Ecological Limitations and Potentials of Artificial Aquatic Systems By

Home , Buglife

Ecological Limitations and Potentials of Artificial Aquatic Systems

Chelsea Connair Clifford

Environment Duke University

Date:______Approved:

______James B. Heffernan, Supervisor

______Martin W. Doyle

______Dean L. Urban

______Daniel D. Richter

Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Environment in the Graduate School of Duke University

2018

ABSTRACT

Ecological Limitations and Potentials of Artificial Aquatic Systems

Chelsea Connair Clifford

Environment Duke University

Date:______Approved:

______James B. Heffernan, Supervisor

______Martin W. Doyle

______Dean L. Urban

______Daniel D. Richter

An abstract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Environment in the Graduate School of Duke University

2018

Abstract

As humans increasingly alter the surface geomorphology of the Earth, a multitude of artificial aquatic systems have appeared, both deliberately and accidentally.

Human modifications to the hydroscape range from alteration of existing waterbodies to construction of new ones. The extent and ecosystem services of these systems are underexplored, but likely substantial and changing. Instead of simply accepting that artificial ecosystems have intrinsically low values, environmental scientists should determine what combination of factors, including setting, planning and construction, subsequent management and policy, and time, impact the condition of these systems.

Scientists, social scientists, and policymakers should more thoroughly evaluate whether current study and management of artificial aquatic systems is based on the actual ecological condition of these systems, or judged differently, due to artificiality, and consider resultant possible changes in goals for these systems. The emerging recognition and study of artificial aquatic systems presents an exciting and important opportunity for science and society.

Irrigation ditches are ubiquitous features of water networks in rural and urban settings in drylands, and are thus potentially important habitats within the modern hydroscape. The habitat value of ditches and other artificial systems depends on whether these systems respond to local and watershed-scale land use in similar ways to

natural features, or whether artificial origin inherently constrains a system's ecological condition. The ditches and creeks of Bishop, California are fed by water from the same minimally developed watershed on the Eastern slope of the Sierra Nevada Mountains, and so served to test whether artificial and natural waters in the same watershed setting and with shared water can provide similar habitat. We sampled benthic macroinvertebrates at 52 sites within the town, stratified by substrate and season.

Communities varied by substrate and season as expected, but did not differ significantly between artificial and natural streams. Instead, both types of streams changed as water flowed from undeveloped desert through town, suggesting that irrigation ditches respond to local urbanization in much the same way that natural streams do. Differences in finer-scale spatial structure of community similarity suggest that community assembly processes may differ between natural and artificial channels, but potential mechanisms for these differences are unclear. This study demonstrates that artificial aquatic systems may have substantial ecological value, and suggests that the poor condition of many artificial aquatic systems may reflect stressful watershed settings rather than something intrinsic to their artificiality.

The drainage ditches of the North Carolina Coastal Plain do not merely degrade wetlands; they themselves have ecological structural characteristics of wetlands. We surveyed 32 agricultural, freeway, and forested ditch reaches across this region for hydrologic indicators, soil organic matter, and plants. All showed some

hydrologic indicators and had some soil organic matter, with easterly, swampy forests having the most, though with substantial variation across all and few significant differences between types. All had hydrophytic herbaceous plant communities in the sense of at least half their percent cover belonging to obligate, facultative wetland, or facultative taxa. These herbaceous communities differed significantly across site types

(F=3.25, d.f.= 2, p=0.001), and responded to both landscape-level factors like nearby development coverage and local-level factors like apparent mowing. Sample sites were not well mapped in well-used federal aquatic databases; the National Hydrography

Dataset only included one on a “CanalDitch” flow line, and the National Wetlands

Inventory only included two within “partially drained/ditched” areas, and none as individual “excavated” features. Others were mis-categorized, but neither database included any highway sampling sites. Despite this limited information about extent, variation and management impact suggests that human potential to impact wetland structure of these manmade aquatic ecosystems throughout the North Carolina Coastal

Plain, and beyond, could be large.

Artificial lakes are a dominant aquatic ecosystem type, but the processes controlling their condition are under-explored. Here we use structural equations modeling to compare the formation of algal blooms and associated water quality issues in 1,045 artificial and 870 natural lakes in the United States using the U.S. Environmental

Protection Agency’s National Lakes Assessment data. We compare chemical and

physical measurements associated with water quality and the relationships between them, and find that the processes are significantly different between natural and artificial lakes, in a way that suggests impacts of interference with thermal stratification through dam management in reservoirs. However, both the overall processes and the distributions of the data are roughly similar between the two origin types, and between

2007 and 2012 sample years. Artificial lakes are lakes, and process-based explorations of their behavior can help us better know management options.

Taken together, this dissertation examines an artificial version of each of the major aquatic ecosystem types: stream, wetland, and lake. It examines the processes controlling their ecological condition with increasing intricacy with each chapter, and finds ways that artificial aquatic ecosystems are both similar to and different from natural ones. This dissertation provides a new way of looking at the constraints and opportunities that artificial waterbodies afford those in charge of them and interested in their conservation potential.

vii

Contents

Abstract ...... iv

Contents ...... viii

List of Figures ...... xi

Acknowledgements ...... xiii

0. Introduction ...... 1

0.1 Invertebrate communities in Bishop, California (Ch.2) ...... 1

0.2. Conceptual Review of Artificial Aquatic Systems (Ch. 1) ...... 2

0.3. Ditches of the North Carolina Coastal Plain (Ch.3) ...... 4

0.4. Algal Blooms in the National Lakes Assessment (Ch. 4) ...... 5

1. Artificial aquatic ecosystems...... 6

1.1 .Introduction ...... 6

1.2. Degrees and Axes of Artificiality ...... 8

1.3. The Ecological Significance of Artificial Aquatic Systems ...... 14

1.3.1. The Extent and Dynamics of Artificial Aquatic Systems ...... 14

1.4. The Condition of Artificial Aquatic Systems and Its Drivers ...... 19

1.4.1. Setting ...... 24

1.4.2. Time ...... 26

1.4.3. Design...... 28

1.4.3.1. Design Goals ...... 29

1.4.3.2. Planning and Construction ...... 31

1.4.3.3. Management and policy ...... 33

viii

1.4.3.4. Monitoring, learning, and iteration ...... 36

1.5. Artificiality and Perception of Ecosystem Value ...... 38

1.5.1. Interactions Between Perception and Condition in Artificial Aquatic Systems ...... 40

1.6. Conclusions - Artificial Aquatic Ecosystems in Hybrid Hydroscapes ...... 44

2. Urban Ditches and Natural Creeks Support Similar Invertebrate Communities in a Minimally Developed California Watershed ...... 48

2.1. Introduction ...... 48

2.2. Materials and Methods ...... 51

2.2.1. Study area ...... 51

2.2.2. Study design ...... 54

2.2.3. Data analysis ...... 56

2.3. Results ...... 59

2.4. Discussion ...... 64

2.5. Conclusion ...... 70

3. North Carolina Coastal Plain ditch types support distinct hydrophytic communities 71

3.1. Introduction ...... 71

3.2. Materials & Methods ...... 73

3.2.1. Sample site selection ...... 73

3.2.2. Field and Lab Methods ...... 75

3.2.3. Site characterization by remote sensing ...... 79

3.3. Results ...... 80

3.4. Discussion ...... 87

3.5. Conclusion ...... 92

4. Similarities and Differences in Algal Bloom Processes in Artificial and Natural Lakes ...... 94

4.1. Introduction ...... 94

4.2. Materials & Methods ...... 96

4.3. Results ...... 99

4.4. Discussion ...... 102

4.5. Conclusion ...... 103

5. Conclusion ...... 105

References ...... 108

Biography ...... 133

List of Figures

Figure 1: Classification of Artificial Aquatic Systems ...... 11

Figure 2: Extent of Artificial as Compared to Natural Waterbodies in the US ...... 15

Figure 3: Inherent (Left) vs. Process-Based Model (Right) of the Condition of Artificial Aquatic Systems ...... 24

Figure 4: Value Axes for Aquatic Ecosystems ...... 44

Figure 5: Conceptual Diagram of the Role of Artificiality in the Management and Services Provided by an Aquatic Ecosystem ...... 47

Figure 6: Bishop, California ...... 53

Figure 7: Weighted Averages of Taxa and Vectors of Independent Variable Factors Within Nonmetric Multidimensional Scaling Analysis ...... 60

Figure 8: Nonmetric Multidimensional Scaling Analysis Plots ...... 61

Figure 9: Impact of Longitude on Invertebrate Communities ...... 63

Figure 10: Decay of Community Similarity with Distance Differs Between Natural and Artificial Sites ...... 64

Figure 11: Map of Sites Sampled ...... 75

Figure 12: Example Subsampling Layout, Within a Freeway Sample Site ...... 76

Figure 13: Site Average Water and Soil Values by Ditch Type ...... 80

Figure 14: Site Average Total Percent Cover Per Quadrat of Taxa in Each Wetland Indicator Status by Ditch Type ...... 81

Figure 15: Site Average Characteristics of Herbaceous Plant Communities by Ditch Type ...... 82

Figure 16: Nonmetric Multidimensional Scaling Analysis of Herbaceous Plant Communities at Different Ditch Types, Overlain with Variable Vectors ...... 84

Figure 17: Inclusion of Sampled Sites in Federal Databases ...... 85

Figure 18: Map of Croatan National Forest Sample Site and Lidar ...... 87

Figure 19: Structural equations model of the algal bloom process in natural and man-made lakes ...... 99

Figure 20: Difference Between Man-Made and Natural Structural Equations Model ....100

Figure 21: Box Plots Representing Variable Data Distributions ...... 102

xii

Acknowledgements

Thanks first to my advisor, Jim Heffernan, for putting up with and doing too much to list here. My entire committee has also helped intellectually and materially with this dissertation; they all can think through problems that take me months in a matter of seconds. Dean Urban loaned me equipment and helped with experimental design.

Martin Doyle has talked me through court cases, law, politics, geomorphology, and a great deal more that I would never have found on my own. Dan Richter trekked through the floodplains with me repeatedly beyond endurance, and let me tag along with his lab.

Similarly, my coauthors, Jeff Holmquist, Jutta Schmidt Gengenbach, and Kris Voss, mentored and befriended me through tribulations, and responded to my requests for timely feedback for this document helpfully and with good cheer.

The Duke River Center and the Nicholas School of the Environment have been a wonderful intellectual community, and too many people to name, particularly in the

River Center, Landscape Ecology Lab, and Wetland Center have shared ideas, equipment, and time with me; thank you all so much. Outside of the Nicholas School,

Priscilla Wald and Matthew Taylor deserve similar credit for ideas and editing, and outside of Duke, the same goes to Alex Webster, Monica Palta, and Lauren McPhillips.

Thanks additionally to Sarah Ludwig Monty, Danilo Meyer, and Megan Fork for field

xiii

assistance, and to Meg Stevens Avery, Danielle Wiggins, Dawn Becker, and Denise

Waterbury for excessive administrative assistance.

The California Department of Fish and Game, the occupants of Bishop,

California, the Los Angeles Department of Water and Power, the North Carolina

Department of Transportation, the North Carolina Department of Agriculture &

Consumer Services, the North Carolina Wildlife Resources Commission, the North

Carolina Department of Parks and Recreation, the North Carolina Department of

Environmental Quality, the U.S. Forest Service, the U.S. Fish and Wildlife Service, and

The Conservation Fund all provided site access and/or information. I received funding from the Nicholas School, an NSF Graduate Research Fellowship, a Kolenkow-Reitz

Fellowship from Carleton College, the Earth Stewardship Initiative, the Duke Wetlands

Center, and the Kenan Institute for Ethics.

I also owe gratitude to family, friends, my animals, the Mountain to Sea Trail, and the Little River Waterfowl Impoundment for helping keep me sane and functioning enough to get to this point. Likewise, I would not have finished, or indeed started, without the help of numerous medical professionals, including Nathan Zasler, Sarah

Bryce, Aurelio Alonso, Mandi Singer, Elizabeth Kearney, Dmitri Putulin, Martha

Golden, Sammy Banawan, Kathryn Walker, and Lori Fendell. Past mentors, including

Phil Camill, Mark McKone, Betsey Boughton, and Gene Lollis, took the time to instrumentally shaped my path. Thank you all!

xiv

0. Introduction

This dissertation explores the wet parts of the world that environmental scientists usually do not explore, as they are too close to us; it applies the observational and analytical tools of ecology to manmade analogs for streams, wetlands, and lakes, to deduce some of what processes make these systems what they are. Each chapter is a somewhat more intricate look at some aspect of how an artificial aquatic ecosystem works. This dissertation suggests that there is much more to these ecosystems than most ecologists currently realize, beginning with the ontological conundrum of where humanity takes off and nature begins, if indeed there is a border. Better understanding what makes these ecosystems, these ditches, reservoirs, canals, and similar, exhibit or not exhibit which ecological structures and functions, provides a better understanding of how the world works at the human-nature-water interface, and expands the known management possibilities within these systems. I believe that my research adds some information to general knowledge of what the potential opportunities and limitations are for artificial aquatic systems, and what factors extend and constrain these options, but in large part it is a call to action, ideas and examples of what could be more widespread as other scientists also become interested in this realm.

0.1 Invertebrate communities in Bishop, California (Ch.2)

While I had grown up playing in both ditches and “real” creeks and wetlands, my first research encounter with ditches occurred during a college summer job. While

working in an ecology lab that focused on invertebrates, I got to do a project of my own, and opted to look at benthic macroinvertebrates in the decorative ditches diverted from irrigation ditches diverted from Bishop Creek. I was not sure if I would find anything, but I set out to collect what I could at accessible sites, randomizing within those and creek sites for comparison. Surprisingly, while the invertebrate communities differed strongly by substrate, they were statistically not differentiable between ditches and creeks. This result intrigued me enough that I secured a Kolenkow-Reitz Fellowship from my college to return and expand my sample collection, which yielded the same result, with some seasonal differences. Further analysis suggested that invertebrate communities changed, losing diversity and sensitive taxa, as waters flowed through town, which suggested an urbanization impact. This project suggests that at least in some cases, artificiality need not mean that an ecosystem houses less tolerant organisms.

This project got me to thinking about all the irrigation ditches in the western U.S., all the ditches generally, and wondering if they, too, harbored communities interacting in complex ways with “natural” ecosystems.

0.2. Conceptual Review of Artificial Aquatic Systems (Ch. 1)

My beginning my PhD at Duke’s River Center in 2013, funded by an NSF

Graduate Research Fellowship, coincided with the USEPA’s release of drafts of the

Clean Water Rule. Having researched ditches before, I noticed that the then-current draft prioritized the exclusion of most ditches and other artificial aquatic systems from

jurisdiction, over all factors that could have resulted in their inclusion, such as tributary status. While I understood, roughly, the political necessity of this choice, I also thought it telling that not just policymakers, but most conservationists and environmental scientists, seemed willing to accept it. It indicates to me that artificiality tends to automatically devalue ecosystems, which I think has affected both science and actual waterbodies. So, I wrote a paper that defines artificiality in terms of degree of human modification and level of intentionality, the interaction of which predicts probability of regulation; we are least likely to regulate systems that humans have constructed de novo and on purpose as “water.” I review the probably vast but basically unknown extent of artificial aquatic systems, and examples of known ecosystem services and disservices they provide, which are probably substantial and below their potential. I suggest process-based alternatives to “artificiality” as an explanation for the poor condition of some artificial aquatic systems, including watershed setting, design, which includes both initial planning and ongoing management and policy, and time for succession and/or naturalization. Finally, I suggest that artificiality influences how people perceive the condition of aquatic systems, which impacts how we manage and regulate these systems, such that low expectations for artificial aquatic systems could be self-fulfilling.

This paper is very close to submission for publication, and I hope that other scientists find the frameworks I present to be useful launching points for their own research. I

started writing it largely to structure my own dissertation; I am trying to disentangle those drivers of condition of artificial aquatic ecosystems somewhat, in some systems.

0.3. Ditches of the North Carolina Coastal Plain (Ch.3)

I sought to disentangle the influences of setting and some design choices in drainage ditches in the North Carolina Coastal Plain by examining plant communities, morphology, soils, and waters present in forested, agricultural, and roadside ditches. All three types of ditch generally had hydrophytic plant communities, by U.S. Army Corps of Engineer wetland delineation standards, but the actual taxonomic composition of these communities differed significantly among the three types. Mowing of roadside ditches, as suggested by herbaceous vegetation height, influenced plant community composition, suggesting some management influence. Agricultural ditches supported the most herbaceous plants and the most biodiverse plants, and overall, the plant communities resembled those of young wetland restorations in the region. Wetter forested sites had soil carbon comparable to local undisturbed, natural wetlands, but roadside and agricultural sites had relatively little. All sites showed some U.S. Army

Corps of Engineers wetland hydrology indicators, and some had substantial flow, blackwater, and other interesting aquatic characteristics. In spite of all of these qualities of interest to aquatic science, almost none of these sites were included in the National

Hydrography Dataset or National Wetland Inventory, which supports my suspicion that

our inventories of artificial waters with structures and functions of ecological interest are missing a great deal.

0.4. Algal Blooms in the National Lakes Assessment (Ch. 4)

The USEPA collected a great deal of data in over 1,000 lakes around the U.S., roughly half natural and half artificial, in surveys in 2007 and 2012. In the report on the first survey, when comparing artificial and natural lakes, they simply concluded that artificial lakes are in worse condition than natural ones. This trove of physical, chemical, biological, and recreational data allowed me to explore along many more axes than I could alone, through structural equations modeling, how the process of algal bloom formation and associated water quality issues differs and not between artificial and natural lakes. Artificial lakes decline in quality through similar, but not identical, processes to natural ones. It appears that deeper, colder manmade lakes may be more likely to respond largely to changes in their thermal stratification by humans through dam management, but other explanations are possible for the patterns observed. This model suggests both that the fundamental ecological congruence between artificial and natural aquatic ecosystems, and the unique ways that humans influence the ones that we build and manage.

1. Artificial aquatic ecosystems

By Chelsea C. Clifford and James B. Heffernan

1.1 .Introduction

Humans alter geomorphology on an ever-increasing scale (Hooke 2000), one comparable with (Haff 2010), and in some ways exceeding (Wilkinson 2005), rates of natural processes. Every change people make to Earth’s surface has the potential to affect the flow and accumulation of water. People have dug ditches, impounded streams and rivers, and otherwise shifted Earth’s surface to direct and store water for human use, especially agriculture, for over 5,000 years (Hooke 2000). Today, land use matrices that entail human-engineered waterbodies, such as urban settlements, rice villages, and irrigated cropland, cover significant fractions of terrestrial Earth (Ellis and Ramankutty

2008), and patterns of surface water extent are tightly linked to land use (Steele et al.

2014). Man-made and -modified aquatic systems have become ubiquitous landscape features (Downing et al. 2006).

In spite of their commonness, artificial aquatic systems remain poorly understood. Indeed, it is often unclear which waterbodies even belong in the category of

“artificial” or “anthropogenic.” To date, the limited study of different artificial aquatic systems has been fragmented among various domains of ecology and other environmental sciences, and more often discussed on the margins of science and management of natural ecosystems than in conjunction with them, as part of a complete

hydroscape (Sayer 2014). Because of their marginalized status, the abundance and extent of artificial systems are poorly quantified, and the conditions of artificial aquatic ecosystems and their various causes are similarly understudied. As a result, we lack a scientific basis for assessing the ecological value of artificial aquatic systems, or determining how management and policy might improve that value. The ubiquity of artificial aquatic systems, the potential commonalities among them and with natural aquatic ecosystems, and our limited understanding of their central and evolving role in the modern hydroscape all argue for more integrated study of the waterbodies created and transformed by human activity.

In this paper, we propose a framework for aquatic ecosystem artificiality that includes both the intent and magnitude of modification, and argue that policy and management often implicitly use these characteristics to differentiate in their treatment of aquatic systems. We assemble current estimates of the extent of some types of artificial waterbodies in the US, and review established knowledge of their ecological condition, including the ecosystem services and disservices that they provide. We argue that the condition of artificial aquatic systems, as for their natural counterparts, likely reflects ecological processes and human decisions both in place and within their watersheds, and that the often poor condition of these systems (Biggs, von Fumetti, and

Kelly-Quinn 2016) is not necessarily inherent to their anthropogenic origin. Finally, we posit that undervaluation of artificial aquatic systems may lead scientists, managers, and

policymakers to treat artificial waterbodies in ways that perpetuate poor ecological conditions. To manage the rapidly changing and increasingly anthropogenic hydroscape, aquatic scientists need to better inventory its myriad artificial components, evaluate their current conditions, and link these to causes and ecological function.

Future assessment of artificial aquatic systems that clarifies what their real and perceived values are, and how to purposefully change those values, will require more deliberate, intensive, systematic and mechanistic study, and, perhaps, a shift in our perspective regarding what counts as nature.

1.2. Degrees and Axes of Artificiality

What does it mean to identify aquatic ecosystems as ‘artificial’? Designer ecosystems, like rain gardens and ponds, swales, and wetlands conceived and built for water treatment (Anderson et al. 2017), are obviously deliberately constructed for human purposes, often where no waterbody existed before, and thus undisputedly artificial (Higgs 2017). A stream or lake in a protected watershed far removed from intensive human land use might serve, in a traditional ecological study, as a “completely natural” or “pristine” reference site. However, many ecosystems fall between these extremes. For example, scientists and policymakers often differentiate between a channelized stream and a wholly manmade trench, but either may be colloquially called a “ditch” or “artificial,” and the two may look and even function quite similarly, particularly within a highly modified agricultural landscape; all natural waterbodies do

not necessarily maintain better structure and function than all artificial ones (Crail,

Gottgens, and Krause 2011). Restoration projects similarly blur the bounds of “natural” and “artificial” (Palmer and Bernhardt 2006, Higgs 2017). While restoration typically has the goal of returning an ecosystem to some more natural state (Palmer and Filoso 2009), the process of restoration necessitates human intervention, which is often sustained through maintenance (Ross et al. 2015), an implicit acknowledgement that many ecosystems cannot withstand human disturbance without purposeful human assistance.

Meanwhile, undirected human actions, like stormwater efflux, can accidentally “restore” natural systems (Bateman et al. 2014, Palta et al. 2016, Sueltenfuss et al. 2013). Even mostly or wholly manmade systems, like retention ponds, can “naturalize,” or become more biotic and ecosystem-like, in time, without deliberate human effort (Chester and

Robson 2013).

These test cases suggest that artificial aquatic ecosystems can be usefully organized along two axes: the degree to which their existence and characteristics depend on human activity, and the degree to which that activity is specifically intended to produce those changes (Fig. 1). The first axis spans from moderate alteration of systems with initially geologic origins, to the wholesale creation of waterbodies on formerly dry land, sometimes away from topographic lows. The latter axis ranges from aquatic systems that came to exist as inadvertent or accidental by-products of other human activities on the Earth’s surface to deliberate and intentional products of such activity

(Kueffer and Kaiser-Bunbury 2013). These two artificiality axes are gradients, not discrete categories, and can be difficult to parse, especially for the many waterbodies with complex, multi-layered histories of modification (Hupy and Koehler 2012, Meade and Wheeler 1990, LAF 2016). Depending on the purpose, it may be appropriate to define “artificial” water bodies broadly or narrowly within this space.

