UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Bankfull and effective discharge in small mountain streams of British Columbia Brayshaw, Drew Devoe 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_brayshaw_drew.pdf [ 29.9MB ]
Metadata
JSON: 24-1.0072555.json
JSON-LD: 24-1.0072555-ld.json
RDF/XML (Pretty): 24-1.0072555-rdf.xml
RDF/JSON: 24-1.0072555-rdf.json
Turtle: 24-1.0072555-turtle.txt
N-Triples: 24-1.0072555-rdf-ntriples.txt
Original Record: 24-1.0072555-source.json
Full Text
24-1.0072555-fulltext.txt
Citation
24-1.0072555.ris

Full Text

BANKFULL AND EFFECTIVE DISCHARGE IN SMALL MOUNTAIN STREAMS OF BRITISH COLUMBIA  by Drew Devoe Brayshaw  B. Sc., The University of British Columbia, 1994 M. Sc., The University of British Columbia, 1998  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  January 2012  © Drew Devoe Brayshaw, 2012  Abstract In this thesis, the concept of channel-forming discharge developed for large lowland rivers is critically re-evaluated for small mountain streams. In large lowland rivers under equilibrium conditions, the effective discharge (the discharge interval that transports the greatest proportion of sediment) approximates the bankfull discharge. The effective and bankfull discharges therefore provide measurable analogues for the theoretical channelforming discharge, responsible for the form and dimensions of the stream channel. In small mountain streams this concordance of bankfull and effective discharges has been suspected to break down, in part because the channel form and dimensions are determined by the nonfluvial sediment supply as well as by fluvial processes. This interaction of non-fluvial sediment supply with fluvial processes makes the application of models and concepts developed for lowland rivers to mountain streams difficult. The history of past glaciation in mountain environments adds complexity in the form of increased sediment supply and changes in streamflow regime over time. Bankfull and effective discharges are measured and estimated for small mountain streams in three distinct hydroclimatic zones in southern British Columbia. These discharges are determined using surveyed stream cross-sections, photogrammetric analysis of in-channel sediment, long-term Water Survey of Canada gauging records, and a two-fraction sediment transport model. Results indicate multiple orders of magnitude of variability in the frequency and magnitude of bankfull and effective discharge for the studied streams, and little correlation between the bankfull and effective discharges. Bankfull frequency can be related to hydraulic and hydroclimatic variables while effective discharge frequency cannot. Most of the studied  ii  streams are incised, with bankfull discharges one or more orders of magnitude greater than their effective discharge, a condition attributable to the legacy of Pleistocene glaciation. In these streams the effective discharge is not a channel-forming discharge; at best, it is a channel-maintaining flow, and at worst it is geomorphically meaningless. In a few streams, the bankfull and effective discharges are both large, rare events; these are the only streams in the study in which the bankfull and effective discharges approximate a truly channel-forming discharge.  iii  Table of Contents Abstract .................................................................................................................................... ii Table of Contents ................................................................................................................... iv List of Tables ......................................................................................................................... vii List of Figures ....................................................................................................................... viii List of Symbols and Abbreviations ..................................................................................... xii Acknowledgements .............................................................................................................. xiv Chapter 1: Introduction ........................................................................................................ 1 1.1  Small Mountain Streams as Transitional Between Alluvial and Non-Alluvial  Environments ........................................................................................................................ 1 1.2  Statement of Research Hypothesis and Goals .......................................................... 6  1.3  Format of Thesis ....................................................................................................... 7  Chapter 2: Background ......................................................................................................... 8 2.1  Channel-forming Discharge in Threshold Alluvial Streams .................................... 8  2.2  Bankfull Discharge ................................................................................................. 10  2.3  Effective Discharge ................................................................................................. 22  Chapter 3: Study Area and Selection of Streams ............................................................. 31 Chapter 4: Methods and Data Collection .......................................................................... 39 4.1  Field Measurements ................................................................................................ 39  4.1.1  Field Site Selection ............................................................................................. 39  4.1.2  Measurements ..................................................................................................... 39  4.2 4.2.1  Other Data Collection ............................................................................................. 46 Flow Frequency Data .......................................................................................... 46  iv  4.2.2  Physiographic Data From Maps.......................................................................... 46  4.2.3  Sediment Source Analysis .................................................................................. 46  4.3  Methods of Analysis ............................................................................................... 47  4.3.1  Flow Frequency Analysis ................................................................................... 47  4.3.2  Sediment Photo Processing and Analysis ........................................................... 48  4.3.3  Volumetric Sediment Sample Processing ........................................................... 49  4.3.4  Survey Data GIS Processing ............................................................................... 49  4.4  Estimation of Bankfull Discharge ........................................................................... 50  4.5  Estimation of Sediment Transport and Bedload Effective Discharge .................... 52  4.5.1  Estimation of Sediment Transport ...................................................................... 53  4.5.2  Division of Recorded Discharge into Discharge Classes and Qeff Estimation ... 56  4.6  Estimation of Size of Mobile Sediment Fraction .................................................... 58  4.7  Calculation of Half-load Discharge ........................................................................ 60  4.8  Statistical Significance and Physical Meaning ....................................................... 60  4.9  Sources of Error ...................................................................................................... 61  4.9.1  Sources of Error in Estimation of Bankfull Discharge ....................................... 61  4.9.1.1  Errors in Field Measurement ...................................................................... 61  4.9.1.2  Estimation of Manning’s n Value and Reach-averaged Channel Slope. .... 63  4.9.1.3  Estimation of Sediment Transport, Effective Discharge and Half-load  Discharge .................................................................................................................... 64 4.9.1.4  Estimation of Grain Shear Stress ................................................................ 64  4.9.2  Estimation of Sediment Size ............................................................................... 64  4.9.3  Estimation of Half-load and Effective Discharge ............................................... 65  v  4.9.4  Estimation of Frequency of Bankfull, Effective and Half-load Discharge ......... 67  Chapter 5: Results................................................................................................................ 69 5.1 5.1.1  Bankfull Discharge ................................................................................................. 69 Factors Affecting Relative Magnitude of Bankfull Discharge ........................... 77  5.2  Effective Discharge ................................................................................................. 94  5.3  Half-load Discharge .............................................................................................. 114  5.4  Other Results ......................................................................................................... 120  Chapter 6: Discussion ........................................................................................................ 122 6.1  Bankfull Discharge ............................................................................................... 122  6.1.1  Comparison of Bankfull Discharge Values to Other Regions .......................... 122  6.1.2  Interrelation of Factors Affecting Bankfull Discharge ..................................... 124  6.1.3  Bankfull Discharge in Transitional Alluvial Streams ....................................... 135  6.2  Effective (and Half-load) Discharge ..................................................................... 136  6.3  Linking the Bankfull and Effective Discharges .................................................... 141  Chapter 7: Conclusions ..................................................................................................... 150 7.1  Frequency and Magnitude of Bankfull and Effective Discharge.......................... 150  7.2  Bankfull and Effective Discharge as Proxies for Channel-Forming Discharge ... 153  7.3  Suggestions for Further Research ......................................................................... 157  Bibliography ........................................................................................................................ 159 Appendix A: Hydrologic Data from East Creek, Malcolm Knapp Research Forest ... 177 Appendix B: Examples of Channel Conditions and Sediment Size Distributions ........ 180 Appendix C: Example of Bankfull Discharge Calculation ............................................. 215  vi  List of Tables Table 3-1: Watershed name, WSC gauge number, years of record, and hydroclimatic region. ................................................................................................................................................. 35 Table 5-1: Measured parameters by watershed. ..................................................................... 70 Table 5-2: Estimated and computed parameters for each drainage basin. .............................. 71 Table 5-3: Comparison of bankfull discharge values from other studies to those determined in this study. ............................................................................................................................ 72 Table 5-4: Predicted and observed discharge thresholds for onset of sand and gravel sediment transport in benchmark streams. ............................................................................................. 95 Table 5-5: Fraction of time effective discharge is exceeded, by hydrologic region and overall. ................................................................................................................................................. 96 Table 5-6: Qeff/Qmad ratio by hydroclimatic region and overall. ............................................. 98 Table 5-7: Mean, median, and standard deviation of mobility ratios (D50mobile/D50 bed). ...... 105 Table 5-8: Bankfull, effective and half-load discharges for the studied streams.................. 115 Table 6-1: Mean (and median) Qbf/Q2, Qbf/Qmad, and Qbf return period determined from the annual maximum flood series for the entire study area and by hydrologic region. .............. 135  vii  List of Figures Figure 3-1: Hydrologic regions and locations of sampled watersheds for study. ................... 32 Figure 4-1: Bankfull range and best estimates by cross-section for Greata Creek. ................ 52 Figure 4-2: Histogram of discharge frequency for Cabin Creek. ........................................... 57 Figure 4-3: Sample effectiveness plot for Redfish Creek, cross-section 11. .......................... 58 Figure 5-1: Histogram of bankfull frequency exceedence based on flow duration. ............... 73 Figure 5-2a: Frequency-magnitude relation for bankfull discharge based on ratio of bankfull to mean annual discharge and flow duration series. ............................................................... 75 Figure 5-2b: Frequency-magnitude relation for bankfull discharge based on ratio of bankfull to Q2 and annual maximum series. ......................................................................................... 76 Figure 5-3a: Qbf/Q2 variation by hydrologic region. ............................................................... 78 Figure 5-3b: Qbf/Qmad variation by hydrologic region. ........................................................... 78 Figure 5-4a: Plot of Qbf vs. Q2 by hydrologic region, p=0.0000. ........................................... 79 Figure 5-4b: Qbf/Q2 ratio vs. unit runoff. ................................................................................ 80 Figure 5-5a: Qbf/Qmad ratio vs. drainage basin area by region. ............................................... 81 Figure 5-5b: Qbf/Q2 ratio vs. drainage basin area. .................................................................. 82 Figure 5-6a: Qbf/Q2 ratio vs. reach gradient by hydrologic region. ........................................ 83 Figure 5-6b: Qbf/Qmad ratio vs. reach gradient by hydrologic region. ..................................... 84 Figure 5-7a: Qbf/Q2 ratio vs. Cv of peak flow series by hydrologic region. ........................... 85 Figure 5-7b: Qbf/Qmad ratio vs. Cv of annual maximum series of peak flows. ....................... 86 Figure 5-8a: Box and whisker plot of Qbf/Qmad grouped by Q2 streampower/D84 class. ... 87 Figure 5-8b: Box and whisker plot of Qbf/Q2 grouped by Q2 streampower/D84 class. ........... 87  viii  Figure 5-9: Box and whisker plot of Qbf/Qmad grouped by wood density per unit channel length class. ............................................................................................................................. 88 Figure 5-10a: Box and whisker plot of Qbf/Qmad grouped by channel width class. ................ 89 Figure 5-10b: Box and whisker plot of Qbf/Q2 grouped by channel width class. ................... 89 Figure 5-11a: Albert River: lower end of reach above woody debris jam with aggraded sediment. ................................................................................................................................. 90 Figure 5-11b: Albert River, lower end of degraded/armoured reach below woody debris jam. ................................................................................................................................................. 91 Figure 5-12a: Relation between dimensionless bankfull width and dimensionless bankfull discharge. ................................................................................................................................ 92 Figure 5-12b: Relation between channel slope and dimensionless bankfull discharge. ......... 94 Figure 5-13: Histogram of effective discharge exceedance frequencies. ............................... 96 Figure 5-14: Histogram of Qeff to Qmad ratio using arithmetic bins. ....................................... 97 Figure 5-15: Effective discharge frequency-magnitude relation. ........................................... 98 Figure 5-16: Example of Type I stream effectiveness diagram. ............................................. 99 Figure 5-17: Example of Type II stream effectiveness diagram........................................... 100 Figure 5-18: Example of effectiveness diagram for Type III stream.................................... 101 Figure 5-19: Cluster analysis of effective discharge types. .................................................. 103 Figure 5-20: Magnitude-frequency analysis of effective discharge stratified by stream type. ............................................................................................................................................... 104 Figures 5-21a and Figure 5-21b: Qeff/Qmad ratio vs. D50mob/D50bed ratio for Q2 and Q20 discharges. ............................................................................................................................. 107  ix  Figures 5-21c and Figure 5-21d: Qeff/Qmad ratio vs. D50mob/D50bed ratio for bankfull and effective discharge. ............................................................................................................... 108 Figure 5-22: Unscaled Qeff vs Qbf, plotted by stream effective discharge type..................... 110 Figure 5-23: Q2 Streampower/D84 ratio vs. unscaled Qeff. .................................................... 111 Figures 5-24a and 5-24b: Qeff/Qbf ratio vs. relative magnitudes of bankfull discharge. ....... 112 Figure 5-25: Regional trend in Qeff with drainage basin area. .............................................. 113 Figure 5-26: Absolute (unscaled) relation of Qbf and Qh by hydrologic region, p=0.0002. . 116 Figure 5-27: Absolute (unscaled) relation of Qeff and Qh by hydrologic region, p=0.0000. . 117 Figures 5-28a and 5-28b: Relation between relative magnitudes of Qh and Qeff. ................. 118 Figure 5-29: Relation between relative magnitudes of Qh and Qbf . ..................................... 119 Figure 5-30: Relative magnitude of Qh vs. Q2 streampower/D84 ratio. ................................ 119 Figure 5-31: Qh/Qeff ratio vs area, stratified by hydrologic region. ...................................... 120 Figure 6-1: Qbf/Qmad ratio vs. area values from this study compared to results from three other comprehensive studies. ......................................................................................................... 124 Figure 6-2a: Relation between reach gradient and drainage basin area for each hydroclimatic region, p=0.0043. .................................................................................................................. 125 Figure 6-2b: Relation between Melton ratio and drainage basin area for each hydroclimatic region, p= 0.0008. ................................................................................................................ 126 Figure 6-3: Flood duration series frequency of bankfull discharge vs. Qbf/Q2 ratio, p=0.0701. ............................................................................................................................................... 127 Figure 6-4a: Relation between Q2 streampower/D84 ratio and bankfull channel width, p=0.0000. .............................................................................................................................. 129  x  Figure 6-4b: Relation between bankfull width and drainage basin area for each hydroclimatic region, p-values listed by region: KC p = 0.00000, TO p = 0.0007, CM p = 0.0033. ......... 130 Figure 6-5a: Cv of annual maximum peak flow series vs. unit mean runoff by region. ...... 132 Figure 6-5b: Relation between Q2 streampower/D84 and Cv of annual maximum series of floods, p=0.0174. .................................................................................................................. 133 Figure 6-6: Replot of Figure 5-17 (Daves Creek effectiveness diagram, example of Type II stream) using fixed number of classes. ................................................................................. 140 Figure 6-7: Qeff vs. Qbf values from this study compared to results from other published studies from western North America. ................................................................................... 143  xi  List of Symbols and Abbreviations B~  dimensionless bankfull width  Bbf  bankfull width (m)  CFA  Consolidated Frequency Analysis program  CM  Coast and Mountains region  Cv  coefficient of variation of annual maximum series  D50  diameter of median particle (m)  D50b  diameter of median bed particle (m)  D50m  diameter of median mobile bed particle (m)  D84  diameter of 84th percentile bed particle (m)  Dg  median diameter of gravel fraction (particles coarser than 8mm) (m)  Ds  median diameter of sand fraction (particles finer than 8mm) (m)  FFD  flow frequency distribution  Fg  fractional proportion of bed material coarser than 8mm  Fs  fractional proportion of bed material finer than 8mm  g  gravitational acceleration (9.8 m/s2)  hʹ  grain-effective depth (m),  h  mean depth of flow (m),  KC  Kootenay-Columbia region  Ks  roughness parameter (m), equal to 3D84  LWD large woody debris n  Manning’s coefficient (s m-⅓)  PDF  probability density function  Q^  dimensionless bankfull discharge,  Q2  discharge equaled or exceeded once in 2 years on average (m3 s-1)  Q20  discharge equaled or exceeded once in 20 years on average (m3 s-1)  Qi  discharge equaled or exceeded once in i years on average (m3 s-1)  Qbf  bankfull discharge (m3 s-1)  Qbi  fractional unit bedload transport rate of fraction i (sand or gravel) (m3 s-1)  Qeff  effective discharge (m3 s-1)  Qh  half-load discharge (m3 s-1) xii  Qmad  mean annual discharge (m3 s-1)  s  ratio of grain density to water density (grain density is assumed to be the density of quartz, so s-1 = 1.65)  S  channel gradient (m/m),  TO  Thompson-Okanagan region  U  mean velocity (m s-1),  u*  shear velocity (m s-1, equal to [τ/ρ]0.5)  W i*  dimensionless transport rate  WSC Water Survey of Canada θc  critical Shields number  ρ  density of water (1000 kg m-³)  τ  total shear stress (kg m-1 s-2)  τ*ri  fractional reference shear stress  τg  grain shear stress (kg m-1 s-2).  ϕʹ  1.27  ϕ  τ/τri  Ω  reference stream power at Q2 discharge (kg m s-3)  xiii  Acknowledgements This research was funded by the BC Forest Sciences Program. The Water Survey of Canada provided a discount on data services. The City of Nelson, City of Creston, Central Fraser Valley Water Commission, and Greater Vancouver Watershed District provided access to watersheds normally off-limits to the public. Dan Hogan provided data from Carnation Creek and facilitated access to the watershed. Dr. Michael Feller provided streamflow data for East Creek. Dr. David Graham provided assistance with troubleshooting the Digital Gravelometer. Fred Touche, Steven Harng, and Miranda Huron assisted with fieldwork; Hale Jones-Cox, Jennifer Wardle and Tony Lagemaat assisted with sediment analysis and photo processing. Eryne Croquet provided valuable help with GIS data as well as thesis formatting. I would particularly like to thank Scott Weston for providing me the stimulus to begin this project, Dr. Younes Alila and Dr. Marwan Hassan of the University of British Columbia together with Dan Hogan and Dr. Ned Andrews for their dedicated supervision, and Ken Hughes-Adams and Gordon Butt of Madrone Environmental Services Ltd. for financial and intellectual support.  xiv  Chapter 1: Introduction Small gravel-bed mountain streams, particularly those in formerly glaciated regions, comprise a domain that is transitional between alluvial (self-formed) and non-alluvial streams. These mountain streams are not as well-studied as are lowland alluvial rivers, but concepts and expectations developed from studies of lowland rivers are often applied to the mountain streams for lack of more appropriate knowledge. Large, lowland alluvial rivers are often in a condition of equilibrium or dynamic equilibrium, while small gravel-bed mountain streams often depart from equilibrium conditions. The misapplication of expectations developed from lowland rivers to these mountain streams can therefore result in increased risks to human life and property, design failures, and/or mismanagement of riparian areas. This thesis examines one such concept, the channel-forming discharge, and its measurable proxies, the bankfull discharge and the effective discharge, in formerly glaciated gravel-bedded mountain streams. In lowland alluvial rivers, the bankfull discharge often approximates the effective discharge, and the magnitude of both is comparable to that of the mean annual flood. These expectations are challenged for the studied streams. The ranges of frequencies and relative magnitudes of bankfull and effective discharges present in the studied streams are characterized and used to draw conclusions about the channel-forming processes. 1.1 Small Mountain Streams as Transitional Between Alluvial and Non-Alluvial Environments Stream channels transport both water and sediment. Factors which influence the form and development of the stream channel include the frequency and magnitude of streamflow, the nature of the channel bed and banks, the supply (timing, volume, nature and size) of  1  sediment to the channel, the topography of the environment the stream flows through, the nature and abundance of riparian vegetation, and the effects of contingency and history (the age of the stream and of the landscape in which it flows, and the history of anthropogenic disturbance of the channel and its upstream catchment). Common types of streams are alluvial, colluvial, and bedrock streams (Montgomery and Buffington, 1997). Alluvial streams are those streams which flow over sediments they are capable of transporting; because the stream bed and banks consist of mobile sediment, these streams are said to be self-formed. Colluvial streams flow over sediments derived from adjacent hillslopes, where the nature of sediment supplied to the stream means that the stream bed and banks are generally immobile, or persist because the supply of colluvial sediment to the channel greatly exceeds the ability of the channel to transport that sediment away. Bedrock streams flow over bedrock, which is similarly immobile. Colluvial and bedrock streams may transport alluvial sediment, but they are not self-formed in the same manner as alluvial stream channels. Other, rarer types of stream channels are possible, such as stream channels in peat wetlands that flow entirely over organic material and transport water but no clastic sediment (Nanson et al, 2010). Fluvial sediment transport can be characterized as either supply-limited or transportlimited (Nordin and Beverage, 1965; Schumm, 1985; Nordin, 1985). In transport-limited systems, sediment is easily transported, and the limit on sediment transport is the availability of streamflow. As stream discharge increases in a transport-limited system, so does the amount of sediment transported. In supply-limited systems, by contrast, sediment transport is limited by the availability of sediment rather than streamflow, and theoretically two stream discharges of exactly the same magnitude can transport wildly different amounts of sediment  2  depending on whether sediment is available for transport or not. An intermediate case between purely supply-limited and purely transport-limited conditions is represented by those streams in which there is one or more discharge thresholds for sediment transport: below the threshold, little or no sediment is transported, while above the transport threshold, sediment becomes mobile. Large rivers in which alluvial characteristics are well-developed tend to be characterized as transport-limited, and the sediment they transport tends to be either relatively fine-grained, or the size of mobile sediment is small relative to the size of the channel. By contrast, exclusively supply-limited conditions are found almost exclusively in non-alluvial bedrock streams (Bravo-Espinosa et al., 2003). In gravel-bed streams, particularly smaller streams where the sizes of the largest clasts present in the channel are some appreciable fraction of the total channel width, the intermediate condition occurs, and one or more thresholds for sediment transport are present (Schumm, 1979; Church, 2002). Such small gravel-bed streams are most commonly found in headwater and/or mountain environments (Montgomery and Buffington, 1997). They are characterised by relatively coarse sediment (ranging from gravel to boulders), threshold-limited sediment transport conditions, relatively high thresholds for the initiation of sediment transport due to bed armouring and surface structures, and a high degree of geomorphic connectivity to the surrounding hillslopes. These streams are transitional between channels in which the bed and banks are non-alluvial, composed entirely of bedrock, colluvium or glaciogenic material, and streams in which fluvial processes dominate and which have self-formed bed and banks composed entirely of alluvium. In cases where riparian vegetation is well-established, there  3  may be a significant discrepancy between the mobility of the bed sediment and the mobility of the bank sediment, with the bed more mobile than the banks. In large rivers, where sediment is generally fine-textured (or where the size of sediment is very small relative to the size of the channel) and sediment transport is transportlimited, conditions of dynamic equilibrium predominate because the system is capable of rapidly moving and redistributing the available sediment. The response time of the system to disturbance is therefore short. In smaller gravel-bed streams, the episodic sediment supply and potentially high thresholds for sediment transport result in much longer response times to disturbances, ranging from tens to thousands of years; equilibrium conditions and processes are therefore rarer and non-equilibrium conditions correspondingly common (Hassan and Zimmermann, 2011). Studies of river hydrology and fluvial geomorphology have tended to focus on the larger rivers, where transport-limited conditions predominate and in which dynamic equilibrium prevails. The knowledge gained from such studies may not be applicable to the small or mountain streams with gravel beds and transitional alluvial conditions in which equilibrium conditions are rare or form a transient state. Likewise, although the end points are well known, the exact nature of the transition between alluvial and non-alluvial conditions is difficult to characterize from our current knowledge, as are the transitions between purely transport-limited conditions, threshold transport-limited conditions, and supply-limited conditions. Large alluvial lowland rivers are entirely self-formed. They flow over beds and have banks composed of sediment they have transported over time and that they are competent to transport. Purely bedrock and colluvial streams are not self-formed. They flow over sediment  4  or bedrock that they are unable to transport and they may not have definable banks. In between these two extremes lies a transitional alluvial environment such as occurs in small mountain streams. In this transitional alluvial environment, the stream is capable of transporting some but not all of the sediment in the channel, and the channel is at least partly formed by both present and past streamflow and sediment transport. These transitional alluvial streams can have relatively high or quite low rates of sediment transport, but are generally supply-limited or threshold supply-limited rather than transport-limited. Furthermore, because of the importance of sediment supply from outside the channel, purely equilibrium conditions are rare: cycles of aggradation and degradation are the most common condition, and these can last for hundreds to thousands of years. In these transitional alluvial streams, one might expect that there will be a wide range of frequencies and relative magnitudes for bankfull and effective discharge. For streams which are close to fully alluvial, there should be relatively close agreement between the frequencies and magnitudes of bankfull and effective discharge, and these discharges should provide a good approximation of the channel-forming discharge. For streams that are minimally alluvial in nature, there should be a wide divergence between the values of bankfull and effective discharge, and the relevance of the theoretical channel-forming discharge may be low because the channel is only minimally formed by alluvial processes. Between these extremes, it is expected that the range of frequencies and relative magnitudes for bankfull and effective discharge will provide information relating to the relative importance of alluvial processes in forming the channel and the degree to which the channelforming discharge concept is relevant to the stream.  