAHS Publication no. 190, F03013. doi:10.1029/2007JF000893, 16 p. pp. 93–100. Aufdenkampe, A.K., E. Mayorga, P.A. Raymond, J.M. Battin, T.J., L.A. Kaplan, S. Findlay, C.S. Hopkinson, E. Melack, S.C. Doney, S.R. Alin, R.E. Aalto, K. Yoo. 2011. Marti, A.I. Packman, J.D. Newbold, F. Sabater. 2008. Riverine coupling of biogeochemical cycles between Biophysical controls on organic carbon fluxes in fluvial land, oceans, and atmosphere. FrontiersinEcologyand networks. Nature Geoscience 1, 95–100. the Environment 9, 53–60. Battin, T.J., S. Luyssaert, L.A. Kaplan, A.K. Aufdenkampe, Aylsworth, J.M., A. Duk-Rodkin, T. Robertson, J.A. A. Richter, L.J. Tranvik. 2009. The boundless carbon Traynor. 2000. Landslides of the Mackenzie valley and cycle. Nature Geoscience 2, 598–600. adjacent mountainous and coastal regions. In: L.D. Baumgart-Kotarba, M., J.-P. Bravard, M. Chardon, Dyke,G.R.Brooks,eds.,The Physical Environment of V.Jomelli,S.Kedzia,A.Kotarba,P.Pech,Z. the Mackenzie Valley, Northwest Territories: A Base Line Raczkowska. 2003. [High-mountain valley floors for the Assessment of Environmental Change. Geological evolution during recession of alpine glaciers in the Survey of Canada, Bulletin 547, pp. 167–176. Massif des Ecrins, France]. Geographia Polonica 76, Bacon, S.N., E.V. McDonald, S.E. Baker, T.G. Caldwell, 65–87. G.K. Dalldorf. 2010. Timing and distribution of allu- Bayley, P.B. 1991. The flood-pulse advantage and the vial fan sedimentation in response to strengthening of restoration of river-floodplain systems. Regulated late Holocene ENSO variability in the Sonoran Desert, Rivers Research and Management 6, 75–86. southwestern Arizona, USA. Quaternary Research 73, Beaumont, C., G. Quinlan. 1994. A geodynamic frame- 425–438. work for interpreting crustal-scale seismic reflectivity Bagnold, R.A. 1977. Bedload transport by natural waters. patterns in compressional orogens. International Jour- Water Resources Research 13, 303–312. nal of Geophysics 116, 754–783. Baker, V.R. 1978. Large-scale erosional and depositional Beauvais, A.A., D.R. Montgomery. 1997. Are channel net- features of the Channeled Scabland. In: V.R. Baker, works statistically self-similar? Geology 25, 1063–1066. D. Nummedal, eds., The Channeled Scabland.National Beckman, N.D. 2013. Crossing a Threshold: The legacy of Aeronautics and Space Administration, Washington, 19th Century Logging on Log Jams and Carbon Storage DC, pp. 81–115. in Front Range Headwater Streams. Unpublished PhD Baker, V.R. 1988. Flood erosion. In: V.R. Baker, R.C. dissertation, Colorado State University, Fort Collins, Kochel,P.C.Patton,eds.,Flood Geomorphology.John CO. Wiley & Sons, New York, pp. 81–95. Bednarek, A.T., D.D. Hart. 2005. Modifying dam opera- Baker, V.R. 2013. Global late Quaternary fluvial paleo- tions to restore rivers: ecological responses to Tennessee hydrology: with special emphasis on paleofloods and River dam mitigation. Ecological Applications 15, 997– megafloods. In: E. Wohl, ed., Treatise on Fluvial Geo- 1008. morphology. Elsevier, Amsterdam, pp. 512–527. Belmont, P., K.B. Gran, S.P. Schottler, P.R. Wilcock, S.S. Baker, D.B., P.R. Richards, T.T. Loftus, J.W. Kramer. 2004. Day, C. Jennings, J.W. Laurer, E. Viparelli, J.K. Willen- A new flashiness index: characteristics and applications bring, D.R. Engstrom, G. Parker. 2011. Large shift in to Midwestern rivers and streams. Journal of the Amer- source of fine sediment in the Upper Mississippi River. ican Water Resources Association 40, 503–522. Environmental Science and Technology 45, 8804–8810. 258 References

Beltaos, S. 2002. Effects of climate on mid-winter ice jams. tude of arsenic pollution in the Mekong and Red River Hydrological Processes 16, 789–804. deltas – Cambodia and Vietnam. Science of the Total Benda, L.E. 1990. The influence of debris flows on chan- Environment 372, 413–425. nels and valley floors in the Oregon Coast Range, USA. Bergkamp, G., 1998. A hierarchical view of the inter- Earth Surface Processes and Landforms 15, 457–466. actions of runoff and infiltration with vegetation and Benda, L., K. Andras, D. Miller, P. Bigelow. 2004. microtopography in semiarid shrublands. Catena 33, Confluence effects in rivers: interactions of basin 201–220. scale, network geometry, and disturbance regimes. Berner, E.K., R.A. Berner. 1987. The Global Water Cycle: Water Resources Research 40, W05402. doi:10.1029/ Geochemistry and Environment. Prentice Hall, Engle- 2003WR002583, 15 p. wood Cliffs, NJ, 397 p. Benda, L.E., T. Dunne. 1987. Sediment routing by debris Bernhardt, E.S., M.A. Palmer, J.D. Allan, G. Alexander, flow. In: R.L. Beschta, T.Blinn, G.E. Grant, F.J. Swanson, K.Barnas,S.Brooks,J.Carr,S.Clayton,C.Dahm,J. G.G. Ice, eds., Erosion and Sedimentation in the Pacific Follstad-Shah, D. Galat, S. Gloss, P. Goodwin, D. Hart, Rim. IAHS Publication no. 165, pp. 213–223. B. Hassett, R. Jenkinson, S. Katz, G.M. Kondolf, P.S. Benda, L.E., M.A. Hassan, M. Church, C.L. May. 2005. Lake, R. Lave, J.L. Meyer, T.K. O’donnell, L. Pagano, Geomorphology of steepland headwaters: the transi- B. Powell, E. Sudduth. 2005. Ecology. Synthesizing U.S. tion from hillslopes to channels. Journal of the American river restoration efforts. Science 308, 636–637. Water Resources Association 41, 835–851. Bernhardt,E.S.,E.B.Sudduth,M.A.Palmer,J.D.Allan, Benda,L.E.,D.Miller,J.Sais,D.Martin,R.Bilby,C.Veld- J.L.Meyer,S.Brooks,J.Carr,S.Clayton,C.Dahm, huisen, T. Dunne. 2003b. Wood recruitment processes J. Follstad-Shah, D. Galat, S. Gloss, P.Goodwin, D. Hart, and wood budgeting. In: S.V.Gregory, K.L. Boyer, A.M. B. Hassett, R. Jenkinson, S. Katz, G.M. Kondolf, P.S. Gurnell, eds., The Ecology and Management of Wood Lake, R. Lave, J.L. Meyer, T.K. O’Donnell, L. Pagano, in World Rivers. American Fisheries Society, Bethesda, B. Powell, E. Sudduth. 2007. Restoring rivers one reach MD, pp. 49–73. at a time: results from a survey of U.S. river restoration Benda, L.E., N.L. Poff, C. Tague, M.A. Palmer, J. Piz- practitioners. Restoration Ecology 15, 482–493. zuto, S.D. Cooper, E. Stanley, G. Moglen. 2002. How to Bertoldi,W.,L.Zanoni,S.Miori,R.Repetto,M.Tubino. avoid train wrecks when using science in environmen- 2009. Interaction between migrating bars and bifurca- tal problem solving. BioScience 52, 1127–1136. tions in gravel bed rivers. Water Resources Research 45, Benda, L.E, J.C. Sias. 2003. A quantitative framework W06418. doi:10.1029/2008WR007086 forevaluatingthemassbalanceofin-streamorganic Beschta, R.L., W.J. Ripple. 2012. The role of large debris. Forest Ecology and Management 172, 1–16. predators in maintaining riparian plant communities Benda, L., C. Veldhuisen, J. Black. 2003a. Debris flows and river morphology. Geomorphology 157–158, 88– as agents of morphological heterogeneity at low-order 98. confluences, Olympic Mountains, Washington. Geolog- Best, J.L. 1986. The morphology of river channel conflu- ical Society of America Bulletin 115, 1110–1121. ences. Progress in Physical Geography 10, 157–174. Benedict, J.B. 1992. Along the great divide: paleoindian Best, J.L. 1987. Flow dynamics at river channel conflu- archaeology of the high Colorado Front Range. In: D.J. ences: implications for sediment transport and bed Stanford,J.S.Day,eds.,Ice Age Hunters of the Rockies. morphology. In: F.G. Ethridge, R.M. Flores, M.D. Har- University Press of Colorado, Niwot, CO, pp. 343–359. vey, eds., Recent Developments in Fluvial Sedimentology. Benito, G., M.J. Machado, A. Perez-Gonzalez. 1996. Cli- SEPM Special Publication 39, pp. 27–35. mate change and flood sensitivity in Spain. In: J. Bran- Best, J.L. 1992. On the entrainment of sediment and initi- son, A.G. Brown, K.J. Gregory, eds., Global Continen- ation of bed defects: insights from recent developments tal Changes: The Context of Palaeohydrology. Geological within turbulent boundary layer research. Sedimentol- Society Special Publications 115, pp. 85–98. ogy 39, 797–811. Benito, G., J. O’Connor. 2013. Quantitative paleoflood Best, J.L. 1993. On the interactions between turbulent flow hydrology. In: E. Wohl, ed., Treatise on Fluvial Geomor- structure, sediment transport and bedform develop- phology. Elsevier, Amsterdam, pp. 459–474. ment.In:N.J.Clifford,J.R.French,J.Hardistry,eds., Berg, M., C. Stengel, P.T.K. Trang, P.H. Viet, M.L. Samp- Turbulence: Perspectives on Flow and Sediment Trans- son, M. Lang, S. Samreth, D. Fredericks. 2007. Magni- port. John Wiley & Sons, Chichester, pp. 61–92. References 259

Best, J.L. 2005. The fluid dynamics of river dunes: a review Bledsoe, B.P., C.C. Watson. 2001. Effects of urbaniza- and some future research directions. Journal of Geo- tion on channel instability. Journal of American Water physical Research 110, F04S02. Resources Association 37, 255–270. Best, J.L., B.L. Rhoads. 2008. Sediment transport, bed Bloesch, J. 2003. [Flood plain conservation in the Danube morphology and the sedimentology of river channel River basin, the link between hydrology and limnol- confluences. In: S. Rice, A. Roy, B.L. Rhoads, eds., River ogy.] Archiv fur¨ Hydrobiologie, Supplement 147, 347– Confluences, Tributaries and the Fluvial Network.John 362. Wiley & Sons, Chichester, pp. 45–72. Bloesch, J., U. Sieber. 2003. The morphological destruc- Best, J.L., A.G. Roy. 1991. Mixing-layer distortion at the tion and subsequent restoration programmes of large confluence of channels of different depth. Nature 350, rivers in Europe. Archives fur¨ Hydrobiologie Supplement 411–413. 147, 363–385. Beven, K.J., A. Binley. 1992. The future of distributed Bocchiola, D., M.C. Rulli, R. Rosso. 2008. A flume models: model calibration and uncertainty prediction. experiment on the formation of wood jams in rivers. Hydrological Processes 6, 279–298. Water Resources Research 44, W02408. doi:10.1029/ Beven, K., Germann, P.,1982. Macropores and water flow 2006WR005846 in soils. Water Resources Research 18(5), 1311–1325. Bombino, G., A.M. Gurnell, V. Tamburino, D.A. Zema, Bierman, P.R., K.K. Nichols. 2004. Rock to sediment – S.M. Zimbone. 2009. Adjustments in channel form, slope to sea with 10Be – rates of landscape change. sediment caliber and vegetation around check-dams in Annual Review of Earth and Planetary Sciences 32, 215– the headwater reaches of mountain torrents, Calabria, 255. Italy. Earth Surface Processes and Landforms 34, 1011– Bilby, R.E., J.W. Ward. 1989. Changes in characteristics 1021. and function of woody debris with increasing size of Bonnell, M., Gilmour, D.A., 1978. The development of streams in western Washington. Transactions of the overland flow in a tropical rain-forest catchment. Jour- American Fisheries Society 118, 368–378. nal of Hydrology 39, 365–382. Birken, A.S., D.J. Cooper. 2006. Processes of Tamarix Bookhagen, B., D.W. Burbank. 2010. Toward a complete invasion and floodplain development along the lower Himalayan hydrological budget: spatiotemporal distri- Green River, Utah. Ecological Applications 16, 1103– bution of snowmelt and rainfall and their impact on 1120. river discharge. Journal of Geophysical Research 115, Bishop, P., T.B. Hoey, J.D. Jansen, I.L. Artza. 2005. Knick- F03019. doi:10.1029/2009JF001426 point recession rate and catchment area: the case of Booth, A.L., C.P. Chamberlain, W.S.F. Kidd, P.K. Zeitler. uplifted rivers in eastern Scotland. Earth Surface Pro- 2009a. Constraints on the metamorphic evolution cesses and Landforms 30, 767–778. of the eastern Himalayan syntaxis from geochrono- Bityukov, V.P. 1990. Downstream of the Novosibirsk logic and petrologic studies of Namche Barwa. hydroelectric station on the Ob River. Hydrotechnical Geological Society of America Bulletin 121, 385– Construction 23, 587–591. 407. Blair, T.C. 1999. Alluvial fan and catchment initiation by Booth, A.M., J.J. Roering, J.T. Perron. 2009b. Automated rock avalanching, Owens Valley, California. Geomor- landslide mapping using spectral analysis and high- phology 28, 201–221. resolution topographic data: Puget Sound lowlands, Blair, T.C. 2001. Outburst flood sedimentation on the Washington, and Portland Hills, Oregon. Geomorphol- proglacial Tuttle Canyon alluvial fan, Owens Valley, ogy 109, 132–147. California, USA. Journal of Sedimentary Research 71, Botter, G., S. Basso, A. Porporato, I. Rodriguez-Iturbe, 657–679. A. Rinaldo. 2010. Natural streamflow regime alter- Blair, T.C., J.G. McPherson. 2009. Processes and forms of ations: damming of the Piave River basin (Italy). alluvialfans.In:A.J.Parsons,A.D.Abrahams,eds.,Geo- Water Resources Research 46, W06522. doi:10.1029/ morphology of Desert Environments, 2nd edn. Springer, 2009WR008523 pp. 413–467. Boucher, E., Y. Begin,´ D. Arsenault, T.B.M.J. Ouarda. Blanton, P., W.A. Marcus. 2009. Railroads, roads and lat- 2012. Long-term and large-scale river-ice processes in eral disconnection in the river landscapes of the conti- cold-region watersheds. In: M. Church, P.M. Biron, nental United States. Geomorphology 112, 212–227. A.G. Roy, eds., Gravel-bed Rivers: Processes, Tools, 260 References

Environments.JohnWiley&Sons,Hoboken,NJ, ment. Earth Surface Processes and Landforms 22, 669– pp. 546–554. 683. Boulton, A.J. 1999. An overview of river health assess- Braudrick, C.A., G.E. Grant, Y. Ishikawa, H. Ikeda. 1997. ment: philosophies, practice, problems and prognosis. Dynamics of wood transport in streams: a flume exper- Freshwater Biology 41, 469–479. iment. Earth Surface Processes and Landforms 22, 669– Bovee, K.D., R. Milhous. 1978. Hydraulic Simulation in 683. Instream Flow Studies: Theory and Techniques. Instream Brayshaw, A.C. 1984. Characteristics and origin of clus- flow information paper no. 5, FWS/OBS-78/33, U.S. ter bedforms in coarse-grained alluvial channels. In: Fish and Wildlife Service, Fort Collins, CO. E.H.Koster,R.J.Steel,eds.,Sedimentology of Gravels Boyer, E.W., G.M. Hornberger, K.E. Bencala, D.M. McK- and Conglomerates– Memoir 10. Canadian Society of night. 2000. Response characteristics of DOC flush- Petroleum Geologists, pp. 77–85. inginanalpinecatchment.Hydrological Processes 11, Brice, J.C. 1964. Channel Patterns and Terraces of the Loup 1635–1647. Rivers in Nebraska. U.S. Geological Survey Professional Boyer,E.W.,R.W.Howarth,J.N.Galloway,F.J.Den- Paper 422D, pp. 1–41. tener, P.A. Green, C.J. Vor¨ osmarty.¨ 2006. Riverine Brice, J.C., J.C. Blodgett. 1978. Countermeasures for nitrogen export from the continents to the coasts. Hydraulic Problems at Bridges.FederalHighway Global Biogeochemical Cycles 20, GB1S91. doi:10.1029/ Administration Report FHWA-RD-78–162, volumes. 2005GR002537, 9 p. 1, 2, Washington, DC. Braatne, J.H., S.B. Rood, L.A. Goater, C.L. Blair. 2008. Bridge, J.S. 2003. Rivers and Floodplains: Forms, Processes, Analyzing the impacts of dams on riparian ecosystems: and Sedimentary Record. Blackwell Publishing, Malden, a review of research strategies and their relevance to MA. the Snake River through Hells Canyon. Environmental Brierley, G.J., K.A. Fryirs. 2005. Geomorphology and River Management 41, 267–281. Management: Applications of the River Styles Frame- Bracken, L.J., J. Croke. 2007. The concept of hydrolog- work. Blackwell, Oxford, 398 p. ical connectivity and its contribution to understand- Brierley, G.J., K.A. Fryirs. 2009. Don’t fight the site: ing runoff-dominated geomorphic systems. Hydrolog- three geomorphic considerations in catchment-scale ical Processes 21, 1749–1763. river rehabilitation planning. Environmental Manage- Bradshaw, P. 1985. An Introduction to Turbulence and its ment 43, 1201–1218. Measurement.PergamonPress,Oxford,218p. Brierley, G.J., K.A. Fryirs, V. Jain. 2006. Landscape con- Braga, G., S. Gervasoni. 1989. Evolution of the Po River: nectivity: the geographic basis of geomorphic applica- an example of the application of historic maps. In: G.E. tions. Area 38, 165–174. Petts, H. Moller,¨ A.L. Roux, eds., Historical Change of Brocard, G.Y., P.A.van der Beek, D.L. Bourles, L.L. Siame, Large Alluvial Rivers: Western Europe.JohnWiley& J.-L. Mugnier. 2003. Long-term fluvial incision rates Sons, Chichester, pp. 113–126. and postglacial river relaxation time in the French west- Bragg, D.C., J.L. Kershner, D.W. Roberts. 2000. Model- ern Alps from 10Be dating of alluvial terraces with ing Large Woody Debris Recruitment for Small Streams assessment of inheritance, soil development and wind of the Central Rocky Mountains.USDAForestService ablation effects. Earth and Planetary Science Letters 209, General Technical Report RMRS-GTR-55, Fort Collins, 197–214. CO. Brooker, M.P. 1981. The impact of impoundments on the Braudrick, C.A., W.E. Dietrich, G.T. Leverich, L.S. Sklar. downstream fisheries and general ecology of rivers. In: 2009. Experimental evidence for the conditions nec- T.H. Coaker, ed., Advances in Applied Biology, vol. 6. essary to sustain meandering in coarse-bedded rivers. Academic Press, London, pp. 91–152. Proceedings of the National Academy of Science 106, Brooks, S.M. 2003. Slopes and slope processes: research 16936–16941. over the past decade. Progress in Physical Geography 27, Braudrick, C.A., G.E. Grant. 2000. When do logs 130–141. move in rivers? Water Resources Research 36, 571– Brooks, A.P., G.J. Brierley. 2004. Framing realistic river 583. rehabilitation targets in light of altered sediment sup- Braudrick, C.A., G.E. Grant. 2001. Transport and deposi- ply and transport relationships: lessons from East Gipp- tion of large woody debris in streams: a flume experi- sland, Australia. Geomorphology 58, 107–123. References 261

Brooks, A.P., G.J. Brierley, R.G. Millar. 2003. The long- Bull, W.B. 1977. The alluvial fan environment. Progress in termcontrolofvegetationandwoodydebrisonchannel Physical Geography 1, 222–270. and flood-plain evolution: insights from a paired catch- Bull, W.B. 1991. Geomorphic Responses to Climatic ment study in southeastern Australia. Geomorphology Change.OxfordUniversityPress,NewYork. 51, 7–29. Bunn, S.E., A.H. Arthington. 2002. Basic principles and Brotherton, D.I. 1979. On the origin and characteristics ecological consequences of altered flow regimes for of river channel patterns. Journal of Hydrology 44, 211– aquatic biodiversity. Environmental Management 30, 230. 492–507. Bruijnzeel, L.A. 2005. Tropical montane cloud for- Bunn, S.E., P.M. Davies, T.D. Mosisch. 1999. Ecosystem est: a unique hydrological case. In: M. Bonell, L.A. measures of river health and their response to ripar- Bruijnzeel, eds., Forests, Water and People in the ian and catchment degradation. Freshwater Biology 41, Humid Tropics. Cambridge University Press, pp. 462– 333–345. 483. Bunte, K., S.R. Abt. 2001. Sampling Surface and Subsur- Brummer, C.J., D.R. Montgomery. 2003. Downstream face Particle-size Distributions in Wadable Gravel- and coarsening in headwater channels. Water Resources Cobble-bed Streams for Analyses in Sediment Transport, Research 39. doi:10.1029/2003WR001981, 14 p. Hydraulics, and Streambed Monitoring.USDAForest Brunsden, D., J.B. Thornes. 1979. Landscape sensitivity Service General Technical Report RMRS-GTR-74, Fort and change. Transactions of the Institute of British Geog- Collins, CO, 428 p. raphers 4(4), 463–484. Burbank,D.W.,J.Leland,E.Fielding,R.S.Ander- Bryan, R.B., Jones, J.A.A., 1997. The significance of soil son, N. Brozovic, M.R. Reid, C. Duncan. 1996. piping processes: inventory and prospect. Geomorphol- Bedrock incision, rock uplift and threshold hillslopes ogy 20, 209–218. in the northwestern Himalayas. Nature 379, 505– Buffin-Belanger,´ T., A.G. Roy, S. Demers. 2013. Turbu- 510. lence in river flows. In: E. Wohl, ed., Treatise on Fluvial Burchsted, D., M. Daniels, R. Thorson, J. Vokoun. 2010. Geomorphology. Elsevier, Amsterdam, pp. 70–86. The river discontinuum: applying beaver modifications Buffin-Belanger,´ T., A.G. Roy, A.D. Kirkbride. 2000. On to baseline conditions for restoration of forested head- large-scale flow structures in a gravel-bed river. Geo- waters. BioScience 60, 908–922. morphology 32, 417–435. Burge, L.M. 2005. Wandering Miramichi rivers, Buffington,J.M.,T.E.Lisle,R.D.Woodsmith,S.Hilton. New Brunswick, Canada. Geomorphology 69, 253– 2002. Controls on the size and occurrence of pools in 274. coarse-grained forest rivers. River Research and Appli- Burke, B.C., A.M. Heimsath, A.F. White. 2007. Coupling cations 18, 507–531. chemical weathering with soil production across soil- Buffington, J.M., D.R. Montgomery. 1997. A systematic mantled landscapes. Earth Surface Processes and Land- analysis of eight decades of incipient motion studies, forms 32, 853–873. with special reference to gravel-bedded rivers. Water Burke, M., K. Jorde, J.M. Buffington. 2009. Application of Resources Research. 33 (8), 1993–2029. a hierarchical framework for assessing environmental Buffington, J.M., D.R. Montgomery. 1999. Effects of impacts of dam operation: changes in streamflow, bed hydraulic roughness on surface textures of gravel-bed mobility and recruitment of riparian trees in a western rivers. Water Resources Research 35, 2507–3521. North American river. Journal of Environmental Man- Buffington, J.M., D.R. Montgomery. 2013. Geomorphic agement 90 (suppl 3), S224–S236. classification of rivers. In: E. Wohl, ed., Treatise on Flu- Burkham, D.E. 1972. Channel changes of the Gila River vial Geomorphology. Elsevier, Amsterdam, pp. 730–766. in Safford Valley, Arizona 1846–1970. U.S. Geological Bull, W.B. 1962. Relations of alluvial-fan size and slope to Survey Professional Paper 655-G, G1–G24. drainage-basin and lithology in western Fresno County, Burns,D.A.,J.J.McDonnell,R.P.Hooper,N.E.Peters,J.E. California. U.S. Geological Survey Professional Paper Freer, C. Kendall, K. Beven. 2001. Quantifying contri- 450-B, 51–53. butions to storm runoff through end-member mixing Bull, W.B. 1964. Geomorphology of segmented alluvial analysis and hydrologic measurements at the Panola fans in western Fresno County, California. U.S. Geolog- Mountain Research Watershed (Georgia, USA). Hydro- ical Survey Professional Paper 352-E, 89–129. logical Processes 15, 1903–1924. 262 References

Butler, D.R. 2006. Human-induced changes in animal Carson, M.A., M.J. Kirkby. 1972. Hillslope form and Pro- populations and distributions, and the subsequent cess. Cambridge University Press, London, 475 p. effects on fluvial systems. Geomorphology 79, 448– Carson, M.A., M.F. Lapointe. 1983. The inherent asym- 459. metry of river meander planform. Journal of Geology 91, Butzer, K.W.1976. Geomorphology from the Earth.Harper 41–55. and Row Publishers, New York. Castillo, V.M., W.M. Mosch, C.C. Garcia, G.G. Barbera, Cadol, D., E. Wohl. 2011. Wood retention and transport in J.A.N. Cano, F. Lopez-Bermudez. 2007. Effectiveness tropical, headwater streams, La Selva Biological Station, and geomorphological impacts of check dams for soil Costa Rica. Geomorphology 123, 61–73. erosion control in a semiarid Mediterranean catch- Campbell, E.P. 2005. Physical-statistical models for pre- ment: El Carcavo (Murcia, Spain). Catena 70, 416–427. dictions in ungauged basins. In: S. Franks, M. Siva- Castro, J., 2008. Certificate in river restoration: one palan, K. Takeuchi, Y. Tachikawa, eds., Predictions in approach to training and professional development. The Ungauged Basins: International Perspectives on the State Stream Restoration Networker 2 (1), 2–4. of the Art and Pathways Forward. IAHS Publication no. Castro, J.M., P.L. Jackson. 2001. Bankfull discharge recur- 301, pp. 292–298. rence intervals and regional hydraulic geometry rela- Candy, I., S. Black, B.W. Sellwood. 2004. Interpreting the tionships: patterns in the Pacific Northwest, USA. Jour- response of a dryland river system to late Quaternary nal of American Water Resources Association 37, 1249– climate change. Quaternary Science Reviews 23, 2513– 1262. 2523. Cavalli, M., S. Trevisani, F. Comiti, L. Marchi. 2013. Geo- Canestrelli, A., S. Fagherazzi, A. Defina, S. Lanzoni. 2010. morphometric assessment of spatial sediment connec- Tidal hydrodynamics and erosional power in the Fly tivity in small Alpine catchments. Geomorphology 188, River delta, Papua New Guinea. Journal of Geophysical 31–41. Research 115. doi:10.1029/2009JF001355 Cenderelli, D.A. 2000. Floods from natural and artificial Cannon, S.H., R.M. Kirkham, M. Parise. 2001. Wildfire- damfailures.In:E.E.Wohl,ed.,Infand Flood Hazards: related debris-flow initiation processes, Storm King Human, Riparian, and Aquatic Communities.Cam- Mountain, Colorado. Geomorphology 39, 171–188. bridge University Press, Cambridge, pp. 73–103. Canuel, E.A., S.S. Cammer, H.A. McIntosh, C.R. Pondell. Cenderelli, D.A., B.L. Cluer. 1998. Depositional pro- 2012. Climate change impacts on the organic carbon cesses and sediment supply in resistant-boundary chan- cycle at the land-ocean interface. Annual Reviews Earth nels: examples from two case studies. In: K.J. Tinkler, Planetary Science 40, 685–711. E.E. Wohl, eds., Rivers Over Rock: Fluvial Processes in Carling, P.A. 1988. Channel change and sediment trans- Bedrock Channels. American Geophysical Union Press, port in regulated U.K. rivers. Regulated Rivers 2, 369– Washington, DC, pp. 105–131. 388. Cenderelli, D.A., E.E. Wohl. 2001. Peak discharge esti- Carling, P.A.1995. Flow-separation berms downstream of mates of glacier-lake outburst floods and “normal” cli- a hydraulic jump in a bedrock channel. Geomorphology matic floods in the Mount Everest region, Nepal. Geo- 11, 245–253. morphology 40, 57–90. Carling, P.A., N.A. Reader. 1982. Structure, composition Cenderelli, D.A., E.E. Wohl. 2003. Flow hydraulics and andbulkpropertiesofuplandstreamgravels.Earth Sur- geomorphiceffectsofglacial-lakeoutburstfloodsinthe face Processes and Landforms 7, 349–365. Mount Everest region, Nepal. Earth Surface Processes Carling, P.A., N. Wood. 1994. Simulation of flow over and Landforms 28, 385–407. pool-riffle topography: a consideration of the velocity- CENR. 2000. Integrated Assessment of Hypoxia in the reversal hypothesis. EarthSurfaceProcessesandLand- Northern Gulf of Mexico. National Science and Technol- forms 19, 319–332. ogy Council Committee on Environment and Natural Carson, M.A. 1984. Observations on the meandering- Resources, Washington, DC. braided river transition, the Canterbury plains, New Chang, H.H. 1978. Fluvial processes in river engineering. Zealand: part two. New Zealand Geographer 40, 89–99. John Wiley & Sons, New York, 432 p. Carson, M.A., J.N. Jasper, F.M. Conly. 1998. Magnitude Chang, H.H. 1988. On the cause of river meandering. and sources of sediment input to the Mackenzie Delta, In: W.R. White, ed., International Conference on River Northwest Territories, 1974–94. Arctic 51, 116–124. Regime. Hydraulics Research, Wallingford, pp. 83–93. References 263

Changnon, S.A. 2002. Frequency of heavy rainstorms on and/or climatic deterioration? Geomorphology 84, 317– areas from 10 to 100 km2, defined using dense rain 343. gauge networks. Journal of Hydrometeorology 3, 220– Chorley, R.J. 1962. Geomorphology and General Sys- 223. tems Theory. U.S. Geological Survey Professional Paper Chanson, H. 1996. Comment on ‘Step-pool streams: 500-B, 10 p. adjustment to maximum flow resistance’ by A.D. Abra- Chow, V.T. 1959. Open-channel Hydraulics. McGraw-Hill, hams, G. Li, J.F. Atkinson. Water Resources Research 32, New York, 679 p. pp. 3401–3402. Church, M. 1983. Pattern of instability in a wander- Chanson, H. 1999. The hydraulics of Open Channel Flow: ing gravel bed channel. In: J.D. Collinson, J. Lewin, An Introduction. Arnold, London, 495 p. eds., Modern and Ancient Fluvial Systems.International Chappell, J., C. Woodroffe. 1994. Macrotidal estuar- Association of Sedimentologists, Blackwell Scientific, ies. In: C. Woodroffe, R. Carter, eds., Coastal Evo- Oxford, pp. 169–180. lution: Late Quaternary Shoreline Morphodynamics. Church, M. 2005. Bed material transport and the mor- Cambridge University Press, Cambridge, pp. 187–218. phology of alluvial river channels. Annual Reviews of Chatanantavet, P., E. Lajeunesse, G. Parker, L. Malverti, Earth and Planetary Science 34, 325–354. P. Meunier. 2010. Physically based model of down- Church, M., M.A. Hassan. 1992. Size and distance of stream fining in bedrock streams with lateral input. travel of unconstrained clasts on a streambed. Water Water Resources Research 46, W02581. doi:10.1029/ Resources Research 28, 299–303. 2008WR007208 Church, M.A., D.G. McLean, J.F. Wolcott. 1987. River bed Chatanantavet, P.,G. Parker. 2009. Physically based mod- gravels: sampling and analysis. In: C.R. Thorne, J.C. eling of bedrock incision by abrasion, plucking, and Bathurst,R.D.Hey,eds.,Sediment Transport in Gravel- macroabrasion. Journal of Geophysical Research 114, bed Rivers. John Wiley & Sons, Chichester, pp. 43–88. F04018. doi:10.1029/2008JF001044, 22 p. Clayton, J.A., J. Pitlick. 2007. Spatial and temporal vari- Chen, J., A. Ohmura. 1990. On the influence of Alpine ations in bed load transport intensity in a gravel bed glaciersonrunoff.In:H.Lang,A.Musy,eds.,Hydrol- river bend. Water Resources Research 43, W02426. ogy in Mountainous Regions I. Hydrological Measure- doi:10.1029/2006WR005253 ments, the Water Cycle. IAHS Publication no. 193, Clayton, J.A., J. Pitlick. 2008. Persistence of the surface pp. 117–125. texture of a gravel-bed river during a large flood. Earth Chiew, F.H., T.A. McMohan. 2002. Modelling the impacts Surface Processes and Landforms 33, 661–673. of climate change on Australian streamflow. Hydrologi- Clifford, N.J. 1993. Differential bed sedimentology and cal Processes 16, 1235–1245. the maintenance of riffle-pool sequences. Catena 20, Chin, A. 2006. Urban transformation of river landscapes 447–468. in a global context. Geomorphology 79, 460–487. Clifford, N.J., J.R. French. 1993a. Monitoring and mod- Chin, A., M.D. Daniels, M.A. Urban, H. Piegay,´ K.J. Gre- eling turbulent flows: historical and contemporary per- gory, W. Bigler, A.Z. Butt, J.L. Grable, S.V. Gregory, spectives.In:N.J.Clifford,J.R.French,J.Hardistry,eds., M. Lafrenz, L.R. Laurencio, E. Wohl. 2008. Perceptions Turbulence: Perspectives on Flow and Sediment Trans- of wood in rivers and challenges for stream restora- port. John Wiley & Sons, Chichester, pp. 1–34. tion in the United States. Environmental Management Clifford, N.J., J.R. French. 1993b. Monitoring and analysis 41, 893–903. of turbulence in geophysical boundaries: some analyti- Chin, A., A.P. O’Dowd, K.J. Gregory. 2013. Urbaniza- cal and conceptual issues. In: N.J. Clifford, J.R. French, tion and river channels. In: E. Wohl, ed., Treatise on J. Hardistry, eds., Turbulence: Perspectives on Flow and Fluvial Geomorphology. Elsevier, Amsterdam, pp. 809– Sediment Transport. John Wiley & Sons, Chichester, 826. pp. 93–120. Chin, A., E.Wohl. 2005. Toward a theory for step pools Clifford, N.J., K.S. Richards, A. Robert. 1992. Estimation in stream channels. Progress in Physical Geography 29, of flow resistance in gravel-bedded rivers: a physical 275–296. explanation of the multiplier of roughness length. Earth Chiverrell, R.C., A.M. Harvey, G.C. Foster. 2007. Hill Surface Processes and Landforms 17, 111–126. slope gullying in the Solway Firth, Morecambe Bay Clow, D.W., G.P. Ingersoll, M.A. Mast, J.T. Turk, D.H. region, Great Britain: responses to human impact Campbell. 2002. Comparison of snowpack and 264 References

winter wet-deposition chemistry in the Rocky Moun- pool channels. Water Resources Research 45, W04424. tains, USA: implications for winter dry deposition. doi:10.1029/2008WR007259 Atmospheric Environment 36, 2337–2348. Comiti, F., L. Mao, E. Preciso, L. Picco, L. Marchi, M. Clow, D.W., L. Schrott, R. Webb, D.H. Campbell, A. Tor- Borga. 2008. Large wood and flash floods: evidence izzo, M.M. Dornblaser. 2003. Ground water occurrence from the 2007 event in the Davcaˇ basin (Slovenia). In: and contributions to streamflow in an alpine catch- D.deWrachien,C.A.Brebbia,M.A.Lenzi,eds.,Mon- ment, Colorado Front Range. Ground Water 41, 937– itoring, Simulation, Prevention and Retention of Dense 950. and Debris Flow II. WIT Transactions on Engineering Cohn, T.A., W.L. Lane, J.R. Stedinger. 2001. Confidence Sciences 60, pp. 173–182. intervalsforExpectedMomentsAlgorithmfloodquan- Constantine, J.A., S.R. McLean, T. Dunne. 2010. A mech- tile estimates. Water Resources Research 37, 1695–1706. anism of chute cutoff along large meandering rivers Coleman, J., O. Huh, D. Braud, H.H. Roberts. 2005. Major with uniform floodplain topography. Geological Society world delta variability and wetland loss. Gulf Coast of America Bulletin 122, 855–869. Association of Geological Societies Transactions 55, 102– Constantine, C.R., J.F. Mount, J.L. Florsheim. 2003. The 131. effects of longitudinal differences in gravel mobility Collier, M.P., R.H. Webb, E.D. Andrews. 1997. Experi- on the downstream fining pattern in the Cosumnes mental flooding in Grand Canyon. Scientific American River, California. Journal of Geology 111, 233– 276, 66–73. 241. Collins, D.N. 1998. Rainfall-induced high-magnitude Constantz, J., D.A. Stonestrom. 2003. Heat as a tracer of runoff events in highly-glacierized Alpine basins. In: water movement near streams. In: D.A. Stonestrom, J. K. Kovar, U. Tappeiner, N.E. Peters, R.G. Craig, eds., Constantz, eds., Heat as a Tool for Studying the Move- Water Resources and Ecology in Headwaters.IAHSPub- ment of Ground Water Near Streams.U.S.Geological lication no. 248, pp. 69–78. Survey Circular 1260, pp. 1–6. Collins, R.O. 2002. The Nile.YaleUniversityPress,New Cooke, R.U., R.W. Reeves. 1976. Arroyos and Environ- Haven, CT, 260 p. mental Change in the American South-West.Clarendon Collins, B.D., D.R. Montgomery. 2002. Forest develop- Press, Oxford, 213 p. ment, wood jams and restoration of floodplain rivers Cooper, J.A.G. 2001. Geomorphological variability in the Puget Lowland, Washington. Restoration Ecology among micro-tidal estuaries from the wave-dominated 10, 237–247. South African coast. Geomorphology 40, 99–122. Collins, D.N. 2006b. Variability of runoff from Alpine Cooper, J.A.G., A.N. Green, C.I. Wright. 2012. Evolution basins.In:S.Demuth,A.Gustard,E.Planos,F.Scatena, of an incised valley coastal plain estuary under low sedi- E. Servat, eds., Climate Variability and Change – Hydro- ment supply: a ‘give-up’ estuary. Sedimentology 59, 899– logical Impacts. IAHS Publication no. 308, pp. 466–472. 916. Collins, B.D., D.R. Montgomery, K.L. Fetherston, T.B. Corbel, J. 1959. Vitesse de l erosion. Zeitschrift Geomor- Abbe. 2012. The floodplain large-wood cycle hypoth- phologie 3, 1–28. esis: a mechanism for the physical and biotic struc- Costa, J.E. 1984. Physical geomorphology of debris turing of temperate forested alluvial valleys in the flows.In:J.E.Costa,P.J.Fleisher,eds.,Developments North Pacific coastal ecoregion. Geomorphology 139– and Applications of Geomorphology. Springer, Berlin, 140, 460–470. pp. 268–317. Collins, D.N., D.P. Taylor. 1990. Variability of runoff Costa, M.H. 2005. Large-scale hydrological impacts from partially-glacierised Alpine basins. In: H. Lang, A. of tropical forest conversion. In: M. Bonell, L.A. Musy, eds., Hydrology in Mountainous Regions I. Hydro- Bruijnzeel, eds., Forests, Water and People in the logical Measurements, the Water Cycle. IAHS Publica- Humid Tropics. Cambridge University Press, Cam- tion no. 193, pp. 365–372. bridge, pp. 590–597. Comiti, F. 2012. How natural are Alpine mountain rivers? Costa, J.E., J.E. O’Connor. 1995. Geomorphically effec- Evidence from the Italian Alps. Earth Surface Processes tive floods. In: J.E. Costa,A.J. Miller, K.W. Potter, P.R. and Landforms 37, 693–707. Wilcock, eds., Natural and Anthropogenic Influences in Comiti, F., D. Cadol, E. Wohl. 2009. Flow regimes, bed Fluvial Geomorphology. American Geophysical Union morphology, and flow resistance in self-formed step- Press, Washington, DC, pp. 45–56. References 265