The most clearly artificial aquatic systems are those humans construct where none existed before. Deliberate examples include fountains, many roadside ditches, rain gardens, stormwater treatment areas, many farm ponds, and all designer ecosystems

(Higgs 2017). Water conservation regulation typically exempts such constructs outright

(Zollitsch and Christie 2014, USEPA and USACE 2015, Sayer 2014). Accidental examples include logging ruts, erosional gullies in building sites and agricultural fields, poorly drained impervious surfaces, and even bomb craters (Vad et al. 2017). Left unmaintained, in time, such accidental waterbody construction in relative uplands can

“naturalize” to a seemingly “wild” ecosystem (Vad et al. 2017, Hawley 2014, Palta,

Grimm, and Groffman 2017). Waterbodies that humans have constructed accidentally, but that appear relatively free from human intent, are more likely to be regulated than waterbodies that humans have constructed on purpose (Zollitsch and Christie 2014,

USEPA and USACE 2015).

Figure 1: Classification of Artificial Aquatic Systems. Level of deliberateness of modification increases from left to right, and degree of modification increases from bottom to top. These two axes combined yield a third axis, likelihood of legal standing as a water for regulatory purposes in the U.S., running from unlikely among deliberately constructed ecosystems like swimming pools and upland farm ponds in the upper right, to likely among accidentally altered ecosystems, like a lake bisected by a railroad causeway, in the lower left. Other potentially influential characteristics, such as size and permanence, may influence valuation in both natural and artificial systems.

Transformation occurs when human intervention changes waterbodies from one type to another, fundamentally different in morphology and flow, such as from a lake to a wetland, or a wetland to a stream. Deliberate transformations include ditching of wetlands for agriculture, conversion of wetlands to ponds during development, damming of streams to build reservoirs, piping of creeks, and many restorations. The

impoundment of the Everglades behind Tamiami Trail, a road that interrupted sheet flow, is one prominent such example (Boler and Sikkema 2010). Similarly, humans accidentally transformed the bed of the Salt River, dried through damming upstream, into wetlands, at stormwater outflows (Bateman et al. 2014). Transformed waters generally retain their regulatory status (USEPA and USACE 2015, Zollitsch and Christie

2014) regardless of intent. Such transformations are often considered potentially degrading and thus may require regulatory permission (Zollitsch and Christie 2016,

2014, Hough and Robertson 2009).

The least modification that might make a waterbody appear artificial is alteration, in which fundamental morphology and flow are retained. Deliberately, people straighten and channelize rivers, harden riverbanks and shorelines, and dredge lakes. Accidentally, sedimentation from agriculture, mining, or construction makes streams and lakes shallower (Gregory 2006); mill dams clogged river valleys all over the eastern U.S. (Walter and Merritts 2008). Meanwhile, “urban stream syndrome,” including incision, flashiness, and other changes largely in response to stormwater drainage, has become a well-known issue in developed areas everywhere (Walsh et al.

2009). In one particularly grand example of accidental alteration, a railroad causeway divides the Great Salt Lake into mostly independent halves with very different chemistry and community assemblages (Naftz et al. 2008). Such modifications generally do not remove jurisdiction of regulation from waterbodies, and instead likely invoke

regulatory oversight (Zollitsch and Christie 2016, 2014, Hough and Robertson 2009).

These are the waterbodies often easiest to imagine in their “natural” state, likeliest to be labeled as simply “degraded”, most likely to attract conservation, restoration, and related scientific interest, and, depending on context, possibly most appropriate for restoration (Wohl, Lane, and Wilcox 2015, Hallett et al. 2013, Hobbs 2016).

Together, these two axes of artificiality - degree and intent of modification - may help explain the regulatory protection afforded to various aquatic ecosystems (Fig. 1) in

US water law (Zollitsch and Christie 2016, 2014, Hough and Robertson 2009), in which increasing modification with increasing intent render water bodies less ‘natural’, less

‘valuable’, and less likely to be protected (Crifasi 2005). Scientists, regulators, legislators, and other policymakers often do not explicitly acknowledge the value judgement implied by differential treatment of waterbodies according to their type of artificiality.

Other traits such as technology, purpose, age, size, and permanence may also figure into value judgments and policy decisions people make about aquatic systems, and so might serve as a basis for further regulatory classification of artificial waterbodies. Some of these attributes, like small size and impermanence, may disproportionately characterize artificial aquatic systems, but also apply to most natural waterbodies (Downing et al.

2012, Downing et al. 2006). Regulatory standards that omit smaller, less permanent waterbodies may do so as much because of their biological, geomorphological, and chemical features (Zollitsch and Christie 2014), or because their dense distribution

inconveniences property and land use considerations (Rabkin 2014), as because of their human origins per se.

1.3. The Ecological Significance of Artificial Aquatic Systems

Understanding the ecological and socio-ecological value of artificial aquatic systems requires that we understand their extent and distribution, their physical and chemical condition and how they relate to biotic communities, and the range of ecosystem services that they provide, but considerable uncertainty surrounds all of these characteristics. Artificial aquatic systems are likely to be ecologically important, owing to their extent, which may rival that of natural drainage systems and water bodies. The ecological functions of artificial systems are likely to be socially important due to their frequent placement near large numbers of people. Moreover, the extent, distribution, and characteristics of artificial water bodies are likely changing rapidly. Interdisciplinary understanding of the services and disservices of artificial aquatic systems, the factors that influence them, and their distribution in space and time could foster decisions that increase their ecological value.

1.3.1. The Extent and Dynamics of Artificial Aquatic Systems

Our understanding of the extent of artificial aquatic systems is piecemeal, and to our knowledge, no previous study has assembled estimates of the extent of various artificial aquatic systems. Available estimates are largely limited to the US, and largely for intentionally designed aquatic features that are ubiquitous in agricultural, industrial,

urban, and recreational land uses. Even with these incomplete inventories, it is clear that the deliberately constructed or altered fraction of the hydroscape is both large and growing, and must be included in any comprehensive assessment of aquatic resources.

Figure 2: Extent of Artificial as Compared to Natural Waterbodies in the US. Data drawn from a variety of sources, shown in parentheses. Areal features are shown in panel a. at left; linear features are shown in panel b. at right.

Deliberately constructed and transformed channels constitute a significant portion of the US hydroscape (Fig. 2). The U.S. National Hydrography Database includes 5,525 km of ditches and canals, or approximately the length of the Missouri

River through the Mississippi to the Gulf of Mexico. This aggregation is likely a substantial underestimate, as many smaller ditches do not appear in the database. In

2010, ditches in U.S. agricultural lands occupied about 115,760 km2 (USDA 2015), a surface area similar to that of Lakes Superior and Huron combined, or 2.67 times the

surface area of all U.S. streams combined (Downing et al. 2012). Channelization, another deliberate transformation, has altered the geomorphology of upwards of 26,550 km of rivers and streams in the U.S. (Leopold 1977, as cited in Gregory 2006), more than seven times the length of the Mississippi River. Humans have channelized more than 500,000 km of rivers worldwide, and built more than 63,000 km of canals (1995, as cited in

Gleick, Singh, and Shi 2001, as cited in Meybeck 2003). The U.S. also has 6.5 million km of roads (2014c), many of which have ditches or gutters along both sides that contain water at least occasionally. Thus, it is likely that road drainage in the US, which effectively serve as urban headwaters (Hobbie 2014), is of comparable length to the 5.3 million km of the country's rivers and streams. The US EPA estimates that 77% of the approximately 1.8 million km of wadeable streams in the US are in poor (42%) or fair

(25%) condition (Paulsen et al. 2004), meaning that anthropogenic channels are of comparable or perhaps even greater extent than natural channels whose condition has been affected by human activity.

The extent of constructed lakes and ponds are similarly significant in comparison to natural water bodies (Fig. 2). The U.S. has about 22,000 km2 of deliberately constructed or transformed farm ponds an area similar to that of Lake Michigan; the world as a whole has about 76,830 km2 of farm ponds (Downing et al. 2006). In total for the U.S., the U.S. Army Corps of Engineers inventories an impounded (deliberately transformed) area in excess of 180,000 km2 (USACE 2013), or almost double the area of

all the Great Lakes, for purposes including irrigation, hydroelectricity, flood control, navigation, water supply, and recreation (USACE 2013). The world had approximately

258,570 km2 of impounded water in the mid-2000s (Downing et al. 2006), before the completion of the Three Gorges Dam in China and other projects. In 2009, The U.S. Fish and Wildlife Service estimated that only 31% of the country’s 27,151 km2 of freshwater ponds were natural. Of the artificial pond area, about the size of Lake Ontario, farm ponds took up about 1.5 times the space as natural ponds; urban ponds occupied about half the area of natural ponds despite the relatively small amount of urban space, and industrial and aquaculture ponds made up significant fractions as well (Dahl 2011). This estimated aquaculture pond area, of just over a thousand square kilometers, is relatively much smaller than that of many countries’; globally, 1.7 million km2 of the world’s 2.7 million km2 of irrigated land goes to rice production, and is flooded seasonally, at least

(Ellis and Ramankutty 2008).

The abundance and extent of accidentally created aquatic systems is extremely poorly quantified. Human earth movement has risen in the last 150 years, from a historic background level of less than 5 tons per capita to more than 30 tons per capita annually in the U.S. (Hooke 2000), creating the potential exists for the formation of local low areas and water accumulation. Moreover, earth movement has become more common in wet spaces (Wohl, Lininger, and Baron 2017), where the potential for accidental creation of water bodies is higher. Water infrastructure, such as stormwater or water supply pipes,

can create accidental wetlands wherever leakage occurs (Palta, Grimm, and Groffman

2017). Accidental creation of waterbodies is perhaps most likely in abandoned areas, where anthropogenic depressions and impoundments that accumulate water may remain, often with minimal human interference. Better information about the density of accidental water bodies, combined with estimates of the land area over which they might occur, would allow us to estimate their extent, but that information is currently lacking.

The distribution of artificial waterbodies, like many natural systems, is dynamic in time, owing to both seasonal and event-driven hydrologic change as well as longer- term changes in land cover. Between 1984 and 2015, North America as a whole, home to

52% of the world’s permanent surface water, added a net 17,000 km2 to this area. The area of permanent surface water in the U.S. as a whole grew 0.5%, even as six of its western states lost 33%, or over 6,000km2, of their permanent surface water (Pekel et al.

2016). Farm ponds are highly dynamic in use, creation, and abandonment (Berg et al.

2016). In commercial and residential developments, ponds and other stormwater features too small to appear in the above analysis, wink in and out of existence too quickly for inventorying, or further scientific study (Oertli 2018).

Emerging technologies and modeling approaches have the potential to improve inventories of small and accidental water bodies and better characterize the distribution and dynamics of hydroscape change. Advances in remote-sensing technology, such as

the increasing availability of high-resolution lidar data, may very soon yield much better maps of small and otherwise over-looked waterbodies over broad extents (Rapinel et al.

2015, Bailly et al. 2008, Cazorzi et al. 2013).

1.4. The Condition of Artificial Aquatic Systems and Its Drivers

The perceived poor condition of artificial aquatic systems matches the reality of poor water quality and altered ecological structure in many man-made water bodies

(Herzon and Helenius 2008, USEPA 2009). Many artificial waterbodies support species- poor (Rosenvald, Jarvekulg, and Lohmus 2014) or otherwise undesirable communities or organisms, including disease vectors (Herzon and Helenius 2008, Mattah et al. 2017,

Wang 2012), and can spread pest species to natural habitats (Maheu-Giroux and de Blois

2007, O'Connor and Rothermel 2013). Some have also contributed to, accelerated, or facilitated flow of excess nutrients and other pollutants (Mallin and McIver 2012, Bailey et al. 2015), activation of toxicants (Herzon and Helenius 2008), interruption of movement and dispersal of desirable species (Naito, Sakai, and Morimoto 2012), increased greenhouse gas emissions (Palta, Grimm, and Groffman 2017) (Hergoualc'h and Verchot 2012), bad smells (Mao et al. 2012), and even concealed crime (Perring et al.

2013). While natural water bodies can possess the same undesirable characteristics, we are more likely to assume that artificial water bodies have a negative influence without investigation (Salomon Cavin 2013).

Are artificial aquatic systems intrinsically less biologically diverse and less functional than natural ones? It is at least plausible that humans cannot create a waterbody that supports communities as diverse or provides as many ecosystem services. One important constraint on artificial aquatic systems is that with their anthropogenic origin comes a severely shortened evolutionary, ecological, and geophysical history (Robert et al. 2017, Miller, Soule, and Terborgh 2014). To the extent that diversity and other aspects of ecosystem structure depend on slow processes of physical change and community assembly, the recent origin of most artificial systems will likely limit their function. Moreover, there are other potential limits on the condition and value of artificial aquatic systems. First, imperfect understanding of how differing designs and constructions affect ecological outcomes, and imperfect ability to reproduce natural structures and conditions, may constrain the most ecologically- motivated projects, as may be the case for many ecosystem restorations (Palmer and

Filoso 2009). Second, artificial aquatic systems are often embedded in intensively used landscapes, potentially exposing them to anthropogenic stressors and disconnecting them from diverse natural populations (Palmer and Filoso 2009). Finally, and perhaps most importantly, many artificial aquatic systems may support limited diversity and ecosystem function because they are not designed or managed to do so; in many cases their intended function may preclude or limit the provision of other services (Palta,

Grimm, and Groffman 2017). As a result, scientists, policy-makers, and the general

public have tended to accept that artificial aquatic systems will necessarily and inherently have limited value. However, these assumptions are often not subject to the same critical assessment and process-based explanations that are applied to explanations of the condition of other aquatic systems.

The poor condition of artificial aquatic systems is far from universal, and at least some, perhaps many artificial aquatic systems also have clear ecological value.

Constructed and transformed aquatic systems, whether agricultural, industrial, urban, or recreational, can sustain biodiversity (Ridder 2007), sometimes including rare and desirable species (Chester and Robson 2013). In Europe, manmade farm ponds serve as primary or important habitat for amphibians (Reid et al. 2014), birds (Sebastian-

Gonzalez, Sanchez-Zapata, and Botella 2010), invertebrates (Ruggiero et al. 2008,

Cereghino et al. 2008, Abellan et al. 2006), plants, and other species (Davies et al. 2008,

Biggs et al. 2005); indeed, European conservation proceeds from the assumption that

"artificial, ‘man-made’ ponds are not fundamentally ecologically different from ‘natural’ ones” (Cereghino et al. 2014). Some species now apparently depend primarily on deliberate artificial aquatic ecosystems for habitat (Kadoya, Suda, and Washitani 2009,

Canessa et al. 2013, Casas et al. 2011); even species new to science continue to emerge from ditches (Taylor and Robison 2016, Zang 2017, Gan and Li 2016). Equally in the U.S., habitats that we presently tend to overlook, such as a stormwater treatment wetlands, can sometimes be the best sites for reproduction of amphibians and other species with

specific hydrologic needs (Brand and Snodgrass 2010). Artificial aquatic systems, whether designed for the purpose or not, have improved water quality in critical watersheds like the Mississippi River basin (Roley et al. 2016, Perez-Gutierrez, Paz, and

Tagert 2017, Cooper et al. 2002). Without more intensive and systematic study, it remains unclear whether good ecological conditions, and the desirable ecosystem services that derive from them, are a negligible, rare, or even commonplace occurrence in artificial aquatic systems.

Making artificial aquatic systems more functional and valuable will require a mechanistic and predictive understanding of their condition and their capacity to provide ecosystems services. We propose that the science of artificial ecosystems should entertain and evaluate hypotheses about what drives variation among them as well as their differences from their natural counterparts. Like their natural counterparts, the ecological characteristics of artificial aquatic systems are likely to depend on their physical structure, the characteristics of the watershed and landscape in which they are embedded, their age and trajectory over time, and the ongoing interventions of humans for various purposes (Fig. 3). While all of these mechanisms are shaped by human design decisions, they also have clear analogs to factors commonly invoked to explain the condition of natural aquatic systems. This re-casting of rationales for why artificial aquatic systems are assumed to be in poor condition as testable alternative mechanisms allows us to consider how different decisions about the design, placement, and longevity

of artificial aquatic systems might improve their condition and value. The poor condition and seemingly inherent limitations of artificial aquatic systems could be simply a syndrome of those decisions. In our exploration of these possible causal variables for the ecological condition of artificial aquatic systems here, we focus on this management-oriented way of examining the causal variables, and barely begin to explore the possible interactions between them. We recognize that scientists with different foci could propose other valid testable hypotheses, and indeed invite them to do so, but we consider setting, time, physical design, and subsequent management of artificial waterbodies to be good intellectual places to start trying to understand these ecosystems.

Figure 3: Inherent (Left) vs. Process-Based Model (Right) of the Condition of Artificial Aquatic Systems. We suggest that scientists and policy-makers have too readily accepted model A., in which the inherent qualities of artificiality negatively impact ecosystem structure and function, rather than scientifically exploring a mechanistic model such as the one we propose in panel B, which breaks the influence of artificiality down into multiple processes.

1.4.1. Setting

Understanding how the watershed setting of artificial aquatic systems affects their ecology is important both because that understanding will be essential for better policy and management that might protect and improve their ecological condition, and because the effects of setting may obscure the ecological effects of other factors such as design, management, and time. The condition of the watershed and landscape around any waterbody influences its condition (Palmer 2009), and artificial aquatic systems should be no different in this respect. Since humans tend to create artificial aquatic systems in and around heavily modified landscapes with substantial chemical inputs like agricultural fields, roads, and parking lots, artificial aquatic systems such as ditches

tend to have lower water quality than their natural counterparts (Stewart et al. 2001,

Dudgeon et al. 2006, Burcher, Valett, and Benfield 2007). The communities of artificial aquatic systems also tend to reflect the local and regional species pools, which may include more exotic species in a restoration in a developed area (Suding 2011).

Available evidence suggests that setting does exert a strong and often overwhelming influence on artificial water bodies, and that these effects are similar to those observed in natural systems. Water quality of artificial aquatic systems such as ditches responds to catchment land use in much the same way as that of waterbodies of natural origin (Biggs, von Fumetti, and Kelly-Quinn 2016) (Veraart et al. 2017), and agricultural land cover impacts reservoirs food webs (Bremigan et al. 2008). In the Salt

River in Arizona, level of urbanization explained much of the variance in communities of plants, birds, and herptiles, for reference, restored, and accidentally restored river reaches alike (Bateman et al. 2014). A study in the Florida panhandle found that natural streams, altered streams, and ditches had similar macroinvertebrate and fish assemblages (Simon and Travis 2011). Agricultural intensification around fishponds has contributed to the rapid decline in breeding populations of black-headed gulls

(Chroicocephalus ridibundus) in central France (Francesiaz et al. 2017). More such comparisons between artificial and natural waterbodies in similar settings are needed to disentangle the effects of watershed setting from other factors that influence the condition of artificial aquatic systems.

The predictable responses of artificial aquatic systems to their watershed setting has implications for how these systems are managed and how that management could be improved. The importance of watershed land cover for reservoir water quality shapes economically motivated conservation, like New York City’s efforts to prevent development in the watersheds of its reservoirs upstate (Mehaffey et al. 2005). More generally, stream restoration is more effective in undeveloped than developed catchments (Bernhardt and Palmer 2007). While more studies are needed, the available evidence suggests that the condition of artificial aquatic systems depends strongly on their setting, and that those conditions, and the ecosystems services that depend on them, could be improved by the same watershed-scale policy and management that protects natural water bodies.

1.4.2. Time

Most artificial aquatic systems are young, both because most land change and earth moving has occurred within the past few hundred years (Hooke 2000) and because artificial aquatic systems turn over quickly than natural ones (Palmer 2009). Given the timescales over which community assembly occurs in newly formed natural streams and lakes (Brown and Milner 2012), it is likely that limited diversity of some artificial aquatic systems simply reflects their recent origin. Understanding the consequences of recent formation requires that we understand the timescales over which newly created ecosystems develop, whether they arose from anthropogenic or from geologic processes,

and potentially attain the characteristics of their older counterparts. At present, we lack both general and system-specific models of these trajectories, as well as criteria with which to judge that an artificial aquatic system has 'naturalized'.

A relatively limited set of long-term and chronosequence studies indicate that artificial aquatic systems can change in important ways over timescales that are similar to natural ecosystems, and that are relevant to decision-making (Sayer et al. 2012,

Arthaud et al. 2013, Drake et al. 2010). In restored wetlands, for example, ecological structure and function such as carbon sequestration can improve with time since intervention, and soils characteristics approach natural properties over decades

(Ballantine and Schneider 2009). Re-configured two-stage ditches can achieve soil formation and a geomorphological “quasi-equilibrium” within a decade (D'Ambrosio,

Ward, and Witter 2015). Agricultural ditches also undergo relatively predictable succession of plants and associated invertebrate communities (Palmer, Drake, and

Stewart 2013). Accidentally created water bodies also change over time, often acquiring more 'natural' characteristics. For example, gravel quarries develop more structurally complex and diverse vegetation over several decades (Rehounkova and Prach 2008). At some point in time, artificial aquatic systems may be difficult to distinguish from natural systems. Many of the small, ephemeral wetlands that sustain populations of amphibians in the Piedmont of the U.S. Southeast are likely legacies of historical human disturbance

(Hawley 2014). Such examples suggest that time eventually erases many signatures of

anthropogenic origin, and that this naturalization may change how aquatic systems are perceived and valued.

Better understanding of how time constrains the characteristics of artificial aquatic systems, and the mechanisms by which they evolve, could improve our ability to manage them, individually and as part of the broader aquatic landscape (Oosting 1942).

Goals and expectations of restorations and other interventions might need to reflect the differential responses to the same management technique, as has been observed in young and old artificial aquatic systems (Zhang et al. 2017). Deliberate management of successional stages could be, and has been used to increase the abundance and diversity of desirable species (Drake et al. 2010). Wider adoption of such approaches will require better models of succession and its dependence on design, setting, and management.

1.4.3. Design

Design is a goal-oriented process with multiple stages, including the establishment of goals, a plan to achieve those goals given constrains, implementation

(including initial construction and subsequent maintenance), and, ideally, subsequent iterations of goal-setting and redesign (2015). Decisions, whether unconscious or deliberate, at each of these stages have the potential to shape the outcomes of later stages, and ultimately the ecological character of artificial aquatic systems, including their trajectories over time and how they respond to the forcings imposed by their watershed setting. The physical structure and management of accidentally created water

bodies obviously does not depend directly on goal-oriented design, though their structure may reflect design decisions and management regimes of which they are not the object.

1.4.3.1. Design Goals

Historically, many deliberate artificial aquatic systems have been designed and maintained to provide one or a few services, such as water conveyance and storage

(Palta, Grimm, and Groffman 2017, Kaushal et al. 2015). The exclusion of many such artificial water bodies from protection within the US reflects the fact that policymakers and legal frameworks value these systems almost exclusively for their intended, fully human-oriented purposes (Zollitsch and Christie 2014). Planning for only one or a few ecosystem services, such as water storage and conveyance for flood control, can limit the ability of a deliberate artificial aquatic ecosystem to provide other services, especially when designers overbuild that system for its given purpose(s) (Palta, Grimm, and

Groffman 2017). In many cases, the design goal itself can inherently produce a major ecological cost, as in wetland drainage by agricultural ditches (Li et al. 2012, Vincent,

Burdick, and Dionne 2013), or can result in unintended disservices arising from synergies and trade-offs in ecosystem services (Bennett, Peterson, and Gordon 2009,

Power 2010). Nonetheless, many designed artificial aquatic systems also provide a range of additional ecosystem services beyond the purpose of their design (e.g. Bateman et al.