5  1.2 Statement of Research Hypothesis and Goals This thesis evaluates bankfull and effective discharge frequencies and magnitudes for a number of small streams in southern British Columbia, a study area which offers a wide and highly heterogenous range of relief, physiography, geologic and climatic environments. All of the drainage basins share the following characteristics: they are relatively small, are located in mountainous or upland areas, and were glaciated during the Pleistocene; all the streams are gravel-bedded. The streams were chosen to cover the range of discharges and sediment sizes over which the transition from poorly-developed to well-developed alluvial character should occur, and over which it is therefore hypothesized that the channel-forming discharge will go from being a concept of little relevance to a strongly relevant concept. A range of frequencies and relative magnitudes should be present for both bankfull and effective discharge in the studied streams. The specific research goals are to: •  characterize the range of frequencies and relative magnitudes of bankfull and effective discharge present within the study area;  •  evaluate the significant geomorphic and hydroclimatic factors (drainage basin area, channel gradient, hydroclimate, flood generation mechanisms, sediment size, etc.) that appear to explain variation in the frequency and magnitude of bankfull and effective discharge; and,  •  use this information to draw conclusions about the relevance of the concept of channelforming discharge to the streams in the study area.  6  1.3 Format of Thesis Chapter 2 presents background detail that complements this introduction, including definitions and a comprehensive literature review tracing the development of and arguments concerning the definitions of channel-forming, bankfull, and effective discharge. Chapter 3 describes the study area in detail, as well as presenting details related to each of the three hydroclimatic regions present within the study area. Chapter 4 presents a detailed exposition of the measurements made and computational methods used to determine values of bankfull and effective discharge for each stream in the study area, and Chapter 5 presents the results from those observations and methods, including characterization of both the frequency and magnitude of bankfull and effective discharge as well as of factors affecting the frequency and magnitude of bankfull and effective discharge. Chapter 6 discusses those results and places them within the context of other applicable research. Chapter 7 then draws conclusions from the results and discussion about the meaning of the results and the implications for channel-forming processes in the studied streams. A detailed bibliography follows.  7  Chapter 2: Background This chapter traces the development of the concepts of and research concerning channel-forming, bankfull, effective, and half-load discharges in the hydrologic and geomorphic literature, with a particular focus on their applicability to threshold alluvial environments such as small, gravel-bedded mountain streams. Where different authors have proposed variant definitions of, or ways of computing, one of the discharges of interest, the issues underlying the choices of definition or method are presented and a rationale given for choosing the most applicable criterion for definition or method for this study. 2.1 Channel-forming Discharge in Threshold Alluvial Streams The concept of the channel-forming discharge provides a specific example of the way in which ideas developed for large, transported-limited, fully alluvial river systems may not be directly applicable to smaller, gravel-bedded, threshold transport-limited or supply-limited streams. The idea that the shape and dimensions of a stream channel in dynamic equilibrium with its surroundings can be represented by a single reference discharge, called the channelforming or dominant discharge (Inglis, 1949), is deeply engrained in the hydrologic literature, and arose initially from studies of artificial channels, specifically canals (Kennedy, 1895; Schaffernak, 1922; Lacey, 1933, 1939; Inglis 1949; Leopold and Maddock, 1953; Wolman and Miller, 1960; Ackers, 1972). A poorly designed canal, over- or undersized for the flow diverted into it, would erode or suffer from sedimentation, whereas a canal that was “in regime” would maintain its dimensions. Extension of the same concept from canals to natural stream channels gave rise to idea that a single representative discharge or narrow range of discharges, if maintained indefinitely, would maintain an equilibrium stream channel and hence could be used as an index to represent the wider range of discharges  8  experienced by a natural stream. In other words, out of the wide range of discharges experienced by a natural channel, a smaller subset of discharges was most characteristic of the equilibrium form and dimensions of the channel and therefore had formed the channel or would, if maintained indefinitely, result in a channel of the same form as the existing natural channel. This channel-forming discharge is simple conceptually but difficult to characterize by direct measurement, leading to efforts to approximate the channel-forming discharge by an index discharge that could be directly or indirectly measured. Examples of such index discharges include those based on a specified return period, such as the 1-in-2 year flood (Q2); the bankfull discharge (Qbf), based on the point at which water overtops the channel banks (Wolman and Leopold, 1957); or the effective discharge (Qeff), that discharge interval that over time transports the most sediment (Wolman and Miller, 1960). The approximation of the channel-forming discharge by a specified return period discharge is the simplest in practice, requiring only the analysis of flow frequency to define. It is also the least justifiable theoretically, since the entire rationale for using a specified return period to approximate the channel-forming discharge rests on the argument that either bankfull or effective discharge (or both) recur at some characteristic interval. Thus the argument that a specified return period approximates the channel-forming discharge requires that either bankfull discharge, effective discharge or both have a characteristic return period, and that one or both of these discharges must be the channel-forming discharge. Both the effective and bankfull discharge have been studied extensively, and there is a correspondingly large body of literature related to each of these subjects. Bankfull and effective discharge, as analogues of channel-forming discharge, have philosophical  9  connections to more general considerations of the frequency and magnitude of geomorphic processes, specifically the debate between gradualism (uniformitarianism) and catastrophism in geology and geomorphology (Baker, 1988; 1998). The idea that stream channels are dominantly formed by relatively rare, large-magnitude events is a part of the larger catastrophist idea that the landscape is formed by rare, large-magnitude events, and likewise uniformitarianism supposes relatively frequent, small-magnitude events form the landscape. The modern philosophic synthesis of these ideas has moved beyond promoting one and demoting the other, to evaluating the circumstances under which each predominates. With respect to channel-forming discharge, or measures which approximate the theoretical channel-forming discharge, therefore, the important thing is to understand the circumstances under which bankfull discharge and effective discharge are respectively frequent or rare events. 2.2 Bankfull Discharge Bankfull discharge is defined as that discharge at which a stream is full to the top of the streambanks without overflowing onto the floodplain (Wolman, 1955; Wolman and Leopold, 1957; Nixon, 1959; Woodyer, 1968; Williams, 1979; Johnson and Heil, 1996). The identification of the banks of a stream as distinct features, and the recognition that flooding in a stream involves overbank flow, predate the science of geomorphology; the adoption of quantitative methods in geomorphology in the 1950s led to the refinement of these concepts, and identified the bankfull discharge specifically as the threshold for the beginning of overbank flow, rather than defining flooding subjectively as the stage at which, for instance, dangerous flow conditions or damage to property occurred (Wolman and Leopold, 1957). The similarity of bankfull hydraulic geometry relations across a wide range of climate and  10  landscape types (Leopold et al., 1964; Parker et al., 2007) underlies much of the science of quantitative hydrology. The identification of the bankfull stage is conceptually simple: the stream bank is a recognizable geomorphic feature which separates the stream channel from the adjacent landscape - for instance, from the floodplain. In practice, identification of the bankfull stage is more difficult than it may first appear. A part of this difficulty arises from the prevalence of multiple, overlapping methods of defining and identifying the banks of a stream, including methods based on morphologic, geometric, sedimentological, vegetative/ecological, statistical, and experiential criteria (Williams, 1979; Johnson and Heil, 1996; RadeckiPawlik, 2002; Navratil et al, 2006). Published methods of defining bankfull stage include: a) the elevation of the valley flat (Nixon, 1959; Woodyer, 1968) b) the elevation of the active floodplain (Wolman and Leopold, 1957); c) the elevation of the lowest bench (Schumm, 1960; Bray, 1972); d) the elevation of the middle or an intermediate bench, where more than one bench is present (Woodyer, 1968); e) the elevation of the most prominent bench (Kilpatrick and Barnes, 1964) f) the elevation of the top surface of channel bars (Wolman and Leopold, 1957); g) the elevation of the lowest limit of perennial vegetation (Schumm, 1960; Leopold and Skibitzke, 1967; Bray, 1972); h) the elevation of the lowest limit of small organic debris such as leaves, needles and cones (Leopold and Skibitzke, 1967; Castro and Jackson, 2001); i) the elevation corresponding to the community of riparian vegetation subject to periodic inundation (Harris, 1999);  11  j) the elevation of the upper limit of sand-sized particles in the bank sediment (Leopold and Skibitzke, 1967); k) the elevation of another textural change in the bank sediment, such as from rounded to angular clasts (Radecki-Pawlik, 2002); l) the elevation at which the width to depth ratio of the channel cross-section is at a minimum (Wolman, 1955; Pickup and Warner, 1976; Navratil et al, 2006); m) the area at which the area to width ratio of the channel changes abruptly (Williams, 1978; Navratil et al, 2006); n) the first maximum of the Riley bench index (Riley, 1972); o) the second, third, or subsequent maximum of the Riley bench index (RadeckiPawlik, 2002); p) the elevation at which the stage-discharge relation displays a pronounced break in slope (Williams, 1978) ; q) the stage corresponding to the discharge corresponding to a fixed return period (Dury, 1976); r) the stage corresponding to the discharge computed from a regional or general equation predicting bankfull discharge from parameters such as drainage basin area (Williams, 1978; Radecki-Pawlik, 2002) s) the elevation at which an experienced fluvial geomorphologist, who examines the stream channel and conducts an internal heuristic assessment of the weighting of each of the preceding factors, determines the top of bank is located (USDA Forest Service, 1995).  12  This listing comprises 19 methods. While not all methods can be applied to any one stream cross-section, and in many cases the use of different bankfull criteria will produce a comparable result, the high number of potentially overlapping definitions results in a corresponding degree of uncertainty in the estimation of bankfull discharge (Williams, 1978; Johnson and Heil, 1996; Radecki-Pawlik, 2002; Navratil et al, 2006). It is evident when the methods are compared that for a given reach some will be more appropriate than others; most often problems arise in practice when some of the best indicators of bankfull stage are not present or are ambiguous, and a determination of bankfull consequently must be made by choosing from one or more of the secondary indicators (Leopold and Skibitzke, 1967). The morphologically-based (items a-f) and geometrically-based (items l-o) bankfull indicators presented above are rendered more difficult to use in practice because of the issue of inactive floodplains, or terraces. Fluvial terraces are segments of inactive floodplain located adjacent to but higher than the active floodplain, and are diagnostic of episodes of degradation following tectonic uplift or changes in streamflow or sediment supply (Merritts et al., 1994). The distinction between an active and inactive floodplain is not always easy to make in the field. Wolman and Leopold (1957) adopted an a posteriori definition: surveying a range of reported bankfull discharge frequencies ranging from 1.01 to 200 years on the annual maximum series, they noted that most were less than 2 years, and concluded that therefore, any surface inundated less frequently than once in two years was a terrace and not an active floodplain. This definition has been criticized as unnecessarily narrow: Schmudde (1968) revised the cutoff between active and inactive floodplains upwards to 10 years, while more recent summaries have used a synthesis of risk-based, ecological, and geomorphic approaches to suggest that the active floodplain may include, for instance, everything below  13  the 100-200 year flood level, as long as it is relatively flat, largely composed of fluvial sediments, and located in the valley bottom (Marriott and Alexander, 1999). Definitions which use a frequency of inundation to define and divide active and inactive floodplains are of low utility in the field, where frequency of inundation must be inferred from available data. Furthermore, like any of the other criteria for defining bankfull stage, they are of low utility when considered singly. Given, for instance, a choice between two benches (to use the terminology of Riley (1972)), one a low bench with a relatively high frequency of inundation, say 1.01 year on the annual maximum series, which is flooded for up to two months of the year at a time, and the other a high bench, much wider in extent but only a meter higher up the bank which is flooded once every four years on average for a few days at a time – which bench best represents the active floodplain and hence the best value of bankfull discharge? Conundrums of this sort cannot be answered by defaulting to a single geometric or morphological criterion; rather, the fluvial geomorphologist must integrate multiple criteria, including the values of bankfull determined for adjacent cross-sections, to come up with a probable best value or range of values (Johnson and Heil, 1996) that typify bankfull discharge for that cross-section. The consideration of the frequency of bankfull discharge is ultimately conditioned by the statistical method used to analyse the time series of flood frequencies. There are several methods for evaluating the frequency of one flood out of a series of floods, and the choice of statistical distribution used once a particular method has been specified will also affect the results. Floods occur over time, but the recording of these events is broken down into discrete intervals with a representative discharge characterizing each interval. In small streams where high-intensity rainfall is the dominant flood generating mechanism, a single flood may last a  14  few hours only; in large rivers where floods result from snowmelt, a single flood may last for a month, with several local peaks and recessions during that month. The methods used to evaluate flood frequency have differing abilities to represent these events. The three main methods for splitting a sequence of discharges into a number of discrete flood events to be analyzed are the annual maximum series, the partial duration series, and the flow duration method. In the literature, reported frequencies of bankfull discharge are most often presented with reference to the annual maximum flood series, less frequently with respect to the partial duration series or flow duration method. In the annual maximum flood series, the largest discharge occurring in each year is recorded and forms the sequence of floods that are analyzed (Langbein, 1949). Permutations of the annual maximum flood series lie in the choice of instantaneous maximum or daily maximum values, and the choice of annual interval (calendar year or water year). Each of these variations offers specific strengths and weaknesses with respect to characterization of flood frequency, but the choice of method used is ultimately affected by data availability. If only daily maximum discharge values have been recorded, for instance, instantaneous values cannot be used. Likewise, the choice of water year rather than calendar year may be appropriate for streams that have a distinct wet season and dry season, where the largest flood is always associated with the wet season, but for streams in which floods can occur at any time of year as a result of multiple flood generation mechanisms, the utility of water year over calendar year may be minimal. In any case, when using the annual maximum flood series, only the largest flood per year is considered, so some information is lost. In the rare case where two unusually large floods, perhaps a snowmelt flood followed by a purely rainfall-driven flood, occur in one calendar year, for instance, only one would be considered,  15  even if both corresponded in magnitude to Q10 or larger events. Likewise, information about flood duration is lost when the annual maximum series is used – a flood that lasts 10 days and one that lasts 1 day, which peak at the same maximum value, are treated identically. In the partial duration flood series, some of the problems with the annual maximum series are addressed. A flood threshold value is identified, and all floods that exceed that threshold are then recorded and analyzed (Langbein, 1949). Difficulties arise in the choice of threshold value and in the choice of methods of separating floods with multiple peaks into individual events. If a suspected value of bankfull discharge is known, the expected bankfull value can be used as the threshold, but the problems with counting events remain: is a flood that exceeds bankfull for two days, drops below it for two days, then rises and exceeds it again for another two days, properly counted as one or two floods for the purposes of analysis? The flood duration method is a different method which analyses the continuous or semi-continuous record of streamflow over the entire interval of flow monitoring. Rarely is streamflow gauging truly continuous, but the use of even daily values over a long period of record creates a very large amount of data. From that data, the frequency of exceedance of a discharge of any given magnitude may be determined. Using the flow duration method may create difficulties with estimation of frequency for flows larger than the largest flow on record, particularly in cases of mixed distributions where the largest peak flows (tail of the distribution) follow a different distribution than do the daily flows and more frequent peak flows. Partial duration series and annual maximum series estimates of flood frequency converge above a return period of approximately 5-10 years, given a long enough record of  16  flows, and return periods calculated from either method may also be converted to or from expected frequencies of occurrence based on the flow duration method (Langbein, 1949; Chow, 1950 Takeuchi, 1984). For instance, a daily maximum peak with a return period of 2 years on the annual maximum series, or 1 year on the partial duration series, might be expected to be equalled or exceeded with a frequency of between 0.14% and 0.27% on the flow duration series. A flow duration series composed of daily discharges (24-hour average values) may not be suitable for estimating the duration of exceedance of floods that exceed the daily maximum for some fraction of the day, such as occurs in small drainage basins where the instantaneous peak discharge may be two or more times greater than the daily discharge. The issue of the frequency of bankfull discharge is confounded by the consideration of the frequency of the mean annual flood. For instance, Wolman and Leopold (1957) reported that most streams in their pioneering study had a frequency of bankfull discharge of between 1 and 2 years, based on the annual maximum series, with a mean of approximately 1.5 years, but did not specifically report the frequency distribution used to fit the flood series in order to extract the bankfull frequency; reference is made to Langbein (1949), so the distribution is probably the generalized extreme value (GEV) distribution. Later, Dury (1961, 1977) and Roberts (1989) identify this bankfull flood as specifically being the mean annual flood, based on the GEV; however, Dury (1961) gives the expected frequency of the mean annual flood as 2.33 years whereas Dury (1977) gives the expected frequency of the mean annual flood as 1.58 years. The discrepancy between Dury’s two values arises from the assumptions made in fitting the observed record of floods to a frequency distribution. Rao (1981) provides examples of the way in which various skews in data can change the mean  17  annual flood estimate relative to the mode. Under a normal distribution, and with a long enough period of record, in general the mean annual flood will have a return period of two years. Skewed data, short periods of record, and the choice of two, three, or four-parameter frequency distributions to model expected flood frequencies can result in estimates of the frequency of the mean annual flood that vary from the Q2 to either lower or higher return periods. A final consideration when discussing the frequency of bankfull discharge is that frequency and magnitude are related in a nonlinear fashion. Wolman and Leopold (1957), for instance, arrived at the estimation of the average return period of bankfull discharge as approximately 1.5 years by noting that most of the bankfull discharges they recorded were in the interval of 1 to 2 years on the annual maximum flood series. Based on the annual maximum series, however, there is no event more frequent than 1 year, and in fact the most frequent bankfull discharges reported in Wolman and Leopold (1957) have a frequency of 1.01 year. The 1.01 year event on the annual maximum series corresponds to a very frequent flood on the partial duration series – one with a return period significantly less than 0.4 year (Langbein, 1949; Takeuchi, 1984). Because of the nonlinear relation between frequency and magnitude, if two floods in the same reach of the same river are to be averaged, one with a frequency of 1.01 year on the annual maximum series and the other with a frequency of 2 years, averaging by frequency and averaging by magnitude will give different results. Specifically, the magnitude determined by averaging frequencies will not equal the frequency determined by averaging magnitudes. Thus the average value of Q1.5 for the frequency of bankfull discharge determined by Wolman and Leopold (1957) does not necessarily correspond meaningfully to the expected magnitude of the average bankfull flood, which  18  might be larger or smaller than the mean annual flood, because it averages together floods somewhat smaller than the mean annual together with floods very much smaller. Even if there is uncertainty over the average return period of bankfull discharge, it may still be determined from reviewing the literature that the frequency of bankfull discharge varies from stream to stream and within a drainage basin. In studies in which the bankfull stage is measured in the same way and the frequency of bankfull flow is also evaluated in the same fashion for all the surveyed streams, the potential errors and uncertainties discussed above with respect to estimation of bankfull frequency and magnitude are lessened, at least with respect to relative frequency and magnitude of the streams in the study; comparisons between studies which use different criteria will still be affected. Factors which have been suggested to affect the frequency of bankfull discharge include: •  climate or hydroclimatic regime (Wolman and Gerson, 1978; Castro and Jackson, 2001)  •  elevation (Jarrett, 1990)  •  stream channel or drainage basin overall gradient (Wolman and Leopold, 1957; Kilpatrick and Barnes, 1964; Williams, 1978)  •  drainage basin size (Wolman and Leopold, 1957; Wolman and Gerson, 1978; Andrews, 1984; Petit and Pauquet, 1997; Castro and Jackson, 2001; Dodov and Foufoula-Georgiou, 2005)  •  sediment supply and size/nature of material forming bed and banks (Nolan et al, 1987; Millar, 2005; Nanson et al, 2010)  •  lithology (Harvey 1968; Petit and Pauquet, 1997)  •  anthropogenic development (Whelan, 2000)  19  In general, the reported trends of these factors in the cited studies is in agreement, such that relatively drier, steeper, undeveloped, higher-elevation streams, with coarse gravel banks, and a low (or episodic) sediment supply, will have less frequent bankfull discharge. The situation for drainage basin size is less clear, with some studies finding a general trend to less frequent bankfull discharge as basin size decreases (e.g. Wolman and Gerson, 1978; Castro and Jackson, 2001) and others the opposite trend (e.g. Dodov and Foufoula-Georgiou, 2005). The reported specific effects of lithology are indirect; evidently what matters is not the exact type of rock, but whether the processes resulting in peak flows are driven by increased baseflow or quickflow (Harvey 1968; Petit and Pauquet, 1997), with baseflowpeak rivers having less frequent bankfull discharges than those with quickflow-driven peaks. Underlying many of these disparate effects is a common theme, sometimes evident and sometimes not: the variability of flooding affects the frequency of bankfull discharge. Expressed statistically, both the variability (coefficient of variation, Cv) (Wolman and Gerson, 1978) and skewness (Andrews, 1980) of the flow frequency distributions (annual maximum flood series) have been correlated with the frequency of bankfull discharge. The skewness and the Cv measure different aspects of the same physical relation, namely the size and frequency of the largest floods relative to the mean annual flood. In streams where the rare, large floods are proportionally larger in relation to the mean annual flood, more work is done by the less frequent events, and so the frequency of the conceptual channel-forming event is likely to be less and its magnitude greater. If the conceptual measure (channelforming discharge) is a larger, less frequent event, then the measurable discharges (bankfull and effective) which approximate the channel-forming discharge should also be larger and less frequent events.  