Chapter 10

Channel Geomorphology: Fluvial Forms, Processes, and Forest Management Effects

Dan L. Hogan and David S. Luzi

Introduction

This chapter deals with fluvial geomorphology—that that result from human influence on the landscape. is, with the landforms developed by flowing water. Streams of all sizes have the same basic func- This seemingly simple subject encompasses a rather tion: they move water, sediment, and other matter complex set of interrelated processes that produce (organic, chemical, biological materials, etc.) over a diverse array of fluvial forms. We restrict the the land surface and, ultimately, empty all of these content of this compendium chapter to those aspects materials into the ocean or a lake. The details regard- most relevant to the professional hydrologist and ing time and space scales are numerous; to enable us geomorphologist working in British Columbia, a to consider these factors, we approach this chapter province that is physiographically diverse (Chapter within a watershed context. A watershed includes 2, “Physiography of British Columbia”) and, there- the entire stream network upstream of a point on the fore, includes many types of streams and rivers. For mainstem, as well as the hillslopes contributing wa- simplicity, we use “streams” as an all-encompassing ter, sediment, and woody debris to the network. The term that includes all channels with flowing water fluvial dynamics at a given point along a stream are (creeks, brooks, tributaries, rivers, etc.), regardless influenced by the various processes occurring in the of the absolute size or the timing and frequency of watershed upstream, which is why it is necessary to flows carried by the channels. This chapter provides consider processes that function at a watershed scale. an overview of the key factors determining the mor- Although most management decisions are made to phology and dynamics of forested watershed streams protect a particular stream reach, it is important to in British Columbia and shows how forest manage- understand that this protection can only be achieved ment influences these factors. It is not intended to by maintaining properly functioning conditions provide a complete and comprehensive treatise of the within the contributing watershed. A stream reach is subject; the reader is referred to the many textbooks defined here as a length of channel with homogene- and technical journals for further, detailed cover- ous morphology, discharge, and hillslope-channel age of fluvial geomorphology (e.g., Leopold et al. coupling. 1964; Schumm 1977; Calow and Petts [editors] 1992; For this chapter, we do not intend to prescribe Knighton 1998; Naiman and Bilby [editors] 1998; best management practices for forestry operations, Wohl 2000; Bridge 2003; Bennett and Simon [edi- as these will vary regionally, but rather provide a tors] 2004). We do, however, show the links to foundation on which decisions can be made to mini- other chapters in this volume and then consider the mize channel impacts. To do this, we consider the effects on stream channel conditions and dynamics following points, which underlie our objective.

331 • The watershed is the fundamental landscape unit • Depending on the type of watershed, the produc- that must be understood when considering stream tion, delivery, and calibre of sediment supplied to channel processes. the channel are often the most significant of the • Important differences exist among watersheds potential factors influencing channel morphology. in British Columbia that are related to variations • Forest management activities can interfere with in climate, physiography, and land use, as well as some factors that determine channel morphology ecologic and geomorphic legacies. and can, if not conducted in a prudent manner, • Specific factors within a watershed determine have long-term and widespread harmful impacts channel morphology; some of these factors are on stream environments. determined by the conditions of the surrounding terrain and others are associated with the proper- These points are expanded in the following sections. ties of the fluvial network.

Factors Controlling Channel Morphology

Factors controlling channel morphology can be magnitude, and duration of streamflows. Both divided into those that are imposed on the watershed temporal and spatial variability in discharge can (i.e., independent) and those that adjust to the im- have a large influence on channel morphology (for posed conditions (i.e., dependent). Only the depen- a discussion of flood-generating mechanisms in dent factors can be influenced by forest management British Columbia, see Chapter 4, “Regional Hydrol- activities. The independent landscape factors con- ogy”). Riparian vegetation has an important influ- trolling channel morphology are geology, climate, ence on bank erodibility and near-bank hydraulic and human (Figure 10.1). The geology of a watershed conditions, and is also a source of in-channel large is determined by processes acting at the landscape woody debris (LWD). Classic conceptual models de- scale, and can include volcanism, tectonics, and, to a picted channel morphology as primarily a function lesser extent, surficial processes such as glacial ero- of streamflow and sediment transport rate, where sion and deposition (see Chapter 2, “Physiography transport rate equals sediment supply for equilibri- of British Columbia”). Within a watershed, these um conditions (e.g., Lane 1955; Blench 1957; Schumm processes control the distribution, structure and 1971). However, these models did not explicitly ad- type of bedrock, surficial materials, and topography dress the role of vegetation or other boundary condi- (Montgomery 1999). Climate is considered an inde- tions, which often play a critical role in determining pendent factor at the landscape scale, as it is driven channel morphology. by synoptic conditions related to global atmospheric In addition to riparian vegetation, important circulation patterns (see Chapter 3, “Weather and boundary conditions include elements found within Climate”). Human alteration of the landscape can the stream channel, as well as those that may influ- also significantly change watershed conditions. ence the channel’s ability to migrate laterally and The geologic, climatic, and human conditions (or) build vertically. The most important boundary to which a watershed is subjected determine the conditions include: dependent landscape variables of sediment supply, stream discharge, and vegetation (Montgomery and • bank composition and structure, which influence Buffington 1993; Buffington et al. 2003). Channel bank erodibility as determined by the sedimentol- morphology is the result of the combined influ- ogy and geotechnical properties of the material ence of the dependent landscape variables, and the bounding the channel; channel responds to changes in these variables by • bedrock and other non-erodible units (such as adjustments in one or many of the dependent chan- colluvial material, compact tills, and lag glacioflu- nel variables (Figure 10.1). An additional important vial deposits), which may limit lateral and vertical independent variable is time since disturbance. channel migration and determine stream channel Sediment supply is determined by the frequency, alignment; volume, and calibre of material delivered to the • erodible sediment stored in valley bottoms in channel. Stream discharge includes the frequency, floodplains, fans, or terraces (including alluvial

332 Independent Landscape Geology Climate Human Variables

Sediment supply Stream discharge Vegetation

Dependent Frequency, volume Frequency, volume Riparian vegetation: bank Landscape and calibre of and duration of stability and local flow Variables sediment streamflow hydraulics; in-channel large woody debris (orientation and position)

Dependent width, depth, bed slope, Channel grain size, bedforms, Variables sinuosity, scour depth

FIGURE 10.1 Governing conditions as independent landscape and watershed variables and the dependent channel variables (modified from Montgomery and Buffington 1993 and Buffington et al. 2003).

sediments; lacustrine, marine, and glacial out- well as the history of human intervention. Church wash deposits; and fine-textured colluvium); and Slaymaker (1989) suggested that, because many • valley slope, which, although related to the vol- streams in British Columbia are still constrained by ume of stored sediment in the valley, represents boundary sediments deposited during glaciation, the the short- to medium-term maximum possible streams have not completely adjusted to post-glacial gradient that a stream channel can attain; and conditions. The current morphology of a stream is, • human channel alternations, such as culverts, rip- therefore, a product of both present-day and historic rap bank protection, bridge crossings, and flood watershed processes. Thus, an awareness of both protection works. present and historical influences on current stream morphology is important (see the Yakoun case study These boundary conditions are primarily influ- below for an example). enced by the geomorphic history of a landscape, as

Channel Types, Morphology, and Indicators of Disturbance

Channel type classifications can be based on the type channel bed and bank strength and the channel’s of material through which streams flow and in which threshold of erodibility (Kellerhals et al. 1976). Three channels form. Schumm (1985) proposed a chan- categories of materials are used in this chapter: (1) nel classification that included three categories: (1) non-erodible, (2) semi-erodible, and (3) erodible. bedrock, (2) semi-controlled, and (3) alluvial; how- These terms (as opposed to the conventional “non- ever, this classification does not address the variable alluvial” and “alluvial”) are more useful from a geotechnical properties associated with the glaciated forest-operations perspective. Although, by defini- landforms found across British Columbia. Catego- tion, all contemporary alluvial material is erodible, ries should be based on the materials that determine many non-alluvial materials are also highly erodible

333 (e.g., marine, lacustrine, and glaciofluvial deposits). Alluvial channels comprise the major type of Similarly, some alluvial materials are far less erodible channel within the erodible material category. than others; for instance, armoured channel beds Alluvial channels often develop within larger alluvial developed by fluvial processes are far more resistant landscapes, such as along main valleys with fan com- to movement than other alluvium such as gravel-bar plexes and floodplain features. This type of channel deposits, which are rearranged on an annual basis. frequently flows through erodible material that has Here, we are interested in a classification scheme that been previously eroded, transported, and deposited distinguishes features on the basis of their suscepti- by flowing water. Streams bounded by alluvial bility to changes from forest management activities. sediment are active, and relatively major pattern For channels that flow through non-erodible changes may occur as the channel migrates laterally materials (e.g., bedrock, coarse colluvium, and non- across the alluvial deposit. Major pattern changes erodible glacial deposits), boundary conditions tend may also occur because of changes in the governing to dominate the channel morphology. This type of factors, such as those produced by upstream land channel usually has a limited sediment supply and a use. Stream channels may form in other erodible, but morphology that is largely determined by the struc- non-alluvial, materials and these warrant careful ture and composition of the material through which consideration when planning forestry activities. it flows. Bedrock channels, for example, frequently Channels can also be classified according to plan- run along faults or other geologic planes of weakness form pattern, which in turn is a function of water- within the rock. Overall, these channels are rela- shed properties. To classify fluvial landforms during tively insensitive to disturbances, including distur- air photo interpretation, Mollard (1973) identified bances from changes occurring upstream (i.e., the 17 planform channel types that were related to both channel is relatively stable), but bedrock channels the physiographic environment in which channels are very effective at transferring disturbances from flowed, and the materials that made up the channel upstream to downstream reaches. Although bedrock bed and banks. In general, Mollard (1973) based this channels resist erosion, Montgomery et al. (1996) ob- channel pattern classification on the factors control- served that LWD could promote sedimentation that ling morphology, specifically streamflow, sediment potentially causes bedrock reaches to change from supply, the relative dominance of fluvial transport non-erodible to “forced” erodible (alluvial) zones. processes, and the materials within which the chan- The opposite is also evident where long expanses of nel is formed. former erodible (alluvial) channels are degraded to Church (1992) classified channel patterns on the non-erodible bedrock zones downstream of logjams, basis of the calibre and volume of sediment supply, which impede the downstream transfer of sediment separating the patterns into phases related to how (Hogan and Bird 1998). These situations support the supplied sediment was then transported (Figure watershed and riparian protection initiatives, par- 10.2). These patterns include phases dominated by ticularly in environments where sediment transfer bed material supply and wash material supply, and a changes can occur, even though the particular zone transitional phase where neither bed nor wash mat- is non-erodible. erial dominates. Bed material generally forms the Channels flowing throughsemi-erodible material coarser part of the sediment load a channel is able to may have reaches that alternate between zones transport and constitutes the bed and lower banks flowing through non-erodible, partially erodible, or of the stream. The wash material is the finer part of fully erodible materials. This becomes a scale issue the sediment load and is generally transported long and leads to classification problems; for instance, how distances, being deposited in the upper banks and extensive should a non-erodible channel section be on the floodplain. Generally, bed material consists before it is classified as “non-erodible”? The degree of of coarser sands, gravel, cobbles, and boulders; wash erodibility can also vary along either the channel’s material consists of finer sands, silts, and clays. The banks or bed, depending on local boundary condi- flow energy of the particular channel determines tions (i.e., the degree of erodibility of the boundary). whether sand-sized sediment is assigned to bed or Although the transitional nature of this type of wash material (Church 2006). channel can make classification problematic, its For applied or operational purposes, much of identification is important because these channels are the work of Mollard (1973), Schumm (1977), and relatively sensitive to changes in the governing factors. Church (1992) can be used to assess channel form

334 Decreasing channel stability Increasing sediment supply

Boulders, cobbles

Gravel dominant phase dominant Bed material supply Bed material

Sand

Braided channels Wandering channels

Sand phase Transitional

Gravel

Anastomosed channels Meandering channels

Fine sand, silt dominant phase dominant Wash material supply material Wash Increasing channel gradient channel stabilityDecreasing calibre Increasing sediment

FIGURE 10.2 Channel form (B.C. Ministry of Forests and B.C. Ministry of Environment, Lands and Parks 1996a, after Church 1992).