2014, Brand and Snodgrass 2010).

The design of aquatic ecosystems, including both newly constructed water bodies and restoration of degraded systems, increasingly seek to provide a portfolio of ecosystem services and functions through redesign of physical structure as well as changes in management (Hale et al. 2017, Landers 2007). Urban dwellers appreciate open expanses of water in spaces where they go for recreation (Polat and Akay 2015), and even modified or constructed waterbodies can mitigate pollutants and floods, cool the air, and provide spaces for recreational, spiritual, and community-building activities

(Perring et al. 2013). For example, the Los Angeles River, converted to a concrete flood chute and movie set for car chases in the mid-20th century, has recently become the focus of an ambitious revitalization project to improve water quality and sustain wildlife while also providing a greenway and other recreational opportunities (Landers 2007).

Similar redesigns of channelized rivers have already demonstrated the benefits of design for a range of ecosystem services (Smith et al. 2016). The Landscape Architecture

Foundation has endorsed projects throughout the U.S. and around the world with similar methods and goals, specifically including stormwater management, water conservation, water quality, flood protection, and groundwater recharge alongside other environmental, social, and economic goals (LAF 2016). Deliberate artificial aquatic ecosystems like these tend to remain primarily human-oriented, and not ecologically- oriented, in their goals, however. Even restoration designs often explicitly and unapologetically include human-specific concerns, such as ease of maintenance,

accessibility, recreational appeal, aesthetics, regulatory standards, finances, and property lines, alongside more ecologically oriented values (Lave 2012, Doyle et al. 2015,

Podolak and Kondolf 2016).

1.4.3.2. Planning and Construction

The reduced physical complexity of many artificial aquatic systems, such as concrete-lined channels, obviously limits their value as habitat and potential for improvements in water quality (Dollinger et al. 2015). Restorations that seek to improve these values therefore often focus on the (re-) introduction of heterogeneous structures that are more similar to natural systems (Landers 2007). Such designs (and their implementation) can be constrained or flawed in ways that limit their ecological value, including by insufficient scientific understanding of how design features and subsequent management influence eventual outcomes (Palmer 2009, Wohl, Lane, and

Wilcox 2015, Palmer, Menninger, and Bernhardt 2010, Bernhardt and Palmer 2011).

However, many such systems are also affected by intensive land use and short lifespans

(Bernhardt and Palmer 2007); artificial systems whose structure mimics that of natural systems can support similar biotic communities when water quality is high (Simon and

Travis 2011).

Conversely, engineering research on designer ecosystems constructed for a very specific subset of aquatic ecosystem services, such as water quality improvement, clearly demonstrates that design plays a role in how effectively these systems achieve their

purpose. For example, plant species choice in wastewater treatment wetlands affects speed and removal efficiency of different forms of nitrogen (Kumwimba and Zhu 2017).

In wetlands constructed to remove pharmaceuticals from water, design choices of substrate, plants, and regimes of hydrology, temperature, oxygen, and light all affected removal efficiency, which varied from compound to compound in ways apparently related to microbial processes (Verlicchi and Zambello 2014, Li et al. 2014). While much variation remained unexplained even in these relatively controlled systems, they do demonstrate that how an ecosystem is constructed affects its ecological behavior.

Physical, legal, and cultural constraints exert strong control on goals and resulting designs. For example, restoration efforts are typically constrained and otherwise impacted by funding, land ownership, and other social and economic variables (Palmer 2009, Palmer and Filoso 2009, Bernhardt and Palmer 2007); restorations can have a wide range of intended outcomes (Suding 2011). Morphology of stream restorations depends in predictable ways upon funding source and legal purpose, and whether the metric for success is stream length (resulting in very sinuous designs) or some other characteristic (Doyle et al. 2015). Stream restoration in general has tended towards a single-channel, S-shaped, meandering morphology that conforms to longstanding aesthetic concerns (Podolak and Kondolf 2016), reduces maintenance

(Lave 2012), and maximizes mitigation credits than to local natural history (Doyle et al.

2015). One indication of the limitations of many restoration projects is the finding that

accidental aquatic systems can sometimes provide equal or greater services compared to deliberate, designed systems. For example, “accidentally restored” wetlands at stormwater outflows in the dry bed of the Salt River in Arizona had greater wetland plant richness and cover than comparable actively restored sites, though the reverse was true for birds and herptiles (Bateman et al. 2014).

Changes in goals often dictate substantial changes in the physical structure of artificial aquatic ecosystems. Two-stage ditches, in which miniature floodplains are constructed alongside existing conveyances (Powell et al. 2007), can significantly reduce concentrations of phosphorus and other nutrients, turbidity, and total suspended sediments (Davis et al. 2015, Mahl et al. 2015). Their nutrient removal efficiency compares well with, and can complement, other farmland best practices, like planting cover crops (Roley et al. 2016). When properly constructed according to fluvial principles, these ditches can remain functionally stable, without maintenance, for years

(D'Ambrosio, Ward, and Witter 2015). Thus, water infrastructure of agricultural landscapes can be designed, and successfully re-built, to achieve a wider range of goals than water conveyance, but additional land area and design and construction effort may be required.

1.4.3.3. Management and policy

Ecosystem function and services of artificial waterbodies likely depends on the management they receive after construction as well, just as in natural waterbodies. In

reservoirs, the ability of an artificial aquatic system to provide ecosystem services may depend more heavily on ongoing policy and management than on the specifics of the initial design (Cordell and Bergstrom 1993, Branche 2017). In ditches, management strategies, including dredging, mowing, chemical weeding, burning, and regulation of water depth can have significant impacts on ditch biodiversity and water quality (Twisk,

Noordervliet, and ter Keurs 2003, Needelman et al. 2007). However, ongoing management and maintenance, like initial design, often does not include these potential outcomes in its considerations, instead opting to continue to focus on relatively few, highly human-oriented goals (Dollinger et al. 2015). Such decisions about ongoing management and policy, however, can, at least in theory, be revised to reflect changing goals.

In complex landscapes, achieving a portfolio of ecosystem services often requires both structural changes and ongoing active management of artificial aquatic systems.

Ditches and canals draining ranchlands in the watershed of Lake Okeechobee were not traditionally managed for water quality or conservation purposes, even though they house large native animals and other species of interest (personal observation, Smith,

Boughton, and Pierre 2015). Establishment of Total Maximum Daily Loads for phosphorus in the lake and its tributaries (Fluck, Fonyo, and Flaig 1992, USEPA 2008) prompted creative responses including regional collaborations among various government agencies, nonprofit conservation organizations, a local scientific research

station, and ranchers to raise and actively manage water levels to flood ditched wetlands. This strategy removed phosphorous more cost-effectively than did constructed storm water treatment areas (Johns 2011), while also increasing wetland vegetated area and vertebrate abundance (Bohlen et al. 2014). Multi-stakeholder management of artificial aquatic systems with ecologically oriented goals could prove a cost-effective way to increase ecosystem services at a similarly regional scale in other locations, a complementary tool to more traditional restoration.

Accidental waters, which often receive little to no management attention, can provide comparable but non-overlapping ecosystem services to both deliberate and more natural waterbodies. Abandoned features, especially within broader abandoned landscapes and even in the hearts of cities, can provide habitat for urban-avoiders and other organisms that survive best away from humans and human intervention (Hawley

2014). Abandoned areas can contain accidental artificial waterbodies sustaining both human and nonhuman life, and functioning as little pockets of biodiversity (Palta,

Grimm, and Groffman 2017). Accidental urban wetlands can also mitigate nutrient pollution flowing from cities to downstream in natural waterbodies (Palta, Grimm, and

Groffman 2017). Two European species of damselfly were believed extinct for decades, until rediscovered, separately, in former industrial and mining areas “not usually explored by biologists.” Notably, conservation interventions “focused on returning habitats to a ‘natural’ state” intended to boost one of those damselfly populations

actually backfired (Harabis and Dolny 2015). These observations suggest that active intervention, for non-ecological and even ecological goals, can limit the ecological value of artificial aquatic systems; other neglected artificial habitats in our midst could hold similar surprises.

1.4.3.4. Monitoring, learning, and iteration

One of the criticisms of many restorations is that they require ongoing, often expensive management to avoid reverting to a degraded state, which some scientists consider a failure of resilience. Part of the problem with declaring restoration success or failure is that goals for a specific restoration are often unclear and may change through time (Baker and Eckerberg 2016), but for most aquatic restorations are never evaluated

(Palmer et al. 2007, Suding 2011). Monitoring protocols often focus on easily quantifiable measures that ensure mitigation credit, rather than landscape-scale and long term ecological contributions (Doyle et al. 2015). Current practices of stream and wetland restoration may not be well configured for learning and for adapting designs to improve environmental outcomes (Ross et al. 2015). More broadly, the exclusion of artificial aquatic systems from policy protections eliminates an important motive for monitoring.

In the UK, a recent precipitous decline in farm pond numbers and services in the U.K. sparked conservation concern and action (Jeffries 2012). The U.S. lacks the monitoring data necessary to characterize trends in its small artificial aquatic systems and respond accordingly.

Moving forward, adaptive management, designed experiments (Felson and

Pickett 2005), reconciliation ecology (Rosenzweig 2003b), and other ecologically based ways of improving designs may change the outlook for deliberate waterbodies. A future increase in the acceptance of novel ecosystems might allow the creation of new types of waterbodies designed to provide similarly novel suites of ecosystem services (Hobbs,

Higgs, and Harris 2009, Higgs 2017). Even the tendency for less regulation of more highly and accidentally modified (Zollitsch and Christie 2016, 2014, Hough and

Robertson 2009) and smaller waterbodies (Sayer 2014, Biggs, von Fumetti, and Kelly-

Quinn 2016), particularly in the U.S., constitutes an opportunity; this quality could make them comparatively easy and low-cost systems in which to study, test, and implement novel ecological design ideas (similar to Felson and Pickett 2005, De Meester et al. 2005,

Koetsier and McCauley 2015). Together with the repetitive design and construction of many such waterbodies, like ditches and ponds, the manipulability of artificial aquatic ecosystems makes them prime sites for natural experiments (Diamond 1983) and designed experiments (Felson and Pickett 2005). Irrigation canals can serve as “lotic mesocosms,” ideal because of their known histories, predictability, and accessibility

(Koetsier and McCauley 2015). Ditch network structure recently served as a good system within which to model possible alternative stable states in primary producer structure

(van Gerven et al. 2017). Science in artificial aquatic ecosystems could contribute substantially to broader ecological theory and practice. While win-win design decisions

to support multiple desired ecosystem services and other goals can prove challenging to envision and implement, even in artificial aquatic systems, these waterbodies remain sufficiently understudied that exploration of the many remaining questions around them likely has many win-wins, in terms of furthering both applied and theoretical science, left to yield. Judging from the apparent recent uptick in conference presentations and papers pertaining to artificial aquatic systems, many scientists have realized as much; we hope that the conceptual structures introduced in this article will assist in future such work.

1.5. Artificiality and Perception of Ecosystem Value

The concept of artificiality, its associated dichotomy between human and nature, and its connotations for valuation, have deep roots. One of the earliest abstract concepts

American children master is the difference between artificial and natural, in terms of origin; they learn to tell whether an object is “made by people or something that people can't make” (Petrovich 1999). Accordingly, Western culture has a long tradition of elaborating upon the natural/artificial dichotomy (Chakrabarty 2014) and including it in value systems (Bensaude-Vincent and Newman 2007). In American history, wild nature provided a divine purpose for European settlers, a spiritual rejuvenation for

Romanticists and early conservationists like Muir, a source of strength to manage for technocrats, and a rallying point for complex unity among environmentalists (Purdy

2015). Today, untouched wilderness “exists nowhere but in the imagination” (2007), yet

the concept of pristine wild nature continues to exert a strong pull. A recent psychological study found that subjects preferred environments when told that they were natural (McMahan et al. 2016), perceiving them “less dangerous, cleaner, and more plentiful” than those already exploited by other humans. We argue that research and policy-making about artificial aquatic systems reflects this cultural subordination of artificial things to the natural and wild, inherited from broader contemporary Western culture (McMahan et al. 2016, Rozin et al. 2004).

The past several decades have seen ferocious clashes in environmental philosophy and conservation biology over the role of artificial ecosystems. In the 1980s, prominent ethicists lampooned restoration on the grounds that “faked nature is inferior” in much the same way that an art forgery is inferior, because it is “a product of contrivance,” lacking “causal continuity with the past” (1982), and that man-made natural areas represent “domination, the denial of [the] freedom and autonomy” that defines nature (1983). Some philosophers have some softened this dichotomy, viewing it as a gradient or broadening the criteria that constitute a necessary fidelity to nature

(Callicott 1995, Light, Thompson, and Higgs 2013, Higgs 2003). While naturalness remains valuable by all standards of environmental ethics, many ethicists increasingly distinguish categories (or dimensions) of naturalness, including “as a physical property of species and ecosystems,” such as native biodiversity, and “as a quality of processes that are free of human intervention,” (Ridder 2007). This particular pair of categories,

posited as a distillation of values already in wide use, corresponds well with our proposed axes of degree of modification and level of intentionality, which U.S. policy tends to reflect (Zollitsch and Christie 2016, 2014, Hough and Robertson 2009).

Discussions of humanity’s relation to authentic nature intermingle with and parallel debates within conservation biology and broader environmental science and policy. Proponents of traditional wilderness- and biodiversity-based conservation have reacted with alarm to “new conservation,” a loose grouping of movements that include human and socio-economic goals, such as poverty reduction, in their conservation plans

(Soule 2013, Miller, Soule, and Terborgh 2014). While restoration has become a widely accepted practice, novel and designed ecosystems are at the battle lines between “new conservationists” who would like to include them in conservation plans and more traditional conservationists who would not (Hobbs et al. 2006). The difficulty in reconciling these perspectives (Tallis and Lubchenco 2014, Matulis and Moyer 2017) may arise in part from different priorities between the physical and ecological condition of ecosystems versus their freedom from human intervention, and in part from conflicting views about whether human intervention inherently degrades ecological condition.

1.5.1. Interactions Between Perception and Condition in Artificial Aquatic Systems

The presumption that artificial aquatic systems have little ecological value is important because it promotes neglect. People make management decisions about aquatic systems not on the basis of perfect factual knowledge of the state of these 40

systems, but instead upon how they perceive them (Gelcich and O'Keeffe 2016). Natural aquatic ecosystems in poor condition often retain perceived value, which restoration seeks to regain, however little remains (Francis and Hoggart 2008). However, traditionally, scientists and policymakers have regarded artificial ecosystems as relatively low in ecological value (Salomon Cavin 2013), regardless of their actual function and services (Fig. 3). Conversely, a high-functioning artificial system may be overlooked (Chester and Robson 2013). For example, the 250km of canals of the North

Poudre Irrigation Company near Fort Collins, Colorado, supported 92% of wetland area in the 23,300-hectare service area through leakage. In spite of the ecosystem services these wetlands provide, this leakage is considered inefficient use of a scarce resource, and may cease as irrigation practices change (Sueltenfuss et al. 2013). Perceived value influences design and management, which, along with any other more direct impacts of artificiality, in turn influence ecological condition (Fig. 5). If, as appears prevalent, perceived artificiality downgrades perceived value of aquatic ecosystems, and management and policy decisions reflect this lower valuation in low expectations and low protections for artificial waterbodies, then assumptions of the low quality status of artificial aquatic systems could be self-fulfilling.

The divergent consequences of this positive feedback loop may be illustrated by examining water management policy in the U.S., where the Clean Water Rule prioritized the exclusion of many artificial aquatic systems from jurisdiction under the Clean Water

Act (USEPA and USACE 2015). In contrast, the European Union’s 1996 Water

Framework Directive resolved to gradually expand protection “to all waters, surface waters and groundwater” (2014a). In line with this inclusive view of aquatic ecosystems, pond degradation (Declerck et al. 2006) and loss is a stated conservation concern for the

EU (Jeffries 2012) and NGOs (2013). Freshwater Habitats Trust’s Million Ponds Project aims to “to reverse a century of pond loss, ensuring that once again the UK has over one million countryside ponds”, and claims more than 1,000 ponds created in 2008-2012, housing about 50 rare and declining species (2014b). Meanwhile, European researchers continue to explore pond conservation measures (Cereghino et al. 2014), including management options that improve habitat quality in existing ponds (Sayer et al. 2012,

Canals et al. 2011). Similar research and conservation activity is progressing for British and other European ditches (Clarke 2015, Shaw et al. 2015). Assuming even modest success of such efforts, the condition of artificial aquatic systems in the EU is likely to improve, while the quality of artificial water bodies in the US is likely to decline.

Europe’s example suggests that how people regulate, perceive and manage farm ponds, and other artificial aquatic systems, does impact conservation outcomes.

Getting the policy, management, and science around these systems right matters not just ecologically and in terms of ecosystem services provisioning, but for social reasons as well. Under-managed artificial waterbodies, particularly environmentally hazardous ones, may often occur in already at-risk communities. Two well-studied

examples of 20th-century environmental injustice in the United States, in Anniston,

Alabama, and Hyde Park, Georgia, both involved predominantly black communities contaminated and sickened in part by ditches bearing water laced with toxic industrial waste (Spears 2014, Checker 2005). Recently, hog waste lagoons associated with industrial swine facilities in eastern North Carolina have proved resistant to regulation despite repeated flooding during hurricanes and tropical storms and persistent strong detrimental effects on the health and quality of life of neighbors, who are disproportionately black and low in income (Setzer and Domino 2004, Kim, Goldsmith, and Thomas 2010). Thus, what artificial aquatic systems go unregulated may say as much about what we socially undervalue as what we ecologically undervalue.

Relatedly, the same accidental wetlands that host birds and remove nitrogen in the Salt

River in Phoenix, Arizona, provide somewhat unsafe, legally unauthorized sources of water and cool places to rest for homeless people (Palta et al. 2016, Palta, Grimm, and

Groffman 2017, Bateman et al. 2014), which calls into question the design of non-aquatic infrastructure whose functions may have been deputized to or externalized on artificial aquatic ecosystems. When science and policy overlooks artificial aquatic systems, it risks overlooking the people impacted by them as well.

Figure 4: Value Axes for Aquatic Ecosystems. The solid line represents a traditional axis for the value of aquatic ecosystems. The dashed lines parse out artificiality from ecosystem services provisioning along this traditional axis.

1.6. Conclusions - Artificial Aquatic Ecosystems in Hybrid Hydroscapes

Artificial aquatic systems comprise a substantial, perhaps predominant, and likely enduring component of the modern hydroscape. Because the sheer extent of artificial aquatic ecosystems may, by some measures, rival that of natural systems, they have the potential to play an important role in both conservation and in the provision of ecosystem services within these hybrid aquatic landscapes. The premise underlying reconciliation ecology (Rosenzweig 2003b) is the insufficient extent of relatively undisturbed habitats to preserve anything but a fraction of extant species. In some regions, it may be difficult to enact any sufficiently wide-reaching biodiversity

conservation policy without them (Rosenzweig 2003a). Because artificial aquatic systems are interwoven with, rather than separate from, natural elements of the hydroscape, improvements in the condition of artificial systems may benefit natural water bodies as well (Reid et al. 2014). Thus, plans to improve land and water management should target artificial aquatic systems as well as those of natural origin (Rosenzweig 2003a).

To realize greater socio-ecological benefits from artificial aquatic systems, we need to understand not just their current value, but the possible provisioning of ecosystem services. This understanding will require, first and foremost, better assessments of the extent and condition of artificial aquatic systems. Improving that condition will require that we suspend our conventional assumption that artificial aquatic systems are intrinsically inferior; instead, we need more hypothesis-driven study that evaluates the factors, such as watershed setting, physical structure and design, time, and management, that influence their ecological condition. We will need to move beyond this initial exploration to more thoroughly consider interactions among these drivers and alternative ways of framing the mechanisms underlying artificiality (e.g., physical vs. biological), first conceptually and then through well-controlled studies.

Because the very way we perceive artificial aquatic systems may affects their ultimate condition and value, effective management of the modern hybrid hydroscape may require reconsidering cultural norms about the concept of artificiality, even undoing our deeply held notions about a human/nature dichotomy. Environmental

scientists, and our cross-disciplinary collaborators, must first take on such efforts in support of our own work, but can also play a role in helping policy-makers and others meet these challenges.

Figure 5: Conceptual Diagram of the Role of Artificiality in the Management and Services Provided by an Aquatic Ecosystem. Top part a, the currently prevailing process model, depicts an approximation of how environmental scientists appear to typically think of the role of artificiality in impacting ecosystems. The lower part b, our proposed replacement, suggests that while artificiality may impact ecosystem function directly through mechanisms yet little elucidated, we are more certain that it impacts the perceived value of ecosystem services. Because perception impacts policy, policy affects reality, and reality impacts perception, this proposed replacement process model for the role of artificiality in aquatic ecosystems sets up a positive feedback loop.

2. Urban Ditches and Natural Creeks Support Similar Invertebrate Communities in a Minimally Developed California Watershed

By Chelsea Clifford1, Jeffrey G. Holmquist2, Kristofor Voss3, Jutta

Schmidt Gengenbach2, James B. Heffernan1

1. Duke University

2. UCLA White Mountain Research Center

3. Regis University

2.1. Introduction

Irrigation ditches are a dominant feature of the US hydroscape. In 2012, the

United States had 27,967 km2 of farmland irrigated by 42.7 km2 unlined ditches, and

8,469 km2 of farmland irrigated by 150 km2 of lined ditches (e.g. Allan 2004, Meyer, Paul, and Taulbee 2009, Stewart et al. 2001), yielding a total irrigated farmland area somewhat larger than Lake Tanganyika. These figures only account for farmland, and not for irrigation ditches and canals that traverse urban or undeveloped systems; the actual footprint of these systems is likely much larger. Agricultural drainage ditches occupy a much larger area (USDA 2014), and some ditches function for both drainage and irrigation at different times.

Despite the commonness of these irrigation ditches, ecologists have traditionally considered them primarily in terms of degradation of their natural water sources, via

abstraction’s impact on natural flows (USDA 2015, Bunn and Arthington 2002).

Scientists have rarely evaluated irrigation ditches and other artificial aquatic systems as ecosystems in their own right, and attempted to explain their condition in a process- based manner (Chapter 1). Similarly, federal (e.g. Dynesius and Nilsson 1994, Malmqvist and Rundle 2002) and state (USEPA and USACE 2015) regulations tend to deprioritize ditches as legally protected waters (Chapter 1). Neither scientists nor policymakers have much information about ecological structure and function of US irrigation ditches yet.

Globally, conservation biologists increasingly recognize that artificial water bodies can also provide valuable habitat (Zollitsch and Christie 2014, 2016) and other ecosystem services such as water purification and pest control (Chester and Robson

2013). Irrigation ditches specifically can host birds (Herzon and Helenius 2008), herptiles

(김미란 et al. 2013), fish, arthropods (Katano et al. 2003), turtles (Bataller, Forteza, and

Sancho 2008), mussels, (Nakano et al. 2015), and plants (Meier et al. 2017). including rare species (Langheinrich et al. 2004, Kadoya, Suda, and Washitani 2009). The net effect of irrigation ditches on aquatic habitat thus includes both negative impacts on existing aquatic systems and novel habitat provision.