20  Ultimately the bankfull discharge only has meaning value for streams that have banks. In a purely bedrock stream, with both channel bed and banks composed of rock, there are not really any banks in the sense of bankfull discharge; vegetation trimlines may be present, corresponding to the age of various high water flows that scoured the sidewalls of the channel, but these are not banks. Small or gravel-bedded mountain streams, especially formerly glaciated or headwater drainages, can display a range of channel types and stream morphologies ranging from colluvial, through various types of alluvial morphology, to bedrock, depending on sediment supply (Montgomery and Buffington, 1997). At both extremes of sediment supply (colluvial and bedrock channel types), alluvial processes play little role in stream bank formation. The exact definition of stream type can be as difficult as the definition of bankfull stage, particularly in the case of bedrock streams. The narrow definition of bedrock channels includes only those streams with discontinuous sediment present and extensive bedrock in the channel bed and banks (Gilbert, 1877; Howard, 1980, Whipple, 2004) while more recent inclusive definitions of bedrock channels (e.g. Turowski et al, 2008) include all streams that cannot substantially deepen, widen or shift their bed without eroding bedrock, even if their bed and banks are primarily composed of alluvial materials. The presence of paraglacial sediments (Church and Ryder, 1972; Church and Slaymaker, 1989), such as till and glaciofluvial materials, may further complicate this transition, especially in streams that developed under a glaciofluvial regime but that have transitioned to a purely fluvial regime (Miller, 1958). Limits to alluvial character have also been proposed that are independent of channel material type, relating instead to the resistance of channel materials to hydraulic forces (Wohl, 2004). Therefore, except for the limited extreme cases of purely rock-banked bedrock streams and purely colluvial streams flowing  21  over immobile boulders, there should exist a transitional regime in small mountain streams where alluvial character becomes less well-developed, and correspondingly banks become intermittent, overbank alluvial sediment reduces in volume, and the significance of the bankfull discharge as a unique value decreases. While bankfull discharge in many large or lowland streams may approximate the mean annual flood, in small or mountain gravel-bed streams, as the limits of alluvial character are approached, the stream banks and associated bankfull discharge should not abruptly disappear. Rather, it is reasonable to expect that as the alluvial nature of the stream decreases, the relative magnitude of the bankfull discharge should increase and the frequency should correspondingly decrease, until some limit is reached where there are no longer any signs of alluvial character to the stream and the banks are correspondingly non-existent. In the case of bedrock channels where the closest approximation of a top of bank is far removed from the active channel, the effective magnitude of the corresponding discharge is likely to be so large as to be meaningless and the corresponding return period, so rare as to approximate the age of the stream. 2.3 Effective Discharge The effective discharge is most commonly defined, after Wolman and Miller (1960) as the discharge (or narrow range of discharges) which transports the most sediment over time in the stream, and hence does the most (or greatest proportion of) work in the stream channel. The effective discharge is more closely conflated with the channel-forming discharge than is the bankfull discharge in the hydrologic literature because of the work of Wolman and Miller. Therefore, it is critical to differentiate between papers which use the effective discharge to mean “the discharge which transports the most sediment in a stream”  22  and those which use the effective discharge to mean “the channel-forming discharge”. The linkage between effective discharge and sediment transport is clear and measurable. Another, conceptual step is required to go on from measurements or predictions of the discharge or range of discharges that transport the most sediment in a stream to also identify that discharge as the discharge which is responsible for the form of the channel, requiring a fundamental assumption which may or may not be testable. Wolman and Miller (1960) applied the concept of flow frequency and magnitude to the formation and maintenance of natural, self-formed alluvial stream channels. Leopold and Wolman (1957) had already examined the bankfull dimensions of natural channels and found substantial similarities across a range of scales and hydroclimates. The Wolman-Miller model extended this consideration of similarity to the frequency range of streamflows and the sediment transport potential for each. Effectiveness of sediment transport was formally defined as the product of flow frequency and sediment transport by flows of that frequency. Very low flows occur frequently but transport little or no sediment. Very large flows transport large quantities of sediment, but occur rarely, so their contribution over time to the total sediment transported is low. The most effective discharge, that which transports the most sediment when averaged over time, is an intermediate flow. For rivers transporting mostly fine sediment, and assuming a lognormal flow frequency distribution, Wolman and Miller (1960) found that the most effective discharge was approximately equal to the bankfull discharge, and that both corresponded to the mean annual flood. Therefore, they concluded that the effective discharge was the channel-forming discharge. At a basic conceptual level, the idea that the bankfull discharge should be the most effective discharge is initially attractive and reasonable, since it explains the geomorphic  23  significance of the banks (the channel is adjusted to the most effective discharge). There are likewise differences between the sediment-transporting capabilities of within-channel and overbank flow arising from velocity changes and the presence of vegetation outside the channel banks which suggest that the bankfull flow should be the most effective flow, at least for suspended sediment. However, when the subject is examined more closely, enough discrepancies between the magnitude and frequency of bankfull and effective discharge are documented to indicate caution in assuming that the effective discharge must always be the channel-forming discharge. Subsequent to the paper of Wolman and Miller (1960) which dealt with effective discharge mostly at a conceptual level, as part of a general consideration of magnitude and frequency in geomorphic processes, numerous investigations have considered effective discharge (Benson and Thomas, 1966; Kennedy, 1972; Dury, 1973; Pickup and Warner, 1976; Wolman and Gerson, 1978; Baker, 1977; Pickup and Reiger, 1979; Andrews, 1980; Newson, 1980; Beven, 1981; Gupta, 1983; Nolan et al., 1987; Baker and Pickup, 1987; Ashmore and Day, 1988; Carling, 1988; Kochel, 1988; Pitlick, 1988; Miller, 1990; Nash, 1994; Costa and O’Connor, 1995; Andrews and Nankervis, 1995; Rosgen, 1996; Hey, 1997; Sichingabula, 1999; Biedenharn et al., 2001; Soar and Thorne, 2001; Emmett and Wolman, 2001; Vogel et al., 2003; Torizzo and Pitlick, 2004; Goodwin, 2004; Crowder and Knapp, 2005; Lenzi et al., 2006; Barry et al. 2008; Bernedo, 2009). Results from these investigations have been mixed. Some (for instance Dury, 1973; Andrews, 1980; Rosgen, 1996; Torizzo and Pitlick, 2004) have reached conclusions similar to those of Wolman and Miller (1960), finding a relatively close agreement between effective and bankfull discharge, while others (for instance Benson and Thomas 1966; Pickup and Warner 1976; Baker, 1977; Nolan et al.,  24  1987; Ashmore and Day 1988; Pitlick, 1988; Nash 1994) have documented significant divergence. Variation in the frequency and magnitude of effective discharge has been ascribed to basin size (Wolman and Gerson, 1978), shape of flow regime hydrograph (Andrews, 1980), channel bed and bank material (Nolan et al., 1987; Emmett and Wolman, 2001), bioclimate and riparian vegetation (Wolman and Gerson, 1978; Werrity, 1997), channel gradient (Pitlick, 1988) and other factors. As well, both the sequence of occurrence (Pickup and Reiger, 1979; Beven, 1981) and duration (Costa and O’Connor, 1995) of geomorphically effective floods have been demonstrated as being of perhaps greater importance than simply their magnitude. Some of the basic assumptions of Wolman and Miller (1960) have also been challenged. Nash (1994) noted many streams with bimodal, multimodal, and heavily skewed flood probability distribution functions, as well as PDFs best fit to distributions other than the lognormal. These variable flood frequency distributions, when combined with sediment rating curves, produce effectiveness relations of considerable variety, ranging from streams where the smallest and most frequent flows are the most effective to streams where the effectiveness increases steadily with increasing discharge. Likewise, cases where more than one peak in the effective discharge relation occurs have been reported (Nash, 1994). The Wolman-Miller model explicitly identifies the effective discharge as the channel-forming discharge; in cases where multiple discharges of equal or approximately equal effectiveness occur, the question of whether one can be identified as the channel-forming discharge at the expense of another is less clear. Defining a geomorphic meaning for the remaining effective peaks is also problematic without additional interpretation.  25  The conflation of the discharge that is the most effective (transports the greatest amount of sediment over time) with the channel-forming discharge has also been criticized by Vogel et al. (2003) who point out that only in the case of a symmetrical effectiveness distribution is the peak of the effectiveness relation also indicative of the median discharge for sediment transport. Vogel et al. (2003) argue that for asymmetric (skewed) effectiveness distributions, much of the total sediment transported may be moved by rare events. This leads Vogel et al. (2003) to question the geomorphic significance of the effective discharge, preferring instead to use the half-load discharge (the discharge above which half the total load is transported) to summarize the effectiveness of rare floods. Vogel et al. (2003) note that for many streams the half-load discharge is significantly higher than the effective discharge. However, the half-load discharge itself may not have a unique geomorphic significance in the same fashion the discharge associated with the peak of an effectiveness diagram has been considered to have. The original Wolman and Miller (1960) model dealt only with suspended sediment transport, primarily because of the availability of such data and the relative paucity of bedload transport measurements. Subsequent approaches incorporating bedload transport have considered bedload to represent a fixed or negligible proportion of the suspended sediment transport (e.g. Benson and Thomas, 1966), or have used an empirical relation for bedload transport (e.g. Andrews, 1980) or a nominally physically based bedload transport equation (e.g. Torizzo and Pitlick, 2004). In the latter two approaches, if bedload transport rate does not vary identically with suspended transport rate (in the case of power law relations, if the exponents of the bedload and suspended load rating curves are not identical), two separate effective discharges occur, one for suspended sediment transport and one for  26  bedload transport. This situation (two separate effective discharges, one for suspended load and one for bedload) is observed when suspended load and bedload transport are sampled relatively directly (e.g. Lenzi et al., 2006). An overall effective discharge based on total load (summed suspended load and bedload) will also occur, but given the relative magnitudes of volumes of sediment moved as bedload and suspended load, is likely to correspond closely to the suspended load effective discharge. Geomorphically, the concept of multiple effective discharges has been interpreted as representing respectively channel-maintaining (Andrews and Nankervis, 1995) and channelchanging (Wolman and Gerson, 1978; Phillips, 2002) discharges, with more frequent flows transporting enough sediment to maintain the form of the channel and prevent aggradation, degradation, or vegetative encroachment, while rarer floods that exceed some geomorphic threshold (Schumm, 1979; Church, 2002) cause significant channel changes including realignment of banks, redistribution of bedforms and woody debris jams (Grant et al., 1990), and potentially lateral channel motions. Although appealing conceptually, the concept of channel-maintaining and channel-changing discharges may not directly explain many streams with multiple effective discharges. Specifically, in gravel-bed, gravel-bank rivers, 90% or more of the total sediment transported may be suspended sediment, but the suspended sediment load moves through the channel without significantly affecting the channel; only the movement of bed-material load is of significance for potentially channelforming discharges (Leopold, 1992; Nash, 1994; Andrews and Nankervis, 1995; Phillips, 2002; Lenzi et al., 2006). Thus, although all streams have an effective discharge for suspended sediment transport, in gravel bed channels with gravel banks the suspended effective discharge is argued to have little geomorphic meaning beyond an association with  27  denudation rates and landscape evolution. Intuitively, purely bedrock or colluvial (i.e. nonalluvial) channels must experience a similar condition, in which there are specific suspended and bedload effective discharges but neither is related to channel-forming processes because the channel is not self-formed. A further source of difficulty in the determination of effective discharge comes from the underlying methodology. Sediment transport is most commonly measured for suspended sediment. Physical measurement of bedload transport of gravels is limited by the capacity of the sampling apparatus (Church et al., 1987; Bunte et al., 2004), by the frequency of sampling (Emmett, 1980), and by the possibility of damage to sampling equipment and personnel during large or sediment-concentrated floods capable of moving cobble to boulder sized material. Therefore, the most common approach to estimating sediment transport has been to use equations, either empirically or theoretically based. One advantage to this approach is that the resultant determination of effective discharge is not particularly sensitive to the choice of transport equation used (Pickup and Warner, 1976; Barry et al., 2008). The area in which the greatest source of potential error arises in the determination of effective discharge is in the method of classifying or representing the frequency of discharge (Sichingabula, 1999; Crowder and Knapp, 2005; Lenzi et al., 2006). Discharge is a continuous variable, but is rarely sampled continuously; often averages over time, such as daily mean discharges, are all the data that is available. The original Wolman and Miller (1960) paper stated that effective discharge could be determined mathematically, but did not present a specific example. Subsequently, three main approaches have been used to render discharge frequency computationally tractable when determining effective discharge (Biedenharn et al., 2001; Soar and Thorne, 2001; Bernedo, 2009):  28  a) group the range of reported discharges into a fixed number of classes (for example, 25) of equal width in magnitude, and determine the sediment moved by the mean discharge for each class interval (Benson and Thomas, 1966; Pickup and Warner, 1976). This is the approach most commonly used in the literature; b) represent the observed frequency of flow with a theoretical flow frequency distribution which approximates it, such as a lognormal distribution, and determine the continuous product of the sediment transport function and the flow frequency distribution; this is the approach conceptually presented by Wolman and Miller (1960) and has been used by Nash (1994) and others; c) determine the sediment transported by each reported instance of flow, and determine the effective discharge from the steepest point of the cumulative sediment transport curve (Emmett and Wolman, 2001). In practice, however, a smoothing algorithm is used, which renders the result similar to using a fixed-class size, with the width of the smoothing interval replacing the class width (Emmett and Wolman, 2001; Soar and Thorne, 2001). The class-width method has been harshly criticized (Sichingabula, 1999; Crowder and Knapp, 2005; Lenzi et al, 2006) because the result obtained is highly dependent on the number of classes used, and in fact applied references such as Biedenharn et al. (2000) and Soar and Thorne (2001) explicitly recommend using an iterative method of varying the class width in the case of unusual, outlying, or unexpected results until something acceptably close to the preconceived result is obtained. These same criticisms apply to method c) above because the smoothing interval stands in for the fixed class width.  29  Method b) has its own difficulties, in that the observed distribution of flows will be more well or less well fit by whatever flow frequency distribution is used, hence the estimate of the effective discharge becomes dependent on the goodness of fit of the chosen flow frequency distribution (FFD) (Goodwin, 2004). This becomes a particular problem for bimodal, multimodal, heavy-tailed or heavily skewed flow frequencies, especially where distributions are chosen for reasons of computational convenience or legal specification rather than for sound underlying theoretical reasons (Alila and Mtiraoui, 2002; Kidson and Richards, 2005). Likewise, a chosen flow frequency distribution may provide a systematic bias, as in the case where the mode of the observed flow frequencies is well fit by a simple FFD such as a lognormal distribution but the tails are poorly fit. When all of the above factors are integrated into an expectation about the frequency of effective discharge and the relation of the effective discharge to the channel-forming discharge, the situation is most clear for relatively large rivers in equilibrium with sediment supply in humid temperate drainages, where, in the absence of anomalous conditions, one would generally expect the effective discharge to approximate the mean annual flood and the bankfull discharge, and thus provide a reasonable estimate of the channel-forming discharge. With respect to small gravel-bed mountain streams, however, there is not necessarily any reason to expect that effective discharge will approximate bankfull discharge, nor to expect that either value will have a magnitude close to either the mean annual flood or the channelforming discharge. Instead, as drainage area decreases, the size of in-channel sediment increases, and the self-formed alluvial character of the stream decreases, it seems likely that effective discharge will decrease in utility as a surrogate for the channel-forming discharge.  30  Chapter 3: Study Area and Selection of Streams The study area for this thesis consisted of southern British Columbia (Figure 3-1). Southern British Columbia was extensively glaciated during the Pleistocene (Clague and James, 2002), with the Cordilleran Ice Sheet flowing in multiple directions (but generally southwards in southern BC) from centers in the Coast and Columbia Mountains. Ice depths at glacial maximum were generally greater than 2,000 meters, resulting in substantial postglacial isostatic adjustment (Clague and James, 2002). The legacy of glacial erosion resulted in derangement of preglacial valley long profiles, leaving hanging valleys which disrupt the idealized downstream pattern of river process domains (Brardinoni and Hassan, 2006), with streams in hanging valleys developing towards fully alluvial character but then steepening and entering bedrock reaches downstream, or in the most extreme cases beginning as low-gradient headwater streams on upland plateaus and then steepening progressively downslope as they approach an incised major valley. The legacy of Pleistocene glaciation continues to affect streams in the study area through the mechanism of paraglacial sedimentation (Church and Slaymaker, 1989; Collins and Montgomery, 2011), with small, mountain and upland streams currently degrading through deposits of glacial sediments while large rivers are aggrading as reworked glacial sediments are transported to valley bottom. Active stream piracy of headwaters is another mechanism which has caused extensive changes in the directions of flow and rates of erosion in some streams in southern British Columbia (Riedel et al., 2007). Transitional alluvial conditions should therefore be present in small mountain streams in southern British Columbia. “Small” in the sense of “small streams” is difficult to define in a way that can be determined in advance from maps and GIS data, rather than after observing the stream in  31  Figure 3-1: Hydrologic regions and locations of sampled watersheds for study. 32  question in the field. For example, Wohl’s (2004) definition of the transition between welland poorly-developed alluvial character references Q2 streampower and particle size; neither parameter can be measured from a map. A simple means of identifying candidate streams for this study was to use a scale parameter: the chosen scale parameter had to be large enough to include some streams with well-developed alluvial character as well as others with transitional alluvial character or poorly-developed alluvial characteristics. The entire Water Survey of Canada database of streamflow records for southern British Columbia was examined, using a scale parameter of 100 km² as a spatial cutoff and a minimum 20 years of recorded streamflow data. Every unregulated drainage basin with area <100 km² and >20 years of recorded Water Survey of Canada streamflow information in southern British Columbia was considered. The drainage basin size limit of 100 km² was chosen as a relatively arbitrary upper limit for size. The 100 km² upper threshold for drainage basin size provided a practical means of defining a number of candidate drainage basins for analysis, some of which would have well-developed alluvial characteristics, some poorly-developed or non-alluvial, and some transitional alluvial streams. The selection of 20 years as a minimum length of record was intended to provide confidence in the estimates of infrequent return periods as it was anticipated that in some of the studied streams either bankfull or effective discharge would be rare events. In addition to the <100 km² drainage basins described above, Harris Creek was included as a study site because it is a well-studied gravel bed stream with known hydrologic characteristics (Hassan and Church, 2000, 2001). Harris Creek at the studied reach has a drainage area of 225 km², making it useful as a reference site at the upper size bound of the studied streams.  33  Streams with regulated flow were excluded because the process of flow regulation could introduce a confounding factor to the consideration of channel-forming discharges, if the discharges that formed the channel were reduced or modified by the regulation of flow. Central and Northern British Columbia were excluded from the study area because there are very few drainage basins there that are both small and have long-term records of streamflow. The remaining study area covers approximately 157,000 km², primarily in the Columbia River and Fraser River drainage basins. Every candidate stream was visited and inspected. Candidate streams were excluded if they were obviously subject to non-fluvial process such as debris flow, if they were purely bedrock streams without obvious banks, or if their channels or stream banks had been subject to significant anthropogenic alteration that rendered it difficult to determine the natural location of the streambank. Several candidate streams were excluded because forest road deactivation had made former gauging locations too difficult to access. One stream was also excluded because severe windthrow of beetle-killed timber that had almost completely obscured both banks for several hundred meters upstream of the gauging station. Finally, one stream was excluded because of very high runoff resulting from glacial melt during a summer heat wave, which prevented safe access to the channel. After completing this exclusion process, thirty-six stream reaches in thirty-four drainage basins remained. The location of the selected stream gauges is shown in Figure 3-1, A summary of the geomorphic and hydrologic characteristics of the drainage basins is presented in Table 3-1. The drainage basins group naturally into three hydroclimatic zones: the Coast and Mountains region, the Thompson-Okanagan Region, and the Kootenay-Columbia region. Each hydroclimatic zone is internally homogenous and distinct with respect to flood  34  generation mechanisms, and the ranges of magnitude of annual precipitation within each hydroclimatic zone also varies notably (Beckers et al, 2002). Table 3-1: Watershed name, WSC gauge number, years of record, and hydroclimatic region. Watershed Code 08NP004 08NK026 08NF005 08NF005 08NH115 08NH084 08NB016 08NJ061 08NE114 08NJ130 08NJ129 08NE087 08NM142 08NM137 08NM242 08NM241 08NM240 08LE077 08NM173 08LG064 08NM134 08LG016 08LG056 08LF081 08LF084 08MF062 08MH150 08MH076 08MH141 08GA065 08HB069 08HB048 EASTCRK  Watershed Name Cabin Creek near the mouth Hosmer Creek above diversions Albert River at 1310 m contour above jam Albert River at 1310 m contour below jam Sullivan Creek near canyon Arrow Creek near Erickson Split Creek at the mouth Redfish Creek near Harrop Hidden Creek near the mouth Anderson Creek near Nelson Fell Creek near Nelson Deer Creek at Deer Park Coldstream Creek above municipal intake Daves Creek near Rutland Dennis Creek near 1780 m contour Two Forty-one Creek near Penticton Two Forty Creek near Penticton Corning Creek near Squilax Greata Creek near the mouth Beak Creek at the mouth Camp Creek at mouth near Thirsk Pennask Creek near Quilchena Guichon Creek above Tunkwa Lake diversion Ambusten Creek near the mouth Anderson Creek above diversions Harris Creek near Lumby Coquihalla River below Needle Creek Norrish Creek above Rose Creek Kanaka Creek near Webster Corners Coquitlam River above Coquitlam Lake Noons Creek at Meridian Substation Road Carnation Creek at 150 m contour Carnation Creek at the mouth East Creek in Malcolm Knapp Research Forest  Area (km²) 93.2 6.4 69.7 69.7 6.22 78.7 81.3 26.2 56.7 9.07 4.4 80.5 58.5 31.1 3.73 4.5 5 26.2 40.7 85 33.9 87 78.2 32.9 31.9 221 85.5 79 47.7 54.7 1.6 2.53 10.1 1.21  Record Length (y) 28 24 26 26 42 45 32 35 33 43 29 47 36 21 20 21 20 22 34 19 39 69 37 22 20 9 34 19 43 23 16 19 32 30  Hydroclimatic Region KC KC KC KC KC KC KC KC KC KC KC KC TO TO TO TO TO TO TO TO TO TO TO TO TO TO CM CM CM CM CM CM CM CM  35  The Coast and Mountains (CM) hydroclimatic region is located in southwest British Columbia, within the ranges of the Coast Mountains, Cascade Mountains, and Vancouver Island Mountains, and is typified by high annual precipitation (2000 to 4000 mm/yr for the studied drainages), with relatively dry summers and wet winters. Generally 60-75% of annual precipitation in the region falls during the period between October and March. Valley-bottom temperatures generally remain above freezing throughout the winter months. Although mountain snowpacks can be quite deep, and snowmelt produces a pronounced spring freshet, for the smaller drainage basins such as those in this study the largest peak flows are usually produced by either pure rainfall, or rain-on-snow events, and tend to occur in late fall or winter (Melone, 1985). The Thompson-Okanagan (TO) region lies on British Columbia’s Interior Plateau, in the south-central part of the province. Annual precipitation is low, ranging from 300mm or less in valley bottoms to 1000mm at plateau highpoints. Summer temperatures can be very high; winter temperatures are usually below freezing for several months of the year. Precipitation amounts are distributed approximately evenly throughout the year. Peak flows are predominantly caused by snowmelt, but rain-on-snow events also occur, and in the driest parts of the region for streams of the scale studied the largest floods on record result from summer convective storms. (Waylen and Woo, 1982, 1984). The Kootenay-Columbia (KC) region is located in southeast British Columbia, within the ranges of the Columbia Mountains and Rocky Mountains. Annual precipitation ranges from 500mm in valley bottoms to 2500mm on west-facing mountain slopes in this hydroclimatic region, but is generally 1000-2500mm in the studied drainage basins. Winter valley-bottom temperatures are generally below freezing for several months. The majority of  36  precipitation falls during the winter months. Peak flows are predominantly caused by spring snowmelt, but very rare instances of fall or spring rain-on-snow floods have been observed (Beckers et al, 2002). In each watershed, study reaches were selected as close as possible to the Water Survey of Canada (WSC) stream gauging location. All locations had stream reaches selected that were single-thread, gravel-bed channels with well-defined banks. Stream morphology ranged from riffle-pool through plane-bed and step-pool to cascade-pool in the steepest reaches (Montgomery and Buffington, 1997). Most studied reaches had some associated valley flat adjacent to the banks, but the extent of the valley flats were generally limited; direct inputs of colluvial sediment from adjacent hillslopes to stream channels were possible in almost all the studied reaches, but none of the reaches selected had purely colluvial banks. All of the studied drainage basins had some level of past or present forest harvesting. None were extensively logged (>50% clearcut) at the time of the study, but many had previously been harvested and were forested to some degree with mature or maturing second growth. In the TO and KC regions, many watersheds had some level of mountain pine beetle (Dendroctonus ponderosae) infestation as well (Hélie et al, 2003). Riparian vegetation, including mature trees, was present along the banks of all the studied reaches. The relatively homogenous hydroclimatic regions encompassed by this study do not imply homogeneity in other parameters. Bedrock geology, for instance, is heterogenous throughout the study area, encompassing rocks of a wide range of sedimentary, metamorphic, intrusive and extrusive lithologies, and with ages ranging from Proterozoic to Miocene. Similar heterogeneity is present in the physiography: the Coast and Mountains and Kootenay-Columbia regions are generally rugged and mountainous while the Thompson-  37  Okanagan region has upland plateaux and broad, incised valley bottoms. The physiographies of the sampled basins within each region are not always representative of the overall physiographic character of the regions: for instance, several of the drainage basins within the Coast and Mountains physiographic region are located along the boundary between the uplifted Coast Mountains and the adjacent sedimentary basin of the Georgia Lowland. Similarly, the Coast Mountains, the Columbia Mountains, and the Rocky Mountains are all typified by the presence of large alpine glaciers and neves, but only one of the streams in this study, Split Creek in the Kootenay-Columbia region, had a significant glacier in its headwaters. In the results and discussion that follow, the relative homogeneity of the hydroclimatic regions can be contrasted to the heterogeneity of geology and physiography within and between each region. The design of this study anticipated that observed regional differences would be attributable to the contrasting hydroclimatic regimes between regions, while some of the scatter in the data within each region would be attributable to the variability in geology and physiography as well as to the differences in elevation and precipitation in the studied streams within each region.  