335 and function. Figure 10.2 shows general trends in or more individual channels. When the channel is channel form as these relate to the governing factors, not too active it can divide and recombine around and the direction of channel stability as it relates to stable, vegetated islands; these are called “wandering channel form. In this context, “stability” refers to a channels” (Figure 10.2). In other situations, the chan- channel’s propensity for vertical or lateral movement nel becomes too active for stable vegetated islands to (Church 2006). This particular diagram was modi- develop, and the system divides into numerous indi- fied for the province’s Channel Assessment Proce- vidual channels that divide and recombine around dure (B.C. Ministry of Forests and B.C. Ministry of unstable gravel bars; these are called “braided Environment, Lands and Parks 1996a) and is used channels.” Characterized by rapid lateral migration as a preliminary assessment tool to establish chan- rates, and often undergoing net vertical aggradation, nel attributes and to document channel pattern braided channels are amongst the most active of all changes over time. As sediment supply increases the stream channels in British Columbia. above the transport capacity of the channel, sedi- Detectable changes in channel pattern indicate ment is deposited (aggradation), which increases important changes in both the watershed and the the channel width-to-depth ratio, and the level of factors controlling morphology. Managers can use channel stability decreases. Channel aggradation the evidence of channel changes, as prepared by a is evident on aerial photographs as an increase in hydrologist and (or) geomorphologist, as an indica- the size, number, and extent of sediment accumula- tor of the environmental health of the watershed. In tions within the channel, when compared to earlier general, as sediment supply or streamflow increases, photographs. For channels with moderate-sized bed channel pattern becomes straighter (Figure 10.3). material (such as gravel-bed streams), channels with Changes in the patterns of in-channel sediment moderate sediment supply usually have a straight storage in bars and islands are also an early indicator or sinuous planform. As the supply rate approaches of future channel problems. Bars are non-vegetated or exceeds the channel’s capacity to transport the accumulations of sediment typically exposed above additional sediment, the channel may break into two the low-water level that often develop on the sides

Direction of increasing sediment supply

Tortuous Regular Irregular meanders meanders meanders Wandering Sinuous Straight

Confined pattern

FIGURE 10.3 Channel pattern classification (modified from Kellerhals et al. 1976).

336 of the channel, although these accumulations may increase in the number of islands generally indicates also form in the middle of the channel (Figure 10.4). increased sediment supply. Bars are aggregate features, the stability of which Changes in the lateral activity of the channel (i.e., is a function of the interlocking nature of many displacement of the channel laterally across a valley smaller particles into a larger feature. Changes in flat surface) may also indicate variations in the con- bar morphology over time usually indicate varia- ditions upstream (Figure 10.6). Lateral movement is tions in upstream sediment supply. For example, if often caused by progressive bank erosion or channel the bars of a stream reach are predominantly medial avulsion. Progressive bank erosion can be the result bars when historically they had been point bars, this of sediment aggradation within the channel or can may indicate a general increase in sediment sup- occur simply from natural meandering processes. In plied to the reach. In contrast, channel islands are contrast to bank erosion, channel avulsion is usually vegetated with the top surfaces occurring at or above a relatively sudden and major shift in the position of bankfull channel height (Figure 10.5). Islands are the channel to a new part of the floodplain (first- relatively stable over time, but expand and contract order avulsion), a sudden re-occupation of an old in response to long-term sediment supply rates; an channel on the floodplain (second-order avulsion),

Longitudinal and crescentric bars

D ire ct io n o f in c re Transverse bars as in g

se d im e n t s u Medial bars p p ly

Point or lateral bars

Diagonal bars

FIGURE 10.4 Channel bars (B.C. Ministry of Forests and B.C. Ministry of Environment, Lands and Parks 1996a, after Church and Jones 1982).

337 Direction of increasing sediment supply

Occasional Infrequent, irregular Frequent, regular Split Anastomosing No overlapping of Infrequent Not overlapping, Islands overlap Continuously islands, average overlapping, average average spacing less frequently or overlapped islands, spacing being 10 or spacing less than 10 than 10 channel continuously, usually with 2 or more flow more channel widths channel widths widths 2 or 3 flow branches branches

FIGURE 10.5 Channel islands (after Kellerhals et al. 1976). or a relatively minor switching of channels within a movement and enable a stable channel morphology braided channel or other similarly active channels to exist in an environment in which it otherwise may (third-order avulsion) (Nanson and Knighton 1996). not occur. Millar (2000) illustrated how Slesse Creek Logjams accumulating along certain streams are an evolved from a stable sinuous gravel-bed river to an interesting aspect in British Columbia’s forest lands. unstable braided morphology after the removal of These features, discussed later in the chapter, can riparian vegetation. For additional information on have a dramatic influence on both bank erosion and lateral channel movement, refer to Rapp and Abbe avulsion processes over long time periods and large (2003) who present a detailed discussion of channel areas. migration and the methods and tools to delineate Lateral channel movement influences the riparian boundaries for historic, current, and potential lateral zone, eroding some areas and building up others. channel movement. The channel’s boundary conditions and the relation- Several systems are used to differentiate the vari- ship between the stream and the valley through ous channel types found in British Columbia. The which it flows will determine the limit of lateral Channel Assessment Procedure (B.C. Ministry of channel movement. If no imposed constraints are Forests and B.C. Ministry of Environment, Lands present, such as valley confinement,1 bridges, or and Parks 1996a, 1996b) uses aerial photographs dykes, and the valley flat is filled with erodible mat- followed by field verification, or just field studies for erial, then the channel is usually capable of eroding streams not reliably visible on photographs. Both ap- across the entire extent of its floodplain. Wherever proaches rely on obtaining data on basic channel di- valley width exceeds channel width, a potential for mensions (gradient, width, depth, and sediment size) lateral channel movement exists, although in con- to provide a systematic, repeatable, and objective fined systems in which the valley is only marginally method of channel type determination. For inter- wider, the extent of lateral movement is limited. In mediate- and smaller-sized streams (bankfull width forested valleys, the additional bank strength pro- < 20 m), the procedure identifies three morphologies vided by riparian vegetation can limit lateral channel at low flow conditions: (1) riffle-pool, (2) cascade-

1 Valley confinement refers to the degree to which a channel is deflected by the valley walls or by resistant terraces (Kellerhals et al. 1976).

338 Downstream progression Point bar deposits

Point bars

Progression and cutoffs Oxbow lake

Mainly cutoffs

Slip off slope Entrenched loop development

Terrace scarp

Irregular lateral activity Side channel or slough

Chute

Avulsion

Former channels

FIGURE 10.6 Lateral activity associated with a large channel (after B.C. Ministry of Forests and B.C. Ministry of Environment, Lands and Parks 1996a, and Kellerhals et al. 1976).

339 pool, and (3) step-pool (Table 10.1). In addition, bed material, frequently resulting in a bed surface of the three morphologies are further subdivided by finer texture. A reduction in sediment supply causes dominant bed material clast size and whether LWD is extensive riffles and bars, reduced pool volumes and functioning (i.e., influencing morphology), present, depths, and coarser bed surface. or absent in the channel. The riffle-pool morphol- At the majority of streamflows occurring in rif- ogy consists of riffle, bar, and pool units, with the fle-pool channels in a normal year, bedload does not bar representing the major storage site for sediment move; as streamflow stage rises, sediment eventually storage. In general, pools are topographically low becomes entrained, usually at or near the stream’s areas with relatively slow-moving water, and riffles bankfull discharge. Bed material is initially en- are topographically high areas with locally steeper trained from the riffle surface, then from the pool, gradients and faster-flowing water (Figure 10.7a). and as discharge approaches bankfull, sediment Riffles are accumulations of sediment that extend is transported over, or deposited on, the next riffle diagonally across the channel to the head of a bar, downstream (Pyrce and Ashmore 2003). Therefore, which extends downstream on the opposite side of sediment does not move at most flow rates, but does the channel. As bars are typically deposited on alter- move during infrequent, annually occurring high nate sides of the stream, riffles will cross the channel flows. On the Coast, bankfull discharges commonly in alternate directions, shifting from one side to the occur during the late fall and winter as a result of other as water flows downstream. On average, the heavy rainfalls and rain-on-snow events; in the Inte- distance between riffles is about two to seven times rior, these discharges occur in spring and early sum- the channel width (Leopold et al. 1964; Hogan 1986; mer as a result of snowmelt (see Chapter 4, “Regional Montgomery et al. 1995). Pools occur upstream of Hydrology”). each riffle, and are both narrower and deeper than Similar to riffle-pool morphology, cascades are the riffles at low flow. Pools are often modified by the aggregate structures (generally a series of repeat- scouring action of water flowing around obstruc- ing stone lines), but have cobble- and boulder-sized tions such as bedrock outcrops, large boulders, chan- particles, with water flowing over and around each nel bends, and often wood in forested watersheds clumped feature (Figure 10.7b). Pools located be- (Lisle 1986; Montgomery et al. 1995). tween the cascades are usually as long as the chan- Although riffle-pool morphologies are stable nel is wide and tend to be of lower-gradient. The configurations, they are not static. As sediment cascade-pool morphologies, which are considered supply is increased, channel bars expand into the partially erodible features (fully alluvial to semi-allu- centre of the channel, become less stable, and move vial), represent a transitional phase between condi- more frequently (Figure 10.4). As the bar expands, tions found in lower gradient riffle-pool channels the riffle attached to the bar expands, and the pool and the higher-gradient step-pool channels. In- extent is reduced, creating a simplified morphology creased sediment supply can result in fewer distinct with minimal depth variability. Bar expansion can pools and lead to localized bank erosion. Decreased also lead to bank erosion, which further increases sediment supply can lead to the erosional displace- sediment supply to the reach. The increased sedi- ment of the stone lines, leaving no recognizable ment supply also changes the composition of the pattern. In step-pool morphology, steps are created through the interlocking of a few large particles TABLE 10.1 Channel types and associated characteristics (usually < 10 stones) aligned across the channel (Fig- (modified from B.C. Ministry of Forests and B.C. ure 10.7c). The steps consist of diagonally arranged Ministry of Environment, Lands and Parks 1996b) stone lines in diamond- or oval-shaped cells and Morphology Sub-code Bed material LWD represent abrupt breaks in the longitudinal profile. Pools with finer-textured sediment are positioned riffle-pool RP gravel functioning g-w between steps. Steps form by the progressive move- riffle-pool RP cobble functioning c-w ment of large stones over short distances. These cascade-pool CP cobble present c-w stones eventually jam together, producing very stable cascade-pool CP boulder absent b features. Step formation also depends on the relative step-pool SP boulder present b-w Shields number (i.e., the ratio between the applied step-pool SP boulder absent b shear stress and the stress needed to mobilize the step-pool SP boulder-block absent r bed) and the ratio between bed material supply and

340 a discharge (Church and Zimmermann 2007). Large Riffle Pool Riffle Pool woody debris also contributes to step formation; if it is incorporated into the step riser, then step heights increase along with channel resistance (Curran and Wohl 2003). Once steps are established, large storm floods with recurrence intervals of 30–50 years (Grant et al. 1990) and debris flows are required to disturb them, although the actual time frame of step-pool disturbance can vary greatly (Church and Riffle-pool sequences Zimmermann 2007). Channels exhibiting step-pool morphology are partially erodible to non-erodible features (semi- or non-alluvial). Channel banks are composed of similar materials (large interlocking clasts) and bank strength is less dependent on ripar- b ian vegetation than in strictly alluvial zones. Cascade Pool Cascade Each channel type responds differently to changes in sediment supply or discharge. General responses include either vertical shifts (aggradation or degra- dation, evident by the upward or downward position of the channel bed) and (or) lateral shifts (sideways movement of the bed and banks, evident by old or abandoned channels on a floodplain). The riffle- and Step-pool sequences cascade-pool types are free to move both vertically and horizontally in the erodible deposits, but the step-pool type is usually restricted to vertical shifts within its non-erodible boundaries (except in steep Pool Step fans). The Channel Assessment Procedure considers the expected response of each channel type (Figure 10.8). See the Channel Assessment Procedure Field c Guidebook (B.C. Ministry of Forests and B.C. Min- Step Step Step Step istry of Environment, Lands and Parks 1996b) for Pool Pool Pool Pool Pool a detailed description, including photographs, field examples, and indicators of disturbance. The Channel Assessment Procedure is intended to evaluate a channel’s response to changes in the forces that shape its morphology and does not explicitly assign causes to these changes; the responses will Step-pool sequence be the same whether produced by natural or hu- man-related influences on sediment supply, riparian Pool Step vegetation, or streamflow. However, understanding both the cause and result of the response is critical to all aspects of forest management, ranging from initial planning and operational practices to restora- tion activities (see Chapter 18, “Stream, Riparian, FIGURE 10.7 Channel morphological units (B.C. Ministry of and Watershed Restoration”). Forests and B.C. Ministry of Environment, Lands Now that the driving factors determining channel and Parks 1996b): (a) riffle-pool morphology; morphology have been placed in a watershed con- (b) cascade-pool morphology (after Grant et text, we next discuss how these factors, watersheds, al. 1990); and (c) step-pool morphology (after and stream channels vary spatially and temporally Church 1992). across the province.

341 a b

FIGURE 10.8 Channel morphology matrix showing levels of disturbance (aggradation and degradation): (a) cascade-pool (CPb ,

CPc-w ) and riffle-pool (RPc-w , RPg-w ) morphologies; and (b) step-pool (SPr , SPb , SPb-w ) morphology. See Table 10.1 for morphology definitions; S = channel slope, D = largest stone moved by flowing water, Wb = bankfull channel width, d = bankfull channel depth (after B.C. Ministry of Forests and B.C. Ministry of Environment, Lands and Parks 1996b).