Costa, J.E., R.L. Schuster. 1988. The formation and failure Dalrymple, T. 1960. Flood-frequency analyses. U.S. Geo- of natural dams. Geological Society of America Bulletin logical Survey Water-Supply Paper 1543-A, 80 p. 100, 1054–1068. Dalrymple, T., M.A. Benson. 1967. Measurement of Peak Coulthard, T.J., M.J. Van de Wiel. 2013. Numerical mod- Discharge by the Slope-Area Method.Techniquesof elinginfluvialgeomorphology.In:E.Wohl,ed.,Trea- water-resources investigations of the U.S. Geological tise on Fluvial Geomorphology. Elsevier, Amsterdam, Survey, Book 3, Chapter A1, pp. 1–12. pp. 694–710. Dalrymple, R.W., K. Choi. 2007. Morphologic and Cowie,P.A.,M.Attal,G.E.Tucker,A.C.Whittaker,M. faciestrendsthroughthefluvial-marinetransition Naylor, A. Ganas, G.P. Roberts. 2006. Investigating the in tide-dominated depositional systems: a schematic surface process response to fault interaction and linkage framework for environmental and sequence strati- using a numerical modelling approach. Basin Research graphic interpretation. Earth-Science Reviews 81, 135– 18, 23–266. 174. Cowie,P.A.,A.C.Whittaker,M.Attal,G.Roberts,G.E. Dalrymple, R.W., B.A. Zaitlin, R. Boyd. 1992. Estuarine Tucker, A. Ganas. 2008. New constraints on sediment- facies models: conceptual basis and stratigraphic impli- flux-dependent river incision: implications for extract- cations. Journal of Sedimentary Petrology 62, 1130– ing tectonic signals from river profiles. Geology 36, 535– 1146. 538. Daniel, J.F. 1971. Channel Movement of Meandering Indi- Croke, J., S. Mockler. 2001. Gully initiation and road-to- ana Streams. U.S. Geological Survey Professional Paper stream linkage in a forested catchment, southeastern 732-A. Australia. Earth Surface Processes and Landforms 26, Daniels, M.D., B.L. Rhoads. 2004. Effect of large woody 205–217. debris configuration on three-dimensional flow struc- Crosby, B.T., K.X. Whipple. 2006. Knickpoint initiation ture in two low-energy meander bends at vary- and distribution within fluvial networks: 236 waterfalls ing stages. Water Resources Research 40, W11302. in the Waipaoa River, North Island, New Zealand. Geo- doi:10.1029/2004WR003181, 14 p. morphology 82, 16–38. Darby, S., D.A. Sear. 2008. River Restoration: Managing the Crosby, B.T., K.X. Whipple, N.M. Gasparini, C.W.Wobus. Uncertainty in Restoring Physical Habitat.JohnWiley& 2007. Formation of fluvial hanging valleys: theory Sons, Chichester. and simulation. Journal of Geophysical Research 112, Darby, S.E., M.J. Van De Wiel. 2003. Models in fluvial geo- F03S10. doi:10.1029/2006JF000566 morphology. In: G.M. Kondolf, H. Piegay,´ eds., Tools in Crossland, C.J., H.H. Kremer, H.J. Lindeboom, J.I.M. Fluvial Geomorphology. John Wiley & Sons, Chichester, Crossland, M.D.A. LeTissier, eds. 2005. Coastal Fluxes pp. 503–537. in the Anthropocene. Springer, Berlin. Darcy, H., 1856. Les Fontaines Publiques De La Ville De Crowder, D.W., H.V. Knapp. 2005. Effective discharge Dijon.VictorDalmont,Paris. recurrence intervals of Illinois streams. Geomorphology David, G.C.L., B.P. Bledsoe, D.M. Merritt, E.Wohl. 2009. 64, 167–184. Theimpactsofskislopedevelopmentonstreamchan- Curran, J.H., E.E. Wohl. 2003. Large woody debris and nel morphology in the White River National Forest, flow resistance in step-pool channels, Cascade Range, Colorado, USA. Geomorphology 103, 375–388. Washington. Geomorphology 51, 141–147. David, G.C.L., E. Wohl, S.E. Yochum, B.P. Bledsoe. 2010. Curtis, K.E., C.E. Renshaw, F.J. Magilligan, W.B. Dade. Controls on at-a-station hydraulic geometry in steep 2010. Temporal and spatial scales of geomorphic headwater streams, Colorado, USA. Earth Surface Pro- adjustments to reduced competency following flow reg- cesses and Landforms 35, 1820–1837. ulation in bedload-dominated systems. Geomorphology Davies, N.S., M.R. Gibling. 2011. Evolution of fixed- 118, 105–117. channel alluvial plains in response to Carboniferous Dale, V.H., S. Brown, M.O. Calderon,´ A.S. Montoya, R. E. vegetation. Nature Geoscience 4, 629–633. Mart´ınez. 2005. Projected land-use change for the east- Davies, T.R.H., A.J. Sutherland. 1983. Extremal hypothe- ern Panama Canal watershed and its potential impact. ses for river behavior. Water Resources Research 19, In: R.S. Harmon, ed., The Rio Chagres, Panama: A Mul- 141–148. tidisciplinary Profile of a Tropical Watershed.Springer, Davies, G., D.C. Woodroffe. 2010. Tidal estuary width Dordrecht, The Netherlands, pp. 337–345. convergence: theory and form in north Australian 266 References

estuaries. Earth Surface Processes and Landforms 35, port laws for predicting landscape form and dynam- 737–749. ics. In: P.R. Wilcock, R.M. Iverson, eds., Prediction in Davis, W.M. 1899. The geographical cycle. Geographical Geomorphology. American Geophysical Union Press, Journal 14, 481–504. Washington, DC, pp. 103–132. Davis, W.M. 1902a. Base level, grade, and peneplain. Jour- Dietrich, W.E., G. Day, G. Parker. 1999. The Fly River, nal of Geology 10, 77–111. Papua New Guinea: inferences about river dynamics, Davis, W.M. 1902b. River terraces in New England. Bul- floodplain sedimentation and fate of sediment. In: A.J. letin of the Museum of Comparative Zoology 38, 281– Miller,A.Gupta,eds.,Varieties of Fluvial Form.John 346. Wiley & Sons, Chichester, pp. 345–376. Davis, J.A. 2006. Naples and Napoleon: Southern Italy and Dietrich, W.E., G. Day, J.C. Rowland. 2007. The Depo- the European Revolutions 1780–1860.OxfordUniver- sitional Web on the Fly River Floodplain, Papua New sity Press, New York. Guinea. Abstracts, Annual Meeting of the American Dean, D.J., J.C. Schmidt. 2011. The role of feedback mech- Association of Petroleum Geologists, AAPG, Tulsa, anisms in historic channel changes of the lower Rio Oklahoma, v. 2007, p. 37. Grande in the Big Bend region. Geomorphology 126, Dietrich, W.E., T. Dunne. 1978. [Sediment budget for a 333–349. small catchment in mountainous terrain]. Zeitschrift de Rham, L.P., T.D. Prowse, S. Beltaos, M.P. Lacroix. Geomorphologie 29, 191–206. 2008a. Assessment of annual high-water events for the Dietrich, W.E., Dunne, T., 1993. The channel head. In: K. Mackenzie River basin, Canada. Hydrological Processes Beven, M.J. Kirkby, eds., Channel Network Hydrology. 22, 3864–3880. John Wiley & Sons, Hoboken, NJ, pp. 175–219. de Rham, L.P., T.D. Prowse, B.R. Bonsal. 2008b. Tempo- Dietrich, W.E.,J.W.Kirchner, H. Ikeda, F.Iseya. 1989. Sed- ral variations in river-ice break-up over the Macken- iment supply and the development of the coarse surface zie River Basin, Canada. Journal of Hydrology 349, 441– layer in gravel-bedded rivers. Nature 340, 215–217. 454. Dietrich, W.E., C.J. Wilson, D.R. Montgomery, J. McK- Derksen, C., R. Brown. 2012. Spring Snow Cover Extent ean. 1993. Analysis of erosion thresholds, channel net- Reductions in the 2008–2012 Period Exceeding Climate works, and landscape morphology using a digital ter- Model Projections. Geophysical Research Letters 39, rain model. Journal of Geology 101, 259–278. L19504. doi:10.1029/2012GL053387 Dietrich, W.E., C.J. Wilson, D.R. Montgomery, J. McK- Deroanne, C., F. Petit. 1999. Longitudinal evaluation of ean, R. Bauer. 1992. Erosion thresholds and land sur- the bed load size and of its mobilization in a gravel face morphology. Geology 20, 675–679. bedriver.In:R.Casale,C.Margottini,eds.,Floods and Dingwall, P.R. 1972. Erosion by overland flow on an alpine Landslides. Springer, Berlin, pp. 335–342. debris slope. In: H.O. Slaymaker, H.J. McPherson, eds., de Scally, F.A., I.F. Owens. 2004. Morphometric controls Mountain Geomorphology: Geomorphological Processes and geomorphic responses on fans in the Southern in the Canadian Cordillera. Tantalus Resources Ltd., Alps, New Zealand. Earth Surface Processes and Land- Vancouver, pp. 113–120. forms 29, 311–322. Diplas, P., C. L. Dancey, A.O. Celik, M. Valyrakis, K. de Scally, F.E.S., O. Slaymaker, I. Owens. 2001. [Mor- Greer, T. Akar, 2008. The role of impulse on the ini- phometric controls and basin response in the Cas- tiation of particle movement under turbulent flow con- cade Mountains]. Geografiska Annaler 83A, 117– ditions. Science 322, 717–720. 130. Dixon, R.K., S. Brown, R.A. Houghton, A.M. Solomon, Dethier, D.P. 2001. Pleistocene incision rates in the west- M.C. Trexler, J. Wisniewski. 1994. Carbon pools and ern United States calibrated using Lava Creek B tephra. flux of global forest ecosystems. Science 263, 185– Geology 29, 783–786. 190. Dietrich, W.E. 1987. Mechanics of flow and sediment Do Carmo, J.S.A. 2007. The environmental impact and transport in river bends. In: K.S. Richards, ed., River risks associated with changes in fluvial morphody- Channels: Environment and Process.Blackwell,Oxford, namic processes. In: J.P. Lobo Ferreira, J.M.P. Viera, pp. 179–227. eds., Water in Celtic Countries: Quantity, Quality and Dietrich, W.E., D.G. Bellugi, L.S. Sklar, J.D. Stock, Climate Variability. IAHS Publication no. 310, pp. 307– A.M.Heimsath, J.J. Roering. 2003. Geomorphic trans- 319. References 267

Dolan, R., A. Howard, D. Trimble. 1978. Structural con- Dunne, T., D.V. Malmon, S.M. Mudd. 2010. A rain splash troloftherapidsandpoolsoftheColoradoRiverinthe transport equation assimilating field and laboratory Grand Canyon. Science 202, 629–631. measurements. Journal of Geophysical Research 115, Domenico, P.A.,F.W.Schwartz. 1998. Physical and Chem- F01001. doi:10.1029/2009JF001302, 16 p. ical Hydrogeology. John Wiley & Sons, New York. Dunne, T., L.A.K. Mertes, R.H. Meade, J.E. Richey, B.R. Douglas, I. 2009. Hydrological investigations of forest dis- Forsberg. 1998. Exchanges of sediment between the turbance and land cover impacts in South-East Asia: a flood plain and channel of the Amazon River in Brazil. review. PhilosophicalTransactionsoftheRoyalSociety Geological Society of America Bulletin 110, 450–467. B: Biological Sciences 354, 1725–1738. Dunne, T., K.X. Whipple, B.F. Aubry. 1995. Microtopog- Douglass, J., M. Schmeeckle. 2007. Analogue modeling raphy of hillslopes and initiation of channels by Hor- of transverse drainage mechanisms. Geomorphology 84, ton overland flow. In: J.E. Costa,A.J. Miller, K.W. Pot- 22–43. ter, P.R. Wilcock, eds., Natural and Anthropogenic Influ- Downs, P.W., G. Priestnall. 2003. Modelling catchment ences in Fluvial Geomorphology.AmericanGeophysical processes.In:G.M.Kondold,H.M.Piegay,´ eds., Tools Union Press, Washington, DC, pp. 27–44. in Fluvial Geomorphology.JohnWiley&Sons,Chich- Dunne, T., Zhang, W.,Aubry, B.F.,1991. Effects of rainfall, ester, pp. 205–230. vegetation, and microtopography on infiltration and Doyle, M.W., S.H. Ensign. 2009. Alternative reference runoff. Water Resources Research 27(9), 2271–2285. frames in river system science. BioScience 59, 499–510. Dury, G.H. 1965. Theoretical Implications of Underfit Drever, J.I. 1988. The geochemistry of natural waters,2nd Streams. U.S. Geological Survey Professional Paper 452- edn. Prentice Hall, Englewood Cliffs, NJ. C, Washington, DC, 43 p. Drigo, R. 2005. Trends and patterns of tropical land use Dust, D., E. Wohl. 2012b. Conceptual model for complex change. In: M. Bonell, L.A. Bruijnzeel, eds., Forests, river responses using an expanded Lane’s relation. Geo- Water and People in the Humid Tropics. Cambridge Uni- morphology 139–140, 109–121. versity Press, Cambridge, pp. 9–39. Dust, D., E. Wohl. 2012a. Characterization of the Dubinski, I.M., E. Wohl. 2012. Relationships between hydraulics at natural step crests via weir flow concepts. block quarrying, bed shear stress, and stream power: a Water Resources Research 48, W09542. doi:10.1029/ physical model of block quarrying of a jointed bedrock 2011WR011724. channel. Geomorphology 180–181, 66–81. Duvall, A., E. Kirby, D. Burbank. 2004. Tectonic and Dunne, T. 1978. Field studies of hillslope flow processes. lithologic controls on bedrock channel profiles and In: M.J. Kirkby, ed., Hillslope Hydrology.JohnWiley& processes in coastal California. Journal of Geophysical Sons, Toronto, pp. 227–293. Research 109, F03002. doi:10.1029/2003JF000086, 18 p. Dunne, T. 1980. Formation and controls of channel net- Dyke,L.,J.Aylsworth,N.Burgess,F.Nixon,F.Wright. works. Progress in Physical Geography 4, 211–239. 1997. Permafrost in the Mackenzie basin, its influence Dunne, T., 1990. Hydrology, mechanics, and geomorphic on land-altering processes, and its relationship to cli- implications of erosion by subsurface flow. In: C.G. mate change. In: S. Cohen, ed., Mackenzie Basin Impact Higgins, D.R. Coates, eds., Groundwater Geomorphol- Study Final Report. Environment Canada, pp. 125–145. ogy; The Role of Subsurface Water in Earth-Surface Pro- Dykes, A.P., J.B. Thornes. 2000. Hillslope hydrology in cesses and Landforms. Geological Society of America, tropical rainforest steeplands in Brunei. Hydrological Special Paper 252, Boulder, CO, pp. 1–28. Processes 14, 215–235. Dunne, T., R.E. Aalto. 2013. Large river floodplains. In: E. Dynesius, M., C. Nilsson. 1994. Fragmentation and flow Wohl, ed., Treatise on Fluvial Geomorphology. Elsevier. regulation of river systems in the northern third of the Dunne, R., R.D. Black. 1970a. An experimental investi- world. Science 266, 753–762. gation of runoff production in permeable soils. Water Eaton, B.C. 2006. Bank stability analysis for regime mod- Resources Research 6(2), 478–490. els of vegetated gravel bed rivers. Earth Surface Processes Dunne, R., R.D. Black. 1970b. Partial area contributions to and Landforms 31, 1438–1444. storm runoff in a small New England watershed. Water Eaton, B.C. 2013. Hydraulic geometry: empirical inves- Resources Research 6(5), 1296–1311. tigations and theoretical approaches. In: E. Wohl, ed., Dunne, T., L.B. Leopold. 1978. Water in Environmental Treatise on Fluvial Geomorphology. Elsevier, Amster- Planning.W.H.Freeman,NewYork. dam, pp. 313–329. 268 References

Eaton, B.C., M. Church, T.R.H. Davies. 2006. A con- Elsenbeer, H., Vertessy, R.A., 2000. Stormflow gener- ceptual model for meander initiation in bedload- ation and flowpath characteristics in an Amazonian dominated streams. Earth Surface Processes and Land- rainforest catchment. Hydrological Processes 14, 2367– forms 31, 875–891. 2381. Eaton, B.C., M. Church, R.G. Millar. 2004. Rational Ely, L.L., Y. Enzel, V.R. Baker, D.R. Cayan. 1993. A 5000- regime model of alluvial channel morphology and year record of extreme floods and climate change in the response. Earth Surface Processes and Landforms 29, southwestern United States. Science 262, 410–412. 511–529. England,J.F.,J.E.Godaire,R.E.Klinger,T.R.Bauer, Eaton, B.C., M.F. Lapointe. 2001. Effects of large floods P.Y. Julien. 2010. Paleohydrologic bounds and extreme on sediment transport and reach morphology in the flood frequency of the Upper Arkansas River, Colorado, cobble-bed Sainte Marguerite River. Geomorphology USA. Geomorphology 124, 1–16. 40, 291–309. England, J.F., R.D. Jarrett, J.D. Salas. 2003. Data-based Eaton, B.C., R.G. Millar. 2004. Optimal alluvial channel comparisons of moments estimators using historical width under a bank stability constraint. Geomorphology and paleoflood data. Journal of Hydrology 278, 172–196. 62, 35–45. Ergenzinger, P., C. de Jong. 2002. Perspectives on bed Eaton, L.S., B.A. Morgan, R.C. Kochel, A.D. Howard. load measurement. In: J. Bogen, ed., Erosion and Sed- 2003. Role of debris flows in long-term landscape iment Transport Measurement in Rivers: Technological denudation in the central Appalachians of Virginia. and Methodological Advances. IAHS Publication no. Geology 31, 339–342. 283, pp. 113–125. Eckhardt, K., U. Ulbrich. 2003. Potential impacts of cli- Erslev, E.A. 2001. Multistage, multidirectional tertiary mate change on groundwater recharge and streamflow shortening and compression in north-central New in a central European low mountain range. Journal of Mexico. Geological Society of America Bulletin 113, Hydrology 284, 244–252. 63–74. Eckley, M.S., D.L. Hinchliff. 1986. Glen Canyon Dam’s Ettema, R., S.F. Daly. 2004. Sediment Transport Under Ice. quick fix. Civil Engineering/American Society of Civil U.S. Army Corps of Engineers, Cold Regions Research Engineers 56, 46–48. and Engineering Laboratory, Hanover, New Hamp- Edmonds, D.A., D.C.J.D. Hoyal, B.A. Sheets, R.L. Slinger- shire, ERDC/CRREL TR-04–20. land. 2009. Predicting delta avulsions: implications for Ettema, R., E.W. Kempema. 2012. River-ice effects on coastal wetland restoration. Geology 37, 759–762. gravel-bed channels. In: M. Church, P.M. Biron, A.G. Edmonds, D.A., J.B. Shaw, D. Mohrig. 2011. Topset- Roy, eds., Gravel-bed Rivers: Processes, Tools, Environ- dominated deltas: a new model for river delta stratig- ments. John Wiley & Sons, Hoboken, NJ, pp. 525–540. raphy. Geology 39, 1175–1178. European Commission. 2000. Directive 2000/60/EC of the Eggertsson, O. 1994. Mackenzie River driftwood – a den- European Parliament and of the Council of 23 October drochronological study. Arctic 47 128–136. 2000 Establishing a Framework for Community Action Egozi, R., P. Ashmore. 2008. Defining and measuring in the Field of Water Policy. Official Journal of the Euro- braiding intensity. EarthSurfaceProcessesandLand- pean Communities, L 327. forms 33, 2121–2138. Evette, A., S. Labonne, F. Rey, F. Liebault, O. Jancke, J. Ehlen, J., E. Wohl. 2002. Joints and landform evolution in Girel. 2009. History of bioengineering techniques for bedrock canyons. Transactions of the Japanese Geomor- erosion control in rivers in western Europe. Environ- phological Union 23, 237–255. mental Management 43, 972–984. Einstein, H.A., R.B. Banks. 1950. Fluid resistance of com- Fagherazzi, S. 2008. Self-Organization of Tidal Deltas.Pro- posite roughness. Transactions of the American Geo- ceedings of the National Academy of Sciences 105, physical Union 31, 603–610. 18692–18695. Einstein, H.A., H.W. Shen. 1964. A study of meander- Fagherazzi, S., D.J. Furbish. 2001. On the shape and ing in straight alluvial channels. Journal of Geophysical widening of salt marsh creeks. Journal of Geophysical Research 69, 5239–5247. Research 106, 991–1003. Elsenbeer, H. 2001. Hydrological flowpaths in tropical Fagherazzi, S., I. Overeem. 2007. Models of deltaic and rainforest soilscapes – a review. Hydrological Processes inner continental shelf landform evolution. Annual 15, 1751–1759. Reviews of Earth and Planetary Sciences 35, 685–715. References 269

Fahnestock, R.K. 1963. Morphology and Hydrology Ferguson, R. 2013. Reach-scale flow resistance. In: E. of a Glacial Stream – White River, Mount Rainier, Wohl, ed., Treatise on Fluvial Geomorphology. Elsevier, Washington. U.S. Geological Survey Professional Amsterdam, pp. 50–68. Paper 422A, 61 p. Finnegan, N.J. 2013. Interpretation and downstream Fairweather, P.G.1999. State of environment indicators of correlation of bedrock river terrace treads created ‘river health’: exploring the metaphor. Freshwater Biol- from propagating knickpoints. Journal of Geophysical ogy 41, 211–220. Research: Earth Surface 118. doi:10.1029/2012JF002534 Falke, J.A., K.D. Fausch, R. Magelky, A. Aldred, D.S. Finnegan, N.J., L.S. Sklar, T.K. Fuller. 2007. Interplay Durnford, L.K. Riley, R. Oad. 2011. The role of ground- of sediment supply, river incision, and channel mor- water pumping and drought in shaping ecological phology revealed by the transient evolution of an futures for stream fishes in a dryland river basin of experimental bedrock channel. Journal of Geophysical the western Great Plains, USA. Ecohydrology 4, 682– Research 112, F03S11. doi:10.1029/2006JF000569, 17 p. 697. Flemming, B.W. 1988. Zur klassifikation subaquatischer, Fanelli, R.M., L.K. Lautz. 2008. Patterns of water, heat, stromungstrans versaler transportkorper. Bochumer and solute flux through streambeds around small dams. Geologische und Geotechnische Arbeiten 29, 44–47. Ground Water 46, 671–687. (in German, cited in Venditti, 2013) Farnsworth, K.L., J.D. Milliman. 2003. Effects of cli- Flener,C.,E.Lotsari,P.Alho,J.Kayhk¨ o.¨ 2012. Com- maticandanthropogenicchangeonsmallmountainous parison of empirical and theoretical remote sensing rivers; the Salinas River example. Global and Planetary based bathymetry models in river environments. River Change 39, 53–64. Research and Applications 28, 118–133. Fausch, K.D., K.R. Bestgen. 1997. Ecology of fishes indige- Flick, R.E. 1993. The myth and reality of southern nous to the central and southwestern Great Plains. In: California beaches. Shore and Beach 61, 3–13. F.L. Knopf, F.B. Samson, eds., Ecology and Conservation Flint, J. 1974. Stream gradient as a function of order, of Great Plains Vertebrates. Springer-Verlag, New York, magnitude, and discharge. Water Resources Research 10, pp. 131–166. 969–973. Fenn, C.R. 1987. Sediment transfer processes in alpine Flint, A.L., L.E. Flint, M.D. Dettinger. 2008. Modeling soil glacier basins. In: A.M. Gurnell, M.J. Clark, eds., Glacio- moisture processes and recharge under a melting snow- fluvial Sediment Transfer.JohnWiley&Sons,Chich- pack. Vadose Zone Journal 7, 350–257. ester, pp. 59–85. Florsheim, J.L., E.A. Keller, D.W. Best. 1991. Fluvial sed- Feo, M.L., A. Ginebreda, E. Eljarrat, D. Barcelo. 2010. iment transport in response to moderate storm flows Presence of pyrethroid pesticides in water and sedi- following chaparral wildfire, Ventura County, southern ments of Ebro River delta. Journal of Hydrology 393, California. Geological Society of America Bulletin 103, 156–162. 504–511. Ferguson, R.I. 1986. Hydraulics and hydraulic geometry. Foley,J.A.,R.DeFries,G.P.Asner,C.Barford,G.Bonan, Progress in Physical Geography 10, 1–31. S.R. Carpenter,F.S. Chapin, M.T. Coe, G.C. Daily, H.K. Ferguson, R.I. 1987. Hydraulic and sedimentary con- Gibbs, J.H. Helkowski, T. Holloway, E.A. Howard, C.J. trols of channel pattern. In: K.S. Richards, ed., River Kucharik, C. Monfreda, J.A. Patz, I.C. Prentice, N. Channels: Environment and Process.Blackwell,Oxford, Ramankutty, P.K.Snyder. 2005. Global consequences of pp. 129–158. land use. Science 309, 570–574. Ferguson, R.I. 2003. Emergence of abrupt gravel to Folk, R.L. 1980. Petrology of sedimentary rocks.Hemphill sand transitions along rivers through sorting processes. Publishing Company, Austin, TX, 182 p. Geology 31, 159–162. Fonstad, M.A. 2003. Spatial variation in the power of Ferguson, R.I. 2007. Flow resistance equations for gravel- mountain streams in the Sangre de Cristo Mountains, and boulder-bed streams. Water Resources Research 43, New Mexico. Geomorphology 55, 75–96. W05427. Fox, R.W., A.T. McDonald. 1978. Introduction to Fluid Ferguson, R.I. 2012. River channel slope, flow resis- Mechanics,2ndedn.JohnWiley&Sons,NewYork, tance, and gravel entrainment thresholds. Water 684 p. Resources Research 48, W05517. doi:10.1029/2011WR0 Francis, R.A., D. Corenblit, P.J. Edwards. 2009b. Perspec- 10850 tives on biogeomorphology, ecosystem engineering and 270 References

self-organisation in island-braided fluvial ecosystems. Fryirs, K., G.J. Brierley. 2011. Variability in sediment Aquatic Sciences 71, 290–304. delivery and storage along river courses in Bega catch- Francis, J.A., D.M. White, J.J. Cassano, W.J. Gutowski, L.D. ment, NSW, Australia: implications for geomorphic Hinzman, M.M. Holland, M.A. Steele, C.J. Vor¨ osmarty.¨ river recovery. Geomorphology 38, 237–265. 2009a. An arctic hydrologic system in transition: Fryirs, K., G.J. Brierley, W.D. Erskine. 2012. Use of ergodic feedbacks and impacts on terrestrial, marine, and reasoning to reconstruct the historical range of variabil- human life. Journal of Geophysical Research – Oceans ity and evolutionary trajectory of rivers. Earth Surface 114, G4. Processes and Landforms 37, 763–773. Fredlund, D.G., N.R. Morgenstern, R.A. Widger. 1978. Fryirs, K.A., G.J. Brierley, N.J. Preston, J. Spencer. 2007. The shear strength of unsaturated soils. Canadian Catchment scale (dis)connectivity in sediment flux in Geotechnical Journal 15, 313–321. the upper Hunter catchment, New South Wales, Aus- Freeman, M.C., C.M. Pringle, C.R. Jackson. 2007. Hydro- tralia. Geomorphology 84, 297–316. logic connectivity and the contribution of stream head- Furbish, D.J., E.M. Childs, P.K. Haff, M.W. Schmeeckle. waters to ecological integrity at regional scales. Journal 2009a. Rain splash of soil grains as a stochastic of the North American Water Resources Association 43, advection-dispersion process, with implications for 5–14. desert plant-soil interactions and land-surface evolu- Frey, P., M. Church. 2012. Gravel transport in granular tion. Journal of Geophysical Research, 114, F00A03. perspective. In: M. Church, P.M. Biron, A.G. Roy, eds., doi:10.1029/2009JF001265, 18 p. Gravel-bedRivers:Processes,Tools,Environments.John Furbish,D.J.,P.K.Haff,W.E.Dietrich,A.M.Heimsath. Wiley & Sons, pp. 39–55. 2009b. Statistical description of slope-dependent Frey, K.E., J.W. McClelland. 2009. Impacts of permafrost soil transport and the diffusion-like coefficient. degradationonarcticriverbiogeochemistry.Hydrolog- Journal of Geophysical Research 114, F00A05. ical Processes 23, 169–182. doi:10.1029/2009JF001267, 19 p. Friedman, J.M., V.J. Lee. 2002. Extreme floods, channel Galat, D.L., L.H. Fredrickson, D.D. Humburg, K.J. change, and riparian forests along ephemeral streams. Bataille, J.R. Bodie, J. Dohrenwend, G.T. Gelwicks, Ecological Monographs 72, 409–425. J.E. Havel, D.L. Helmers, J.B. Hooker, J.R. Jones, M.F. Frings, R.M., B.M. Berbee, G. Erkens, M.G. Kleinhans, Knowlton, J. Kubisiak, J. Mazourek, A.C. McColpin, M.J.P. Gouw. 2009. Human-induced changes in bed R.B. Renken, R.D. Semlitsch. 1998. Flooding to restore shear stress and bed grain size in the River Waal (The connectivity of regulated, large-river wetlands. Bio- Netherlands) during the past 900 years. Earth Surface Science 48, 721–733. Processes and Landforms 34, 503–514. Galay, V.J. 1983. Causes of river bed degradation. Water Fripp, J.B., P. Diplas. 1993. Surface sampling in gravel Resources Research 19, 1057–1090. streams. Journal of Hydraulic Engineering 119, 473–490. Galloway, J.N., F.J. Dentener, D.G. Capone, E.W. Boyer, Frissell, C.A., W.J. Liss, C.E. Warren, M.D. Hurley. 1986. R.W. Howarth, S.P. Seitzinger, G.P. Asner, C.C. Cleve- A hierarchical framework for stream habitat classifica- land, P.A. Green, E.A. Holland, D.M. Karl, A.F. tion: viewing streams in a watershed context. Environ- Michaels, J.H. Porter, A.R. Townsend, C.J. Vor¨ osmarty.¨ mental Management 10, 199–214. 2004. Nitrogen cycles: past, present, and future. Biogeo- Froehlich, W., L. Starkel. 1987. [Normal and extreme chemistry 70, 153–226. monsoon rains – their role in the shaping of the Dar- Gallaway, J.M., Y.E.Martin, E.A. Johnson. 2009. Sediment jeeling Himalaya]. Studia Geomorphologica Carpatho- transport due to tree root throw: integrating tree pop- Balcanica 21, 129–158. ulation dynamics, wildfire, and geomorphic response. Froese, D.G., D.G. Smith, D.T. Clement. 2005. Character- Earth Surface Processes and Landforms 34, 1255–1269. izing large river history with shallow geophysics; mid- Galy, A., C. France-Lanord, B. Lartiges. 2008. [Load- dle Yukon River, Yukon Territory and Alaska. Geomor- ing and fate of particulate organic carbon from the phology 67, 391–406. Himalays to the Ganges-Brahmaputra delta]. Geochim- Fryirs, K., G.J. Brierley. 2009. Naturalness and ica et Cosmochimica Acta 72, 1767–1787. place in river rehabilitation. Ecology and Society Gannett, M.W., M. Manga, K.E. Lite. 2003. Groundwa- 14(1): 20, http://www.ecologyandsociety.org/vol14/ ter hydrology of the Upper Deschutes basin and its iss1/art20 influence on streamflow. In: J.E. O’Connor, G.E. Grant, References 271

eds., A Peculiar River: Geology, Geomorphology, and Gillan, B.J., J.T. Harper, J.N. Moore. 2010. Timing of Hydrology of the Deschutes River, Oregon.American present and future snowmelt from high elevations Geophysical Union Press, Washington, DC, pp. 31–49. in northwest Montana. Water Resources Research 46, Gao, Y., R.M. Vogel, C.N. Kroll, N.L. Poff, J.D. Olden. W01507. doi:10.1029/2009WR007861 2009. Development of representative indicators of Gilman, K., M.D. Newson. 1980. Soils Pipes and Pipeflow – hydrologic alteration. Journal of Hydrology 374, 136– A Hydrological Study in Upland Wales.Geobooks,Nor- 147. wich. Garcia, A.F. 2006. Thresholds of strath genesis deduced Giorgi, F., C. Shields-Brodeur, G.T. Bates. 1994. Regional from landscape response to stream piracy by Pancho climate change scenarios over the United States pro- Rico Creek in the Coast Ranges of central California. duced with a nested regional climate model: spatial and American Journal of Science 306, 655–681. seasonal characteristics. Journal of Climatology 7, 375– Gaston, K.J., J.I. Spicer. 2004. Biodiversity: An Introduc- 399. tion, 2nd edn. Blackwell Publishing, 191 p. Glock, W.S. 1931. The development of drainage systems: Gaudet, J.M., A.G. Roy. 1995. Effects of bed morphology asynopticview.Geographical Review 21, 475–482. on flow mixing length at river confluences. Nature 373, Goldrick, G., P.Bishop. 2007. Regional analysis of bedrock 138–139. stream long profiles: evaluation of Hack’s SL form, and Geleynse, N., J.E.A. Storms, D.-J.R. Walstra, H.R.A. formulation and assessment of an alternative (the DS Jagers, Z.B. Wang, M.J.F. Stive. 2011. Controls on form). Earth Surface Processes and Landforms 32, 649– river delta formation; insights from numerical mod- 671. eling. Earth and Planetary Science Letters 302, 217– Goldsmith,S.T.,A.E.Carey,W.B.Lyons,S.Kao,T. 226. Lee, J. Chen. 2008. Extreme storm events, landscape Gellis, A.C., R.M.T. Webb, S.C. McIntyre, W.J. Wolfe. denudation, and carbon sequestration: Typhoon Min- 2006. Land-use effects on erosion, sediment yields, and dulle, Choshui River, Taiwan. Geology 36, 483–486. doi: reservoir sedimentation: a case study in the Lago Lo´ıza 10.1130/G24624A.1 basin, Puerto Rico. Physical Geography 27, 39–69. Goldsmith, E., N. Hildyard. 1984. The Social and Envi- Genereux, D.P., M. Jordan. 2006. Interbasin groundwater ronmental Effects of Large Dams. Sierra Club, San Fran- flow and groundwater interaction with surface water in cisco, CA, 404 p. a lowland rainforest, Costa Rica: a review. Journal of Gomez, B., R.L. Naff, D.W. Hubbell. 1989. Temporal vari- Hydrology 320, 385–399. ations in bedload transport rates associated with the Germann, P.F., 1990. Preferential flow and the genera- migration of bedforms. Earth Surface Processes and tion of runoff 1. Boundary layer flow theory. Water Landforms 14, 135–156. Resources Research 26, 3055–3063. Goode, J.R., C.H. Luce, J.M. Buffington. 2012. Enhanced Germanoski, G., S.A. Schumm. 1993. Changes in braided sediment delivery in a changing climate in semi-arid river morphology resulting from aggradation and mountain basins: implications for water resource man- degradation. Journal of Geology 101, 451–466. agement and aquatic habitat in the northern Rocky Gerrard, J. 1990. Mountainous Environments: An Exami- Mountains. Geomorphology 139–140, 1–15. nation of the Physical Geography of Mountains. The MIT Goode, J.R., E. Wohl. 2010a. Coarse sediment transport Press, Cambridge, MA, 317 p. in a bedrock channel with complex bed topography. Gilbert, G.K. 1877. Report on the Geology of the Henry Water Resources Research 46, W11524. doi:10.1029/ Mountains. U.S. Geological Survey, Washington, DC. 2009WR008135 Gilbert, G.K. 1885. The Topographic Features of Lake Goode, J.R., E. Wohl. 2010b. Substrate controls on the lon- Shores.U.S.GeologicalSurveyAnnualReport5, gitudinal profile of bedrock channels: implications for pp. 69–123. reach-scale roughness. Journal of Geophysical Research Gilbert, G.K. 1914. Transportation of Debris by Running Earth Surfaces 115, F03018. doi:10.1029/2008JF Water. U.S. Geological Survey Professional Paper 86, 001188 221 p. Goolsby, D.A., W.A. Battaglia, G.B. Lawrence, R.S. Artz, Gilbert, G.K. 1917. Hydraulic-Mining Debris in the B.T. Aulenbach, R.P. Hooper, D.R. Keeney, G.J. Stens- Sierra Nevada. U.S. Geological Survey Professional land. 1999. Flux and Sources of Nutrients in the Paper 105. Mississippi-Atchafalaya River Basin: Topic 3 Report for 272 References