38  Chapter 4: Methods and Data Collection 4.1 Field Measurements 4.1.1  Field Site Selection Fieldwork was mostly conducted during periods of low flow in the summers of 2006  and 2007, when unrestricted access to the stream channels could be obtained. East Creek was surveyed in both September and December 2006, and Carnation Creek fieldwork was conducted in summer 2009. At each site, a representative stream reach of approximately 100200m length (generally 10-30 times the bankfull channel width) was selected, with relatively homogenous width, gradient, sediment characteristics and riparian vegetation. In each watershed, study reaches were selected as close as possible to the Water Survey of Canada (WSC) stream gauging location. All selected reaches were alluvial single-thread, gravel-bed channels with well-defined banks. In cases where the WSC gauge was at a location without natural banks, such as at a location with a bridge over the stream and riprapped banks, the closest homogenous reach with natural banks was chosen instead. In the cases where suitable reaches existed both upstream and downstream of the unsuitable gauged reach, the upstream reaches were chosen over the downstream to avoid potential effects from the artificial banks upstream in the unsuitable gauged reaches. The gauge was located within the surveyed reach in 17 of the 36 stream reaches studied. 4.1.2  Measurements From nine to 16 cross-sections on each stream reach were surveyed using a total  station (Dutnell, 2000; McCandless and Everett, 2002). In one case, on the Albert River, where a massive LWD jam was present in the channel, and where the channel reaches above and below the jam seemed visibly different, twenty-four cross-sections were surveyed,  39  twelve above the jam and twelve below. The cross-sections were generally located at least one channel width apart, spaced more or less evenly. The need to obtain clear lines of sight meant that perfectly even spacing could not be obtained. Small obstacles such as brush were cut or pruned to obtain a clear line of sight, but large trees were not cut down to this end. Survey data was recorded digitally for export to GIS. Position and elevation data was recorded for surveyed points on each cross section and the survey tied to a control point located using GPS. The number of surveyed points on each cross-section varied with the size of the stream; the smallest streams had less than 20 points per cross-section while the largest had more than 70. The number and spacing of survey points was sufficient in all cases to capture the microtopography of the stream bed and banks as well as the overbank topography. In general, the cross-sections were surveyed well beyond the channel banks. For most of the smaller streams this meant that the entire width of the valley flat was surveyed. For some of the larger streams, especially those with inactive terraces and wide valley flats, it was impossible to survey the entire width of the valley flat because of dense vegetation that prevented line of sight from valley wall to valley wall. In these cases, the survey was extended to the limits of the line of sight away from the channel in both directions. Typically, the survey included one of the valley walls at the edge of the survey, to a height above all alluvial sediment. Identification of channel banks can be a contentious subject, as discussed in Chapter 1. Rather than selecting one indicator of bankfull and applying that to every cross-section, even where it might be locally inappropriate, the surveys were designed to capture multiple definitions of bankfull (Johnson and Heil, 1996; Radecki-Pawlik, 2002, Navratil et al.; 2006).  40  In practice, geomorphic and morphologic indicators such as prominent slope breaks at the top of bank, or the elevation of point bars, can be distinguished directly from the surveyed crosssection and require no special geomorphic training to recognize. In the field, however, care was also taken to identify on the surveyed cross-section the most apparent locations of the left and right bank on each cross section. Where two locations appeared equally valid as the top of the bank, or where the bank seemed indefinite and upper and lower limits of a potential bank needed to be defined, more than one potential location for a single bank were noted. The water surface elevation was also noted on the channel cross-sections so that it might be possible to check measured stage against recorded discharge for those watersheds with active stream gauges. In East Creek, surveyed channel cross-sections with bank locations identified were provided by Andre Zimmermann and Joshua Calkins. In Carnation Creek, twenty years of sequentially surveyed channel cross-sections with bank locations identified were provided by Dan Hogan. Measurement of in-channel sediment within each reach was subdivided into two components; measurement of surface sediment and measurement of subsurface sediment. For both surface and subsurface sediment measurement and sampling, the methodology and rationale of Bunte and Abt (2001) was used as a standard reference. The primary method used to measure surface sediment was photogrammetric, using the Digital Gravelometer photo analysis program (Graham et al., 2005). To collect surface sediment photos for analysis, in addition to a Canon A620 digital camera, a frame and tripod were used. The frame had approximate dimensions of 1.0m by 0.7m, and was placed on exposed sediment. The tripod was used to position the camera above the frame at a sufficient  41  distance to image the entire frame. Before taking each sediment photograph, the area within the frame was hand cleared of leaf litter, twigs and other debris that might obscure the sediment. A nylon tarpaulin was used to shade the camera, frame and tripod from direct sunlight, thus preventing unwanted shadow/sunlight contrasts in the photo that could be mistaken for grain boundaries. Between three and forty photographs were taken in each studied reach. The original objective was to take between seven and ten photographs in each reach. In the studied reaches with narrow channels and no bars, so little exposed sediment was present that only three or four locations could be found in which a sizeable enough accumulation of sediment existed to place the frame. In other studied reaches where significant expanses of exposed sediment were present, many more than ten sediment photos were taken, including some duplicate photos of the same sediment, to allow for analysis of factors such as microscale variability and effects of photo processing. In reaches where significant expanses of exposed sediment were present (primarily the larger watersheds in the study), a pebble count of 100 stones from a 10mx10m grid, 5mx20m grid, or linear count was made in addition to the photographs (Wolman, 1954). The purpose of making the Wolman pebble count was to allow comparison of results from the Graham et al. (2005) photogrammetric method to the older Wolman method. Wolman counts were not made in some of the studied reaches, however, because the areas of exposed sediment were too small to permit a sufficient grid for counting stones to be laid out. The original intent of the study was to sample subsurface sediment once per studied reach, but the same grain size constraints which affected the measurement of surface sediment also affected the measurement of subsurface sediment. In the reaches with little  42  sediment exposed, it was not possible to measure subsurface sediment because too little sediment existed to volumetrically sample in a statistically significant manner. In these reaches, the largest stones in the surface sediment were usually in direct contact with bedrock or with immobile lag beneath the channel, meaning that an armour layer of coarser surface sediment overlying finer subsurface sediment did not exist and thus obviating the need for a subsurface sediment sample. At the other end of the scale, in some reaches with abundant exposed surface sediment, subsurface sediment samples were not taken because of the size of the largest exposed clasts. Bunte and Abt (2001) recommend that the largest clast in a volumetric sample comprise no more than 0.5% of the total mass, or 3% of the mass of a subsample if subsamples are taken and combined. In some of the studied reaches, the largest stones had diameters > 2.0m, and a significant fraction of stones present had diameters greater than 256mm. This posed two difficulties: firstly, samples containing these stones would have to be unfeasibly large in order to meet the Bunte and Abt criteria, and secondly, the largest stones were too large and heavy for a two person field crew to move. After eliminating those streams where not enough sediment was present to take a statistically volumetric sample, and those streams where the sediment was too coarse to permit volumetric sampling, approximately one-third of the studied stream reaches remained. In these reaches, volumetric samples were taken. The average mass of the samples was approximately 350-400 kg. Clasts larger than 32mm b-axial diameter were hand sieved and weighed in the field. The remaining sediment was mixed to randomize it and then a subsample of approximately 10kg was taken for laboratory analysis.  43  In addition to volumetric samples of in-channel sediment, the original intent was to also sample the bank material, because it was anticipated that there might be cases where the texture of the bank material and bed material was significantly different, for instance, a gravel bed but sandy banks. In practice, this was only observed in two or three of the studied watersheds. In the remaining watersheds, the bed and bank material appeared to be similar, or the bank material contained significant amounts of non-fluvial material such as bedrock or compact basal till that would have been difficult to sample. Because of the small number of watersheds where bed and bank material appeared to differ, any sample of bank material would not be statistically significant, so bank material samples were not taken. In addition to the surface sediment photographs, between 20 and 100 oblique photographs were taken of each studied reach. The purpose of the oblique photographs was to document bank materials, sediment patchiness, channel pattern, bankfull indicators, channel morphology, and to record any anomalies present in the channel (such as anthropogenic debris, large glacial erratics, or other similar features). In any case where possibly interesting or unusual features occurred in or along the channel, one or more oblique photos were taken. A sketch map was drawn by hand of the studied reach for each watershed. The location of landmarks such as bedrock outcrops, LWD jams, bridges, roads, and stream gauges was indicated on the sketch maps. Each surface sediment photograph and oblique photograph was assigned a number, and the numbers were added to the sketch map in order that the location of a particular photograph could be determined at later dates. The amount of LWD in the channel can be an important control on sediment transport, and adds to flow resistance. The purpose of this LWD survey was to allow  44  comparison of woody debris abundance between studied watersheds. Therefore, the sampling procedure was primarily designed to be internally consistent rather than designed so as to compare results to those of other studies primarily concerned with LWD function. The primary reference used for the methodology of sampling LWD was the TFW Ambient Monitoring Program Manual (1994). Over the length of the studied reach, the total number of pieces of large woody debris was counted and the length and median diameter of each piece were measured. The survey distinguished three types of LWD: logs, rootwads and LWD jams. Logs have a minimum diameter of 0.1m and a minimum length of 1.0m. Rootwads are less than 2m long with a minimum diameter of 0.2m where the stem and roots meet, and the roots must be detached from the ground. Where the definitions of rootwads and logs overlap, pieces meeting both criteria were defined as logs. LWD jams are composed of multiple logs and/or rootwads together with smaller debris. The overall dimensions of a jam (length, width and breadth) and its porosity were measured to determine the volume of the jam. Length was measured with a 30m nylon measuring tape. Diameter was measured at the approximate midpoint of each piece of LWD using a 5m steel tape. Jam porosity was visually estimated to the nearest 10%. The distance of each piece upstream of the start of the channel survey was measured using a hip chain in order to allow evaluation of local variations in LWD density along the studied reaches. Measurements distinguished between LWD which is within the active channel (Zones 1 and 2 of TFW, 1994), and LWD which overhangs the channel but is not within the active channel (Zones 3 and 4 of TFW, 1994). Where LWD was partially buried in bank or channel sediments, only the exposed portion was measured. Likewise, where a large piece of LWD  45  such as a fallen tree occurred with the majority of its length well outside the active channel and only a small portion within, the volumes of the two portions were measured separately. No attempt was made to account for decay classes of material. 4.2 Other Data Collection 4.2.1  Flow Frequency Data In each studied drainage basin, daily average and instantaneous peak flow data over  the period of record were obtained from WSC. Record length at individual gauges ranged from a minimum of twenty years to over sixty years, with the exception of Harris Creek, where there are only nine years of record. In East Creek, thirty years of recorded discharge measurements were provided by Dr. Michael Feller, and digitized from the analog (drum graph) recordings by Joshua Caulkins and Andre Zimmermann (Appendix A) 4.2.2  Physiographic Data From Maps British Columbia government online mapping tools (iMapBC and BC Basemap) were  used to measure physiographic data related to each drainage basin, including area above studied reach, total relief (elevation differential between low point of studied reach and highest poin along watershed divide), and drainage basin length (distance from low point of studied reach to furthest part of drainage area). These measured values were then used to compute additional characteristics, including Melton ruggedness ratio (relief/area0.5), relief ratio (relief/length) and shape parameter (area/length2). 4.2.3  Sediment Source Analysis Two different types of sediment sources were evaluated in each drainage basin to  serve as indexes of sediment supply to the stream channel: landslides and unpaved roads. The total number of landslides in each drainage was counted using the most recent provincial  46  aerial photography at 1:20,000 scale supplemented with high resolution Google™ Earth satellite imagery. The total length of unpaved roads in each drainage area was measured from the 1:20,000 scale TRIM maps. The raw numbers (number of landslides and km of roads) were divided by drainage basin area to create indexes of number of landslides and km of roads per km² of area. The total width of the valley flat (sidewall to sidewall) in the surveyed reach was also measured from the TRIM maps and compared to the surveyed cross-sections. The width of valley flat was of interest as a proxy value for evaluating the connectivity of the adjacent hillslopes to the channel. The presence of upstream lakes and wetlands that could serve as sediment sinks was also noted and quantified. Representative examples of surveyed cross-sections, sediment photos, overview photos showing banks and channels, and rating curves for each studied stream are presented in Appendix B. 4.3 Methods of Analysis 4.3.1  Flow Frequency Analysis The annual maximum series of instantaneous peak flows was extracted from the data  for each stream and analyzed using the Consolidated Frequency Analysis (CFA) program (Pilon and Harvey, 1994). CFA fits six distributions to the data: generalized extreme value, three-parameter lognormal, log-Pearson Type III, Wakeby, Weibull and nonparametric. For each stream, the distribution that best fit the observed data was selected and used to evaluate the return periods for instantaneous peak flow events. For a few of the streams, the total number of years with recorded instantaneous peak flows numbered less than twenty, but the records of instantaneous peaks were supplemented by a longer record of daily maximum flow. For these streams, the dates for which  47  instantaneous and daily maximum floods coincided over the overlapping period of record was used to estimate a ratio of daily maximum to instantaneous peak flow for each of the streams, and this ratio was used to estimate the missing instantaneous peak flow values from the observed daily maximum values. For two streams (Harris Creek and Noons Creek) the WSC gauge was some distance from the surveyed reach – in Harris Creek because the gauge is located on private farm land below the forested watershed, and in Noons Creek because suburban development has obliterated the WSC gauge site and an undisturbed adjacent reach was surveyed. For both of these streams, the ratio of drainage area at the gauged site to drainage area at the surveyed site was calculated, and the observed peak flows scaled by this ratio raised to the 0.75 power (based on a regionalization of peak discharge values for British Columbia by Eaton et al., 2002), in order to estimate the magnitude of peak flows at the surveyed reach relative to those observed at the gauge. 4.3.2  Sediment Photo Processing and Analysis Photos of bed material were analyzed with the Digital Gravelometer program  (Graham et al., 2005). In cases where sediment contained appreciable amounts of granitic or gneissic rock, with a speckled appearance, analysis of unprocessed photos tended to interpret a single large clast as a pile of coarse sand. Therefore, sediment photos were preprocessed with Adobe Photoshop, manually blurring the visible portions of speckled casts into a single homogenous colour. Once preprocessed in this manner, values from the Digital Gravelometer output corresponded reasonably well to the manually collected Wolman counts for reaches in which both measurements were made.  48  Digital Gravelometer has a lower size bound below which it cannot effectively differentiate grain size. For these photos, this corresponded to 8 mm on average. While this reduces its effectiveness in evaluating proportions of sand and fine gravel, the description of the distribution of the coarse fraction of sediment was the primary goal of this research; the effect of the lower truncation on the proportions of the D75, D84 and D90 was minimal. To characterize each surveyed reach, the individual photos were combined for analysis, with the resultant grain size parameters recorded; thus the D50 value for the reach represents a reachaveraged value, while the D100 recorded corresponds to the largest grain in any of the photos. There is likewise a practical upper bound to the Digital Gravelometer analysis imposed by the size of the photo and frame, somewhere >800mm; these large and largely immobile clasts were measurable by Wolman counts and uppermost size classes of the Digital Gravelometer output adjusted accordingly. 4.3.3  Volumetric Sediment Sample Processing The subsurface samples were weighed, dried, reweighed and sieved at 0.5 phi  intervals. Field-measured data for >32mm clasts was added to the sieved fractional weights and the size distribution of the entire sample determined. The proportion of the total sample coarser than 8mm was recorded and the remaining sediment analyzed to determine the proportionate weight and size distribution of each size class in the <8mm sediment. 4.3.4  Survey Data GIS Processing The survey data from the total stations (X and Y coordinates, elevations, and notes  such as bank indicator) was imported into ArcMap 9.2, and a beta version of HEC-GeoRAS 4.2 (USACE, 2008) was used to designate straight cross-sections perpendicular to the direction of flow through the surveyed cross-sections, which were never exactly straight.  49  Corrected cross-sections were analyzed using WinXSPro 3.0 (Hardy et al, 2005). Reachaveraged bank elevation slope (bankfull gradient) was used as channel gradient. Manning’s n was estimated using the Jarrett method (Jarrett, 1984, 1990); some of the streams in this study outside the range of widths and gradients for which Jarrett’s method was originally developed, but are within the range for which subsequent work has demonstrated the Jarrett method to provide a reasonable estimate of n (Marcus et al, 1992; Curran and Wohl, 2003). Estimates of Manning’s n were cross-referenced against WSC stage-discharge curves for the gauging sites to ensure that the values of Manning’s n estimated approximated real conditions; the agreement was generally good over the range of stage-discharge curves developed by WSC, and use of this method allowed estimation of n values for stages not directly measured by WSC. 4.4 Estimation of Bankfull Discharge For each cross-section, using the previously determined reach-averaged slope and Manning’s n values, the range of bankfull discharges was estimated with WinXSPro, using the uppermost and lowermost of the surveyed bankfull indicators as bounding stages. Within the bankfull range, coincident bankfull indicators, as well as graphical indicators such as discontinuities in the slope of the calculated stage-discharge relation from WinXSPro, were used as indicators to evaluate a best estimate of bankfull stage and discharge for each crosssection. The range of bankfull values for each cross-section was then plotted in sequence for each reach and a magnitude of bankfull discharge for the reach determined which best fit the ranges of bankfull estimated over each individual cross-section (Figure 4-1). A detailed example illustrating how the reach-averaged value of bankfull discharge was determined is presented in Appendix C.  50  To reduce the dependence of bankfull discharge on watershed area and provide a dimensionless value for comparison between watersheds, it is useful to scale the observed bankfull discharge against a reference discharge or discharges. Reference discharges used were the two-year peak discharge (Q2) as determined from the best-fit flow frequency analysis of the annual instantaneous maximum series, and the mean annual discharge (Qmad) as determined from the daily discharge series and annual flow statistics. Both of these reference discharges are useful, and each represents a complementary physical parameter. The ratio of Qbf/Q2 is a broad measure of how the bankfull discharge compares to the average peak flow in the watershed, while the Qbf/Qmad ratio is a similarly broad measure of how the bankfull discharge compares to the average flow in the watershed. For the classic Wolman and Miller (1960) model, where the bankfull discharge is approximately equal to the mean annual flood, Qbf/Q2 should be approximately equal to, or less than 1 (depending on whether the magnitude of the mean annual flood is assumed to be equal to Q1.5 or Q2), while the Qbf/Qmad should depend on the watershed size, climate and other factors but should generally be larger than Qbf/Q2 and greater than 1. The flood frequency analysis and the record of daily discharges were used to evaluate the return period of the bankfull discharges compared to the instantaneous peak discharge annual maximum series and the daily discharge record (flow duration method). For the smaller and rainfall-dominated watersheds where the instantaneous peak discharge is significantly larger than the daily discharge value, it would be preferable to use a flow duration series based on a shorter duration than daily values, such as 15-minute intervals; however, this information is not available from WSC. Comparison of the instantaneous peaks  51  against the daily discharge arge series provides an acceptable method of overcoming o this limitation.  Figure 4-1: Bankfull range and nd be best estimates by cross-section for Greata Creek. Best-fit range of bankfull discharge e va value for entire channel, 1.1 to 1.3 m³/s, falls within estimat imated range of bankfull determined for 7 out of 12 cross-sections tions.  iment Transport and Bedload Effective Dischar scharge 4.5 Estimation of Sedimen To determine thee effec effective discharge it is necessary to estimate the rate r of sediment transport by the range of ob observed and predicted discharges and then en determine de which interval of discharges transpo ransports the most sediment. The methods of estimating estim sediment  52  transport and of grouping the range of observed discharges into intervals are essentially independent. 4.5.1  Estimation of Sediment Transport The output from WinXSPro includes depth, hydraulic radius, Manning’s n, velocity,  discharge and total shear stress for a range of stages calculated in 0.01m increments from 0 to the limits of the surveyed channel, above bankfull stage. The first stage in calculating sediment transport from the WinXSPro output is to evaluate the fraction of total shear stress available for sediment transport. Einstein (1950) provided a method for determining the proportion of grain stress from total shear stress. The following method provides an approximation of the logarithmic resistance equation (Whiting and Dietrich, 1980; Pitlick, 1992; Torizzo and Pitlick, 2004): U/(ghʹS)1/2 = 8.1(hʹ/Ks)1/6  (4-1)  where U = mean velocity (ms-1), g = gravity (ms-2), hʹ = grain-effective depth (m), S = channel gradient (m/m), and Ks = roughness parameter (m). Various authors have proposed differing methods of evaluating Ks; for the streams in this study, which are generally coarsegravel bedded and have a wide range of sediment sizes, Ks = 3D84 was used (Torizzo and Pitlick, 2004). Because of the approximation of the logarithmic equation, the terms can be rearranged and solved for hʹ without requiring an iterative solution (unlike the original Einstein (1950) equation): hʹ = (1/gS)3/4(U/8.1)3/2(Ks)1/4  (4-2)  The proportion of total shear stress available for sediment transport is then hʹ/h where h is mean depth, ie. τg = τhʹ/h where τ = total shear stress (kgm-1s-2) and τg = grain shear stress (kgm-1s-2).  