Streams of British Columbia

The factors governing channel morphology do ments leave a distinct imprint on its streams. The not differ geographically, and are thus considered physiography, climate, hydrology, soils, forests, and universal. Nevertheless, each factor’s relative im- other components of provincial geography are dis- portance to a specific channel does vary, as does the cussed elsewhere in this compendium (see Chapters factor’s internal attributes, which are determined by 1–4); all provide valuable background for under- local watershed characteristics. Although the impor- standing stream environments. Also important to tance of sediment supply is universal, its actual attri- stream development are other landscape features, butes—that is, whether coarse-textured sediment is particularly those strongly associated with the local delivered directly (but episodically) by landslides, or hydrology. Cheong (1996) explored specific geomor- whether fine-textured sediment is delivered continu- phic and hydrologic variables that influence channel ally from streambank erosion—have a fundamental morphology in British Columbia and identified 11 effect on channel morphology. Therefore, it is the distinctly different watershed types. Table 10.2 re- relative differences in the attributes of each factor classifies these into four types and Figure 10.9 shows that lead to the diverse nature of streams in British examples. The watershed types are differentiated on Columbia. The key to understanding these different the basis of the percent area covered with: functions lies in grasping the nature of watershed characteristics; that is, the type of watershed will • perennial snow or ice, which influences the determine factor characteristics. stream discharge regime, especially summer British Columbia’s diverse biophysical environ- flows;

342 TABLE 10.2 Watershed types reclassified and summarized according to connectivity of hillslope and channel sediments (based on 1:50 000 maps)

Watershed type (%)a Terrain attributesb Physiographic zonesc I(25) Steep, coupled (gullies, fans) Coast Mountains, Northern and Rocky Mountains, Kootenay and Columbia II(16) Steep, decoupled (gullies, floodplains) Exposed Coast, Southern Rockies, Northeast Mountains III(33) Flat, coupled (incised plateau) Northern Interior, Okanagan, Cariboo and Monashee IV(26) Flat, decoupled (floodplains) Northern Plains, Northern Interior, Exposed Coast a The percent of the total number of watersheds N( = 87). b For details of the dominant morphological setting, see Cheong (1996). c Zones taken from Cheong (1996); the first zone has the greatest proportion of the particular watershed type.

• steep lands (greater than 60% gradient), which nels are characterized by riffle-pool morphology, influence timing of discharge as well as erosion and LWD is fundamental in providing structure potential; and physical strength and influencing form (Figure • lakes or open water, which modulate stream dis- 10.10b). charge and are sediment sinks; Type II and III watersheds exist in other areas • valley flats (gradients < 7% and connected to the of the province with channel longitudinal pro- channel network), which can store both water and files reflecting the imposed conditions (Table 10.2; sediment; and Figures 10.9b and 10.9c). Longitudinal profiles are a • other landforms (extensive gully networks, fan blend of alternating convex-up and concave-up pat- complexes, terraces, etc.), which are sources or terns—profiles common in the province’s low-relief sinks for sediment. interior—with frequent hanging valleys produced by glacial erosion. Channel morphology is a function of The different watershed types are more prevalent slope (Buffington et al. 2004; Brardinoni and Hassan in certain physiographic zones; for example, type I 2007) and can be predicted on the basis of watershed occurs most frequently in the Coast Mountains and type. In addition, watershed types provide informa- type III in the Northern Interior (Table 10.2). The tion on other terrain attributes, such as the degree of watershed attributes determine the stream channel hillslope-channel coupling, which is another impor- boundary conditions. British Columbia’s topography tant factor in determining channel morphology. is often thought of as primarily steep and coastal. In As the absolute size of a particular watershed fact, large areas of the province—almost 60% of the increases, the overall shape of the composite drain- landscape—are covered with low-gradient topogra- age will depend on the arrangement of internal sub- phy. basins. Each sub-basin can be a different watershed The shape of a watershed controls the overall type and, since each watershed type has specific longitudinal profile of stream channels. Type I topographic characteristics (steep headwaters, flat watersheds have a concave-up longitudinal profile valleys, lakes, etc.) with associated hillslope-channel (Table 10.2; Figure 10.9a). In this setting, streams are coupling properties, channel morphology may vary expected to exhibit riffle-pool morphology in the greatly along the watercourse. Depending on scale, wider, lower-gradient, and finer-textured reaches the possible combinations are virtually limitless, as near the drainage outlet, and step-pool morphology is reflected in British Columbia’s extremely diverse in the narrow, higher-gradient, and coarser-textured range of stream types. reaches near the headwaters. Within a watershed, stream sediment supply has a Conversely, type IV watersheds have streams with critical influence over channel morphology; how- convex-up longitudinal profiles (Table 10.2; Figure ever, sediment delivery to, and movement within, a 10.9d). The same channel morphologies exist in this stream has only been implied thus far. A sediment watershed type, but at different positions along the budget addresses this issue and is commonly defined profile. For example, the steepest channels with as an accounting of the sources, storage, transfer, step-pool morphology are located near the drainage and fate of sediment within a watershed. (For more outlet (Figure 10.10a). The smaller headwater chan- information on constructing a formal sediment

343 a b

Plate XIb. (20) Coast Mountains. Chilcotin Ranges. Looking northwestward across Taseko River toward the abrupt front of the Chilcotin Ranges against the Fraser Plateau. Elevation of Taseko River is just below 4,500 feet. Mount Tatlow (10,058 feet) is in the left distance and Mount Waddington is the high peak on the skyline. Photo B.C. 654:35.

Plate XIIb. (22) Coast Mountains, Pacific Ranges. Looking southeast down the d glaciated valley of Tingle Creek to Stave Lake near the southern edge of the Pacific Ranges. Mountain Baker (10,778 feet), a volcanic cone in the Cascade Mountains of Washington, is in the right distance. Photo B.C. 499:82.

c

Plate XXXVIII. (71) Alberta Plateau. Looking southwest across a remnant of the Plate XLa. (74) Fort Nelson Lowland. Looking east across the Fort Nelson Lowland, elevation 1,300 upland surface of the Alberta Plateau at an elevation of 2,500 to 3,000 feet to 1,400 feet, from the junction of the Kahntah and Fontas Rivers. The large meltwater channel on between the Fort Nelson and Muskwa Rivers. Notice the scarp, which is the the left runs southwestward from Ekwan Lake. The relief on the surface is not more than 300 feet. outcrop of a flat-lying sandstone member. Photo R.C.A.F. T27R-196. Photo B.C. 1198:71.

FIGURE 10.9 Examples of watershed types from Holland (1976): (a) type I watershed; (b) type II watershed; (c) type III watershed; and (d) type IV watershed. budget, see Dietrich and Dunne 1978 and Reid and the watershed processes most important to chan- Dunne 1996.) A simplified sediment budget includes nel morphology in a particular basin. The simpli- both terrestrial and aquatic sources, and storage fied budget (Figure 10.11) illustrates three important and transfer (hillslopes and channels) components points. (Figure 10.11). Sediment is delivered to the channel through three main hillslope processes (landslides, 1. Several process types exist, with each type pro- soil creep, sheetwash) or it can be stored as colluvi- ducing different sediment amounts and textures; um along valley floors and floodplains until it is then sediment delivery to the channel depends on the transferred to the channel by colluvial and fluvial channel’s location within the watershed. In the processes (Figure 10.11; see also Table 2.1 in Chapter hillslope zones, landslides produce the greatest 2, “Physiography of British Columbia”). amounts of sediment (relative to other processes) Although quantitative sediment budgets are rare- and the textures vary by several orders of magni- ly constructed for management purposes, conceptual tude (from boulders to sand), although the mat- sediment budgets can be developed that identify erial is primarily coarse textured. Lesser amounts

344 a

b

FIGURE 10.10 Examples of channels in type IV watersheds: (a) step-pool morphology near drainage outlet; and (b) pool-riffle channel near drainage divide. (Photos: D. Hogan)

345 Landslides Soil creep Sheetwash

60 15 60 15 45 Colluvial deposits 10 Hillslope 40

Bank erosion

Streambank 15 40 5 5 10 5

tr Suspension First-order 5 channel storage Traction 5 Channel 30

Debris flows 5 Mouth of first-order basin 35 tr 25

FIGURE 10.11 Hypothetical sediment budget for a first-order basin. Processes are noted as ovals, storage elements as rectangles, and transfers as arrows; streambank and basin mouth are noted as dashed lines. Sediment transfers values are given in t/km2 per year (after Reid and Dunne 1996).

(i.e., less than a quarter of landslide production) Consequently, all aspects of a sediment budget, as of finer-textured, in-channel sediment are pro- constrained by the outlet of the drainage basin, will duced by soil creep and sheetwash and through depend on watershed type. The delivery of landslide, road-related erosion. Much of this material is soil creep, and sheetwash material from upslope stored as colluvium in fans or valley fill along the to the stream network is conditioned by watershed hillside footslopes. In the channel zone, sediment properties. In type I watersheds, where steep slopes produced through streambank erosion is stored are directly coupled to the stream network, land- for differing durations within the channel margin slides will clearly be the primary sediment source, and LWD-related storage areas. if the materials are susceptible to mass wasting (see 2. As the watershed becomes larger and the distance Chapter 8, “Hillslope Processes,” and Chapter 9, from the headwater zone becomes greater, the “Forest Management Effects on Hillslope Process- amounts of sediment produced and delivered to es”). The nature of sediments derived from landslides the channel by soil creep and sheetwash increases depends on the parent material, but the textures will relative to that delivered by landslides; the cou- generally be coarser than those derived from other pling of the hillslope and stream channel becomes input mechanisms such as from upstream reaches. less direct (the channel becomes increasingly In these cases, source mechanisms other than land- isolated from the hillslope due to the presence of slides are secondary. However, if watersheds are less a valley flat). steep and (or) hillslopes are not coupled to the chan- 3. Nearby sediment sources, such as floodplains, nel (types IV and II), then soil creep and sheetwash channel banks, and in-channel sediment storage, are relatively more prevalent. These mechanisms will increase downstream. deliver finer-textured material to the stream system

346 than those derived from mass wasting. In other to a multiple-thread channel that is either anastomo- watershed types (types III and II), the relative rates sing at lower gradients or with braids and chutes or of sediment production and delivery to the channel cascade morphologies at steeper gradients. will vary. For larger watersheds, the configuration In addition to sediment, LWD is an important and type of individual sub-basins can also influence component influencing stream morphology and is sediment budget dynamics. For example, a steep, common in many forested streams in British Co- coupled sub-basin (type I) flowing into a channel lumbia. Although characterized in various ways (see that originates in a flat, uncoupled sub-basin type( review by Hassan et al. 2005), it is most frequently IV) can strongly influence sediment dynamics at and defined as wood material 1 m or longer with a mean downstream of the confluence. diameter of greater than 0.1 m. Large woody debris Time is another aspect of sediment supply im- enters a stream section by several mechanisms, plicitly included in a sediment budget. For example, including as inputs from landslides (Figure 10.12a), landslides occur episodically and thus deliver large windthrow or blow-down (Figure 10.12b), bank ero- volumes of material infrequently over a given time sion (Figure 10.12c), tree mortality and fall (Figure period. In watersheds prone to episodic landslide 10.12d), and flotation from upstream (Figure 10.12e). inputs, the channel must constantly adjust to the The type of watershed largely controls the dominant natural rate of landslide disturbance. Soil creep and LWD input mechanism. Woody debris from land- sheetwash occur chronically and thus deliver rela- slides will predominate in steep, coupled watersheds, tively smaller volumes to the channel. This sediment whereas windthrow and bank erosion are important is much finer and will likely not alter channel mor- in lower-gradient, decoupled basins. In the Interior, phology (Figure 10.2), but it can have adverse effects tree mortality is important, especially in areas where on aquatic biota. If channel gradient is constant but vast expanses of insect-infested forest and (or) forest sediment supply is increased, then the stream will fires occur. Downstream flotation ofLWD depends change from a single-thread channel (with meanders on stream size and related scale factors, but this in low gradients and step-pools in steeper gradients) mechanism is generally more important in larger

a b c

d e

FIGURE 10.12 Large woody material input mechanisms: (a) landslides, (b) windthrow, (c) streambank erosion, (d) tree mortality, and (e) flotation from upstream. (Photos: D. Hogan)

347 channels where the typical tree height is less than the well as sediment movement and storage. Where LWD channel width (Montgomery et al. 2003). is oriented across or perpendicular to the channel, After its delivery to the channel,LWD has a range channels are wide, sediment textures are highly vari- of effects that will depend on the relative size of able spatially, and banks are undercut. Where LWD wood compared to channel dimensions and the ar- is lying parallel to the channel, channels are narrow, rangement of wood within and along the channel. beds are scoured, and banks are vertical or sloping The input mechanism and dominant tree species away from the channel. This LWD architecture is often determine the size of wood entering a stream typical in all forested watershed streams across all channel. For example, woody debris introduced forest types (Bird et al. 2004). by landslides has a range of sizes, including both Landslides and windthrow are commonly intact and broken stems, and many smaller pieces; responsible for the entry of large numbers of trees tree mortality, blowdown, and streambank erosion to the channel at a single point, although bank deliver mostly intact stems resulting in larger pieces. erosion and mortality will also deliver many trees The size of these trees varies by tree species and for- to a single location. As channel size increases and est type. becomes less connected to the hillslope, LWD is Figure 10.13 shows the range of tree heights and more commonly floated in from upstream. This bole diameters for the dominant species in all the episodic delivery of substantial amounts of LWD has biogeoclimatic ecosystem classification BEC( ) zones particular significance to the spatial and temporal in British Columbia. Larger trees affect a greater variability of stream channel conditions. Hogan et range of stream sizes than smaller trees and this al. (1998a) linked landslide frequency to the presence influence depends on biogeoclimatic factors. Chan- of logjams, showing that logjams invariably occur nel complexity and diversity are attributed to LWD where landslides in the forest enter stream channels. characteristics. When the wood is large compared Accumulations of LWD that interfere with water to channel width and is predominantly oriented flow and the transfer and storage of sediment within across the channel, it directly affects water flow as the channel are also referred to as “jams.” Jams are

LWD Input processes Landslides Windthrow Bank erosion 100

90

80

70 0 6 m Stem diameter 60

50

40

30 Maximum tree height (m) height Maximum tree 20

10

0 Tsuga Pinus Populus Picea Pseudotsuga Abies Abies Tsuga Pinus Picea Thuja Picea Pseudotsuga mertensiana contorta tremuloides glauca menziesii lasiocarpa amabilis heterophylla ponderosa engelmannii plicata sitchensis menziesii interior coastal Mountain Lodgepole Trembling White Douglas-fir Subalpine Amabilis Western Ponderosa Engelman Western Sitka Douglas-fir hemlock pine aspen spruce interior fir fir hemlock pine spruce redcedar spruce coastal

Dominant Tree Species (< 700 m) (< 500 m) Elevation snow rain Precipitation cold, dry, Northern Interior warm, wet, Southern Coast Climate

FIGURE 10.13 Large woody debris input processes and maximum tree height for dominant tree species in all the bio- geoclimatic zones in British Columbia. General patterns of LWD input processes based on preliminary data analysis and regional physiography. All sizes are approximate and for illustrative purposes only.