the Integrated Assessment on Hypoxia in the Gulf of Mex- tions: channel response to basin-wide sediment load- ico. NOAA Coastal Ocean Program Decision Analysis ingatMountPinatubo,Philippines.Geological Society Series No. 17. NOAA Coastal Ocean Program, Silver of America Bulletin 117, 195–211. Spring, MD, 130 p. Gran, K.B., D.R. Montgomery, J.C. Halbur. 2011. Long- Gooseff, M.N. 2010. Defining hyporheic zones – advanc- term elevated post-eruption sedimentation at Mount ing our conceptual and operational definitions of where Pinatubo, Philippines. Geology 39, 367–370. stream water and groundwater meet. Geographical Gran, K.B., D.R. Montgomery, D.G. Sutherland. 2006. Compass 4, 945–955. Channel bed evolution and sediment transport under Gorrick, S., J.F. Rodr´ıguez. 2012. Sediment dynam- declining sand inputs. Water Resources Research 42, ics in a sand bed stream with riparian vegetation. W10407. doi:10.1029/2005WR004306 Water Resources Research 48, W02505. doi:10.1029/ Grant, G.E. 1997. Critical flow constrains flow hydraulics 2011WR011030 in mobile-bed streams: a new hypothesis. Water Goulding, H.L., T.D. Prowse, S. Beltaos. 2009b. Spatial and Resources Research 33, 349–358. temporal patterns of break-up and ice-jam flooding in Grant, G.E., J.E. O’Connor, M.G. Wolman. 2013. A river the Mackenzie Delta, NWT. Hydrological Processes 23, runs through it: conceptual models in fluvial geomor- 2654–2670. phology. In: E. Wohl, ed., Treatise on Fluvial Geomor- Goulding, H.L., T.D. Prowse, B. Bonsal. 2009a. Hydrocli- phology. Elsevier, Amsterdam, pp. 6–20. matic controls on the occurrence of break-up and ice- Grant, G.E., J.C. Schmidt, S.L. Lewis. 2003. A geological jam flooding in the Mackenzie Delta, NWT, Canada. framework for interpreting downstream effects of dams Journal of Hydrology, 379, 251–267. onrivers.In:J.E.O’Connor,G.E.Grant,eds.,APeculiar Graf, W.L. 1978. Fluvial adjustments to the spread of River: Geology, Geomorphology, and Hydrology of the tamarisk in the Colorado Plateau region. Geological Deschutes River, Oregon. American Geophysical Union Society of America Bulletin 89, 1491–1501. Press, Washington, DC, pp. 203–219. Graf, W.L. 1983. Flood-related channel change in an arid Green,C.,A.Fernandez-Bilbao.´ 2006. Implementing the region river. Earth Surface Processes and Landforms 8, water framework directive: how to define a “competent 125–139. authority.” Journal of Contemporary Water Research and Graf, W.L. 1988. Fluvial Processes in Dryland Rivers. Education 135, 65–73. Springer-Verlag, Berlin. Green, P.A., C.J. Vor¨ osmarty,¨ M. Meybeck, J.N. Galloway, Graf, W.L. 1996. Geomorphology and policy for restora- B.J. Peterson, E.W.Boyer. 2004. Pre-industrial and con- tion of impounded American rivers: what is ‘natu- temporary fluxes of nitrogen through rivers: a global ral?’ In: B.L. Rhoads, C.E. Thorn, eds., The Scientific assessment based on typology. Biogeochemistry 68, 71– Nature of Geomorphology.JohnWiley&Sons,New 105. York, pp. 443–473. Gregory, K.J. 2006. The human role in changing river Graf, W.L. 2001. Damage control: restoring the physical channels. Geomorphology 79, 172–191. integrity of America’s rivers. Annals of the Association Gregory, K.J., C.C. Park. 1974. Adjustment of river of American Geographers 91, 1–27. channelcapacitydownstreamfromareservoir.Water Graf, W.L.2006. Downstream hydrologic and geomorphic Resources Research 10, 870–873. effects of large dams on American rivers. Geomorphol- Gregory, S.V.,F.J. Swanson, W.A.McKee, K.W.Cummins. ogy 79, 336–360. 1991. An ecosystem perspective of riparian zones. Bio- Graf, W.L., J. Stromberg, B. Valentine. 2002. Rivers, Science 41, 540–551. dams, and willow flycatchers: a summary of their sci- Griffin,E.R.,J.W.Kean,K.R.Vincent,J.D.Smith,J.M. ence and policy connections. Geomorphology 47, 169– Friedman. 2005. Modeling effects of bank friction and 188. woody bank vegetation on channel flow and bound- Grahame, T.J., R.B. Schlesinger. 2007. Health effects of ary shear stress in the Rio Puerco, New Mexico. Jour- airborne particulate matter: do we know enough to nal of Geophysical Research 110, F04023. doi:10.1029/ consider regulating specific particle types or sources? 2005JF000322 Inhalation Toxicology 19, 457–481. Griffiths, G.A. 1979. Recent sedimentation history of the Gran, K.B., D.R. Montgomery. 2005. Spatial and tempo- Waimakariri River, New Zealand. Journal of Hydrology ral patterns in fluvial recovery following volcanic erup- New Zealand 18, 6–28. References 273

Griffiths, G.A.1987. Form resistance in gravel channels Gurnell, A.M., M. Lee, C. South. 2007. Urban rivers: with mobile beds. Journal Hydraulic Engineering 115, hydrology, geomorphology, ecology and opportunities 340–355. for change. Geography Compass 1, 1–20. Griffiths, G.A., M.A. Carson. 2000. Channel width for Gurnell, A.M., J.-L. Peiry, G.E. Petts. 2003. Using histor- maximum bedload transport capacity in gravel-bed ical data in fluvial geomorphology. In: G.M. Kondolf, rivers, South Island, New Zealand. Journal of Hydrology H. Piegay,´ eds., Tools in Geomorphology.JohnWiley& New Zealand 39, 107–126. Sons, Chichester, pp. 77–101. Griffiths, P.G., R. Hereford, R.H. Webb. 2006. Sediment Gurnell, A.M., G. Petts. 2006. Trees as riparian engineers: yield and runoff frequency of small drainage basins the Tagliamento River, Italy. Earth Surface Processes and in the Mojave Desert, USA. Geomorphology 74, 232– Landforms 31, 1558–1574. 244. Guthrie, R.H., S.G. Evans. 2007. Work, persistence, and Grimm, N.B., S.G. Fisher. 1992. Responses of arid-land formativeevents;thegeomorphicimpactoflandslides. streams to changing climate. In: P. Firth, S.G. Fisher, Geomorphology 88, 266–275. eds., Global Climate Change and Freshwater Ecosystems. Gyalistras, D., C. Schaer, H.C. Davies, H. Wanner. 1997. Springer, New York, pp. 211–233. FutureAlpineclimate.In:P.Cebon,U.Dahinden,H.C. Guneralp,¨ I.,̇ R. Marston. 2012. Process-form linkages in Davies,D.Imboden,C.C.Jaeger,eds.,AViewfrom meander morphodynamics: bridging theoretical mod- the Alps: Regional Perspectives on Climate Change.MIT eling and real world complexity. Progress in Physical Press, Boston, MA, pp. 171–224. Geography 35, 718–746. Habersack,H.M.,H.Piegay.´ 2008. River restoration in the Guo, L., C.-L. Ping, R.W. Macdonald. 2007. Mobilization Alps and their surroundings: past experience and future pathways of organic carbon from permafrost to arctic challenges. In: H. Habersack, H. Piegay,´ M. Rinaldi, rivers in a changing climate. Geophysical Research Let- eds., Gravel-bed Rivers VI: From Process Understanding ters 34, L13603. to River Restoration. Elsevier, Amsterdam, pp. 703–737. Gupta, A. 1995. Magnitude, frequency, and special fac- Hack, J.T. 1957. Studies of Longitudinal Profiles in Vir- tors affecting channel form and processes in the sea- ginia and Maryland.U.S.GeologicalSurveyProfes- sonal tropics. In: J.E. Costa,A.J. Miller, K.W.Potter, P.R. sional Paper 294-B. Wilcock, eds., Natural and Anthropogenic Influences in Hack, J.T. 1960. Interpretation of erosional topography in Fluvial Geomorphology. American Geophysical Union humid temperate regions. American Journal of Science Press, Washington, DC, pp. 125–136. 258-A, 80–97. Gupta, A., A. Dutt. 1989. [The Auranga: description of a Hack, J.T. 1973. Stream-profile analysis and stream- tropical river]. Zeitschriftur f¨ Geomorphologie 33, 73– gradient index. Journal of Research of the U.S. Geological 92. Survey 1, 421–429. Guralnik, B., A. Matmon, Y. Avni, N. Porat, D. Fink. Haga, H., T. Kumagai, K. Otsuki, S. Ogawa. 2002. Trans- 2010. Constraining the evolution of river terraces port and retention of coarse woody debris in moun- with integrated OSL and cosmogenic nuclide data. tain streams: an in situ field experiment of log trans- Quaternary Geochronology. doi:10.1016/j.quageo. port and a field survey of coarse woody debris distri- 2010.06.002 bution. Water Resources Research 38, 1126. doi:10.1029/ Gurnell, A.M. 1995. Sediment yield from alpine glacier 2001WR001123 basins.In:I.Foster,A.Gurnell,B.Webb,eds.,Sediment Hales, R.C., J.J. Roering. 2007. Climatic controls on and Water Quality in River Catchments.JohnWiley& frost cracking and implications for the evolution of Sons, Chichester, pp. 407–435. bedrock landscapes. Journal of Geophysical Research Gurnell, A.M. 2013. Wood in fluvial systems. In: E. 112, F02033. doi:10.1029/2006JF000616, 14 p. Wohl, ed., Treatise on Fluvial Geomorphology. Elsevier, Hales, T.C., J.J. Roering. 2009. A frost “buzzsaw” mech- Amsterdam, pp. 163–188. anism for erosion of the eastern Southern Alps, New Gurnell, A.M., W. Bertoldi, D. Corenblit. 2012. Chang- Zealand. Geomorphology 107, 241–253. ing river channels: the roles of hydrological processes, Hall, R.O., J.L. Tank. 2003. Ecosystem metabolism con- plants and pioneer fluvial landforms in humid temper- trolsnitrogenuptakeinstreamsinGrandTeton ate,mixedload,gravelbedrivers.Earth-Science Reviews National Park, Wyoming. Limnology and Oceanography 111, 129–141. 48, 1120–1128. 274 References

Ham, D.G. 2005. Morphodynamics and Sediment Trans- Harmon, R.S. 2005b. The geological development of port in a Wandering Gravel-Bed Channel: Fraser River, Panama. In: R.S. Harmon, ed., The Rio Chagres, British Columbia.PhDdissertation,Universityof Panama: A Multidisciplinary Profile of a Tropical British Columbia, Vancouver. Watershed. Springer, Dordrecht, The Netherlands, Hamilton, L.S., L.A. Bruijnzeel. 1997. Mountain water- pp. 45–62. sheds – integrating water, soils, gravity, vegetation, and Harris, J.H., R. Silveira. 1999. Large-scale assessments of people. In: B. Messerli, J.D. Ives, eds., Mountains of river health using an Index of Biotic Integrity with low- the World: A Global Priority. The Parthenon, London, diversity fish communities. Freshwater Biology 41, 235– pp. 337–370. 252. Hancock, G.S., R.S. Anderson. 2002. Numerical model- Hart, D.D., T.E. Johnson, K.L. Bushaw-Newton, R.J. Hor- ing of fluvial strath-terrace formation in response to witz, A.T. Bednarek, D.F. Charles, D.A. Kreeger, D.J. oscillating climate. Geological Society of America Bul- Velinsky. 2002. Dam removal: challenges and opportu- letin 114, 1131–1142. nities for ecological research and river restoration. Bio- Hancock, G.S., R.S. Anderson, K.X. Whipple. 1998. Science 52, 669–681. Beyond power: bedrock river incision process and Harvey, A.M. 1997. [Coupling between hillslope gully sys- form. In: K.J. Tinkler, E.E. Wohl, eds., Rivers Over temsandstreamchannelsintheHowgillFells,north- Rock: Fluvial Processes in Bedrock Channels.Amer- west England: temporal implications]. Geomorphologie: ican Geophysical Union Press, Washington, DC, Relief, Processus, Environnement 3, 3–19. pp. 35–60. Harvey, A.M. 2002. Effective timescales of coupling Hancock, P.J., A.J. Boulton, W.F. Humphreys. 2005. within fluvial systems. Geomorphology 44, 175–201. Aquifers and hyporheic zones: towards an ecological Harvey, A.M. 2007. Differential recovery from the effects understanding of groundwater. Hydrogeology Journal of a 100 year storm: significance of long term hills- 13, 98–111. lope channel coupling: Howgill Fells, northwest Eng- Hancock, G.S., E.E. Small, C. Wobus. 2011. Modelling land. Geomorphology 84, 192–208. the effects of weathering on bedrock-floored channel Harvey, A.M. 2008. Developments in dryland geomor- geometry. Journal of Geophysical Research 116, F03018. phology. In: T.P. Burt, R.J. Chorley, D. Brunsden, N.J. doi:10.1029/2010JF001908, 13 p. Cox, A.S. Goudie, eds., TheHistoryoftheStudyofLand- Harbor, D.J., A. Bacastow, A. Heath, J. Rogers. 2005. forms or the Development of Geomorphology, Volume [Capturing variable knickpoint retreat in the central 4: Quaternary and Recent Processes and Forms (1890– Appalachians, USA]. Geografia Fisica e Dinammica 1965) and The Mid-Century Revolutions.TheGeologi- Quaternaria. 28, 23–36. cal Society, London, pp. 729–765. Harden, D.R. 1990. Controlling factors in the distribution Harvey, J.W., C.C. Fuller. 1998. Effect of enhanced man- and development of incised meanders in the central ganese oxidation in the hyporheic zone on basin-scale Colorado Plateau. Geological Society of America Bulletin geochemical mass balance. Water Resources Research 102, 233–242. 34, 623–636. Harden, C.P. 2006. Human impacts on headwater fluvial Harvey, M.D., R.A. Mussetter, E.J. Wick. 1993. A physical systems in the northern and central Andes. Geomor- process-biological response model for spawning habi- phology 79, 249–263. tat formation for the endangered Colorado squawfish. Harden, C.P.,P.D. Scruggs. 2003. Infiltration on mountain Rivers 4, 114–131. slopes: a comparison of three environments. Geomor- Haschenburger, J.K. 2006. Observations of event-based phology 55, 5–24. streambed deformation in a gravel bed channel. Harding, J.S., E.F. Benfield, P.V. Bolstad, G.S. Helfman, Water Resources Research 42, W11412. doi:10.1029/ E.B.D. Jones. 1998. Stream biodiversity: the ghost of 2006WR004985 land use past. Proceedings of the National Academy of Haschenburger, J.K. 2013. Bedload kinematics and fluxes. Sciences USA 95, 14843–14847. In:E.Wohl,ed.,Treatise on Fluvial Geomorphology. Harmon, R.S. 2005a. An introduction to the Panama Elsevier, Amsterdam, pp. 104–123. Canal watershed. In: R.S. Harmon, ed., The Rio Chagres, Hasnain, S.I. 2002. Himalayan glaciers meltdown: impact Panama: A multidisciplinary profile of a tropical water- on South Asian rivers. In: H.A.J. van Lanen,S. Demuth, shed. Springer, Dordrecht, The Netherlands, pp. 19–28. eds., FRIEND 2002: Regional Hydrology; Bridging the References 275

Gap Between Research and Practice.IAHSPublication ment of Environmental Change. Geological Survey of no. 274, pp. 417–423. Canada, Bulletin 547, pp. 31–39. Hassan, M.A., M. Church, P.J. Ashworth. 1992. Virtual Heimsath, A.M., W.E. Dietrich, K. Nishiizumi, R.C. rateandmeandistanceoftravelofindividualclastsin Finkel. 1997. The soil production function and land- gravel-bed channels. Earth Surface Processes and Land- scape equilibrium. Nature 388, 358–361. forms 17, 617–627. Heimsath, A.M., W.E. Dietrich, K, Nishiizumi, R.C. Hassan, M.A., R. Egozi, G. Parker. 2006. Experiments on Finkel. 2001. Stochastic processes of soil production the effect of hydrograph characteristics on vertical grain and transport: erosion rates, topographic variation and sortingingravelbedrivers.Water Resources Research cosmogenic nuclides in the Oregon Coast Range. Earth 42, W09408. doi:10.1029/2005WR004707 Surface Processes and Landforms 26, 531–552. Hassan, M.A., A.S. Gottesfeld, D.R. Montgomery, J.F. Hendrickx, J.M.H., D. Vega, J.B.J. Harrison, L.E. Calvo- Tunnicliffe, G.K.C. Clarke, G. Wynn, H. Jones-Cox, Gobbeti, P.Rojas, T.W.Miller. 2005. Hydrology of hills- R. Poirier, E. MacIsaac, H. Herunter, S.J. Macdonald. lope soils in the Upper Rio Chagres watershed, Panama. 2008. Salmon-driven bed load transport and bed mor- In: R.S. Harmon, ed., The Rio Chagres, Panama: A Mul- phologyinmountainstreams.Geophysical Research tidisciplinary Profile of a Tropical Watershed.Springer, Letters 35, L04405. doi:10.1029/2007GL032997, 6 p. Dordrecht, The Netherlands, pp. 113–138. Hassan, M.A., D.L. Hogan, S.A. Bird, C.L. May, T. Gomi, Henkle, J.E., E. Wohl, N. Beckman. 2011. Locations of D. Campbell. 2005. Spatial and temporal dynamics of channel heads in the semiarid Colorado Front Range, wood in headwater streams of the Pacific Northwest. USA. Geomorphology 129, 309–319. Journal of the American Water Resources Association 41, Heritage, G.L., M.E. Charlton, S. O’Regan. 2001. Morpho- 899–919. logical classification of fluvial environments: an inves- Hassan, M.A., E.L. Petticrew, D.R. Montgomery, A.S. tigation of the continuum of channel types. Journal of Gottesfeld, J.F. Fex. 2011. Salmon as biogeomorphic Geology 109, 21–33. agents in gravel bed rivers: the effect of fish on sed- Herron, N., C. Wilson. 2001. A water balance approach to imentmobilityandspawninghabitat.In:A.Simon, assessing the hydrologic buffering potential of an allu- S.J. Bennett, J.M. Castro, eds., Stream Restoration in vial fan. Water Resources Research 37, 341–351. Dynamic Fluvial Systems: Scientific Approaches, Anal- Hester, E.T., M.W. Doyle. 2008. In-stream geomor- yses, and Tools. American Geophysical Union Press, phic structures as drivers of hyporheic exchange. Washington, DC, pp. 337–352. Water Resources Research 44, W03417. doi:10.1029/ Hassan, M.A., I. Reid. 1990. The influence of microform 2006WR005810, 17 p. bed roughness elements on flow and sediment trans- Hester, E.T., M.N. Gooseff. 2010. Moving beyond port in gravel bed rivers. Earth Surface Processes and the banks: hyporheic restoration is fundamental to Landforms 15, 739–750. restoring ecological services and functions of streams. Hassan, M.A., R.D. Woodsmith. 2004. Bed load transport Environmental Science and Technology 44, 1521– in an obstruction-formed pool in a forest, gravelbed 1525. stream. Geomorphology 58, 203–221. Hettler, T., G. Irion, B. Lehmann. 1997. [Environmen- Hauer,F.R.,J.S.Baron,D.H.Campbell,K.D.Fausch, tal impact of mining waste disposal on a tropical low- S.W.Hostetler, G.H. Leavesley, P.R. Leavitt, D.M. McK- land river system; a case study on the Ok Tedi Mine, night, J.A. Stanford. 1997. Assessment of climate change Papua New Guinea]. Mineralium Deposita 32, 280– and freshwater ecosystems of the Rocky Mountains, 291. USA and Canada. Hydrological Processes 11, 903– Heuer, K., K.A. Tonnessen, G.P. Ingersoll. 2000. Com- 924. parison of precipitation chemistry in the central Rocky Heezen, B.C., D.B. Ericson, M. Ewing. 1954. Further evi- Mountains, Colorado, USA. Atmospheric Environment dence for a turbidity current following the 1929 Grand 34, 1713–1722. Banks earthquake. Deep Sea Research 1, 193–202. Hewitt, K. 1998. Catastrophic landslides and their effects Heginbottom, J.A. 2000. Permafrost distribution and on the Upper Indus streams, Karakoram Himalaya, ground ice in surficial materials. In: L.D. Dyke, G.R. northern Pakistan. Geomorphology 26, 47–80. Brooks, eds., The Physical Environment of the Mackenzie Hewlett, J.D., A.R. Hibbert. 1967. Factors affecting the Valley, Northwest Territories: A Base Line for the Assess- response of small watersheds to precipitation in humid 276 References

areas.In:W.E.Sopper,H.W.Lull,eds.,Forest Hydrology, Hirschboeck, K.K. 1987. Hydroclimatically-defined Pergamon Press, New York, pp. 275–290. mixed distributions in partial duration flood series. Hickin, E.J. 1984. Vegetation and river channel dynamics. In:V.P.Singh,ed.,Hydrologic Frequency Modeling.D. Canadian Geographer 28, 111–126. Reidel, pp. 199–212. Hickin, E.J., G.C. Nanson. 1984. Lateral migration rates Hirschboeck, K.K. 1988. Flood hydroclimatology. In: V.R. of river bends. Journal of Hydraulic Engineering 110, Baker, R.C. Kochel, P.C. Patton, eds., Flood Geomor- 1557–1567. phology. John Wiley & Sons, New York, pp. 27–49. Hicks, F., S. Beltaos. 2008. River ice. In: M. Woo, ed., Cold Hjulstrom,¨ F. 1935. Studies of the morphological activ- Region Atmospheric and Hydrologic Studies.Springer- ity of rivers as illustrated by the River Fyris. Bulletin Verlag, Berlin, pp. 281–305. Mineral. Geol. Inst. University of Uppsala 25, 221– Hicks,D.M.,M.J.Duncan,S.N.Lane,M.Tal,R.Westaway. 258. 2008. Contemporary morphological change in braided Hobley, D.E.J., H.D. Sinclair, P.A. Cowie. 2010. Pro- gravel-bed rivers: new developments from field and lab- cesses, rates, and time scales of fluvial response in an oratory studies, with particular reference to the influ- ancient postglacial landscape of the northwest Indian ence of riparian vegetation. In: H. Habersack, H. Piegay,´ Himalaya. Geological Society of America Bulletin 122, M. Rinaldi, eds., Gravel-bed Rivers VI: From Process 1569–1584. Understanding to River Restoration. Elsevier, Amster- Hodge, R.A., T.B. Hoey, L.S. Sklar. 2011. Bed load dam, pp. 557–586. transport in bedrock rivers: the role of sediment Hildebrandt,A.,M.A.AlAufi,M.Amerjeed,M.Sham- cover in grain entrainment, translation, and depo- mas, E.A.B. Eltahir. 2007. Ecohydrology of a sea- sition. Journal of Geophysical Research 116, F04028. sonal cloud forest in Dhofar: 1. Field experiment. doi:10.1029/2011JF002032 Water Resources Research 43, W10411. doi:10.1029/ Hodgkins, R., R. Cooper, J. Wadham, M. Tranter. 2009. 2006WR005261 The hydrology of the proglacial zone of a high-Arctic Hilderbrand, R.H., A.C. Watts, A.M. Randle. 2005. The glacier (Finsterwalderbreen, Svalbard): atmospheric myths of restoration ecology. Ecology and Society 10(1), and surface water fluxes. Journal of Hydrology 378, 150– 19, http://www.ecologyandsociety.org/vol10/iss1/art19 160. Hilmes, M.M., E.E. Wohl. 1995. Changes in channel mor- Hoffman, P.F., J.P. Grotzinger. 1993. Orographic precipi- phology associated with placer mining. Physical Geog- tation, erosional unloading, and tectonic style. Geology raphy 16, 223–242. 21, 195–198. Hilton, R.G., A. Galy, N. Hovius. 2008a. Riverine par- Hohensinner,S.,H.Habersack,M.Jungwirth,G.Zauner. ticulate organic carbon from an active mountain belt: 2004. Reconstruction of the characteristics of a natu- importance of landslides. Global Biogeochemical Cycles ral alluvial river-floodplain system and hydromorpho- 22, GB1017. doi:10.1029/2006GB002905, 12 p. logical changes following human modifications: the Hilton,R.G.,A.Galy,N.Hovius,M.-C.Chen,M.-J. Danube River (1812–1991). River Research Applications Horng, H. Chen. 2008b. Tropical-cyclone-driven ero- 20, 25–41. sion of the terrestrial biosphere from mountains. Holbrook, J., G. Kliem, C. Nzewunwah, Z. Jobe, R. Gobie. Nature Geoscience 1, 759–762. 2006. Surficial alluvium and topography of the Over- Hilton,R.G.,A.Galy,N.Hovius,M.-J.Horng,H.Chen. ton Bottoms North Unit, Big Muddy National Fish and 2011a. Efficient transport of fossil organic carbon to Wildlife Refuge in the Missouri River Valley and its the ocean by steep mountain rivers: an orogenic carbon potential influence on environmental management. In: sequestration mechanism. Geology 39, 71–74. R.B. Jacobson ed., Science to Support Adaptive Man- Hilton, R.G., P. Meunier, N. Hovius, P.J. Bellingham, agement – Overton Bottoms North Unit, Big Muddy A. Galy. 2011b. Landslide impact on organic carbon National Fish and Wildlife Refuge. U.S. Geological Sur- cycling in a temperate montane forest. Earth Surface vey Scientific Investigations Report 2006–5086, Mis- Processes and Landforms 36, 1670–1679. souri, pp. 17–31. Hirsch, R.M., J.F. Walker, J.C. Day, R. Kallio. 1990. The Holmes,R.M.,J.W.McClelland,B.J.Peterson,I.A. influence of man on hydrologic systems. In: M.G. Wol- Shiklomanov, A.I. Shiklomanov, A.V. Zhulidov, V.V. man, H.C. Riggs, eds., Surface Water Hydrology.Geo- Gordeev, N.N. Bobrovitskaya. 2002. A circumpolar logical Society of America, Boulder, CO, pp. 329–359. perspective on fluvial sediment flux to the Arctic References 277

Ocean. Global Biogeochemical Cycles 16, 1098. pre-steady-state mountain belt: Finisterre Mountains, doi:1010.1029/2001GB001849 Papua New Guinea. Geology 26, 1071–1074. Holmes, R.M., J.W. McClelland, P.A. Raymond, B.B. Howard, A.D. 1980. Thresholds in river regimes. In: D.R. Frazer, B.J. Peterson, M. Stieglitz. 2008. Lability of DOC Coates,J.D.Vitek,eds.,Thresholds in Geomorphology. transported by Alaskan rivers to the Arctic Ocean. Geo- George Allen and Unwin, London, pp. 227–258. physical Research Letters 35, L03402. Howard, A.D. 1987. Modelling fluvial systems: rock-, Hood, W.G. 2010. Tidal channel meander formation by gravel- and sand-bed channels. In: K. Richards, ed., depositional rather than erosional processes: examples River Channels: Environment and Process.Blackwell, from the prograding Skagit River delta (Washington, New York, pp. 69–94. USA). Earth Surface Processes and Landforms 35, 319– Howard, A.D. 1994. A detachment-limited model of 330. drainage basin evolution. Water Resources Research 30, Hooke, J. 2003. Coarse sediment connectivity in river 2261–2285. channel systems: a conceptual framework and method- Howard, A.D. 1998. Long profile development of bedrock ology. Geomorphology 56, 79–94. channels: interaction of weathering, , bed Hooke, J.M. 2013. Meandering rivers. In: E. Wohl, ed., erosion, and sediment transport. In: K.J. Tinkler, E.E. Treatise on Fluvial Geomorphology. Elsevier, Amster- Wohl, eds., Rivers Over Rock: Fluvial Processes in dam, pp. 260–287. Bedrock Channels. American Geophysical Union Press, Hooke, R.L.B. 2000. On the history of humans as geomor- Washington, DC, pp. 297–319. phic agents. Geology 28, 843–846. Howard, A.D., W.E.Dietrich, M.A. Seidl. 1994. Modelling Hooke, R.L.B., W.L. Rohrer. 1979. Geometry of alluvial fluvial erosion on regional to continental scales. Journal fans:effectofdischargeandsedimentsize.Earth Sur- of Geophysical Research 99(B7), 13971–13986. face Processes 4, 147–166. Huang, H.Q., G.C. Nanson. 2000. Hydraulic geometry Hopp, L., J.J. McDonnell. 2009. Connectivity at the hills- and maximum flow efficiency as products of the prin- lope scale: identifying interactions between storm size, ciple of least action. Earth Surface Processes and Land- bedrock permeability, slope angle and soil depth. Jour- forms 25, 1–16. nal of Hydrology 376, 378–391. Huang, H.Q., G.C. Nanson. 2007. Why some alluvial Hori, K., Y. Saito. 2007. An early Holocene sea-level jump rivers develop an anabranching pattern. Water and delta initiation. Geophysical Research Letters 34. Resources Research 43, W07441. doi:10.1029/ doi:10.1029/2007GL031029 2006WR005223 Hori,K.,S.Tanabe,Y.Saito,S.Haruyama,V.Nguyen,A. Hughes, F.M.R., W.M. Adams, E. Muller, C. Nilsson, K.S. Kitamura. 2004. Delta initiation and Holocene sea-level Richards, N. Barsoum, Henri Decamps, R. Foussadier, change; example from the Song Hong (Red River) delta, Jacky Girel, Hel´ eneGuilloy,A.Hayes,M.Johansson,` Vietnam. Sedimentary Geology 164, 237–249. LucLambs,G.Pautou,J.-L.Peiry,M.Perrow,F.Vautier, Hornung, J., D. Pflanz, A. Hechler, A. Beer, M. Hinderer, M. Winfield. 2001. The importance of different scale M. Maisch, U. Bieg. 2010. 3-D architecture, deposi- processes for the restoration of floodplain woodlands. tional patterns and climate triggered sediment fluxes of Regulated Rivers: Research and Management 17, 325– an alpine alluvial fan (Samedan, Switzerland). Geomor- 345. phology 114, 202–214. HughesF.M.R., A. Colston, J.O. Mountford. 2005. Horton, R.E. 1932. Drainage-basin characteristics. Trans- Restoring Riparian Ecosystems: The Challenge of actions of the American Geophysical Union 13, 350–361. Accommodating Variability and Designing Restoration Horton, R.E. 1945. Erosional development of streams and Trajectories. Ecology and Society 10, http://www. their drainage basins: hydrophysical approach to quan- ecologyandsociety.org/vol10/iss1/art12 titative morphology. Geological Society of America Bul- Hultine, K.R., S.E. Bush. 2011. Ecohydrological conse- letin 56, 275–370. quences of non-native riparian vegetation in the south- Houjou, K., Y. Shimizu, C. Ishii. 1990. Calculation of westernUnitedStates:areviewfromanecophysiologi- boundary shear stress in open channel flow. Journal of cal perspective. Water Resources Research 47, W07542. Hydroscience and Hydraulic Engineering 8, 21–37. doi:10.1029/2010WR010317 Hovius, N., C.P. Stark, M.A. Tutton, L.D. Abbott. 1998. Humborg,C.,V.Ittekkot,A.Cociasu,B.V.Bodungen. Landslide-driven drainage network evolution in a 1997. Effect of Danube River dam on Black Sea 278 References

biogeochemistry and ecosystem structure. Nature 386, Jaeger, K.L., J.D. Olden. 2012. Electrical resistance sensor 385–388. arrays as a means to quantify longitudinal connectiv- Hupp, C.R., G. Bornette. 2003. Vegetation as a tool in ity of rivers. River Research and Applications 28, 1843– the interpretation of fluvial geomorphic processes and 1852. landforms in humid temperate areas. In: G.M. Kondolf, James, L.A. 1997. Channel incision on the lower American H. Piegay,´ eds., Tools in Geomorphology.JohnWiley& River, California, from streamflow gage records. Water Sons, Chichester, pp. 269–288. Resources Research 33, 485–490. Ijjasz-V´ asquez,´ E.I., R.L. Bras. 1995. Scaling regimes of James, L.A., S.A. Lecce. 2013. Impacts of land-use and local slope versus contributing area in digital elevation land-cover change on river systems. In: E. Wohl, ed., models. Geomorphology 12, 299–311. Treatise on Fluvial Geomorphology. Elsevier, Amster- Inman, D.L., C.E. Nordstrom. 1971. On the tectonic and dam, pp. 768–793. morphologic classification of coasts. Journal of Geology James, L.A., W.A. Marcus. 2006. The human role in chang- 79, 1–21. ing fluvial systems: retrospect, inventory and prospect. IPCC (Intergovernmental Panel on Climate Change). Geomorphology 79, 152–171. 2008. Climate Change 2007. IPCC Fourth Assessment Jansen, J.D. 2006. Flood magnitude-frequency and litho- Report. Cambridge University Press, Cambridge. logic control on bedrock river incision in post-orogenic Istanbulluoglu, E., D.G. Tarboton, R.T. Pack, C. Luce. terrain. Geomorphology 82, 39–57. 2002. A probabilistic approach for channel initi- Jansen,J.D.,D.Fabel,P.Bishop,S.Xu,C.Schnabel,A.T. ation. Water Resources Research 38. doi:10.1029/ Codilean. 2011. Does decreasing paraglacial sediment 2001WR000782, 14 p. supply slow knickpoint retreat? Geology 39, 543–546. Iverson, R.J. 2005. Debris-flow mechanics. In: M. Jakob, Jansen, J.D., G.C. Nanson. 2004. Anabranching and O. Hungr, eds., Debris-flow Hazards and Related Phe- maximum flow efficiency in Magela Creek, north- nomena. Springer, Berlin, pp. 105–134. ern Australia. Water Resources Research 40, W04503. Jabaloy-Sanchez, A., F.J. Lobo, A. Azor, P. Barcenas, J.M. doi:10.1029/2003WR002408 Fernandez-Salas, V.Diaz del Rio, J.V.Perez-Pena. 2010. Jansson, R., C. Nilsson, M. Dynesius, E. Andersson. 2000. Human-driven coastline changes in the Adra River Effects of river regulation on river-margin vegetation: deltaic system, southeast Spain. Geomorphology 119, 9– a comparison of eight boreal rivers. Ecological Applica- 22. tions 10, 203–224. Jacobson, R.B., K.B.Gran. 1999. Gravel sediment routing Japanese Ministry of Construction. 1993. Sabo.Ministry from widespread, low-intensity landscape disturbance, of Construction, Kobe, 36 p. Current River basin, Missouri. Earth Surface Processes Jaquette, C., E. Wohl, D. Cooper. 2005. Establishing a and Landforms 24, 897–917. context for river rehabilitation, North Fork Gunnison Jacobson, R.B., J.P. McGeehin, E.D. Cron, C.E. Carr, J.M. River, Colorado. Environmental Management 35, 593– Harper, A.D. Howard. 1993. Landslides Triggered by the 606. Storm of November 3–5, 1985, Wills Mountain Anti- Jarrett, R.D. 1987. Errors in slope-area computations cline, West Virginia and Virginia.U.S.GeologicalSurvey ofpeakdischargesinmountainstreams.Journal of Bulletin 1981, Part C, C1–C33. Hydrology 96, 53–67. Jacobson, R.B., K.A. Oberg. 1997. Geomorphic Changes on Jarrett, R.D. 1989. Hydrology and paleohydrology used to the Mississippi River Flood Plain at Miller City, Illinois, improve the understanding of flood hydrometeorology as a Result of the Flood of 1993. U.S. Geological Survey in Colorado. In: M.L. Albertson, R.A. Kia, eds., Design Circular 1120-J, Denver, CO, 22 p. of Hydraulic Structures 89, Proceedings of the Second Jacobson, R.B., J.E. O’Connor, T. Oguchi. 2003. Surfi- International Symposium on Design of Hydraulic Struc- cial geologic tools in fluvial geomorphology. In: G.M. tures. A.A. Balkema, Rotterdam, pp. 9–16. Kondolf, H. Piegay,´ eds., Tools in Geomorphology.John Jarrett, R.D., J.F. England. 2002. Reliability of paleostage Wiley & Sons, Chichester, pp. 25–57. indicators for paleoflood studies. In: P.K. House, R.H. Jaeger, K.L., D.R. Montgomery, S.M. Bolton. 2007. Chan- Webb,V.R.Baker,D.R.Levish,eds.,Ancient Floods, nel and perennial flow initiation in headwater streams: Modern Hazards: Principles and Applications of Pale- management implications of variability in source-area oflood Hydrology. American Geophysical Union Press, size. Environmental Management 40, 775–786. Washington, DC, pp. 91–109. References 279