53  To calculate expected sediment transport from grain shear stress, the two-fraction surface-based model of Wilcock and Kenworthy (2002) was used. The advantages of this model are firstly that it is surface-based and secondly that the main requirement for size characterization is in the coarser (gravel) class. This works well with the limitations of the photogrammetric sediment characterization method, because of the Digital Gravelometer’s lower size cutoff. The model requires as input the specification of the two size fractions, designated as “sand” and “gravel” although there is no requirement for the boundary between the two to correspond to the sedimentological definition of sand. Because the Digital Gravelometer lower cutoff approximated 8mm, this was used as the boundary between the two size fractions. Once the two size fractions are specified, the Wilcock-Kenworthy model requires the specification of Fg and Fs (fraction of gravel and sand), Ds (median diameter of sand fraction, m) and Dg (median diameter of gravel fraction, m), where Fg, Fs, Ds and Dg are measured using the grain-by-number method. Digital Gravelometer does not provide a size distribution for sediment below its lower cutoff, but it does indicate what proportion of total sediment is above (coarser than) the cutoff; this value was used for Fg and thus Fs = (1-Fg). The size distribution of the remaining (gravel fraction) sediment was then recalculated using the gravelometer and Dg set equal to the D50 value of this truncated sample. An estimation of Ds is also necessary in order to use the Wilcock and Kenworthy (2002) transport equation, and estimation of the size fractions of the sediment <8mm in size could not be completed with Digital Gravelometer.. Therefore the size distributions of the volumetric subsurface sediment samples, specifically the <8mm fraction, were examined. All subsurface samples had D50 (<8mm) in the range 1.41 – 2.83mm with a mean at 2.00mm. Wilcock and Kenworthy (2002) assert that sand-fraction D50 is more variable over time than  54  is gravel-fraction D50 in response to inputs of fine sediment to the stream. Therefore it seemed reasonable to set Ds as 2.00 mm for all sites as an acceptable approximation of longterm average value. Once Fs, Fg, Ds and Dg are known, the Wilcock-Kenworthy model requires calculation of fractional reference shear stresses, using fractional Shields numbers (Shields, 1936): τ*ri = (τ*ri)1 + [(τ*ri)0 -(τ*ri)1]e-14Fs  (4-3)  where (τ*rg)0 at Fs = 0 is 0.035, (τ*rg)1 at Fs = 1 is 0.01, (τ*rs)1 at Fs = 1 is 0.065, and (τ*rs)0 at Fs = 0 is equal to (τ*rg)0(Dg/Ds). The fractional Shields numbers are used to evaluate the fractional reference shear stresses as follows: τ*ri = τri/((s-1)ρgDi)  (4-4)  where s is equal to the ratio of grain density to water density (grain density is assumed to be the density of quartz, so s-1 = 1.65.) Then dimensionless transport rate Wi* is calculated for each size fraction and used to evaluate bedload transport for each size fraction: Wi* = 0.002ϕ7.5 for ϕ < ϕʹ  (4-5)  Wi* = 115[1-(0.923/ϕ0.25)]4.5 for ϕ ≥ ϕʹ  (4-6)  where ϕ = τ/τri and ϕʹ = 1.27. The dimensionless transport rate is then used to solve for the fractional bedload transport rate: Wi* = (s-1)gQbi/Fiu*3  (4-7)  where Qbi is fractional unit bedload transport rate of fraction i (sand or gravel) (m3s-1) and u* is shear velocity (ms-1, equal to [τ/ρ]0.5). In practice, the equations were solved for (τ*rs)0, then hʹ, τg, ϕs and ϕg, then Ws* and Wg*, then Qbs and Qbg.  55  For each stage increment, Qbs and Qbg were calculated and multiplied by channel width to obtain total sediment output from WinXSPro. Total sediment discharge was plotted as a function of stream discharge (Qsed vs. Q) for each cross-section. Equations of the powerlaw form (Qsed = αQβ) best fit the resultant curves. From this plot two power-law relations and two discharge thresholds were obtained for each cross-section: below the first threshold discharge, no transport; above the first but below the second threshold discharge, generally, partial transport with only sand fraction mobile; above the second threshold discharge, both sand and gravel mobile. These sediment transport rating curves were then plotted against the flow duration curve to estimate sediment transport for each discharge class (Biedenharn et al, 2001). 4.5.2  Division of Recorded Discharge into Discharge Classes and Qeff Estimation Dividing the flow duration curve into discharge classes is critical for evaluating  effective discharge, with effective discharge magnitude for each class dependant on class width and median discharge value for the class. The nature of the daily discharge data are such that the data must be divided into flow classes: daily discharge values from WSC are recorded to three significant figures, and for most streams cover two or more orders of magnitude (such as from 0.01 to 10 m³/s). If the discharge data are not divided into flow classes, the frequency distribution of discharges shows peaks at orders of magnitude (Figure 4-2), with one or two recorded days of discharge at each of 9.96, 9.97, 9.98, and 9.99 m³/s and then a spike to ten or more days of discharge at 10.0 m³/s. In essence, the problem is that because the raw data come with significant figures, there are already discharge classes present of unequal width (0.01 m³/s between 1 and 9.99 m³/s, and 0.1 m³/s from 10.0 m³/s on up). To deal with this problem, but not artificially reduce the apparent frequency of the  56  largest flows, flow class width widths were set equal to the width of the largest st class clas present in the data, rather than defining ng a sset number (such as 25 or 50) of flow classes, classe an approach consistent with the event-based based flow class methodology of Sichingabula (1999). (1999 This resulted in a total number of flow w cla classes in the hundreds or thousands for each ach stream. st For each discharge class, the average erage discharge magnitude for the class was determined, determ and the sediment transport ratingg curv curve used to evaluate the sediment transportt for that th class by the class-averaged discharge. Eac ach classes’ sediment transport was multiplied by the frequency of that discharge class (numbe umber of days with discharge divided by total number numbe of days in the length of record); the effective fective discharge was then the discharge class with the th largest value in the product of sediment ent tr transport and frequency, or the peak in the plot pl of sediment transport times frequencyy vs. ddischarge (Figure 4-3).  Figure 4-2: Histogram of discharg charge frequency for Cabin Creek. Artificial peaks in frequency at 1.0 and 10 m³/s are a result of changes in discharge class width size resulting from WSC’s reporting of data with three significant nt fig figures.  57  Figure 4-3: Sample effectiveness ness p plot for Redfish Creek, cross-section 11. Threshold discharges for are 0.75 m³/s ³/s for sand transport and 2.40 m³/s for gravel. Effective discharg harge is 2.64 m³/s. Halfload discharge is 4.45 m³/s.  4.6 Estimation of Size of M Mobile Sediment Fraction One limitation off the W Wilcock and Kenworthy (2002) bedload transport transp function is that it provides an estimate ate of the total volume of sediment moving within hin each ea size fraction for a specified discharge, e, but does not specifically estimate the particle size of the mobile sediment relative to the particl particle size of the bed sediment at a specified dischar ischarge. To evaluate the size of the mobile sedimen diment, it is possible to estimate the D50 of the transpo ransported material if the Shields transport relation tion is solved using a specified critical Shields number, numbe θc: D50m = τg/(s-1)ρgθc  (4-8)  where D50m is the median ian gr grain size (m) of the transported sediment. The T result of the equation is thus dependent ent on the specified value of θc. Values of 0.045 5 and 0.060 are often used, after Shields (1936), 6), to represent expected values of θc for well- and poorly-sorted 58  gravel beds; measurements from gravel-bed channels, however, often support higher values of θc, with a range between 0.052 and 0.086 reported for relatively coarse-bedded mountain channels (Buffington and Montgomery, 1997), and values as high as 0.1 not unknown. These high values for θc may represent whole-channel conditions; scaling the reported θc values by the hʹ/h relation (from Eq. 4-2) to determine the proportion of grain-effective dimensionless shear stress capable of transporting sediment may be necessary (Buffington and Montgomery, 1997). To constrain the possible values of θc for this study, D50m was computed for the entire range of potential discharges, from zero flow to the limits of the surveyed cross-sections, for the three stream reaches in the study with extensive previous measurements of sediment transport (Harris Creek, Carnation Creek near the mouth, and East Creek), for four potential θc values: 0.045, 0.052, 0.060, and 0.086. The resulting discharge vs. D50m relations for each θc value were then compared to the published size of sediment transported values from the literature for the studied streams. The low end of the range of θc values (0.045-0.052) provided reasonable approximations of published size of sediment transported values. Therefore, a value of θc = 0.045 was used as a reasonable reference value for all the streams in the survey and the D50m calculated for Q2 and Q20 return period floods for each stream as well as the D50m for Qbf and Qeff for each stream. The Q2, Q20, Qbf and Qeff D50m were then compared to the bed surface D50 for each studied reach to estimate the size of mobile sediment relative to bed surface sediment under conditions of common and relatively infrequent flooding, bankfull, and effective discharge. The purpose of the comparisons was to evaluate the relative mobility of the bed surface. As the calculated D50m approximates the bed D50 it indicates that all bed sediment is in general motion, whereas a D50m much smaller  59  than the bed material D50 suggests finer material moving over and past immobile coarse stones. 4.7 Calculation of Half-load Discharge Vogel et al (2003) distinguish between the theoretical half-load discharge calculated from a distribution fitted to the observed daily discharge record, and the empirical half-load discharge calculated directly from the observed daily discharge record. For the streams in this study, the empirical half-load discharge for bedload (Qh) was calculated from the daily discharge records and the sediment rating curves by calculating the total bedload moved by each discharge class over the period of record, summing the load over all the classes to determine the total load moved by the stream over the period of record, and then dividing the load moved by each class by the total load to find the fractional proportion of load moved by each class. The fractional proportions were then cumulatively summed and the Qh determined from the discharge that moved the 50th percentile of the total load. Although the computation was made using the class-grouped discharges for ease of computation, the determination of Qh is not sensitive to the number or width of the classes. 4.8 Statistical Significance and Physical Meaning In this thesis, developing an understanding of the underlying processes by determining the factors contributing to variations in the magnitude and frequency of bankfull and effective discharge is considered to be of primary importance. Tests of statistical significance are not ignored: where appropriate, r-squared values and p-values are presented when values are graphed. However, the overreliance on tests of statistical significance, and on hypothesis testing as the only means of answering questions (Klemes, 1974; Schmidt, 1996; Elliot and Brook, 2007; Alila et al., 2009) has led some modern science into  60  philosophical errors or dead ends. In the past, overreliance on hypothesis testing has sometimes led to the exclusion of meaningful data as outliers or to the wrongful rejection of physically plausible hypotheses. In this thesis, the presentation of results and subsequent discussion and conclusions are based on considerations of the underlying meaning of observed correlations rather than on acceptance or rejection of an arbitrarily defined statistical hypothesis. 4.9 Sources of Error This analysis of the frequency and magnitude of effective and bankfull discharge has been computationally based by design, since actual measurements of bedload sediment transport in the studied streams have generally been unavailable. As previously noted, for the three stream reaches (Harris Creek, East Creek, and Carnation Creek at the mouth) for which previously published estimates of bankfull magnitude and thresholds for sand and gravel transport were available, the values estimated in this study were generally relatively close to the previously published values. For many of the studied streams there are no such published values available for comparison. There are several stages in the processes of estimation of bankfull and effective discharges in which errors could be introduced and affect the final estimates; each is discussed in turn. 4.9.1 4.9.1.1  Sources of Error in Estimation of Bankfull Discharge Errors in Field Measurement Potential errors in field measurement of bankfull discharge essentially consist of the  potential for accidental or systematic misidentification of indicators of the top of bank. Accidental misidentifications of the top of bank in any one cross-section, while possible, are unlikely to have had a noticeable effect on the estimation of reach-averaged bankfull  61  discharge because of the deliberate identification of a bankfull range on each bank and the determination of bankfull value for the stream reach as the value which best fit the greatest number of individual cross-sections. Thus, the potential effect of individual outliers arising from misidentification of top of bank in any one cross-section was minimized. Systematic misidentifications of the top of bank could have two forms. Firstly, choice of an inappropriate indicator (such as misidentifying an inactive terrace top as an active top of bank) could result in a widened range of bankfull for multiple cross-sections and hence distort the range of bankfull discharges estimated for each cross-section and ultimately the overall estimate of bankfull discharge for the stream reach. Secondly, and related to the first, an indicator could be chosen appropriately but could still be wrong. For instance, in a stream which had recently experienced a 1 in 100 year flood, and associated channel change, but which had not yet had time to experience modification by smaller subsequent floods the height of the top of bank and bars might be related only to the magnitude of the rare large flood and not to the full range of floods, thus distorting upwards the estimate of bankfull magnitude. For either situation, errors which result in estimation of a wrongfully lower bankfull stage, such as through the identification of a frequently-inundated bed surface as a bar top, could also occur. The potential for systematic errors of bankfull identification were minimized through examination of the record of past hydrologic events and through the solicitation of independent estimates of the location of bankfull stage, both in the field and later, from inspection of the oblique channel photographs. Therefore, the probability that systematic errors in the identification of bankfull stage affected the magnitudes of bankfull discharge estimated in this study is small.  62  4.9.1.2  Estimation of Manning’s n Value and Reach-averaged Channel Slope. Converting the surveyed cross-sections with bankfull indicators into estimates of  bankfull discharge (or other discharges, for the calculation of sediment transport) required the specification of channel slope and Manning’s n values. Channel slope was specified as the average change in bank elevation from the top to the bottom cross-section of the reach. Manning’s n was estimated by the Jarrett (1984, 1990) method. Because of the nature of the stream channels, the local channel slope between any two or three adjacent cross-sections can be steeper or gentler than the reach-averaged gradient. Jarrett’s method estimates Manning’s n as a function of the gradient, so this can result in an effective n value at any one crosssection that is higher or lower than that estimated by the Jarrett method using the reachaverage gradient. Furthermore, the Jarrett method used to estimate n is insensitive to local factors such as woody debris in the channel or to specific characteristics of overbank vegetation which can increase or decrease flow resistance. Thus, it is almost certain that some of the variation in estimates of bankfull discharge from one cross-section to the next within a single stream reach is a result of this uncaptured variability in slope and n. Using the best-fit value of bankfull discharge for the whole reach accounts for and removes some of the variability in the bankfull discharges calculated for each individual cross-section. The ability of the Jarrett (1984) estimate of Manning’s n to capture actual channel conditions was evaluated and calibrated against the Water Survey of Canada stage-discharge relations for each studied stream, but for streamflows outside the range of calibrated WSC discharges, or for those streams where the gauge is located in a location with artificial banks and the studied reach was the nearest adjacent reach with natural banks, there remains an unavoidable uncertainty.  63  4.9.1.3  Estimation of Sediment Transport, Effective Discharge and Half-load  Discharge The potential uncertainty in Manning’s n also applies to the determination of sediment transport and consequently the half-load and effective discharge, because the same stage-discharge relations were constructed for each cross-section using WinXSPro. The estimation of sediment transport also had additional sources of potential error arising from the construction of sediment transport rating curves for each cross-section. 4.9.1.4  Estimation of Grain Shear Stress The partitioning of total shear stress in order to evaluate the fraction, grain shear  stress, available for sediment transport was presented as Eq. 4-1 and Eq. 4-2 in Chapter 4. The logarithmic approximation of the Einstein (1950) equation rather than an iterative solution for each cross-section, the choice of a grain size and a constant to multiply it by (3D84) to represent the roughness parameter Ks, and the use of reach-averaged channel gradient rather than cross-section to cross-section local gradient are all potential sources of error in this process. However, they are also consistent between cross-sections and streams for this study, so if the results are in error they are consistently so. 4.9.2  Estimation of Sediment Size The estimation of sediment size represents one of the greatest potential sources of  error because of the dependence of the sediment transport equations on the proportions of sand-and gravel-sized sediment (Fs, Fg) and the characteristic sizes of each fraction (Ds, Dg). The use of Digital Gravelometer is arguably less subjective than the Wolman pebble count method in characterizing the sediment present in the stream bed, and allowed multiple photographs to be taken of the bed material of each stream. Information on the sediment size  64  distribution from individual photographs was pooled to evaluate sediment characteristics for each stream reach. However, this pooling results in a loss of information about the variability in sediment size distribution between cross-sections. In the stream bed, during low-flow conditions, a riffle may have coarser sediment than a pool. No attempt was made to account for this spatial variation when calculating sediment transport in order to evaluate effective discharge and half-load discharge. Both the Wolman method and the Digital Gravelometer analysis have difficulty in characterizing the absolute size and size distribution of very large stones that form the largest clasts in the channel. For streams like Coquitlam River, where the bed sediment was generally small, this is probably not a significant source of error; for streams like Norrish Creek, where boulders up to approximately 8m in diameter were present in the channel, it is likely that even the Digital Gravelometer results manually adjusted by the results from the Wolman pebble counts underestimated the size fractions at the coarse end of the grain size distribution. While these very large stones appear to remain immobile under almost all discharges, and so the range of sediment evaluated by the modified Digital Gravelometer results may reasonably approximate the range of mobile sediment in the channel, the very large clasts present may exert effects (hiding of sediment, additional form drag, mobility at certain very high discharges) which were not captured. 4.9.3  Estimation of Half-load and Effective Discharge While previous studies have found (Pickup and Warner, 1976; Barry et al, 2008) that  the calculation of effective discharge is not particularly sensitive to the choice of sediment transport equation, the transport equations utilized in those studies were generally not of the two-fraction Wilcock and Kenworthy (2002) type. The two-fraction sediment transport  65  equation models large increases in the volume of sediment transported when the sand transport and gravel transport thresholds are exceeded, and hence the determination of effective discharge is sensitive to the location of these thresholds, particularly the threshold for gravel transport, in cases where the threshold is close to the calculated effective discharge. In cases where the calculated effective discharge is much greater than the threshold for gravel transport, for instance, there is much less sensitivity to the gravel transport threshold than when the effective discharge is close to the threshold. The calculated half-load discharge is generally less significantly affected by the location of the thresholds, but can be affected, particularly by large changes in the location of the gravel transport threshold. As an example of the sensitivity of the calculation of the effective and half-load discharges to the grain size distribution and sand/gravel proportions, Norrish Creek is analyzed with altered sediment size distribution parameters. For Norrish Creek, the actual measured and calculated parameters are as follows: Fs =0.1225, Fg = 0.8775, Ds = 2mm, Dg = 69.7mm. Sand transport discharge threshold is 0.65 m3/s and sand transport sediment rating curve is Qsed = 0.0049Q1.2105; gravel transport threshold is 3.75 m3/s and combined gravel and sand sediment transport rating curve is Qsed= 0.039Q1.1189. Calculated Qeff is 13 m3/s and Qh is 15 m3/s. To illustrate the effects of the sensitivity of the transport equations to changes in the sediment size parameters, altered calculations were made supposing that Fs =0.1, Fg = 0.9. Ds = 2mm, Dg = 100mm, which is within the range of the data measured by Digital Gravelometer from photos of Norrish Creek albeit at the coarser end of the scale. These changes result in estimates of a new sand transport threshold of 2.52 m3/s and a new gravel transport threshold of 9.5 m3/s. The new sediment rating curve equations are respectively Qsed  66  = 0.0023Q1.3502 for sand transport and Qsed = 0.0321Q1.1438 for combined sand and gravel transport. Calculated Qeff after the changes are made is still 13 m3/s, but calculated Qh increases to 20.6 m3/s. Qeff did not change because the calculated Qeff magnitude is well outside the range affected by the increased gravel transport threshold. The increase in magnitude of Qh from 15 m3/s to 20.6 m3/s corresponds to a decrease in frequency (of equaled or exceeded discharge) from 46 to 26 days/year on average. 4.9.4  Estimation of Frequency of Bankfull, Effective and Half-load Discharge There is an uncertainty in the estimation of the frequency of a discharge of interest  (bankfull, effective, or half-load) from its magnitude that depends on the length of record at the WSC gauge and the goodness of fit of the observed record of streamflows to the best theoretical distribution. For instance, if the length of record at the adjacent stream gauge is 20 years, the Q2 flood based on the annual maximum series has a magnitude of 5.0 m3/s, and the largest flood on record has a magnitude of 12 m3/s, there will be a much greater degree of confidence associated with estimating the frequency for a discharge of 7.0 m3/s than there will be for a discharge of 20 m3/s. Because of the relative magnitudes of the various bankfull, effective and half-load discharges determined in this study, this uncertainty is more of a problem for the estimated bankfull discharges, some of which were well outside the range of magnitude of observed floods, than it is for the calculated effective (or half-load) discharges, which were generally smaller and more frequent events. The nonlinear relations between sediment size and effective discharge magnitude, and between discharge magnitude and frequency, also create the possibility for relatively small errors or uncertainties to translate into larger errors during the process of calculating the effective discharge frequency, especially for rare large-magnitude events. For instance, a  67  difference of 20mm in the estimated D50 of the gravel fraction might result in a difference of 2 m3/s in the estimated magnitude of the effective discharge, which might result in a difference of decades in the estimated frequency of the effective discharge. As noted above, most of the effective discharges observed in this study were frequent events, for which this uncertainty is minimized.  68  Chapter 5: Results This chapter presents results from the analyses reported in Chapter 4. Results pertaining to bankfull discharge are presented first, followed by results pertaining to effective discharge, then results pertaining to half-load discharge. Factors which affect or which explain some of the observed variation in the frequency and magnitude of bankfull, effective, and half-load discharge are listed in turn. Discussion of the meaning of these relations is generally presented in Chapter 6, although some discussion of interrelation between factors is presented in this chapter in order to aid in interpretation of the results. 5.1 Bankfull Discharge The observed ranges of magnitude and frequency of bankfull discharge for the studied streams are highly variable. In the following section, that variability is first described, and then the factors which contribute to and explain some of the variability in bankfull frequency and magnitude are presented. Measured and calculated hydrologic characteristics for the studied streams and drainage basins are presented in Tables 5-1 and 5-2. Measured parameters are those which were measured either directly in the field or from maps of each drainage basin. Measured parameters (Table 5-1) include area, relief, length, Melton ratio, shape parameter, relief ratio, bankfull channel width, channel gradient, LWD density (expressed as m³/m of channel length and m³/m² of channel area across flow) and channel sediment surface D50 and D84. Calculated or estimated parameters are those that were determined from the Water Survey of Canada discharge data and from the surveyed cross-sections. Calculated or estimated parameters (Table 5-2) include estimated Q2, estimated Qbf magnitude and frequency, Qbf/Q2 and Qbf/Qmad ratios, and estimated total stream power Ω at Q2 reference  69  Table 5-1: Measured parameters by watershed. Watershed Code 08NP004 08NK026 08NF005 08NF005 08NH115 08NH084 08NB016 08NJ061 08NE114 08NJ130 08NJ129 08NE087 08NM142 08NM137 08NM242 08NM241 08NM240 08LE077 08NM173 08LG064 08NM134 08LG016 08LG056 08LF081 08LF084 08MF062 08MH150 08MH076 08MH141 08GA065 08HB069 08HB048 EASTCRK  Watershed Name  Cabin Creek Hosmer Creek Albert River above jam Albert River below jam Sullivan Creek Arrow Creek Split Creek Redfish Creek Hidden Creek Anderson Creek (Nelson) Fell Creek Deer Creek Coldstream Creek Daves Creek Dennis Creek Two Forty-one Creek Two Forty Creek Corning Creek Greata Creek Beak Creek CampCreek Pennask Creek Guichon Creek Ambusten Creek Anderson Creek (Hat) Harris Creek near Lumby Coquihalla River Norrish Creek Kanaka Creek Coquitlam River Noons Creek (Scott Cr.) Carnation Creek Upper Carnation Creek Mouth East Creek MKRF  Area (km²)  Relief (km)  Length (km)  Melton Ratio  Width (m)  Hydraulic mean depth (m)  Channel Gradient (m/m)  LWD (m3/m )  D50 (mm)  D84 (mm)  93.2 6.4 69.7 69.7 6.22 78.7 81.3 26.2 56.7 9.07 4.4 80.5 58.5 31.1 3.73 4.5 5 26.2 40.7 85 33.9 87 78.2 32.9 31.9 221 85.5 79 47.7 54.7 1.6 2.53 10.1 1.21  1.08 1.14 1.93 1.93 1.40 1.43 1.89 1.75 1.16 1.32 1.30 1.77 1.01 0.42 0.38 0.43 0.43 1.45 0.73 0.69 0.93 0.55 0.82 0.90 0.98 1.42 1.35 1.11 1.05 1.63 0.51 0.77 0.92 0.10  15.3 4.08 13.3 13.3 4.69 17.1 15 9.17 7.15 5.05 4.62 14.3 11.2 5.66 2.47 3.42 3.22 11.4 7.35 15.4 8.68 13.1 12.1 8.08 6.55 21.7 12.1 11.6 12.1 10.5 2.48 2.96 6.55 1.1  0.112 0.452 0.231 0.231 0.561 0.162 0.209 0.342 0.154 0.438 0.620 0.197 0.132 0.075 0.197 0.203 0.192 0.283 0.114 0.075 0.159 0.059 0.093 0.157 0.173 0.095 0.146 0.125 0.152 0.220 0.403 0.484 0.289 0.091  16.54 5.04 16.68 15.62 4.79 15.22 17.74 11.36 12.56 4.97 4.97 13.04 8.12 5.19 5.31 2.72 2.93 7.71 4.19 11.45 5.57 9.20 4.36 3.85 4.67 12.60 30.06 48.34 27.36 26.18 4.95 11.60 20.10 4.53  0.90 0.81 0.57 0.76 0.39 0.89 0.35 0.81 0.79 0.36 0.47 0.78 0.45 0.54 0.36 0.30 0.34 0.52 0.38 0.72 0.35 0.48 0.38 0.35 0.56 0.77 0.79 1.37 0.73 1.06 0.56 0.62 0.96 0.42  0.0174 0.0849 0.0117 0.0046 0.170 0.0278 0.0358 0.072 0.0441 0.158 0.184 0.0255 0.0347 0.0721 0.0916 0.0345 0.0453 0.0501 0.0238 0.0271 0.0191 0.00777 0.00426 0.0676 0.0688 0.0135 0.0117 0.0312 0.0101 0.00711 0.0598 0.0190 0.0122 0.0157  0.018 0.035 0.654 0.922 0.214 0.11 1.25 0.008 0.015 0.065 0.058 1.02 0.329 0.137 0.152 0.247 0.146 0.183 0.201 1.05 0.05 0.069 0.353 0.076 0.232 0.088 0.415 0.389 0.199 0.406 0.072 0.030 0.015 0.018  77.33 38.76 22.14 41.53 40.66 51.74 28.33 80.34 118.02 53.18 59.66 56.21 31.89 60.40 119.99 53.73 80.93 61.97 25.59 41.25 63.58 27.65 39.51 14.78 54.01 55.01 62.77 69.70 88.99 28.64 38.99 40.24 23.29 44.61  176.0 113.0 40.71 91.73 99.40 110.39 78.95 163.19 236.76 124.7 138.89 124.17 74.77 156.75 202.58 132.84 166.55 160.58 54.08 127.1 138.06 50.64 134.61 32.33 186.71 105.58 135.54 185.96 195.93 46.27 88.18 150.03 48.26 82.08  70  Table 5-2: Estimated and computed parameters for each drainage basin. Watershed Code 08NP004 08NK026 08NF005 08NF005 08NH115 08NH084 08NB016 08NJ061 08NE114 08NJ130 08NJ129 08NE087 08NM142 08NM137 08NM242 08NM241 08NM240 08LE077 08NM173 08LG064 08NM134 08LG016 08LG056 08LF081 08LF084 08MF062 08MH150 08MH076 08MH141 08GA065 08HB069 08HB048 EASTCRK  Watershed Name Cabin Creek Hosmer Creek Albert River above jam Albert River below jam Sullivan Creek Arrow Creek Split Creek Redfish Creek Hidden Creek Anderson Creek (Nelson) Fell Creek Deer Creek Coldstream Creek Daves Creek Dennis Creek Two Forty-one Creek Two Forty Creek Corning Creek Greata Creek Beak Creek CampCreek Pennask Creek Guichon Creek Ambusten Creek Anderson Creek (Hat) Harris Creek near Lumby Coquihalla River Norrish Creek Kanaka Creek Coquitlam River Noons Creek (Scott Cr.) Carnation Creek Upper Carnation Creek Mouth East Creek MKRF  Area (km²)  Q2 (m3/s)  Q100 (m3/s)  Qbf (m3/s)  Qmad (m3/s)  Qbf Return Pd (y)  Qbf/Q2  Qbf/Qmad  Q2 Stream Power (kg m/s3)  93.2 6.4 69.7 69.7 6.22 78.7 81.3 26.2 56.7 9.07 4.4 80.5 58.5 31.1 3.73 4.5 5 26.2 40.7 85 33.9 87 78.2 32.9 31.9 221 85.5 79 47.7 54.7 1.6 2.53 10.1 1.21  22.5 1.25 15.2 15.2 0.43 17.5 10.7 8.99 13.7 0.86 0.45 7.27 1.79 1.42 1.05 1.13 1.06 5.31 0.52 5.50 1.42 9.21 1.10 0.15 1.05 15.6 25.6 101 61.6 130 5.11 7.70 28.1 2.91  44.0 6.01 41.5 41.5 1.70 33.3 20.9 15.6 42.0 3.60 1.07 15.1 6.80 3.34 2.23 4.19 2.30 14.5 2.85 20.9 3.70 25.8 3.19 1.23 5.10 24.0 70.8 243 215 210 11.6 38.7 75.5 5.66  27.6 6.63 11.3 21.4 2.44 21.0 7.16 15.6 18.1 2.89 2.32 12.2 4.61 1.25 1.55 0.685 1.29 3.60 1.13 8.03 1.98 5.30 1.19 1.34 2.57 14.2 32.7 142 31.0 65.1 3.03 9.27 17.7 1.77  1.93 0.12 1.59 1.59 0.06 1.72 1.66 0.86 1.61 0.11 0.06 0.88 0.25 0.12 0.05 0.06 0.06 0.37 0.08 0.14 0.47 0.15 0.73 0.13 0.02 0.09 1.20 3.29 7.90 2.73 6.58 0.25 0.22 0.81  5.1 154 1.25 7 200 3.5 1.1 100 4 35 500 21 21 1.9 10 1.15 3.5 1.13 5.5 3.6 3.8 1.2 2.75 200 12 1.5 3.3 8 1.05 1.01 1.1 2 1.1 1.25  1.227 5.304 0.743 1.408 5.674 1.200 0.669 1.735 1.321 3.360 5.156 1.678 2.575 0.880 1.476 0.606 1.217 0.678 2.173 1.460 1.394 0.575 1.082 8.933 2.448 0.910 1.277 1.406 0.503 0.501 0.593 1.204 0.577 0.680  14.3 54.3 7.10 13.5 40.7 12.2 4.30 18.1 11.2 26.0 36.3 13.8 18.1 10.8 29.8 12.5 21.5 9.70 14.0 16.9 13.2 7.30 9.20 60.9 28.9 11.8 9.90 18.0 11.4 9.90 12.3 34.7 19.9 16.5  3837 1040 1743 685.2 716.4 4768 3754 6343 5921 1331 811 1817 608.7 1003 942.6 382.1 470.6 2607 121.3 1461 265.8 701.3 45.92 99.37 708.0 2064 2935 30880 6067 9058 29950 1434 3359 447.7  71  discharge (kg m s-3, calculated as ρgQS, where ρ= density of water, 1000 kg m-3; g = gravity, 9.8 m s-2, Q = Q2 discharge in m3s-1 and S= channel gradient in m/m). The most important result, and therefore the result presented first, is the comparison of calculated bankfull discharge values from this study to the values of bankfull discharge previously determined for these streams. This result is a direct test of the accuracy of the methods used in this study to determine the bankfull discharge. The three streams with previously published or known values for bankfull discharge are East Creek, lower Carnation Creek, and Harris Creek, and they bracket the size range of the studied drainage basins: East Creek and lower Carnation Creek are amongst the smallest of the studied drainage basins, and Harris Creek is the largest. The comparison is presented in Table 5-3. The values for  Table 5-3: Comparison of bankfull discharge values from other studies to those determined in this study. Stream Name Harris Creek East Creek Carnation Creek near Mouth  Published Bankfull Discharge (m³/s) 15.91  Estimated Bankfull Discharge (m³/s) 14.2  1.6-1.72  1.77  3  30  17.7  bankfull discharge estimated in this study are very close to the values of bankfull discharge previously determined for Harris Creek and for East Creek. The value for bankfull discharge determined at the station in lower Carnation Creek was lower than a previously published  1  Hassan and Church, 2001. Published value of 19.0 m³/s is not scaled by ratio of watershed area at study site to watershed area at gauge. Scaled value is presented here.  2  Zimmerman, A. and Caulkins, J. (2010) personal communications  3  Haschenburger and Rice (2004). Method of determining bankfull not given. May have taken Q2 discharge to be bankfull discharge.  