348 either wide and low structures in zones with free formation reverse—the upstream channel begins to lateral movement (as in streams with a channel downcut, pools develop, surface sediment textures migration zone or floodplain), or narrow and high become coarser, previously buried LWD becomes structures in zones confined by erosion-resistant exhumed, and the extent of riffles decreases. Down- materials, thereby creating areas of vertical sediment stream processes are also reversed with increased deposition in an otherwise erosional or bedrock bar development, pool formation around LWD, and reach, the “forced” alluvial reaches of Montgomery textural fining. This trend continues and channels and Buffington (1997). gradually return to the complex, diverse environ- Formation of in-channel logjams at the terminus ments that existed before the landside inputs. The of landslide run-out paths leads to many channel temporal adjustments last for about 50 years. Similar modifications. Soon after the jam forms, the chan- trends are evident in non-coastal areas, although nel both upstream and downstream of the structure documentation is not yet completed (D. Hogan, B.C. undergoes major changes (Hogan 1989). For exam- Ministry of Forests and Range, unpublished data). ple, the channel tends to fill with sediment upstream In British Columbia, debris budgets (wood input, as the intact barrier of wood interrupts the down- storage, and output) have been developed for several stream transfer of sediment. In addition, surface decades; however, many difficulties are associated gradients decrease, channel banks erode as the with the conclusive determination of input mecha- channel expands to accommodate increased storage nisms. Preliminary results suggest that input mecha- of sediment, surface sediment textures become finer, nisms vary by BEC zone, with landslide inputs more and pools in-fill and decrease in overall extent with a prevalent in the steeper coastal zones, windthrow- corresponding increase in riffles and braided zones. related inputs being more common in northern inte- Overall, the channel comes to resemble a simple run rior zones and streambank erosion more important or glide. Downstream of the logjam, the channel is in interior and northern areas (Figure 10.13). Early deprived of sediment from upstream and adjusts by results also indicate that the pattern of in-stream downcutting, which causes the loss of pools, locally storage, that is LWD predominantly stored in jams, is steeper gradients, coarse surface sediment textures, similar across the province. Little is known about the and fewer pieces of functional LWD. The combined output of woody material from the stream system; zone of influence (both upstream and downstream) this budget term is commonly deduced from the may exceed distances equivalent to 100 channel other two terms and includes the inherent uncer- widths in length (Hogan et al. 1998b). tainties of both. In addition to the spatial influence of logjams, This section considered the factors controlling temporal adjustments occur (Hogan 1989). In the channel morphology from the context of British Co- first decade after jam formation, wood begins to lumbia’s different watershed types and the processes deteriorate and the jam structure gradually becomes occurring within each. This background is neces- more open, allowing sediment and woody material sary for the following discussion of the influence of to pass around or through the jam. After a decade, forestry activities on stream conditions. the processes occurring immediately after jam

Forest Management Influences on Channel Morphology

Resource managers have been interested in the influ- fish habitat protection, and several government and ence of forestry activities on watershed conditions, industry agencies have implemented forest practices and streams in particular, for decades. Of prime regulations and guidelines to both safeguard aquatic importance is the need to protect public and worker environments and fish habitat, and ensure channel safety and the environment. Safety issues are usually integrity. related to landslides, but also involve stream cross- A great deal of research has dealt with manage- ings and roadways on floodplains. Terrain specialists ment and channel morphology (see Fish–Forestry deal with many of these issues, but considerations Interaction Program [FFIP] references at end of regarding structure location and flood mitigation chapter). Here, we restrict our discussion to the remain important. The federal Fisheries Act regulates influences of management practices on the factors

349 controlling channel morphology (Figure 10.1). Case nel and riparian zone character over both time and studies illustrate many of the points included in Ta- space. The role of forests, in terms of their natural ble 10.3, which lists a ranking of these factors accord- disturbance patterns and susceptibility to manage- ing to the potential of forest management activities ment modifications, is a fundamental key to under- to alter a channel. The factor ranking is based on our standing channel morphology in British Columbia. assessment of forest management impact causing Another coastal case study, from Carnation Creek stream channel change, although its validity is the on Vancouver Island, provides finer temporal and subject of an ongoing debate amongst geomorpholo- spatial resolution on the nature of channel altera- gists and hydrologists. This debate centres around tions documented on Haida Gwaii. Carnation Creek the relative importance of management activities is part of a comprehensive fish–forestry interaction on the dependent factors (sediment supply, riparian program, and channel monitoring began in 1971 (see vegetation, and streamflow changes), and not the www.for.gov.bc.ca/hre/ffip/CarnationCrk.htm). Ex- independent, or geologically imposed, factors. tensive upslope logging and landslides have occurred within this watershed, and the channel experienced Case Studies riparian logging that has affected streambank stabil- ity and LWD supply. The five case studies reviewed here are drawn from The Donna Creek case study considers channel diverse physiographic and biogeoclimatic settings. response to a single, large landslide event in north- The focus is on watersheds, with an emphasis on central British Columbia, and addresses several relatively low-gradient, riffle-pool streams. These geomorphic and anthropogenic factors. The water- are generally the most sensitive to the effects of for- shed has undergone specific land use changes (for est management and have been the subject of B.C. details, see Chapter 9, “Forest Management Effects Forest Service research efforts over the past three on Hillslope Processes”), and logging activities have decades. influenced runoff patterns, sediment delivery to the The first case study concentrates on a suite of -ex channel, and riparian zone disturbance. The Fubar amples from Haida Gwaii (Queen Charlotte Islands). Creek case study summarizes a monitoring program The first example provides an overview of the range that was designed to document channel adjustment of different channel types found within a single old- and recovery of small, interior streams following growth forested watershed. The next three examples riparian logging and disturbance to the channel illustrate the importance of LWD in modifying chan- banks. The Yakoun River case study illustrates the

TABLE 10.3 Factors governing channel morphology according to the potential of forest management activities to influence channel conditions (in descending order of influence)

Factor 1. Sediment supply - source (size; timing) a. Landslides (coarse sediment and LWD; episodic) b. Roads – function of road use/maintenance (fine sediment; chronic) c. Gullies (mixed sediment and LWD; episodic and chronic) d. Fans (fine sediment; episodic) 2. Riparian a. Streamside vegetation removal b. Streambank disturbance 3. Land use a. Road drainage b. Road crossings (culverts and bridges) c. Road protection (e.g., riprap) d. Timber harvesting 4. Streamflow a. Peak flows b. Low flows c. Direct channel capture 5. Climate a. Streamflows b. Indirect land use effects

350 importance of watershed type on sediment produc- watershed. Figure 10.15 illustrates the morphological tion and delivery to a larger mainstem channel and characteristics of a typical steep-coupled watershed the spatial response of the larger channel. It also (Government Creek) in an old-growth forest in the highlights the importance of historical management Coastal Western Hemlock (CWH) biogeoclimatic strategies and practices, a factor that is often misi- zone. The channel is diverse, with complex longitu- dentified or simply overlooked when considering dinal and planimetric forms (Figure 10.15a). The long- forest management practices and stream channel itudinal profile has riffles and distinct, well-defined integrity. pools, which account for 65% of the overall chan- nel area. The channel width is variable, alternating Government Creek Government Creek is an between narrow sections with stable banks and wide intermediate-sized (17 km2), old-growth, coastal sections where the channel becomes locally unstable. watershed on the northern tip of Moresby Island The channel banks are commonly undercut and bars in the Haida Gwaii archipelago (seen from aloft in consist of cobble, gravel, and sand-textured sedi- Figure 10.14a). The watershed has steep hillslopes ment. The complexity and diversity of the channel in the headwater zones and lower-gradient terrain is attributed to LWD characteristics (Figure 10.15a). closer to the stream’s mouth (type I watershed). The The wood is large compared to channel width and is channel network has a single gravel-bed stream at its predominantly diagonally oriented across the chan- outlet; a series of smaller tributaries join the main- nel, directly influencing water flow and sediment stem upstream and away from the stream mouth. movement and storage (Figure 10.15b). The main sediment source for Government Creek is Another Haida Gwaii example (Mosquito Creek) landslides that deliver sediment directly to the chan- illustrates the influence of riparian logging practices nel network. Streamflow, channel width and depth, on channels not affected by landslides (see exam- and sediment storage all increase as the watershed ple in Chapter 9, “Forest Management Effects on area becomes larger, and channel gradient and bed Hillslope Processes”). Besides logging—57% of the sediment textures decrease. watershed was logged during the 1940s and 1960s Figure 10.14b shows the channel morphology near by skidder and high-lead methods without ripar- the stream’s outlet. Both the channel and near-bank ian leave strips—Mosquito Creek (Figure 10.16) is sediments are erodible (alluvial), and the morphol- similar in most biogeophysical aspects to Govern- ogy is characterized by riffle-pools and associated ment Creek. The two creeks are similar in watershed bars. Continuing upstream along the mainstem area, drainage density and shape, average channel channel and entering the first main tributary (Figure and hillslope gradient (including uncoupled sec- 10.14c), the channel gradient and bed material size tions), geology, forest type (CWH), and climate. The increases and channel width decreases along the post-harvest channel of Mosquito Creek is relatively tributary. The morphology changes from riffle-pool morphologically simple, with minimal variability to cascade-pool (Figure 10.14d). In zones near the in longitudinal, planimetric, and sedimentologic headwaters, the channel is characterized by step- characteristics. Its longitudinal profile shows long pool morphology (Figure 10.14e). pools with relatively uniform depths. Although The Government Creek watershed includes all riffles and glides are more prevalent in the logged three of the commonly found stream morphologi- Mosquito Creek channel (Figure 10.16a) compared cal types. This is typical and expected if the entire to the forested Government Creek channel (Figure length of channel (from the stream mouth to the 10.15a), their shapes are commonly long and shallow. basin’s headwater) is considered, although the stream Channel width is not only wider than expected for order along the channel length may vary according the drainage size and hydrological conditions, it is to watershed conditions. constantly wide with minimal variability. Channel banks are rarely undercut and most channel bars Channel structure and large woody debris To consist of uniformly textured gravels. consider the role of LWD in controlling channel TheLWD characteristics associated with each structure and to provide background from which to channel explain the underlying differences in the consider forestry activities later in this compendium, channel morphologies of the logged and forested this Haida Gwaii example compares an unlogged, watershed streams. The most important difference old-growth forest stream with a stream in a logged is a shift inLWD orientation—significantly more

351 a b

c d

e

FIGURE 10.14 A typical coastal basin, Government Creek, Haida Gwaii: (a) view looking upstream to watershed at the mouth of Government Creek; (b) view looking downstream at the channel bed near the stream outlet (note riffle-pool morphology); (c) view looking upstream at the confluence of two tributaries; (d) view looking upstream at one of the tributary streams (note cascade-pool morphology); and (e) view looking upstream near headwaters (note step-pool morphology). (Photos: D. Hogan)

LWD is oriented parallel to the channel banks of the therefore the same amount of woody material has logged channel, whereas the forested stream shows less influence on scouring and trapping of sedi- a predominantly diagonal arrangement of LWD. The ment in the logged stream (Hogan 1986). Although shift in orientation reduces the interaction among the total volume of LWD is similar in each stream, LWD, streamflow, and sediment transport, and the size distribution also shows a shift, with more

352 Top of bank a Large organic debris Bottom of channel bank Stream backwater P R Pool and riffle divisions Undercut bank Deposition boundary

LWD Buried LWD

LWD cluster

Morphology 0 10 30 metres N 5 20 50

Reach profile

) 12.0

11.0

10.0 y elevation (m 9.0 20 40 60 80 100 120 140 160 180 200 220 240 260 Arbitrar Horizontal distance (m)