Jasechko, S., Z.D. Sharp, J.J. Gibson, S.J. Birks, Y. Yi, P.J. Jun, P.,S. Chen, P.Dong. 2010. Temporal variation of sed- Fawcett. 2013. Terrestrial water fluxes dominated by iment load in the Yellow River basin, China, and its transpiration. Nature 496, 347–351. impactsonthelowerreachesandtheriverdelta.Catena Jefferson, A., G.E. Grant, S.L. Lewis, S.T. Lancaster. 83, 136–147. 2010. Coevolution of hydrology and topography on Jungers, M.C., P.R. Bierman, A. Matmon, K. Nichols, abasaltlandscapeintheOregonCascadeRange, J. Larsen, R. Finkel. 2009. Tracing hillslope sediment USA. Earth Surface Processes and Landforms 35, 803– production and transport with in situ and meteoric 816. 10Be. Journal of Geophysical Research 114, F04020. Jeffries, R., S.E. Darby, D.A. Sear. 2003. The influence of doi:10.1029/2008JF001086 vegetation and organic debris on flood-plain sediment Junk, W.J., P.B. Bayley, R.E. Sparks. 1989. The flood pulse dynamics: case study of a low-order stream in the New concept in river-floodplain systems. Canadian Special Forest, England. Geomorphology 51, 61–80. Publication of Fisheries and Aquatic Sciences 106, 110– Jerolmack, D.J., D. Mohrig. 2007. Conditions for branch- 127. ing in depositional rivers. Geology 35, 463–466. Kaiser, K., G. Guggenberger, L. Haumeier. 2004. Changes Jimenez Sanchez, M. 2002. Slope deposits in the upper in dissolved lignin-derived phenols, neutral sugars, NalonRiverbasin(NWSpain):anapproachtoaquan- uronic acids, and amino sugars with depth in forested titative comparison. Geomorphology 43, 165–178. Haplic Arenosols and Rendzic Leptosols. Biogeochem- John, S., A. Klein. 2004. Hydrogeomorphic effects of istry 70, 135–151. beaver dams on floodplain morphology: avulsion pro- Kale, V.S.,V.R. Baker, S. Mishra. 1996. Multi-channel pat- cesses and sediment fluxes in upland valley floors terns of bedrock rivers: an example from the central (Spessart, Germany). Quaternaire 15, 219–231. Narmada basin, India. Catena 26, 85–98. Johnson, A.M., J.R. Rodine. 1984. Debris flows. In: D. Kale, V.S., P.S. Hire. 2007. Temporal variations in the spe- Brunsden,D.B.Prior,eds.,Slope Instability.JohnWiley cific stream power and total energy expenditure of a & Sons, New York, pp. 257–361. monsoonal river: the Tapi River, India. Geomorphology Jones, J.A.A. 1981. The Nature of Soil Piping: A Review of 92, 134–146. Research.Geobooks,Norwich. Kampf, S.K., B.B. Mirus. 2013. Subsurface and surface Jones, J.A.A. 2010. Soil piping and catchment response. flow leading to channel initiation. In: E. Wohl, ed., Trea- Hydrological Processes 24, 1548–1566. tise on Fluvial Geomorphology. Elsevier, Amsterdam, Jones, A.F., P.A. Brewer, E. Johnstone, M.G. Macklin. pp. 23–42. 2007. High-resolution interpretative geomorphological Kao, S.J., J.D. Milliman. 2008. Water and sediment mapping of river valley environments using airborne discharge from small mountainous rivers, Tai- LiDAR data. Earth Surface Processes and Landforms 32, wan: the roles of lithology, episodic events, and 1574–1592. human activities. Journal of Geology 116, 431– Jones, W.K.,D.C. Culver, J.S. Herman. 2003. Epikarst.Pro- 448. ceedings of the 2003 Epikarst Symposium. Special Pub- Karr, J.R. 1991. Biological integrity: a long-neglected lication 9, Karst Waters Institute, Inc, Charles Town, aspect of water resource management. Ecological Appli- WV, 160 p. cations 1, 66–84. Jones, J.A., F.J. Swanson, B.C. Wemple, K.U. Snyder. 2000. Karr, J.R. 1999. Defining and measuring river health. Effects of roads on hydrology, geomorphology, and Freshwater Biology 41, 221–234. disturbance patches in stream networks. Conservation Kasran, B. 1988. Effect of logging on sediment yield in a Biology 14, 76–86. hill dipterocarp forest in Peninsular Malaysia. Journal Jorde,K.,M.Burke,N.Scheidt,C.Welcker,S.King,C. of Tropical Forest Science 1, 56–66. Borden. 2008. Reservoir operations, physical processes, Kattenberg,A.,F.Giorgi,H.Grassl,G.A.Meehl,J.F.B. and ecosystem losses. In: H. Habersack, H. Piegay,´ Mitchell, R.J. Stouffer, T. Tokioka, A.J. Weaver, T.M.L. M. Rinaldi, eds., Gravel-bed Rivers VI: From Process Wigley. 1996. Climate models – projections of future Understanding to River Restoration. Elsevier, Amster- climate. In: J.T. Houghton, L.G. Meiro Filho, B.A. dam, pp. 607–636. Callander, eds., Climate Change 1995 – the Science of Julien, P.Y. 1998. Erosion and Sedimentation.Cambridge Climate Change, Cambridge University Press, Cam- University Press, Cambridge, 280 p. bridge, pp. 285–357. 280 References

Katz, G.L., J.M. Friedman, S.W. Beatty. 2005. Delayed Kirchner, J.W. 1993. Statistical inevitability of Horton’s effects of flood control on a flood-dependent riparian laws and the apparent randomness of stream channel forest. Ecological Applications 15, 1019–1035. networks. Geology 21, 591–594. Kean, J.W., J.D. Smith. 2006a. Form drag in rivers due Kirchner, J. 2003. A double paradox in catchment hydrol- to small-scale natural topographic features: 1. Regu- ogy and geochemistry. Hydrological Processes 17, 871– lar sequences. Water Resources Research 111, F04009. 874. doi:10.1029/2006JF000467 Kirkby, M.J. 1967. Measurement and theory of soil creep. Kean, J.W., J.D. Smith. 2006b. Form drag in rivers due Journal of Geology 75, 359–378. to small-scale natural topographic features: 2. Irregu- Kirkby, M. 1988. Hillslope runoff processes and models. lar sequences. Water Resources Research 111, F04010. Journal of Hydrology 100, 315–339. doi:10.1029/2006JF000490 Klimek, K. 1987. [Man’s impact on fluvial processes in Keller, E.A. 1971. Areal sorting of bedload material, the the Polish western Carpathians]. Geografiska Annaler hypothesis of velocity reversal. Geological Society of 69A(1), 221–226. America Bulletin 82, 279–280. Knighton, A.D. 1974. Variation in width-discharge rela- Keller, E.A., W.N. Melhorn. 1978. Rhythmic spacing and tion and some implications for hydraulic geometry. origin of pools and riffles. Geological Society of America Geological Society of America Bulletin 85, 1069–1076. Bulletin 89, 723–730. Knighton, D. 1984. Fluvial Forms and Processes.Edward Keller, T., C. Pielmeier, C. Rixen, F. Gadient, D. Gustafs- Arnold, London, 218 p. son, M. Stahli. 2004. Impact of artificial snow and ski- Knighton, D. 1989. River adjustment to changes in sed- slope grooming on snowpack properties and soil ther- iment load: the effects of tin mining on the Ringa- mal regime in a sub-alpine ski area. Annals of Glaciology rooma River, Tasmania, 1875–1984. Earth Surface Pro- 38, 314–318. cesses and Landforms 14, 333–359. Kellerhals, R., M. Church. 1990. Hazard management Knighton, D. 1998. Fluvial Forms and Processes: A New on fans, with examples from British Columbia. Perspective.Arnold,London. In: A.H. Rachocki, M. Church, eds., Alluvial Fans: Knighton, A.D. 1999. Downstream variation in stream AfieldApproach. John Wiley & Sons, pp. 335– power. Geomorphology 29, 293–306. 354. Knighton, A.D., G.C. Nanson. 2002. Inbank and overbank Kellerhals, R., M. Church, D.I. Bray. 1976. Classification velocity conditions in an arid zone anastomosing river. of river processes. Journal of the Hydraulics Division, Hydrological Processes 16, 1771–1791. ASCE 102, 813–829. Kobor, J.S., J.J. Roering. 2004. Systematic variation of Kellerhals, R., M. Church, L.B. Davies. 1979. Morphologi- bedrock channel gradients in the central Oregon Coast cal effects of interbasin river diversions. Canadian Jour- Range: implications for rock uplift and shallow lands- nal of Civil Engineering 6, 18–31. liding. Geomorphology 62, 239–256. Kelly, S. 2006. Scaling and hierarchy in braided rivers and Kochel, R.C. 1988. Geomorphic impact of large floods: their deposits: examples and implications for reservoir review and new perspectives on magnitude and fre- modeling.In:G.H.SambrookSmith,J.L.Best,C.S.Bris- quency. In: V.R. Baker, R.C. Kochel,P.C. Patton, eds., tow, G.E. Petts, eds., Braided Rivers: Process, Deposits, Flood Geomorphology. John Wiley & Sons, New York, Ecology and Management.Blackwell,Oxford,pp.75– pp. 169–187. 106. Kondolf, G.M. 1997. Hungry water: effects of dams and Kesel, R.H. 2003. Human modifications to the sediment gravel mining on river channels. Environmental Man- regime of the Lower Mississippi River flood plain. Geo- agement 21, 533–551. morphology 56, 325–334. Kondolf, G.M. 2001. Planning approaches to mitigate Kieffer, S.W. 1989. Geologic nozzles. Reviews of Geo- adversehumanimpactsonland/waterecotones.Inter- physics 27, 3–38. national Journal of Ecohydrology and Hydrobiology 1, Kim, H.J., R.C. Sidle, R.D. Moore, R. Hudson. 2004. 111–116. Throughflow variability during snowmelt in Kondolf, G.M., A.J. Boulton, S. O’Daniel, G.C. Poole, a forested mountain catchment, coastal British F.J. Rahel, E.H. Stanley, E. Wohl, A. Bang,˚ J. Carl- Columbia, Canada. Hydrological Processes 18, 1219– strom,C.Cristoni,H.Huber,S.Koljonen,P.Louhi, 1236. K. Nakamura. 2006. Process-based ecological river References 281

restoration: visualizing three-dimensional connectiv- Kovacs, A.G. Parker. 1994. A new vectorial bedload for- ity and dynamic vectors to recover lost linkages. Ecol- mulation and its application to the time evolution of ogy and Society 11(2): 5, http:www.ecologyandsociety straight river channels. Journal of Fluid Mechanics 267, .org/vol11/iss2/art5 153–183. Kondolf, G.M., R.R. Curry. 1986. Channel erosion Kramer, P.J., Boyer, J.S., 1995. Water Relations of Plants along the Carmel River, Monterey County, Califor- and Soils.AcademicPress,SanDiego,CA,495p. nia. EarthSurfaceProcessesandLandforms11, 307– Kramer, N., E.E. Wohl, D.L. Harry. 2012. Using ground 319. penetrating radar to ‘unearth’ buried beaver dams. Kondolf, G.M., M. Larson. 1995. Historical channel anal- Geology 40, 43–46. ysisanditsapplicationtoriparianandaquatichabitat Krigstrom, A. 1962. [Geomorphological studies of sandur restoration. Aquatic Conservation: Marine and Freshwa- plains and their braided rivers in Iceland]. Geografiska ter Ecosystems 5, 109–126. Annaler 44, 328–346. Kondolf, G.M., M.W. Smeltzer, S.F. Railsback. 2001. Krom, M.D., J.D. Stanley, R.A. Cliff, J.C. Woodward. 2002. Design and performance of a channel reconstruction Nile River sediment fluctuations over the past 7000 yr project in a coastal California gravel-bed stream. Envi- and their key role in sapropel development. Geology 30, ronmental Management 28, 761–776. 71–74. Konrad, C.P. 2012. Reoccupation of floodplains by rivers Kuhnle, R.A. 2013. Suspended load. In: E. Wohl, ed., Trea- and its relation to age structure of floodplain vege- tise on Fluvial Geomorphology. Elsevier, Amsterdam, tation. Journal of Geophysical Research 117, F00N13. pp. 124–136. doi:10.1029/2011GJ001906 Kuo, C.-W.,G.J. Brierley. 2013. The influence of landscape Konrad, C.P., J.D. Olden, D.A. Lytle, T.S. Melis, J.C. configuration upon patterns of sediment storage in a Schmidt, E.N. Bray, M.C. Freeman, K.B. Gido, N.P. highly connected river system. Geomorphology 180– Hemphill, M.J. Kennard, L.E. McMullen, M.C. Mims, 181, 255–266. M. Pyron, C.T. Robinson, J.G. Williams. 2011. Large- Lacey, R.W.J., P. Legendre, A.G. Roy. 2007. Spatial-scale scale flow experiments for managing river systems. Bio- partitioning of in situ turbulent flow data over a pebble Science 61, 948–959. cluster in a gravel-bed river. Water Resources Research Koons, P.O., P.K. Zeitler, C.P. Chamberlain, D. Craw, 43, W03416. doi:10.1029/2006WR005044 A.S. Meltzer. 2002. Mechanical links between ero- Laenen, A., J.C. Risley. 1997. Precipitation-runoff and sion and metamorphism in Nanga Parbat, Pakistan Streamflow-routing Models for the Willamette River Himalaya. The American Journal of Science 302, 749– Basin, Oregon.U.S.GeologicalSurveyWaterResources 773. Investigations Report 95–4284, 197 p. Korup, O. 2013. Landslides in the fluvial system. In: Lagasse, P.F., B.R. Winkley, D.B. Simons. 1980. Impact E. Wohl, ed., Treatise on Fluvial Geomorphology.Else- of gravel mining on river system stability. Journal of vier, Amsterdam, pp. 244–259. Waterway, Port, Coastal and Ocean Division, American Korup, O., A.L. Densmore, F. Schlunegger. 2010. The role Society of Civil Engineering 106, 389–404. of landslides in mountain range evolution. Geomor- Lai, C., Katul, G., 2000. The dynamic role of root-water phology 120, 77–90. uptake in coupling potential to actual transpiration. Korup, O., M.J. McSaveney, T.R.H. Davies. 2004. Sedi- Advances in Water Resources 23, 427–439. ment generation and delivery from large historic land- Lamb, M.P., A.D. Howard, W.E. Dietrich, J.T. slides in the Southern Alps, New Zealand. Geomorphol- Perron. 2007. Formation of ampitheater-headed val- ogy 61, 189–207. leys by waterfall erosion after large-scale slumping on Korup, O., A.L. Strom, J.T. Weidinger. 2006. Fluvial Hawaii. Geological Society of America Bulletin 119, 805– response to large rock-slope failures; examples from the 822. Himalayas,theTienShan,andtheSouthernAlpsin Lamb,M.P.,A.D.Howard,J.Johnson,K.X.Whipple,W.E. New Zealand. Geomorphology 78, 3–21. Dietrich, J.T. Perron. 2006. Can springs cut canyons Koster, E.H. 1978. Transverse ribs: their characteris- into rock? Journal of Geophysical Research 111, E07002. tics, origin and paleohydraulic significance. In: A.D. doi:10.1029/2005JE002663 Miall, ed., Fluvial Sedimentology. Canadian Society of Lancaster, S.T., N.E. Casebeer. 2007. Sediment storage Petroleum Geologists Memoir 5, pp. 161–186. and evacuation in headwater valleys at the transition 282 References

between debris-flow and fluvial processes. Geology 35, Lautz, L.K., R.M. Fanelli. 2008. Seasonal biogeochemi- 1027–1030. cal hotspots in the streambed around restoration struc- Lancaster, S.T., G.E. Grant. 2006. Debris dams and the tures. Biogeochemistry 91, 85–104. relief of headwater streams. Geomorphology 82, 84–97. Lautz, L.K., D.I. Siegel. 2007. The effect of transient stor- Lancaster, S.T., S.K. Hayes, G.E. Grant. 2003. Effects age on nitrate uptake lengths in streams: an inter-site of wood on debris flow runout in small moun- comparison. Hydrological Processes 21, 3533–3548. tain watersheds. Water Resources Research 39, 1168. Lawler, D.M. 1992. Process dominance in bank erosion doi:10.1029/2001WR001227 systems. In: P.A. Carling, G.E. Petts, eds., Lowland Lancaster, S.T., E.F. Underwood, W.T. Frueh. 2010. Sed- Floodplain Rivers: Geomorphological Perspectives.John iment reservoirs at mountain stream confluences: Wiley & Sons, Chichester, pp. 117–143. dynamics and effects of tributaries dominated by Lebedeva, M.I., R.C. Fletcher, S.L. Brantley. 2010. A math- debris-flow and fluvial processes. Geological Society of ematical model for steady-state regolith production at America Bulletin 122, 1775–1786. constant erosion rate. EarthSurfaceProcessesandLand- Lane, E.W.1955. The importance of fluvial morphology in forms 35, 508–524. hydraulic engineering. American Society of Civil Engi- Leeder, M.R. 1983. On the interactions between turbu- neers Proceedings 81(745), 1–17. lent flow, sediment transport and bedform mechan- Langbein, W.B. 1964. Geometry of river channels. Journal ics in channelized flows. In: J.D. Collinson, D. Lewin, of Hydraulics Division, ASCE 90, 301–312. eds., Modern and Ancient Fluvial Systems.International Langbein, W.B., L.B. Leopold 1964. Quasi-equilibrium Association of Sedimentologists Special Publication 6, states in channel morphology. American Journal of Sci- pp. 5–18. ence 262, 782–794. Legleiter, C.J., L.R. Harrison, T. Dunne. 2011. Effect of Langbein, W.B., L.B. Leopold. 1966. River Meanders – The- point bar development on the local force balance gov- ory of Minimum Variance. U.S. Geological Survey Pro- erning flow in a simple, meandering gravel bed river. fessional Paper 422H. Journal of Geophysical Research – Earth Surface 116, Langbein, W.B., L.B. Leopold. 1968. River Channel Bars F01005. doi:10.1029/2010JF001838 and Dunes – Theory of Kinematic Waves.U.S.Geological Legleiter, C.J., D.A. Roberts, R.L. Lawrence. 2009. Spec- Survey Professional Paper 422L. trally based remote sensing of river bathymetry. Earth Langbein, W.B., S.A. Schumm. 1958. Yield of sediment Surface Processes and Landforms 34, 1039–1059. in relation to mean annual precipitation. Transac- Leier, A.L., P.G. DeCelles, J.D. Pelletier. 2005. Mountains, tions of the American Geophysical Union 39, 1076– monsoons, and megafans. Geology 33, 289–292. 1084. Lenzi, M.A. 2001. Step-pool evolution in the Rio Cordon, Larsen, M.C., A.S. Roman.´ 2001. Mass wasting and sedi- northeastern Italy. Earth Surface Processes and Land- ment storage in a small montane watershed: an extreme forms 26, 991–1008. case of anthropogenic disturbance in the humid trop- Lenzi, M.A. 2002. Stream bed stabilization using boulder ics. In: J.M. Dorava, F. Fitzpatrick, B.B. Palcsak, D.R. check dams that mimic step-pool morphology features Montgomery, eds., Geomorphic Processes and Riverine in northern Italy. Geomorphology 45, 243–260. Habitat. American Geophysical Union Press, Washing- Lenzi, M.A. 2004. Displacement and transport of marked ton, DC, pp. 119–138. pebbles, cobbles and boulders during floods in a steep Latocha, A., P.Migon.´ 2006. Geomorphology of medium- mountain stream. Hydrological Processes 18, 1899– high mountains under changing human impact, from 1914. managed slopes to nature restoration: a study from the Leonard, E.M. 2002. Geomorphic and tectonic forcing of Sudetes, SW Poland. Earth Surface Processes and Land- late Cenozoic warping of the Colorado Piedmont. Geol- forms 31, 1657–1673. ogy 30, 595–598. Latrubesse, E.M. 2008. Patterns of anabranching chan- Leopold, L.B. 1976. Reversal of erosion cycle and climatic nels: the ultimate end-member adjustment of mega change. Quaternary Research 6, 557–562. rivers. Geomorphology 10, 130–145. Leopold, L.B. 1994. A View of the River. Harvard Univer- Latrubesse, E.M., E. Franzinelli. 2002. The Holocene allu- sity Press, Cambridge, MA. vial plain of the middle Amazon River. Geomorphology Leopold, L.B., R.A. Bagnold, M.G. Wolman, L.M. Brush. 44, 241–257. 1960. Flow Resistance in Sinuous or Irregular Channels. References 283

U.S. Geological Survey Professional Paper 282-D, Liu,J.,Y.Saito,X.Kong,H.Wang,L.Xiang,C.Wen,R. Washington, DC, pp. 111–134. Nakashima. 2010. Sedimentary record of environmen- Leopold, L.B., W.B. Langbein. 1962. The Concept of tal evolution off the Yangtze River estuary, East China Entropy in Landscape Evolution. U.S. Geological Survey Sea, during the last ∼13,000 years, with special refer- Professional Paper 500-A. ence to the influence of the Yellow River on the Yangtze Leopold, L.B., T. Maddock. 1953. The Hydraulic Geome- River delta during the last 600 years. Quaternary Science try of Stream Channels and Some Physiographic Impli- Reviews 29, 2424–2438. cations. U.S. Geological Survey Professional Paper 252, Lopez-Moreno,´ J.I., S. Goyette, M. Beniston. 2009. Impact Washington, DC, 56 p. of climate change on snowpack in the Pyrenees: hori- Leopold, L.B., M.G. Wolman. 1957. River Channel Pat- zontal spatial variability and vertical gradients. Journal terns – Braided, Meandering and Straight.U.S.Geolog- of Hydrology 374, 384–396. ical Survey Professional Paper 282B, pp. 39–85. Lowrance, R., R. Todd, J. Fail, O. Hendrickson, R. Leopold, L.B., M.G. Wolman, J.P.Miller. 1964. Fluvial Pro- Leonard. 1984. Riparian forests as nutrient filters in cesses in Geomorphology. W.H. Freeman, San Francisco, agricultural watersheds. BioScience 34, 374–377. CA. Loye, A., A. Pedrazzini, J.I. Theule, M. Jaboyedoff, F. Liebault,´ F., Clement,´ H. Piegay,´ C.F. Rogers, G.M. Kon- Liebault,´ R. Metzger. 2012. Influence of bedrock struc- dolf, N. Landon. 2002. Contemporary channel changes tures on the spatial pattern of erosional landforms in in the Eygues basin, southern French Prealps: the rela- small alpine catchments. EarthSurfaceProcessesand tionship of subbasin variability to watershed character- Landforms 37, 1407–1423. istics. Geomorphology 45, 53–66. Lu, H., C.J. Moran, M. Sivapalan. 2005. A theoreti- Lienkaemper, G.W.,F.J. Swanson. 1987. Dynamics of large cal exploration of catchment-scale sediment delivery. woody debris in streams in old-growth Douglas-fir Water Resources Research 41, W09415. doi:10.1029/ forests. Canadian Journal of Forest Research 17, 150– 2005WR004018, 15 p. 156. Lu, X.X., S. Zhang, J. Xu. 2010. Climate change and sed- Ligon, F.K., W.E. Dietrich, W.J. Trush. 1995. Downstream iment flux from the Roof of the World. Earth Surface ecological effects of dams. BioScience 45, 183–192. Processes and Landforms 35, 732–735. Link, T.E., G.N. Flerchinger, M. Unsworth, D. Marks. Lugt, H.J. 1983. Vortex Flow in Nature and Technology. 2005. Water relations in an old-growth Douglas fir JohnWiley&Sons,NewYork,297p. stand. In: C. de Jong, D. Collins, R. Ranzi, eds., Cli- Lundquist, J.D., M.D. Dettinger, D.R. Cayan. 2005. Snow- mate and Hydrology in Mountainous Areas.JohnWiley fed streamflow timing at different basin scales: case & Sons, Chichester, pp. 147–159. study of the Tuolumne River above Hetch Hetchy, Lisle, T.E. 1982. Effects of aggradation and degradation Yosemite, California. Water Resources Research 41, on riffle-pool morphology in natural gravel channels, W07005. doi:10.1029/2004WR003933 northwestern California. Water Resources Research 18, Lyons, W.B., C.A. Nezat, A.E. Carey, D.M. Hicks. 2002. 1643–1651. Organiccarbonfluxestotheoceanfromhigh-standing Lisle, T.E. 1986. Stabilization of a gravel channel by large islands. Geology 30, 443–446. streamside obstructions and bedrock bends, Jacoby MacFarlane, W.A., E. Wohl. 2003. Influence of step Creek, northwestern California. Geological Society of composition on step geometry and flow resistance America Bulletin 97, 999–1011. in step-pool streams of the Washington Cascades. Lisle, T.E. 2008. The evolution of sediment waves influ- Water Resources Research 39, 1037. doi:10.1029/ enced by varying transport capacity in heterogeneous 2001WR001238 rivers. In: H. Habersack, H. Piegay,M.Rinaldi,eds.,´ Mackay, J.R. 1963. The Mackenzie Delta area, N.W.T. Gravel-bed Rivers VI: From Process Understanding Geographical Branch, Mines and Technical Surveys, to River Restoration. Elsevier, Amsterdam, pp. 443– Ottawa,Memoir8,Queen’sPrinter,Ottawa,189p. 472. Mackin, J.H. 1937. Erosional history of the Big Horn Lisle, T.E., H. Ikeda, F. Iseya. 1991. Formation of station- Basin, Wyoming. Geological Society of America Bulletin ary alternate bars in a steep channel with mixed-size 48, 813–893. sediment: a flume experiment. Earth Surface Processes Mackin, J.H. 1948. Concept of the graded river. Geological and Landforms 16, 463–469. Society of America Bulletin 59, 463–512. 284 References

Macklin, M.G., P.A. Brewer, K.A. Hudson-Edwards, G. southeastern Minnesota (USA) streams. Environmental Bird,T.J.Coulthard,I.A.Dennis,P.J.Lechler,J.R.Miller, Management 42, 377–390. J.N. Turner. 2006. A geomorphological approach to the Maitre,V.,A.-C.Cosandey,E.Desagher,A.Parriaux. management of rivers contaminated by metal mining. 2003. Effectiveness of groundwater nitrate removal in a Geomorphology 79, 423–447. river riparian area: the importance of hydrogeological Macklin, M.G., K.A. Hudson-Edwards, E.J. Dawson. conditions. Journal of Hydrology 278, 76–93. 1997. The significance of pollution from historic metal Manners, R.B., M.W. Doyle, M.J. Small. 2007. Struc- mining in the Pennine orefields on river sediment con- ture and hydraulics of natural woody debris jams. taminant fluxes to the North Sea. The Science of the Total Water Resources Research 43, W06432. doi:10.1029/ Environment 194/195, 391–397. 2006WR004910 Macklin, M.G., J. Lewin. 2008. Alluvial responses to the Mao, L., A. Andreoli, F. Comiti, M.A. Lenzi. 2008. Geo- changing Earth system. Earth Surface Processes and morphic effects of large wood jams on a sub-Antarctic Landforms 33, 1374–1395. mountain stream. River Research and Applications 24, MacVicar, B., H. Piegay.´ 2012. Implementation and val- 249–266. idation of video monitoring for wood budgeting in Marchenko, S.S., A.P. Gorbunov, V.E. Romanovsky. 2007. a wandering piedmont river, the Ain River (France). Permafrost warming in the Tien Shan Mountains, cen- Earth Surface Processes and Landforms 37, 1272–1289. tral Asia. Global and Planetary Change 56, 311–327. Maddock, I. 1999. The importance of physical habitat Marcus, W.A., R.A. Marston, C.R. Colvard, R.D. Gray. assessmentforevaluatingriverhealth.Freshwater Biol- 2002. Mapping the spatial and temporal distributions ogy 41, 373–391. of woody debris in streams of the Greater Yellowstone Maddy,D.,T.Demir,D.R.Bridgland,A.Veldkamp,C. ecosystem, USA. Geomorphology 44, 323–335. Stemerdink, T. van der Schriek, R. Westaway. 2005. Marcus, W.A., G.A. Meyer, D.R Nimmo. 2001. Geomor- An obliquity-controlled early Pleistocene river terrace phic control of persistent mine impacts in a Yellowstone record from western Turkey? Quaternary Research 63, Park stream and implications for the recovery of fluvial 339–346. systems. Geology 29, 355–358. Madej, M.A., V. Ozaki. 2009. Channel response to sed- Marinucci,M.R.,F.Giorgi,M.Benitson,M.Wild,P. iment wave propagation and movement, Redwood Tschuck, A. Bernasconi. 1995. High resolution simula- Creek, California, USA. Earth Surface Processes and tion of January and July climate over the western Alpine Landforms 21, 911–927. region with a nested regional modelling system. Theo- Madole, R.D., D.P. Van Sistine, J.A. Michael. 1998. Pleis- retical and Applied Climatology 51, 119–138. tocene Glaciation in the Upper Platte River Drainage Markham, A.J., C.R. Thorne. 1992. Geomorphology of Basin, Colorado. Map, scale 1:500,000, U.S. Geological gravel-bed river bends. In: P. Billi, R.D. Hey, C.R. Survey, Denver, CO. Thorne, P.Tacconi, eds., Dynamics of Gravel-bed Rivers. Magilligan, F.J. 1992. Thresholds and the spatial variability John Wiley & Sons, Chichester, pp. 433–450. of flood power during extreme floods. Geomorphology Marks, K., P. Bates. 2000. Integration of high-resolution 5, 373–390. topographic data with floodplain flow models. Hydro- Magilligan, F.J., P.F. McDowell. 1997. Stream channel logical Processes 14, 2109–2122. adjustments following elimination of cattle grazing. Marks,J.C.,R.Parnell,C.Carter,E.C.Dinger,G.A. Journal of American Water Resources Association 33, Haden. 2006. Interactions between geomorphology and 867–878. ecosystem processes in travertine streams: implications Magilligan, F.J., K.H. Nislow. 2001. Long-term changes in for decommissioning a dam on Fossil Creek, Arizona. regional hydrologic regime following impoundment in Geomorphology 77, 299–307. a humid-climate watershed. Journal of American Water Marsh, G.P. 1864. Man and Nature: Physical Geography Resources Association 37, 1551–1569. as Modified by Human Action.CharlesScribner,New Magilligan, F.J., K.H. Nislow, C.E. Renshaw. 2013. Flow York. regulation by dams. In: E. Wohl, ed., Treatise on Fluvial Marston,R.A.,J.D.Mills,D.R.Wrazien,B.Bassett,D.K. Geomorphology. Elsevier, Amsterdam, pp. 794–807. Splinter. 2005. Effects of Jackson Lake Dam on the Magner, J.A., B. Vondracek, K.N. Brooks. 2008. Grazed Snake River and its floodplain, Grand Teton National riparian management and stream channel response in Park,Wyoming,USA.Geomorphology 71, 79–98. References 285

Martin, D.J., L.E. Benda. 2001. Patterns of instream McDonnell,J.J.,B.McGlynn,K.Kendall,J.Shanley,C. wood recruitment and transport at the watershed scale. Kendall. 1998. The role of near-stream riparian zones in Transactionsof the American Fisheries Society 130, 940– thehydrologyofsteepuplandcatchments.In:K.Kovar, 958. U. Tappeiner, N. E. Peters, R.G. Craig, eds., Hydrology, Martin, R.L., D.J. Jerolmack, R. Schumer. 2012. The phys- Water Resources and Ecology in Headwaters.IAHSPub- ical basis for anomalous diffusion in bed load trans- lication no. 248, pp. 173–180. port. JournalofGeophysicalResearch–EarthSurface McDonnell,J.J.,B.McGlynn,K.Vache,I.Tromp-Van 117, F01018. doi:10.1029/2011JF002075 Meerveld. 2005. A perspective on hillslope hydrology Martin, D.A., Moody, J.A., 2001. Comparison of soil infil- inthecontextofPUB.In:S.Franks,M.Sivapalan,K. tration rates in burned and unburned mountainous Takeuchi, Y. Tachikawa, eds., Predictions in Ungauged watersheds. Hydrological Processes 15, 2893–2903. Basins: International Perspectives on the State of the Mattingly, R.L., E.E. Herricks, D.M. Johnston. 1993. Art and Pathways Forward. IAHS Publication no. 301, Channelization and levee construction in Illinois: pp. 204–212. review and implications for management. Environmen- McDonnell, J.J., M.K. Stewart, I.F. Owens. 1991. Effect of tal Management 17, 781–795. catchment-scale subsurface mixing on stream isotopic May, C.L., R.E. Gresswell. 2003. Processes and rates of response. Water Resources Research 27, 3065–3073. sediment and wood accumulation in headwater streams McEwen, L.J., J.A. Matthews, R.A. Shakesby, M.S. Berris- of the Oregon Coast Range, USA. Earth Surface Pro- ford. 2002. Holocene gorge excavation linked to boul- cesses and Landforms 28, 409–424. der fan formation and frost weathering in a Norwe- May, C.L, R.E. Gresswell. 2004. Spatial and temporal pat- gian alpine periglaciofluvial system. Arctic, Antarctic terns of debris-flow deposition in the Oregon Coast and Alpine Research 34, 345–357. Range, USA. Geomorphology 57, 135–149. McEwan, I., Sørensen, M., Heald, J., Tait, S., Cunningham, McCabe, G.J., M.P. Clark, L.E. Hay. 2007. Rain-on-snow G., Goring, D., Willets, B., 2004. Probabilistic modeling events in the western United States. Bulletin of the of bed-load composition. Journal of Hydraulic Engineer- American Meteorological Society 88, 319–328. ing 130 (2), 129–139. McClain, M.E., E.W. Boyer, C.L. Dent, S.E. Gergel, N.B. McGinness, H.M., M.C. Thoms, M.R. Southwell. 2002. Grimm, P.M. Groffman, S.C. Hart, J.W. Harvey, C.A. Connectivity and fragmentation of flood plain-river Johnston, E. Mayorga, W.H. McDowell, G. Pinay. 2003. exchanges in a semiarid, anabranching river system. In: Biogeochemical hot spots and hot moments at the inter- F.J.Dyer,M.C.Thoms,J.M.Olley,eds.,The Structure, face of terrestrial and aquatic ecosystems. Ecosystems 6, Function, and Management Implications of Fluvial Sedi- 301–312. mentary Systems. IAHS Publication no. 276, pp. 19–26. McClain, M.E., G. Pinay, R.M. Holmes. 1999. Contrasting McGlynn, B., J.J. McDonnell. 2003. The role of discrete Biogeochemical Cycles of Riparian Forests in Temperate, landscape units in controlling catchment dissolved Wet Tropical, and Arid Regions. 1999 Annual Meeting organic carbon dynamics. Water Resources Research 39, abstracts, Ecological Society of America, p. 26. 3–18. McCormick, B.C., K.N. Eshleman, J.L. Griffith, P.A. McKean, J.A., W.E. Dietrich, R.C. Finkel, J.R. Southon, Townsend. 2009. Detection of flooding responses at M.W. Caffee. 1993. Quantification of soil production the river basin scale enhanced by land use change. and downslope creep rates from cosmogenic 10Be accu- Water Resources Research 45, W08401. doi:10.1029/ mulations on a hillslope profile. Geology 21, 343–346. 2008WR007594 Meade, R.H., ed. 1996. Contaminants in the Mississippi McDonald, A., S.N. Lane, N.E. Haycock, E.A. Chalk. 2004. River, 1987–92. U.S. Geological Survey Circular 1133, Riversofdreams:onthegulfbetweentheoreticaland Washington, DC. practical aspects of an upland river restoration. Trans- Meade, R.H. 2007. Transcontinental moving and storage: actionsoftheInstituteofBritishGeographers29, 257– theOrinocoandAmazonRiverstransfertheAndesto 281. the Atlantic. In: A. Gupta, ed., Large Rivers: Geomor- McDonnell, J.J. 2003. Where does water go when it phology and Management.JohnWiley&Sons,Chich- rains? Moving beyond the variable source area concept ester, pp. 45–63. of rainfall-runoff response. Hydrological Processes 17, Meade,R.H.,T.Dunne,J.E.Richey,U.M.Santos,E. 1869–1875. Salati. 1985. Storage and remobilization of suspended 286 References

sediment in the lower Amazon River of Brazil. Science A. Gupta, ed., Large Rivers.JohnWiley&Sons, 228, 488–490. Chichester, pp. 115–144. Meade, R.H., T.R. Yuzyk, T.J. Day. 1990. Movement and Mertes, L.A.K., T. Dunne, L.A. Martinelli. 1996. Channel- storage of sediment in rivers of the United States and floodplain geomorphology along the Solimoes-˜ Canada.In:M.G.Wolman,H.C.Riggs,eds.,Surface Amazon River, Brazil. Geological Society of America Water Hydrology. Geological Society of America, Boul- Bulletin 108, 1089–1107. der, CO, pp. 255–280. Meyer, G.A., S.G. Wells. 1997. Fire-related sedimentation Mei-e, R., Z. Xianmo. 1994. Anthropogenic influences events on alluvial fans, Yellowstone National Park, USA. on changes in the sediment load of the Yellow River, Journal of Sediment Research 67, 776–791. China, during the Holocene. The Holocene 4, 314– Meyer-Peter, E., R. Mueller. 1948. Formulas for Bedload 320. Transport. Proceedings of the 2nd International Asso- Meixner, T., R.C. Bales, M.W. Williams, D.H. Camp- ciation for Hydraulic Research Congress, Stockholm, bell, J.S. Baron. 2000. Stream chemistry modeling of pp. 39–65. two watersheds in the Front Range, Colorado. Water Miall, A.D. 1977. A review of the braided-river deposi- Resources Research 36, 77–87. tional environment. Earth Science Reviews 13, 1–62. Melis, T.S., ed. 2011. Effects of three high-flow experi- Milhous, R.T. 1973. Sediment Transport in a Gravel- ments on the Colorado River ecosystem downstream bottom Stream. PhD Dissertation.OregonStateUniver- from Glen Canyon Dam, Arizona. U.S. Geological Sur- sity,Corvallis,OR,238p. vey Circular 1366, 147 p. Miller, A.J. 1995. Valley morphology and boundary con- Merritt, D.M. 2013. Reciprocal relations between riparian ditions influencing spatial variations of flood flow. vegetation, fluvial landforms, and channel processes. In:J.E.Costa,A.J.Miller,K.W.Potter,P.R.Wilcock, In: E. Wohl, ed., Treatise on Fluvial Geomorphology. eds., Natural and Anthropogenic Influences in Fluvial Elsevier, Amsterdam, pp. 220–243. Geomorphology. American Geophysical Union Press, Merritts, D.J., K.R. Vincent. 1989. Geomorphic response Washington, DC, pp. 57–81. of coastal streams to low, intermediate, and high rates Miller, D.J., L. Benda. 2000. Effects of punctuated sedi- of uplift, Mendocino triple junction region, northern ment supply on valley-floor landforms and sediment California. Geological Society of America Bulletin 100, transport. Geological Society of America Bulletin 112, 1373–1388. 1814–1824. Merritts, D.J., K.R. Vincent, E.E. Wohl. 1994. Long river Miller, J., R. Barr, D. Grow, P. Lechler, D. Richardson, K. profiles, tectonism, and eustasy: a guide to interpreting Waltman, J. Warwick. 1999. Effects of the 1997 flood fluvial terraces. Journal of Geophysical Research 99(B7), on the transport and storage of sediment and mercury 14031–14050. within the Carson River valley, west-central Nevada. Merritt, D.M., E.E. Wohl. 2003. Downstream hydraulic Journal of Geology 107, 313–327. geometry and channel adjustment during a flood along Miller, A.J., D.J. Parkinson. 1993. Flood hydrology and an ephemeral, arid-region drainage. Geomorphology 52, geomorphic effects on river channels and floodplains: 165–180. the flood of November 4–5, 1985, in the South Potomac Merritt, D.M., E.E. Wohl. 2006. Plant dispersal along River basin of West Virginia. In: R.B. Jacobson, ed., Geo- rivers fragmented by dams. River Research and Appli- morphic Studies of the Storm and Flood of November 3– cations 22, 1–26. 4, 1985, in the Upper Potomac and Cheat River Basins in Mertes, L.A.K. 1997. Documentation and significance of West Virginia and Virginia. U.S. Geological Survey Bul- the perirheic zone on inundated floodplains. Water letin 1981, pp. E1–E96. Resources Research 33, 1749–1762. Milliman, J.D., R.H. Meade. 1983. World-wide delivery of Mertes, L.A.K. 2000. Inundation hydrology. In: E.E. river sediment to oceans. Journal of Geology 91, 1–21. Wohl, ed., Inland Flood Hazards: Human, Riparian, Milliman, J.D., J.P.M.Syvitski. 1992. Geomorphic/tectonic and Aquatic Communities. Cambridge University Press, control of sediment discharge to the ocean: the impor- Cambridge, pp. 145–166. tance of small mountainous rivers. The Journal of Geol- Mertes, L.A.K., T. Dunne. 2007. The effects of tec- ogy 100, 525–544. tonics,climatichistory,andsea-levelhistoryonthe Milly,P.C.D.,J.Betancourt,M.Falkenmark,R.M.Hirsch, form and behavior of the modern Amazon River. In: Z.W. Kundzewicz, D.P. Lettenmaier, R.J. Stouffer. 2008. References 287