To understand the ecology of a whole, partially abstracted hydroscape, we need to understand not just degradation, but also the habitat provisioning of irrigation ditches. We need to know whether these systems respond to the same drivers as natural systems, such as watershed setting, disturbance and succession, watershed policy, and

local management, and/or to drivers unique to artificial systems, such as initial construction design, or other intrinsic influences on artificial water bodies (Chapter 1).

These variables are often conflated. We might expect that irrigation ditches, like streams

(Allan 2004, Walsh et al. 2009), would respond negatively to urban and agricultural stressors in their watersheds, and indeed, acequia irrigation waters in New Mexico apparently do respond to human land uses in their watersheds (Fernald et al. 2012). It is often difficult to separate these impacts from the impacts of artificiality itself, though, because artificial aquatic systems like irrigation ditches typically occur in watersheds with agricultural and development impacts (Doubek and Carey 2017). Few studies have compared adjacent and connected irrigation ditches and rivers to control for differences between the two system types in water, except to describe their connectivity in different seasons, as for the passage of fish (Cowley, Wissmar, and Sallenave 2007) or the processing of nutrients (Mortensen et al. 2016).

Bishop, California provided a unique system to test these questions, because a natural creek and its ditched diversions run in a broadly parallel fashion, with the same water flowing back and forth between them. After this water flows out of the Sierra

Nevada Mountain snowmelt and through largely undeveloped desert into the town of

Bishop, in both streams and ditches, it is equally subject to the small-town version of urban stream syndrome (Tessier, Sfreddo, and Ouin 2009, Langheinrich et al. 2004) and agricultural influences (Walsh et al. 2009). Moreover, the ditches of Bishop were

constructed over the same benthic sediments. The primary difference between the channels, aside from the processes that produced them, is age; the ditches range in age on the order of decades, whereas the creeks arose with the Sierra Nevada mountains.

Age, and time for succession and naturalization, is another factor that can affect both natural (Brown and Milner 2012) and artificial (Ballantine and Schneider 2009) aquatic systems. Here we could compare habitat provisioning by ditches and their natural analogs while minimizing and controlling for the influence of upland land use, providing a rare opportunity to directly assess the role of channel origin. We sampled benthic macroinvertebrate communities, and anticipated that they would vary by season and substrate, as in natural systems (Moss 2008), but also between artificial and natural waterways.

2.2. Materials and Methods

2.2.1. Study area

All samples were collected within the town limits of Bishop, California, in Bishop

Creek and its water network (Fig. 6). Water flows from the Sierra Nevada Mountains eastward through Bishop Creek and a complex system of irrigation ditches that channel water from the creek to a canal, through which most of it eventually travels to Los

Angeles. This system creates many kilometers of artificial, small streams. Within the town, the water flows through a variety of agricultural and residential lands, including under and alongside roads. In suburban-style subdivisions especially, this water has

significant aesthetic value. Residents landscape their small streams as they wish, creating ponds and waterfalls, lining and edging them with stones, brick, concrete, lawn, or other plants, and creating space for ducks and fountains. Even here, people put their share of water to practical use as well, and dam or pump it to flood yards, just as nearby farmers likewise irrigate pastures. Beyond irrigation, this ditch water is often the sole source of water for the stock grazed there. Thus, on its journey through Bishop, this water is exposed to fossil fuel hydrocarbons, lawn clippings, garden gnomes, pesticide, herbicide, fertilizer, manure, and other anthropogenic run-off. Irrigation may result in temporary lowering or drying of downstream reaches.

Figure 6: Bishop, California. All 50 samples included in all analyses, across years and substrate types, are shown, together will all area waterways in the National Hydrography Database. Water flows from the Sierra Nevada Mountains to the left and west of this image, towards the Owens River to the right and east.

A preliminary survey of Bishop’s waterways in January 2010 suggested that substrates in the particle sizes of sand or cobble on the Wentworth scale {Wentworth,

1922 #1540} predominated, sometimes covered with macrophytes or leaf litter. Substrate dominance by larger, smaller, or intermediate particles was rare.

The natural streams of the branches of Bishop Creek were generally wider and deeper than artificial streams, with an average width and depth at winter sample sites of

6.6m and 23.2cm, respectively, compared to winter artificial sites’ average width of 3.8m and depth of 18.0cm. Natural streams also flowed slightly faster and had somewhat more vegetative cover at sampling sites in winter than artificial streams. While average stream temperature was similar across all stream sites in winter, about 4.5°C, and probably 15-20°C on average in summer based on a few 2010 measurements, the smaller artificial streams appeared to have slightly more variable temperatures than natural streams did.Temperature and distance east of the fork of Bishop Creek, down the valley, were not strongly correlated (r2=0.102) in winter.

2.2.2. Study design

Twelve summer samples were collected in late July 2008, and supplemented with

40 winter samples in late January and February 2010. In each season, samples were evenly distributed between predominantly sandy sites and rocky, predominantly cobble sites, and between natural creek sites and human-dug “artificial” sites. This sampling scheme yielded four evenly sampled categories, artificial sand (AS), artificial rock (AR), natural sand (NS), and natural rock (NR.) Suitable sampling sites featured at least 50% of the desired dominant substrate, either cobble for rock sites, or sand for sand sites. In winter, on average, the rock sites featured about 75% cobble and 10% pebble, and sand sites featured about 70% sand and 15-20% gravel.

Summer sites were selected randomly from properties with access already granted, and concentrated in West Bishop. Properties constituting prospective sampling

locations were visually assessed for quantity of suitable sand and cobble sample sites.

This information was used to determine what approximate proportion of the total potential sample area of each of the four categories occurred at each property. Three samples per category were distributed among the properties, using a weighted random distribution based upon these proportions.

In winter, sampling sites were randomly selected from a Bishop Water

Associations map of the streams and properties within the system. Two areas not sampled were the Bishop Paiute Reservation, where the origin of streams was not always clear, and Los Angeles Department of Water and Power land, where access was not granted. If permission to sample a randomly selected property was not granted, then permission was sought on the adjoining properties through which the selected stream flowed, in order of proximity to the randomly selected sample site, before the site was discarded for a new random selection.

In both seasons, within each randomly selected property, the precise sample site was the closest suitable substrate to the point selected randomly from ten roughly equally spaced points along the length of the stream. In broad streams with suitable substrate across much of their width, a similar procedure was followed to determine where in the width of each stream to sample. The order of sampling was also randomized, but in rounds of one of each of the four substrate/origin categories at a time, in random order, before repeating categories. All sampling occurred in daylight.

Each sample was collected within a 0.093m2 area of a Surber net with 1mm2 mesh

(*). Substrates were disturbed by hand to a depth of approximately 10cm, until all substrate particles had been roughly brushed. Organisms were live-sorted from the resultant samples, and killed and stored by freezing for later identification.

Identification of each organism in melted samples, to order and often to family, was executed using Merritt and Cummins’s Introduction to the Aquatic Insects of North America

{, 2008 #1541} and a dissecting microscope, generally at 10x magnification. Individuals in each taxon in each sample were counted. Because Oligochaeta often broke apart during sample collection and freezing, rendering their counts dubious, Oligochaeta were excluded from analyses. After identification, all organisms were preserved in 70% ethanol.

2.2.3. Data analysis

Raw counts of individuals per Surber sample were converted to individuals per square meter by multiplying by 10.75. Mean and standard deviations of individuals per square meter per taxon for each independent variable factor category were calculated. In one sample, from a summer, natural, and cobble site, incomplete preservation prevented more detailed identification of Trichoptera, so this sample was excluded from all calculations of Trichoptera sub-taxa and from all further, distance matrix-based analysis.

Abundance data were log-transformed (log10(x+1) in order to achieve normality.

One sample, from a natural, sand site in winter with a great deal of leaf litter and

organic matter, returned exclusively Oligochaeta, leaving values of zero for all taxa analyzed, so this sample was omitted from all further, distance matrix-based, analysis.

Thus, 50 samples remained for community-level data analyses in R.

The Bray-Curtis distance between sample invertebrate communities was calculated using function “vegdist” (1996). Contrasts between winter and summer, cobble and sand, and artificial and natural were set up using the functions “contrasts” and “model.matrix” in package “stats” (Oksanen et al. 2015). PERMANOVAs comparing invertebrate communities within each pair were executed using the “adonis” function (Bates et al. 2014), with 999 permutations each. Gower distance between samples in treatment, based solely on the independent factor variables of origin, substrate, and season, was calculated using the “daisy” function (Oksanen et al. 2015).

Physical Euclidean distance between samples, based on UTM coordinates, was determined using function “earth.dist” (Goslee and Urban 2007).

A non-metric multidimensional scaling analysis (NMS), function “nmds”

(Vavrek 2011), further elucidated differences between invertebrate samples based on origin, substrate, season, and longitude. Longitude served as a post-hoc proxy for watershed urbanization, with the assumption of more urban influence upstream as waters flowed east (Fig. 6). A stepdown NMS with 1-6 dimensions, 10 ordinations, and

1,000 maximum iterations revealed that a minimum number of dimensions for low stress and high r2 was three, so three axes were used for the rest of the analysis. This

NMS, with three dimensions, 1000 ordinations, and 10,000 maximum iterations, was rotated so that power to explain variability was highest for axis 1 and lowest for axis 3 through a Principle Component Analysis (PCA) using the function “princomp” (Goslee and Urban 2007). The proportion of variability explained by each axis was multiplied by the overall r2 to calculate the percentage of overall variability explained by each axis.

The weighted average score of each taxon within the NMS was calculated using function

“wascores,” and the vector of each factor, of origin, substrate, and season, was calculated using function “envfit” (Bates et al. 2014).

We assessed raw count of taxa included in the above analyses, and count of

Ephemeroptera, Plecoptera, and Trichoptera individuals, in samples as a function of longitude via Poisson regressions, “glm” (Davies, Ihaka, and Team 2015). We first converted longitude to degrees east of a major upstream fork in Bishop Creek, located at

118.459811°W, and took the negative natural log of these values, to meet assumptions of heteroscedasticity. We started with full models including this transformed longitude variable and factor variables of season, substrate, and artificiality as predictor variables, and removed least contributing variables in a stepwise fashion until all variables contributed significantly.

Finally, we used a simple ANCOVA, “aov” (Oksanen et al. 2015) to assess the relative impacts of spatial distance between sites, artificiality, and the interaction of the two on Bray-Curtis community distance.

2.3. Results

The three most common invertebrate taxa across all samples were Elmidae,

Chironimidae, and Ephemerellidae. As expected, Plecoptera and other cobble- dependent taxa clustered with the cobble sites (Fig. 7). All leeches (Rhynchobdellida) and most Ephemerellidae were caught in winter, whereas Simuliidae and Tipulidae occurred primarily in summer. A few taxa occurred primarily in a unique grouping of factor variables. The four Caenegrionidae captured all came from three artificial cobble sites in the southwestern half of Bishop in winter. All Amphipoda occurred in one natural sand site, in Bishop City Park, in winter, which also supported the one

Ceratopogonid caught. All Ostracoda and Copepoda came from one western artificial cobble site in winter. Of the 17 Gastropoda caught, all but two came from artificial cobble.

Invertebrate assemblages differed by substrate (F= 5.43, df=1, p=0.001) and by season (F= 2.88, df=1, p=0.015), but not by artificiality of origin (F= 1.34, df=1, p= 0.235), according to PERMANOVAs. Similarly, in NMS results (Fig. 8), vectors (Fig. 7) for substrate (r2=0.35, p=0.001) and season (r2=0.25, p=0.005) were significant, but not the vector for artificiality (r2=0.08, p=0.285). Communities clustered by substrate on axis 1 and season on all axes, especially 2 and 3, but assemblages in natural and artificial streams overlapped almost entirely on all axes (Fig. 8).

Longitude, with which increased urbanization, ultimately best explained

variation in invertebrate assemblages according to the NMS (F=8.77, df=1, p=0.001).

Figure 7: Weighted Averages of Taxa and Vectors of Independent Variable Factors Within Nonmetric Multidimensional Scaling Analysis. Rows represent different combinations of axes from the same NMS. Vectors for substrate (r2=0.34, p=0.001), season (r2=0.25, p=0.005), and longitude (r2=0.39, p=0001) were significant, but not the vector for origin (r2=0.08, p=0.285). Note that axis scaling varies.

Similarly, communities with high scores on the first NMS axis correlated with higher,

Figure 8: Nonmetric Multidimensional Scaling Analysis Plots. Each row is the same plot repeated thrice, with different colors and symbols in each column identifying the samples by different factors, by origin, substrate, and season. The rows represent different combinations of axes from the same NMS. Assemblages differed by substrate (F= 5.43, df=1, p=0.001) and by season (F= 2.88, df=1, p=0.015), but not by artificiality of origin (F= 1.34, df=1, p= 0.235).

more easterly longitude, as well as with sandy substrate and summer (Fig. 7). Raw count of the taxa used in the NMS, and number of EPT individuals, also declined asymptotically moving from west to east in the primary direction of water flow (Fig.6).

Declines were nearly identical for both indicators across artificial and natural sites, and the EPT decline was identical across seasons; the remaining curves (Fig. 9) reflected fewer taxa and less EPT in sand, and somewhat less taxonomically rich communities in winter. Neither substrate nor season affected the rates of decline as reflected by the shapes of the curves (Fig. 9). Taxa richness was best explained by degrees longitude east of the fork of Bishop Creek (z=3.335, df=49, p<0.001) and the factor of sand vs. cobble (z=

-3.020, df=49, p=0.002), and marginally by season (z=1.900, df=49, p=0.056). EPT count was strongly influenced by degrees longitude east of the fork of Bishop Creek (z=-6.788, df=49, p<0.001) and substrate (z=-5.743, df=49, p<0.001).

Natural and artificial communities did differ in one respect, namely how different communities at each site were from communities at sites nearby (Fig. 10). The

ANCOVA indicated that artificiality of origin does not itself directly affect the log of the

Bray-Curtis distance between communities (F=0.055, df=1, p=0.815), and all sites become less similar in macroinvertebrate community with distance (F=281, df=1, p<0.001); however, artificial-artificial pairs and natural-natural pairs differed in how community similarity decayed over spatial distance (F=72.0, df=1, p<0.001). Dissimilarity among

artificial sites increased over a distance of 1 km, but not beyond. At short distances,

natural sand summer

10 artificial 10 cobble 10 winter 8 8 8 6 6 6 Taxa Taxa count Taxa count Taxa count 4 4 4 2 2 2

-118.45 -118.42 -118.39 -118.45 -118.42 -118.39 -118.45 -118.42 -118.39

Longitude Longitude Longitude

natural sand summer 15 artificial 15 cobble 15 winter 10 10 10 5 5 5 EPT per square meter EPT square per meter EPT square per meter EPT square per 0 0 0

-118.45 -118.42 -118.39 -118.45 -118.42 -118.39 -118.45 -118.42 -118.39

Longitude Longitude Longitude

Figure 9: Impact of Longitude on Invertebrate Communities. Longitude axis corresponds to the flow of water from west to east through the town of Bishop, i.e., with upstream urbanized area and heated valley floor. Taxa count is a raw count of the taxa represented in Figure 3 . EPT is Ephemeroptera, Plecoptera, and Trichoptera. Each point corresponds to one sample. Regression lines on the graphs reflect simple Poisson models of response variables as functions of transformed longitude. natural-natural pairs were more similar than artificial-artificial pairs, and their similarity declined over a range of 2-3 km, ultimately reaching a greater level of dissimilarity than that of distant artificial -artificial pairs.

Figure 10: Decay of Community Similarity with Distance Differs Between Natural and Artificial Sites. Only natural-natural and artificial-artificial pairings were included. Lines are LOWESS-smoothed curves, using the function “lowess” (Davies, Ihaka, and Team 2015) based on the scatter plots, likewise corresponding to natural and artificial pairings.

2.4. Discussion

The ditches of Bishop, CA provided a rare opportunity to evaluate the ecological effects of artificial vs. natural origin in aquatic systems. This comparison was possible because the channels of different origin occupy, and are interconnected within, a relatively undisturbed watershed setting, and because comparable benthic substrates are present within both natural and artificial channels. Invertebrate communities in ditches were not significantly different from those of natural creeks, even as communities varied as expected by substrate and season. Instead, communities lost sensitive EPT taxa and 64

supported reduced species richness as they moved downstream through Bishop. These patterns suggest that relative watershed urbanization, or some other environmental gradient accrued flowing through the town, drove the habitat quality of the aquatic systems, regardless of their origin. A variety of characteristics of urban streams could account for this gradient (Walsh et al. 2009).

Our findings support the ideas that artificial aquatic systems can house biodiversity (Chester and Robson 2013) and that ditches can provide important ecosystem services (Herzon and Helenius 2008). More specifically, our results are consistent with a growing body of evidence that the condition of setting, of watershed- and local-level land use, affects ecological structure and function of artificial waterbodies similarly to natural ecosystems. Some similar results in published literature suggest that our finding of lack of difference in community structure between artificial and natural ditches, and stronger response to setting, may not be as unusual as it first appears (Simon and Travis 2011, Harabis and Dolny 2015, Verdonschot, Keizer-vlek, and

Verdonschot 2011, Leslie and Lamp 2017). The broader implication of these results is that the poor condition of artificial aquatic systems in other settings cannot necessarily be attributed to their origin per se, but rather may be explained by material and identifiable aspects of their setting and structure.

In natural streams, aquatic macroinvertebrate communities respond to substrate, season (Suren and Jowett 2006), urban stressors like water pollution (Walsh et al. 2009,

Roy et al. 2003), and spatial distance to source populations of other macroinvertebrates

(Galic et al. 2013), so we expected to see variation in invertebrate communities in response to these variables. However, we did not anticipate the lack of differentiation between artificial ditch and natural creek communities. It is possible that insufficient sampling or taxonomic resolution explains this null result, and the subtle differences among different treatment groups suggests that the full story of this system may be more complex than we documented. However, that we found differences in response to known ecological drivers, substrate and season, suggests that our methods sufficed to detect substantial differences in aquatic macroinvertebrate communities, i.e., that irrigation ditches and natural creeks really did host relatively similar macroinvertebrate assemblages. We infer from the decline in taxonomic richness and sensitive taxa with increased upstream urbanization across town from west to east (Fig. 6) that this ecosystem likely responds to the impacts of watershed development much like any other, apparently the same whether in one of the ditches or in Bishop Creek. Water quality monitoring and more detailed watershed analysis could clarify this influence, though delineating watersheds around irrigation ditches with variably regulated flow can be complicated.

We recognize that the interconnectedness of ditches and creeks sharing water fresh out of the Sierra Nevada Mountains makes our system somewhat of a special case, and that all ditches most likely do not function as well as all streams. Many irrigation

ditches and similar likely suffer poor water quality from their watersheds, but since this system shares the same basic watershed and, indeed, the same primary water source, fresh from the Sierra Nevada Mountains, it largely controlled for this variable. We used this situation to our advantage in this study, but it is likely not entirely normal.

However, it is also not unique; irrigation water is collected from snowpack-fed streams as they run down mountains in other places throughout the US West, and even globally.

For all the similarity between artificial and natural streams in our study, they did apparently differ in community assembly. Communities at natural sites are more similar to other natural sites across space, and apparently connect over longer distances, than artificial site pairs (Fig. 10). Artificial site pair dissimilarity was almost constant no matter how close or far the sites. It appears that communities at natural sites are well spatially sorted, likely in response to environmental conditions, whereas the artificial site communities are not. This pattern could be some kind of sampling artifact relating to the environmental gradient across town (Fig. 9). We have not tracked the dispersal or extinction of these invertebrates isotopically or otherwise, across time or space, so our inference as to the reasons for these patterns must be limited. Perhaps the ditch sites, which range from a few years to several decades old, have not had time to environmentally sort as thoroughly as the natural sites. These macroinvertebrates’ life cycles are short, often with winged adults between generations, and the populations are closely connected to natural sources, but dispersal and disturbance response ability

varies by species (Galic et al. 2013). Perhaps the ditch sites are more homogenous in some other key way, as manmade urban ecosystems tend to be to each other, despite often high richness at any given site (Galic et al. 2013). If ditch invertebrate communities remain perpetually new, they could even comprise sink populations to the local macroinvertebrate metapopulation, which would be a very important caveat to our finding that artificial streams support similar invertebrate communities to creeks’. These ditches have begun to run dry in recent years, which could make this possibility more likely, though the life cycles of some benthic macroinvertebrates are also adapted to seasonal dryness. Investigating dispersal and extinction directly, and getting permission to sample in locations where there are gaps in our data and potentially different design

(McKinney 2006), such as on land belonging to the Los Angeles Department of Water and Power and to members of the Bishop Paiute Tribe, could help resolve these unknowns.

Experimentation in ditches like these could inform basic ecological science, as well. The notion that human influences may make a ditch stream-like, or, conversely, a stream ditch-like, freshly reiterates the question of “What is a stream?” (Moyle 2014) for science, and perhaps eventually for policy. We have much left to learn about not just irrigation ditches, but all manmade aquatic systems. Unfortunately for further experimentation in Bishop, due to drought in California, these ditches have since run dry in all but the recent wettest of winters, and the macroinvertebrate samples described

here were destroyed in a winter wildfire. Anecdotally in Bishop, but supported by broader trends within the region, drought and fire are becoming increasingly common here and throughout the US Southwest, as well as many other arid parts of the world, and will continue to do so in future (Boisrame et al. 2017). Climate change, then, and how these ecosystems can adapt, should be part of new research questions here and in similar systems (Doyle and Bernhardt 2011).

Going forward, iIrrigation ditches and other manmade water bodies could represent an important and under-appreciated opportunity for conservation. Humans have the capacity to change design and management of these common and relatively unregulated systems in ways that benefit biodiversity and optimize ecosystem services

(sensu Ross et al. 2015). Indeed, due to limited water availability and high demand, manmade irrigation waterbodies and heavily modified streams prevail in California, so we have little choice there but to accept the premise of reconciliation ecology

(Rosenzweig 2003), that not enough land and water can be preserved to save all the biodiversity and ecosystem services. Waters like the ditches of Bishop and the Los

Angeles Aqueduct into which they ultimately flow are a necessary consideration in the conservation portfolio of the west (Dollinger et al. 2015, Cowley, Wissmar, and Sallenave

2007). One may justifiably view abstracted waters as degradation of natural flows, but they, and even the accidental systems their leakages sometimes create (Sueltenfuss et al.

2013, Palta et al. 2016, Bateman et al. 2014), may be the best aquatic ecosystems left in

some places. Learning the basics of their ecology, including what structure and function they can and cannot provide under what conditions, can only help with this process.

2.5. Conclusion

Our findings suggest that under similar watershed conditions, artificial and natural streams can have similar habitat value for benthic macroinvertebrates. Further work could determine if this finding holds true in other artificial aquatic systems beyond irrigation-style ditches, and what watershed and waterbody traits enable natural function in artificial water bodies. How well macroinvertebrates can disperse to and persist in constructed ecosystems remains unclear. The answers to these questions hold important implications for ecology, conservation, and water policy.