72  value (Haschenburger and nd Ric Rice, 2004). Details of the method by which ch bankfull ban discharge was determined were not ot pub published by Haschenburger and Rice (2004)) but their fieldwork was conducted in the early-- to mid- 1990s. Therefore, it is possible that they th determined bankfull discharge accurately urately at the time but subsequent aggradation ion has ha reduced the channel’s bankfull capacity, city, oor that they determined the bankfull discharge arge inaccurately in and the channel dimensions have remained relatively constant. Their reported ted value va of bankfull discharge is very close to the calculated Q2 discharge; it is possible that at they the assumed that bankfull discharge wouldd equa equal Q2 and did not test this assumption.  Figure 5-1: Histogram of bankfull nkfull frequency exceedence based on flow duration.  73  The calculated return periods (from Table 5-2) for bankfull discharge range from 1.01 year to >200 years based on the annual maxima series, and between 0.00001 and 12 days per year based on the flow duration series (Figure 5-1). Seven of the studied streams have never recorded a daily discharge exceeding bankfull, and five have never recorded an instantaneous peak discharge exceeding bankfull (although for one of those five, the maximum instantaneous peak discharge recorded is just slightly less than the calculated bankfull discharge), so the return period of the bankfull discharge is estimated from the annual maximum flood series only. An approximate equivalent return period on the flow duration series is then estimated by converting the annual maximum series return period into the expected flow duration by taking the inverse of the yearly return period in days. (The conversion is only approximate because the duration of a rare flood might be shorter or longer than one day). The magnitude of bankfull discharge is typically in the range of 0.5 to 2 times Q2, and 10 to 20 times Qmad, although there are a significant number of watersheds with larger bankfull magnitudes, with the total range observed between 0.5 and 8.9 times Q2 and 4.3 to 60.9 times Qmad. Absolute (unscaled) magnitude of bankfull discharge is not as useful for comparison between watersheds as are measures of relative magnitude. Two measures of relative magnitude are presented in Table 5-2: the ratio of bankfull discharge to the one-in-2-year flood (Qbf/Q2), and the ratio of bankfull discharge to the mean annual discharge of the stream (Qbf/Qmad). The two relative measures of magnitude are complementary; Qbf/Q2 expresses the size of the bankfull discharge relative to the average flood event in the stream, while Qbf/Qmad expresses the size of the bankfull discharge relative to the average discharge in the stream.  74  The two measures of relative ative bbankfull magnitude are linked by a third, implicit implic relation – the relative magnitude of the Q2 discharge compared to the mean annual discharge charge. There is a generally rally strong inverse relation between the frequenc equency and relative magnitude of bankfull dischar ischarge (Figures 5-2a and 5-2b), as might be expected, exp with the most scatter in the data occurr occurring in the area of high-magnitude, low frequenc quency events where the estimation of frequency ncy hhas lower precision than the estimate of magnitude. magnit The same general relation is observed erved with both flow duration (Fig 5-2a) and nd annual an maximum  Figure 5-2a: Frequency-magnitud gnitude relation for bankfull discharge based on ratio io of bankfull to mean annual discharge and flow duratio uration series. Relations are divided by hydrologic region egion. p=0.00126.  75  series (Figure 5-2b) estimates imates of frequency, with the annual maximum series giving a better fit at the frequent end off the sscale and a poorer, or equally bad, fit forr the rare r events. The difference in the strength th of tthe relation and the data scatter between the two tw measures of frequency is attributable to the variable flood durations possible for a given iven flood fl magnitude. The choice of the comparison arisons between measures of relative magnitudee and frequency also plays a role in the strength th of tthe relation: Qbf/Q2 is compared to the annual nual maximum m series, from which Q2 is derived, d, and Qbf/Qmad to the flow duration series, from m which whi the estimate  Figure 5-2b: Frequency-magnitud gnitude relation for bankfull discharge based on ratio o of bankfull b to Q2 and annual maximum series. Relations are divided by hydrologic region egion, p=0.00007.  76  of mean annual discharge is derived. If Qbf/Q2 is compared to the flow duration series estimate of bankfull frequency, or if Qbf/Qmad is compared to the annual maximum series estimate of bankfull frequency, the inverse relation is still apparent but there is much more scatter in the data because of the mismatch of measures of frequency. 5.1.1  Factors Affecting Relative Magnitude of Bankfull Discharge Given the strong inverse relation between frequency and relative magnitude of  bankfull discharge, a variable that explains some of the variation in the relative magnitude of bankfull discharge can therefore also be expected to explain variation in the frequency of bankfull discharge. Variables that demonstrated a correlation when tested against Qbf/Q2 and/or Qbf/Qmad included hydrologic region, watershed area, unit runoff, coefficient of variation of peak flow annual maximum series, channel gradient, Melton ratio, wood volume per meter of channel, bankfull width and Wohl’s (2004) Q2 stream power/D84 ratio. These variables combine naturally into conceptual hydroclimatic and hydraulic (driving/resisting forces) groupings; the combination and interrelation of the variables is discussed in detail in Chapter 5. The three hydroclimatic regions are the Coast and Mountains (fall and winter rain and rain-on-snow dominated), Thompson-Okanagan (spring snowmelt and summer convective rain-dominated) and Kootenay-Columbia (spring snowmelt dominated). The three hydroclimatic regions exhibit statistically significant distinct median values of Qbf/Q2 (Figure 5-3a), with the KC region having the highest values, the TO intermediate, and the CM lowest (ANOVA p-value of 0.01). However, the same is not true for the Qbf/Qmad values (Figure 5-3b); although there is a trend, with the KC region highest and CM region lowest,  77  Figure 5-3a: Qbf/Q2 variation by hydrologic region. Region 1 = Kootenay-Columbia. Region 2 = Thompson-Okanagan. Region 3=Coast and Mountains  Figure 5-3b: Qbf/Qmad variation by hydrologic region. Hydrologic regions are the same as Figure 5-2a.  78  the difference between the th three subgroups is not statistically significant when w tested with ANOVA (p=0.9). Takenn toge together, these two results indicate that the size of the bankfull discharge relative to the avera average peak flow varies with hydroclimatic region, region while the size of the bankfull dischargee relati relative to the mean annual discharge is essentially tially the t same across the three regions. Examining this varia variability further, can some of the variability iability between the hydroclimatic regions bee ascri ascribed to differences in precipitation or runoff off alone? alo The plot of Qbf vs. Q2 (Figure 5-4a)) show shows increased variability and a leftward deviatio eviation at the lower end; correspondingly, Qbf/Q2 ratio shows a decrease with both increasing g unit runoff (Figure 5-4b) and increasing mean an ann annual precipitation, though the effect is stronger onger for runoff than  Figure 5-4a: Plot of Qbf vs. Q2 by h hydrologic region, p=0.0000.  79  precipitation. The plot of Qbf vs. Qmad does not show a similar deviation and Qbf/Qmad ratio does not vary with unit it runo runoff. Essentially, drier drainage basins have ave higher hi values of bankfull discharge relative ive to Q2 but not to mean annual discharge. The he TO region is the driest and the CM region is the wettest. Compare this to Figure 5-3a: thee KC region is wetter than the TO region and yet ha has a higher overall median Qbf/Q2, so not all of the differences between the three hydroclima oclimatic regions can be ascribed purely to thee relative rela wetness or dryness of the regions.  Figure 5-4b: Qbf/Q2 ratio vs. unit rrunoff.  Drainage basin area rea is another variable that has been reported to affect affec the frequency and magnitude of bankfull ull disc discharge, as discussed in Chapter 1, with smaller aller drainage basins reported as having largerr and lless frequent bankfull discharge. For the drainag rainage basins in this study, area exerts a strong ng eff effect on the ratio of Qbf/Qmad (Figure 5-5a) and a much weaker 80  effect on the Qbf/Q2 ratio (Fig Figure 5-5b). All three regions exhibit a trend nd to lower Qbf/Qmad with increasing watershed ed siz size, but only the Kootenay-Columbia region on exhibits ex the same trend with area for Qbf/Q2; the TO and CM regions do not exhibit any noticeable noticea trend. This is in keeping with the results sults ppresented in Figure 5-3a and Figure 5-3b;; the effect e of area on the relative magnitude off bank bankfull discharge observed here is independent dent of o hydroclimatic region.  Figure 5-5a: Qbf/Qmad ratio vs. dra drainage basin area by region.  81  Figure 5-5b: Qbf/Q2 ratio vs. drain drainage basin area.  In the literature,, strea stream channel gradient is often linked with h the frequency and magnitude of bankfull discha discharge, as discussed in Chapter 1, with steeper eeper streams having larger and less frequent bankf bankfull discharge than low-gradient streams.. There is an obvious correlation with drainagee basin area; large drainage basins are usually low-gra gradient, but small drainages can be steeperr and so mean gradient increases with decreasing sing area. a This study evaluated two complementary entary measures of gradient: local channel gradient ent over ov the length of the studied reach, and drainage ainage basin Melton ratio, which measures the overall gradient of the drainage basin. Both measures asures of relative bankfull magnitude (Qbf/Qmad and Qbf/Q2) increase with increasing reach gradient, adient, and this is true for the study as a whole and also als for individual hydrologic regions (Figures res 5-6a and 5-6b), although the Coast and Mountai ountains (CM) region shows little effect on Qbf/Q2 (but a strong effect for Qbf/Qmad) with increase creased gradient. The 82  effect is not as strong for drainage basin Melton ratio. All hydrologic gic regions re show an increase in Qbf/Qmad with ith an increase in Melton ratio, but only the KC region r shows a corresponding increase in Qbf/Q2 with an increase in Melton ratio.  Figure 5-6a: Qbf/Q2 ratio vs. reach gradient by hydrologic region.  The coefficient of vari variation of the annual maximum series of peak flows (Cv) has been less frequently linked ed dir directly in the literature to the relative magnitude itude or frequency of bankfull discharge than have ssome of the previous variables, but as noted d in Chapter C 1, many of the factors previouslyy disc discussed can result in higher Cv. For any hypothetical hypoth drainage basin, the Cv would be expec expected to be higher than average if the watershe tershed was smaller, steeper, and drier than averag average. Hence, it was expected that there would ould be b a correlation between higher Cv and highe higher relative magnitude of bankfull discharge. ge. This T expectation  83  Figure 5-6b: Qbf/Qmad ratio vs. s. rea reach gradient by hydrologic region.  was fulfilled: higher Cv result resulted in higher relative bankfull magnitude for both bo measures of relative magnitude (Figures 5-7a and 5-7b). The Cv data show considerable erable scatter in both figures; since multiple variabl ariables can influence Cv, and these parameters rs do not n always vary in unison (for instance inn the case of a drainage basin which is smallerr but also a wetter than average) this scatter wass not uunexpected. In addition to the Cv, the L-Cv Cv (based (b on the Lmoment) was evaluated,, but the Cv showed the stronger effect. Cv and L-Cv are both measures of the dispersion sion oof a probability distribution, and in physical ysical terms can be considered as being analogous logous to the size of rare flood events relative to the th mean. Of the two, however, L-Cv is less ss affe affected by extreme outliers than Cv.  84  Figure 5-7a: Qbf/Q2 ratio vs. Cv of peak flow series by hydrologic region.  As discussed in Chapte hapter 1, Wohl (2004) identified the value of Q2 streampower/D stre 84 = 10,000 kg/s3 as an empirical pirical threshold between stream reaches with th well-developed wel or poorly-developed alluvial al cha character. Wohl’s variable represents the balance balanc of hydraulic driving forces to resistance nce fo forces, represented by streampower at a reference ference discharge and grain size respectively. The tthreshold value separates those streams found to have welldeveloped downstream hydrau hydraulic geometry from those with poorly develop veloped downstream hydraulic geometry, andd hen hence has been taken in this study as an indicator indica of alluvial character, since non-alluvial uvial reaches would be expected to have very ry poorly po developed hydraulic geometry relation elations. Wohl (2004)’s threshold value proved prov significant  85  Figure 5-7b: Qbf/Qmad ratio vs. s. Cv of annual maximum series of peak flows.  (p-values between 0.005 and 0.01 with ANOVA) when tested against Qbf/Qmad and Qbf/Q2, with both values higher for poorly developed alluvial character reaches es (Ω/ (Ω D84 < 10,000 kg/s3) than for well-developed eloped alluvial reaches (Figures 5-8a and 5-8b). Although Al Wohl’s suggested threshold valuee prov proved significant, there was not a well-defined ed trend tren of increasing Ω/ D84 and decreasing relative lative bankfull magnitude for either Qbf/Qmad or Qbf/Q2. Woody debris appeared peared at first to exert a significant effect on Qbf/Qmad ma and Qbf/Q2 but this effect did not hold upp und under scrutiny, with the apparent effect of wood ood actually ac being an effect of channel width.. Spec Specifically, two measures of woody debris abundance, abund low total  86  Figure 5-8a: Box and whisker plot of Qbf/Qmad grouped by Q2 streampower/D84 class. Class 1 = Ω/D84 < 10,000 kg/s³. Class 2 = Ω/D84 > 10,000 kg/s³.  Figure 5-8b: Box and whisker plot of Qbf/Q2 grouped by Q2 streampower/D84 class. Class 1: Ω/D84 < 10,000 kg/s³. Class 2: Ω/D84 > 10,000 kg/s3.  87  Figure 5-9: Box and whisker plot of Qbf/Qmad grouped by wood density per unit channel length class. Class 1 = wood density < 0.1 m3/m. Class 2 = wood density >0.1 m3/m  volume (m³) of woody debris in the channel and low volume of wood per unit length of channel (m³/m) correlated to higher ratios of Qbf/Qmad and Qbf/Q2 (Figure 5-9) although the degree of correlation was not strong (p-value of 0.09 with ANOVA). However, neither total volume of wood nor volume of wood per unit length of channel account for channel width, because a wider channel can have a greater amount of wood per unit length. When mean wood depth (total volume of wood divided by planimetric area of channel) is compared to Qbf/Qmad and Qbf/Q2, neither measure varies significantly. This suggests that it is channel  88  Figure 5-10a: Box and whisker plot of Qbf/Qmad grouped by channel width class. Width Class 1 is channels narrower than 8m. Width Class 2 is channels wider than 8m.  Figure 5-10b: Box and whisker plot of Qbf/Q2 grouped by channel width class. Width Class 1 is channels narrower than 8m. Width Class 2 is channels wider than 8m  89  width rather than wood abundance that is the true reason for the results presented in Figure 5-8. This is borne out when Qbf/Qmad and Qbf/Q2 are compared to channel width, with narrow channels (narrower than approximately 8m) having higher relative magnitudes of bankfull discharge (Figures 5-10a and 5-10b); the effect is stronger (lower p-value) for Qbf/Qmad than for Qbf/Q2. Despite the overall lack of effect of woody debris on the relative magnitude of bankfull discharge, there is evidence that accumulations of woody debris can still significantly influence local bankfull conditions. At the gauging site in Albert River (WSC  Figure 5-11a: Albert River: lower end of reach above woody debris jam with aggraded sediment. View upstream. Jam is approximately 50m downstream  90  gauge 08NF005), two adjacent reaches were surveyed, immediately above and below a very large woody debris jam. The reach above the jam (Figure 5-11a) was apparently aggrading, with much sediment trapped behind the jam, while the downstream reach (Figure 5-11b) was apparently in a degrading state, with its upstream sediment supply cut off by the jam. The bankfull magnitude above the jam was 11.3 m³/s, corresponding to a Qbf/Q2 ratio of 0.73 and a frequency of exceedance of 4.1 days/yr. In the reach below the jam, the bankfull magnitude was 21.4 m³/s, corresponding to a Qbf/Q2 ratio of 1.41 and an exceedance frequency of 0.6 days per year. In Albert River, therefore, the woody debris jam between the two reaches  Figure 5-11b: Albert River, lower end of degraded/armoured reach below woody debris jam. View upstream. Jam is approximately 200m upstream (out of sight).  91  interrupted sediment transpor ansport and caused local accumulation and depletio epletion of sediment which resulted in a doubling bling of the magnitude of bankfull discharge and a corresponding approximate sevenfold reducti eduction in the frequency of bankfull discharge from above to below the jam. These results, taken aken in aggregate, indicate that differences exist between these transitional alluvial streams ams an and other streams in which alluvial character ter is well-developed. w To further explore and charact haracterize the nature of these transitional alluvial ial streams, st they were compared to the quasi-univers niversal dimensionless hydraulic geometry formulatio ulations of Parker et al. (2007). Parker et al. l. (20 (2007) develop relations for dimensionlesss bankfull bank discharge,  Figure 5-12a: Relation between een di dimensionless bankfull width and dimensionless bankfu ankfull discharge. Gray line (crosses) is Parker et al. (2007) 007)’s proposed quasi-universal relation. Data from this study is stratified stra by hydroclimatic region.  92  dimensionless width and dimensionless depth for single-thread gravel bed channels. Because the small channels in the present study were generally of irregular cross-section, without a generally uniform depth across the width of the channel, comparisons were made only to Parker et al. (2007)’s formulations for dimensionless bankfull discharge and dimensionless bankfull width. Parker et al. (2007) present the following equations for dimensionless bankfull discharge and dimensionless bankfull width: Q^ = Qbf/(gD50)0.5(D50)2  (5-1)  B~ = (g0.2Bbf)/(Qbf)0.4  (5-2)  where Q^ = dimensionless bankfull discharge, B~ = dimensionless bankfull width, g = gravitational acceleration (9.8 m/s2), D50 = median grain size of bed surface sediment (m), Qbf is bankfull discharge (m3/s), and Bbf is bankfull width (m). Figures 5-12a and 5-12b show the relations between Q^ and B~ and between Q^ and slope for the stream reaches in this study, compared to the Parker et al. (2007) quasiuniversal relations. The slope of the two relations for the streams in this study is similar to the quasi-universal relations, but there are systematic offsets. Streams in this study are generally narrower (lower dimensionless bankfull width) than the streams in the Parker et al. (2007) study for a given dimensionless bankfull discharge. Likewise, streams in this study are all steeper for a given dimensionless discharge than streams in the Parker et al. (2007) study, or alternately, for a given slope, the streams in this study have a much higher dimensionless discharge than the streams in the Parker et al. (2007) study. There is possibly some small degree of systematic variation between the CM, KC, and TO regions for the width-discharge relation, or at least the data points for the three regions form recognizable  93  groups (Figure 5-12a) but ut the there is no such regional variation and less recogn ecognizable grouping for the slope-discharge relation elation.  Figure 5-12b: Relation between een ch channel slope and dimensionless bankfull discharge. Crosses and line are Parker et al. (2007) 007)’s proposed quasi-universal relation. Data from this study is stratified strat by hydroclimatic region.  5.2 Effective Discharge The evaluation off sedi sediment transported by the studied streams was largely la based on the sediment transport equatio quations, with only a few actual measurements of sediment sed transport available. Therefore the result results which evaluate the effectiveness of the equations equat predicting sediment transport are prese presented before any results relating to the calculated calcu effective 94  discharges. Observations of sediment transport were available, either as published data or as personal communications from other researchers, for three of the studied streams: East Creek, Carnation Creek, and Harris Creek. A sensitivity analysis was conducted that suggested that the calculated effective discharge was largely insensitive to variations in the slope or power of the sediment transport rating curve, but was sensitive in some cases to the estimated threshold values for initiation of sand and gravel fraction mobility, especially where the threshold value for mobility was close to the calculated effective discharge. Table 5-4 shows the calculated and observed thresholds for sand and gravel fraction motion for the three watersheds. There is a generally close agreement between the calculated and observed thresholds, although observed values are slightly closer to predicted values for the threshold of gravel transport than for sand transport. The degree of agreement is the strongest for Harris Creek, where the published threshold for gravel motion is also specifically based on a size of >8mm. Table 5-4: Predicted and observed discharge thresholds for onset of sand and gravel sediment transport in benchmark streams.  Stream  Harris Creek East Creek Carnation Creek near mouth  sand gravel sand gravel sand gravel  Predicted Discharge Threshold for transport (m³/s) 3.4 6.6 0.6 1.6 1.5 4.4  4  Hassan and Church, 2001  5  Zimmerman, A. and Caulkins, J. (pers. comm, 2010)  6  Hassan, M. (pers. comm, 2010)  7  Haschenburger and Rice, 2004  Observed Discharge Threshold for transport (m³/s) 34 74 0.35 1.45 36 47  95  Effective discharge rge exc exceedance frequencies determined for the streams stream in this study were widely variable, ranging nging over four orders of magnitude, from 0.01% % to 47% of the total length of record (0.04 too 173 days per year). (Figure 5-13). Mean and nd median me values for effective discharge exceedance edance frequency, subdivided by hydroclimatic region, region are presented in Table 5-5. In general,, the C Coast and Mountains region had a higher frequency frequen of effective discharge than the other er two regions. Overall, for the streams in this study, s effective discharge was a relatively ely fre frequent event, with flows equal to or larger er than tha the effective discharge occurring on average verage 26 days per year (median exceedance 17 days per year).  Figure 5-13: Histogram of effectiv fective discharge exceedance frequencies. Data bins are logarithmic (compare to ari arithmetic bins in Figure 5-14). Frequency is based on flow durati uration series.  Table 5-5: Fraction of time effectiv ffective discharge is exceeded, by hydrologic region and nd overall. ove  Region KC TO CM Combined  Mean 0.053 0.071 0.105 0.073  Median edian 0.051 0.060 0.045 0.053  St Dev 0.039 0.072 0.160 0.092  First quartile 0.021 0.018 0.000 0.013  Third quartile 0.083 0.094 0.123 0.095  96  Relative magnitude de of effective discharge was determined by dividin ividing Qeff by mean annual discharge (Qmad) for ea each stream. The use of Qeff/Qmad was chosen en as the best way of representing relative magnitud agnitude over Qeff/Q2 because of the frequency y of Qeff, which was generally much more frequent equent an event and hence significantly smallerr than Q2, unlike the bankfull discharge, for which both Q2 and Qmad provided reasonable variabl ariables to determine relative magnitude. The rang range of Qeff /Qmad for the studied watersheds ranged nged from 0.66 to 41 (Figure 5-14). Mean andd med median values for Qeff /Qmad for each hydroclimatic limatic region and for the entire study area are re pres presented in Table 5-6. In general, the ratio tio of Qeff /Qmad was greatest for the Coast andd Mou Mountains region and declined for the other two wo regions. reg As might be expected, there is a strong trong inverse relation between the relative magnitude nitude and frequency of exceedance for the effecti ffective discharge, with the relation essentially similar simil for all three hydroclimatic regions (Figure igure 5-15).  Qeff/Qma Figure 5-14: Histogram of Qeff to Qmad ratio using arithmetic bins.  97  Table 5-6: Qeff/Qmad ratio by hydro hydroclimatic region and overall.  Region KC TO CM All  Mean 5.37 9.03 12.4 8.53  Qeff/Qmad ratio Median SD First quartile 4.38 3.45 3.10 4.60 10.2 2.49 4.70 14.8 1.76 4.38 9.93 2.80  Third quartile 5.83 10.4 20.7 8.06  Several different type bbehaviours of sediment transport and effective tive discharge d can be distinguished from the applica pplication of the sediment transport equation and d the hydrographs of the studied streams, emerging erging from the way in which sediment transport port increases i by an  Figure 5-15: Effective discharge rge fr frequency-magnitude relation. Frequency of effective discharge is based ased on flow duration series (number of days Qeff is equalled or exceeded), excee p=0.0032.  98  order of magnitude or more when transport of the gravel fraction begins. gins. The first, most common type (here designated ignated Type I) is the stream, exemplified by Split plit Creek C or Norrish Creek, where the range of stre streamflow is enough to transport both sand d and gravel over the course of an average year. ar. In sstreams of this type, the effective discharge ge is greater g than, but generally close to the thresho hreshold of gravel transport (Figure 5-16), resulting ulting in a relatively frequent Qeff exceeded multipl ultiple days per year.  Figure 5-16: Example of Type I sstream effectiveness diagram. Split Creek cross-section 11 effectivenes eness diagram demonstrating Type I behaviour. Threshold of sand and transport is 0.1 m³/s (due to high abundance of fine sedime diment: headwaters are glaciated); threshold of gravel transport ort is i 1.2 m³/s. Effective discharge is 5.15 m³/s, exceeded 8% of the time or 29 days per year on average. Qh is 4.54 m³/s.  99  Type II streams, exemp exemplified by Daves Creek (Figure 5-17) and Twoforty Twof Creek, are streams that have only excee exceeded the threshold for gravel transport a few times over the period of record, but thatt have a large difference between the magnitude of the rates of gravel and sand transport. In these ese st streams, the amount of sand transported by more frequent flows is small, and the few instances tances of gravel transport therefore dominate, resultin sulting in a relatively infrequent Qeff (discharge ge exc exceeded less than one day per year, and generally genera only a few times over the period of record record).  