Stationarity is dead: whither water management? Sci- landformsinanold-growthforestvalley,QueetsRiver, ence 319, 573–574. Washington, USA. Quaternary Research 65, 147–155. Milton, L.E. 1966. The geomorphic irrelevance of some Montgomery, D.R., T.B. Abbe, J.M. Buffington, N.P.Peter- drainage net laws. Australian Geographical Studies 4, son, K.M. Schmidt, J.D. Stock. 1996. Distribution of 89–95. bedrock and alluvial channels in forested mountain Minshall, G.W. 1984. Aquatic insect-substratum relation- drainage basins. Nature 381, 587–589. ships. In: V.H. Resh, D.M. Rosenberg, eds., The Ecol- Montgomery, D.R., J.M. Buffington. 1997. Channel-reach ogy of Aquatic Insects. Praeger Publishers, New York, morphology in mountain drainage basins. Geological pp. 358–400. Society of America Bulletin 109, 596–611. Mirus, B.B., Ebel, B.A., Loague, K., Wemple, B.C., 2007. Montgomery, D.R., J.M. Buffington, R.D. Smith, K.M. Simulated effect of a forest road on near-surface hydro- Schmidt, G. Pess. 1995. Pool spacing in forest channels. logic response: redux. Earth Surface Processes and Water Resources Research 31, 1097–1105. Landforms 32, 126–142. Montgomery, D.R., B.D. Collins, J.M. Buffington, T.B. Mitsch, W.J., J.W. Day, J.W. Gilliam, P.M. Groffman, D.L. Abbe. 2003a. Geomorphic effects of wood in rivers. In: Hey, G.W. Randall, N. Wang. 2001. Reducing nitrogen S.V. Gregory, K.L. Boyer, A.M. Gurnell, eds., The Ecol- loading to the Gulf of Mexico from the Mississippi River ogy and Management of Wood in World Rivers.Ameri- basin: strategies to counter a persistent ecological prob- can Fisheries Society, Bethesda, MD, pp. 21–47. lem. BioScience 51, 373–388. Montgomery, D.R., W.E. Dietrich. 1988. Where do chan- Moir, H.C., C.N. Gibbins, C. Soulsby, J. Webb. 2004. nels begin? Nature 336, 232–234. Linking channel geomorphic characteristics to spatial Montgomery, D.R., W.E. Dietrich. 1989. Source areas, patterns of spawning activity and discharge use by drainage density, and channel initiation. Water Atlantic salmon (Salmo salar L.). Geomorphology 60, Resources Research 25, 1907–1918. 21–35. Montgomery, D.R., W.E. Dietrich. 1992. Channel initia- Molnar,´ P. 2013. Network-scale energy distribution. In: tionandtheproblemoflandscapescale.Science 255, E. Wohl, ed., Treatise on Fluvial Geomorphology.Else- 826–830. vier, Amsterdam, pp. 43–49. Montgomery, D.R., W.E. Dietrich. 2002. Runoff genera- Molnar, P., P. England. 1990. Late Cenozoic uplift of tion in a steep, soil-mantled landscape. Water Resources mountain ranges and global climate change: chicken or Research 38(9), 1168. doi:10.1029/2001WR000822 egg? Nature 346, 29–34. Montgomery, D.R., W.E. Dietrich, J.T. Heffner. 2002. Molnar,´ P., J.A. Ram´ırez. 1998. An analysis of energy Piezometric response in shallow bedrock at CB1: expenditure in Goodwin Creek. Water Resources implications for runoff generation and landsliding. Research 34, 1819–1829. Water Resources Research 38(12), 1274. doi:10.1029/ Montgomery, D.R. 1994. Road surface drainage, chan- 2002WR001429 nel initiation, and slope instability. Water Resources Montgomery, D.R., W.E. Dietrich, K. Sullivan. 1998. The Research 30, 1925–1932. role of GIS in watershed analysis. In: S.N. Lane, K.S. Montgomery, D.R. 1999. Process domains and the river Richards, J.H. Chandler, eds., Landform Monitoring, continuum. Journal of the American Water Resources Modeling and Analysis. John Wiley & Sons, Chichester, Association 35, 397–410. pp. 241–261. Montgomery, D.R. 2002. Valley formation by fluvial and Montgomery, D.R., E. Foufoula-Georgiou. 1993. Channel glacial erosion. Geology 30, 1047–1050. network source representation using digital elevation Montgomery, D.R. 2004. Observations on the role of models. Water Resources Research 29, 3925–3934. lithology in strath terrace formation and bedrock chan- Montgomery, D.R., K.B. Gran. 2001. Downstream varia- nel width. American Journal of Science 304, 454– tions in the width of bedrock channels. Water Resources 476. Research 37, 1841–1846. Montgomery, D.R. 2007. Is agriculture eroding civiliza- Montgomery, D.R., T.M. Massong, S.C.S. Hawley. 2003b. tion’s foundation? Geological Society of America Today Influence of debris flows and log jams on the loca- 17(10), 4–9. tion of pools and alluvial channel reaches, Oregon Montgomery, D.R., T.B. Abbe. 2006. Influence of logjam- Coast Range. Geological Society of America Bulletin 115, formed hard points on the formation of valley-bottom 78–88. 288 References

Montgomery, D.R., K.M. Schmidt, W.E.Dietrich, J. McK- Naiman, R.J., H. Decamps,´ M.E. McClain. 2005. Riparia: ean. 2009. Instrumental record of debris flow initiation Ecology, Conservation, and Management of Streamside during natural rainfall: implications for modeling slope Communities. Elsevier, Amsterdam. stability. Journal of Geophysical Research 114, F01031. Naiman, R.J., C.A. Johnston, J.C. Kelley. 1988. Alteration doi:10.1029/2008JF001078 of North American streams by beaver. BioScience 38, Montgomery, D.R., D.B. Stolar. 2006. Reconsidering 753–762. Himalayan river anticlines. Geomorphology 82, 4–15. Naiman, R.J., J.M. Melillo, J.E. Hobbie. 1986. Ecosystem Moody, J.A., R.H. Meade. 2008. Terrace aggradation dur- alteration of boreal forest streams by beaver (Castor ing the 1978 flood on Powder River, Montana, USA. canadensis). Ecology 67, 1254–1269. Geomorphology 99, 387–403. Nakamura, F., F.J. Swanson. 1993. Effects of coarse woody Morgan, J. 1970. Deltas – a resum´ e.´ Journal of Geological debris on morphology and sediment storage of a moun- Education 18, 107–117. tain stream system in western Oregon. Earth Surface Morgan, P., G.H. Aplet, J.B. Haufler, H.C. Humphries, Processes and Landforms 18, 43–61. M.M. Moore, W.D. Wilson. 1994. Historical range Nanson, G.C. 1986. Episodes of vertical accretion and of variability: a useful tool for evaluating ecosystem catastrophic stripping: a model of disequilibrium flood- change. Journal of Sustainable Forestry 2, 87–111. plain development. Geological Society of America Bul- Moshe, L.B., I. Haviv, Y. Enzel, E. Zilberman, A. Matmon. letin 97, 1467–1475. 2008. Incision of alluvial channels in response to con- Nanson, G.C. 2013. Anabranching and anastomosing tinuous base level fall: field characterization, modeling, rivers. In: E. Wohl, ed., Treatise on Fluvial Geomorphol- and validation along the Dead Sea. Geomorphology 93, ogy. Elsevier, Amsterdam, pp. 330–344. 524–536. Nanson, G.C., J.C. Croke. 1992. A genetic classification of Mosley, M.P.1976. An experimental study of channel con- floodplains. Geomorphology 4, 459–486. fluences. Journal of Geology 84, 535–562. Nanson, G.C., H.Q. Huang. 2008. Least action principle, Mosley, M.P., D.S. Tinsdale. 1985. Sediment variability equilibrium states, iterative adjustment and the stability and bed material sampling in gravel-bed rivers. Earth of alluvial channels. Earth Surface Processes and Land- Surface Processes and Landforms 10, 465–482. forms 33, 923–942. Mote, P.W., A.F. Hamlet, M.P. Clark, D.P. Lettenmaier. Nanson, G.C., A.D. Knighton. 1996. Anabranching rivers: 2005. Declining mountain snowpack in western North their cause, character and classification. Earth Surface America. Bulletin of the American Meteorological Soci- Processes and Landforms 21, 217–239. ety 86, 39–49. Nanson, G.C., B.R. Rust, G. Taylor. 1986. Coexistent Moyle, P.B., J.F. Mount. 2007. Homogenous Rivers, mud braids and anastomosing channels in an arid-zone Homogenous Faunas. Proceedings of the National river: Cooper Creek, central Australia. Geology 14, 175– Academy of Sciences USA 104, pp. 5711–5712. 178. Murle,¨ U., J. Ortlepp, M. Zahner. 2003. Effects of experi- Nanson, G.C., S. Tooth, A.D. Knighton. 2002. A global mental flooding on riverine morphology, structure and perspective on dryland rivers: perceptions, misconcep- riparian vegetation: the River Spol,¨ Swiss National Park. tions and distinctions. In: L.J. Bull, M.J. Kirkby, eds., Aquatic Science 65, 191–198. Dryland Rivers: Hydrology and Geomorphology of Semi- Myers, T.J., S. Swanson. 1996. Long-term aquatic habitat Arid Channels. John Wiley & Sons, Chichester, pp. 17– restoration: Mahogany Creek, Nevada, as a case study. 54. Water Resources Bulletin 32, 241–252. National Academy. 1992. Restoration of Aquatic Ecosys- Nadler, C.T., S.A. Schumm. 1981. Metamorphosis of tems: Science, Technology, and Public Policy.National South Platte and Arkansas Rivers, eastern Colorado. Academy Press, Washington, DC, 552 p. Physical Geography 2, 95–115. Navratil, O., M.-B. Albert, E. Herouin,´ J.-M. Gresillon. Nageswara Rao, K., P. Subraelu, K.C.V. Naga Kumar, G. 2006. Determination of bankfull discharge magnitude Demudu, B. Hema Malini, A.S. Rajawat. 2010. Impacts andfrequency:comparisonofmethodson16gravel- of sediment retention by dams on delta shoreline reces- bed river reaches. Earth Surface Processes and Land- sion: evidences from the Krishna and Godavari deltas, forms 31, 1345–1363. India. Earth Surface Processes and Landforms 35, 817– Nelson, J. M., R.L. Shreve, S.R. McLean, T.G. Drake. 827. 1995. Role of near-bed turbulence structure in bed-load References 289

transport and bed form mechanics. Water Resources NOAA (National Oceanic and Atmospheric Adminis- Research 31(8), 2071–2086. tration). 1998. NCDC (National Climate Data Center) Newson, M.D., A.R.G. Large. 2006. ‘Natural’ rivers, Summary of the Day Weather Data. EarthInfo, Boulder, ‘hydromorphological quality’ and river restoration: a CO. challengingnewagendaforappliedfluvialgeomor- Nonaka, E., T.A.Spies. 2005. Historical range of variability phology. Earth Surface Processes and Landforms 31, in landscape structure: a simulation study in Oregon, 1606–1624. USA. Ecological Applications 15, 1727–1746. Nichols, K., P. Bierman, R. Finkel, J. Larsen. 2005. Sedi- Norris, R.H., M.C. Thoms. 1999. What is river health? ment generation rates for the Upper Rio Chagres basin: Freshwater Biology 41, 197–209. evidence from cosmogenic 10Be.In:R.S.Harmon,ed., O’Connor, J.E., J.J. Clague, J.S. Walder, V. Manville, R.A. The Rio Chagres, Panama: A Multidisciplinary Profile of Beebee. 2013. Outburst floods. In: E. Wohl, ed., Trea- aTropicalWatershed. Springer, Dordrecht, The Nether- tise on Fluvial Geomorphology. Elsevier, Amsterdam, lands, pp. 297–313. pp. 475–509. Nicholson, K., N.V.C. Polunin, P.L. Osborne. 1993. Sed- O’Connor, J.E., J.H. Hardison, J.E. Costa. 2001. Debris imentary geochemistry of Lake Daviumbu, part of Flows from Failures of Neoglacial-Age Moraine Dams the Fly River system, Papua New Guinea; impact in the Three Sisters and Mount Jefferson Wilderness of upstream copper-gold mining. Environmental Geo- Areas, Oregon.U.S.GeologicalSurveyProfessional chemistry and Health 15, 186–187. Paper 1606, 93 p. Niedzialek, J.M., F.L. Ogden. 2005. Runoff production O’Connor, J.E., M.A. Jones, T.L. Haluska. 2003. Flood in the upper Rio Chagres watershed, Panama. In: plain and channel dynamics of the Quinault and Queets R.S. Harmon, ed., The Upper Rio Chagres, Panama: Rivers, Washington, USA. Geomorphology 51, 31–59. A Multidisciplinary Profile of a Tropical Watershed. O’Connor, J.E., R.H. Webb. 1988. Hydraulic modeling for Springer, Dordrecht, The Netherlands, pp. 149– paleoflood analysis. In: V.R. Baker, R.C. Kochel, P.C. 168. Patton, eds., Flood Geomorphology.JohnWiley&Sons, Nienow, P., M. Sharp, I. Willis. 1998. Seasonal changes New York, pp. 393–402. in the morphology of the subglacial drainage system, O’Connor, J.E., R.H. Webb, V.R. Baker. 1986. Paleohy- Haut Glacier d’Arolla, Switzerland. Earth Surface Pro- drology of pool-and-riffle pattern development: Boul- cesses and Landforms 23, 825–843. der Creek, Utah. Geological Society of America Bulletin Nihlgard, B.J., W.T. Swank, M.J. Mitchell. 1994. Biolog- 97, 410–420. ical processes and catchment studies. In: B. Moldan, Oerlemans, J., E.J.L. Klok. 2004. Effect of summer snowfall J. Cerny, eds., Biogeochemistry of Small Catchments: A on glacier mass balance. Annals of Glaciology 38, 97– tool for Environmental Research.JohnWiley&Sons, 100. Chichester, pp. 133–161. Oguchi, T.,C.T.Oguchi. 2004. Late Quaternary rapid talus Nijssen,B.,G.M.O’Donnell,A.F.Hamlet,D.P.Letten- dissectionanddebrisflowdepositiononanalluvialfan maier. 2001 Hydrologic sensitivity of global rivers to cli- in Syria. Catena 55, 125–140. mate change. Climatic Change 50, 143–175. Oguchi, T., H. Ohmori. 1994. [Analysis of relationships Nilsson, C., K. Berggren. 2000. Alterations of riparian among alluvial fan area, source basin area, basin slope, ecosystems caused by river regulation. BioScience 50, and sediment yield]. Zeitschrift fur Geomorphologie 38, 783–792. 405–420. Nilsson, C., R. Jansson, U. Zinko. 1997. Long-term Oguchi, T., T. Wasklewicz, Y.S. Hayakawa. 2013. Remote responses of river-margin vegetation to water-level reg- data in fluvial geomorphology: characteristics and ulation. Science 276, 798–800. applications. In: E. Wohl, ed., Treatise on Fluvial Geo- Nilsson, C., C.A. Reidy, M. Dynesius, C. Revenga. 2005. morphology. Elsevier, Amsterdam, pp. 711–729. Fragmentation and flow regulation of the world’s large Ohmori, H. 1991. Change in the mathematical func- river systems. Science 308, 405–408. tion type describing the longitudinal profile of a river Nittrouer, J.A., D. Mohrig, M.A. Allison, A.B. Peyret. through an evolutionary process. Journal of Geology 99, 2011. The lowermost Mississippi River: a mixed 97–110. bedrock-alluvial channel. Sedimentology. doi:10.10111/ Ollesch, G., I. Kistner, R. Meissner, K.-E. Lindenschmidt. j.1365–3091.2011.01245.x 2006. Modelling of snowmelt erosion and sediment 290 References

yieldinasmalllow-mountaincatchmentinGermany. availability on cluster evolution. Earth Surface Processes Catena 68, 161–176. and Landforms 28, 69–86. Onda, Y., Y. Komatsu, M. Tsujimura, J. Fujihara. 2001. Park, C.C. 1977. World-wide variations in hydraulic Theroleofsubsurfacerunoffthroughbedrockon geometry exponents of stream channels: an analysis and storm flow generation. Hydrological Processes 15, 1693– some observations. Journal of Hydrology 33, 133–146. 1706. Parker, G. 1976. On the cause and characteristic scales Osborne, L.L., P.B. Bayley, L.W.G. Higler, B. Statzner, of meandering and braiding in rivers. Journal of Fluid F. Triska, T.M. Iversen. 1993. Restoration of lowland Mechanics 76, 457–480. streams: an introduction. Freshwater Biology 29, 187– Parker, G. 1990. Surface-based bedload transport relation 194. for gravel rivers. Journal of Hydraulic Research 28, 417– Osman, A.M., C.R. Thorne. 1988. Riverbank stability 436. analysis I: theory. Journal of Hydraulic Engineering 114, Parker, G. 1991. Downstream variation of grain size 134–150. in gravel rivers: abrasion versus selective sorting. In: Oswald, E.B., E. Wohl. 2008. Wood-mediated geomorphic A. Armanini, G. DiSilvio, eds., Fluvial Hydraulics effects of a jokulhlaup¨ in the Wind River Mountains, of Mountain Regions. Springer, Berlin, pp. 347– Wyoming. Geomorphology 100, 549–562. 360. Oswood, M.W., A.M. Milner, J.G. Irons. 1992. Climate Parker, G. 2008. Transport of gravel and sediment mix- change and Alaskan rivers and streams. In: P. Firth, tures.In:Garcia,M.,ed.,Sedimentation Engineering: S.G. Fisher, eds., Global Climate Change and Freshwater Theory, Measurements, Modeling and Practice (ASCE Ecosystems. Springer, New York, pp. 192–210. Manuals and Reports on Engineering Practice No. 110). Palmer, M.A., R.F. Ambrose, N.L. Poff. 1997. Ecological American Society of Civil Engineers, Reston, VA, theory and community restoration ecology. Restoration pp. 165–251. Ecology 5, 291–300. Parker, G., Y. Cui, J. Imran, W.E. Dietrich. 1997. Flood- Palmer, M.A., E.S. Bernhardt, W.H. Schlesinger, K.N. ing in the Lower Ok Tedi, Papua New Guinea due to the Eshleman,E.Foufoula-Georgiou,M.S.Hendryx,A.D. Disposal of Mine Tailings and its Amelioration.Inter- Lemly, G.E. Likens, O.L. Loucks, M.E. Power, P.S. national seminar on recent trends of floods and their White, P.R. Wilcock. 2010a. Mountaintop mining con- preventive measures in commemoration of the Tenth sequences. Science 327, 148–149. Anniversary of the Foundation of Hokkaido River Dis- Palmer, M.A., H.L. Menninger, E. Bernhardt. 2010b. River aster Prevention Research Center, 20–21 June 1996 Sap- restoration, habitat heterogeneity and biodiversity: a poro, pp. 21–48. failure of theory or practice? Freshwater Biology 55, Parker, G., P.C. Klingeman, D.G. McLean. 1982. Bedload 205–222. and size distribution in paved gravel-bed streams. Jour- Pan,B.,D.Burbank,Y.Wang,G.Wu,J.Li,Q.Guan. nal of the Hydraulic Division, American Society of Civil 2003. A 900 k.y. record of strath terrace formation dur- Engineering 108, 544–571. ing glacial-interglacial transitions in northwest China. Parker, G., T. Muto, Y. Akamatsu, W.E. Dietrich, J.W. Geology 31, 957–960. Lauer. 2008. Unravelling the conundrum of river Painter,T.H.,J.Deems,J.Belnap,A.Hamlet,C.C. response to rising sea-level from laboratory to field: Landry, B. Udall. 2010. Response of Colorado River part II, the Fly-Strickland River system, Papua New Runoff to Dust Radiative Forcing in Snow. Proceed- Guinea. Sedimentology 55, 1657–1686. ings of the National Academy of Sciences 10.1073/ Parker, G., C. Paola, S. Leclair. 2000. Probabilistic Exner pnas.0913139107. sediment continuity equation for mixtures with no Paola, C. 2001. Modelling stream braiding over a range of active layer. Journal of Hydraulic Engineering 126, 818– scales. In: M.P. Mosley, ed., Gravel-bed Rivers V.New 826. Zealand Hydrological Society, Wellington, pp. 11–46. Parker, G., C.M. Toro-Escobar. 2002. Equal mobil- Paola, C., D. Mohrig. 1996. Paleohydraulics revisited: ityofgravelinstreams:theremainsoftheday. paleoslope estimation in coarse-grained braided rivers. Water Resources Research 38, 1264. doi:10.1029/ Basin Research, 8, 243–254. 2001WR000669 Papanicolaou, A.N., K. Strom, A. Schuyler, N. Talebbey- Parker, G., C.M. Toro-Escobar, M. Ramey, S. Beck. 2003. dokhti. 2003. The role of sediment specific gravity and Effect of floodwater extraction on mountain stream References 291

morphology. Journal of Hydraulic Engineering 129, using a spatial autocorrelation technique for the hyp- 885–895. sometric integral. Journal of Geophysical Research 114, Parrish, J.D., D.P. Braun, R.S. Unnasch. 2003. “Are we F02018. doi:10.1029/2008JF001092 conserving what we say we are?” Measuring ecologi- Personius, S.F., H.M. Kelsey, P.C. Grabau. 1993. Evidence cal integrity within protected areas. BioScience 53, 851– for regional stream aggradation in the Central Oregon 860. CoastRangeduringthePleistocene-Holocenetransi- Parvis, M. 1950. Drainage pattern significance in air- tion. Quaternary Research 40, 297–308. photo identification of soils and bedrock. U.S. National Petroski, H. 2006. Levees and other raised ground. Amer- Academy of Sciences National Research Council. High- ican Scientist 94, 7–11. way Research Board Bulletin 28, 36–62. Petts, G.E. 1979. Complex response of river channel mor- Pasternack, G.B. 2013. Geomorphologist’s guide to par- phology subsequent to reservoir construction. Progress ticipating in river rehabilitation. In: E. Wohl, ed., Trea- in Physical Geography 3, 329–362. tise on Fluvial Geomorphology. Elsevier, Amsterdam, Petts, G.E. 1984. Impounded Rivers.JohnWiley&Sons, pp. 843–860. Chichester. Patrick, R., ed. 1995. Chemicals in riverine water. In: Petts, G.E., A.M. Gurnell. 2005. Dams and geomorphol- Rivers of the United States, Volume. 2, Chemical and ogy: research progress and future directions. Geomor- Physical Characteristics. John Wiley & Sons, New York, phology 71, 27–47. pp. 195–228. Phillips, J.D., L. Park. 2009. Forest blowdown impacts of Patton, P.C. 1988. Geomorphic response of streams to Hurricane Rita on fluvial systems. Earth Surface Pro- floods in the glaciated terrain of southern New Eng- cesses and Landforms 34, 1069–1081. land. In: V.R. Baker, R.C. Kochel, P.C. Patton, eds., Pickup, G., R.F. Warner. 1976. Effects of hydrologic Flood Geomorphology. John Wiley & Sons, New York, regime on magnitude and frequency of dominant dis- pp. 261–277. charge. Journal of Hydrology 29, 51–75. Pazzaglia, F.J. 2013. Fluvial terraces. In: E. Wohl, ed., Piegay,´ H., A.M. Gurnell. 1997. Large woody debris and Treatise on Fluvial Geomorphology. Elsevier, pp. 379– river geomorphological pattern: examples from S.E. 412. France and S. England. Geomorphology 19, 99–116. Pazzaglia, F.J., M.T. Brandon. 2001. A fluvial record Piegay,´ H., P.-G.Salvador. 1997. Contemporary floodplain of long-term steady-state uplift and erosion across forest evolution along the middle Ubaye River, South- the Cascadia forearc high, western Washington State. ern Alps, France. Global Ecology and Biogeography Let- American Journal of Science 301, 385–431. ters 6, 397–406. Pederson, J.L., M.D. Anders, T.M. Rittenour, W.D. Sharp, Pierce, J., G. Meyer. 2008. Long-term fire history from J.C. Gosse, K.E. Karlstrom. 2006. Using fill ter- alluvial fan sediments: the role of drought and climate races to understand incision rates and evolution of variability, and implications for management of Rocky the Colorado River in eastern Grand Canyon, Ari- Mountain forests. International Journal of Wildland Fire zona. Journal of Geophysical Research 111, F02003. 17, 84–95. doi:10.1029/2004JF000201 Pierson, T.C. 2005. Hyperconcentrated flow – transitional Pelletier,J.D.,L.Mayer,P.A.Pearthree,P.K.House,K.A. process between water flow and debris flow. In: M. Demsey, J.E. Klawon, K.R. Vincent. 2005. An inte- Jakob, O. Hungr, eds., Debris-flow Hazards and Related grated approach to flood hazard assessment on allu- Phenomena. Springer, Berlin, pp. 159–202. vial fans using numerical modeling, field mapping, and Pierson, T.C. 2007. Dating young geomorphic surfaces remote sensing. Geological Society of America Bulletin using age of colonizing Douglas fir in southwestern 117, 1167–1180. Washington and northwestern Oregon, USA. Earth Penkman,K.E.H.,R.C.Preece,D.H.Keen,D.Maddy, Surface Processes and Landforms 32, 811–831. D.C. Schreve, M.J. Collins. 2007. Testing the aminos- Pike, A.S., F.N. Scatena. 2010. Riparian indicators of flow tratigraphy of fluvial archives: the evidence from intra- frequency in a tropical montane stream network. Jour- crystalline proteins within freshwater shells. Quater- nal of Hydrology 382, 72–87. nary Science Reviews 26, 2958–2969. Pike, A.S., F.N. Scatena, E.E. Wohl. 2010. Longitudinal Perez-Pe´ na,˜ J.V., J.M. Azan˜on,G.Booth-Rea,A.Azor,J.´ patterns in stream channel morphology in the tropical Delgado. 2009. Differentiating geology and tectonics mountain streams of the Luquillo Mountains, Puerto 292 References

Rico. Earth Surface Processes and Landforms 35, 1402– Poole, G.C. 2002. Fluvial landscape ecology: addressing 1417. uniqueness within the river discontinuum. Freshwater Piˇsut,´ P. 2002. Channel evolution of the pre-channelised Biology 47, 641–660. Danube River in Bratislava, Slovakia (1712–1886). Potter, P.E. 1978. Significance and origin of big rivers. Earth Surface Processes and Landforms 27, 369–390. Journal of Geology 86, 13–33. Pitlick, J. 1994. Relation between peak flows, precipita- Poulos, S.E., G. Ghionis, H. Maroukian. 2009. The conse- tion, and physiography for five mountainous regions in quences of a future eustatic sea-level rise on the deltaic the western USA. Journal of Hydrology 158, 219–240. coasts of Inner Thermaikos Gulf (Aegean Sea) and Pitlick, J., P.R. Wilcock. 2001. Relations between stream- Kyparissiakos Gulf (Ionian Sea), Greece. Geomorphol- flow,sedimenttransport,andaquatichabitatinregu- ogy 107, 18–24. lated rivers. In: J.M. Dorava, D.R. Montgomery, B.B. Powell, J.W. 1875. Exploration of the Colorado River of Palcsak, F.A. Fitzpatrick eds., Geomorphic Processes and the West (1869–72). U.S. Government Printing Office, Riverine Habitat. American Geophysical Union Press, Washington, DC. Washington, DC, pp. 185–198. Powell, J.W.1876. Report on the Geology of the Eastern Por- Pizzuto, J. 2002. Effects of dam removal on river form and tion of the Uinta Mountains. U.S. Government Printing process. BioScience 52, 683–691. Office, Washington, DC. Pizzuto, J. 2003. Numerical modeling of alluvial land- Powell, D.M. 1998. Patterns and processes of sediment forms.In:G.M.Kondolf,H.Piegay,´ eds., Tools in Flu- sortingingravel-bedrivers.Progress in Physical Geog- vial Geomorphology. John Wiley & Sons, Chichester, raphy 22, 1–32. pp. 577–595. Prestegaard, K.L. 1983. Variables influencing water- Pizzuto, J., M. O’Neal, S. Stotts. 2010. On the retreat surface slopes in gravel-bed streams at bankfull stage. of forested, cohesive riverbanks. Geomorphology 116, Geological Society of America Bulletin 94, 673–678. 341–352. Priesnitz, K., E. Schunke. 2002. The fluvial morpho- Poff, N.L., J.D. Allan, M.B. Bain, J.R. Karr, K.L. Preste- dynamics of two small permafrost drainage basins, gaard, B.D. Richter, R.E. Sparks, J.C. Stromberg. 1997. Richardson Mountains, northwestern Canada. Per- The natural flow regime: a paradigm for river conser- mafrost Periglacial Processes 13, 207–217. vation and restoration. BioScience 47, 769–784. Pringle, C.M. 2001. Hydrologic connectivity and the man- Poff, N.L., J.D. Olden, D.M. Merritt, D.M. Pepin. 2007. agement of biological reserves: a global perspective. Homogenization of regional river dynamics by dams Ecological Applications 11, 981–998. and global biodiversity implications. Proceedings of the Pringle, C.M., R.J. Naiman, G. Bretschko, J.R. Karr, M.W. National Academy of Sciences 104, 5732–5737. Oswood, J.R. Webster, R.L. Welcomme, M.J. Winter- Pollen, N., A. Simon. 2005. Estimating the mechanical bourn. 1988. Patch dynamics in lotic systems: the effects of riparian vegetation on stream bank stability stream as a mosaic. Journal of the North American Ben- usingafiberbundlemodel.Water Resources Research thological Society 7, 503–524. 41, W07025. doi:10.1029/2004WR003801, 11 p. Procter,J.,S.J.Cronin,I.C.Fuller,G.Lube,V.Manville. Pollen-Bankhead, N., A. Simon. 2010. Hydrologic and 2010. Quantifying the geomorphic impacts of a lake- hydraulic effects of riparian root networks on stream- breakout lahar, Mount Ruapehu, New Zealand. Geology bank stability: is mechanical root reinforcement the 38, 67–70. whole story? Geomorphology 116, 353–362. Prospero, J.M. 1999. Long-range transport of mineral dust Polvi, L.E., E. Wohl. 2012. The beaver meadow complex in the global atmosphere: impact of African dust on revisited–theroleofbeaversinpost-glacialfloodplain the environment of the southeastern United States. Pro- development. Earth Surface Processes and Landforms ceedings of the National Academy of Science USA 96, 37, 332–346. 3396–3403. Polvi, L.E., E. Wohl. 2013. Biotic drivers of river planform Prosser, I.P., B. Abernethy. 1996. Predicting the topo- – implications for understanding the past and restoring graphic limits to a gully network using a digital ter- the future. BioScience 63, 439–452. rain model and process thresholds. Water Resources Polvi, L.E., E.E. Wohl, D.M. Merritt. 2011. Geomorphic Research 32, 2289–2298. and process domain controls on riparian zones in the Prosser, I.P., W.E. Dietrich, J. Stevenson. 1995. Flow Colorado Front Range. Geomorphology 125, 504–516. resistance and sediment transport by concentrated References 293

overland flow in a grassland valley. Geomorphology 13, lennial time scales. In: P.K. House, R.H. Webb, V.R. 71–86. Baker, D.R. Levish, eds., Ancient Floods, Modern Haz- Prowse, T., K. Alfredsen, S. Beltaos, B.R. Bonsal, W.B. ards: Principles and Applications of Paleoflood Hydrol- Bowden, C. Duguay, A. Korhola, J. McNamara, W.F. ogy. American Geophysical Union Press, Washington, Vincent, V.Vuglinsky, K.M. Walter Anthony, G.A. Wey- DC, pp. 21–45. henmeyer. 2011. Effects of changes in Arctic lake and Reid, D.E., E.J. Hickin, S.C. Babakaiff. 2010. Low-flow river ice. Ambio 40, 63–74. hydraulic geometry of small, steep mountain streams Prowse, T.D., S. Beltaos. 2002. Climatic control of river- in southwest British Columbia. Geomorphology 122, ice hydrology: a review. Hydrological Processes 16, 805– 39–55. 822. Rengers, F.J., E. Wohl. 2007. Trends of grain sizes on gravel Prowse, T.D., T. Carter. 2002. Significance of ice-induced bars in the Rio Chagres, Panama. Geomorphology 83, storage to spring runoff: a case study of the Mackenzie 282–293. River. Hydrological Processes 16, 779–788. Restrepo, J.D., A. Kettner. 2012. Human-induced dis- Prowse, T.D., M.G. Ferrick. 2002. Hydrology of ice- charge diversions in a tropical delta and its environ- covered rivers and lakes: scoping the subject. Hydrolog- mental implications: the Patia River, Colombia. Journal ical Processes 16, 759–762. of Hydrology 424–425, 124–142. Pyrce, R.S., P.E.Ashmore. 2003. The relation between par- Reynolds, L.V., D.J. Cooper, N.T. Hobbs. 2012. Drivers of ticlepathlengthdistributionsandchannelmorphology riparian tree invasion on a desert stream. River Research in gravel-bed streams: a synthesis. Geomorphology 56, and Applications. doi: 10.1002/rra.2619 167–187. Rhea, S. 1993. Geomorphic observations of rivers in the Rachold, V., H. Eicken, V.V. Gordeev, M.N. Grigoriev, Oregon Coast Range from a regional reconnaissance H.-W.Hubberten,A.P.Lisitzin,V.P.Shevchenko,L. perspective. Geomorphology 6, 135–150. Schirrmeister. 2004. Modern terrigenous organic car- Rhoads, B.L., D. Wilson, M. Urban, E.E. Herricks. 1999. bon input to the Arctic Ocean. In: R. Stein, R.W. Mac- Interaction between scientists and nonscientists in donald, eds., TheOrganicCarbonCycleintheArctic community-based watershed management: emergence Ocean. Springer, pp. 33–55. of the concept of stream naturalization. Environmental Radecki-Pawlik, A. 2002. Bankfull discharge in mountain Management 24, 297–308. streams: theory and practice. Earth Surface Processes Rice, S. 1994. Towards a model of changes in bed material and Landforms 27, 115–123. texture at the drainage basin scale. In: M.J. Kirkby, ed., Raff, D.A., J.A. Ram´ırez, J.L. Smith. 2004. Hillslope Process Models and Theoretical Geomorphology.John drainage development with time: a physical experi- Wiley & Sons, Chichester, pp. 159–172. ment. Geomorphology 62, 169–180. Rice, S.P. 1998. Which tributaries disrupt downstream Rao, P.G. 1995. Effect of climate change on streamflows fining along gravel-bed rivers? Geomorphology 22, in the Mahanadi River basin, India. Water International 39–56. 20, 205–212. Rice, S.P, M. Church. 1996. Sampling surficial fluvial Rathburn, S.L., D.M. Merritt, E.E. Wohl, J.S. Sander- gravels: the precision of size distribution percentile son, H.A.L. Knight. 2009. Characterizing environmen- estimates. Journal of Sedimentary Research 66A, 654– tal flows for maintenance of river ecosystems: North 665. ForkCachelaPoudreRiver,Colorado.In:L.A.James, Richards, K.S. 1982. Rivers: Form and Process in Alluvial S.L. Rathburn, G.R. Whittecar, eds., Management and Channels. Methuen, London, 348 p. Restoration of Fluvial Systems with Broad Historical Richardson, K., P.A.Carling. 2005. A typology of sculpted Changes and Human Impacts. Geological Society of forms in open bedrock channels. Geological Society of America Special Paper 451, Boulder, CO, pp. 143– America Special Paper 392.Boulder,CO. 157. Richey,J.E.,L.A.K.Mertes,T.Dunne,R.Victoria,B.R., Raymo, M.E., W.F. Ruddiman, P.N. Froelich. 1988. Influ- Forsberg, A. Tancredi, E. Oliveira. 1989. Sources and ence of late Cenozoic mountain building on geochemi- routing of the Amazon River flood wave. Global Bio- cal cycles. Geology 16, 649–653. geochemical Cycles 3, 191–204. Redmond, K.T., Y. Enzel, P.K. House, F. Biondi. 2002. Cli- Richmond, A.D., K.D. Fausch. 1995. Characteristics and mate variability and flood frequency at decadal to mil- function of large woody debris in subalpine Rocky 294 References