3. North Carolina Coastal Plain ditch types support distinct hydrophytic communities

By Chelsea C. Clifford and James B. Heffernan

3.1. Introduction

Drainage ditches have gotten plenty of recognition for the wetland ecosystems, habitat, and ecosystem services destroyed through their drainage and other negative effects (e.g. Vincent, Burdick, and Dionne 2013, Zedler and Kercher 2005), but relatively little follow-up study on their role in the landscape as possible wet refugia (Remm et al.

2013, e.g. Biggs, von Fumetti, and Kelly-Quinn 2016). This oversight loses something ecologically, as drainage ditches are common and can have an important ecological role within the broader hydroscape and landscape (Sayer 2014, Herzon and Helenius 2008, e.g. Favre-Bac et al. 2014), including, in some cases, supporting species of conservation concern (Harabis and Dolny 2015, Chester and Robson 2013). Better understanding of ditches can lead to better management outcomes (Dollinger et al. 2015), so it is worthwhile to try to understand if ditches really can function at least partially like the wetlands that they often supplant (Ch. 1). We sought to answer this question in the

Coastal Plain of North Carolina, which is a flat, humid landscape heavily drained for forestry, agriculture, and development, particularly since European settlement (Phillips

1997).

We examined water, soil, and plants of agricultural, forested, and freeway roadside ditches in North Carolina’s Coastal Plain for wetland characteristics, and attempted to assess what conditions allowed for wetland-like structure. We predicted that some ditches will function like wetlands, as natural and manmade topography and soils will permit water accumulation in them. Such wetland-like ditches will be evident by their domination by wetland indicator plant species, their higher soil carbon, and their greater evidence of water presence over time, similarly to standard indication of wetland status in the region (USACE 2010). In ditches functioning like wetlands, we predicted that developed land cover and agricultural land cover around the ditches would limit the development of specialist wetland-like vegetation communities and organic matter storage due to stress, similarly to natural wetlands and other aquatic systems in developed areas (Ormerod et al. 2010, Brooks et al. 2005). In other words, we expected forested ditches to show more wetland-like conditions, because we expected them to largely separate the anthropogenic stressors of setting from the anthropogenic construction of all ditches, whereas we expected agricultural and freeway roadside sites to respond to stressors, and exhibit less wetland characteristics on average. Having anecdotally observed some very wetland-looking agricultural and roadside ditches locally, however, we also expected that at least some of the freeway and agricultural sites would also exhibit some wetland characteristics.

Once we found wetland-like ditches, after we had selected, sampled, and evaluated them, we checked to see if the National Hydrography Dataset, or NHD+

(USGS 2015), and National Wetlands Inventory, or NWI (USFWS 2017), NHD+ and NWI accounted for them as aquatic features. We expected that some substantial fraction of the ditches would appear in these datasets in some form. Finally, we looked at aerial imagery, the currently available statewide DEM, and the state’s publicly available, finer scale lidar as means to potentially update maps to greater resolution of ditches. In total, this dataset allowed us to understand that freeway, forested, and agricultural ditches can function like wetlands, gave us some ideas as to the drivers of community composition and wetland status, and suggested ways that mapping of ditches could improve.

3.2. Materials & Methods

3.2.1. Sample site selection

In summer 2015, we sampled 32 ditch reaches scattered across the coastal plain of North Carolina (Fig. 11) for plants, soil, and morphology. We aimed to sample approximately equal numbers of sites with dominantly agricultural, forested, and highway or freeway roadside watersheds, and generally sampled sites where we could gain access, and then randomly within ditch reaches available at those sites. We ensured that each sampling site was at least 10 km from any other site of the same category.

In the end, we sampled 10 agricultural ditches at North Carolina Department of

Agriculture & Consumer Resources Research Stations, 10 forested ditches at lands belonging to or managed by the North Carolina Department of Environmental Quality,

North Carolina Wildlife Resources Commission, North Carolina Division of Parks and

Recreation, U.S. Forest Service, U.S. Fish and Wildlife Service, and The Conservation

Fund, and 12 freeway sites maintained by the North Carolina Department of

Transportation. For the freeway sites, we simply randomly selected points along the largest class of highways in the North Carolina highway shapefile, clipped to the Coastal

Plain physiographic province, with at least 10km space in between, using ESRI’s ArcGIS and Hawth’s Tools’ “Generate Random Points” (Beyer 2004). The forested sites skewed easterly, towards the coast, and wet, towards bottomlands (Fig. 11), relative to the probable native forest distribution of the North Carolina Coastal Plain, probably because what remains forested today is areas less desirable as farmland, not dry uplands. The freeway sites skewed westerly, towards cities and ports, including the fall line dividing the Coastal Plain and Piedmont physiographic provinces, the end of tidal navigability, which the I-95 corridor roughly traces.

Figure 11: Map of Sites Sampled. Each dot represents one ditch reach sampled at one site. The pop-out boxes show photographic examples of one ditch reach of each type, with their locations indicated.

3.2.2. Field and Lab Methods

At each sampling point (e.g. A, Fig. 12), we randomly selected which direction to proceed in designating a 50-meter sampling reach (e.g. A-E, Fig. 12), subdivided into

10m increments (e.g. A-B, B-C, Fig. 12) at which we established subsampling transects. If a culvert, ditch ending, etc. broke that reach, we shifted the sampling reach just far enough to exclude the interruption. The subsampling transects spanning the width of

the sampling reach (vertical dotted lines, Fig. 12) extended as far as clear physical boundaries (e.g. roads, fences), property lines (e.g. highway right-of-way corridors), or until the ground leveled out and quit sloping towards the ditch, whichever came first.

At each site, we noted regional USACE wetland indicators (USACE 2010), like crayfish burrows and water-stained leaves. We also recorded the location of the center (point C in Fig. 2) of the reach with a GPS, for later analysis of how proximity and watershed setting impact sample measurements.

Figure 12: Example Subsampling Layout, Within a Freeway Sample Site. Blue rectangle represents a ditch; dark gray rectangle at bottom represents a roadside; row of black boxes above ditch represents a fence line. Letters represent midpoints of subsampling transects; subsampling transects are indicated by dashed white lines. White boxes symbolize randomized herbaceous plant quadrats.

At each sampling site, we used laser level to determine the slope of the ditch over the 50m sampling reach. Perpendicularly to the ditch, along the transects across its width, we measured the distance and changes in elevation from the center of the ditch out to the water’s edge, the bank toes (if distinguishable), bank tops, the ends of the

transects, and any points in between where the slope changed significantly, using the laser level and measuring tapes.

At each of three of the ditch transect center points per sample site, the ends and centers, (A, C, and E in Fig. 12), we collected a 10cm soil core, using a wetland style auger at wet sites and a standard terrestrial auger at dry sites. We combined these subsamples within a plastic sandwich bag for each sampling site, and chilled these samples until analysis. Back in the lab, we recorded wet, dried, and ashed weights of initially approximately 50g soil subsamples. Drying occurred in a drying oven at 100°C for 12 hours, and ashing occurred in a muffle furnace at 500 °C for 15 hours. We calculated percent water by subtracting dry weight from wet weight and dividing that remainder by wet weight for each sample, and we calculated organic matter by subtracting ashed weight from dry weight and dividing that remainder by dry weight for each sample.

We sampled vegetation at each of five width transects (dotted vertical lines A-E in Fig. 12) at each sample site. At the center of each transect (points A-E in Fig. 12), we measure canopy cover using a spherical crown densiometer. In a belt one meter to the side of each transect, we recorded the species and diameter at breast height of any tree or shrub of breast height or taller.

Along each transect, we estimated percent cover of each herbaceous vegetation species on the Daubenmire scale (Daubenmire 1959) within two quadrats, one on either

side of the center point of the ditch. The quadrats measured 1m by 0.5m quadrat, and we positioned them length parallel to the ditch. We placed them at a random distances from the ditch center based on a normal distribution with a standard deviation of five meters from center for ditches of 10m or more in width, or half that, a standard deviation of

2.5m, for smaller ditches. Thus, more quadrats were relatively near the ditch center

(white rectangles, Fig. 12). At each quadrat, we also measure maximum vegetation height.

Later, we categorized all plant taxa recorded by their Wetland Indicator Status according to the USDA Plant Database (USDA 2015), and calculated percent cover of wetland indicator statuses in each subsample. We log-transformed average total coverage in each status by site, and compared across site types. We compared herbaceous communities, in terms of site level mean taxon proportion cover across quadrats, to each other using a nonmetric multidimensional scaling (NMS) analysis with a Bray-Curtis distribution and a PERMANOVA (Goslee and Urban 2007). Then we calculated the vectors along which variables like mean canopy cover and soil percent organic matter (OM) mapped within the NMS (Oksanen et al. 2015). We inverted the vector for site mean maximum herbaceous vegetation height, to make it more obviously reflect greater mowing influence rather than greater growth.

3.2.3. Site characterization by remote sensing

We looked up each sampled ditch reach in the National Hydrography Dataset, or

NHD+ (USGS 2015), and National Wetlands Inventory, or NWI (USFWS 2017), to see if each database included the ditch at all, and if so, how, under what category, using the field-recorded GPS coordinates of our ditches verified with aerial imagery from ESRI and Google. We counted ditches as included even if a linear feature, i.e. a ditch or a stream, was clearly intended as the one we sampled but mapped slightly differently, i.e. roughly traced the shape of the ditch within a few meters of its actual location and could not potentially represent any other nearby linear feature, but did not count reaches near but outside polygons, e.g. of wetlands, as included.

Then, we mapped the center of the reach and the NHD against ESRI’s aerial imagery base layer, the North Carolina state digital elevation model (DEM) with grid cells 20’ on each side (2013), and 1km square DEMs around each site with 8’ grid cells that we made ourselves using publicly available state lidar data (2014). We made these

DEMs by converting the raw .las files we got from the state dataset to LAS Datasets, evaluating their statistics to determine that an 8’ grid cell resolution would work best, converting the LAS to a Multipoint, adding the Multipoint to a Terrain, and converting the Terrain to a Raster via Natural Neighbors. Once we had all of these map layers, we visually, qualitatively examined each one to see which types of data layers tended to

reveal our ditch reaches, to assess what publicly available data could be used in future to add to better map ditches like the ones we sampled.

3.3. Results

All the site types, agricultural, forested, and freeway, showed wetland characteristics in terms of hydrology, soils, and plants, but the site types also differed.

Water and soil showed similar patterns across the sites (Fig. 13); soil water and OM content were highly correlated (r2=0.79 with OM percentage log-transformed), and hydrologic indicator count roughly corresponded to soil water content. Forested sites had the greatest range of values and by far the extreme highest counts of U.S. Army

Corps of Engineers hydrologic indicators, soil water content, and soil organic matter

(Fig. 13). All soils contained at least some organic matter, if generally little along freeways; maximum soil OM content, at forested North River Game Land (pictured, Fig.

Figure 13: Site Average Water and Soil Values by Ditch Type. Red boxes are 95% confidence intervals about the mean.

11), was 61%. The soils varied substantially in terms of color and apparent texture; there was no one ditch soil.

In terms of herbaceous plant coverage, in these quadrats weighted towards the middles of the ditch, all of these sites met wetland criteria of at least 50% of the taxa by average coverage across quadrats at each site falling within obligate (OBL), facultative wetland (FACW), or facultative (FAC) indicator statuses, with some also in facultative upland status (FACU), but little plain upland (UPL) taxa coverage (Fig. 14). The variation in coverage by indicator status was high for wetter indicator statuses across all

Figure 14: Site Average Total Percent Cover Per Quadrat of Taxa in Each Wetland Indicator Status by Ditch Type. Red boxes are 95% confidence intervals about the mean. site types, dipping as low as zero for almost all sites in OBL and FACW categories.

Spreads tightened for drier indicator statuses. Agricultural sites had the highest percent coverage values for both OBL and, at the other extreme, FACU and UPL taxa (Fig. 14); they had the maximum and typically higher percent cover in general, and maximum

Figure 15: Site Average Characteristics of Herbaceous Plant Communities by Ditch Type. Red boxes are 95% confidence intervals about the mean. and typically higher taxon count (Fig. 15). None of these differences were statistically significant, though mean total percent herbaceous cover came close, and agricultural sites did tend to be densely vegetated and apparently less often mown than freeway roadsides. Freeway roadsides showed signs of mowing like piles of cut grass, and had a lower average maximum herbaceous vegetation height. While this difference was not statistically significant, the range of site average height of herbaceous vegetation on freeway roadsides was about half that of agricultural and forested sites (Fig. 15). Forests did have higher percent canopy than any other category (Fig. 15).

The ditch types had consistently different taxa of herbaceous plant communities

(F=3.25, d.f.= 2, p=0.001), and these variations appeared to follow both local, site-level predictors and wetland indicator status plant traits, and somewhat broader, landscape- level predictors (Fig. 16). Forested site plant communities separated most strongly from the other two site types in ordination space and clustered slightly more strongly; none of 82

the plant communities of any of the site types looked much more homogenous across sites than any other. Sites included a mixture of nonnative and native herbaceous and tree species. Forested site plant communities tended to have higher measured percent canopy, but did not correspond significantly with percent forest within a radius of either

100 or 500m in the NLCD; instead they tended strongly to occur in areas with wetland cover within 500m, and somewhat with areas with water cover within 100m. Similarly, plant communities at these forested sites tended to have high soil water content and percent organic matter, and high counts of USACE wetland hydrologic indicators.

Communities corresponding to greater water within 100m also tended to be farther east in longitude, closer to the coast.

Figure 16: Nonmetric Multidimensional Scaling Analysis of Herbaceous Plant Communities at Different Ditch Types, Overlain with Variable Vectors. Each point represents one ditch site, and the ellipses represent averages across sites. Differences among sites were statistically significant according to a PERMANOVA (F=3.25, d.f.= 2, p=0.001). The two plots are the same; the left includes local, site-level vectors, and the right includes landscape-level vectors. All vectors are significant at p<0.05; tested vectors that were not significant are not shown.

Agricultural and freeway roadside plant communities overlapped somewhat and shared a tendency to occur with more developed area within 500m, and at higher elevation (Fig. 16). Agricultural site plant communities typically occurred in areas with cultivation within 500m and within 100m, though some freeway roadside communities also appeared to respond to the 500m cultivation influence. Freeway roadside communities typically occurred within the influence of more development within 100 and 500m, though some agricultural communities apparently responded to these drivers as well. Agricultural plant communities, and some freeway roadside communities, typically featured the most FACU and UPL taxa, and the greatest herbaceous percent 84

cover, whereas freeway roadside communities tended to have FACW taxa and shorter vegetation.

Figure 17: Inclusion of Sampled Sites in Federal Databases. Colors designate site types. On the left are categories under which the NHD+ included sampled ditch reaches; on the right are categories under which the National Wetland Inventory included them. The bands connecting the two represent sites and which designations they had under each database, and the width of the band against each database represents how many sites each database included in each category. Most sites were not included at all in either database, including all freeway sites; these are labeled “absent.” The full terminology for “Riverine Intermittent Streambed” in the NWI was “Riverine Intermittent Streambed Seasonally Flooded,” and “Palustrine wetland” subsumes a few different subcategories. CanalDitch and StreamRiver are linear features; the rest of the categories are larger areas that included ditches within them.

These sampled ditch reaches generally did not appear in the National

Hydrography Dataset or National Wetlands Inventory in any form (Fig. 17). When the databases did feature them as some kind of aquatic feature, they generally did not recognize them as ditches (Fig. 17). The National Hydrography Dataset only included one sample reach, in Croatan National Forest, as a “CanalDitch,” whereas the National

Wetlands Inventory included two reaches within areas labeled “partially

ditched/drained,” but none as independent excavated features. A few more, usually the same few, were mislabeled as streams or simply wetlands (Fig. 17), but the NHD+ excluded 26 sites entirely, and the NWI excluded 22, both leaving out all of the freeway sites.

Within the NHD+ and NWI more broadly, in surrounding areas, sometimes neighboring ditches and other small waterbodies were mapped, and sometimes not; a ditch adjoining the one mapped “CanalDitch in the NHD+ was not mapped. Some of the

NHD+ maps around our sample sites showed plain issues in categorization, with canals switching to streams and back, for example, and multiple streams tributaries to the canals. Visually, unmapped ditches are sometimes apparent from readily available aerial imagery, often apparent from the statewide 20’-sided grid cell DEM, and usually apparent from the 8’ grid cell DEMs we created with publicly available lidar (Fig. 18).

Figure 18: Map of Croatan National Forest Sample Site and Lidar. This ditch site was the only one correctly included in the NHD+ as CanalDitch, as shown by the purple lines, but as the gray elevation of the lidar suggests, ditches just south and west of this point were not mapped. StreamRiver is shown as blue lines, and acknowledge wetland by the blue-gray stippling in the lower left corner of the image.

3.4. Discussion

Remarkably, all of these ditch sites exhibited at least some wetland structure, especially in terms of their hydrophytic plant communities. All of them are storing at least some organic matter in their soil, and all of them are somewhat wet. Very few of them, however, are accounted for in the NHD+ or the NWI, which means that we are

failing to account for these ditches as a mapped aquatic resource through the traditional means. Ditches may be narrow, but their net length is likely very long, particularly here in the North Carolina Coastal Plain, so this under-accounting could add up to a substantial area, especially in terms of connectivity across the landscape. For any given variable that we measured, all three ditch types had fairly wide ranges of values, suggesting a diversity of habitat types and possibilities among Coastal Plain ditches. The vector of average maximum herbaceous vegetation height (Fig. 16), suggests that roadside herbaceous plant communities do respond to management interventions.

Indeed, when we looked at which sites of which site types had herbaceous communities mixed with or toward the margins of clusters of communities of other site types in the

NMS, permitted growth, through mowing and herbicide, and shading through trees, appeared to be strong drivers towards the edges of the clusters. This suggests that people do have a choice in steering the ecological structure of these miniature wetlands, as backed by earlier findings elsewhere (Dollinger et al. 2015), so policy-makers, farmers, civil engineers, foresters, and landscape architects could factor this thinking into their choices about these systems here.

Freeway roadside sites, in particular, are managed by the North Carolina

Department of Transportation, with environmental scientists and regional divisions; this system could be a relatively easy place to implement experimental ditch management factoring in wetland criteria. Designs would have to allow water to flow without

creating a flooding risk or adding a maintenance burden, but this may well be possible.

Systems like this built for one or few ecosystem services, just flood control and water transport, are often overbuilt for those purposes (Palta, Grimm, and Groffman 2017), and designed experiments incorporating natural elements into infrastructure have reduced maintenance burden before (Felson and Pickett 2005). We did not sample private ditches along smaller roads, though these can be just as interesting (McPhillips et al. 2016).

Agricultural sites had the most vigorous herbaceous plant communities, including the most dense herbaceous cover (Fig. 15) and the most upland taxa (Fig. 15).

The relative lack of trees, typical interventions of mowing and dredging (Dollinger et al.

2015), and often high nutrient inputs could make them great places to experiment (De

Meester et al. 2005, Koetsier and McCauley 2015) with manipulating competitive interactions between plants and disturbance responses. In spite of the stress of surrounding development and agriculture (Fig. 16), these sites often had high taxa richness (Fig. 15), albeit generally featuring more upland taxa as well as more obligate taxa (Fig. 14). Their palette of plants there was broad, without deliberate intervention to make it so. For farmers interested in influencing the ecosystem services received from their ditches, there are opportunities here too. As the actual sites we sampled were agricultural research stations, these could be good places to begin experimenting.

Forested sites proved more complicated than the less impacted control we had imagined they could be; they were more swampy than simply forested. In hindsight, that the forests were wetlands was to be expected for several reasons: they were drained, their easterly, coast-ward distribution from which we were forced to choose, and that the forest that remains in conservation is likely the land least suited to farming decades and centuries ago. So, their ditches were much more wetland-like in terms of soil carbon and water, but this could be an artifact of the surrounding landscape that has little to do with the behavior of the ditches themselves.

We left a great deal yet to explore about ditch soils, but the variation we observed among them suggests that there is no one ditch soil. We did not sample soil carbon in transects across the ditches to see if carbon accumulation is greatest at their thalwegs where we sampled, as is likely. Even so, none of the ditches had as much soil carbon as the 95% in pocosins and 75% in gum swamps measured at Croatan (Bridgham and Richardson 1993), as a comparison to a natural swamp in the area, but a few came somewhat close, and it is notable that ditches, forested and even agricutltural, have high carbon in their soils at all. The high variability in the ditch soils suggests that their composition likely correspond more to their local parent soil than to anything about ditches per se, except perhaps construction imports of soil in some cases. Soil horizon formation in ditches can begin fairly quickly (Needelman, Ruppert, and Vaughan 2007),

however, so deeper coring and sectioning could also reveal further information about changes since ditching occurred.

We left even more to explore about ditch hydrology. Without repeat site visits and monitoring, we lack precise measurements of site seasonal hydrology. However, the correspondence between USACE hydrologic indicators, soil water content, soil organic matter (Fig. 13), and to some degree known wetland and water landscape surroundings

(Fig. 16) suggests that our rough water metrics did capture something about general wetness regime of the ditches studied. Both shorter term monitoring through the establishment of wells over the course of a year to examine seasonal inundation and flows, and longer term proxies of hydrology from soil (USACE 2010) and trees (Hupp,

Dufour, and Bornette 2016), could better inform as to the role of hydrology in shaping these ditches.

The national databases most commonly used to find aquatic features do not appear to account for the wetland-like ditches that we found well, but the technology and even data is available to improve maps. Our sampling of ditch inclusion in the NHD and NWI is far from random, but we did not consult these databases before site selection, and even a casual glance at either database superimposed on aerial imagery for this region makes clear that many ditches, and other small water features, are left out, or mis-categorized. In fairness, the oddities in canals with stream tributaries or segments in the NHD+ could make sense if a channelized stream counts as a canal, but

given that some of the stream sections were as straight as some canal sections, mismapping or confusion over definition of “CanalDitch” versus “StreamRiver,” or artificial versus natural flowlines, appears highly likely. This lack of information and misinformation on extent and distribution of ditches makes it difficult to know quite what is and is not getting management attention at the moment, or to design how best to target interventions. However, our lidar-based dabbling, albeit without applying a model to extract ditches, suggests that finding ditches to add to maps should be doable, given enough computing power to deal with the large amount of data the lidar entails.

Better remote sensing scientists than us have made significant progress on this task

(Rapinel et al. 2015, Bailly et al. 2008, Cazorzi et al. 2013), so perhaps better ditch maps for the North Carolina Coastal Plain will emerge soon regardless of our encouragement.

3.5. Conclusion

The forested, freeway, and agricultural ditches that we sampled throughout the

North Carolina Coastal Plain all showed some ecological structural characteristics of wetlands. They all had herbaceous hydrophytic plant communities towards their thalwegs, all stored at least a little organic matter in their soil, and all appeared at least somewhat wet. There was a large range of variation within these characteristics, though, and distinct plant communities across site types that appeared to respond to management intervention, indicating that humans may have the ability to shape these communities if we so choose. The actual extent of these ditch wetland ecosystems is

unclear, as they appear to be poorly mapped in the NHD+ and NWI. A great deal of research remains to be done on these systems, including experimenting with management techniques, and greater initial characterization of hydrology, soils, biota beyond plants, and other ecosystem services. Our findings raise the possibility that drainage features and other artificial aquatic systems questions well beyond the North

Carolina Coastal Plain, as drainage ditches and other artificial aquatic systems are widespread.