Figure 5-17: Example of Typee II st stream effectiveness diagram. Effectiveness diagram for Daves Creek, eek, cross-section 5, representative of Type II behavior. Threshold hold for sand transport is 0.56 m³/s; threshold for gravel transpo nsport is 1.56 m³/s. Only 10 days out of the 7598 days of recor ecord have exceeded the threshold for gravel transport; effective ive d discharge is the largest flow on record, 2.72 m³/s. Qh is 1.13 m³/s. m³/  Type III streams,, obser observed in a few cases, are the case where thee threshold thres for gravel transport is only exceeded ed du during large, rare peak flows, or is not ever er exceeded. exc In these streams, effective discharge arge ddepends on the frequency of sand-fraction transport. trans If sand is transported relatively frequent quently, it can dominate over the rare or gravel-trans transport events and the result is an effective ve dis discharge dominated by the sand-size fraction, ction, occurring with 100  similar frequency and magnitu agnitude to the Type I streams but as sand transport ort only, on with gravel immobile. Examples of this type of stream are Guichon Creek and Camp Cam Creek. The estimated threshold for grave gravel transport in Camp Creek (Figure 5-18)) is 4.9 4 m³/s but the largest flood on record (out oof 15500 days of record) is 3.43 m³/s; rarer rer floods flo that move gravel do not occur frequently uently enough to move as much sediment as thee effective effe discharge of 0.77 m³/s.  Figure 5-18: Example of effectiven ctiveness diagram for Type III stream. Camp Creek, cross-section 6, effectivene iveness diagram, representative of Type III behavior. Threshold for sand s transport is 0.62 m³/s; threshold for gravel transportt is 4 4.9 m³/s, which approximates the Q200. Effective discharge of 0.77 m³/s represents sand moving over an essentially immobile obile gravel bed. Qh is 1.1 m³/s.  In addition to these ese th three main stream types, there were three outlier outlie streams, each representing a condition simila similar to but not identical to one of the three main ain types ty of streams. In Kanaka Creek the proportio portion of sand present in the surface is very small. The Wilcock and Kenworthy (2002) equations tions contain a hiding function that applies when the sand sized  101  fraction of sediment is minimal. This hiding function models a behaviour in which the sandsize fraction is hidden in the voids between the gravel-sized particles and only begins to move after the gravel-sized fraction begins to move. Kanaka Creek is the only stream in this study in which the sand fraction is low enough that this hiding function comes into play when modelling the transport of sediment, and consequently, the equations suggest that the threshold for the movement of gravel in Kanaka Creek is lower than the threshold for the movement of sand. Otherwise, the stream behaves much like a Type II stream. In Coquitlam River, the reach studied was several kilometers below a recent landslide and had been overloaded with fine sediment, resulting in localized aggradation. The proportion of sand present was high, and the median size of the gravel present was small. The equations suggest that both sand and gravel transport occur frequently in this reach, and that the effective discharge occurs in the sand-only transport phase due to the amount of sand present in the bed. This is a special case of the Type III stream, where gravel transport is common instead of rare, but where sand still dominates due to the abundance and mobility of the sand-sized fraction. In essence it is a Type III stream that behaves like a Type I stream, or vice versa. Finally, in Dennis Creek, which is a small, subalpine, transitional bedrock stream, the thresholds for transport of both sand and gravel are high. Over the period of record, sand transport has only occurred a few times, and gravel transport has never occurred, resulting in a Qeff that is an infrequent event, but in sand rather than gravel (a special combination of the Type II and III stream, where sediment transport of any kind is rare). A cluster analysis, with the frequency of effective discharge compared to the ratio of the proportion of time sand and gravel are transported to the proportion of time only sand is  102  transported, confirms that at Typ Type I through Type III behaviour falls into three distinct regions (Figure 5-19) with the outliers utliers Kanaka Creek and Dennis Creek falling closest closes to Type II and Coquitlam River falling close to Type I. Sand appears to replace gravel avel as a the dominant sediment transported when en the frequency and magnitude of gravel transport sport decreases d past a threshold (from Figure 5-19 19, when the ratio of the frequency of gravel transport transp falls below approximately 1% of the frequ frequency of sand transport).  Figure 5-19: Cluster analysiss of eff effective discharge types. Frequency of effective discharge is plotte lotted against ratio of sand to gravel transport frequencies.  When Figure 5-15,, the frequency-magnitude relation, is examined ned stratified st by type rather than by hydrologic ic reg region (Figure 5-20), a clearer picture of the relation re between  103  frequency and magnitudee of eeffective discharge and type of effective discharge discha is revealed. For both Type I and Typee III, effective discharge tends to be frequent and nd of low magnitude, while Type II streams have ave inf infrequent, large discharges. If only the frequency uency and magnitude of the effective discharge ge is known, but not the type of sediment that hat is moved at that effective discharge, Typee I an and Type III streams cannot be distinguished. hed. Because B of this overlap of frequency andd magn magnitude of effective discharge between stream am reaches rea with very different types of effective ve disc discharge, it is not surprising that very few measure easured or calculated variables (watershed area, ea, ch channel gradient, woody debris abundance, e, Melton Me ratio, etc.) exhibited significant correlatio rrelation when compared against frequency or relative relativ magnitude of effective discharge.  Figure 5-20: Magnitude-frequency uency analysis of effective discharge stratified by stream am type. ty Compare Figure 5-20 to Figure 5-15 (whi (which displays the same relation, but stratified by hydroclimatic region). regi  104  As previously discussed in Section 3, the Wilcock and Kenworthy (2002) equations indicate the volume of sediment moved but not the size of the sediment that is moving. Estimates of the relative mobility of the bed were made using a representatively estimated value of the critical Shields number for specified discharge magnitudes and comparing the resultant estimated median size of the mobile sediment (D50m) to the D50 of the bed. The calculated D50m values for, respectively, Q2, Q20, Qbf, and Qeff (Table 5-7) indicate that for the majority of the studied streams, the D50 of the mobile sediment approximates the D50 of the bed sediment at discharges greater than or equal to the Q2, but there are also streams for which even at the Q20 discharge or the bankfull discharge the D50 of the mobile sediment is less than the D50 of the bed sediment. For seven of the studied streams, the D50 of the mobile sediment has been less than the D50 of the bed sediment under all recorded streamflows, while for some other streams D50m equals the bed D50 at very low flows. At the effective discharge, the D50m is generally smaller than the D50 of the bed, because the effective discharge is most commonly a frequent flow smaller than the Q2. As might be expected, mean and median D50m increase from Q2 to Q20 for all the streams. There are differences between the hydrologic regions when it comes to sediment mobilization, with size of  Table 5-7: Mean, median, and standard deviation of mobility ratios (D50mobile/D50 bed). Information is provided for four reference discharges (Q2, Q20, Qbf and Qeff) by hydroclimatic region and overall.  Q2 D50mob ratio Region Mean Median SD KC 0.92 0.90 0.30 TO 0.75 0.66 0.32 CM 1.46 1.26 0.75 All 0.99 0.86 0.53  Q20 D50mob ratio Mean Median SD 1.24 1.23 0.43 1.03 1.09 0.45 1.90 1.54 0.85 1.33 1.25 0.66  Qbf D50mob ratio Mean Median SD 1.26 1.21 0.44 0.88 0.76 0.46 1.25 1.06 0.69 1.11 1.06 0.54  Qeff D50mob ratio Mean Median SD 0.68 0.67 0.21 0.60 0.56 0.24 0.66 0.62 0.26 0.64 0.62 0.23  105  mobilized sediment increasing with increasing precipitation (so CM>KC>TO region) for all four mobilization estimates. Stratifying the D50m values by stream effective discharge type reveals some interesting details (Figures 5-21a - 5-21d). A wide range of behaviour occurs in the Type I streams, from streams where D50m= bed D50 at Q<Q2 to streams that even at Q20 or Qbf (Figures 5-21b and c) have a largely immobile bed, with D50m<<D50. The Type II streams fall into the latter category, although generally approaching D50m=D50bed at flows less than Q20 or Qbf. The Type III streams are uniformly of low relative mobility even at high flows, suggesting that these streams are indeed cases where sand dominates effective discharge because of the inability of flows able to transport the bed gravels rather than cases where sand dominates the effective discharge because of the supply of sand to and in the channel. This contrasts strongly with Coquitlam River, which has some of the most mobile sediment of any of the streams studied. Coquitlam River is an example of a stream in which the stream is capable of transporting gravel as well as sand, and transports both frequently; sand dominates the effective discharge because of the very high supply of fine sediment rather than an inability of the stream to transport gravel-sized sediment.  106  45 40  Qeff/Qmad  35 30  Type I  25  Type II  20  Type III  15  Kanaka  10  Coquit  5  Dennis  0 0.00  0.50  1.00  1.50  2.00  2.50  3.00  3.50  Q2 D50m/D50bed  45 40  Qeff/Qmad  35 30  Type I  25  Type II  20  Type III  15  Kanaka  10  Coquit  5  Dennis  0 0.00  0.50  1.00  1.50  2.00  2.50  3.00  3.50  Q2 D50m/D50bed  Figures 5-21a and Figure 5-21b: Qeff/Qmad ratio vs. D50mob/D50bed ratio for Q2 and Q20 discharges. Figure 5-21a: Mobility ratio for Q2 discharge Figure 5-21b: Mobility ratio for Q20 discharge  D50 mobile/D50 bed ratios greater than 1 suggest only that the bed is generally mobile and should not be overinterpreted.  107  45 40  Qeff/Qmad  35 30  Type I  25  Type II  20  Type III  15  Kanaka  10  Coquit  5  Dennis  0 0.00  0.50  1.00  1.50  2.00  2.50  3.00  3.50  Qbf D50m/D50bed  45 40  Qeff/Qmad  35 30  Type I  25  Type II  20  Type III  15  Kanaka  10  Coquit  5  Dennis  0 0.00  0.50  1.00  1.50  2.00  2.50  3.00  3.50  Qeff D50m/D50bed  Figures 5-21c and Figure 5-21d: Qeff/Qmad ratio vs. D50mob/D50bed ratio for bankfull and effective discharge. Figure 5-21c: Mobility ratio for bankfull discharge Figure 5-21d: Mobility ratio for effective discharge  108  The effective discharges determined for the studied streams can be compared directly to the bankfull discharge values determined for the same streams. Based on the WolmanMiller model, fully alluvial equilibrium streams should plot close to the 1:1 line when effective discharge is plotted against bankfull discharge. When unscaled effective discharges are compared against unscaled bankfull discharge values for the same streams (Figure 5-22), all the Type II streams, as well as Kanaka Creek and Dennis Creek which are Type II-like, plot close to or above the 1:1 line (Qeff>Qbf), while the remaining streams plot below it (Qeff<Qbf). Some of the Type I and Type I-like streams plot well below the line, indicating magnitudes of bankfull discharge over an order of magnitude lower than the corresponding effective discharges. Wohl (2004) proposed a threshold value for well-developed alluvial conditions based on the ratio of Q2 streampower to the size of the large sediment in the channel (represented by the sediment D84), as discussed in Chapter 1 and earlier in this chapter with respect to bankfull discharge. In order to evaluate how well this parameter influenced the effective discharge, the relative magnitude of Qeff was plotted against the Wohl (2004) Q2 streampower/D84 parameter (Figure 5-23). All of the Type II and Type III streams, as well as Dennis Creek, fall below Wohl’s threshold value of 10,000 kg/s3 for well-developed alluvial conditions, while the Type I streams are distributed both above and below this value and Kanaka and Coquitlam both fall above. Overall, there is a pattern of an increase in mean Q2 streampower/D84 parameter from Type III to Type II to Type I streams.  109  Figure 5-22: Unscaled Qeff vs Qbf, p plotted by stream effective discharge type.  Although diagramss like Figure 5-22 are common in the literature, re, by itself Figure 522 is insufficient to indicate cate w whether stream reaches in which Qbf and Qeff values valu are identical or relatively similar are anoma anomalous or ordinary. A stream in which Qbf and Qeff are both 1 in 100 year events, for instance, ance, w will plot as closely to the 1:1 line as a stream eam in which Qbf and Qeff are both 1 in 2 year ar eve events. Therefore, to distinguish between these two cases, the relation between the ratioo of eeffective to bankfull discharge, and relativee bankfull bank magnitude is also presented (expresse pressed as both Qbf/Q2 and Qbf/Qmad) (Figur Figures 5-24a and  110  Figure 5-23: Q2 Streampower/D84 ratio vs. unscaled Qeff. Wohl (2004)’s threshold value for poorly orly developed alluvial character is 10,000 kg/s³.  5-24b). In these two graphs, phs, w where the 1:1 line from Figure 5-22 becomes mes a horizontal line, there is a strong inverse relati relation between Qeff/Qbf ratio and relative bankfull nkfull magnitude for Type I streams. The other er stre stream types show no obvious relation, although ugh the th sample sizes are much smaller. The inverse verse relation between the effective to bankfulll ratio and the relative bankfull magnitude suggests ests th that in Figure 5-22, the Type I streams which ich plot pl furthest from the 1:1 line do so because se the they have relatively large, infrequent bankfull ull discharges dis – they are incised.  111  Figures 5-24a and 5-24b: Qeff/Qbf ratio vs. relative magnitudes of bankfull discharge. Relative magnitude of bankfull expressed ssed as Qbf/Qmad (Figure 5-23a, p=0.0007) and Qbf/Q2 (Fig 5-23b,, p=0.0071). p=  112  Figure 5-25: Regional trend in Qeff with drainage basin area.  Hydroclimatic regime gime iis one of the only variables in this study which hich has any notable effect on the magnitude of effe effective discharge. In general, the CM region n has larger effective discharge for a given areaa tha than do TO and KC regions (Figure 5-25), while the TO region has the greatest variability. ity. In part, this is likely due to the smaller sample mple size in the CM region, and in part to hydroclim droclimatic regime: the higher the precipitation,, the larger la the runoff and consequently the larger the Qeff, despite all other factors.  113  5.3 Half-load Discharge Vogel et al. (2003) proposed that the half-load discharge provides a better means than the effective discharge of evaluating the contributions of large flows to the overall transport of sediment. The half-load discharge is the discharge above which half of the total sediment load is transported. Vogel et al. (2003) discussed cases where the half-load discharge was a larger and rarer flood than the effective discharge. The calculated empirical half-load discharge, Qh for each stream reach is compared to the mean annual discharge, bankfull and effective discharges in Table 5-8. In general, Qh is more frequent and smaller than both the Q2 and Qbf, but less frequent and larger than the mean annual discharge. The Qh and Qeff are related: on the plot of the effectiveness relation, the Qeff is the peak of the curve (the mode), while the Qh is the median value. In the model presented by Wolman and Miller (1960), (Figure 1 in their paper, or Figure 1 of Andrews and Nankervis (1995)) the effectiveness relation has an approximately normal or lognormal distribution, and the peak of the curve, corresponding to the mean annual flood, also approximates the median value, so Qh approximates Qeff. This relation does not hold true for more strongly skewed effectiveness distributions, and in fact the Qh and Qeff together are descriptive of the skew: for Qh > Qeff the effectiveness relation is positively skewed (the effective discharge is relatively small) and for Qh < Qeff the relation is negatively skewed (the effective discharge is relatively large).  114  Table 5-8: Bankfull, effective and half-load discharges for the studied streams.  Watershed Code 08NP004 08NK026 08NF005 08NH115 08NH084 08NB016 08NJ061 08NE114 08NJ130 08NJ129 08NE087 08NM142 08NM137 08NM242 08NM241 08NM240 08LE077 08NM173 08LG064 08NM134 08LG016 08LG056 08LF081 08LF084 08MF062 08MH150 08MH076 08MH141 08GA065 08HB069 08HB048 EASTCRK  Name CABIN CREEK HOSMER CREEK ALBERT RIVER ABOVE JAM ALBERT RIVER BELOW JAM SULLIVAN CREEK ARROW CREEK SPLIT CREEK REDFISH CREEK HIDDEN CREEK ANDERSON CREEK (Nelson) FELL CREEK DEER CREEK COLDSTREAM CREEK DAVES CREEK DENNIS CREEK TWO FORTY-ONE CREEK TWO FORTY CREEK CORNING CREEK GREATA CREEK BEAK CREEK CAMP CREEK PENNASK CREEK GUICHON CREEK AMBUSTEN CREEK ANDERSON CREEK (Hat Cr) HARRIS CREEK COQUIHALLA RIVER NORRISH CREEK KANAKA CREEK COQUITLAM RIVER NOONS CREEK CARNATION CREEK AT 150 M CARNATION CREEK AT THE MOUTH EAST CREEK IN MKRF (September) EAST CREEK (December)  Qmad (m3/s) 1.93 0.12 1.59 1.59 0.06 1.72 1.66 0.86 1.61 0.11 0.06 0.88 0.25 0.12 0.05 0.06 0.06 0.37 0.08 0.47 0.15 0.73 0.13 0.02 0.09 1.20 3.29 7.90 2.73 6.58 0.25 0.22  Qbf (m3/s) 27.6 6.63 11.3 21.4 2.44 21 7.16 15.6 18.1 2.89 2.32 12.2 4.61 1.25 1.55 0.685 1.29 3.6 1.13 8.03 1.98 5.3 1.19 1.34 2.57 14.2 32.7 142 31 65.1 3.03 7.7  Qeff (m3/s) 15.5 0.315 6.43 23.9 0.213 5.135 5.15 2.65 9.93 0.634 0.345 4.16 0.694 2.72 1.85 1.04 0.234 1.5 0.654 1.14 0.774 4.64 0.214 0.045 0.084 13.45 12.1 13 68.8 4.34 0.445 9.18  Qh (m3/s) 16.7 0.632 7.36 8.35 0.275 8.24 4.54 4.44 10.7 0.61 0.326 3.95 1.22 1.13 1.09 0.7 0.435 1.94 0.814 2.55 1.1 5.25 0.655 0.122 0.566 9.95 15.1 15 31.1 11.9 0.88 2.32  0.81  16.2  4.65  7.25  0.12  1.98  2.31  1.97  0.12  1.98  0.64  0.725  115  There are relatively vely sstrong correlations between the Qh and d both bot the absolute (unscaled) Qeff (Figure 5-26 26) and the Qbf (Figure 5-27). While correlation ation is i expected and hence unsurprising here beca because unscaled discharges are being compared, pared, the degree of correlation between both Qh and Qbf and Qh and Qeff is greater than thatt directly direc between Qbf and Qeff (Figure 5-22). When scaled (relative) magnitudes are compared, red, there th is a strong correlation between relative tive m magnitudes of Qh and Qeff for both Qh/Qmad vs. Qeff/Qmad , and Qh/Q2 vs. Qeff/Q2 (Figures 5-28 a-b). For relative magnitudes of Qh and Qbf, there is no apparent correlation between ween Qh/Qmad and Qbf/Qmad, but some apparentt correlation corre between Qh/Q2 and Qbf/Q2 (Figure 5-229).  Figure 5-26: Absolute (unscaled) led) relation of Qbf and Qh by hydrologic region, p=0.0002. 0002.  116  Figure 5-27: Absolute (unscaled) led) relation of Qeff and Qh by hydrologic region, p=0.0000 0000.  The relative magnitud gnitude of Qh is generally not correlated with th variables var such as watershed area, Cv of annual nnual maximum peak flow series, or reach gradient. dient. However, both Qh/Qmad and Qh/Q2 show w a nnegative correlation with the Wohl (2004) 4) alluvial all character (streampower/D84) parameter, eter, with the relation stronger for Qh/Q2 (Figure 5--30). When the Qh is compa compared to the typology of streams, it is apparent ent that, tha like the Qeff, the Qh is unable by itselff to di distinguish between the Type I streams in which hich gravel g transport is a frequent process and nd the Type III streams in which sand moves over ver a gravel g bed. In general, because of the correl correlation between Qeff and Qh, the streams that at have high relative magnitudes of Qh (Type pe II sstreams, plus Type II-like streams Kanaka Creek and Dennis  117  Figures 5-28a and 5-28b: Relation lation between relative magnitudes of Qh and Qeff. Relation between relative (scaled) magn agnitudes of Qh and Qeff, using Qmad and Q2 respectively for scaling aling, stratified by stream effective discharge type. Figure 5.28a a p= p=0.0000 and Figure 5.28b p=0.0003.  118  Figure 5-29: Relation between n rela relative magnitudes of Qh and Qbf . Relative magnitudes scaled by Q2 and d str stratified by stream type, p=0.0745.  Figure 5-30: Relative magnitude tude oof Qh vs. Q2 streampower/D84 ratio. Relation between relative magnitude e of Qh (scaled by Q2) against Wohl (2004) alluvial characterr parameter, pa stratified by hydrologic region. p=0.0009.  119  Creek) also have high relati relative magnitudes of Qeff. Furthermore, when hen Qh/Qeff ratio is considered, where Qh/Qeff rati ratios equal to 1 indicate no skew of the effectiv ffectiveness relation, values less than 1 indicate te neg negative skew and values greater than 1 indicate cate positive po skew, all the Type II and Type II-like like sstreams have values less than one, and the he remaining rem streams have values close to or greater reater than 1 (Figure 5-31).  Figure 5-31: Qh/Qeff ratio vs area, stratified by hydrologic region.  5.4 Other Results In general, results ts for measured values which did not show notable table correlation c with the frequency or relativee magn magnitude of Qbf, Qeff or Qh have not been present resented. The lack of correlation for one such group of values deserves to be presented becausee the absence a of such correlation can be as meaning eaningful as the presence of correlation. Specifically, ically, drainage-basin scale parameters which it ha had been thought might provide useful indices indice for sediment supplied to the stream chan channel – measures of number of landslides es and roads in the drainages, and density off such features per square kilometer of drainage basin area, as well as  120  the width of the valley flat (which had been anticipated to measure connectivity of the sidewalls to the channel) – did not demonstrate any apparent correlation to the frequency or relative magnitude of bankfull, effective, or half-load discharge.  121  Chapter 6: Discussion The results presented in Chapter 4 reveal a fundamental disconnect between the bankfull and effective discharge for the studied streams. The frequency and relative magnitude of bankfull discharge show correlation with drainage-basin and reach-scale hydroclimatic and hydraulic variables while the frequency and relative magnitude of effective discharge show limited correlation with any variables beyond the most basic hydroclimatic distinction between rainfall- and snowmelt-dominated streams. In order to place these results in context, this chapter discusses the interrelation between the hydroclimatic and hydraulic variables affecting bankfull discharge. The results from this study for bankfull and effective discharge are next compared to the results from other studies in order to place the results in a wider context. Differences which arise from the circumstances common to the drainage basins in this study (such as glacial history) are distinguished from differences which arise because of assumptions or methodological differences particular to other studies, such as the class-based method of defining the effective discharge. The first part of this chapter discusses the bankfull discharge results. This is followed by a discussion of the results concerning effective and half-load discharge. Finally, the meaning of the bankfull and effective discharge results are discussed together, and are followed by suggestions for future research. Conclusions follow in Chapter 7. 6.1 Bankfull Discharge 6.1.1  Comparison of Bankfull Discharge Values to Other Regions The results presented in Chapter 4 indicated that the relative magnitudes of bankfull  discharge observed in this study were higher than what might have otherwise been expected. Specifically, it appeared that many of the streams with Type I and Type III effective  122  discharge were incised to some degree. In order to place the bankfull discharge values from this study into a broader context, several other previously published studies were considered in which both Qbf and a reference discharge were presented for a range of drainage basin areas. In the literature, reference discharges are often omitted from publication, which reduces the possibility for this sort of comparison between studies. Andrews (1980) and Navratil et al. (2006) present both Qbf and Qmad for their studied streams, while Castro and Jackson (2001) present only Qbf, but reference stream gauge numbers for which the USGS publishes streamflow statistics online. Values were calculated for Qmad from the USGS data for all the streams studied by Castro and Jackson (2001) and Qbf/Qmad ratios computed. The ratios from the three other studies were then compared to the measured Qbf/Qmad ratio values from this study in order to evaluate how such ratios vary across a much larger range of areas than the 1-100 km² drainage basins in this study (Figure 6-1). The graph of combined results suggests that the range of Qbf/Qmad values, and by extension the Qbf frequencies and magnitudes, observed in this study are broadly within the range of variability of values observed in other studies, particularly since the range of Qbf/Qmad values observed in the other studies for drainage areas <100 km² overlaps the range of Qbf/Qmad values observed in this study. While within the broad range of variability reported from other studies, the results from this study are in general at the upper end of that range of variability, and the streams in this study with the highest Qbf/Qmad ratios have relative bankfull magnitudes that are nearly twice as large as the highest values from the other studies. This correlates with the findings concerning the narrower dimensionless bankfull width B~ and steeper gradient of the streams in this study relative to those studied by Parker et al. (2007). The streams in this study are, as  123  a group, steeper and narrowe arrower than other streams, and they have high gh ratios rat of bankfull discharge relative to reference rence discharges.  Figure 6-1: Qbf/Qmad ratio vs. aarea values from this study compared to results ults from f three other comprehensive studies. Data from other studies is from Andrews rews (1980) for Colorado, Navratil et al (2006) for France, and Castr astro and Jackson (2001) for US Pacific Northwest.  6.1.2  actors Affecting Bankfull Discharge Interrelation of Facto A number of variables iables were presented in Chapter 4 which demonstrat nstrated a correlation  with the frequency or relative elative magnitude of bankfull discharge. There are also a correlations between the variables affec affecting bankfull discharge frequency and d magnitude; ma those correlations are discussedd here here. It was suggested in Chapter 4 that these se variables vari represent 124  two subgroups, hydroclimati limatic and hydraulic forces. This consideration tion underlies the discussion of the interrelation lation oof the individual variables. The first interrelation lation to consider is that between drainage basin asin area a and stream gradient. Figures 5-5a and 55-5b indicated that drainage basin area wass inversely inve correlated with relative bankfull magnitu agnitude for all three hydroclimatic regions for or Qbf/Qmad, but only showed similar correlation on in tthe Kootenay-Columbia region for Qbf/Q2. Figures Figu 5-6a and 56b indicated a positive correla correlation between stream gradient and relative bankfull bank magnitude for both Qbf/Qmad and Qbf/Q2.  Figure 6-2a: Relation between en re reach gradient and drainage basin area for each hydroclimatic hyd region, p=0.0043. 125  The general interrelati rrelation between stream channel gradient and drainage drain basin area can be summarized by saying aying that within the constraints imposed by tectonics ctonics and landscape evolution, small drainages es can have a wide range of gradients, from gentle to steep, while large drainages are constrain nstrained to have gentle gradients. Small drainage inages are therefore typically steeper on average erage than large ones. Within the study area, there were negative correlations between drainage ainage basin area and stream channel gradient for each ea of the three hydroclimatic regions (Figur Figure 6-2a) but the degree of correlation was greatest g for the Kootenay-Columbia (KC)) reg region and lowest for the Coast and Mountains ins (CM) (C region. All  Figure 6-2b: Relation between een M Melton ratio and drainage basin area for each hydroclimatic hydr region, p= 0.0008.  126  three regions have a similar ilar ra range of channel gradients for the larger streams treams (70-100 km2), but for the small streams, s, the KC region has much steeper gradients than the TO region, which in turn is steeper than tthe CM region. The small streams in the KC region re are almost an order of magnitude steeper teeper than the small streams in the CM region. n. Note Not that the same relation does not hold forr drain drainage basin Melton ratio (Figure 6-2b) which hich measures m overall drainage basin gradient rathe rather than local channel gradient: the CM streams stream have higher Melton ratios than do thee TO streams, and are almost as steep overall ass the KC streams for many of the studied drainages. nages.  