Mountain streams in northern Colorado. Canadian ture of river basins. Water Resources Research 28, 1095– Journal of Fisheries and Aquatic Sciences 52, 1789– 1103. 1802. Rodriguez-Iturbe, I., J.B. Valdes. 1979. The geomorpho- Richter, B.D., J. Baumgartner, J. Powell, D. Braun. 1996. logic structure of hydrologic response. Water Resources A method for assessing hydrologic alteration within Research 15, 1409–1420. ecosystems. Conservation Biology 10, 1163–1174. Rodr´ıguez-Rodr´ıguez, M., J. Benavente, F.J. Alcala,´ M. Rickenmann, D. 1991. Bed load transport and hyper- Paracuellos. 2011. Long-term water monitoring in two concentrated flow at steep slopes. In: A. Armanini, G. Mediterranean lagoons as an indicator of land-use DiSilvio, eds., Fluvial Hydraulics of Mountain Regions. changes and intense precipitation events (Adra, south- Springer, Berlin, pp. 429–441. eastern Spain). Estuarine, Coastal and Shelf Science 91, Riggsbee, J.A., M.W. Doyle, J.P. Julian, R. Manners, J.D. 400–410. Muehlbauer, J. Sholtes, M.J. Small. 2013. Influence Roe, G.H., D.R. Montgomery, B. Hallett. 2002. Effects of of aquatic and semi-aquatic organisms on channel orographic precipitation variations on the concavity of forms and processes. In: E. Wohl, ed., Treatise on Flu- steady-state river profiles. Geology 50, 143–146. vial Geomorphology. Elsevier, Amsterdam, pp. 189– Roering, J.J., P. Almond, P. Tonkin, J. McKean. 2002. Soil 201. transport driven by biological processes over millennial Rigon, E., F. Comiti, L. Mao, M.A. Lenzi. 2008. Rela- time scales. Geology 30, 1115–1118. tionships among basin area, sediment transport mech- Roering, J.J., M. Gerber. 2005. Fire and the evolution anisms and wood storage in mountain basins of the of steep, soil-mantled landscapes. Geology 33, 349– Dolomites (Italian Alps). WIT Transactions on Engi- 352. neering Sciences 60, 163–172. Roering, J.J., J.W. Kirchner, W.E. Dietrich. 2005. Char- Riihimaki, C.A., R.S. Anderson, E.B. Safran, D.P. Dethier, acterizing structural and lithologic controls on deep- R.C. Finkel, P.R. Bierman. 2006. Longevity and pro- seated landsliding: Implications for topographic relief gressive abandonment of the Rocky Flats surface, Front and landscape evolution in the Oregon Coast Range, Range, Colorado. Geomorphology 78, 265–278. USA. Geological Society of America Bulletin 117, 654– Rinaldi, M., S.E. Darby. 2008. Modelling river-bank ero- 668. sion processes and mass failure mechanisms: progress Roering, J.J., K.M. Schmidt, J.D. Stock, W.E.Dietrich, D.R. towards fully coupled simulations. In: H. Habersack, Montgomery. 2003. Shallow landsliding, root reinforce- H. Piegay,M.Rinaldi,eds.,´ Gravel-bed Rivers VI: From ment, and the spatial distribution of trees in the Oregon Process Understanding to River Restoration. Elsevier, Coast Range. Canadian Geotechnical Journal 40, 237– Amsterdam, pp. 213–239. 253. Ritter, D.F., R.C. Kochel, J.R. Miller. 2011. Process geo- Rood, S.B., J. Pan, K.M. Gill, C.G. Franks, G.M. Samuel- morphology, 5th edn. Waveland Press, Long Grove, IL, son, A. Shepherd. 2008. Declining summer flows of 652 p. Rocky Mountain rivers: changing seasonal hydrology Robert, A. 2003. River Processes: An Introduction to and probable impacts on floodplain forests. Journal of Fluvial Dynamics.Arnold,London. Hydrology 349, 397–410. Robert, A., A.G. Roy, B. De Serres. 1996. Turbulence at Rosgen, D.L., H.L. Silvey. 1996. Applied River Morphology. a roughness transition in a depth limited flow over a Wildland Hydrology, Pagosa Springs, CO, 390 p. gravel bed. Geomorphology 16, 175–187. Roy, A.G., A.D. Abrahams. 1980. Rhythmic spac- Roberts, G.G., N. White. 2010. Estimating uplift rate ing and origin of pools and riffles: discussion and histories from river profiles using African exam- reply. Geological Society of America Bulletin 91, 248– ples. Journal of Geophysical Research 115, B02406. 250. doi:10.1029/2009JB006692 Roy,A.G.,P.Biron,T.Buffin-Belanger,´ M. Levasseur. Rodr´ıguez-Iturbe, I., A. Rinaldo. 1997. Fractal River 1999. Combined visual and quantitative techniques in Basins: Chance and Self-Organization.Cambridge the study of natural turbulent flows. Water Resources University Press, 547 p. Research 35, 871–877. Rodr´ıguez-Iturbe, I., A. Rinaldo, R. Rigon, R.L. Bras, A. Ruban, G.I. 1997. Species structure, contemporary distri- Marani, E. Ijjasz-V´ asquez.´ 1992. Energy dissipation, bution and status of the Siberian sturgeon, Acipenser runoff production, and the three-dimensional struc- baerii. Environmental Biology of Fishes 48, 221–230. References 295

Rubin, D.M., J.M. Nelson, D.J. Topping. 1998. Relation Carp in America. Unpublished PhD dissertation, Uni- of inversely graded deposits to suspended-sediment versity of Illinois at Urbana-Champaign. grain-size evolution during the 1996 flood experiment Savenije, H. 2005. Salinity and Tides in Alluvial Estuaries. in Grand Canyon. Geology 26, 99–102. Elsevier, Amsterdam. Rubin, Z., S.L. Rathburn, E. Wohl, D.L. Harry. 2012. His- Sayama, T., J.J. McDonnell. 2009. A new time-space toric range of variability in geomorphic processes as a accounting scheme to predict stream water residence context for restoration: Rocky Mountain National Park, time and hydrograph source components at the water- Colorado, USA. Earth Surface Processes and Landforms shed scale. Water Resources Research 45, W07401. 37, 209–222. doi:10.1029/2008WR007549 Rustomji, R., I. Prosser. 2001. Spatial patterns of sediment Saynor, M.J., R.J. Loughran, W.D. Erskine, P.F. Scott. delivery to valley floors: sensitivity to sediment trans- 1994. Sediment movement on hillslopes measured port capacity and hillslope hydrology relations. Hydro- by caesium-137 and erosion pins. In: L.J. Olive, R.J. logical Processes 15, 1003–1018. Loughran,J.A.Kesby,eds.,Variability in Stream Ero- Ryan, S.E. 1997. Morphologic response of subalpine sion and Sediment Transport. IAHS Publication no. 224, streamstotransbasinflowdiversion.Journal of the pp. 87–93. American Water Resources Association 33, 839–854. Scanlon, B.R., I. Jolly, M. Sophocleous, L. Zhang. 2007. Ryder, J.M. 1971. The stratigraphy and morphology Global impacts of conversions from natural to agri- of paraglacial alluvial fans in south-central British cultural ecosystems on water resources: quantity ver- Columbia. Canadian Journal of Earth Sciences 8, 279– sus quality. Water Resources Research 43, W03437. 298. doi:10.1029/2006WR005486, 18 p. Safford, H.D., J.L. Betancourt, G.D. Hayward, J.A. Scatena, F.N., A. Gupta. 2013. Streams of the mon- Wiens, C.M. Regan. 2008. Land management in tane humid tropics. In: E. Wohl, ed., Treatise on Flu- the Anthropocene: is history still relevant? Trans- vial Geomorphology. Elsevier, Amsterdam, pp. 595– actions of the American Geophysical Union 89(37), 610. 343. ScatenaF.N., A.E. Lugo. 1995. Geomorphology, distur- Salant, N.L., C.E. Renshaw, F.J. Magilligan. 2006. Short bance, and the soil and vegetation of two subtropical and long-term changes to bed mobility and bed compo- wet steepland watersheds of Puerto Rico. Geomorphol- sition under altered sediment regimes. Geomorphology ogy 13, 199–213. 76, 43–53. Schmeeckle, M.W., J.M. Nelson, R.L. Shreve. 2007. Samadi, A., E. Amiri-Tokaldany, S.E. Darby. 2009. Identi- Forces on stationary particles in near-bed turbulent fying the effects of parameter uncertainty on the relia- flows. Journal of Geophysical Research 112, F02003. bility of riverbank stability modelling. Geomorphology doi:10.1029/2006JF000536 106, 219–230. Schmidt, K.-H., P. Ergenzinger. 1992. Bedload entrain- Sambrook Smith, G.H., A.P. Nicholas. 2005. Effect ment,travellengths,steplengths,restperiods–studied onflowstructureofsanddepositiononagravel with passive (iron, magnetic) and active (radio) tracer bed: results from a two-dimensional flume exper- techniques. Earth Surface Processes and Landforms 17, iment. Water Resources Research 41, W10405. 147–165. doi:10.1029/2004WR003817 Schmidt, J.C., J.B. Graf. 1990. Aggradation and Degra- Sanborn, S.C., B.P. Bledsoe. 2006. Predicting streamflow dation of Alluvial Sand Deposits, 1965 to 1986, regime metrics for ungauged streams in Colorado, Colorado River, Grand Canyon National Park, Arizona. Washington, and Oregon. Journal of Hydrology 325, U.S. Geological Survey Professional Paper 1493, 241–261. 48 p. Sanderson,J.S.,N.Rowan,T.Wilding,B.P.Bledsoe, Schmidt, K.-H., D. Morche. 2006. Sediment output and W.J. Miller, N.L. Poff. 2011. Getting to scale with effective discharge in two small high mountain catch- environmental flow assessment: the watershed flow ments in the Bavarian Alps, Germany. Geomorphology evaluation tool. River Research and Applications. 80, 131–145. doi:10.1002/rra.1542 Schmidt, J.C., D.M. Rubin, H. Ikeda. 1993. Flume simu- Sandiford, G. 2009. Transforming an Exotic Species: lation of recirculating flow and sedimentation. Water Nineteenth-Century Narratives About Introduction of Resources Research 29, 2925–2939. 296 References

Schmocker-Fackel, P.,F. Naef. 2010. More frequent flood- U.S.Geological Survey Circular 437, Washington, DC, ing? Changes in flood frequency in Switzerland since 14 p. 1850. Journal of Hydrology 381, 1–8. Schumm, S.A., M.D. Harvey, C.C. Watson. 1984. Incised Schoof, R. 1980. Environmental impact of channel modi- Channels: Morphology, Dynamics and Control.Water fication. Water Resources Bulletin 16, 697–701. Resources Publications, Littleton, CO, 200 p. Schumm, S.A. 1956. Evolution of drainage systems Schumm, S.A., M.P. Mosley, W.E. Weaver. 1987. Exper- and slopes in badlands at Perth Amboy, New Jer- imental Fluvial Geomorphology.JohnWiley&Sons, sey. Geological Society of America Bulletin 67, 597– Chichester. 646. Schumm, S.A., R.S. Parker. 1973. Implications of complex Schumm, S.A. 1960. The shape of alluvial channels in rela- response of drainage systems for Quaternary alluvial tion to sediment type. U.S. Geological Survey Profes- stratigraphy. Science 243, 99–100. sional Paper 352B, 17–30. Schumm, S.A., I.D. Rutherfurd, J. Brooks. 1994. Pre-cutoff Schumm, S.A. 1963. Sinuosity of alluvial rivers on the morphology of the lower Mississippi River. In: S.A. Great Plains. Geological Society of America Bulletin 74, Schumm, B.R. Winkley, eds., The Variability of Large 1089–1100. Alluvial Rivers. ASCE Press, New York, pp. 13–44. Schumm, S.A. 1967. Meander wavelength of alluvial Schuster, P.F., D.P. Krabbenhoft, D.L. Naftz, L.D. Cecil, rivers. Science 157, 1549–1550. M.L. Olson, J.J. DeWild, J.R. Green. 2002. A 270-year Schumm, S.A. 1969. River metamorphosis. Journal of the ice core record of atmospheric mercury deposition to Hydraulics Division, ASCE 95, 255–273. western North America. Environmental Science and Schumm, S.A. 1973. Geomorphic thresholds and complex Technology 36, 2303–2310. response of drainage systems. In: M. Morisawa, ed., Flu- Scott, K.M., G.C. Gravlee. 1968. Flood Surge on the vial geomorphology. SUNY Binghamton Publications in Rubicon River, California – Hydrology, Hydraulics, and Geomorphology, pp. 299–310. Boulder Transport.U.S.GeologicalSurveyProfessional Schumm, S.A. 1977. The fluvial system.JohnWiley& Paper 422-M, 38 p. Sons, New York, 338 p. Sear, D.A., C.E. Millington, D.R. Kitts, R. Jeffries. 2010. Schumm, S.A. 1979. Geomorphic thresholds: the con- Logjam controls on channel:floodplain interactions in cept and its applications. Transactions of the Institute of wooded catchments and their role in the formation British Geographers 4, 485–515. of multi-channel patterns. Geomorphology 116, 305– Schumm, S.A. 1981. Evolution and response of the flu- 319. vial system, sedimentologic implications. Society of Eco- Sear, D.A., J.M. Wheaton, S.E. Darby. 2008. Uncertain nomic Paleontologists and Mineralogists Special Publica- restorationofgravel-bedriversandtheroleofgeo- tion 31, 19–29. morphology. In: H. Habersack, H. Piegay,´ M. Rinaldi, Schumm, S.A. 1985. Patterns of alluvial rivers. Annual eds., Gravel-bed Rivers VI: From Process Understand- Reviews of Earth and Planetary Sciences 13, 5–27. ing to River Restoration. Elsevier, Amsterdam, pp. 739– Schumm, S.A. 1991. To interpret the Earth: Ten Ways 761. to be Wrong. Cambridge University Press, Cambridge, Sedell, J.R., J.L. Froggatt. 1984. [Importance of stream- 133 p. side forests to large rivers: the isolation of the Willamet- Schumm, S.A., J.F. Dumont, J.M. Holbrook. 2000. Active ter River, Oregon, USA, from its floodplain by snag- Tectonics and Alluvial Rivers.CambridgeUniversity ging and streamside forest removal]. Verhandlun- Press, Cambridge. gen des Internationalen Verein Limnologie 22, 1828– Schumm, S. A., V. J. Galay. 1994. The River Nile in Egypt. 1834. In:S.A.Schumm,B.R.Winkley,eds.,The Variability of Segura, C., J. Pitlick. 2010. Scaling frequency of channel- Large Alluvial Rivers. American Society of Civil Engi- forming flows in snowmelt-dominated streams. neers, New York, pp. 75–102. Water Resources Research 46, W06524. doi:10.1029/ Schumm, S.A., R.F. Hadley. 1957. Arroyos and the semi- 2009WR008336 arid cycle of erosion. American Journal of Science 25, Seibert, J., K. Bishop, A. Rodhe, J.J. McDonnell. 2003. 161–174. Groundwaterdynamicsalongahillslope:atestofthe Schumm, S.A., R.F. Hadley. 1961. Progress in the Applica- steady-state hypothesis. Water Resources Research 39, tion of Landform Analysis in Studies of Semiarid Erosion. 1014. doi:10.1029/2002WR001404 References 297

Seidl, M.A., W.E. Dietrich, J.W. Kirchner. 1994. Longi- Larimer and Weld Counties, Colorado. U.S. Geological tudinal profile development into bedrock: an analy- Survey Professional Paper 1115, pp. 87–152. sis of Hawaiian channels. Journal of Geology 102, 457– Sidle,R.C.,S.Noguchi,Y.Tsuboyama,K.Laursen. 474. 2001. A conceptual model of preferential flow systems Seidl, M.A., R.C. Finkel, M.W. Caffee, G.B. Hudson, W.E. in forested hillslopes: evidence of self-organization. Dietrich. 1997. Cosmogenic isotope analyses applied to Hydrological Processes 15, 1675–1692. river longitudinal profile evolution: problems and inter- Sieben, J. 1997. Modeling of Hydraulics and Morphology in pretations. Earth Surface Processes and Landforms 22, Mountain Rivers. PhD dissertation, Technical Univer- 195–209. sity of Delft, The Netherlands, 223 p. Selby, M.J. 1982. Hillslope Materials and Processes.Oxford Simco, A.H., D.B. Stephens, K. Calhoun, D.A. Stephens. University Press, Oxford, 264 p. 2010. Historic irrigation and drainage at Priestley Farm Seminara, G. 2006. Meanders. Journal of Fluid Mechanics by Joseph Elkington and William Smith. Vadose Zone 554, 271–297. Journal 9, 4–13. Seminara, G., M. Tubino. 1989. Alternate bars and mean- Simon, A. 1994. Gradation Processes and Channel Evo- dering: free, forced and mixed interactions. In: S. Ikeda, lution in Modified West Tennessee Streams: Process, G. Parker, eds., River Meandering.AmericanGeophys- Response, and Form. U.S. Geological Survey Profes- ical Union Press, Washington, DC, pp. 267–320. sional Paper 1470, 84 p. SER (Society for Ecological Restoration). 2002. The SER Simon, A., J. Castro. 2003. Measurement and analysis of primer on ecological restoration. www.ser.org. alluvial channel form. In: G.M. Kondolf, H. Piegay,´ eds., Shafroth, P.B., A.C. Wilcox, D.A. Lytle, J.T. Hickey, Tools in Fluvial Geomorphology.JohnWiley&Sons, D.C. Andersen, V.B. Beauchamp, A. Hautzinger, L.E. Chichester, pp. 291–322. McMullen, A. Warner. 2010. Ecosystem effects of envi- Simon, A., A.J.C. Collison. 2001. Pore-water pressure ronmental flows: modeling and experimental floods in effects on the detachment of cohesive streambeds: seep- adrylandriver.Freshwater Biology 55, 68–85. age forces and matric suction. Earth Surface Processes Shankman, D., T.B. Pugh. 1992. Discharge response to and Landforms 26, 1421–1442. channelization of a coastal plain stream. Wetlands 12, Simon, A., A. Curini, S.E. Darby, E. Langendoen. 1999. 157–162. Streambank mechanics and the role of bank and near- Sharma, A., L. Marshall, D. Nott. 2005. A Bayesian view bank processes in incised channels. In: S.E. Darby, A. of rainfall-runoff modeling: alternatives for parame- Simon, eds., Incised River Channels: Processes, Forms, ter estimate, model comparison and hierarchical model Engineering and Management.JohnWiley&Sons, development. In, S. Franks, M. Sivapalan, K. Takeuchi, Chichester, pp. 123–152. Y. Tachikawa, eds., Predictions in Ungauged Basins: Simon, A., A. Curini, S.E. Darby, E. Langendoen. 2000. International Perspectives on the State of the Art and Bank and near bank processes in an incised channel. Pathways Forward. IAHS Publication no. 301, 299–311. Geomorphology 35, 193–217. Shieh, C.-L., Y.-R. Guh, S.-Q. Wang. 2007. The application Simon, A., S.E. Darby. 2002. Effectiveness of grade- of range of variability approach to the assessment of a control structures in reducing erosion along incised check dam on riverine habitat alteration. Environmental riverchannels:thecaseofHotophiaCreek,Mississippi. Geology 52, 427–435. Geomorphology 42, 229–254. Shields, A., 1936. Arwendung der Aenlich-keits-mechanik Simon, A., W. Dickerson, A. Heins. 2004. Suspended- and der Turbulenz-forschung auf die Geshienbebewe- sediment transport rates at the 1.5-year recurrence gung. Versuch-sanstalt fur Wasserbau and Schiffsbau, interval for ecoregions of the United States: transport Berlin, Heft 26, 26 p. Mitteilungen der Preussischen. conditions at the bankfull and effective discharge? Geo- Shreve, R.L. 1966. Statistical law of stream numbers. Jour- morphology 58, 243–262. nal of Geology 74, 17–37. Simon, A., M. Rinaldi. 2006. Disturbance, stream incision, Shroba, R.L., P.W.Schmidt, E.J. Crosby, W.R. Hansen, J.M. and channel evolution: the roles of excess transport Soule. 1979. Geologic and Geomorphologic Effects in the capacity and boundary materials in controlling channel Big Thompson Canyon area, Larimer County. Part B, response. Geomorphology 79, 361–383. of Storm and Flood of July 31-August 1, 1976, in the Simon, A., M. Rinaldi. 2013. Incised channels: distur- Big Thompson River and Cache la Poudre River Basins, bance, evolution and the roles of excess transport 298 References

capacity and boundary materials in controlling channel Sklar, L.S., J. Fadde, J.G. Venditti, P. Nelson, M.A. response. In: E. Wohl, ed., Treatise on Fluvial Geomor- Wydzga, Y. Cui, W.E. Dietrich. 2009. Translation phology. Elsevier, Amsterdam, pp. 574–594. and dispersion of sediment pulses in flume experi- Simons, D.B., E.V.Richardson. 1966. Resistance to Flow in ments simulating gravel augmentation below dams. Alluvial Channels. U.S. Geological Survey Professional Water Resources Research 45, W08439. doi:10.1029/ Paper 422J. 2008WR007346 Simons, D.B., E.V. Richardson, W.H. Haushild. 1963. Slattery, M.C., Bryan, R.B., 1994. Surface seal develop- Some Effects of Fine Sediment on Flow Phenomena.U.S. ment under simulated rainfall on an actively eroding Geological Survey Water-Supply Paper 1498G. surface. Catena 22, 17–34. Simˇ unek,˚ J., Jarvis, N.J., van Genuchten, M.Th., Garden¨ as,¨ Slingerland, R., N.D. Smith. 1998. Necessary conditions A., 2003. Review and comparison of models for describ- for a meandering-river avulsion. Geology 26, 435– ing non-equilibrium and preferential flow and trans- 438. port in the vadose zone. Journal of Hydrology 272, 14– Small, E.E., R.S. Anderson. 1998. Pleistocene relief pro- 35. duction in Laramide mountain ranges, western United Singer, M.B., T. Dunne. 2006. Modeling the influ- States. Geology 26, 123–126. ence of river rehabilitation scenarios on bed mate- Smart, J.S. 1968. Statistical properties of stream lengths. rial sediment flux in a large river over decadal Water Resources Research 4, 1001–1014. timescales. Water Resources Research 42, W12415. Smith, C.R. 1996. Coherent flow structures in smooth- doi:10.1029/2006WR004894 wall turbulent boundary layers: facts, mechanisms and Singh, P., M. Arora, N.K. Goel. 2006. Effect of climate speculation. In: P. Ashworth, S. Bennett, J.L. Best, S. change on runoff of a glacierized Himalayan basin. McLelland, eds., Coherent Flow Structures in Open Hydrological Processes 20, 1979–1992. Channels. John Wiley & Sons, Chichester, pp. 1–39. Singh,P.,G.Spitzbart,H.Hubl,¨ H.W. Weinmeister. 1998. Smith, N. 1971. AHistoryofDams. Peter Davies, London. The role of snowpack in producing floods under heavy Smock, L.A., J.E. Gladden, J.L. Riekenberg, L.C. Smith, rainfall. In: K. Kovar, U. Tappeiner, N.E. Peters, R.G. C.R. Black. 1992. Lotic macroinvertebrate production Craig, eds., Hydrology, Water Resources and Ecology in in three dimensions: channel surface, hyporheic, and Headwaters. IAHS Publication no. 248, pp. 89–95. floodplain environments. Ecology 73, 876–886. Sivapalan,M.,T.Wagener,S.Uhlenbrook,E.Zehe,V.Lak- Snow, R.S., R.L. Slingerland. 1987. Mathematical mod- shmi, X. Liang, Y. Tachikawa, P. Kumar. 2006. Predic- elingofgradedriverprofiles.Journal of Geology tions in Ungauged Basins: Promise and Progress.IAHS 95, 15–33. Publication no. 303. Snyder, N.P., K.X. Whipple, G.E. Tucker, D.J. Merritts. Skalak, K., J. Pizzuto. 2010. The distribution and residence 2000. Landscape response to tectonic forcing: digi- time of suspended sediment stored within the channel talelevationmodelanalysisofstreamprofilesinthe margins of a gravel-bed bedrock river. Earth Surface Mendocino triple junction region, northern Califor- Processes and Landforms 35, 435–446. nia. Geological Society of America Bulletin 112, 1250– Sklar, L.S., W.E. Dietrich. 1998. River longitudinal pro- 1263. files and bedrock incision models: stream power and Sobey, I.J. 1982. Oscillatory flows at intermediate Strouhal the influence of sediment supply. In: K.J. Tinkler, number in asymmetry channels. Journal of Fluid E.E. Wohl, eds., Rivers Over Rock: Fluvial Processes in Mechanics 125: 359–373. Bedrock Channels. American Geophysical Union Press, Soldner,M.,I.Stephen,L.Ramos,R.Angus,N.C.Wells, Washington, DC, pp. 237–260. A. Grosso, M. Crane. 2004. Relationships between Sklar, L.S., W.E. Dietrich. 2004. A mechanistic macroinvertebrates fauna and environmental variables model for river incision into bedrock by saltating in small streams of the Dominican Republic. Water bed load. Water Resources Research 40, W06301. Research 38, 863–874. doi:10.1029/2003WR002496, 21 p. Somoza,L.,A.Barnolas,A.Arasa,A.Maestro,J.G.Rees, Sklar, L.S., W.E. Dietrich, E. Foufoula-Georgiou, B. F.J. Hernandez-Molina. 1998. Architectural stacking Lashermes, D. Bellugi. 2006. Do gravel bed river patterns of the Ebro delta controlled by Holocene high- size distributions record channel network structure? frequency eustatic fluctuations, delta-lobe switching Water Resources Research 42, W06D18. doi:10.1029/ and subsidence processes. Sedimentary Geology 117, 2006WR005035 11–32. References 299

Southard, J.B. 1991. Experimental determination of bed- Stanley, J.-D., F. Goddio, T.F. Jorstad, G. Schnepp. 2004. form stability. Annual Reviews of Earth and Planetary Submergence of ancient Greek cities off Egypt’s Nile Sciences 19, 423–455. delta–acautionarytale.Geological Society of America Southard, J.B., L.A. Boguchwal. 1990. Bed configurations Today 14, 4–10. in steady unidirectional water flows. Part 2. Synthesis Stark, C.P. 2006. A self-regulating model of bedrock of flume data. Journal of Sedimentary Petrology 60, 658– river channel geometry. Geophysical Research Letters 33, 679. L04402. doi:10.1029/2005GL023193 Southerland, W.B., F. Reckendorf. 2010. Performance of Statzner, B., U. Fuchs, L.W.G. Higler. 1996. Sand erosion Engineered Log Jams in Washington State – Post Project by mobile predaceous stream insects: implications for Appraisal. 2nd Joint Federal Interagency Conference on ecology and hydrology. Water Resources Research 32, Sedimentation and Hydrologic Modeling, Las Vegas, 2279–2287. Nevada, June 27–July 1, 2010, 11 p. Statzner, B., O. Peltret. 2006. Assessing potential abi- Spaliviero, M. 2003. Historic fluvial development of the otic and biotic complications of crayfish-induced travel Alpine-foreland Tagliamento River, Italy, and conse- transport in experimental streams. Geomorphology 74, quences for floodplain management. Geomorphology 245–256. 52, 317–333. Statzner, B., P. Sagnes. 2008. Crayfish and fish as biotur- Spencer, T., I. Douglas, T. Greer, W. Sinun. 1990. Vege- bators of streambed sediments: assessing joint effects tation and fluvial geomorphic processes in South-east of species with different mechanistic abilities. Geomor- Asian Tropical rainforests. In: J.B. Tornes, ed., Vege- phology 93, 267–287. tation and Erosion: Processes and Environments.John Stedinger, J.R., V.R. Baker. 1987. Surface water hydrology: Wiley & Sons, Chichester, pp. 451–469. historical and paleoflood information. Reviews of Geo- Spence, C., M.K. Woo. 2003. Hydrology of subarctic physics 25, 119–124. Canadian Shiled: bedrock upland. Journal of Hydrology Stephens, D.B., D.A. Stephens. 2006. British land 262, 111–127. drainers: their place among pre-Darcy forefathers of Spence, C., M.K. Woo. 2006. Hydrology of subarctic applied hydrogeology. Hydrogeology Journal 14, 1367– Canadian Shield: soil-filled valleys. Journal of Hydrol- 1376. ogy 279, 151–166. Sternai, P., F. Herman, M.R. Fox, S. Catelltort. 2011. Hyp- Springer, G.S. 2004. A pipe-based, first approach to mod- sometric analysis to identify spatially variable glacial eling closed conduit flow in caves. Journal of Hydrology erosion. Journal of Geophysical Research 116, F03001. 284, 178–189. doi:10.1029/2010JF001823 Springer, G.S., E.E. Wohl. 2002. Empirical and theoreti- Sternberg, H. 1875. Untersuchungen uber¨ Langen-¨ und cal investigations of sculpted forms in Buckeye Creek Querprofile geschiebfuhrender¨ Flusse.¨ Zeitschriftur f¨ Cave, West Virginia. Journal of Geology 110, 469– Bauwesen 25, 483–506. 481. Stock, J.D. 2013. Waters divided: a history of alluvial fan Springer, G.S., E.E. Wohl, J.A. Foster, D.G. Boyer. 2003. research and a view of its future. In: E. Wohl, ed., Trea- Testing for reach-scale adjustments of hydraulic vari- tise on Fluvial Geomorphology. Elsevier, Amsterdam, ables to soluble and insoluble strata: Buckeye Creek pp. 414–457. and Greenbrier River, West Virginia. Geomorphology Stock, J.D., W.E. Dietrich. 2003. Valley incision by 56, 201–217. debris flows: evidence of a topographic signature. Stalnaker,C.,B.L.Lamb,J.Henriksen,K.Bovee,J. Water Resources Research 39, 1089. doi:10.1029/ Bartholow. 1995. The Instream Flow Incremental 2001WR001057 Methodology: A Primer for IFIM. National Biological Stock, J.D., W.E. Dietrich. 2006. Erosion of steepland val- Service, US Department of the Interior, Biological leys by debris flows. Geological Society of America Bul- Report no. 29, Fort Collins, CO. letin 118, 1125–1148. Stanford, J.A., J.V. Ward. 1988. The hyporheic habitat of Stock, J.D., K.M. Schmidt, D.M. Miller. 2007. Controls on river ecosystems. Nature 335, 64–66. alluvial fan long-profiles. Geological Society of America Stanley, E.H., M.W. Doyle. 2003. Trading off: the ecologi- Bulletin 120, 619–640. cal effects of dam removal. Frontiers in Ecology and the Stoddard, J.L., D.P. Larsen, C.P. Hawkins, R.K. John- Environment 1, 15–22. son, R.H. Norris. 2006. Setting expectations for the 300 References

ecological condition of streams: the concept of refer- parison of the Fly and Strickland rivers, Papua New ence condition. Ecological Applications 16, 1267–1276. Guinea. Journal of Geophysical Research 11, F01S03. Stoughton, J.A., W.A. Marcus. 2000. Persistent impacts doi:10.1029/2006JF000623 of trace metals from mining on floodplain grass Syvitski, J.P.M., A.J. Kettner, I. Overeem, E.W.H. Hutton, communities along Soda Butte Creek, Yellowstone M.T.Hannon,G.R.Brakenridge,J.Day,C.Vor¨ osmarty,¨ National Park. Environmental Management 25, 305– Y. Saito, L. Giosan, R.J. Nicholls. 2009. Sinking deltas 320. due to human activities. Nature Geoscience 2, 681–686. Strahler, A.N. 1952. Hypsometric (area-altitude) analysis Syvitski, J.P.M., Y. Saito. 2007. Morphodynamics of deltas of erosional topography. Geological Society of America under the influence of humans. Global and Planetary Bulletin 63, 1117–1142. Change 57, 261–282. Strahler, A.N. 1957. Quantitative analysis of watershed Syvitski, J.P.M., C.J. Vor¨ osmarty,¨ A.J. Kettner, P. Green. geomorphology. Transactions American Geophysical 2005. Impact of humans on the flux of terrestrial sedi- Union 38, 913–920. ment to the global coastal ocean. Science 308, 376–380. Strahler, A.N. 1964. Quantitative geomorphology of Tague,C.,M.Farrell,G.Grant,J.Choate,A.Jefferson. drainage basins and channel networks. In: V.T. Chow, 2008. Deep groundwater mediates streamflow response ed., Handbook of Applied Hydrology. McGraw Hill, New to climate warming in the Oregon Cascades. Climate York, 4–74. Change 86, 1–2. Strecker, M.R., G.E. Hilley, J.R. Arrowsmith, I. Cout. 2003. Tague, C. 2009. Assessing climate change impacts on Differential structural and geomorphic mountain-front alpine stream-flow and vegetation water use: min- evolution in an active continental collision zone: the ing the linkages with subsurface hydrologic processes. NorthwestPamir,southernKyrgyzstan.Geological Soci- Hydrological Processes 23, 1815–1819. ety of America Bulletin 115, 166–181. Tal, M., K. Gran, A.B. Murray, C. Paola, D.M. Hicks. 2004. Surian, N. 1999. Channel changes due to river regulation: Riparian vegetation as a primary control on channel thecaseofthePiaveRiver,Italy.Earth Surface Processes characteristics in multi-thread rivers. In: S.J. Bennett, A. and Landforms 24, 1135–1151. Simon, eds., Riparian Vegetation and Fluvial Geomor- Surian, N. 2002. Downstream variation in grain size along phology. American Geophysical Union Press, Washing- an Alpine river: analysis of controls and processes. Geo- ton, DC, pp. 43–58. morphology 43, 137–149. Tal, M., C. Paola. 2007. Dynamic single-thread channels Sutfin, N.A. 2013. A Geomorphic Classification of maintained by the interaction of flow and vegetation. Ephemeral Streams in Arid Regions. Unpublished MS Geology 35, 347–350. thesis, Colorado State University, Fort Collins, CO. Tamura, T., K. Horaguchi, Y. Saito, N. Van Lap, M. Sutherland, R.A. 1991. [Caesium-137 and sediment Tateishi, K.O.T. Thi, F. Nanayama, K. Watanabe. 2010. budgeting within a partially closed drainage basin]. Monsoon-influenced variations in morphology and Zeitschrift fur Geomorphologie 35, 47–63. sediment of a mesotidal beach on the Mekong River Swales, S., A.W. Storey, I.D. Roderick, B.S. Figa, K.A. delta coast. Geomorphology 116, 11–23. Bakowa. C.D. Tenakanai. 1998. Biological monitoring Tamura, T., Y. Saito, V.L. Nguyen, T.K.O. Ta, M.D. Bate- oftheimpactsoftheOkTedicoppermineonfishpopu- man, D. Matsumoto, S. Yamashita. 2012. Origin and lations in the Fly River system, Papua New Guinea. The evolution of interdistributary delta plains; insights from Science of the Total Environment 214, 99–111. Mekong River delta. Geology 40, 303–306. Swanson, F.J., L.E. Benda, S.H. Duncan, G.E. Grant, Tarolli, P., G. Dalla Fontana. 2009. Hillslope-to-valley W.F. Megahan, L.M. Reid, R.R. Ziemer. 1987. Mass transition morphology: new opportunities from high failures and other processes of sediment production resolution DTMs. Geomorphology 113, 47–56. in Pacific Northwest forest landscapes. In: E.O. Salo, Taylor, M.P., R.G.H. Kesterton. 2002. Heavy metal con- T.W. Cundy, eds., Streamside Management: Forestry and tamination of an arid river environment: Gruben River, Fishery Implications.UniversityofWashington,Insti- Namibia. Geomorphology 42, 311–327. tute of Forest Resources, Seattle, pp. 9–38. Taylor, S., X. Feng, J.W.Kirchner, R. Osterhuber, B. Klaue, Swanson, K.M., E. Watson, R. Aalto, J.W.Lauer, M.T.Bera, C.E. Renshaw. 2001. Isotopic evolution of a seasonal A. Marshall, M.P.Taylor, S.C. Apte, W.EDietrich. 2008. snowpack and its melt. Water Resources Research 37, Sediment load and floodplain deposition rates: com- 759–769. References 301