4. Similarities and Differences in Algal Bloom Processes in Artificial and Natural Lakes

By Chelsea C. Clifford and James B. Heffernan

4.1. Introduction

Humans are by far the greatest source for new lakes and ponds today (Oertli

2018), and most freshwater ponds in the U.S. are artificial (Dahl 2011), yet all too often, scientists, land managers, and policy makers ignore the socio-ecological potentials of these systems (Oertli 2018). Multiple studies have demonstrated that artificial ponds and lakes can provide habitat (Chester and Robson 2013), recreational opportunities (Biggs, von Fumetti, and Kelly-Quinn 2016), and carbon and sediment storage (Renwick et al.

2006, Tranvik et al. 2009, Williamson et al. 2009), as well as the obvious resource of water. The ways in which people tend to construct ponds and lakes, often for a limited purpose and with a correspondingly limited design, can limit their function, however

(Oertli 2018, Palta, Grimm, and Groffman 2017); they may lack the structure to fulfill their potential. Alternatively, exogenous factors relating to where artificial lakes tend to be situated may limit them; artificial lakes can be susceptible to influences from their watersheds just like any other waterbody (Knutson et al. 2004, Mehaffey et al. 2005). The speed with which small artificial waterbodies turn over in existence (Oertli 2018) suggests that age could also be a limiting factor; most of these manmade ponds and lakes are probably relatively young in geologic or successional time (Milner et al. 2011).

In any case, we should better learn about these processes that may control the structure and function of artificial aquatic ecosystems; assuming that these waterbodies are doomed to poor condition will not suffice for science or management (Ch. 1). While many of the ecological processes occurring in artificial aquatic ecosystems will be the very same ones occurring in natural waterbodies, humans typically have greater control, even mandatory control, over artificial waterbodies, which adds to the forcings involved in these systems. We need to know what our interventions do, could do, and cannot achieve, in order to manage effectively. Relatively few studies have explored such questions on more than a case-by-case or small-scale basis.

In the case of lakes, a massive data collection effort that has made broad scale study possible is the National Lakes Assessment (NLA), by the U.S. EPA. Previous studies have begun to use this study to explore artificial lakes in the U.S., notably

Doubek and Carey’s assessment of the first year of NLA data in 2007, in which they found that artificiality both correlated and interacted with latitude across multiple variables, including water quality variables and drivers like land use. They noted that southern reservoirs were deeper and more forested, and southern natural lakes and northern reservoirs had greater total phosphorus and a shallower Secchi depth.

Generally, they noted that natural lakes were more eutrophic, with greater total nitrogen and chlorophyll a, and that reservoirs occurred at higher elevations. While these patterns suggested connections between land cover, geomorphology, and water quality,

statistical methods employed could not parse causes (2017). Earlier, Beaulieu, Pick, and

Gregory-Eaves found that temperature and total nitrogen predicted algal biomass in

2007, particularly in deep natural lakes, using multiple regression (2013). Meanwhile, the

EPA itself has reported higher total phosphorus, and increasing total phosphorus and microcystin algal toxin, in reservoirs (USEPA 2016).

We proposed to use structural equations modeling, which allows assessment of causality across a network of variable relationships (Rosseel et al. 2017), to more thoroughly explore the process through which water quality goes bad via algal blooms potentially leading to microcystin outbreaks in natural as compared to artificial lakes throughout the U.S. using this NLA dataset. Our model suggests that there are differences in how blooms form between natural and artificial lakes that may have to do in part with dam management’s impact on thermal stratification. The overall processes by which blooms form are similar across lake origins, however, as are the actual data distributions. While we were unable to root our model as deeply into the physical features of the lakes and the landscape features of their watersheds as we would have liked, our approach nonetheless highlights the utility of exploring the processes behind ecological outcomes of artificial aquatic systems.

4.2. Materials & Methods

In 2007 and 2012, the U.S. Environmental Protection Agency (EPA) and its collaborators collected an abundance of physical, chemical, biological, and recreation-

related data at over 1,000 lakes both times around the lower 48 United States, as part of its National Lakes Assessment (USEPA 2016). During each sampling event, roughly half of the lakes were natural, of geologic origin, and roughly half were “man-made,” including reservoirs and other impoundments, farm ponds, former quarries, and so on.

The EPA selected lakes through a stratified random sampling design intended to include a variety of sizes and locations of lakes within the National Hydrography Dataset

(NHD+), and collected data on hundreds of indicator variables through field visits and through geospatial combination with the NHD+ and National Land Cover Database

(NLCD), sometimes visiting sites more than once (USEPA 2016). We opted to focus on the first visit to each site for each year for consistency, and on comparatively few variables, those which we judged most essentially physically and chemically related to the formation of algal blooms and associated water quality problems in lakes. Of the variables we selected, some, including dissolved oxygen and temperature, were collected via sensors over profiles every meter or so within the lake, but most were collected as grab samples and processed in a laboratory.

Our final variable list included mean temperature over lake sample profile, range between maximum and minimum temperature in that profile, turbidity, maximum depth, and concentrations of total phosphorus (TP), total nitrogen (TN), dissolved organic carbon (DOC), chlorophyll a (chl a), mean dissolved oxygen (DO) over sampled profile, and microcystin. After removing outliers with temperatures of over 100°C and

samples with missing data, there remained 1,045 man-made and 870 natural sample sites. We log-transformed all variables to achieve normalcy and homoscedasticity for use in a structural equations model (SEM), “sem” (Rosseel et al. 2017), in which we compared the relationships between them in natural and artificial ponds. We originally attempted a larger model including more physical data about the lakes and their watersheds, such as areas, and logit-transformed watershed percent cover data from the

NLCD, in an attempt to better explain the role of watershed setting and physical properties, including design, on our variables of interest, but could not achieve a model that identified properly. We examined D-separation and Fisher’s C (Lefcheck 2016) and then the model function test statistic (Rosseel et al. 2017) to make sure the model was a good fit, and added relationships that we had determined had natural history reasons to belong in the model to our basis set until this fit was achieved. The versions of the SEM tested included one that grouped the model by lake origin, generating separate SEMs for natural and manmade lakes, and one that lumped them all together; we compared the two using an ANOVA (Davies, Ihaka, and Team 2015). We also examined the distributions of the actual data associated with each variable, and computed 95% confidence intervals about the means for each year/origin combination, e.g. 2007 natural,

2007 man-made, etc. Finally, we compared the SEM and variable values to assess forcings by the different variables upon each other.

Figure 1918:: Structural equations model of the algal bloom process in natural and man-made lakes. Each box represents a variable. Arrows scale to size of effect.

4.3. Results

The SEM that separated natural and manmade lakes fit best (Χ2=4.905, df=8, p=0.768); the equivalent model that did not separate the two did not even fit (Χ2=1007.6, df=43, p<0.001), and was significantly different (Χ2=1044.6, df=48, p<0.001). That said, most parts of the models look very similar (Fig. 19). Both show a strong positive effect on chlorophyll concentration by temperature and nitrogen, with a lesser effect by phosphorus. This greater effect of nitrogen rather than phosphorus runs counter to the 99

dominant hypothesis in limnology that phosphorus is the primary limiting nutrient in lakes (Schindler et al. 2016), but it held true across both natural and man-made lake types.Both types also show strong positive effects of depth on temperature range, and total nitrogen on dissolved organic carbon. Turbidity increased somewhat with total nitrogen and phosphorus and with chlorophyll, and was dampened by DOC. DO decreased at least somewhat as temperature and its range increased, and as chlorophyll increased; meanwhile DO reduced DOC. Total nitrogen increased microcystin concentration. Nitrogen and phosphorus concentrations were correlated.

Figure 20: Difference Between Man-Made and Natural Structural Equations Model. This figure was generated by subtracting Fig. 18 a, the coefficients of the natural structural equations model, from Fig. 18 b, the coefficients of the man-made SEM. Arrow line widths correspond to the size of the difference between the two coefficients.

100

There most obvious differences between the natural and man-made models are in terms of size of effect; man-made waterbodies were generally less predictable, with smaller coefficients. Temperature and TN increased chlorophyll somewhat more in artificial waterbodies, while TN had a much smaller effect on DOC there, where there is more DOC (Fig. 20), and depth affected the range of temperature less (Fig. 19). TN had a substantially smaller impact on turbidity in man-made lakes (Fig. 20). Some effects appear or disappear across models (Fig. 20). Temperature range dampened microcystin concentrations in natural waterbodies, while turbidity substantially increased it (Fig. 20); neither of these effects showed up strongly in manmade waterbodies (Fig. 19), which had somewhat less temperature range and more turbidity generally (Fig. 21). DO was less in colder and deeper artificial waterbodies (Fig. 21); man-made waterbodies were generally warmer and a little deeper, and lower in DO (Fig. 21).

In general, the distributions of the data across years and lake origin types were more similar than they were different (Fig. 21). Maximum depth and temperature range decreased somewhat on average between 2007 and 2012, while TP and chlorophyll increased. The spread of the data dwarfs the differences in the averages and medians.

101

Figure 2120:: Box Plots Representing Variable Data Distributions. Notches are 95% confidence intervals about the medians; red rectangles are 95% confidence intervals about the means.

4.4. Discussion

Despite the statistically significant difference in the SEM as grouped for artificial and natural lakes, what the SEM (Fig. 19) shows overall, by starting with the same basic model for artificial and natural lakes and arriving at a solution for each, is that the two go about forming algal blooms and associated water quality issues in much the same way, from the same basic drivers. The mostly overlapping data distributions (Fig. 21) reiterate this similarity. Artificial lakes are still lakes.

That said, the differences between the two lake origins may partially reflect a differing driver, in the form of water releases as dam management interfering with thermal stratification on larger artificial waterbodies. In these man-made waterbodies,

102

depth has a lesser effect on thermal range, but mean temperature has a greater effect on dissolved oxygen, which makes sense if larger artificial waterbodies are most likely managed reservoirs, with controlled releases of water at different levels that impact thermal stratification and lower residence times. Humans impacting temperature could also explain why mean temperature is a stronger driver of algal blooms in artificial aquatic systems. The greater TP to turbidity interaction could reflect clay particles and greater sediments flowing into reservoirs on average, and the tendency of TN to increase the chlorophyll pool more than the DOC pool and turbidity in artificial lakes could also reflect shorter residence times. Even out of reservoirs, in dredged farm ponds, for example, artificial lakes are likely to be disturbed by humans more frequently than natural ones (Oertli 2018), which would trigger a form of turn-over.

Our findings confirm that patterns previous authors (Doubek and Carey 2017,

Beaulieu, Pick, and Gregory-Eaves 2013) had observed in the 2007 NLA data continued into 2012, and add some new information as to why those patterns occur. We suggest that continuing to think in this process-based way could lead to a yet better understanding of those and other trends in this rich dataset.

4.5. Conclusion

The process by which the thousands of lakes sampled throughout the U.S. in

2007 and 2012 form algal blooms and related water quality issues does indeed differ between artificial and natural lakes. These differences may partially reflect water

103

releases in large riverine reservoirs, which impact temperature and thermal stratification. While significant, the differences are less striking than the similarities; which environmental variables drive which environmental variables by approximately how much in what direction are mostly the same. These results suggest a combination, then, of conclusions about artificial lakes, both that they mostly function like natural lakes, and that human intervention may strongly impact them. This finding means that managers and policy-makers should probably assume an artificial lake functions roughly similarly to its natural cousins, until or unless given reason to suspect otherwise, but really, without actual data on the artificial lake itself, we risk missing something important.

104

5. Conclusion

In Chapter 1, a variety of waterbodies may count as artificial in different contexts. Artificiality is an insufficient explanation for their ecological condition; instead we should test process-based alternatives such as setting, design, and age. Better knowledge of these drivers could improve management potential over a potentially large expanse of ecosystems, known to sometimes provide ecosystem services and disservices of concern. Policy based on our current perception of these systems may reinforce negative expectations.

In Chapter 2, irrigation-style ditches in network with a natural creek fresh from the Sierra Nevada Mountains in small Bishop, California, have statistically undifferentiable benthic macroinvertebrate communities from natural creek communities, given similar substrate and same season. Communities do decline in sensitive taxa (mayflies, stoneflies, and caddisflies) and biodiversity downstream across town, presumably as the influence of urban and agricultural land use increases.

Communities in creeks that are close together are more similar than communities in ditches that are close together, suggesting some difference in community assembly.

In Chapter 3, across 32 agricultural, forested, and freeway roadside ditch reaches in the North Carolina Coastal Plain, all supported predominantly wetland plants, and had at least some soil carbon and signs of wetness. Plant communities differed across site types, however, in response to landscape variables like development in the vicinity

105

and local variables like apparent mowing (on freeway roadsides). Forested sites were more easterly and swampy. Agricultural sites had greater plant cover, including of taxa of driest and wettest wetland indicator groups.

In chapter 4, approximately 1,000 lakes sampled for a wide range of variables in

2007 and 1,000 more in 2012, half natural and half manmade, formed algal blooms and some associated water quality issues like low oxygen and cloudiness through similar processes, largely driven by heat and nitrogen. There were some differences in degrees of response, particularly in response to temperature and nutrients, that were partially depth-dependent.

The first chapter provides more research directions than firm answers, and the others began to follow the path it set out, by trying to disentangle the processes through which classic indicator variables for the natural analogs to each system came to vary in artificial systems. As expected, the answer is mixed. All three of the empirical chapters suggest an influence of watershed and the water flowing from it on the systems, through the urbanization gradient in Chapter 2, the land cover associating with plant communities and other variables in Chapter 3, and the responses to nutrient inputs in

Ch. 4. There are hints that design and management could matter, via the apparent mowing effect in Ch.3 and the possible dam management effect in Ch. 4. I did not get to research the impact of time and age in this dissertation. In any case, the artificial aquatic systems show both similarities to natural systems, in all chapters, but especially Ch. 2,

106

and differences, especially in Ch. 4. They show a great variety of ecological structure, even within the seemingly simple category of ditches containing Chapters 2 and 3. In all, there is a potential for artificial aquatic ecosystems, to resemble natural systems at times and to be novel, and limitations, placed by their surrounding and upstream environments if nothing else. This mixture suggests, as posited in Ch. 1, that the artificial/natural dichotomy is more appropriately thought of as a gradient, even as my own studies conformed to the dichotomy for this dissertation.Artificial aquatic systems could be more than what they are (Oertli 2018), if we humans wanted to get more creative with design (sensu Ross 2015); we have not maximized their ecological potential yet. Decisions are open to us, indeed, mandated, in some cases, and as always, more data could improve our chances of making good ones.

107

References

1995. The Freshwater Imperative: A Research Agenda. Washington, D.C.: Island Press.

2013. "History - Statutes." Haute école de Genève, accessed 19 Aug. http://campus.hesge.ch/epcn/history.asp.

2013. 20-Foot DEM Elevation Service. edited by NCDOT: NC OneMap.

2014. "Introduction ot the new EU Water Framework Directive." European Union, Last Modified 18 June, accessed 19 Aug. http://ec.europa.eu/environment/water/water-framework/info/intro_en.htm.

2014. "Million Ponds Project." Freshwater Habitats Trust, accessed 19 Aug. http://www.freshwaterhabitats.org.uk/projects/million-ponds/.

2014. Public Road and Street Mileage in the United States by Type of Surface. U.S. Department of Transportation, Research and Innovative Technology Administration, Bureau of Transportation Statistics.

Base Data. NC Floodplain Mapping Program, Raleigh, NC.

Abellan, P., D. Sanchez-Fernandez, A. Millan, F. Botella, J. A. Sanchez-Zapata, and A. Gimenez. 2006. "Irrigation pools as macroinvertebrate habitat in a semi-arid agricultural landscape (SE Spain)." Journal of Arid Environments 67 (2):255-269. doi: 10.1016/j.jaridenv.2006.02.009.

Anderson, B. S., B. M. Phillips, J. P. Voorhees, and M. Cahn. 2017. "Vegetated Treatment Systems for Removing Contaminants Associated with Surface Water Toxicity in Agriculture and Urban Runoff." Jove-Journal of Visualized Experiments (123):8. doi: 10.3791/55391.

Arthaud, F., D. Vallod, J. Robin, A. Wezel, and G. Bornette. 2013. "Short-term succession of aquatic plant species richness along ecosystem productivity and dispersal gradients in shallow lakes." Journal of Vegetation Science 24 (1):148-156. doi: 10.1111/j.1654-1103.2012.01436.x.

Bailey, C., A. Spielmeyer, R. M. Frings, G. Hamscher, and H. Schuttrumpf. 2015. "From agricultural fields to surface water systems: the overland transport of veterinary

108

antibiotics." Journal of Soils and Sediments 15 (7):1630-1634. doi: 10.1007/s11368- 015-1140-4.

Bailly, J. S., P. Lagacherie, C. Millier, C. Puech, and P. Kosuth. 2008. "Agrarian landscapes linear features detection from LiDAR: application to artificial drainage networks." International Journal of Remote Sensing 29 (12):3489-3508. doi: 10.1080/01431160701469057.

Baker, S., and K. Eckerberg. 2016. "Ecological restoration success: a policy analysis understanding." Restoration Ecology 24 (3):284-290. doi: 10.1111/rec.12339.

Ballantine, K., and R. Schneider. 2009. "Fifty-five years of soil development in restored freshwater depressional wetlands." Ecological Applications 19 (6):1467-1480. doi: 10.1890/07-0588.1.

Bateman, H. L., J. C. Stromberg, M. J. Banville, E. Makings, B. D. Scott, A. Suchy, and D. Wolkis. 2014. "Novel water sources restore plant and animal communities along an urban river." Ecohydrology:n/a-n/a. doi: 10.1002/eco.1560.

Beaulieu, Marieke, Frances Pick, and Irene Gregory-Eaves. 2013. "Nutrients and water temperature are significant predictors of cyanobacterial biomass in a 1147 lakes data set." Limnology and Oceanography 58 (5):1736-1746. doi: doi:10.4319/lo.2013.58.5.1736.

Bennett, Elena M., Garry D. Peterson, and Line J. Gordon. 2009. "Understanding relationships among multiple ecosystem services." Ecology Letters 12 (12):1394- 1404. doi: 10.1111/j.1461-0248.2009.01387.x.

Bensaude-Vincent, Bernadette, and William R. Newman. 2007. "Introduction: The artificial and the natural: State of the problem." In The artificial and the natural, edited by Bernadette Bensaude-Vincent and William R. Newman, 1-19. Cambridge, MA and London, England: The MIT Press.

Berg, M. D., S. C. Popescu, B. P. Wilcox, J. P. Angerer, E. C. Rhodes, J. McAlister, and W. E. Fox. 2016. "SMALL FARM PONDS: OVERLOOKED FEATURES WITH

109

IMPORTANT IMPACTS ON WATERSHED SEDIMENT TRANSPORT." Journal of the American Water Resources Association 52 (1):67-76. doi: 10.1111/1752- 1688.12369.

Bernhardt, E. S., and M. A. Palmer. 2007. "Restoring streams in an urbanizing world." Freshwater Biology 52 (4):738-751. doi: 10.1111/j.1365-2427.2006.01718.x.

Bernhardt, E. S., and M. A. Palmer. 2011. "River restoration: the fuzzy logic of repairing reaches to reverse catchment scale degradation." Ecological Applications 21 (6):1926-1931.

Hawth's Analysis Tools for ArcGIS.

Biggs, J., S. von Fumetti, and M. Kelly-Quinn. 2016. "The importance of small waterbodies for biodiversity and ecosystem services: implications for policy makers." Hydrobiologia 793 (1):3-39. doi: 10.1007/s10750-016-3007-0.

Biggs, J., P. Williams, M. Whitfield, P. Nicolet, and A. Weatherby. 2005. "15 years of pond assessment in Britain: results and lessons learned from the work of Pond Conservation." Aquatic Conservation-Marine and Freshwater Ecosystems 15 (6):693-714. doi: 10.1002/aqc.745.

Bohlen, Patrick J., Elizabeth Boughton, John E. Fauth, David Jenkins, Greg Kiker, Pedro F. Quintana-Ascencio, Sanjay Shukla, and Hilary M. Swain. 2014. Assessing trade-offs among ecosystem services in a payment-for-water services program on Florida ranchlands. In NCER Assistance Agreement Final Report Executive Summary. Lake Placid, FL: Archbold Biological Station.

Boler, R.B., and D.A. Sikkema. 2010. Tamiami Trail Modifications: Next Steps Project Summary of Findings and Draft Environmental Impact Statement. In SFNRC Tenchincal Series. Homestead, Florida: National Park Service.

Branche, Emmanuel. 2017. "The multipurpose water uses of hydropower reservoir: The SHARE concept." Comptes Rendus Physique 18 (7):469-478. doi: https://doi.org/10.1016/j.crhy.2017.06.001.

110

Brand, Adrianne B., and Joel W. Snodgrass. 2010. "Value of Artificial Habitats for Amphibian Reproduction in Altered Landscapes." Conservation Biology 24 (1):295-301. doi: 10.1111/j.1523-1739.2009.01301.x.

Bremigan, M. T., P. A. Soranno, M. J. Gonzalez, D. B. Bunnell, K. K. Arend, W. H. Renwick, R. A. Stein, and M. J. Vanni. 2008. "Hydrogeomorphic features mediate the effects of land use/cover on reservoir productivity and food webs." Limnology and Oceanography 53 (4):1420-1433. doi: 10.4319/lo.2008.53.4.1420.

Bridgham, S. D., and C. J. Richardson. 1993. "Hydrology and nutrient gradients in North-Carolina peatlands." Wetlands 13 (3):207-218.

Brooks, R. P., D. H. Wardrop, C. A. Cole, and D. A. Campbell. 2005. "Are we purveyors of wetland homogeneity? A model of degradation and restoration to improve wetland mitigation performance." Ecological Engineering 24 (4):331-340. doi: 10.1016/j.ecoleng.2004.07.009.

Brown, Lee E., and Alexander M. Milner. 2012. "Rapid loss of glacial ice reveals stream community assembly processes." Global Change Biology 18 (7):2195-2204. doi: 10.1111/j.1365-2486.2012.02675.x.

Burcher, C. L., H. M. Valett, and E. F. Benfield. 2007. "The land-cover cascade: Relationships coupling land and water." Ecology 88 (1):228-242.

Callicott, J. Baird. 1995. "A critique of an an alternative to the wilderness idea." In Environmental Ethics, edited by Andrew Light and Holmes Rolston III, 437-443. Malden, MA: Blackwell Publishing. Original edition, Wild Earth 4:4.

Canals, R. M., V. Ferrer, A. Iriarte, S. Carcamo, L. San Emeterio, and E. Villanueva. 2011. "Emerging conflicts for the environmental use of water in high-valuable rangelands. Can livestock water ponds be managed as artificial wetlands for amphibians?" Ecological Engineering 37 (10):1443-1452. doi: 10.1016/j.ecoleng.2011.01.017.

Canessa, S., F. Oneto, D. Ottonello, A. Arillo, and S. Salvidio. 2013. "Land abandonment may reduce disturbance and affect the breeding sites of an Endangered amphibian in northern Italy." Oryx 47 (2):280-287. doi: 10.1017/s0030605311001542.

111

Casas, J. J., J. S. Sanchez-Oliver, A. Sanz, M. Furne, C. Trenzado, M. Juan, M. Paracuellos, M. D. Suarez, F. Fuentes, I. Gallego, C. Gil, and J. J. Ramos-Miras. 2011. "The paradox of the conservation of an endangered fish species in a Mediterranean region under agricultural intensification." Biological Conservation 144 (1):253- 262. doi: 10.1016/j.biocon.2010.08.023.