FIGURE 10.15 Morphological characteristics of an old-growth b coastal stream, Government Creek, Haida Gwaii: (a) large woody debris location map, planimetric map, and longitudinal profile (Hogan 1986); and (b) photograph from site. (Photo: D. Hogan)

353 Top of bank Large organic debris Bottom of channel bank Stream backwater P R Pool and riffle divisions Undercut bank Deposition boundary

LWD Buried LWD

0 5 10 20 30 50 LWD cluster metres N Morphology

Reach profile ) 12.0

11.0

10.0 y elevation (m 9.0

20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 Arbitrar Horizontal distance (m)

FIGURE 10.16 Morphological characteristics of a logged coastal stream, Mosquito Creek, Haida Gwaii; large woody debris location map, planimetric map, and longitudinal profile (adapted from Hogan 1986). small material evident in the logged stream. The shaped pools and riffles. The forested watershed smaller absolute size of the pieces, and the relative stream is geomorphically complex with diverse size reduction due to the wider channel, makes this features. The differences are attributed to the loss material more mobile at similar flow stages than that of bank strength, which causes channel widening, in the forested stream. The increased mobility leads which can dramatically increase sediment sup- to a reduction in overall channel stability. ply and lead to a decrease in channel complexity if The preceding example serves to illustrate the transport capacity is exceeded. Additionally, the re- differences between two essentially identical streams moval of the riparian vegetation as a source of LWD (with the only exception that one was logged). The input to the stream reduces the channel’s ability to two streams flow through watersheds of similar type store additional sediment and lengthens the chan- and both have steep headwaters with the down- nel’s recovery time. stream zones uncoupled from hillslope processes. The main difference between the channels is related Channel structure and natural disturbance (large to the removal of riparian vegetation and the direct woody debris jams) Much of Haida Gwaii has steep, physical disturbance of streambanks, and not land- unstable terrain and a wet climate with high-in- slides that may have occurred in distal, uncoupled tensity rainstorms. The most prevalent watershed areas of the watershed. Removal of riparian vegeta- type is steep and coupled to the channel. Landslides tion and direct disturbance of streambanks resulted are common in this coastal setting, occurring both in very different channel conditions in each stream. episodically (as infrequently occurring, large-mag- The logged watershed stream is relatively simple nitude events) and on a more frequent basis (as geomorphically, with long, shallow, and uniformly annually occurring, small-magnitude events) (see

354 Chapter 9, “Forest Management Effects on Hills- newly formed jam) coincide with jam formation, lope Processes”). Furthermore, Schwab (1983, 1998) but normal channel processes, such as rhythmic and many others have documented an increase in channel scour and fill, preferential flow paths landslide occurrence as a result of certain forestry (single rather than multi-branches), pool forma- activities (see additional FFIP references). Hogan et tion, bar development, and riparian bank reveg- al. (1998b) summarized the influence of landslides on etation may recover over timespans approaching channel conditions. For channels flowing through 50 years. old-growth watersheds it was found that: This sequence produces a complex and diverse • landslides occur episodically (Figure 10.17), with channel due to the mosaic of channel states, with an the largest generally attributed to a combination approximately equal frequency of recently disturbed of geological and meteorological factors; (major alterations) to old, recovered, and essentially • LWD jams form along the stream channel at or non-disturbed conditions developing over time, all near where landslide materials are deposited into within the same stream. However, along with the the channel; increase in landslide rates after logging, there is a • LWD jams and channel conditions evolve over corresponding increase in the number of recently time as the jams’ influence on sediment transport formed in-channel LWD jams (Figure 10.19). This and storage patterns change (Figure 10.18). Major leads to a second peak of the bimodal logjam age disturbances in the channel (e.g., bed aggrada- distribution. This shift inLWD jam age distribution tion, multi-branched flows, channel widening causes the channel in logged watersheds to have with streambank erosion, infilled pools up- relatively greater channel lengths in a disturbed state stream, and severely scoured channels eliminat- with simplified, less complex channel morphologies ing morphological features downstream of the than the unlogged streams. The reasons for acceler-

a 0.35

0.30 n = 970 landslides 0.25 0.20 0.15 0.10 Percent volume Percent 0.05 0 1800 1850 1865 1870 1875 1886 1887 1891 1905 1908 1917 1925 1935 1940 1942 1945 1957 1961 1962 1964 1972 1974 1975 1976 1977 1978 1979 1980 1982 1983 1984 1985 1987 1988 1989 1990 1991 Repeat 1980–85 Unknow n 1810/1830 b 1891/1917

120 100 80 60 40 20 0

to station record) station to -20 -40 Total precipitation (annual precipitation Total estimate mm/24 hr relative mm/24 hr relative estimate -60 1887 1890 1893 1896 1899 1902 1905 1908 1911 1914 1917 1920 1923 1926 1929 1932 1935 1938 1941 1944 1947 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989

FIGURE 10.17 Historical landslide and precipitation records: (a) landslide events occurring on Haida Gwaii (Queen Charlotte Islands) 1810–1991 (from Schwab 1998); (b) annual maximum 24-hr precipitation records for selected stations (aggregate record for: Port Simpson, 1887–1909; Masset, 1910–1914; Queen Charlotte City, 1915–1948; Sandspit, 1949–1962; Tasu, 1963–1972; Sewell Inlet, 1973–1989) (after Hogan et al. 1998b).

355 Upstream and downstream of LWD jam (a) Undisturbed • complex, diverse channel morphology • high width, depth, and sediment texture variability • pools more extensive than riffles •LWD diagonal to channel • abundant undercut banks • many small LWD steps

(b) Less than 10 years since LWD jam formation Upstream Downstream • braided channel • single thread channel • fine textured sediment • coarse-textured sediment • riffles and glides, few pools • riffles, few pools •LWD in jam with large • LWD parallel to channel volume of sediment • mainly overhanging banks stored upstream • minimal undercut banks

(c) 10–20 years since LWD jam formation Upstream Downstream • reduced number of • one main channel channels • bar development • increased sinuosity (mid-channel) • fine sediment removal • finer sediment and bed coarsening texture • pools associated with • pools associated with LWD LWD • steeper channel gradient

(d) 20–30 years since LWD jam formation

Upstream Downstream • 1 or 2 main channels • 1 or 2 main channels • bed sediment coarser • finer-textured sediment • pools more extensive • pools more extensive • riffles less extensive • riffles less extensive • steeper gradient

(e) 30–50 years since LWD jam formation Upstream and downstream of LWD jam • downcutting continues • high width, depth, and sediment texture variability • pools more extensive than riffles • LWD diagonal to flow • abundant undercut banks

(f) LWD jam formation longer than 50 years ago

Upstream and downstream of LWD jam • side channels • complex, diverse morphology • similar to undisturbed

FIGURE 10.18 Adjustment of channel morphology in response to large woody debris (LWD) jam formation and deterioration (modified from Hogan et al. 1998b). ated landslide rates in logged areas are discussed province, with the following important differences: elsewhere (see Chapter 9, “Forest Management Ef- • material input remains episodic, but may be from fects on Hillslope Processes,” and the additional other mechanisms besides landslides, such as FFIP references). windthrow, particularly in areas affected by fire, Although the connection between landslides, insect infestations, and flood-induced bank ero- material input, LWD jam formation, and channel sion (large snowpack or ice melt); evolution has been documented mainly in coastal • some areas produce smaller trees that are less able watersheds, evidence suggests that similar processes to form jam complexes, which are highly im- and conditions exist in the province’s non-coastal, movable and impermeable barriers to sediment steep, coupled watersheds. The in-channel adjust- transfer (Figure 10.14); and ments to LWD jam formation are similar across the • LWD jam deterioration processes (decomposition,

356 Logjams/Wb

0.12 Forested n = 238 jams Total no. of jams: 620 Logged n = 382 jams Length of survey: 1978 0.10 Forested 1193 W b Logged 1547 W b 0.08 Total 2740 W b 1974

0.06 1891

0.04

1917 1964 0.02

0.00 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 Logjam year of formation

FIGURE 10.19 Large woody debris jam age distributions for forested and logged watershed streams

on Haida Gwaii (Wb = bankfull width; after Hogan et al. 1998a).

physical resistance to abrasion) vary by tree spe- typical of old-growth forests (Bird 1993). The chan- cies and climate (Harmon et al. 1986). nel migrated across the valley flat, cutting channels and building bars, islands, and the floodplain; this Natural disturbance and riparian development sequence of erosion–deposition cycles over centuries Natural disturbances are important determinants produces the old-growth riparian mosaic. The older, of channel morphology, but these disturbances also abandoned channels can convey waters and provide influence riparian stand composition. Figure 10.20 refugia for aquatic biota during flood events. illustrates the interaction between the stream chan- Bird’s (1993) work highlights important opera- nel and the riparian zone for an intermediate-sized tional implications. It shows a link between hill- coastal stream (Gregory Creek; from Bird 1993). slopes and riparian zones that is as important as the The major storm events of 1891, 1917, and 1978 (see link between stream channels and riparian zones. Figure 10.17) induced landslides that introduced large Traditionally, riparian zone management (Chapter quantities of wood into the channel and created log- 15, “Riparian Management and Effects on Func- jams. Subsequent flooding, following these events, tion”) has been based on the maintenance of exist- forced streamflows around the logjams and into the ing streamside vegetation to protect riparian zones floodplain, leading to channel avulsions into the ri- as sources of LWD to streams (e.g., buffer strips); parian zone (Figure 10.20). With the jams no longer however, riparian zones are also controlled by, and within the active channel, sediment wedges that had depend on, hillslope processes. Landslides initiate accumulated upstream were rapidly colonized by the formation of instream logjams that then cre- riparian species. The temporal dynamics (storms→ ate a local base level in the channel, which disrupts landslides→logjams→avulsions) eventually led to a sediment transport and initiates the aggradation of a mosaic of diverse and complex riparian stand ages, sediment wedge. Opportunistic, pioneering riparian ranging from 12 years to over 300 years, which is plant species colonize the infrequently flooded por-

357 tions of the sediment wedge, increasing the stability of the stored sediment. Consequently, the formation of a sediment wedge influences the distribution of riparian vegetation that, in turn, affects the stability of the stream channel. Large floods force water and sediment loads around logjams and into the riparian zone, creating patches of riparian vegetation. His- torically, only the influence of the riparian zone on the channel has been of concern from a management perspective; however, Bird’s (1993) work indicates that the channel can influence the composition of the riparian zone, and therefore riparian zone man- agement should consider hillslope processes as well as stream channels. Riparian zone management has traditionally considered the physical and biological attributes of channels that are not expected to move laterally (i.e., lateral movement is not considered); however, as Bird (1993) has shown, laterally unconfined channels are capable of moving across entire floodplains within a single forest stand rotation. The ability of channels to move laterally depends on both the erodibility of valley-flat material and the relative width of the valley to the channel. Designated riparian manage- ment zones that do not account for potential channel movement can be rendered ineffective by subsequent channel migration. Prudent riparian zone manage- ment should consider the potential for lateral chan- nel movement when delineating riparian buffers.

Carnation Creek Carnation Creek is a relatively small watershed, draining 11.2 km2 on the west coast of Vancouver Island. It has a warm, wet climate; annual precipitation ranges from 2100 to 4800 mm, with 95% falling as rain between November and April. Stream discharges range from 0.025 m3/s in the summer to over 60 m3/s during winter freshets. The watershed consists of two steep-coupled basins FIGURE 10.20 Pathways of fluvial disturbance in a riparian area (after Bird 1993). The arrows identify two that are linked longitudinally. This creates an upper events occurring in 1891 and 1917 (indicated basin, with steep headwaters and a flat, bowl-shaped by Spruce-Alder and the Alder-Spruce patches, section downstream, that leads into a steep-sided respectively), when the channel avulsed into canyon and is then connected to a lower basin, with the riparian area. Logjams C and K formed a broad, valley flat, which extends for approximately during these events and are now abandoned 1 km before the stream enters the ocean (Figure by the channel. A third event in 1978, 10.21). indicated by Alder patches, was responsible Two phases of logging were undertaken. The for the creation of several islands. The riparian first phase occurred from 1976 to 1981, and involved area occupied by Spruce-Hemlock patches has riparian logging; the second phase occurred from been undisturbed for at least 300 years (after 1987 to 1994 in the creek’s headwaters. Three ripar- Hogan et al. 1998b). ian harvesting treatments were applied at Carnation

358 Carnation Creek Channel Morphology Study Study areas and past cutblock locations

■ ² " Study areas

Roads

Streams

Cutblocks

Elevation 20-m contours IX "■

VIII VII "■ VI "■ IV V "■ "■ "■ III "■ II I "■ "■

0 0.5 1 2 kilometres

FIGURE 10.21 Carnation Creek watershed, showing study areas and cutblock locations (A. Zimmermann).