Tharme, R.E. 2003. A global perspective on environmen- Thompson, D.M., E.E. Wohl, R.D. Jarrett. 1999. Veloc- tal flow assessment: emerging trends in the develop- ity reversals and sediment sorting in pools and riffles ment and application of environmental flow method- controlled by channel constrictions. Geomorphology 27, ologies for rivers. River Research and Applications 19, 229–241. 397–441. Thurman, E.M. 1985. Organic Geochemistry of Natural Thom, R. 1975. Structural Stability and Morphogenesis: An Waters.MartinusNijhoff/DrWJunk,Dordrecht,The Outline of a General Theory of Models.W.A.Benjamin, Netherlands, 497 p. Reading, PA, 348 p. Tockner, K., F. Malard, J.V. Ward. 2000. An extension Thomas, M.F., J. Nott, A.S. Murray, D.M. Price. 2007. Flu- of the flood-pulse concept. Hydrological Processes 14, vial response to late Quaternary climate change in NE 2861–2883. Queensland, Australia. Palaeogeography, Palaeoclima- Tockner, K., F. Schiemer, C. Baumgartner, G. Kum, E, tology, Palaeoecology 251, 119–136. Weigand, I. Zweimuller,¨ J.V. Ward. 1999. The Danube Thompson, A. 1986. Secondary flows and the pool-riffle restoration project: species diversity patterns across unit: a case study of the processes of meander develop- connectivity gradients in the floodplain system. Regu- ment. Earth Surface Processes and Landforms 11, 631– lated Rivers: Research and Management 15, 245–258. 641. Tockner, K., F. Schiemer, J.V. Ward. 1998. Conservation Thompson, D.M. 2001. Random controls on semi- by restoration: the management concept for a river- rhythmic spacing of pools and riffle in constriction- floodplainsystemontheDanubeRiverinAustria. dominated rivers. Earth Surface Processes and Land- Aquatic Conservation 8, 71–86. forms 26, 1195–1212. Tonina, D., J.M. Buffington. 2009. Hyporheic exchange Thompson, D.M. 2002. Long-term effect of instream in mountain rivers I: mechanics and environmen- habitat-improvement structures on channel morphol- tal effects. Geography Compass 3. 10.111/j.1749– ogy along the Blackledge and Salmon Rivers, Connecti- 8198.2009.00226.x cut, USA. Environmental Management 29, 250–265. Tooth, S. 1999. Floodouts in central Australia. In: A.J. Thompson, D.M. 2006. Did the pre-1980 use of in-stream Miller, A. Gupta, eds., Varieties of Fluvial Form.John structures improve streams? A reanalysis of historical Wiley & Sons, Chichester, pp. 219–247. data. Ecological Applications 16, 784–796. Tooth, S. 2000. Process, form, and change in dry- Thompson, D.M. 2008. The influence of lee sediment land rivers: a review of recent research. Earth-Science behind large bed elements on bedload transport rates in Reviews 51, 67–107. supply-limited channels. Geomorphology 99, 420–432. Tooth, S. 2013. Dryland fluvial environments: assessing Thompson, D.M. 2012. The challenge of modeling pool- distinctiveness and diversity from a global perspective. riffle morphologies in channels with different densities In: E. Wohl, ed., Treatise on Fluvial Geomorphology. of large woody debris and boulders. Earth Surface Pro- Elsevier, Amsterdam, pp. 613–643. cesses and Landforms 37, 223–239. Tooth, S., T.S. McCarthy. 2004. Anabranching in mixed Thompson, D.M. 2013. Pool-riffle. In: E. Wohl, ed., Trea- bedrock-alluvial rivers: the example of the Orange tise on Fluvial Geomorphology. Elsevier, Amsterdam, River above Augrabies Falls, Northern Cape Province, pp. 364–378. South Africa. Geomorphology 57, 235–262. Thompson, D.M., J.M. Nelson, E.E. Wohl. 1998. Inter- Torres, R., P. Mouginis-Mark, S. Self, H. Garbeil, K. actions between pool geometry and hydraulics. Water Kallianpur, R. Quiambao. 2004. Monitoring the evolu- Resources Research 34, 3673–3681. tion of the Pasig-Potrero alluvial fan, Pinatubo Volcano, Thompson, D.M., G.N. Stull. 2002. [The development and using a decade of remote sensing data. Journal of Vol- historic use of habitat structures in channel restoration canology and Geothermal Research 138, 371–392. in the United States: the grand experiment in fisheries Tooth, S., G.C. Nanson. 2000. Equilibrium and nonequi- management]. GeographiephysiqueetQuaternaire´ 56, librium conditions in dryland rivers. Physical Geogra- 45–60. phy 21, 183–211. Thompson, D.M., E.E. Wohl, R.D. Jarrett. 1996. A revised Toth, L.A., J.T.B. Obeysekera, W.A. Perkins, M.K. Loftin. velocity-reversal and sediment-sorting model for a 1993. Flow regulation and restoration of Florida’s high-gradient, pool-riffle stream. Physical Geography Kissimmee River. Regulated Rivers: Research and Man- 17, 142–156. agement 8, 155–166. 302 References

Trayler, C.R., E.E. Wohl. 2000. Seasonal changes in bed Mayer,D.vanDam,R.W.Howarth,K.J.Nadelhof- elevationinastep-poolchannel,RockyMountains, fer, M. Eve, G. Billen. 2002. Where did all the nitro- Colorado, USA. Arctic, Antarctic and Alpine Research gen go? Fate of nitrogen inputs to large watersheds in 32, 95–103. the northeastern U.S.A. Biogeochemistry 57/58, 267– Trimble, S.W. 1983. A sediment budget for Coon Creek 293. basin in the Driftless Area, Wisconsin, 1853–1977. Van Cleve, K., C.T. Dyrness, G.M. Marion, R. Erick- American Journal of Science 283, 454–474. son. 1993. Control of soil development on the Tanana Trimble, S.W., A.C. Mendel. 1995. The cow as a geomor- River floodplain, interior Alaska. Canadian Journal of phic agent – a critical review. Geomorphology 13, 233– Forestry Research 23, 941–955. 253. Van De Wiel, M.J., S.E. Darby. 2007. A new model to Triska, F.J. 1984. [Role of wood debris in modifying chan- analyse the impact of woody riparian vegetation on the nel geomorphology and riparian areas of a large low- geotechnical stability of riverbanks. Earth Surface Pro- land river under pristine conditions: a historical case cesses and Landforms 32, 2185–2198. study]. Verhandlungen des Internationalen Verein Lim- Van den Berg, J.H. 1995. Prediction of alluvial channel nologie 22, 1876–1892. pattern of perennial rivers. Geomorphology 12, 259– Trumbore, S.E., C.I. Czimczik. 2008. An uncertain future 279. for soil carbon. Science 321, 1455–1456. Van den Berg, M.W., T. van Hoof. 2001. The Maas Tsai, H., Z.-Y.Hseu, W.-S.Huang, Z.-S. Chen. 2007. Pedo- Terrace sequence at Maastricht, SE Netherlands: evi- genicapproachtoresolvingthegeomorphicevolution dence for 200 m of late Neogene and Quater- ofthePakuaRiverterracesincentralTaiwan.Geomor- nary surface uplift. In: D. Maddy, M.G. Macklin, phology 83, 14–28. J.C. Woodward, eds., River Basin Sediment Systems: Tucker, G.E., L. Arnold, R.L. Bras, H. Flores, E. Istanbullu- Archives of Environmental Change.A.A.Balkema,Lisse, oglu,P.Solyom.´ 2006. Headwater channel dynamics in pp. 45–86. semiarid rangelands, Colorado high plains, USA. Geo- Vandenberghe, J., D. Maddy. 2000. The significance of logical Society of America Bulletin 118, 959–974. fluvial archives in geomorphology. Geomorphology 33, Turner, R.E. 1997. Wetland loss in the northern Gulf of 127–130. Mexico: multiple working hypotheses. Estuaries 20, 1– Vandenberghe, J., M.-K. Woo. 2002. Modern and ancient 13. periglacial river types. Progress in Physical Geography Turowski, J.M., D. Rickenmann. 2009. Tools and 26, 479–506. cover effects in bedload transport observations in VanDerSchrier,G.,D.Efthymiadis,K.R.Briffa,P.D. the Pitzbach, Austria. Earth Surface Processes and Jones. 2007. European Alpine moisture variability for Landforms 34, 26–37. 1800–2003. International Journal of Climatology 27, Uchida,T.,N.Ohte,A.Kimoto,T.Mizuyama,L. 415–427. Changhua. 2000. Sediment yield on a devastated hill in Van Der Wal, D., K. Pye. 2003. The use of historical bathy- southern China: effects of microbiotic crust on surface metric charts in a GIS to assess morphological change erosion process. Geomorphology 32, 129–145. in estuaries. The Hydrographic Journal 110, 3–9. Uchida, T., K. Kosugi, T. Mizuyama. 2001. Effects of Van Nieuwenhuyse, E.E., J.D. LaPerriere. 1986. Effects pipeflow on hydrological processes and its relation of placer gold mining on primary production in sub- to landslide: a review of pipeflow studies in forested arctic streams of Alaska. Water Resources Bulletin 22, headwater catchments. Hydrological Processes 15, 2151– 91–99. 2174. Vannote, R.L., G.W.Minshall, K.W.Cummins, J.R. Sedell, Valla, P.G., P.A. van der Beek, D. Lague. 2010. Fluvial C.E. Cushing. 1980. The river continuum concept. incision into bedrock: insights from morphome- Canadian Journal of Fisheries and Aquatic Sciences 37, tric analysis and numerical modeling of gorges 130–137. incising glacial hanging valleys (Western Alps, Veblen, T.T., J.A. Donnegan. 2005. Historical range of vari- France). Journal of Geophysical Research 115, F02010. ability for forest vegetation of the national forests of doi:10.1029/2008JF001079 the Colorado Front Range.Finalreport,USDAForest Van Breemen, N., E.W. Boyer, C.L. Goodale, N.A. Service Agreement 1102–0001–99–033, Golden, CO, Jaworski,K.Paustian,S.P.Seitzinger,K.Lajtha,B. 151 p. References 303

Venditti, J.G. 2013. Bedforms in sand-bedded rivers. In: E. Walling, D.E., P.N. Owens, I.D.L. Foster, J.A. Lees. 2003. Wohl, ed., Treatise on Fluvial Geomorphology. Elsevier, Changes in the fine sediment dynamics of the Ouse and Amsterdam, pp. 138–162. Tweed basins in the UK over the last 100–150 years. Venditti, J.G., M. Church, S.J. Bennett. 2005. Bed form Hydrological Processes 17, 3245–3269. initiation from a flat sand bed. Journal of Geophysical Walling, D.E., P.N. Owens, G.J.L. Leeks. 1999. Finger- Research 110, F01009. doi:10.1029/2004JF000149 printing suspended sediment sources in the catchment Vericat, D., R.J. Batalla, C. Garcia. 2006. Breakup and oftheRiverOuse,Yorkshire,UK.Hydrological Processes reestablishment of the armour layer in a large gravel- 13, 955–975. bed river below dams: the lower Ebro. Geomorphology Walling, D.E., B.W. Webb. 1986. Solutes in river systems. 76 122–136. In: S.T. Trudgill, ed., Solute Processes.JohnWiley& Vileisis, A. 1997. Discovering the Unknown Landscape: A Sons, Chichester, pp. 251–327. History of America’s Wetlands.IslandPress,Washing- Walter, R.C., D.J. Merritts. 2008. Natural streams and ton, DC, 421 p. the legacy of water-powered mills. Science 319, 299– Viseras, C., M.L. Calvache, J.M. Soria, J. Fernandez. 2003. 304. Differential features of alluvial fans controlled by tec- Wang, G., H. Ma, J. Qian, J. Chang. 2004. Impact of land tonic or eustatic accommodation space; examples from use changes on soil carbon, nitrogen and phospho- the Beltic Cordillera, Spain. Geomorphology 50, 181– rus and water pollution in an arid region of northwest 202. China. Soil Use Management 20, 32–39. Vivoni, E.R., R.S. Bowman, R.L. Wyckoff, R.T. Wang, Z., Tuli, A., Jury, W.A., 2003. Unstable flow during Jakubowski, K.E. Richards. 2006. Analysis of a redistribution in homogeneous soil. Vadose Zone Jour- monsoon flood event in an ephemeral tributary and nal 2, 52–60. its downstream hydrologic effects. Water Resources Wang, C., W.Wang, S. He, J. Du, Z. Sun. 2011. Sources and Research 46, W02509. doi:10.1029/2005WR008240 distribution of aliphatic and polycyclic aromatic hydro- Voller, V.R., C. Paola. 2010. Can anomalous diffu- carbons in Yellow River Delta Nature Reserve, China. sion describe depositional fluvial profiles? Journal Applied Geochemistry 26, 1330–1336. of Geophysical Research 115, FA0013. doi:10.1029/ Wang, H., T.J. Wright, Y. Yu, H. Lin, L. Jiang, C.Li, G. 2009JF001278 Qiu. 2012. InSTAR reveals coastal subsidence in the Vor¨ osmarty,C.,D.Lettenmaier,C.Leveque,M.Meybeck,¨ Pearl River delta, China. Geophysical Journal Interna- C. Pahl-Wostl, J. Alcamo, W. Cosgrove, H. Grassl, H. tional 191, 1119–1128. Hoff,P.Kabat,F.Lansigan,R.Lawford,R.Naiman. Warburton, J. 1992. Observations of bed load transport 2004. Humans transforming the global water system. and channel bed changes in a proglacial mountain Transactions American Geophysical Union 84(48), 509– stream. Arctic and Alpine Research 24, 195–203. 520. Ward, J.V. 1997. An expansive perspective of riverine Wainwright, J, Parsons, A.J., 2002. The effect of tempo- landscapes: pattern and process across scales. River ral variations in rainfall on scale dependency in runoff Ecosystems 6, 52–60. coefficients. Water Resources Research 38(12), 1271. Ward, J.V., J.A. Stanford. 1983. The serial discontinuity doi:10.1029/2000WR000188 concept of lotic ecosystems. In: T.D. Fontaine, S.M. Wainwright, J., L. Turnbull, T.G. Ibrahim, I. Lexarta- Bartell, eds., Dynamics of Lotic Ecosystems.AnnArbor Artza, S.F. Thornton, R.E. Brazier. 2011. Linking envi- Science, Ann Arbor, MI, pp. 29–42. ronmental regimes,´ space and time: interpretations of Ward, J.V., J.A. Stanford. 1995. The serial discontinu- structural and functional connectivity. Geomorphology ity concept: extending the model to floodplain rivers. 126, 387–404. Regulated Rivers: Research and Management 10, 159– Walcott, R.C., M.A. Summerfield. 2008. Scale depen- 168. dence of hypsometric integrals: an analysis of Ward, J.V., K. Tockner, U. Uehlinger, F. Malard. 2001. southeast African basins. Geomorphology 96, 174– Understanding natural patterns and processes in river 186. corridors as the basis for effective river restoration. Reg- Walker, J., L. Arnborg, J. Peippo. 1987. [Riverbank erosion ulated Rivers: Research and Management 17, 311–323. in the Colville Delta, Alaska]. Geografiska Annaler 69A, Warburton, J., M. Evans. 2011. Geomorphic, sedimentary, 61–70. and potential palaeoenvironmental significance of peat 304 References

blocks in alluvial river systems. Geomorphology 130, riparian area. Water Resources Research 42, W06404. 101–114. doi:10.1029/2005WR004560 Warne, A.G., L.A. Toth, W.A. White. 2000. Drainage- Westbrook, C.J., D.J. Cooper, B.W. Baker. 2011. Beaver basin-scale geomorphic analysis to determine reference assisted river valley formation. River Research and conditions for ecologic restoration – Kissimmee River, Applications 27, 247–256. Florida. Geological Society of America Bulletin 112, 884– Westerling, A.L., H.G. Hidalgo, D.R. Cayan, R.W. Swet- 899. nam. 2006. Warmer and earlier spring increase in west- Warrick, J.A., L.A.K. Mertes. 2009. Sediment yield from ern U.S. forest wildfire activity. Science 313, 940–943. tectonically active semiarid Western Transverse Ranges Westra, L., P. Mioller, J.R. Karr, W.E. Rees, R.E. Ulanow- of California. Geological Society of America Bulletin 121, icz. 2000. Ecological integrity: integrating environmen- 1054–1070. tal conservation and health. In: D. Pimentel, L. Westra, Waters, J.V., S.J. Jones, H.A. Armstrong. 2010. Climatic R.F. Noss, eds., Ecological integrity and the Aims of the controls on late Pleistocene alluvial fans, Cyprus. Geo- Global Integrity Project.IslandPress,Washington,DC, morphology 115, 228–251. pp. 19–44. Waters, J.M., M. Shirley, G.P. Closs. 2002. Hydroelec- Whipple, K.X. 2001. Fluvial landscape response time: how tric development and translocation of Galaxias bre- plausible is steady-state denudation? American Journal vipinnis:acloudattheendofthetunnel?Cana- of Science 301, 313–325. dian Journal of Fisheries and Aquatic Sciences 59, Whipple, K.X. 2004. Bedrock rivers and the geomorphol- 49–56. ogy of active orogens. Annual Review of Earth and Plan- Wathen, S.J., T.B. Hoey. 1998. Morphological controls on etary Sciences 32, 151–185. the downstream passage of a sediment wave in a gravel- Whipple, K.X., R.A. DiBiase, B.T. Crosby. 2013. Bedrock bed stream. Earth Surface Processes and Landforms 23, rivers. In: E. Wohl, ed., Treatise on Fluvial Geomorphol- 715–730. ogy. Elsevier, Amsterdam, pp. 550–572. Webb, R.H., V.R. Baker. 1987. Changes in hydrologic con- Whipple, K.X., T. Dunne. 1992. The influence of debris- ditions related to large floods on the Escalante River, flow rheology on fan morphology, Owens Valley, Cali- south-central Utah. In: V.P. Singh, ed., Regional Flood fornia. Geological Society of America Bulletin 104, 887– Frequency Analysis. D. Reidel, Dordrecht, The Nether- 900. lands, pp. 309–323. Whipple, K.X., G.S. Hancock, R.S. Anderson. 2000. River Webb, R.H., J.L. Betancourt. 1990. Climatic Variability incision into bedrock: mechanics and relative efficacy and Flood Frequency of the Santa Cruz River, Pima of plucking, abrasion, and cavitation. Geological Society County, Arizona. U.S. Geological Survey Open-File of America Bulletin 112, 490–503. Report 90–553, Tucson, Arizona, 69 p. Whipple, K.X., C.R. Trayler. 1996. Tectonic control of Webb, R.H., R.D. Jarrett. 2002. One-dimensional estima- fan size: the importance of spatially variable subsidence tion techniques for discharges of paleofloods and his- rates. Basin Research 8, 351–366. torical floods. In: P.K. House, R.H. Webb, V.R. Baker, Whitaker, A., Y. Alila, J. Beckers, D. Toews. 2002. D.R. Levish, eds., Ancient Floods, Modern Hazards: Evaluating peak flow sensitivity to clear-cutting in Principles and Applications of Paleoflood Hydrology. different elevation bands of a snowmelt-dominated American Geophysical Union Press, Washington, DC, mountainous catchment. Water Resources Research 38. pp. 111–125. doi:10.1029/WR000514 Webb, R.H., P.T. Pringle, G.R. Rink. 1987. Debris Flows White, W.B. 1999. Karst hydrology: recent developments from Tributaries of the Colorado River, Grand Canyon and open questions. In: B.F. Beck, A.J. Pettit, J.G. National Park, Arizona.U.S.GeologicalSurveyOpen- Herring, eds., Hydrogeology and Engineering Geology File Report 87–118, 64 p. of Sinkholes and Karst. A.A. Balkema, Rotterdam, Werner, B.T.,D.E. McNamara. 2007. Dynamics of coupled pp. 3–20. human-landscape systems. Geomorphology 91, 393– White, W.R., R. Bettess, E. Paris. 1982. Analytical 407. approach to river regime. Journal of Hydraulics Divi- Westbrook, C.J., D.J. Cooper, B.W. Baker. 2006. Beaver sion, ASCE 108, 1179–1193. dams and overbank floods influence groundwater- White, J.Q., G.B. Pasternack, H.J. Moir. 2010. Valley width surface water interactions of a Rocky Mountain variation influences riffle-pool location and persistence References 305

on a rapidly incising gravel-bed river. Geomorphology resistance. Water Resources Research 42, W05419. 121, 206–221. doi:10.1029/2005WR004278, 14 p. Whiting, P.J., W.E.Dietrich, L.B. Leopold, T.G.Drake, R.L. Wilcox, A.C., E.E. Wohl. 2007. Field measurements of Shreve. 1988. Bedload sheets in heterogeneous sedi- three-dimensional hydraulics in a step-pool channel. ment. Geology 16, 105–108. Geomorphology 83, 215–231. Whittaker, A.C., M. Attal, P.A. Cowie, G.E. Tucker, G. Wilcox, A.C., E.E. Wohl, F. Comiti, L. Mao. 2011. Roberts. 2008. Decoding temporal and spatial patterns Hydraulics, morphology, and energy dissipation in an of fault uplift using transient river long profiles. Geo- alpine step-pool channel. Water Resources Research 47, morphology 100, 506–526. W07514. doi:10.1029/2010WR010192, 17 p. Whittaker,A.C.,P.A.Cowie,M.Attal,G.E.Tucker,G.P. Willenbring, J.K., A.T. Codilean, B. McElroy. 2013. Earth Roberts. 2007a. Bedrock channel adjustment to tec- is (mostly) flat: apportionment of the flux of continental tonic forcing: implications for predicting river incision sediment over millennial time scales. Geology 41, 343– rates. Geology 35, 103–106. 346. Whittaker,A.C.,P.A.Cowie,M.Attal,G.E.Tucker,G.P. Willgoose, G., G. Hancock. 1998. Revisiting the hypso- Roberts. 2007b. Contrasting transient and steady-state metric curve as an indicator of form and process in rivers crossing active normal faults: new field observa- transport-limited catchment. Earth Surface Processes tions from the Central Apennines, Italy. Basin Research and Landforms 23, 611–623. 19, 529–556. Williams, G.P. 1978a. Bank-full discharge of rivers. Water Wiberg, P.L., Smith, J. D., 1987. Calculations of the critical Resources Research 14, 1141–1154. shear-stress for motion of uniform and heterogeneous Williams, G.P. 1978b. The case of the shrinking chan- sediments. Water Resources Research 23(8), 1471– nels–theNorthPlatteandthePlatteRiversin 1480. Nebraska. U.S. Geological Survey Circular 781, Wiele, S.M., J.B. Graf, J.D. Smith. 1996. Sand deposition in Arlington, VA. the Colorado River in the Grand Canyon from flooding Williams, G.P. 1983. [Paleohydrological methods and of the Little Colorado River. Water Resources Research some examples from Swedish fluvial environments]. 32, 3579–3596. Geografiska Annaler 65A, 227–244. Wiele, S.M., P.R. Wilcock, P.E. Grams. 2007. Reach- Williams, G.P.1989. Sediment concentration versus water averaged sediment routing model of a canyon river. discharge during single hydrologic events in rivers. Water Resources Research 43, W02425. doi:10.1029/ Journal of Hydrology 111, 89–106. 2005WR004824 Williams, M.W., J.M. Melack. 1991. Solute chemistry of Wilby, R.L. 1995. Greenhouse hydrology. Progress in Phys- snowmelt and runoff in an alpine basin, Sierra Nevada. ical Geography 19, 351–369. Water Resources Research 27, 1575–1588. Wilcock, D.N. 1971. Investigation into the relations Williams, G.P., M.G. Wolman. 1984. Effects of Dams between bedload transport and channel shape. and Reservoirs on Surface-Water Hydrology; Changes in Geological Society of America Bulletin 82, 2159– Rivers Downstream from Dams.U.S.GeologicalSurvey 2176. Professional Paper 1286, 83 p. Wilcock, P.R. 1997. Entrainment, displacement, and Willis, I.C., N.S. Arnold, B.W.Brock. 2002. Effect of snow- transport of tracer gravels. Water Resources Research 22, pack removal on energy balance, melt and runoff in a 1125–1138. small supraglacial catchment. Hydrologic Processes 16, Wilcock, P.R., S.T. Kenworthy. 2002. A two-fraction 2721–2749. model for the transport of sand/gravel mixtures. Winsemann, J., C. Brandes, U. Polom. 2011. Response Water Resources Research 38, 1194. doi:10.1029/ of a proglacial delta to rapid high-amplitude lake- 2001WR000684 level change: an integration of outcrop data and high- Wilcock, P.R., B.W. McArdell. 1997. Partial transport of resolution shear wave seismics. Basin Research 23, a sand/gravel sediment. Water Resources Research 33, 22–52. 235–245. Wishart, D., J. Warburton, L. Bracken. 2008. Gravel Wilcox, A.C., J.M. Nelson, E.E. Wohl. 2006. Flow extraction and planform change in a wandering gravel- resistance dynamics in step-pool channels: 2. Par- bed river: The River Wear, northern England. Geomor- titioning between grain, spill, and woody debris phology 94, 131–152. 306 References

Wittmann, H., F. von Blanckenburg, L. Maurice, J.L. United States. In: A. Simon, S.J. Bennett, J.M. Castro, Guyot, P.W. Kubik. 2011. Recycling of Amazon flood- eds., Stream Restoration in Dynamic Fluvial Systems: plain sediment quantified by cosmogenic 26Al and 10Be. Scientific Approaches, Analyses, and Tools.American Geology 39, 467–470. Geophysical Union Press, Washington, DC, pp. 399– Wohl, E.E. 1992. Bedrock benches and boulder bars: 418. floods in the Burdekin Gorge of Australia. Geological Wohl, E. 2011c. Threshold-induced complex behavior of Society of America Bulletin 104, 770–778. wood in mountain streams. Geology 39, 587–590. Wohl, E.E. 1998. Bedrock channel morphology in relation Wohl, E. 2011d. What should these rivers look like? His- to erosional processes. In: K.J. Tinkler, E.E. Wohl, eds., torical range of variability and human impacts in the Rivers Over Rock: Fluvial Processes in Bedrock Channels. Colorado Front Range, USA. Earth Surface Processes American Geophysical Union Press, Washington, DC, and Landforms 36, 1378–1390. pp. 133–151. Wohl, E. 2013a. Migration of channel heads following Wohl, E.E. 2000a. Anthropogenic impacts on flood haz- wildfire in the Colorado Front Range, USA. Earth Sur- ards. In: E.E. Wohl, ed., Inland Flood Hazards: Human, face Processes and Landforms. Riparian, and Aquatic Communities. Cambridge Uni- Wohl, E. 2013b. Floodplains and wood. Earth-Science versity Press, pp. 104–141. Reviews 123, 194–212. Wohl, E. 2000b. Substrate influence on step-pool Wohl, E. 2013c. Wilderness is dead: whither critical sequences in the Christopher Creek drainage, Arizona. zone studies and geomorphology in the Anthropocene? Journal of Geology 108, 121–129. Anthropocene.InPress. Wohl, E. 2001. Virtual rivers: Lessons from the Colorado Wohl, E., P.L. Angermeier, B. Bledsoe, G.M. Kondolf, Front Range. Yale University Press, New Haven, CT. L. MacDonnell, D.M. Merritt, M.A. Palmer, N.L Poff, Wohl, E. 2004a. Disconnected Rivers: Linking Rivers to D.Tarboton. 2005. River restoration. Water Resources Landscapes. Yale University Press, New Haven, CT, Research 41, W10301. doi:10.1029/2005WR003985 301 p. Wohl,E.,A.Barros,N.Brunsell,N.A.Chappell,M.Coe, Wohl, E. 2004b. Limits of downstream hydraulic geome- T. Giambelluca, S. Goldsmith, R. Harmon, J.M.H. Hen- try. Geology 32, 897–900. drickx, J. Juvik, J. McDonnell, F. Ogden. 2012a. The Wohl, E. 2005. Downstream hydraulic geometry along a hydrology of the humid tropics. Nature Climate Change tropical mountain river. In: R.S. Harmon, ed., The Rio 2, 655–662. Chagres, Panama: A Multidisciplinary Profile of a Trop- Wohl, E., N.D. Beckman. 2014 Leaky rivers: implications ical Watershed. Springer, Dordrecht, The Netherlands, of the loss of longitudinal fluvial disconnectivity in pp. 169–188. headwater streams. Geomorphology 205, 27–35. Wohl, E. 2006. Human impacts to mountain streams. Geo- Wohl,E.,S.Bolton,D.Cadol,F.Comiti,J.R.Goode, morphology 76, 217–248. L. Mao. 2012b. A two end-member model of wood Wohl, E. 2008a. The effect of bedrock jointing on the for- dynamics in headwater neotropical rivers. Journal of mation of straths in the Cache la Poudre River drainage, Hydrology 462–463, 67–76. Colorado Front Range. Journal of Geophysical Research Wohl, E., D. Cadol. 2011. Neighborhood matters: patterns 113, F01007. doi:10.1029/2007JF000817 and controls on wood distribution in old-growth forest Wohl, E. 2008b. Review of effects of large floods in streams of the Colorado Front Range, USA. Geomor- resistant-boundary channels. In: H. Habersack, H. phology 125, 132–146. Piegay,´ M. Rinaldi, eds., Gravel-bed Rivers VI: From Wohl, E.E., D.A. Cenderelli. 1998. Flooding in the Process Understanding to River Restoration. Elsevier, Himalaya Mountains. In: V.S. Kale, ed., Flood studies in Amsterdam, pp. 181–212. India. Memoir Geological Society of India 41, pp. 77– Wohl, E. 2010. Mountain Rivers Revisited. American Geo- 99. physical Union Press, Washington, DC, 573 p. Wohl, E.E., D.A. Cenderelli, M. Mejia-Navarro. 2001. Wohl, E. 2011a. A World Of Rivers: Environmental Change Channel change from extreme floods in canyon rivers. on Ten of the World’s Great Rivers.UniversityofChicago In,D.J.Anthony,M.D.Harvey,J.B.Laronne,M.P. Press, Chicago, IL, 359 p. Mosley, eds., Applying Geomorphology to Environmen- Wohl, E. 2011b. Seeing the forest and the trees: wood tal Management. Water Resources Publications, Little- in stream restoration in the Colorado Front Range, ton, CO, pp. 149–174. References 307