Cazorzi, F., G. Dalla Fontana, A. De Luca, G. Sofia, and P. Tarolli. 2013. "Drainage network detection and assessment of network storage capacity in agrarian landscape." Hydrological Processes 27 (4):541-553. doi: 10.1002/hyp.9224.

Cereghino, R., D. Boix, H. M. Cauchie, K. Martens, and B. Oertli. 2014. "The ecological role of ponds in a changing world." Hydrobiologia 723 (1):1-6. doi: 10.1007/s10750-013-1719-y.

Cereghino, R., A. Ruggiero, P. Marty, and S. Angelibert. 2008. "Biodiversity and distribution patterns of freshwater invertebrates in farm ponds of a southwestern French agricultural landscape." Hydrobiologia 597:43-51. doi: 10.1007/s10750-007-9219-6.

Chakrabarty, D. 2014. "The climate of history: four theses." Teksty Drugie (5):168-199.

Checker, Melissa. 2005. Polluted Promises: Environmental Racism and the Search for Justice in a Southern Town. New York and London: NYU Press.

Chester, E. T., and B. J. Robson. 2013. "Anthropogenic refuges for freshwater biodiversity: Their ecological characteristics and management." Biological Conservation 166 (0):64-75. doi: http://dx.doi.org/10.1016/j.biocon.2013.06.016.

Clarke, S. J. 2015. "Conserving freshwater biodiversity: The value, status and management of high quality ditch systems." Journal for Nature Conservation 24:93-100. doi: 10.1016/j.jnc.2014.10.003.

112

Cooper, Charles M. , Matthew T. Moore, Erin R. Bennett, Jr. Sammie Smith, and Jerry L. Farris. 2002. "Alternative environmental benefits of agricultural drainage ditches." Verhandlungen des Internationalen Verein Limnologie 28:1678-1682.

Cordell, H. Ken, and John C. Bergstrom. 1993. "Comparison of recreation use values among alternative reservoir water level management scenarios." Water Resources Research 29 (2):247-258. doi: doi:10.1029/92WR02023.

Crail, T. D., J. F. Gottgens, and A. E. Krause. 2011. "Fish community response to evolving channel complexity in an agricultural headwater system." Journal of Soil and Water Conservation 66 (5):295-302. doi: 10.2489/jswc.66.5.295.

Crifasi, R. R. 2005. "Reflections in a stock pond: Are anthropogenically derived freshwater ecosystems natural, artificial, or something else?" Environmental Management 36 (5):625-639. doi: 10.1007/s00267-004-0147-1.

D'Ambrosio, J. L., A. D. Ward, and J. D. Witter. 2015. "Evaluating Geomorphic Change in Constructed Two-Stage Ditches." Journal of the American Water Resources Association 51 (4):910-922. doi: 10.1111/1752-1688.12334.

Dahl, T. E. 2011. Status and Trends of Wetlands in the Conterminous United States 2004 to 2009. Washington, D.C.: U.S. Department of the Interior; Fish and Wildlife Service.

Daubenmire, R. F. 1959. "Canopy coverage method of vegetation analysis." Northwest Sci 33:39-64.

Davies, B., J. Biggs, P. Williams, M. Whitfield, P. Nicolet, D. Sear, S. Bray, and S. Maund. 2008. "Comparative biodiversity of aquatic habitats in the European agricultural landscape." Agriculture Ecosystems & Environment 125 (1-4):1-8. doi: 10.1016/j.agee.2007.10.006. stats. The R Project for Statistical Computing.

Davis, Robert T., Jennifer L. Tank, Ursula H. Mahl, Sarah G. Winikoff, and Sarah S. Roley. 2015. "The Influence of Two-Stage Ditches with Constructed Floodplains on Water Column Nutrients and Sediments in Agricultural Streams." Journal of the American Water Resources Association 51 (4):941-955. doi: 10.1111/1752- 1688.12341.

113

De Meester, L., S. Declerck, R. Stoks, G. Louette, F. Van de Meutter, T. De Bie, E. Michels, and L. Brendonck. 2005. "Ponds and pools as model systems in conservation biology, ecology and evolutionary biology." Aquatic Conservation- Marine and Freshwater Ecosystems 15 (6):715-725. doi: 10.1002/aqc.748.

Declerck, S., T. De Bie, D. Ercken, H. Hampel, S. Schrijvers, J. Van Wichelen, V. Gillard, R. Mandiki, B. Losson, D. Bauwens, S. Keijers, W. Vyverman, B. Goddeeris, L. De Meester, L. Brendonck, and K. Martens. 2006. "Ecological characteristic's of small farmland ponds: Associations with land use practices at multiple spatial scales." Biological Conservation 131 (4):523-532. doi: 10.1016/j.biocon.2006.02.024.

Diamond, Jared M. 1983. "Ecology: Laboratory, field and natural experiments." Nature 304 (5927):586-587.

Dollinger, J., C. Dages, J. S. Bailly, P. Lagacherie, and M. Voltz. 2015. "Managing ditches for agroecological engineering of landscape. A review." Agronomy for Sustainable Development 35 (3):999-1020. doi: 10.1007/s13593-015-0301-6.

Doubek, Jonathan P., and Cayelan C. Carey. 2017. "Catchment, morphometric, and water quality characteristics differ between reservoirs and naturally formed lakes on a latitudinal gradient in the conterminous United States." Inland Waters 7 (2):171-180. doi: 10.1080/20442041.2017.1293317.

Downing, J. A., J. J. Cole, C. M. Duarte, J. J. Middelburg, J. M. Melack, Y. T. Prairie, P. Kortelainen, R. G. Striegl, W. H. McDowell, and L. J. Tranvik. 2012. "Global abundance and size distribution of streams and rivers." Inland Waters 2 (4):229- 236. doi: 10.5268/iw-2.4.502.

Downing, J. A., Y. T. Prairie, J. J. Cole, C. M. Duarte, L. J. Tranvik, R. G. Striegl, W. H. McDowell, P. Kortelainen, N. F. Caraco, J. M. Melack, and J. J. Middelburg. 2006.

114

"The global abundance and size distribution of lakes, ponds, and impoundments." Limnology and Oceanography 51 (5):2388-2397.

Doyle, Martin W., Jai Singh, Rebecca Lave, and Morgan M. Robertson. 2015. "The morphology of streams restored for market and nonmarket purposes: Insights from a mixed natural-social science approach." Water Resources Research:n/a- n/a. doi: 10.1002/2015WR017030.

Drake, C.M., N.F. Stewart, M.A. Palmer, and V.L. Kindemba. 2010. The ecological status of ditch systems: an investigation into the current status of the aquatic invertebrate and plant communities of grazing marsh ditch systems in England and Wales. Peterborough: Buglife- the Invertebrate Conservation Trust.

Dudgeon, D., A. H. Arthington, M. O. Gessner, Z. Kawabata, D. J. Knowler, C. Leveque, R. J. Naiman, A. H. Prieur-Richard, D. Soto, M. L. Stiassny, and C. A. Sullivan. 2006. "Freshwater biodiversity: importance, threats, status and conservation challenges." Biol Rev Camb Philos Soc 81 (2):163-82. doi: 10.1017/S1464793105006950.

Elliot, Robert. 1982. "Faking Nature." In Environmental Ethics, edited by Andrew Light and Holmes Rolston III, 381-389. Malden, MA: Blackwell Publishing. Original edition, Inquiry 25: 81-93.

Ellis, E. C., and N. Ramankutty. 2008. "Putting people in the map: anthropogenic biomes of the world." Frontiers in Ecology and the Environment 6 (8):439-447. doi: 10.1890/070062.

Favre-Bac, L., A. Ernoult, C. Mony, Y. Rantier, J. Nabucet, and F. Burel. 2014. "Connectivity and propagule sources composition drive ditch plant metacommunity structure." Acta Oecologica-International Journal of Ecology 61:57-64. doi: 10.1016/j.actao.2014.10.006.

Felson, A. J., and S. T. A. Pickett. 2005. "Designed experiments: new approaches to studying urban ecosystems." Frontiers in Ecology and the Environment 3 (10):549-556. doi: 10.1890/1540-9295(2005)003[0549:denats]2.0.co;2.

115

Fluck, R. C., C. Fonyo, and E. Flaig. 1992. "Land-use-based phosphorus balances for Lake Okeechobee, Florida, drainage basins." Applied Engineering in Agriculture 8 (6):813-820.

Francesiaz, C., E. Guilbault, J. D. Lebreton, J. Trouvilliez, and A. Besnard. 2017. "Colony persistence in waterbirds is constrained by pond quality and land use." Freshwater Biology 62 (1):119-132. doi: 10.1111/fwb.12855.

Francis, R. A., and S. P. G. Hoggart. 2008. "Waste not, want not: The need to utilize existing artificial structures for habitat improvement along urban rivers." Restoration Ecology 16 (3):373-381. doi: 10.1111/j.1526-100X.2008.00434.x.

Gan, Q. L., and X. W. Li. 2016. "Impatiens baokangensis (Balsaminaceae), a new species from Hubei, China." Annales Botanici Fennici 53 (3-4):145-148.

Gelcich, S., and J. O'Keeffe. 2016. "Emerging frontiers in perceptions research for aquatic conservation." Aquatic Conservation-Marine and Freshwater Ecosystems 26 (5):986-994. doi: 10.1002/aqc.2714.

Gleick, P. H., A. Singh, and H. Shi. 2001. Threats to the World's Freshwater Resources. Oakland, California: Pacific Institute for Studies in Development, Environment, and Security.

Goslee, Sarah C., and Dean L. Urban. 2007. "The ecodist package for dissimilarity-based analysis of ecological data." Journal of Statistical Software 22 (7):1-19.

Gregory, K. J. 2006. "The human role in changing river channels." Geomorphology 79 (3- 4):172-191. doi: 10.1016/j.geomorph.2006.06.018.

Haff, P. K. 2010. "Hillslopes, rivers, plows, and trucks: mass transport on Earth's surface by natural and technological processes." Earth Surface Processes and Landforms 35 (10):1157-1166. doi: 10.1002/esp.1902.

Hale, Rebecca, Colden Baxter, Morey Burnham, and Kevin Marsh. 2017. "Co-producing urban river imaginaries over time: an Idaho river and its community." Society for Freshwater Science Annual Meeting, Raleigh, NC.

Hallett, Lauren M., Rachel J. Standish, Kristin B. Hulvey, Mark R. Gardener, Katharine N. Suding, Brian M. Starzomski, Stephen D. Murphy, and James A. Harris. 2013.

116

"Towards a Conceptual Framework for Novel Ecosystems." In Novel Ecosystems, 16-28. John Wiley & Sons, Ltd.

Harabis, F., and A. Dolny. 2015. "Necessity for the conservation of drainage systems as last refugia for threatened damselfly species, Coenagrion ornatum." Insect Conservation and Diversity 8 (2):143-151. doi: 10.1111/icad.12093.

Hawley, Joanna. 2014. "Isolated and ephemeral wetlands of southern Appalachia: Biotic communities and environmental drivers across multiple temporal and spatial scales." Doctor of Philosophy (PhD), Environmental Science, Clemson University.

Hergoualc'h, K. A., and L. V. Verchot. 2012. "Changes in soil CH4 fluxes from the conversion of tropical peat swamp forests: a meta-analysis." Journal of Integrative Environmental Sciences 9:31-39. doi: 10.1080/1943815x.2012.747252.

Herzon, I., and J. Helenius. 2008. "Agricultural drainage ditches, their biological importance and functioning." Biological Conservation 141 (5):1171-1183. doi: 10.1016/j.biocon.2008.03.005.

Higgs, Eric. 2003. Nature by Design. Boston, MA.: Massachusetts Institute of Technology.

Higgs, E. 2017. "Novel and designed ecosystems." Restoration Ecology 25 (1):8-13. doi: 10.1111/rec.12410.

Hobbie, S. E. 2014. "A Watershed Approach to Understanding Urban Eutrophication." Durham, NC, 7 Oct.

Hobbs, Richard J. 2016. "Degraded or just different? Perceptions and value judgements in restoration decisions." Restoration Ecology:n/a-n/a. doi: 10.1111/rec.12336.

117

Hobbs, R. J., S. Arico, J. Aronson, J. S. Baron, P. Bridgewater, V. A. Cramer, P. R. Epstein, J. J. Ewel, C. A. Klink, A. E. Lugo, D. Norton, D. Ojima, D. M. Richardson, E. W. Sanderson, F. Valladares, M. Vila, R. Zamora, and M. Zobel. 2006. "Novel ecosystems: theoretical and management aspects of the new ecological world order." Global Ecology and Biogeography 15 (1):1-7. doi: 10.1111/j.1466- 822x.2006.00212.x.

Hobbs, Richard J., Eric Higgs, and James A. Harris. 2009. "Novel ecosystems: implications for conservation and restoration." Trends in Ecology & Evolution 24 (11):599-605. doi: http://dx.doi.org/10.1016/j.tree.2009.05.012.

Hooke, R. L. 2000. "On the history of humans as geomorphic agents." Geology 28 (9):843- 846. doi: 10.1130/0091-7613(2000)28<843:othoha>2.0.co;2.

Hough, Palmer, and Morgan Robertson. 2009. "Mitigation under Section 404 of the Clean Water Act: where it comes from, what it means." Wetlands Ecology & Management 17 (1):15-33. doi: 10.1007/s11273-008-9093-7.

Hupp, Cliff B., Simon Dufour, and Gudrun Bornette. 2016. "Vegetation as a tool in the interpretation of fluvial geomorphic processes and landforms." In Tools in Fluvial Geomorphology, edited by G. Mathias Kondolf and Hervé Plégay. John Wiley & Sons, Ltd.

Hupy, J. P., and T. Koehler. 2012. "Modern warfare as a significant form of zoogeomorphic disturbance upon the landscape." Geomorphology 157:169-182. doi: 10.1016/j.geomorph.2011.05.024.

Jeffries, M. J. 2012. "Ponds and the importance of their history: an audit of pond numbers, turnover and the relationship between the origins of ponds and their contemporary plant communities in south-east Northumberland, UK." Hydrobiologia 689 (1):11-21. doi: 10.1007/s10750-011-0678-4.

Johns, Grace M. 2011. Compilation of Benefits and Costs of STA and Reservoir Projects in the South Florida Water Management District Hollywood, FL.

Kadoya, T., S. Suda, and I. Washitani. 2009. "Dragonfly crisis in Japan: A likely consequence of recent agricultural habitat degradation." Biological Conservation 142 (9):1899-1905. doi: 10.1016/j.biocon.2009.02.033.

118

Katz, Eric. 1983. "The big lie: Human restoration of nature." In Environmental Ethics, edited by Andrew Light and Holmes Rolston III, 390-397. Malden, MA: Blackwell Publishing. Original edition, Research in Philosophy and Technology 12.

Kaushal, Sujay, William McDowell, Wilfred Wollheim, Tamara Johnson, Paul Mayer, Kenneth Belt, and Michael Pennino. 2015. "Urban Evolution: The Role of Water." Water 7 (8):4063.

Kim, J., P. Goldsmith, and M. H. Thomas. 2010. "Economic impact and public costs of confined animal feeding operations at the parcel level of Craven County, North Carolina." Agriculture and Human Values 27 (1):29-42. doi: 10.1007/s10460-009- 9193-x.

Knutson, M. G., W. B. Richardson, D. M. Reineke, B. R. Gray, J. R. Parmelee, and S. E. Weick. 2004. "Agricultural ponds support amphibian populations." Ecological Applications 14 (3):669-684. doi: 10.1890/02-5305.

Koetsier, Peter, and Luana M. M. McCauley. 2015. "An Irrigation Canal as a Lotic Mesocosm: Examining the Relationship between Macroinvertebrate Benthos and Drift." Western North American Naturalist 75 (3):259-270. doi: 10.3398/064.075.0303.

Kueffer, Christoph, and Christopher N. Kaiser-Bunbury. 2013. "Reconciling conflicting perspectives for biodiversity conservation in the Anthropocene." Frontiers in Ecology and the Environment 12 (2):131-137. doi: 10.1890/120201.

Kumwimba, M. N., and B. Zhu. 2017. "Effectiveness of Vegetated Drainage Ditches for Domestic Sewage Effluent Mitigation." Bulletin of Environmental Contamination and Toxicology 98 (5):682-689. doi: 10.1007/s00128-017-2054-1.

LAF. 2016. Landscape Performance Series. Washington, D.C.: Landscape Architecture Foundation.

119

Landers, J. 2007. "Environmental engineering - Los Angeles aims to combine river restoration, urban revitalization." Civil Engineering 77 (4):11-13.

Lave, Rebecca. 2012. Fields and streams: stream restoration, neoliberalism, and the future of environmental science, Geographies of justice and social transformation ; 12. Athens: University of Georgia Press. Stream Restoration and Natural Channel Design -- The History of Stream Restoration and the Rise of Rosgen -- Capital Conflicts -- Building a Base of Support -- The Political Economy of Stream Restoration -- Conclusions -- Appendix: Interview and Survey Metadata. piecewiseSEM: Piecewise Structural Equation Modeling. R Project.

Leopold, L. B. 1977. "A reverence for rivers." Geology 5 (7):429-430. doi: 10.1130/0091- 7613(1977)5<429:ARFR>2.0.CO;2.

Li, M., L. X. Li, H. S. Liao, Z. W. Huang, and W. D. A. Huang. 2012. "The influence of drainage on wetland degradation in Zoige Plateau." Disaster Advances 5 (4):659- 666.

Li, Y. F., G. B. Zhu, W. J. Ng, and S. K. Tan. 2014. "A review on removing pharmaceutical contaminants from wastewater by constructed wetlands: Design, performance and mechanism." Science of the Total Environment 468:908-932. doi: 10.1016/j.scitotenv.2013.09.018.

Light, Andrew, Allen Thompson, and Eric S. Higgs. 2013. "Valuing Novel Ecosystems." In Novel Ecosystems, 257-268. John Wiley & Sons, Ltd.

Maheu-Giroux, M., and S. de Blois. 2007. "Landscape ecology of Phragmites australis invasion in networks of linear wetlands." Landscape Ecology 22 (2):285-301. doi: 10.1007/s10980-006-9024-z.

Mahl, U. H., J. L. Tank, S. S. Roley, and R. T. Davis. 2015. "Two-Stage Ditch Floodplains Enhance N-Removal Capacity and Reduce Turbidity and Dissolved P in Agricultural Streams." Journal of the American Water Resources Association 51 (4):923-940. doi: 10.1111/1752-1688.12340.

Mallin, M. A., and M. R. McIver. 2012. "Pollutant impacts to Cape Hatteras National Seashore from urban runoff and septic leachate." Marine Pollution Bulletin 64 (7):1356-1366. doi: 10.1016/j.marpolbul.2012.04.025.

120

Mao, I. F., M. R. Chen, L. Wang, M. L. Chen, S. C. Lai, and C. J. Tsai. 2012. "Method development for determining the malodor source and pollution in industrial park." Science of the Total Environment 437:270-275. doi: 10.1016/j.scitotenv.2012.08.056.

Mattah, P. A. D., G. Futagbi, L. K. Amekudzi, M. M. Mattah, D. K. de Souza, W. D. Kartey-Attipoe, L. Bimi, and M. D. Wilson. 2017. "Diversity in breeding sites and distribution of Anopheles mosquitoes in selected urban areas of southern Ghana." Parasites & Vectors 10:15. doi: 10.1186/s13071-016-1941-3.

Matulis, B. S., and J. R. Moyer. 2017. "Beyond Inclusive Conservation: The Value of Pluralism, the Need for Agonism, and the Case for Social Instrumentalism." Conservation Letters 10 (3):279-287. doi: 10.1111/conl.12281.

McMahan, E. A., J. M. Cloud, P. Josh, and M. Scott. 2016. "Nature with a Human Touch: Human-Induced Alteration Negatively Impacts Perceived Naturalness and Preferences for Natural Environments." Ecopsychology 8 (1):54-63. doi: 10.1089/eco.2015.0068.

McPhillips, Lauren E., Peter M. Groffman, Rebecca L. Schneider, and M. Todd Walter. 2016. "Nutrient Cycling in Grassed Roadside Ditches and Lawns in a Suburban Watershed." Journal of Environmental Quality. doi: 10.2134/jeq2016.05.0178.

Meade, R., and B. D. Wheeler. 1990. "Illtyd Pools, Brecon: description and conservation." New Phytologist 115 (1):187-199. doi: 10.1111/j.1469-8137.1990.tb00937.x.

Mehaffey, M. H., M. S. Nash, T. G. Wade, D. W. Ebert, K. B. Jones, and A. Rager. 2005. "Linking land cover and water quality in New York City's water supply watersheds." Environmental Monitoring and Assessment 107 (1-3):29-44. doi: 10.1007/s10661-005-2018-5.

Meybeck, M. 2003. "Global analysis of river systems: from Earth system controls to Anthropocene syndromes." Philosophical Transactions of the Royal Society B- Biological Sciences 358 (1440):1935-1955. doi: 10.1098/rstb.2003.1379.

Miller, B., M. E. Soule, and J. Terborgh. 2014. "'New conservation' or surrender to development?" Animal Conservation 17 (6):509-515. doi: 10.1111/acv.12127.

121

Milner, A. M., A. L. Robertson, L. E. Brown, S. H. Sonderland, M. McDermott, and A. J. Veal. 2011. "Evolution of a stream ecosystem in recently deglaciated terrain." Ecology 92 (10):1924-1935.

Naftz, D., C. Angeroth, T. Kenney, B. Waddell, N. Darnall, S. Silva, C. Perschon, and J. Whitehead. 2008. "Anthropogenic influences on the input and biogeochemical cycling of nutrients and mercury in Great Salt Lake, Utah, USA." Applied Geochemistry 23 (6):1731-1744. doi: 10.1016/j.apgeochem.2008.03.002.

Naito, R., M. Sakai, and Y. Morimoto. 2012. "Negative effects of deep roadside ditches on Pelophylax porosa brevipoda dispersal and migration in comparison with Hyla japonica in a rice paddy area in Japan." Zoological Science 29 (9):599-603. doi: 10.2108/zsj.29.599.

Needelman, Brian A., Peter J. A. Kleinman, Jeffrey S. Strock, and Arthur L. Allen. 2007. "Drainage Ditches: Improved management of agricultural drainage ditches for water quality protection: An overview." Journal of Soil and Water Conservation 62 (4):171-178.

Needelman, B. A., D. E. Ruppert, and R. E. Vaughan. 2007. "The role of ditch soil formation and redox biogeochemistry in mitigating nutrient and pollutant losses from agriculture." Journal of Soil and Water Conservation 62 (4):207-215.

O'Connor, J. H., and B. B. Rothermel. 2013. "Distribution and Population Characteristics of African Jewelfish and Brown Hoplo in Modified Wetlands in South Florida." American Midland Naturalist 170 (1):52-65. doi: 10.1674/0003-0031-170.1.52.

Oertli, Beat. 2018. "Editorial: Freshwater biodiversity conservation: The role of artificial ponds in the 21st century." Aquatic Conservation: Marine and Freshwater Ecosystems 28 (2):264-269. doi: doi:10.1002/aqc.2902. vegan: Community Ecology Package 2.2-1. R Foundation for Statistical Computing, Vienna, Austria.