Creek, but only one “careful” riparian treatment In January 1984, a large storm event resulted in (logged to the streambank, but with no direct in- multiple gully failures that delivered sediment and stream activities) is discussed here. wood into a steep, confined canyon segment of the Figure 10.22a illustrates the pre-logging (1977) creek. This material was transferred and deposited morphology of the study area. At this time, the downstream of the canyon into the study area, where channel exhibited a complex morphology as a result further downstream transport was prevented by the of individual LWD pieces and an intact riparian zone. logjam (Figure 10.22c). This resulted in significant Figure 10.23 shows the bankfull widths of four study aggradation upstream of the jam and started the cy- area cross-sections. cle of jam–channel interaction. All of the cross-sec- This confirms that little change in cross-sectional tions experienced significant channel widening due width occurred between 1971 and 1980. During 1978 to a sediment wedge, which was deposited upstream and 1979, the study area was logged, adding more in- of the jam (Figure 10.23). After 10 years, another channel LWD from slash left along the channel banks peak flow event broke through the jam and allowed upstream and within the study area. A key piece the sediment to be evacuated, returning the width spanning the channel (see Figure 10.22a) trapped this in this area to almost pre-disturbance values (Figure additional material and resulted in the formation of 10.23). a logjam. A majority of the key pieces in the channel The Carnation Creek example highlights the were a legacy from the old-growth stand. It is likely importance of the connectivity of hillslopes to chan- that as these pieces decay, the stability of the jams nels, and the transfer of disturbances to downstream in Carnation Creek will be reduced. As the logjam reaches. Riparian logging added to the scale of the grew in size and began to impede downstream sedi- disturbance by contributing additional LWD to the ment transport, the channel widened (Figure 10.22b); reach and reducing bank strength. The full temporal additional reduction in bank strength caused by extent of the logging impacts in Carnation Creek are riparian tree removal was also a factor. Most of the uncertain, since a number of legacy logs still span cross-sections upstream of the logjam (3, 6, and 9) the channel and act as key pieces for logjams; howev- experienced widening; cross-sections 12 and 18 were er, as these pieces decay, the extent and durability of within and downstream of the jam, respectively current logjams will be compromised and the extent (Figure 10.23). and durability of future jams will be limited.

359 n tio rec Di w Flo b) 1982 a) 1977 Gravel Gravel

Gravel

Gravel Gravel Gravel Gravel

Gravel Bar

Gravel Pool

Vegetation

Vegetation Pool Gravel Vegetation c) 1984

Gravel

Vegetation

Vegetation

Vegetation Legend Vegetation Multiple wood pieces d) 1991 Gravel Steep bank

Gravel

Vegetation

0 6.1 m

FIGURE 10.22 Carnation Creek morphometric maps of study area during summer low flows 1977–1991. Shaded areas indicate wetted surface at time of survey (Luzi 2000).

360 45 XS 3 XS 6 40 XS 9 XS 12 35 XS 18

30

25

20 Bankfull width (m) 15

10

5

0 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 Year

FIGURE 10.23 Bankfull widths of selected cross-sections from Carnation Creek study area; logging in study areas occurred during 1978–1979 (Luzi 2000).

Donna Creek Donna Creek, a relatively large tribu- to the Manson River by debris flood (about 7 km tary of the Manson River, is located in the Omineca downstream from the erosion scar). Mountains 75 km northwest of Mackenzie. Drain- Upstream of the erosion scar (stream input age basin area is 126 km2 and channel width near the location), the banks were stable and undercut, and outlet approaches 30 m. Fisheries values in Man- the channel bars were primarily of the diagonal and son River and its tributaries are significant and, in point bar type; these occupied less than half of the particular, provide habitat for kokanee salmon. The channel width at low flow (Figure 10.24a). Woody upper reaches of Donna Creek are generally coupled debris was abundant, with approximately 12–14 m3 to the hillslopes with only a narrow valley bottom of debris in any 100-m length of channel, and was to buffer sediment transfers from hillslopes to the important in controlling sediment storage and channel. Road-building activities along the upper movement. Woody debris steps and jams were slopes of the watershed altered the natural hillslope critical structural channel elements. Pool-riffle runoff pattern by water capture and routing along sequences were typical of small channels with road ditches, which resulted in a mass wasting event. abundant woody debris; the average pool-riffle A full analysis of the upslope processes, forest prac- spacing was 4.1–4.6 bankfull widths. tices, and resulting mass wasting event is included in Immediately below the erosion scar, most of the Chapter 9 (“Forest Management Effects on Hillslope in-channel debris was either parallel to the channel Processes”) and should be consulted when consider- or elevated above the bed. The channel experienced ing the channel impacts discussed here. extensive widening, and deep, fine sediment deposits During the spring snowmelt of 1992, the altered extended 80 m across the valley flat (Figure 10.24b). runoff patterns resulted in the delivery of an exces- Sediment depth was usually 1 m or less but depths sive amount of water (approximately seven times exceeded 2 m in some areas. Riffles and riffle-glides greater than natural conditions) to a glaciofluvial/ were the most common morphological features, and glaciolacustrine terrace. This initiated a series of pools were infrequent. debris slides and flows that transferred 420 000 m3 The physical features of Donna Creek were radi- of sediment to Donna Creek (see Schwab 2001 for cally altered as a result of the sediment and debris details). Sediment delivered to the channel was ei- introduced to the stream following the hillslope ther stored along the channel margin or transferred failure (Table 10.4). Impacts included severe channel

361 a b

FIGURE 10.24 The impact of a large landslide on Donna Creek, an intermediate-sized interior stream: (a) view of Donna Creek upstream of 1992 landslide (or washout flow, see Chapter 9) entry point; and (b) view of Donna Creek at 1992 landslide (or washout flow) entry point. (Photos: D. Hogan)

TABLE 10.4 Morphological channel conditions observed in the field during 1992 and 2007 (Donna Creek). Note that mean pool

length and pool-riffle spacing are given in bankfull width (Wb) units. Subdominant bar types are shown in parentheses (Schwab et al., unpublished data).

Bankfull Mean pool Pool-riffle Woody debris width (m) length (W ) spacing (m) Bar typea D (mm) (m3/100 m) Reach b 95 number 1992 2007 1992 2007 1992 2007 1992 2007 1992 2007 1992 2007 1 8 10 2.2 1.7 4.6 4.1 sc sc (mc) 196 126 14 12 2 17 19 3.6 - 11.8 - mc sc 163 45 71 120 3 19 29 0.88 0.53 3.0 5.0 mc (sc) sc 203 216 15 64 4 12 19 0.39 0.76 7.3 4.2 sc (mc) sc (mc) 169 179 30 38 5 11 25 - 2.1 - 3.0 sc sc 212 157 5.7 4.9 8 20 18 1.5 1.2 3.1 3.0 pb (sc) sc (pb) 201 164 2.2 8.1 a sc = side channel bar; mc = mid-channel bar; pb = point bar erosion and extensive sedimentation in upstream, processes and the potential to trigger massive geo- near-source areas, and massive sedimentation in morphic change to a river channel (both temporally downstream channel reaches. These impacts have and spatially) by overwhelming natural patterns of persisted in the channel for over 10 years (Figure sediment delivery to stream systems. 10.25); the channel has experienced alternating cycles of net sediment scour and deposition as sediment is Fubar Creek The branching network of stream reworked from temporary storage and transported channels in a watershed ensures that more small downstream. In addition, woody debris that was re- streams are encountered as timber harvesting activi- mobilized from previous accumulations stored in the ties expand from valley-bottom areas into headwater channel and (or) along the channel margin has been areas. Concerns about the effects of forest manage- reorganized and augmented with wood recruited ment activities near small streams became a seri- from the riparian area to form several new logjams. ous issue in the Prince Rupert Forest Region (now The Donna Creek landslide and subsequent chan- the Northern Interior Forest Region) in the early to nel response is an extreme example of environmental mid-1980s. In an attempt to address these concerns, damage that can result from poor forestry practices regional researchers established a program to docu- (specifically, the design and maintenance of forest ment the disturbance and recovery to pre-logging roads). The example illustrates the importance of conditions in a series of small streams, each with a localized alterations to natural hydrological specific logging history. One stream, Fubar Creek,

362 1.0 Upper reaches Lower reaches /yr) 3 m

5 0.5

0

-0.5 Change in sediment storage (x10 storage Change in sediment

-1.0 1989– 1992– 1994– 1996– 2002– 2004– 1992 1994 1996 2002 2004 2007 Period of observation

FIGURE 10.25 Erosion and deposition patterns in Donna Creek (Schwab et al., unpublished data). was selected to evaluate the effect of clearcut logging banks built into the active channel, resulting in a de- an entire cutblock, including the complete removal crease in channel width (by about 1 m between 1992 of all riparian vegetation on both banks of the and 2000). Generally, this indicates that sediment stream. An upstream section of channel (Fubar Up- mobilized in the channel during and shortly after per), biophysically similar to the downstream logged logging has been transferred downstream and (or) section (Fubar Lower), remained unlogged and con- stored overbank as the channel recovers from the stituted a control. No upslope issues (e.g., landslides) initial disturbance. The recovery process may have existed to confound the effect of riparian logging. been prolonged, however, because logging to the Fubar Creek is a small drainage located in the streambanks has affected the supply of woody debris upper Zymoetz River watershed, approximately to the channel. Although more pieces were evident 20 km west of Smithers. Its riffle-pool channel is in the channel, these were generally smaller in length generally less than 3 m in bankfull width. The reach and radius and likely less functional as impediments was logged to the streambanks in the mid-1980s to sediment transport. For example, between 1992 using conventional ground-based skidding tech- and 2000, the number of pieces increased from 41 to niques with equipment operating over and across the 51 and the mean length and diameter both decreased channel. These operations directly altered in-stream from 1.42 to 1.12 m and from 0.13 to 0.10 m, respec- channel features. tively. No statistical tests of difference were con- During logging operations, the streambanks ducted on these data (this operational work does not were disturbed and sediment was transferred to include a complete sample of the stream so the data the channel (Figure 10.26). If channel conditions in are from a single point), but the patterns are similar the upstream reach were representative of those in to those found in other streams (see above and addi- the downstream reach before logging, then channel tional FFIP references at end of this chapter). Because width doubled and sediment storage increased in the treatments practised at this site were considered the channel (note the growth of mid-channel bars). normal for the place and time, the channel differ- Channel recovery has been relatively slow, with evi- ences presented here are attributed to overall logging dence of channel disturbance persisting to 2005 (e.g., activity; no attempt has been made to attribute chan- sediment storage remained relatively high compared nel differences to the effects of individual activities to Fubar Upper). such as roads, stream crossings, yarding methods, The channel bed in Fubar Lower underwent ex- and site degradation. tensive erosion since the initial disturbance, and the This example illustrates the importance of intact

363 Fubar Upper 1989 Fubar Upper 2005

Unlogged

Fubar Lower 1989 Fubar Lower 2005

Flow

Logged

0 2 4 6 Metres

Bar

Pool

Channel bank

FIGURE 10.26 Planimetric maps for unlogged (Fubar Upper) and logged (Fubar Lower) channels in Fubar Creek 1989 and 2005 (after Bird et al. 2010).

364 streamside vegetation. Activities associated with the of Haida Gwaii was isolated until the 1960s and no removal of riparian vegetation resulted in an imme- significant industrial-scale logging had taken place. diate and dramatic change in channel conditions in A detailed discussion of this case, which illustrates this small interior stream. The channel disturbances the importance of historical anthropogenic events in have not been reversed in over a decade. Recent a coastal setting, is included in Hogan (1992). Similar regulations require the maintenance of channel examples could be cited for the interior far north, the condition if harvesting is planned in the riparian Okanagan, and the Kootenays. management area of streams (for current British Co- A common concern when developing forest plans lumbia practices, see Chapter 15, “Riparian Manage- centres around the amount or rate of logging within ment and Effects on Function”). the watershed. Hydrological considerations could be assessed in the Yakoun River system because Yakoun River The hydrological implications of streamflow gauging records were available from the extensive logging began to receive critical con- Canadian government (WSC Station 08Q002). The sideration in the 1960s and serious concerns were station, located close to the river’s mouth, has daily raised in coastal British Columbia in the 1970s. One streamflow records dating back to 1962. An evalua- approach to addressing the hydrological issues was tion of these discharge records in conjunction with to undertake watershed assessments. Several water- local climate data showed that the main change in shed assessment procedures were developed, each the hydrologic response of the Yakoun River during with its own individual strengths and weaknesses. the period of accelerated timber harvest (1962–1989) An advantage of these procedures was that hydro- was a decrease in the ratio of runoff to rainfall; logical processes operating at a watershed level, as there was also a modest increase in local precipita- opposed to site-specific or plot level, were central tion and a slight decrease in annual discharge over to each (Wilford 1987; B.C. Ministry of Forests and the period.2 This is not the expected response based B.C. Ministry of Environment, Lands and Parks on observations in most other research watersheds 1995). A common goal of forest management plans (Chapter 7, “The Effects of Forest Disturbance on was to avoid changes to stream channel morphol- Hydrologic Processes and Watershed Response”), ogy, thereby ensuring the protection of the physi- and the decrease in runoff was attributed to greater cal habitat of fish and other aquatic organisms. An evapotranspiration by the regenerating alder than integral component of watershed assessments was the previous old-growth conifer forest.3 Hydrograph the need to balance different aspects that may lead to analysis further supported the conclusion that the channel change and thus direct the focus toward the basic hydrologic response characteristics of the sys- hydrologic and geomorphic factors operating in a tem appeared unchanged. watershed. The Yakoun River case study provides an A river system’s sediment supply is also a central example of this balancing act. issue when considering forestry plans. The Yakoun The Yakoun River has a large watershed that River watershed has steep headwaters, lower-gradi- drains 477 km2 at its mouth near Port Clements ent mid-slope sections, and a major, wide and long on Haida Gwaii. This case study addresses several valley flat that includes a large lake. It has steep, valuable aspects of stream channel morphology as- connected tributaries and a flat-uncoupled lower sessments; it shows the importance of partitioning mainstem. The overall watershed was portioned into the watershed into multiple internal sub-basins, the 19 sub-basins to distinguish these different types relative weight of hydrologic and geomorphic fac- from the downstream mainstem section. A series of tors, and the consequence of disregarding historical historical aerial photographs helped identify chan- management strategies and practices when evalu- nel changes along the mainstem. These included ating channel conditions. This last aspect is often photographs from 1937 (before logging), 1961 (during neglected because it is easy to assume that if forest an early logging phase), and 1988 (late and post-log- management started in remote areas in the 1940s and ging); additional photographs from 1954 and 1979 1950s, then human-influenced events had surely not were added for selected sections. Photogrammetric occurred before that time. The Yakoun River area analyses of 23 homogeneous reaches included map-

2 Hardy BBT Limited. 1991. Analysis of changes in the discharge of Yakoun River. Consulting Engineering and Environmental Services Report (G.E. Barrett, VG-05673), March 20, 1991. Unpubl. 3 Ibid.