Wohl,E.,D.Cooper,L.Poff,F.Rahel,D.Staley,D. Wohl, E., F.Ogden, J. Goode. 2009. Episodic wood loading Winters. 2007. Assessment of stream ecosystem func- in a mountainous neotropical watershed. Geomorphol- tion and sensitivity in the Bighorn National For- ogy 111, 149–159. est, Wyoming. Environmental Management 40, 284– Wohl, E., M. Palmer, G.M. Kondolf. 2008. River manage- 302. ment in the United States. In: G.J. Brierley, K.A. Fryirs, Wohl, E., G.C.L. David. 2008. Consistency of scal- eds., River Futures: An Integrative Scientific Approach to ing relations among bedrock and alluvial chan- River Repair. Island Press, Washington, DC, pp. 174– nels. Journal of Geophysical Research 113, F04013. 200. doi:10.1029/2008JF000989, 16 p. Wohl, E.E., P.A. Pearthree. 1991. Debris flows as geomor- Wohl, E., D. Dust. 2012. Geomorphic response of a head- phic agents in the Huachuca Mountains of southeastern water channel to augmented flow. Geomorphology 138, Arizona. Geomorphology 4, 273–292. 329–338. Wohl, E., G. Springer. 2005. Bedrock channel incision Wohl, E., K. Dwire, N. Sutfin, L. Polvi, R. Bazan. along the Upper Rio Chagres basin, Panama. In: R.S. 2012c. Mechanisms of carbon storage in mountain- Harmon, ed., The Rio Chagres, Panama: A Multidisci- ous headwater rivers. Nature Communications 3, 1263. plinary Profile of a Tropical Watershed.Springer,Dor- doi:10.1038/ncomms2274 drecht, The Netherlands, pp. 189–209. Wohl, E.E., Y. Enzel. 1995. Data for palaeohydrology. In: Wohl, E.E., D.M. Thompson, A.J. Miller. 1999. Canyons K.J. Gregory, L. Starkel, V.R. Baker, eds., Global Conti- with undulating walls. Geological Society of America nental Palaeohydrology. John Wiley & Sons, Chichester, Bulletin 111, 949–959. pp. 23–59. Wolman, M.G. 1954. A method of sampling coarse river- Wohl, E., J.R. Goode. 2008. Wood dynamics in head- bed material. Transactionsof the American Geophysical water streams of the Colorado Rocky Mountains. Union 35, 951–956. Water Resources Research 44, W09429. doi:10.1029/ Wolman, M.G. 1967a. [A cycle of sedimentation and ero- 2007WR006522 sion in urban river channels]. Geografiska Annaler 49A, Wohl, E.E., N. Greenbaum, A.P. Schick, V.R. Baker. 1994. 385–395. Controls on bedrock channel incision along Nahal Wolman, M.G. 1967b. Two Problems Involving River Paran, Israel. Earth Surface Processes and Landforms 19, Channels and their Background Observations.North- 1–13. western University Studies in Geography 14, pp. 67– Wohl, E.E., H. Ikeda. 1998. Experimental simulation of 107. channel incision into a cohesive substrate at varying Wolman, M.G., R. Gerson. 1978. Relative scales of time gradient. Geology 25, 295–298. and effectiveness of climate in watershed geomorphol- Wohl, E., K. Jaeger. 2009. A conceptual model for the lon- ogy. Earth Surface Processes 3, 189–208. gitudinal distribution of wood in mountain streams. Wolman, M.G., L.B. Leopold. 1957. River Flood Plains: Earth Surface Processes and Landforms 34, 329–344. Some Observations on Their Formation. U.S. Geological Wohl, E., J.N. Kuzma, N.E. Brown. 2004. Reach-scale Survey Professional Paper 282-C. channel geometry of a mountain river. Earth Surface Wolman, M.G., J.P. Miller. 1960. Magnitude and fre- Processes and Landforms 29, 969–981. quency of forces in geomorphic processes. Journal of Wohl, E.E., C.J. Legleiter. 2003. Controls on pool charac- Geology 68, 54–74. teristics along a resistant-boundary channel. Journal of Womack, W.R., S.A. Schumm. 1977. Terraces of Douglas Geology 111, 103–114. Creek, northwestern Colorado: an example of episodic Wohl, E.E., D.M. Merritt. 2001. Bedrock channel mor- erosion. Geology 5, 72–76. phology. Geological Society of America Bulletin 113, Wondzell, S.M., F.J. Swanson. 1996. Seasonal and storm 1205–1212. dynamics of the hyporheic zone of a fourth-order Wohl, E., D.M. Merritt. 2008. Reach-scale channel geom- mountain stream. I. Hydrologic processes. Journal etry of mountain streams. Geomorphology 93, 168–185. of the North American Benthological Society 15, Wohl, E., F.L. Ogden. 2013. Wood and carbon export 3–19. during an extreme tropical storm, Upper Rio Chagres, Wong,M.,G.Parker,P.DeVries,T.M.Brown,S.J. Panama. Earth Surface Processes and Landforms 38, Burges. 2007. Experiments on dispersion of tracer 1407–1416. stones under lower-regime plane-bed equilibrium bed 308 References

load transport. Water Resources Research 43, W03440. Yager, E.M., W.E. Dietrich, J.W. Kirchner, B.W. McArdell. doi:10.1029/2006WR005172 2012. Prediction of sediment transport in step-pool Woo, M.-K. 2010. Cold ocean seas and northern hydrol- channels. Water Resources Research 48, W01541. ogy: an exploratory overview. Hydrology Research 41, doi:10.1029/2011WR010829 439–453. Yager, E.M., H.E. Schott. 2013. The initiation of motion Woo, M.-K., W.R. Rouse. 2008. MAGS contribution to and formation of armor layers. In: E. Wohl, ed., Trea- hydrologic and surface process research. In: M. Woo, tise on Fluvial Geomorphology. Elsevier, Amsterdam, ed., Cold Region Atmospheric and Hydrologic Studies. pp. 87–102. Springer-Verlag, Berlin, pp. 9–38. Yang, C.T. 1976. Minimum unit stream power and fluvial Woo, M.-K., R. Thorne. 2003. Streamflow in the Macken- hydraulics. Journal of Hydraulics Division, ASCE 102, zie Basin, Canada. Arctic 56, 328–340. 919–934. Woods, S.W., L.H. MacDonald, C.J. Westbrook. 2006. Yang, S.L., J.D. Milliman, P. Li, K. Xu. 2011. 50,000 dams Hydrologic interactions between an alluvial fan and a later: erosion of the Yangtze River and its delta. Global slope wetland in the central Rocky Mountains, USA. and Planetary Change 75, 14–20. Wetlands 26, 230–243. Yanites,B.J.,G.E.Tucker,K.J.Mueller,Y.-G.Chen,T. Worner,¨ G., R.S. Harmon, G. Hartmann, K. Simon. 2005. Wilcox, S.-Y. Huang, K.-W. Shi. 2010. Incision and Igneous geology and geochemistry of igneous rocks of channel morphology across active structures along the theUpperRioChagres,Basin.In:R.S.Harmon,ed.,The Peikang River, central Taiwan: implications for the Rio Chagres, Panama: A Multidisciplinary Profile of a importance of channel widths. Geological Society of Tropical Watershed. Springer, Dordrecht, The Nether- America Bulletin 122, 1192–1208. lands, pp. 65–81. Yanosky, T.M., R.D. Jarrett. 2002. Dendrochronologic Wright, S.A., J.C. Schmidt, T.S. Melis, D.J. Topping, D.M. evidence for the frequency and magnitude of pale- Rubin. 2008. Is there enough sand? Evaluating the fate ofloods.In:P.K.House,R.H.Webb,V.R.Baker, of Grand Canyon sandbars. Geological Society of Amer- D.R. Levish, eds., Ancient Floods, Modern Hazards: ica Today 18(8), 4–10. Principles and Applications of Paleoflood Hydrology. Wyatt, A.M., S.W. Franks. 2006. The Multi-model American Geophysical Union Press, Washington, DC, approach to rainfall-runoff modeling. In: M. Sivapalan, pp. 77–89. T.Wagner,S.Uhlenbrook,X.Liang,V.Lakshmi, Yao, S. 1943. The geographic distribution of floods and P. Kumar, E. Zehe, Y. Tachikawa, eds., Predictions droughts in Chinese history 206 B.C.–1911 A.D. Far in Ungauged Basins: Promise and Progress.IAHS Eastern Quarterly 2, 357–378. Publication no. 303, pp. 134–144. Yaru, B.T., R.T. Buckney. 2000. Metal interactions in con- Wynn, T.M., M.B. Henderson, D.H. Vaughan. 2008. taminated freshwater sediments from the Fly River Changes in streambank erodibility and critical shear floodplain, Papua New Guinea. International Journal of stress due to subaerial processes along a headwater Environmental Studies 57, 305–331. stream, southwestern Virginia, USA. Geomorphology Yetemen, O., E. Istanbulluoglu, E.R. Vivoni. 2010. The 97, 260–273. implications of geology, soils, and vegetation on land- Wyzga,´ B. 1991. Present-day downcutting of the Raba scape morphology: influences from semi-arid basins River channel (Western Carpathians, Poland) and its with complex vegetation patterns in central New Mex- environmental effects. Catena 18, 551–556. ico, USA. Geomorphology 116, 246–263. Wyzga,´ B. 2001. Impact of the channelization-induced Yochum, S.E., B.P. Bledsoe, G.C.L. David, E. Wohl. 2012. incision of the Skawa and Wisoka Rivers, southern Velocity prediction in high-gradient channels. Journal Poland, on the conditions of overbank deposition. of Hydrology 424–425, 84–98. Regulated Rivers: Research and Management 17, 85– Young, M.K., D. Haire, M.A. Bozek. 1994. The effect and 100. extent of railroad tie drives in streams of southeastern Xu, J. 1996. Underlying gravel layers in a large sand bed Wyoming. Western Journal of Applied Forestry 9, 125– river and their influence on downstream-dam channel 130. adjustment. Geomorphology 17, 351–359. Zeitler, P.K., A.S. Meltzer, P.O. Koons, D. Craw, B. Hallet, Xu, J. 2004. A study of anthropogenic seasonal rivers in C.P. Chamberlain, W.S.F. Kidd, S.K. Park, L. See- China. Catena 55, 17–32. ber, M. Bishop, J. Shroder. 2001. Erosion, Himalayan References 309

geodynamics, and the geomorphology of metamor- Zimmermann, A., M. Church, M.A. Hassan. 2010. phism. Geological Society of America Today 11(1), 4–8. Step-pool stability: testing the jammed state hypoth- Zernitz, E.R. 1932. Drainage patterns and their signifi- esis. Journal of Geophysical Research 115, F02008. cance. Journal of Geology 40, 498–521. doi:10.1029/2009JF001365 Zhang, W., X. Ruan, J. Zheng, Y. Zhu, H. Wu. 2010. Long Zolezzi, G., A. Bellin, M.C. Bruno, B Maiolini, A. Siviglia. term change in tidal dynamics and its cause in the Pearl 2009. Assessing hydrological alterations at multiple River delta, China. Geomorphology 120, 209–223. temporal scales: Adige River, Italy. Water Resources Zimmermann, A.E. 2013. Step-pool channel features. In: Research 45, W12421. 10.1029/2008WR007266 E. Wohl, ed., Treatise on Fluvial Geomorphology.Else- Zorn, M., K. Natek, B. Komac. 2006. [Mass movements vier, Amsterdam, 347–363. andflash-floodsintheSloveneAlpsandsurround- Zimmermann, A., M. Church. 2001. Channel morphol- ing mountains]. Studia Geomorphologica Carpatho- ogy, gradient profiles and bed stresses during flood in a Balcanica 40, 127–145. step-pool channel. Geomorphology 40, 311–327.

Index

Abrasion, 90, 103, 119, 144, 156, 160, see also Macroabrasion bed configuration, 105, 157, 237 Alluvial channels, 58, 130, 133, 135, 144, 150, 153, 154, 159, dunes, 57, 63, 64, 90, 98, 102, 105–108, 115, 124, 135, 172, 178, 211 140 coarse bed, 41–42, 58, 99, 100, 230, see also Boulder-bed; hysteresis, 108, 236 Cobble-bed; Gravel-bed; Threshold channel infrequently mobile, 83, 108–114 fine bed, 42, 230 lower flow regime, 106 Alternate bar, 99, 109, see also Bars, alternate microforms, mesoforms, macroforms, 105 Amazon River, 4, 97, 166, 170, 173, 174, 230 particle clusters, 108–109, 115 Anabranching, 7, 125, 134, 139, 140–143, 146, 147, 157, 158, pool-riffle, 57, 58, 102, 108–110, 112–114, 128, 136 171, 214 forced pool-riffle, 112 Anastomosing, 116, 143, see also Anabranching velocity reversal hypothesis, 113 Anthropocene, 191–192, 224 readily mobile, 105–108, 110 Antidunes, 98, 105, 106, 108, 109, 124, 135, 157, 183, see also ripples, 63, 64, 90, 98, 105–108, 115, 135 Bedforms step-pool, 57, 58, 100, 102, 105, 108, 110–112, 113, 114, Aquifer, 28, 29, 144, see also Ground water 125 Armor, 99, see also Censored layer; Pavement jammed state, 111 Arroyo, 14, 236, see also Incised channels keystones, 111, 218 Avulsion, 140, 142, 143, 167, 171, 172, 175, 181, 183, 186, transverse ribs, 108, 109 187, 190, 212, 234, 246, see also Braided channels; upper flow regime, 106 Delta; Floodplain wavelength, 112, 113 Bedload, 11–13, 15, 42, 57, 126, 129, 141–144, 156, 158, 160, Banks 161, 183, 193, 206, 209, 236, see also Sediment, hydraulic action, 118 transport mass failure, 118–120, 124 Bedrock rivers, 9, 101, 126, 226 pore water, 91, 117, 118, 120, 124, 191 Berm, 114, 158, 168, 193, see also Bars, berm Bank stabilization, 2, 78, 103, 127, 183, 197, 210, 215, 216, Bifurcation ratio, 38, 40 see also Human impacts to rivers, direct Bioturbation, 22, 23, 30, 89, 182 Bankfull discharge, 125–127, 129, 151 Boulder-bed, 11, 12, 59, 60, 80, 126, 157, see also Alluvial Bars channels, coarse bed alternate, 99, 102, 109, 114, 135, 136, 138, 140 Boundary layer, 55, 61, 65, 105, 107, 108 berm, 114 Braided channels, 114, 127, 134, 139–142, 144, 146, 147, point, 114, 118, 135–139, 169, 172, 173, 175 172, 210, 234, see also Sandur Base level avulsion frequency, 142 eustatic base level change, 178 Burst, 64, 65, 89, 100, 108, 160, see also Turbulence local, 151, 155, 174, 178, 211, 212, 232, 239, 240 relative base level change, 151 Carbon, 4, 23, 92–94, 193, 231, 244, 246, see also Organic ultimate, 151 matter Beaver, 1, 18, 92, 115, 146, 147, 170, 208, 212, 232, 240, 241, Cascade channels, 110, 114 see also Ecosystem engineers Cavitation, 90 Bedforms Censored layer, 99, see also Armor; Pavement amplitude, 58, 59, 110, 112, 113, 130, 131, 139, 140 Channel evolution model, 133, 134, 211 antidunes, 98, 105, 106, 108, 124, 157 Channel head, 8, 17, 34–37, 151

Rivers in the Landscape: Science and Management, First Edition. Ellen Wohl. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Companion Website: www.wiley.com/go/wohl/riverslandscape

311 312 Index

Channel maintenance flow, 197, 222 Davis, W.M., 150, 178, 204, 228 Channelization, 2, 5, 78, 79, 103, 122, 133, 167, 183, 191, Debris flow, 2, 10, 18, 31, 40–43, 109, 114, 116, 147, 148, 197, 203, 210, 211, 250–252, see also Human 153, 160, 162, 174, 176, 178, 181, 182, 185, 190, impacts to rivers, direct 201, 217, 230, 239, 242, 247, 248, see also Chezy equation, 53, see also Flow resistance Landslide; Mass movement Climate change, 72, 116, 122, 198–200, 217, 242, see also Deforestation, 15, 18, 30, 198, 200–204, 241, see also Human Human impacts to rivers, indirect impacts to rivers, indirect Cobble-bed, 11, 14, 183, see also Alluvial channels, coarse Delta, 7, 8, 44, 93, 143, 165, 168, 169, 181, 185–194, 233, bed 245, 246, 251, 252 Coherent flow structure, 64, 108, see also Turbulence; avulsion, 140, 142, 143, 167, 171, 172, 175, 181, 183, 186, Turbulent flow; Vortex 187, 190, 212, 234, 246 Cohesion bottomset beds, 190 on hillslopes, 41, 98 delta front, 187, 188, 190 on stream banks, 91 delta plains, 188–190 on stream beds, 91 delta switching, 187 Colorado Front Range, 9–19, 34, 229, 240 distributary channels, 168, 181, 186–188, 190, 246 Complex response, 132–133, 178, 179, 205 foreset beds, 190 Complex system, 9, 13, 198, see also Nonlinear system foreset-dominated/Gilbert-type deltas, 190 Compound channels, 143 high-constructive/river-dominated deltas, 188 Confluence high-destructive deltas concordant, 148 tide-dominated, 188 discordant, 148 wave-dominated, 188 Connectivity homopycnal, hyperpycnal and hypopycnal flow, biological, 2, 220, 232 188 functional, 2 subdelta, 187 hydrologic, 1, 2, 26, 75 subsidence, 186, 188, 191, 192, 216, 233, 251 landscape, 2, 232 tidal prism, 186, 190, 192 lateral, 2, 175, 209, 211, 251, 252 topset beds, 190 longitudinal, 2, 165, 251, 252 topset-dominated deltas, 190 river/riverine, 2, 75, 252 turbidity current, 190 sediment, 2, 41, 124, 233, 236 Deterministic approaches, 11, 70 structural, 2 Diffusion equation, 15 vertical, 6, 209, 215 Digital elevation model (DEM), 39, 226 Continuity equation, 11, 12, 15, 47, 61, 68, 128, 130 Dimensionless critical shear stress, 86, 88, 89 Corrosion, 90 Dissolved load, 91–94, 123, see also Sediment, transport Creep, 22, 30, 43, 121, see also Mass movement Diversion, 78, 183, 191, 197, 208, 213, 217, see also Human Critical flow, 51–53, 157, 183, see also Subcritical and impacts to rivers, direct supercritical flow Dominant discharge, 127, 129, 222 Critical length of slope, 31 Downstream fining Critical shear stress, 86, 88, 89, see also Dimensionless downstream coarsening, 43 critical shear stress Sternberg’s Law, 160 Critical zone, 7, 224, 250 Drainage density, 38–40, 70, 237, 247 Crops, 3, 7, 23, 186, 200, 203, 204, see also Human impacts Drainage piracy, 226 to river, indirect Dredging, 2, 78, 103, 197, 208, 210, see also Human impacts Cumulative sediment transport curve, 75 to rivers, direct Cycle of erosion, 178, 228 Dryland, 5, 9, 27, 35, 39, 75, 225, 234, 236, 237 Dunes, 57, 63, 64, 90, 98, 102, 105–108, 115, 124, 135, 140, Dam, 3, 4, 7, 18, 68, 78, 79, 90, 92, 96, 99, 103, 115, 116, 146, see also Bedforms 147, 167, 170, 171, 174, 183, 191, 197, 205–212, 214, 215, 218, 221–223, 231–234, 239, 240, 242, Ecological integrity, 75, 222, 223, 242 247, 251, 252, see also Human impacts to rivers, Ecosystem engineers, 1, see also Beaver direct Effective discharge, 95, 102, 125, 127 Danube River, 174, 210, 218, 226, 252 Element threshold concept, 25, 26, 245 Darcy-Weisbach equation, 53, see also Flow resistance Empirical approaches, 11 Index 313

Energy Floodout, 181 equation, 52 Floodplain grade line/energy line, 49, 110, 112 alluvial ridges, 169, 172 specific, 51, 52 avulsion, 172 specific energy curve, 52 backswamp/flood basin, 167, 169 Entrainment, 11, 12, 15, 30, 31, 57, 63, 67, 85–90, 100–104, braid plain, 169 106, 108, 109, 111, 115, 120, 123, 124, 160, 162, channel belt landforms, 167 236, 237, see also Incipient motion fining upward sequence, 172 Environmental flows, 197, 221–223, see also Natural flow floodplain large-wood cycle, 170, 171 regime floodplain turnover, 170, 172–174 Ephemeral flow, 74, 75, 204 lateral accretion, 123, 167, 169–172 Equal mobility, 88 levee (natural), 32, 166, 168, 186 Equifinality, 14, 15, 205 perirheic zone, 166 Equilibrium, 15–17, 31, 104, 130, 157, 229 scroll-bar topography, 169, 171 disequlibrium, 153, 154, 175, 237 splays, 158, 169, 172, 181, 187, 188 dynamic equilibrium, 16, 175, 224, 226, 237 vertical accretion, 97, 167–172, 174, 175, quasi-equilibrium, 40 252 Ergodic reasoning, 17 Flow duration curve, 75, 76, 121 Estuary, 123 Flow pulse, 5, 76 closed estuary, 193 Flow regulation, 5, 10, 18, 47, 77–79, 103, 116, 130, 146, microtidal, 192 167, 197, 205–209, 221, 241, 242, 250–252, normally open estuary, 193 see also Human impacts to rivers, tide-dominated estuary, 193, 194 direct wave-dominated estuary, 193, 194 Flow resistance, see also Chezy equation; Darcy–Weisbach Eulerian framework, 9 equation; Manning equation Evaporation, 14, 23, 26, 34, 75, 118, 244, 249, 250 bank, 236 Exotic organisms, 2 form, 57–59, 111, 129 Extremal hypotheses, 40, 41, 138, 157 grain, 57–59, 129 partitioning, 59, 60 Fan spill, 57 alluvial, 2, 8, 121, 122, 165, 181–185, 211, 229, 232, 249 Flux equation, 11, 15 alluvial apron/bajada, 181 Force equation, 12, 15 debris-flow, 2, 41, 42, 247 Forces fan-delta, 181 buoyant, 188 fanhead trench, 182 drag, 87, 105 megafan, 181 driving, 53, 87, 88, 91, 117, 130, 145, 163 paraglacial, 181 lift, 87, 89, 90 Feedback resisting, 87, 88, 105, 117 self-arresting, 14 weight, 87, 88, 105 self-enhancing, 14, 116 Froude number, 51, 53, 106, 144, 145, 157 Flood annual maximum series, 71 Gaining stream, 5 frequency analysis, 71–72 Geochronology, 178 megaflood, 239 cosmogenic isotopes, 180 outburst, 30, 43, 106, 170, 182, 183, 239 luminescence, 180, 191 partial duration series, 71 radiocarbon, 191, 217 peak, 5, 70, 72, 78, 95, 116, 146, 147, 167, 203, 208, 209 Geomorphic integrity, 215 pulse, 5, 75, 76, 156 Geomorphic transport laws, 18, 19 rainfall, 2, 21, 29, 30, 35, 39, 70, 71, 74, 76, 77, 123, 191, Gilbert, G.K., 13, 15, 113, 150, 151, 178, 190, 213, 226, 228, 228, 235, 236, 239, 241, 243, 244, 248, 250 229 snowmelt, 5, 6, 10–12, 14, 71, 72, 74, 76, 77, 78, 91–93, Graded river, 15, 176 126, 159, 170, 187, 198, 199, 202, 206, 217, Gradual diffusive processes, 29, 30 239–242, 246, 250, 251 Gradually varied flow, 48, 50–52, see also Rapidly varied stationarity, 72, 104 flow; Uniform flow 314 Index

Grain crops, 3, 7, 23, 186, 200, 203, 204 density, 85 deforestation, 15, 18, 30, 198, 200–204, 241 exposure, 89 grazing, 119, 128, 133, 147, 200, 202–203, 213 friction angle, 87–89 land drainage, 203, 204 grain-grain interactions, 98, 104, 106 upland mining, 203–204 kinematics, 103 urbanization, 96, 183, 200, 203, 241 packing, 85, 86, 89 Hydraulic protrusion, 86–89, 101, 105, 129 conductivity, 6, 26 shape, 85 geometry, 16, 149, 154 shielding, 86, 104 at-a-station, 125, 128–129, 130 sorting, 84–86, 101, 102, 160 downstream, 16, 129–130, 243 Grand Canyon (Colorado River), 40, 147–149, 181, 182, jump, 51–53, 57, 109, 111, 114 208, 218, 222 radius, 50, 53–55, 66, 68, 125 Gravel-bed, 9, 80, 84, 98, 108, 116, 125–127, 139, 151, 169, Hydroclimatology, 68, 74, 240 183, 214, 248, see also Alluvial channels, coarse Hydrograph bed; Threshold channel annual, 73, 80, 222 Grazing, 119, 128, 133, 147, 200, 202–203, 213, see also base flow, 73–75, 125 Human impacts to rivers, direct, indirect basin lag, 74 Ground water, 6, 21, 22, 23, 25, 26, 28–31, 78, 92, 123, 126, direct runoff, 73, 74 209, see also Aquifer flashy, 29, 74, 158, 200, 235 flood/event, 73 Hack, J.T., 38, 150, 151, 153, 226 geomorphic unit, 28, 231 Hanging valley, 149, 155, see also Knickpoint peak, 71–74, 77, 78, 95, 96, 99, 122, 127, 146, 166, 170, Headcut, 14, 35, 36, 212 173, 179, 187, 199, 202, 203, 205–207, 209, 232, Helical flow, 136, 137, 147, see also Meandering 233, 235, 241, 243, 245, 246 Hierarchical patch dynamics, 231 rainfall-dominated, 74 Historical range of variability, 217 recession/falling limb, 69, 73, 91, 98, 102, 113, 118, 236 Horton, R.E., 24, 27, 31, 35, 38 rising limb, 5, 69, 73, 74, 91, 98, 102, 236 Hot moment, 3, 92 snowmelt-dominated, 73, 91, 202 Hot spot, 3, 33, 35, 92, 155 unit, 73 Human impacts to rivers, 241 Hydrophobicity, 24 direct, 78, 79, 192, 197 Hyperconcentrated flow, 42, 181, 182 bank stabilization, 2, 78, 79, 103, 127, 183, 197, 210, Hyporheic zone, 4–6, 77, 91, 92 215, 216 Hypsometric curve/integral, 229 channelization, 2, 5, 78, 79, 103, 122, 133, 167, 183, 191, 195, 197, 203, 210, 211, 250, 251, 252 Ice check dams, 209–212 anchor, 77, 101, 245 dredging, 2, 78, 79, 103, 197, 208, 210 break up, 187, 234, 235, 245, 246 flow regulation, 5, 10, 18, 47, 77, 78, 79, 103, 116, 130, frazil, 101 146, 167, 195, 197, 205–208, 221, 241 freeze over, 77, 78, 119, 234, 235, 245, 246 dams, 2, 3, 4, 7, 18, 68, 78, 79, 90, 92, 96, 99, 103, Incipient motion, 85, 109, see Entrainment 115, 116, 146, 147, 167, 170, 171, 174, 183, 191, Incised channels, 35, 121, 133, 147, 182, see Arroyo 197, 205–212, 214, 215, 218, 221–223, 231–234, Incision, 3, 14, 15, 17–19, 30, 31, 35, 36, 39, 98, 103, 112, 132, 239, 240, 242, 247, 251, 252 133, 139, 142, 153–156, 158, 175–183, 203, 206, diversions, 2, 18, 78, 79, 208, 221, 232 209, 211, 214, 226, 228, 230, 238, 244, 248, 249 index of hydrologic alteration, 206 Indirect discharge estimation, 69 milldams, 7, 214, 239, 240 botanical, 69, 70, 216 instream mining, 197, 213 competence, 69, 70, 141, 206 levees (artificial), 158, 166–169, 194, 197, 209, 215 historical, 69, 70, 72, 97, 123, 163, 171, 201, 210, 211, 214, training, 210, 211, 249 216, 217, 219, 224, 241, 242 indirect, 8, 69, 71, 79, 80, 191, 194, 197, 198, 214 paleostage indicator, 49, 69 afforestation, 200, 202 regime-based, 69 climate change, 72, 116, 122, 195, 198–200, 217, 242 Inequality, 1–3 commercial recreational development, 201 Infiltration capacity, 5, 24, 26, 34, 203 Index 315

Instream flow, 197, 222 Matric suction, 91, 117, 120 Instream wood, 13, 17, 18, 59, 89, 99, 102, 103, 109, 115, Meandering, 7, 112, 114, 125, 134, 136–146, 167, 178, 187, 137, 141, 160–163, 171, 174, 201, 207, 208, 212, 213, 214, 234, see also Helical flow 218, 231, 235, 237, 240–243, 247, 248 chute cutoff, 139, 140 alternative stable states, 162, 224 meander geometry, 114, 136–137 congested transport, 163 meander migration, 139 forced alluvial reaches, 161 neck cutoff, 139 logjam, 13, 14, 17, 57, 92, 102, 115, 116, 141, 143, 163, Megaflood, 239, see also Flood 170, 171, 174, 212, 232, 241, 244, 248 Milldams, 7, 214, 239, 240, see also Human impacts to log rafts, 163, 171, 212 rivers, direct Interception, 23, 183, 201 Mining Intermittent flow, 34, 74, 75, 107 lode, 18, see also Human impacts to rivers, indirect placer, 176, 197, 209, 213, 214, 241, see also Human Karst, 28, 75, 144 impacts to rivers, direct Knickpoint, 3, 30, 132, 142, 153–156, 158, 179, 213, 240, Mississippi River, 6, 23, 97, 122, 169, 185, 187–189, 221, 226, 243, see also Hanging valley 238, 240, 251, 252 Knickzone, 154, 160, 240 Mixing length, 49, 149 Mobile bed, 63, 83, 105, 129, 136, 138, 157, see also Alluvial Lagrangian framework, 9 channels Lag time, 14, 15, 74, 179, 205 Morphometric indices, 37–40, 43, 184, see also Bifurcation Laminar flow, 50, see also Reynolds number; Turbulent flow ratio; Drainage density; Length ratio; Relief ratio; Laminar sublayer, 50, 55, 61 Stream order Landslide, 4, 30, 102, 155, 178, 181, 224, 247, 248, see also Debris flow; Mass movement Natural flow regime, 76, 77, 80, 221, see also Environmental Lane’s balance, 130–132, 157 flows Law of divides, 151 Nitrates, 6, 33, 34, 91, see also Nitrogen Law of the wall, 55, 56, 61 Nitrogen, 6, 33, 34, 92, 93, 194, 199, 250, 252, see also Legacy sediments, 93, 214 Nitrates Length ratio, 38 Nonlinear system, 13, see also Complex system Leopold, Luna, 1, 9, 14, 16–18, 40, 57, 61, 98, 113, 126, 128, Numerical models/numerical simulations, 13, 20, 8, 37, 71, 129, 133, 134, 136–138, 141, 144, 145, 170 152, 172, 178, 190, 220, 247 Levees, artificial, 158, 166–169, 194, 197, 209, 215, see also Human impacts to rivers, direct Old water paradox, 27 LiDAR, 2, 8, 181, see also Remote sensing Open system, 14 Link, see also Stream order Optimal channel networks, 40 exterior, 39 Oregon Coast Range, 246–248 interior, 39 Organic matter, 2, 5, 15, 32, 34, 65, 76, 91–94, 107, 116, 161, Logjam, 13, 14, 57, 141, 143, see also Instream wood 168, 169, 180, 193, 209, 212, 233, see also Carbon Longitudinal profile, 14, 17, 18, 105, 112, 134, 149, 150–156, Outburst flood, 2, 5, 15, 32, 34, 65, 76, 91–94, 107, 116, 161, 179, 180, 227, 237, 247 168, 169, 180, 193, 209, 212, 233, see also Flood concavity index, 153 Overland flow DS index, 153 Hortonian, 24, 27, 31, 35, 36 index of profile concavity, 151, 153 infiltration excess, 24 stream gradient (SL) index, 153–154 saturation, 6, 24, 25, 27, 35, 36, 243 Losing stream, 5 Paired watershed approach, 205 Mackenzie River, 187, 244–246 Paleostage indicator, 49, 69, see also Indirect discharge Macroabrasion, 90, see also Abrasion estimation Macroinvertebrate, 4, 5, 89 Particle clusters, 108–109, 115, 124, see also Bedforms Macropore, 5, 26–29, 235 Pavement, 25, 99, 249, 250, see also Armor; Censored layer Manning equation, 53, 55, 68, 69, see also Flow resistance Perennial flow, 34, 74, 75, 204 Mass movement, 19, 22, 29, 30, 32, 34, 41, 43, 122, 148, 161, Permafrost, 28, 93, 119, 244–246 162, 178, 183, 184, 199, 201, see also Creep; Debris Physical experiments, 8, 12, 13, 20, 37, 178, 187, 190 flow; Landslide Physical integrity, 224, 237, see also Geomorphic integrity 316 Index

Piping, 10, 27, 35, 36, 119, 169, see also Soil pipe vegetation, 10, 15, 18, 34, 43, 59, 69, 78, 79, 116–119, 124, Plane bed 127, 128, 141, 143, 146, 147, 170, 206, 207, 210, coarse grained, 110 212, 213, 215, 220, 222, 234, 236, 237, 249, 250 sand, 110 zone, 28, 43, 32–34, 92, 118, 175, 216, 231, 251 Point bar, 114, 136–139, 169, 172, 173, 175, see also Bars, Ripples, 63, 64, 90, 98, 105–108, 115, 124, 135, see also point Bedforms Pool-riffle, 57, 58, 100, 102, 108–110, 112–114, 125, 128, River continuum concept, 231 136, see also Bedforms River health, 198, 223–224 Pothole, 70, 90, see also Sculpted forms River metamorphosis, 146–147 Power function, 16, 37, 39, 129, 153 River styles, 175, 205, 231, see also Process domains Probabilistic approaches, 12 Roughness height, 51, 55, 56 Probable maximum flood, 68 Rouse number, 95 Probable maximum precipitation, 68 Process domains, 34, 37, 41, 42, 217, 225, 230–232, 242, 24, Sand-bed, 9–11, 59, 64, 80, 83, 86, 98, 102–105, 108, 115, see also River styles 123–125, 135, 151, 160, see also Alluvial channels, fine bed Quarrying, 90 Sandur, 234, see also Braided channels Quasi-equilibrium state, 40, see also Equilibrium Sapping, 35, 36 Schumm, S.A., 14, 15, 37, 38, 41, 97, 116, 122, 132, 133–135, Rainsplash, 29, 31, 96 137, 141, 142, 144–146, 159, 174, 178, 179, 226, Rapidly varied flow, 50, 57, see also Gradually varied flow; 230, 238 Uniform flow Sculpted forms, 90, 144 Rating curve, 68, 75, 162 Sedigraph, 96 Reattachment point, 65, 67, 107 Sediment Recirculating flow, 65 budget, 9, 12, 83, 96, 120–124, 162, 185, 250 Reference conditions, 205, 215–217, 221, 224 delivery ratio, 41, 122, 123 Regolith, 17, 22–27, 30, 32, 227, see also Soil rating curve, 75, 162 Relative roughness, 55, 57 residence time, 41, 123, 233 Relative submergence transport form, 87 bed load, 98–104 grain, 87 equal transport mobility, 100 Relief ratio, 38, 39, 43, 202, 237 grain velocity, 98, 100, 103 Remote sensing, 8, 17, 70, 103, see also LiDAR path length, 100 Reservoir, 2, 27, 31, 33, 41, 79, 93, 96, 123, 191, 205, 232, phase I, 100 243, 247, see also Sinks; Storage phase II, 100 associated with dams, 2, 79, 97, 243, see also Human saltation, 94, 98 impacts to rivers, direct sheetflow, 98, 182 Resilience, 114, 224, 231 sheets, 98, 99 Response reaches, 109, 113 step length, 100 Restoration, river streets, 99, 102 field of dreams approach, 218 traction carpets, 99 keystone approach, 111, 218 two-fraction transport model, 104 and monitoring, 218, 220 waves, 100, 102, 106, 113, 186, 192 naturalization, 221 bed-material load, 94, 153 string of beads approach, 221 capacity, 30, 31, 41, 90, 94, 104, 113, 116, 127, 141, 142, system function approach, 218 148, 155, 160–163, 176, 178, 181, 185–188, 198, Reynolds number, 2, 79, 97, 243, see also Laminar flow; 243, 244 Turbulent flow detachment limited, 150 Rill, 31, 35, 123 dissolved load, 91–94, 123, see also Solutes; Total Riparian, 1, 9, 10, 13, 15, 17, 18, 20, 27, 28, 32–43, 59, 69, 75, dissolved solids 77, 78, 92, 114–119, 124, 127, 128, 139, 141, 143, hysteresis effects, 91 146–148, 170, 175, 177, 187, 202, 206–208, rate, 11, 12, 42, 88, 94, 98, 102–106, 118, 157, 235 210–213, 215, 216, 218, 220, 222, 223, 231, 234, supply limited, 94, 98, 113, 150, 163 236, 237, 240, 241, 249–251 suspended load, 94–97 organisms, 1, 9, 20, 139 transport limited, 94, 113, 150, 162, 198 Index 317

wash load, 94, 169 Tectonic aneurysm model, 229 yield, 7, 41, 83, 92, 96, 113, 120, 121–123, 128, 131, 146, Terrace, 175 176–179, 184, 191, 198, 199, 201–203, 205, 217, event-based, 178, 179 220, 224, 247, 251 fill, 7, 176–180, 244 Selective entrainment, 88, 111 and geochronology, 178 Sensitivity, 47, 114, 190, 198, 200, 231 and paleoenvironmental information, 174, 178 Separated flow, 65 riser, 175, 248 Serial discontinuity concept, 231 strath, 43, 155, 176–180, 240 Shear layer, 64, 65, 107, 108, 147 tread, 175, 176, 180 Shear stress, 11, 12, 29–31, 36, 40, 55, 57, 59, 66, 67, 69, 80, Thalweg, 61, 112, 135–138, 146 86–89, 94, 99–100, 104, 106, 117–119, 124, 138, Theoretical approaches, 11 143, 159, 160, 169, 186, 209, 234, 241, 249 Thread flow, 31 Sheet flow (hillslopes), 31, 182, see also Sediment; Threshold Transport; Bedload; Sheetflow external/extrinsic, 14 Shields, A.F., 85, 86, 88, 89, 103, 104, 124, see also Shields internal/intrinsic, 14, 133, 178 number Threshold channel, 83, 157, see also Alluvial channels, Shields number, 86, 89, see also Dimensionless critical shear coarse bed; Bedrock rivers; Boulder-bed; stress Cobble-bed; Gravel-bed apparent Shields number, 86 Throughflow Sinks, 2, 3, 33, 94, 192, see also Reservoir; Storage concentrated, 26, 27 Sinuosity, 125, 131, 134–136, 138, 140, 142, 144, 156, 221, diffuse, 26 230, 238, 241 Total dissolved solids, 32, 91, see also Sediment, transport; Soil, 1, 2, 5, 7, 14, 18, 22–27, 34, 36, 75, 79, 91, 92, 94, 119, Dissolved load; Solutes 120–123, 139, 167, 169, 177, 179, 180–182, 186, Transient landscape, 17 197, 198, 200, 202–204, 207, 235, 242, 243, Transpiration, 23, 26, 34, 75, 79, 118, 201, 213 245–247, 249, 250, see also Regolith Transport reaches, 111, 114 Soil pipe, 14, 26–28, 243 Transverse ribs, 108, 109, 124, see also Bedforms Solutes, 2–6, 8, 19, 21–23, 32, 39, 43, 44, 91, 92, 114, 121, Turbidity current, 190, see also Delta, turbidity 166, 195, 225, 232, 237, see also Sediment, current transport; Dissolved load; Total dissolved solids Turbulence, 11, 13, 47, 49, 51, 52, 55, 56, 57, 60–66, 79, 80, Spawning (fish), 4, 89, 103, 207, 213, 217, 233, 234, 251 86–89, 96, 101, 105, 106, 108, 109, 111, 115, 147, Stage (water surface elevation), 68 166, see also Coherent flow structure; Turbulent Steady flow, 50, see also Unsteady flow flow; Vortex Steady-state landscape, 17 boil, 108 Stemflow, 23 burst, 89, 100, 108 Step-pool, 57, 58, 100, 102, 105, 108–114, 124, 125, 157, 212, intensity, 63, 65, 80 see also Bedforms kolk, 108 Storage, 2, 3, 5, 12, 13, 15, 18, 25, 29, 31, 33, 34, 39, 41, 44, macroturbulence, 108 77, 78, 83, 93, 97, 102, 116, 120, 121–124, 161, sweep, 64, 89 166, 174, 189, 205, 212, 231, 232, 241–243, Turbulent flow, 52, 55, 56, 60–63, 66, see also Coherent flow 245–248, 250, see also Reservoir; Sinks structure; Laminar flow; Reynolds number; Storm transposition, 68 Turbulence; Vortex Strahler, A.N., 38, 39, 74, 229 Stream head, 34, 37, 43 Unchanneled hollows, 34, 41 Stream order, 38, 39, 41, 43, see also Link Uniform flow, 48–50, 53, 54, 57, 66, 68, 69, 87, Stream power, 40, 69, 80, 104, 106, 130, 141, 143, 144, 154, see also Gradually varied flow; Rapidly varied 155, 239, 241, 249 flow per unit area, 88 Unsteady flow, 50, 61, see also Steady flow specific, 67, 175 Upward directed seepage, 91 total, 67 Subcritical flow, 40, 51, 53, 106 Variable source area concept, 25, 26 Supercritical flow, 51, 52, 106, 115, 182 Velocity Suspended load, 94–98, 103, 206, see also Sediment, 1d, 55, 56, 61, 62 transport 2d, 61, 62 Sweep, 64, 89, 160, see also Turbulence 3d, 61, 62, 137, 149 318 Index

Velocity (Continued ) Wandering channels, 139, 143 shear velocity, 55, 56, 64, 94, 95, 104, 115 Wash load, 94, 169, see also Sediment, transport velocity profile, 60–62, 87, 88 Weathering Viscosity, 49, 50, 55, 59, 64, 96 chemical, 19, 22, 36, 90 Vortex, 49, 56, 57, 108, 219, see also Coherent flow structure; physical, 22 Turbulence; Turbulent flow Width/depth ratio, 59, 125, 127–128, 131, 137, 220, 227, horseshoe, 64, 65 249