Oosting, H. J. 1942. "An ecological analysis of the plant communities of Piedmont, North Carolina." Amer Midland Nat 28 ((1)):1-126. doi: 10.2307/2420696.

Ormerod, S. J., M. Dobson, A. G. Hildrew, and C. R. Townsend. 2010. "Multiple stressors in freshwater ecosystems." Freshwater Biology 55:1-4. doi: 10.1111/j.1365- 2427.2009.02395.x.

122

Palmer, M., J. D. Allan, J. Meyer, and E. S. Bernhardt. 2007. "River restoration in the twenty-first century: Data and experiential future efforts." Restoration Ecology 15 (3):472-481. doi: 10.1111/j.1526-100X.2007.00243.x.

Palmer, Margaret A. 2009. "Reforming Watershed Restoration: Science in Need of Application and Applications in Need of Science." Estuaries and Coasts 32 (1):1- 17. doi: 10.1007/s12237-008-9129-5.

Palmer, Margaret A., and Emily S. Bernhardt. 2006. "Hydroecology and river restoration: Ripe for research and synthesis." Water Resources Research 42 (3):n/a-n/a. doi: 10.1029/2005WR004354.

Palmer, M. A., M. Drake, and N. Stewart. 2013. A manual for the survey and evaluation of the aquatic plant and invertebrate assemblages of grazing marsh ditch systems. Peterborough: Buglife- The Invertebrate Conservation Trust.

Palmer, Margaret A., and Solange Filoso. 2009. "Restoration of Ecosystem Services for Environmental Markets." Science 325 (5940):575-576. doi: 10.1126/science.1172976.

Palmer, M. A., H. L. Menninger, and E. Bernhardt. 2010. "River restoration, habitat heterogeneity and biodiversity: a failure of theory or practice?" Freshwater Biology 55:205-222. doi: 10.1111/j.1365-2427.2009.02372.x.

Palta, Monica, Margaret V. du Bray, Rhian Stotts, Amanda Wolf, and Amber Wutich. 2016. "Ecosystem Services and Disservices for a Vulnerable Population: Findings from Urban Waterways and Wetlands in an American Desert City." Human Ecology 44 (4):463-478. doi: 10.1007/s10745-016-9843-8.

Palta, Monica M., Nancy B. Grimm, and Peter M. Groffman. 2017. "“Accidental” urban wetlands: ecosystem functions in unexpected places." Frontiers in Ecology and the Environment:n/a-n/a. doi: 10.1002/fee.1494.

123

Paulsen, Steven G., John L. Stoddard, Susan Holdsworth, Alice Mayio, and Ellen Tarquinio. 2004. Wadeable Streams Assessment: A Colllaborative Survey of the Nation's Streams. Washington D.C.: U.S. Environmental Protection Agency.

Pekel, Jean-François, Andrew Cottam, Noel Gorelick, and Alan S. Belward. 2016. "High- resolution mapping of global surface water and its long-term changes." Nature advance online publication. doi: 10.1038/nature20584 http://www.nature.com/nature/journal/vaop/ncurrent/abs/nature20584.html#supplemen tary-information.

Perez-Gutierrez, J. D., J. O. Paz, and M. L. M. Tagert. 2017. "Seasonal water quality changes in on-farm water storage systems in a south-central US agricultural watershed." Agricultural Water Management 187:131-139. doi: 10.1016/j.agwat.2017.03.014.

Perring, Michael P., Pete Manning, Richard J. Hobbs, Ariel E. Lugo, Cristina E. Ramalho, and Rachel J. Standish. 2013. "Novel Urban Ecosystems and Ecosystem Services." In Novel Ecosystems, 310-325. John Wiley & Sons, Ltd.

Petrovich, Olivera. 1999. "Preschool children's understanding of the dichotomy between the natural and the artificial." Psychological Reports 84 (1):3-27. doi: 10.2466/pr0.1999.84.1.3.

Phillips, Jonathan D. 1997. "A Short History of a Flat Place: Three Centuries of Geomorphic Change in the Croatan National Forest." Annals of the Association of American Geographers 87 (2):197-216. doi: 10.1111/0004-5608.872050.

Podolak, K., and G. M. Kondolf. 2016. "The Line of Beauty in River Designs: Hogarth's Aesthetic Theory on Capability Brown's Eighteenth-Century River Design and Twentieth-Century River Restoration Design." Landscape Research 41 (1):149- 167. doi: 10.1080/01426397.2015.1073705.

Polat, A. T., and A. Akay. 2015. "Relationships between the visual preferences of urban recreation area users and various landscape design elements." Urban Forestry and Urban Greening 14 (3):573-582. doi: 10.1016/j.ufug.2015.05.009.

Powell, G. E., A. D. Ward, D. E. Mecklenburg, and A. D. Jayakaran. 2007. "Two-stage channel systems: Part 1, a practical approach for sizing agricultural ditches." Journal of Soil and Water Conservation 62 (4):277-286.

124

Power, A. G. 2010. "Ecosystem services and agriculture: tradeoffs and synergies." Philosophical Transactions of the Royal Society B-Biological Sciences 365 (1554):2959-2971. doi: 10.1098/rstb.2010.0143.

Purdy, Jedediah. 2015. After nature : a politics for the anthropocene. Cambridge, Massachusetts ; London, England: Harvard University Press. Introduction -- An unequal terrain -- God's avid gardeners -- Nature as teacher -- Natural utopias -- A conservationist empire -- A wilderness passage into ecology -- Environmental law in the anthropocene -- What kind of democracy?

Rabkin, Jeremy A. 2014. "Against the EPA, Absurdity Is No Defense Symposium: The Federal Leviathan: Is There Any Area of Modern Life to Which Federal Government Does Not Extend: The Thirty-Second Annual Federalist Society National Student Symposium on Law and Public Policy - 2013 Section II: Environmental and Property Laws." Harvard Journal of Law & Public Policy 37:41-50.

Rapinel, S., L. Hubert-Moy, B. Clément, J. Nabucet, and C. Cudennec. 2015. "Ditch network extraction and hydrogeomorphological characterization using LiDAR- derived DTM in wetlands." Hydrology Research 46 (2):276-290. doi: 10.2166/nh.2013.121.

Rehounkova, K., and K. Prach. 2008. "Spontaneous vegetation succession in gravel-sand pits: A potential for restoration." Restoration Ecology 16 (2):305-312. doi: 10.1111/j.1526-100X.2007.00316.x.

Reid, N., S. K. Dingerkus, R. E. Stone, J. Buckley, T. J. C. Beebee, F. Marnell, and J. W. Wilkinson. 2014. "Assessing historical and current threats to common frog (Rana temporaria) populations in Ireland." Journal of Herpetology 48 (1):13-19. doi: 10.1670/12-053.

Remm, L., P. Lohmus, M. Leis, and A. Lohmus. 2013. "Long-Term Impacts of Forest Ditching on Non-Aquatic Biodiversity: Conservation Perspectives for a Novel Ecosystem." Plos One 8 (4):13. doi: 10.1371/journal.pone.0063086.

125

Renwick, William H., Richard O. Sleezer, Robert W. Buddemeier, and Steven V. Smith. 2006. "Small artificial ponds in he United States: Impacts on sedimentation and the carbon budget." Proceedings of the Eighth Federal Interagency Sedimentation Conference, Reno, Nevada.

Ridder, B. 2007. "An exploration of the value of naturalness and wild nature." Journal of Agricultural & Environmental Ethics 20 (2):195-213. doi: 10.1007/s10806-006- 9025-6.

Robert, A., C. Fontaine, S. Veron, A. C. Monnet, M. Legrand, J. Clavel, S. Chantepie, D. Couvet, F. Ducarme, B. Fontaine, F. Jiguet, I. le Viol, J. Rolland, F. Sarrazin, C. Teplitsky, and M. Mouchet. 2017. "Fixism and conservation science." Conservation Biology 31 (4):781-788. doi: 10.1111/cobi.12876.

Roley, S. S., J. L. Tank, J. C. Tyndall, and J. D. Witter. 2016. "How cost-effective are cover crops, wetlands, and two-stage ditches for nitrogen removal in the Mississippi River Basin?" Water Resources and Economics 15:43-56. doi: 10.1016/j.wre.2016.06.003.

Rosenvald, R., R. Jarvekulg, and A. Lohmus. 2014. "Fish assemblages in forest drainage ditches: Degraded small streams or novel habitats?" Limnologica 46:37-44. doi: 10.1016/j.limno.2013.12.004.

Rosenzweig, M. L. 2003. "Reconciliation ecology and the future of species diversity." Oryx 37 (2):194-205. doi: 10.1017/s0030605303000371.

Rosenzweig, Michael L. 2003. Win-win ecology : how the earth's species can survive in the midst of human enterprise: Oxford ; New York ; Oxford University Press, 2003.

Ross, Matthew R.V., Emily S. Bernhardt, Martin Doyle, and James B. Heffernan. 2015. "Designer Ecosystems: Incorporating Design Approaches into Applied Ecology." Annual Review of Environment and Resources 40 (1):null. doi: doi:10.1146/annurev-environ-121012-100957. lavaan: Latent Variable Analysis 0.5-23.1097. R Project, CRAN.

Rozin, Paul, Mark Spranca, Zeev Krieger, Ruth Neuhaus, Darlene Surillo, Amy Swerdlin, and Katherine Wood. 2004. "Preference for natural: instrumental and

126

ideational/moral motivations, and the contrast between foods and medicines." Appetite 43 (2):147-154. doi: https://doi.org/10.1016/j.appet.2004.03.005.

Ruggiero, A., R. Cereghino, J. Figuerola, P. Marty, and S. Angelibert. 2008. "Farm ponds make a contribution to the biodiversity of aquatic insects in a French agricultural landscape." Comptes Rendus Biologies 331 (4):298-308. doi: 10.1016/j.crvi.2008.01.009.

Salomon Cavin, Joëlle. 2013. "Beyond prejudice: Conservation in the city. A case study from Switzerland." Biological Conservation 166 (0):84-89. doi: http://dx.doi.org/10.1016/j.biocon.2013.06.015.

Sayer, C., K. Andrews, E. Shilland, N. Edmonds, R. Edmonds-Brown, I. Patmore, D. Emson, and J. Axmacher. 2012. "The role of pond management for biodiversity conservation in an agricultural landscape." Aquatic Conservation-Marine and Freshwater Ecosystems 22 (5):626-638. doi: 10.1002/aqc.2254.

Sayer, Carl D. 2014. "Conservation of aquatic landscapes: ponds, lakes, and rivers as integrated systems." Wiley Interdisciplinary Reviews: Water 1 (6):573-585. doi: 10.1002/wat2.1045.

Schindler, David W., Stephen R. Carpenter, Steven C. Chapra, Robert E. Hecky, and Diane M. Orihel. 2016. "Reducing Phosphorus to Curb Lake Eutrophication is a Success."Environmental Science & Technology 50 (17):8923-8929. doi: 10.1021/acs.est.6b02204.

Sebastian-Gonzalez, E., J. A. Sanchez-Zapata, and F. Botella. 2010. "Agricultural ponds as alternative habitat for waterbirds: spatial and temporal patterns of abundance and management strategies." European Journal of Wildlife Research 56 (1):11-20. doi: 10.1007/s10344-009-0288-x.

Setzer, C., and M. E. Domino. 2004. "Medicaid outpatient utilization for waterborne pathogenic illness following Hurricane Floyd." Public Health Reports 119 (5):472- 478. doi: 10.1016/j.phr.2004.07.004.

127

Shaw, R. F., P. J. Johnson, D. W. Macdonald, and R. E. Feber. 2015. "Enhancing the Biodiversity of Ditches in Intensively Managed UK Farmland." Plos One 10 (10):19. doi: 10.1371/journal.pone.0138306.

Simon, Troy N., and Joseph Travis. 2011. "The contribution of man-made ditches to the regional stream biodiversity of the new river watershed in the Florida panhandle." Hydrobiologia 661 (1):163-177. doi: 10.1007/s10750-010-0521-3.

Smith, C., E. H. Boughton, and S. Pierre. 2015. "Pomacea maculata (Island Apple Snail) Invasion in Seasonal Wetlands on Florida Ranchland: Association with Plant- Community Structure and Aquatic-Predator Abundance." Southeastern Naturalist 14 (3):561-576. doi: 10.1656/058.014.0314.

Smith, R. F., R. J. Hawley, M. W. Neale, G. J. Vietz, E. Diaz-Pascacio, J. Herrmann, A. C. Lovell, C. Prescott, B. Rios-Touma, B. Smith, and R. M. Utz. 2016. "Urban stream renovation: Incorporating societal objectives to achieve ecological improvements." Freshwater Science 35 (1):364-379. doi: 10.1086/685096.

Soule, M. 2013. "The "New Conservation"." Conservation Biology 27 (5):895-897. doi: 10.1111/cobi.12147.

Spears, Ellen Griffith. 2014. Baptized in PCBs: Race, Pollution, and Justice in an All- American Town. Chapel Hill, NC: UNC Press Books.

Steele, M. K., J. B. Heffernan, N. Bettez, J. Cavender-Bares, P. M. Groffman, J. M. Grove, S. Hall, S. E. Hobbie, K. Larson, J. L. Morse, C. Neill, K. C. Nelson, J. O’Neil- Dunne, L. Ogden, D. E. Pataki, C. Polsky, and R. Roy Chowdhury. 2014. "Convergent Surface Water Distributions in U.S. Cities." Ecosystems 17 (4):685- 697. doi: 10.1007/s10021-014-9751-y.

Stewart, J. S., L. Z. Wang, J. Lyons, J. A. Horwatich, and R. Bannerman. 2001. "Influences of watershed, riparian-corridor, and reach-scale characteristics on aquatic biota in agricultural watersheds." Journal of the American Water Resources Association 37 (6):1475-1487. doi: 10.1111/j.1752-1688.2001.tb03654.x.

Suding, K. N. 2011. "Toward an Era of Restoration in Ecology: Successes, Failures, and Opportunities Ahead." In Annual Review of Ecology, Evolution, and Systematics, Vol 42, edited by D. J. Futuyma, H. B. Shaffer and D. Simberloff, 465-487. Palo Alto: Annual Reviews.

128

Sueltenfuss, J. P., D. J. Cooper, R. L. Knight, and R. M. Waskom. 2013. "The Creation and Maintenance of Wetland Ecosystems from Irrigation Canal and Reservoir Seepage in a Semi-Arid Landscape." Wetlands 33 (5):799-810. doi: 10.1007/s13157- 013-0437-6.

Tallis, H., and J. Lubchenco. 2014. "A call for inclusive conservation." Nature 515 (7525):27-28.

Taylor, C. A., and H. W. Robison. 2016. "A new burrowing crayfish of the genus Fallicambarus Hobbs, 1969 (Decapoda: Cambaridae) from the Red River Drainage of the southcentral United States." Zootaxa 4144 (4):575-583. doi: 10.11646/zootaxa.4144.4.7.

Thomas, D. L., C. D. Perry, R. O. Evans, F. T. Izuno, K. C. Stone, and J. W. Gilliam. 1995. "AGRICULTURAL DRAINAGE EFFECTS ON WATER-QUALITY IN SOUTHEASTERN US." Journal of Irrigation and Drainage Engineering-Asce 121 (4):277-282. doi: 10.1061/(asce)0733-9437(1995)121:4(277).

Tranvik, L. J., J. A. Downing, J. B. Cotner, S. A. Loiselle, R. G. Striegl, T. J. Ballatore, P. Dillon, K. Finlay, K. Fortino, L. B. Knoll, P. L. Kortelainen, T. Kutser, S. Larsen, I. Laurion, D. M. Leech, S. L. McCallister, D. M. McKnight, J. M. Melack, E. Overholt, J. A. Porter, Y. Prairie, W. H. Renwick, F. Roland, B. S. Sherman, D. W. Schindler, S. Sobek, A. Tremblay, M. J. Vanni, A. M. Verschoor, E. von Wachenfeldt, and G. A. Weyhenmeyer. 2009. "Lakes and reservoirs as regulators of carbon cycling and climate." Limnology and Oceanography 54 (6):2298-2314. doi: 10.4319/lo.2009.54.6_part_2.2298.

Twisk, W., M. A. W. Noordervliet, and W. J. ter Keurs. 2003. "The nature value of the ditch vegetation in peat areas in relation to farm management." Aquatic Ecology 37 (2):191-209. doi: 10.1023/a:1023944028022.

USACE. 2010. Regional Supplement to the Corps of Engineers Wetland Delineation Manual: Atlantic and Gulf Coastal Plain Region (Version 2.0). edited by Research and Development Center. Vicksburg, MS: U.S. Army Corps of Engineers.

USACE. 2013. National Inventory of Dams. Washington, D.C.: U.S. Army Corps of Engineers.

129

USDA. 2015. Summary Report: 2012 National Resources Inventory. Washington, D.C. and Ames, IA: Natural Resources Conservation Service and Center for Survey Statistics and Methodology, Iowa State University.

USDA. 2015. The Plants Database. Greensboro, NC: U.S. Department of Agriculture Natural Resources Conservation Survey.

USEPA. 2008. Final total maximum daily load for biochemical oxygen demand, dissolved oxygen, and nutrients in the Lake Okeechobee tributaries Osceola, Polk, Okeechobee, Highlands, Glade, and Martin, Florida. In United States Court of Appeals, District of Columbia Circuit. Atlanta, Georgia: U.S. Environmental Protection Association.

USEPA. 2009. National Lakes Assessment: A Collaborative Survey of the Nation's Lakes. Washington, D.C.: U.S. Environmental Protection Agency.

USEPA. 2016. National Lakes Assessment 2012: A Collaborative Survey of Lakes in the United States. Washington, D.C.: U.S. Environmental Protection Agency.

USEPA, and USACE. 2015. “Definition of ‘Waters of the United States’ Under the Clean Water Act” In 40 CFR 230.3. Washington, D.C.: U.S. Army Corps of Engineers and U.S. Environmental Protection Agency.

USFWS. 2017. National Wetlands Inventory. Washington, D.C. : U.S. Department of the Interior, Fish and Wildlife Service.

USGS. 2015. National Hydrography Dataset. edited by USGS. Washington, D.C.: US Geological Survey.

Vad, C. F., A. L. Pentek, N. J. Cozma, A. Foldi, A. Toth, B. Toth, N. A. Bode, A. Mora, R. Ptacnik, E. Acs, K. Zsuga, and Z. Horvath. 2017. "Wartime scars or reservoirs of biodiversity? The value of bomb crater ponds in aquatic conservation." Biological Conservation 209:253-262. doi: 10.1016/j.biocon.2017.02.025. van Gerven, Luuk P. A., Jan J. Kuiper, Jan H. Janse, Annette B. G. Janssen, Michel Jeuken, Wolf M. Mooij, and Jeroen J. M. de Klein. 2017. "How Regime Shifts in Connected Aquatic Ecosystems Are Affected by the Typical Downstream Increase of Water Flow." Ecosystems 20 (4):733-744. doi: 10.1007/s10021-016-0061- 4.

130

Veraart, A. J., M. R. Dimitrov, A. P. Schrier-Uijl, H. Smidt, and J. J. M. de Klein. 2017. "Abundance, Activity and Community Structure of Denitrifiers in Drainage Ditches in Relation to Sediment Characteristics, Vegetation and Land-Use." Ecosystems 20 (5):928-943. doi: 10.1007/s10021-016-0083-y.

Verlicchi, P., and E. Zambello. 2014. "How efficient are constructed wetlands in removing pharmaceuticals from untreated and treated urban wastewaters? A review." Science of the Total Environment 470:1281-1306. doi: 10.1016/j.scitotenv.2013.10.085.

Vincent, R. E., D. M. Burdick, and M. Dionne. 2013. "Ditching and Ditch-Plugging in New England Salt Marshes: Effects on Hydrology, Elevation, and Soil Characteristics." Estuaries and Coasts 36 (3):610-625. doi: 10.1007/s12237-012- 9583-y.

Walsh, Christopher J., Allison H. Roy, Jack W. Feminella, Peter D. Cottingham, Peter M. Groffman, and Raymond P. Morgan. 2009. "The urban stream syndrome: current knowledge and the search for a cure." Journal of the North American Benthological Society 24 (3):706-723. doi: 10.1899/04-028.1.

Walter, Robert C., and Dorothy J. Merritts. 2008. "Natural Streams and the Legacy of Water-Powered Mills." Science 319 (5861):299-304. doi: 10.1126/science.1151716.

Wang, Y. C. 2012. "Examining Landscape Determinants of Opisthorchis viverrini Transmission." Ecohealth 9 (3):328-341. doi: 10.1007/s10393-012-0789-z.

Wilkinson, B. H. 2005. "Humans as geologic agents: A deep-time perspective." Geology 33 (3):161-164. doi: 10.1130/G21108.1.

Williamson, C. E., J. E. Saros, W. F. Vincent, and J. P. Smol. 2009. "Lakes and reservoirs as sentinels, integrators, and regulators of climate change." Limnology and Oceanography 54 (6):2273-2282. doi: 10.4319/lo.2009.54.6_part_2.2273.

131

Wohl, E., S. N. Lane, and A. C. Wilcox. 2015. "The science and practice of river restoration." Water Resources Research 51 (8):5974-5997. doi: 10.1002/2014wr016874.

Wohl, E., K. B. Lininger, and J. Baron. 2017. "Land before water: The relative temporal sequence of human alteration of freshwater ecosystems in the conterminous United States." Anthropocene 18:27-46. doi: 10.1016/j.ancene.2017.05.004.

Zang, D. K. 2017. "Cerasus laoshanensis (Rosaceae), a new species from Shandong, China." Annales Botanici Fennici 54 (1-3):135-137.

Zedler, Joy B., and Suzanne Kercher. 2005. "WETLAND RESOURCES: Status, Trends, Ecosystem Services, and Restorability." Annual Review of Environment and Resources 30 (1):39-74. doi: doi:10.1146/annurev.energy.30.050504.144248.

Zhang, Z. D., R. Brown, J. Bauer, and A. Bedard-Haughn. 2017. "Nutrient dynamics within drainage ditches under recent, medium, and long-term drainage in the Black soil zone of southeastern Saskatchewan." Geoderma 289:66-71. doi: 10.1016/j.geoderma.2016.11.027.

Zollitsch, Brenda, and Jeanne Christie. 2014. Report on State Definitions, Jurisdiction and Mitigation Requirements in State Programs for Ephemeral, Intermittent and Perennial Streams in the United States. Association of State Wetland Managers.

Zollitsch, Brenda, and Jeanne Christie. 2016. Status and Trends Report on State Wetland Programs in the United States. Association of State Wetland Managers.

132

Biography

Chelsea Clifford was born on March 11, 1989 in Williamsburg Virginia, and grew up in Gloucester, Virginia; she originated in the salt marshes of southeastern Virginia.

She got a BA in biology and environmental studies from Carleton College in 2010.Before graduate school, she worked for Chesapeake Environmental Communications, The

Nature Conservancy, MacArthur Agro-Ecology Research Center, White Mountain

Research Station, and the Virginia Department of Conservation and Recreation. She also became a Virginia Master Naturalist. Chelsea began a PhD in Environment at Duke

University in fall 2013, and expects to graduate in summer 2018.She values water, soil, air, life, knowledge, ideas, memory, expression, education, exploration, community, sharing, kindness, integrity, effort, and play.

133