365 ping of channel position and the type and location of of jam development. Figure 10.17 showed that the bars, islands, and logjams. largest storms on record (last 200 years) occurred in Evaluation of the early (1937) aerial photographs 1875, 1891, 1917, 1935 and 1978. These storms undoubt- showed evidence of extensive stream channel dis- edly caused extensive bank erosion, and introduced turbance. This included reaches with indications large volumes of wood from riparian zones into of channel widening, reduced width variability, the stream system. Note that both the large lake unstable bar configuration with many mid-chan- and extensive valley flats of the Yakoun effectively nel bars (longitudinal, crescentric, and medial), decouple the hillslopes, reducing the direct input and island development. Because no logging of any from landslides, although the headwater tributar- importance had occurred, the cause of this distur- ies were most likely severely affected by landslides bance was questioned. A review of early government during the storms. Removal of the logjams seriously documents found a logjam inventory for the Yakoun affected stream channel morphology and stability by River compiled in 1904 by Ells (1906; Table 10.5). changing in-channel sediment storage patterns and This Geological Survey of Canada inventory was reach-scale flow properties. These changes, in turn, compiled to find a route up the Yakoun River from resulted in altered bar patterns with more frequent the ocean for a self-propelled drilling platform (18-m mid-channel bars and reduced bank integrity and barge) to engage in mineral exploration at tributary wider channels. confluences. The jams were characterized as preva- From the available information, the following lent, huge, old, and solid, and represented a seri- conclusions were drawn (Hogan 1992). ous constraint to river navigation. Starting in 1912, • The mainstem channel had undergone significant logjams were systematically removed from the lower changes over the last 50 years, including changes Yakoun River to overcome this impediment (Dalzell in channel width, bar structure, and logjam func- 1968). A second river-clearing program commenced tion. These changes were not evident in all areas; in 1913–1914 when all jams, including three that some reaches had undergone modifications and reportedly extended for 3 km each, were demolished others remained relatively unaltered. by dynamite, enabling access to Wilson Creek (in • Overall, the channel experienced the larg- reach 15 of the recent study). Further clearing oc- est changes during the time when only a small curred in 1958–1959, when roughly 5000 m3 of logs proportion of the Yakoun watershed was logged. and trees were yarded from the channel, largely to It seems unlikely that the amount of watershed protect a bridge. The 31 jams listed by Ells compare logged had a locally significant impact on channel with 51, 69, and 55 jams evident in the 1937, 1961, and changes. More probably, site-specific impacts have 1988 air photographs, respectively. been more important—the channel was prob- Earlier in this chapter, we discussed the im- ably stressed far more by historical removal of portance of logjam influence on streamflow and in-stream LWD through mining-related activities sediment transport, particularly in the early stages (1912–1914) than by sediment derived through spa- tially diffuse sources as a result of contemporary logging activities. Summary of logjam characteristics on the Yakoun TABLE 10.5 • Several factors, besides logging activity, have been River (after Ells 1906) identified as playing important contributing roles Distance upstream Logjam in the documented river behaviour. These include of mouth (mi) inventory the direct impact of early mining activities on channel form and processes, and natural sequenc- 2 – 4 10 jams, logs with diameters > 1 yard es of large, low-frequency storms. • Localized instances appear to contribute to chan- 4 – 7 Few jams nel erosion (e.g., poor road layout and disturbed 7 – 10 10 jams, log diameters up to 60 streambanks). inches 10 – 16 Rare jams The Yakoun River case study highlights the poten- 16 – 19 10 jams, very poor navigational tial range of factors to be considered when develop- potential ing forest plans for an area, including watershed and 19 – 23 Except 1 large jam, clear upstream sub-basin characteristics, location of development to lake within the watershed, natural disturbance regimes,

366 and historic anthropogenic disturbances. These adjust to excess sediment following disturbances can examples show that each watershed responds differ- range from decades in this and other studies (Madej ently to sediment inputs, whether from rapid (acute, and Ozaki 1996), to centuries (Knighton 1989). As episodic) events, or gradual and continuous (chron- Schumm (1977) emphasized, streams are physical ic) processes. The time required for fluvial systems to systems with a history.

SUMMARY

This chapter emphasizes the overarching principles channel, and placed the watershed in relative context that control stream channel geomorphology, in light compared to other processes. of resource management practices and planning. Our case studies illustrate the extent to which We have not attempted to provide a complete and land use management affects a watershed and its comprehensive discussion of fluvial geomorphology channel system. The five case studies dealt with -im or to prescribe best management practices for forest pacts that are both spatially and temporally exten- operations. We have addressed the fluvial geomor- sive; these impacts can be severe or more subdued, phology problems of British Columbia’s forest lands but all contain important knowledge to consider through case studies from various geographic areas with respect to forest management planning. By of the province, with the intention of dealing with considering the factors governing channel morphol- applied problems encountered during management ogy, it is possible to manage land use practices and activities over the last two decades. We hope this of- simultaneously minimize deleterious impacts to fers a foundation upon which decisions can be made stream channels. to minimize channel impacts. Consideration of the processes that operate at We have presented the watershed as the funda- a watershed level in British Columbia led to the mental landscape unit because its composition deter- development of the Channel Assessment Proce- mines many factors that govern channel conditions dure (B.C. Ministry of Forests and B.C. Ministry of and how watersheds respond to active management. Environment, Lands and Parks 1996a and 1996b). Watershed shape is a function of geological proc- This procedure embodies many of the concepts esses including both independent and dependent considered in this chapter and deals with a channel’s factors. These, in turn, strongly influence the nature expected adjustments to changes in materials supply, of the channel’s sediment supply. The reworking of delivery mechanisms, bank integrity, and stream this sediment along the various sections of a water- flow. The procedure endeavours to summarize much shed produces different channel types. In covering of what we have presented in this chapter, attempting these topics, we have detailed the importance of the to resolve past watershed impacts and to avoid future watershed as it influences the temporal and spatial problems. nature of sediment and other materials entering the acknowledgements

This chapter has benefited from the contributions of this chapter, and six anonymous reviewers provided several individuals. Steve Bird assisted with the case thoughtful and constructive comments. All contrib- study sections, Brett Eaton and Michael Church pro- uted comprehensive suggestions that led to major vided very helpful suggestions that greatly improved revisions to this chapter.

367 REFERENCES

B.C. Ministry of Forests and B.C. Ministry of Envi- Buffington, J.M., D.R. Montgomery, and H.M. ronment. 1995. Coastal watershed assessment Greenberg. 2004. Basin-scale availability of procedure guidebook (CWAP), level 1 analysis. salmonid spawning gravel as influenced by B.C. Min. For., For. Pract. Br., Victoria, B.C. channel type and hydraulic roughness in For. Pract. Code B.C. Guideb. www.for.gov. mountain catchments. Can. J. Fish. Aquat. Sci. bc.ca/tasb/legsregs/FPC/fpcguide/COASTAL/ 61(11):2085–2096. CWAPTOC.HTM (Accessed March 2010). Buffington, J.M., R.D. Woodsmith, D.B. Booth, and ______. 1996a. Channel assessment procedure D.R. Montgomery. 2003. Fluvial processes in guidebook. Victoria, B.C. For. Pract. Code Puget Sound Rivers and the Pacific Northwest. B.C. Guideb. www.for.gov.bc.ca/tasb/legsregs/ In: Restoration of Puget Sound rivers. D.R. fpc/fpcguide/channel/chan-toc.htm (Accessed Montgomery, S. Bolton, D.B. Booth and L. Wall March 2010). (editors). Univ. Wash. Press, Seattle, Wash., pp. 46–78. ______. 1996b. Channel assessment procedure field guidebook. Victoria, B.C. For. Pract. Code B.C. Calow, P. and G.E. Petts (editors). 1992. The rivers Guideb. www.for.gov.bc.ca/tasb/legsregs/fpc/ handbook. Blackwell Scientific, Oxford, U.K. fpcguide/CHANFLD/CFLD-TOC.HTM (Ac- Cheong, A.L. 1996. Classifying and comparing cessed March 2010). drainage basins in British Columbia. B.C. Min. Bennett, S.J. and A. Simon (editors). 2004. Riparian Environ., Lands Parks, Victoria, B.C. Water- vegetation and fluvial geomorphology. Am. shed Restoration Guideb. Geophys. Union, Washington, D.C. Water Sci. Church, M. 1992. Channel morphology and typol- Appl. Ser. Vol. 8. ogy. In: The rivers handbook: hydrological and Bird, S.A. 1993. Stream channel and riparian zone re- ecological principles. P. Calow and G. Petts sponse to the development of a lateral sediment (editors). Blackwell Scientific Publications, wedge in the Queen Charlotte Islands, B.C. Oxford, U.K. In: The rivers handbook. P. Calow MSc thesis. Univ. Western Ontario, London, and G.E. Petts (editors). Blackwell Scientific, Ont. Oxford, U.K. Vol. 1, pp. 126–143. Bird, S.A., D.L. Blocka, S.J. Barker, and D.L. Hogan. ______. 2006. Bed material transport and the mor- 2004. Stream channels, large woody debris and phology of alluvial river channels. Annu. Rev. biogeoclimatic zones in managed watersheds: Earth Planet. Sci. 34:325–354. technical report. B.C. Min. For., Res. Br., Van- Church, M. and D. Jones. 1982. Channel bars in couver, B.C. FII Project R04-048. gravel-bed rivers. In: Gravel-bed rivers. R.D. Bird, S., D. Hogan, and J. Schwab. 2010. Photogram- Hey, J.C. Bathurst, and C.R. Thorne (editors). metric monitoring of small streams under a John Wiley and Sons, Chichester, U.K. Wiley, riparian forest canopy. Earth Surf. Process. Chichester, U.K., pp. 291–324. Land. 35, DOI:10.1002/esp.2001. Church, M. and O. Slaymaker. 1989. Disequilibrium Blench, T. 1957. Regime behaviour of canals and of Holocene sediment yield in glaciated British rivers. Butterworths Scientific Publications, Columbia. Nature 337:452–454. London, U.K. Church, M. and A. Zimmermann. 2007. Form Brardinoni, F. and M.A. Hassan. 2007. Glacially- and stability of step-pool channels: research induced organization of channel-reach mor- progress. Water Resour. Res. 43, W03415, phology in mountain streams. J. Geophys. Res. doi:10.1029/2006WR005037. Earth Surf. 112:F03013. Curran, J.H. and E.E. Wohl. 2003. Large woody de- Bridge, J.S. 2003. Rivers and floodplains: forms, bris and flow resistance in step-pool channels, processes, and sedimentary record. Blackwell Cascade Range, Washington. Geomorphology Publishing, Malden, Mass. 51:41–157.

368 Dalzell, K.E. 1968. The Queen Charlotte Islands years of coast research to management solu- 1774–1966. C.M. Adams, Terrace, B.C. tions. D.L. Hogan, P.J. Tschaplinski, and S. Chatwin (editors). B.C. Min. For., Res. Br., Dietrich, W.E. and T. Dunne. 1978. Sediment budget Victoria, B.C., Land Manag. Handb. No. 41, pp. for a small catchment in mountainous terrain. 189–200. www.for.gov.bc.ca/hfd/pubs/Docs/ Zeitschrift für Geomorphologie, Supplemental Lmh/Lmh41.htm (Accessed March 2010). Band 29:191–206. Hogan, D.L., S.A. Bird, and M.A. Hassan. 1998a. Ells, R.W. 1906. Report of Graham Island, B.C. Geol. Spatial and temporal evolution of small coastal Surv. Can., Ottawa, Ont. Annu. Rep., New Ser., gravel-bed streams: influence of forest manage- Vol. 16, No. 940, Pt. B. ment on channel morphology and fish habitats. Grant, G.E., F.J. Swanson, and M.G. Wolman. 1990. In: Gravel-bed rivers in the environment. P.C. Patterns and origin of stepped-bed morpholo- Klingeman, R.L. Beschta, P.D. Komar, and J.B. gy in high gradient streams, Western Cascades, Bradley (editors). Water Resources Publica- Oregon. Geol. Soc. Am. Bull. 102:340–352. tions, Highlands Ranch, Colo., pp. 365–392. Harmon, M.E., J.F. Franklin, F.J. Swanson, P. Sollins, Hogan, D.L., S.A. Bird, and S. Rice. 1998b. Stream S.V. Gregory, J.D. Lattin, N.H. Anderson, S.P. channel morphology and recovery processes. Cline, N.G. Aumen, J.R. Sedell, G.W. Lien- In: Carnation Creek and Queen Charlotte kaemper, K. Cromack, Jr., and K.W. Cummins. Island Fish/Forestry Workshop 1994: Queen 1986. Ecology of coarse woody debris in tem- Charlotte City, B.C. D.L. Hogan, P.J. Tschap- perate ecosystems. Adv. Ecol. Res. 15:133–302. linski, and S. Chatwin (editors). B.C. Min. For., Res. Br., Victoria, B.C. Land Manag. Handb. Hassan, M.A., D.L. Hogan, S.A. Bird, C.L. May, T. No. 41, pp. 77–96. www.for.gov.bc.ca/hfd/pubs/ Gomi, and D. Campbell. 2005. Spatial and tem- Docs/Lmh/Lmh41.htm (Accessed March 2010). poral dynamics of wood in headwater streams of the Pacific Northwest. J. Am. Water Resour. Holland, S.S. 1976. Landforms of British Columbia: a Assoc. 41(4):899–919 physiographic outline. B.C. Dep. Mines Petrol. Resour., Victoria, B.C. Bull. No. 48. Hogan, D.L. 1986. Channel morphology of un- logged, logged and debris torrented streams in Kellerhals, R., M. Church, and D.I. Bray. 1976. Clas- the Queen Charlotte Islands. B.C. Min. For., sification and analysis of river processes. J. Victoria, B.C. Land Manag. Rep. No. 49. www. Hydraul. Div., Am. Soc. Civil Eng. 102:813–829. for.gov.bc.ca/hfd/pubs/Docs/Mr/Lmr049.htm Knighton, A.D. 1989. River adjustment to changes in (Accessed March 2010). sediment load: the effects of tin mining on the ______. 1989. Channel response to mass wasting in Ringarooma River, Tasmania, 1875–1984. Earth the Queen Charlotte Islands, British Colum- Surf. Process. Land. 14:333–359. bia: Temporal and spatial changes in stream ______. 1998. Fluvial forms and processes: a new morphology. In: Proc. Watersheds ‘89: confer- perspective. Arnold, London, U.K. ence on the stewardship of soil, air and water resources, Juneau, Alaska, March 21–23, 1989. Lane, E.W. 1955. The importance of fluvial morphol- U.S. Dep. Agric., For. Serv., Alaska Reg., R10- ogy in hydraulic engineering. Proc. Am. Soc. MB-77, pp. 125–144. Civil Eng. 81:745:1–17. ______. 1992. Yakoun River: rate-of-cut evaluation. Leopold, L.B., M.G. Wolman, and J.P. Miller. 1964. B.C. Min. For., Vancouver For. Reg., Vancou- Fluvial processes in geomorphology. W.H. ver, B.C. Freeman and Company, San Francisco, Calif. Hogan, D.L. and S.A. Bird. 1998. Classification and Lisle, T.E. 1986. Stabilization of a gravel channel by assessment of small coastal stream channels. large streamside obstructions and bedrock In: Carnation Creek and Queen Charlotte bends, Jacoby Creek, northwestern California. Islands Fish/Forestry Workshop: applying 20 Geol. Soc. Am. Bull. 97:999–1011.

369 Luzi, D.S. 2000. Long-term influence of jams and Nanson, G.C. and A.D. Knighton. 1996. Anabranch- LWD pieces on channel morphology, Carnation ing rivers: their cause, character and classifica- Creek, B.C. MSc thesis. Univ. British Colum- tion. Earth Surf. Process. Land. 21:217–239. bia, Vancouver, B.C. Pyrce, R.S. and P.E. Ashmore. 2003. Particle path Madej, M.A. and V. Ozaki. 1996. Channel response length distributions in meandering gravel-bed to sediment wave propagation and movement, streams: results from physical models. Earth Redwood Creek, California, USA. Earth Surf. Surf. Process. Land. 28:951–966. Process. Land. 21(10):911–927. Rapp, C.F. and T.B. Abbe. 2003. A framework for Millar, R.G. 2000. Influences of bank vegetation on delineating channel migration zones. Wash. alluvial channel patterns. Water Resour. Res. State Dep. Ecol., Olympia, Wash. Ecol. Publ. 36(4):1109–1118. No. 03-06-027. www.ecy.wa.gov/pubs/0306027. Mollard, J.D. 1973. Air photo interpretation of fluvial pdf (Accessed May 2010). features. In: Fluvial processes and sedimenta- Reid, L.M. and T. Dunne. 1996. Rapid evaluation tion. Proc. Can. Hydrol. Symp., Edmonton, of sediment budgets. Catena Verlag GMBH, Alta., Nat. Res. Coun. Can., Comm. Geol. Reiskirchen, Germany. Geophys., Subcomm. Hydrol., Ottawa, Ont., pp. 341–380. Schumm, S.A. 1971. Fluvial geomorphology: channel Montgomery, D.R. 1999. Process domains and the adjustment and river metamorphosis. In: River river continuum. J. Am. Water Resour. Assoc. mechanics. H.W. Shen (editor). H.W. Shen, Fort 35(2):397–409. Collins, Colo., pp. 5-1–5-22. Montgomery, D.R., T.B. Abbe, J.M. Buffington, N.P. ______. 1977. The fluvial system. John Wiley and Peterson, K.M. Schmidt, and J.D. Stock. 1996. Sons, New York, N.Y. Distribution of bedrock and alluvial channels ______. 1985. Patterns of alluvial rivers. Annu. Rev. in forested mountain drainage basins. Nature Earth Planet. Sci. 13:5–27. 381:587–589. Montgomery, D.R. and J.M. Buffington. 1993. Schwab, J.W. 1983. Mass wasting: October–November Channel classification, prediction of channel 1978 storm, Rennell Sound, Queen Charlotte response, and assessment of channel condition. Islands, British Columbia. B.C. Min. For., Wash. State Dep. Nat. Resour., Olympia, Wash. Victoria, B.C. Res. Note No. 91. www.for.gov. Rep. No. TFW-SH10-93-002. bc.ca/hfd/pubs/Docs/Mr/Scanned-Rn/Rn067- Rn100/Rn091.pdf (Accessed May 2010). ______. 1997. Channel-reach morphology in moun- tain drainage basins. Bull. Geol. Soc. Am. ______. 1998. Landslides on the Queen Charlotte 109(5):596–611. Islands: processes, rates, and climatic events. In: Carnation Creek and Queen Charlotte Is- Montgomery, D.R., J.M. Buffington, R.D. Smith, lands fish/forestry workshop: applying 20 years K.M. Schmidt, and G. Pess. 1995. Pool spac- of coastal research to management solutions. ing in forest channels. Water Resour. Res. D.L. Hogan, P.J. Tschaplinski, and S. Chatwin 31(4):1097–1105. (editors). B.C. Min. For. Res. Br., Victoria, B.C. Montgomery, D.R., B.D. Collins, J.M. Buffington, Land Manag. Handb. No. 41, pp. 41–46. www. and T.B. Abbe. 2003. Geomorphic effects of for.gov.bc.ca/hfd/pubs/docs/Lmh/Lmh41-1.pdf wood in rivers. In: The ecology and manage- (Accessed May 2010). ment of wood in world rivers. S. Gregory, K. ______. 2001. Donna Creek washout-flow: what did Boyer, and A.M. Gurnell (editors). Am. Fish. we learn? In: Terrain stability and forest man- Soc., Bethesda, Md., pp. 21–47. agement in the interior of British Columbia. P. Naiman, R.J. and R.E. Bilby (editors). 1998. River Jordan and J. Orban (editors). B.C. Min. For., ecology and management: lessons from the For. Sci. Prog., Victoria, Nelson, B.C., pp. 1–13. Pacific Coastal Ecoregion. Springer-Verlag, www.for.gov.bc.ca/hfd/pubs/Docs/Tr/Tr003/ New York, N.Y. Schwab.pdf (Accessed May 2010).

370 Wilford, D.J. 1987. Watershed workbook: forest hy- Levings, C.D., L.B. Holtby, and M.A. Henderson drology sensitivity analysis for coastal British (editors). 1989. Proc., National workshop on ef- Columbia watersheds. B.C. Min. For., Prince fects of habitat alterations on salmonid stocks. Rupert For. Reg., Smithers, B.C. www.for.gov. Can. Spec. Publ. Fish. Aquat. Sci. No. 105. bc.ca/hfd/library/documents/bib11716.pdf (Ac- Lindemayer, D.B., P.J. Burton, and J.F. Franklin. cessed May 2010). 2008. Salvage logging and its ecological conse- Wohl, E. 2000. Mountain rivers. Am. Geophys. quences. Island Press, Washington, D.C. Union, Washington D.C. MacIsaac, E.A. (editor). 2003. Forestry impacts on fish habitat in the northern interior of British References Related to Fish–Forestry Interaction Columbia: a compendium of research from the Program (channel morphology sections included Stuart-Takla Fish-Forestry Interaction Study. in each) Can. Tech. Rep. Fish. Aquat. Sci. No. 2509. Meehan, W.R. (editor). 1991. Influences of forest and Brewin, M.K. and D.M. Monita (editors). 1998. For- rangeland management on salmonid fishes and est-fish conference: land management practices their habitats. Am. Fish. Soc., Bethesda, Md. affecting aquatic ecosystems. Can. For. Serv., Spec. Publ. No. 19. North. For. Cent., Edmonton, Alta. Info. Rep. NOR-X-356. Meehan, W.R., T.R. Merrell, and T.A. Hanley (edi- tors). 1984. Fish and wildlife relationship in Chamberlin, T.W. (editor). 1988. Proceedings of old-growth forests. Proc. Symp., Am. Inst. Fish. the workshop: applying 15 years of Carnation Res. Biol., Juneau, Alaska. Creek results. Carnation Creek Steering Com- mittee. Pacific Biological Station, Nanaimo, Naiman, R.J. and R.E. Bilby (editors). 1998. River B.C. ecology and management: lessons from the Pacific Coastal Ecoregion. Springer-Verlag, Gibbons, D.R. and E.O. Salo. 1973. An annotated New York, N.Y. bibliography of the effects of logging on fish of the western United States and Canada. U.S. Northcote, T.G. and G.F. Hartman. 2004. Fishes and Dep. Agric. For. Serv., Pac. N.W. For Exp. Stn., forestry: worldwide watershed interactions and Portland, Oreg. Gen. Tech. Rep. PNW-10. management. Blackwell Science Ltd., Ames, Iowa. Hartman, G.F. (editor). 1982. Proceedings of the Car- nation Creek workshop: a 10-year review. Pac. Raedeke, K.J. (editor). 1988. Streamside management: Biol. Stn., Nanaimo, B.C. riparian wildlife and forestry interactions. Univ. Washington, Inst. For. Resour., Seattle, Hartman, G.F. and J.C. Scrivener. 1990. Impacts of Wash. Contrib. No. 59. forestry practices on a coastal stream ecosys- tem, Carnation Creek, British Columbia. Dep. Reynolds, P.E. (editor). 1989. Proceedings of the Fish. Oceans, Ottawa, Ont. Can. Bull. Fish. Carnation Creek herbicide workshop. Forest Aquat. Sci. No. 223. Pest Manag. Inst., Sault Ste. Marie, Ont., B.C. Min. For., Res. Br., and For. Can., Pac. For. Hogan, D.L., P.J. Tschaplinski, and S. Chatwin (edi- Cent., Victoria, B.C. FRDA Rep. No. 63. www. tors). 1998. Carnation Creek and Queen Char- for.gov.bc.ca/hfd/pubs/Docs/Frr/Frr063.htm lotte Islands fish/forestry workshop: applying (Accessed May 2010). 20 years of coastal research to management solutions. B.C. Min. For., Res. Br., Victoria, Salo, E.O. and T.W. Cundy (editors). 1987. Streamside B.C. Land Manag. Handb. No. 41. www.for.gov. management: forestry and fishery interactions. bc.ca/hfd/pubs/Docs/Lmh/Lmh41.htm (Ac- Univ. Washington, Inst. For. Resour., Seattle, cessed March 2010). Wash. Contrib. No. 57. Krygier, J.T. and J.D. Hall 1970. Proceedings of a Stednick, J.D. (editor). 2008. Hydrological and bio- symposium forest land uses and stream envi- logical responses to forest practices: the Alsea ronment. October 19–21, 1970. School For. Dep. watershed study. Springer-Verlag, New York, Fish. Wildl., Oreg. State Univ., Corvallis, Oreg. N.Y. Springer Ecolog. Stud. Ser., Vol. 199.

371