Figure 6-3: Flood duration series eries ffrequency of bankfull discharge vs. Qbf/Q2 ratio, p=0.0701. p=0.0 For given Qbf/Q2, frequency of occurrence rence is higher for snowmelt regions (KC, TO) than for rainfall-domina minated region (CM).  127  When Figure 6-2a is compared to Figures 5-5b and 5-6a, it appears that the anomalously strong inverse correlation between Qbf/Q2 and drainage basin area observed for the KC region results from the effects of channel gradient. The smaller KC streams are much steeper than the correspondingly sized CM or TO streams, and therefore experience bankfull discharges that are larger relative to the Q2 discharge. When the effect of gradient is discounted there is still a difference between the snowmelt-dominated regions (KC and TO) and the rainfall-dominated region (CM). In the snowmelt-dominated regions, at the scale of the studied drainage basins, Qbf is on average greater in magnitude and less frequent than the Q2 event; in the rainfall-dominated region, Qbf is on average more frequent and smaller than the Q2 event (Figure 5-3a). Additionally, for the same relative flood magnitude, floods in the CM region are shorter duration (occurring over fewer days per year) than floods in the KC and TO regions (Figure 6-3; compare to Figures 5-2a and 5-2b). These differences are almost certainly the result of the flood-generating mechanisms – at the scale of drainage basin studied, snowmelt floods tend to be of longer duration than do rainfall-generated floods because they are generated by seasonal temperature changes while rainfall-dominated systems tend to have multiple floods per year with shorter individual durations than snowmelt floods (Beckers et al, 2002). A heat wave and corresponding snowmelt flood peak can last for a week or more, while a rainfalldriven flood peak in a small catchment may only last for as long as the day or two of heavy rain which produces it. The effects of rain on snow add a confounding factor, but it can be argued that fall rain on transient snow, which produces the largest floods in the CM region, acts more like intensified rainfall, while spring rain on late-season snowpacks in the interior region (KC and TO) acts more like intensified snowmelt.  128  Of the variables discu discussed above, region is clearly a hydroclimatic matic variable, while gradient is a hydraulic variabl variable since increasing gradient increases stream eam power. p Drainage basin area is less easilyy class classified, but it is probably best considered d as a hydroclimatic variable, since for small draina drainages, the intensity of runoff from localized precipitation preci such as a convective storm is greater ater (C (Church, 2002; Beckers et al, 2002, Buffington, gton, 2011). 2 Another pair of variab variables which proved to be related to each other and to area are bankfull channel width and nd W Wohl’s (2004) ratio of bankfull streampowerr to D84. Figure 6-4a demonstrates the correlation ation bbetween the streampower/D84 ratio and bankfu ankfull width. Based on the observed data, a bankfu ankfull width on the order of 5 to 8m corresponds ds to Wohl’s W (2004)  Figure 6-4a: Relation between n Q2 streampower/D84 ratio and bankfull channel width,, p=0.0000. p=  129  threshold value for poorly orly ddeveloped alluvial conditions of 10,000 kg/s3. Below these threshold values, bankfull ull dis discharge was a generally rare and large event, event while streams above the threshold havee bank bankfull discharges with a median magnitude approximating approx the Q2 and a range generally between tween 0.5 and 2 times Q2. Drainage basin area is strongly stro correlated with channel width forr each region, with the CM region having channe hannels an order of magnitude wider than the he TO region for a given drainage basin area,, and the KC region intermediate between thee two ((Figure 6-4b). The Q2 streampower/D84 ratio is also correlated with drainage basin area for ea each region, although the degree of correlation tion is not as great as  Figure 6-4b: Relation between een ba bankfull width and drainage basin area for each h hydroclimatic hyd region, p-values listed by region: KC p = 00.00000, TO p = 0.0007, CM p = 0.0033. 130  that between area and width. The bankfull width value of 5 to 8m corresponding to the observed transition between well- and poorly-developed alluvial character is similar to, but larger than, a threshold value of 3m used by Church (1992) to distinguish between headwater streams and intermediate streams. The ratio of streampower to sediment size is a hydraulic variable, as is the channel width, but the way in which the channel width and Q2 streampower/D84 ratio vary with area and hydroclimatic region indicate the contribution of hydroclimate to these two variables. The coefficient of variation (Cv) of the annual maximum peak flow series has been correlated with drainage area for British Columbia across a scale range from 1-100,000 km², (multiscaling behaviour; Gupta and Waymire, 1990) with smaller drainages having higher Cv (Beckers et al. 2002). Across the range of drainage areas in this study the trend is not significant and is barely apparent, so for the range of areas examined in this study simple scaling of Cv with area essentially applies. In addition to not changing appreciably with drainage basin area for the studied streams, Cv in this study did not show significant trends with either channel gradient or drainage Melton ratio, but there was a trend. Physically, the Cv of the peak flows indicates the size of large, infrequent floods relative to the frequent ones; for two drainage basins of the same size with the same Q2 but different Cv, the watershed with the higher Cv would be expected to have larger Q20, Q50 and so on. The variability of flood magnitudes has previously been argued to influence the effective discharge and bankfull width of streams (Wolman and Gerson, 1978) as well as the response time of streams to disturbance and hence the time scale required to develop metastable equilibrium conditions (Buffington, 2011). For the studied streams Cv also varies inversely with unit runoff, with the driest drainage basins having the highest Cv (Figure 6-5a), and the  131  ratio of Q2 streampower to sed sediment D84 shows an inverse correlation to Cv, (Figure 6-5b). Note that the overall trend nd of Cv with unit runoff across the three hydroclim roclimatic regions in Figure 6-5a is less steep ep th than the trend within each individual region. ion. Although A flood duration, rather than magnitud gnitude, is the best predictor of geomorphic effectiven ectiveness (Costa and O’Connor, 1995; Buffington, gton, 2011), it nonetheless seems reasonable that hat watersheds wa with a wider range of peak flow magn magnitudes should adjust to that variability.  Figure 6-5a: Cv of annual maximu aximum peak flow series vs. unit mean runoff by region.  The Cv is clearly ly a hhydroclimatic variable, but it can also be related relate to hydraulic variables like the ratio of Q2 streampower to D84. In some ways the Cv integra ntegrates many of the other drainage basin character racteristics such as gradient, hydroclimate and bankfull bankf width into a  132  single number expressingg the vvariability of the flood flows. The problem with using the Cv as an index for a wide range of oother variables is that the other variables do not always vary in the same manner and so th the net effect does not indicate as much h about abo the relative importance of the individual vidual variables as the individual variables themselves thems do. It is interesting that Cv proved ed to show a stronger correlation than L-Cv (the he Cv determined by the method of L-moments) ts) wi with bankfull discharge magnitude and frequency uency; since L-Cv is less affected by outliers in th the distributions (Klemes, 2000), it suggests sts that tha the few very large rare events which may be taken as outliers are important in determini ermining the channel banks.  Figure 6-5b: Relation between een Q2 streampower/D84 and Cv of annual maximum mum series of floods, p=0.0174.  133  The absence of strong effects of woody debris on bankfull discharge magnitude or frequency at scales beyond the local (as exemplified by the Albert River values above and below the massive jam) was somewhat unexpected because in the field there had seemed to be an observable correlation between low woody debris abundance and deeply incised, coarse-armoured channels. This apparent correlation is, however, not supported by the data. It therefore appears that either channels banks are capable of adjusting to the supply of woody debris or that other effects dominate over woody debris abundance in determining the bankfull dimensions of the channel. Discharge magnitude and frequency are nonlinearly related, and the relation varies between streams, even closely adjacent ones. The majority of this discussion has focused on relative magnitude, rather than frequency, if only because it is possible to estimate magnitude with apparently greater accuracy than frequency: a bankfull discharge that is estimated at 7 times the Q2, for instance, has a magnitude which is relatively well constrained, but a much greater uncertainty in the frequency – with only 20 years of recorded discharge, the estimated return period for the bankfull event might vary between 100 and 500 years, for instance. Likewise, while it is possible to calculate mean Qbf/Q2 and Qbf/Qmad values for the total data set and for each hydrologic region, and likewise to calculate mean bankfull return period for each region and the total data set (Table 6-1), the resultant mean relative magnitudes cannot meaningfully be converted to a representative frequency or vice versa because the results depend nonlinearly on whether frequency or magnitudes are averaged. Nonetheless, in general terms, the majority of streams in this study had a return period for bankfull discharge that was greater than that of the mean annual flood.  134  Table 6-1: Mean (and median) Qbf/Q2, Qbf/Qmad, and Qbf return period determined from the annual maximum flood series for the entire study area and by hydrologic region.  All Regions Kootenay-Columbia Thompson-Okanagan Coast And Mountains  Mean (Median) Qbf/Q2 Ratio  Mean (Median) Qbf/Qmad Ratio  1.84 (1.25) 2.46 (1.54) 1.89 (1.31) 0.82 (0.64)  19.09 (13.88) 20.99 (14.06) 18.89 (13.58) 16.57 (14.38)  Mean (Median) Return Period of Qbf, years, annual maximum series 2.51 (3.55) 4.47 (10.5) 2.60 (3.55) 1.46 (1.17)  As noted earlier the effects of drainage basin size (scale) and channel gradient in this study are interrelated, because smaller watersheds tended to be steeper. Although streams with any physical or recorded evidence of debris flow were excluded from the study, it is nonetheless possible that for small drainages with steep gradients, at the extremes of the streams studied, the channel processes include both flood and debris flood (Wilford et al, 2004) and that debris flows could occur in some of the studied streams. Debris flow and debris flood are capable of producing anomalously large discharges (Jakob and Jordan, 2001). Therefore, it is possible that some of the estimated frequencies of rare bankfull discharge based on fluvial floods could underestimate the return period of debris flow or debris flood in the subject streams – what is estimated as a 1 in 200 year bankfull flood event could correspond to a debris flood with a 1 in 50 year return period, for instance. 6.1.3  Bankfull Discharge in Transitional Alluvial Streams The frequency and magnitude of bankfull discharge in the studied streams is  influenced by or related to the size of the drainage basin, the hydroclimatic regime, the coefficient of variation of the annual maximum flood series, the channel slope and overall drainage basin steepness, the ratio of bankfull stream power to the size of coarse sediment in the channel, and the bankfull width of the channel. The interrelation between these factors is 135  complex, but an overall trend is apparent in which, as the alluvial character of the watershed increases, the relative magnitude of bankfull discharge decreases and the frequency of bankfull discharge correspondingly increases. To extend this discussion from bankfull discharge to channel-forming discharge requires consideration of the frequency and magnitude of sediment transport, and hence discussion of the effective discharge. 6.2 Effective (and Half-load) Discharge The range of effective discharge frequencies and magnitudes evaluated in this study reveals substantial variability, although all streams studied were relatively small mountain drainages with gravel-bed streams. This variability arises specifically because these streams are small and gravel-bedded; in larger drainages containing larger streams with finer bed sediment, sediment transport is transport-limited, and hence the effective discharge is a function only of transport capacity, while in these small watersheds, sediment transport (of the coarser material, and in some cases, of the finer material as well) is often supply-limited or threshold transport limited: the watershed has excess transport capacity at some flows (is capable of moving more sediment than actually moves), or cannot move much of the sediment present until a threshold discharge is reached. In the majority of the streams studied, there is a significant disconnect between the bankfull and effective discharges for the stream, with the bankfull discharge being an relatively infrequent, large event and the effective discharge a relatively frequent, small event. For most of the studied streams, neither the bankfull nor the effective discharge corresponds to the mean annual flood. Likewise, in most of the studied streams, the median size of the sediment that moves during the effective discharge is finer than the median size of the bed material. For the Type I and Type III streams the half-load discharge is less frequent  136  than the effective discharge but still generally more frequent than the bankfull discharge. Even though sediment transport is threshold-limited, the threshold for transport in the Type I and III streams is low enough that sediment moves relatively frequently; the sediment that moves is generally much finer than the bed sediment. Hydroclimatic regime is clearly significant in determining the absolute magnitude and frequency of the effective discharge, and the division which proves to be significant is the division between rain-dominated (Coastal) and snowmelt-dominated (Interior) regime. The two Interior subregions, despite being distinct from each other in terms of bioclimate and peak flood generation mechanisms, generally behave more like each other than they behave like the Coastal region. Snowmelt floods tend to be single events with a relatively long duration, while there can be multiple peak events from winter rainfall in the coastal region each rainy season, each of shorter duration than a snowmelt flood. Thus, there are differences in magnitude and duration of floods between regions for a single event of similar frequency. For instance, the Q5 flood in the coastal region has a higher relative magnitude but a lower duration than in the interior region. In addition to the differences shown in Tables 5-4 and 5-5, there are also differences between the regions when the mobilization of sediment is considered, as shown in Table 5-6: for instance, on the Coast, the median size of the sediment mobilized approximates the bed D50 at flows less than the Q2, while in the Interior, the same relation does not occur until much higher flows, Q10 or greater. The use of a single value for effective discharge frequency and magnitude to represent the hydrologic and sediment transport processes occurring in a stream is not optimal, however, because it does not provide enough information about the size and nature of the material moved by the effective discharge relative to the bed material. The same caveat  137  applies to the half-load discharge. Using the half-load discharge and effective discharge together provides more information than using either singly, but still does not distinguish between Type I and Type III streams. The frequencies and magnitudes of effective and halfload discharge in Type I and Type III streams occur within the same ranges, but represent entirely different processes: in Type I streams, the gravel material making up the bed is moving, while in Type III streams, sand is moving over a mostly immobile gravel bed. The Type I, Type II and Type III streams appear represent a continuum of process, with gravel transport relatively common, then rare, and finally exceedingly rare, but the single-value effective discharge does not adequately represent this continuum of process. Single-value effective (and half-load) discharge becomes less frequent as the frequency of gravel discharge decreases (moving from Type I to Type II), then becomes more frequent as sand transport replaces gravel transport as the sediment transport mode that transports the most sediment over time (moving from Type II to Type III). This range of behavior is interesting, both because of the variability, and because it encompasses the whole range of effective discharges mentioned in the literature, from frequent, through annual, to infrequent, and finally exceedingly rare events within the one set of studied streams. The question of the frequency of effective discharge has been a longstanding one in the hydrologic literature, often framed as a question concerning the effectiveness of very large flows. Wolman and Miller (1960), in their conclusion, presented this debate as a fairytale log chopping contest between three woodcutters: a dwarf, a man, and a giant. Many subsequent studies have advanced contentions that one or the other type of effective discharge is the most common general condition. In the present study, examples of  138  all three types (very frequent, moderately frequent, and rare) are present, and some of the conditions under which each type occurs are defined. The Type III stream is also interesting in that it has a channel composed of both sand and gravel, but it is the sand moving through a mostly immobile gravel channel which forms the effective discharge. This illustrates an extreme case of the division between the conceptual ideas of channel-maintaining and channel-forming or channel-changing discharges, in which the channel is maintained by the transport of fine sediment but formed and changed by large rare flows capable of moving coarser material. It is also possible that in some of these streams the large flows occurred more frequently under a previous hydroclimatic regime and that current flows are sufficient to maintain the stream but no longer to change or form it. This issue can be framed as a contrast between effectiveness of sediment transport and geomorphic effectiveness. The effective discharge as defined and measured in this study refers to the effectiveness of stream flow in moving sediment: the effective discharge is the discharge class which transports the maximum amount of sediment over the period of record of stream gauging. For the Type I and II streams, where the stream bed and banks are gravel, and the effective discharge transports gravel, the argument can be made that the effective discharge with respect to the transport of sediment is also, actually or potentially, a geomorphically effective flow. For most Type I streams, however, there is no link between the observed effective discharge and the form of the channel as expressed by the bankfull discharge. Only the Type II streams consistently have an effective discharge that closely approximates the bankfull discharge. For the Type III streams, where the streambed and banks are gravel, but where the effective discharge occurs in the range of discharges that  139  transport only sand-sizedd sedi sediment, it appears that the calculated effective tive discharge di cannot be a geomorphically effectiv fective flow in the sense of forming the channel nnel or altering bed features; at most, it can be a ch channel-maintaining discharge that preventss the accumulation a of fine sediment. The use of a small all nu number (25, 50, or 100, for instance) of classes classe for effective discharge calculation would ould not have affected the classifications of Type I or Type III streams, but by misrepresentin senting the effectiveness of large, rare events, itt would woul have resulted in Type II streams being miscl misclassified as Type I or Type III streams. Ass an example, ex Figure 5-17, the sample effectivene ctiveness relation for a Type II stream, wass replotted rep using a predetermined number off disch discharge classes, in this case 30 (Figure 6-6). ). If this th fixed number of discharge classes had been used, it would have resulted in an estimation ation of the effective  Figure 6-6: Replot of Figure 5-17 17 (Daves Creek effectiveness diagram, example off Type Typ II stream) using fixed number of classes. Class width = 0.1 m, number of classes ses = 30.  140  discharge for Daves Creek as being 0.645 m3/s (within the range of sand-size fraction sediment transport), with a frequency of 14 days/year, instead of the value determined in this study using the event-based class width method of 2.72 m3/s (frequency of 0.05 days/year). This would have resulted in Daves Creek being classified as a Type III-like stream rather than a Type II stream. Since the Type II streams are the ones in which the largest, rare flows are the most effective, the net result of using a fixed width class-based, rather than an eventbased, methodology would have been to misidentify these streams, and to mistakenly assume that all streams in the study had frequent effective discharges. 6.3 Linking the Bankfull and Effective Discharges All of the results obtained in this study must be considered as pertaining to channels that are not in equilibrium or steady-state conditions. The issues of the ordering of sediment transport events, the supply of sediment to the channel from outside sources such as landslides, and the nature of sediment transport in small mountain watersheds, particularly sediment waves or pulses (Gilbert 1917, Cui et al, 2003) mean that the effective discharge magnitudes and recurrence intervals determined in this study are likely specific to the interval of record and to the channel conditions. This position is at odds with the standard definition of effective discharge in the hydrologic literature (as summarized, for instance in Barry et al. (2008)), which holds that the effective discharge, since it moves the most sediment over the long term, is therefore the channel-forming discharge, the discharge or narrow range of discharges in the long term responsible for the form of the channel. The results of this study suggest that instead, for mountain streams, there is not a single channel-forming discharge which is responsible for the long-term shape of the channel, because the channel does not have a long-term shape.  141  There is instead a succession in time of shorter-term channel forms, each in the process of adapting to local fluxes of water and sediment. Within any interval of measuring discharges the channel may be quasi-stable at some times, aggrading at some times and degrading at others. In each of these phases of channel evolution, there may be a separate or multiple conceptual channel forming discharges. It is clear that in only a few of the streams in this study (Type II streams) does the effective discharge approximate the bankfull discharge and hence bear any relation to the theoretical channel-forming discharge; in the rest, at best it approximates a channel-maintaining discharge, and at worst it is geomorphically meaningless. For most of the studied streams, the effective discharge is significantly less than the bankfull discharge. Where the effective discharge is greater than bankfull, it is because of a lack of fine sediment to be transported, resulting in Type II-like conditions. In most of the studied streams, neither the bankfull nor the effective discharge corresponds to mean annual flood. That bankfull discharge is significantly greater than effective discharge for most of the studied streams corresponds to the finding of Church and Slaymaker (1989) that most streams in British Columbia with watershed areas < 300 km² have undergone or are undergoing significant degradation resulting from the relaxing of conditions of paraglacial sedimentation. It is therefore the case that these results may also be expected to apply to small mountain watersheds in other regions that were also extensively glaciated during the Pleistocene. In order to place the results from this study in a larger context, the data from Figure 5-21 were replotted against values for Qbf and Qeff from several other studies that presented values of both bankfull and effective discharge from regions in the mountainous western part  142  of North America, including ding Colorado (Andrews, 1980; Torizzo and Pitlick, Pitlick 2004), Idaho (Whiting et al., 1999), Idaho and Wyoming (Emmett and Wolman, 2001) and California (Nolan et al., 1987) (Figure ure 6-7). In the other studies, data plots closely ly around arou the 1:1 line where Qbf=Qeff. Where the he dat data in the other studies diverges from the 1:1 line, li it does so in the direction Qbf>Qeff, but ut it ddoes not display the range of multiple orders ers of magnitude that data from this study does. s. The range of drainage basin areas, channel gradien radients and sediment sizes in these other studies es are generally comparable to those in this study y (with (wit the exception of Nolan et al. 1987, in which the sediment was much finer), although some of o the drainages areas in the other studiess were in the range 100-1000 km². Some of thee difference diff between  Figure 6-7: Qeff vs. Qbf values es fro from this study compared to results from other published publis studies from western North America.  143  studies can probably be attributed to the use of class-based, rather than event-based, methodology to calculate the effective discharge, which would lead the other studies to underestimate the magnitude and overestimate the frequency of effective discharge for Type II streams. The most immediately apparent difference between this study and the other studies is that in this study the streams were consistently subject to continental ice sheet cover during Pleistocene glaciation, and hence subjected to glacial and paraglacial processes during and after such glaciation. In terms of landscape evolution, the streams in this study are much younger than the streams in the other studies; they have existed in their present form for less than 10,000 years. It must be emphasized again that the effective discharges estimated for the streams in this study apply to only the period of record of discharge, an interval ranging from 20 to 60 years. The conceptual channel-forming discharge, applying to an equilibrium stream that moves sediment in a transport-limited manner, is only poorly related to the effective discharges determined in this study. For a generalized stream of size, steepness and biogeoclimate similar to those in this study, it can be seen that the effective discharge for any short- to medium-term finite interval will and must depend on the magnitude and ordering of flood events and sediment inputs over that interval, and that this effective discharge will therefore be specific to the period of record rather than representing a close approximation of a theoretical channel-forming discharge over the entire lifetime of the channel. The sensitivity of the underlying Wilcock and Kenworthy (2002) sediment transport equations, where relatively small changes in the representative size of the coarse fraction of sediment (D84) and proportion of gravel in the surface (Fg) (a change in D84 of 20mm, or a change in Fg of 10%) might result in a large change in the estimated magnitude and  144  frequency of Qeff, offers further evidence that effective discharge should not be understood as a static property of a stream, or thought of as time-invariant. In relatively small headwater streams such as these, where there is direct or indirect coupling between the stream and adjacent hillslopes, large, episodic inputs of sediment of substantially different composition and size from the equilibrium bed material in the channel can lead to different effective discharges at different intervals as sediment waves move downstream through the system. The effective discharge was evaluated for each stream under the underlying assumption that the sediment conditions observed represented the median long-term values of the entire periods of record, from 20 to 60 years. For most of the studied streams, inputs of finer sediment would increase the amount of sediment transported and lower the thresholds for transport initiation, reducing the calculated frequency of the effective discharge. For instance, a Type II stream with a large amount of fine sediment introduced into it might act as a Type I stream until the introduced sediment had been evacuated from the channel, possibly rapidly or over many years, with a gradual reduction in the frequency of effective discharge and corresponding increase in magnitude over that interval. The time scale for a stream to respond to such a disturbance can range from months to thousands of years (Hassan and Zimmermann, 2011). It is incorrect to assume that the studied streams are in equilibrium with sediment supply; rather, the current (observed) conditions represent the sum of a sequence of sediment supply events over time, with each event causing perturbation and subsequent recovery. Within the studied streams, Coquitlam River and Dennis Creek represent end members of a conceptual spectrum of sediment supply, sediment transport capability, and channel response (Hassan and Zimmermann, 2011). In Coquitlam River, where sediment  145  supply is very high as a result of an upstream landslide, the supply of sediment is greater than the stream can transport, and aggradation has occurred, resulting in a small and frequent bankfull discharge; the notional effective discharge is equaled or exceeded for almost half the year. In Dennis Creek, the supply of sediment is minimal and the channel is transitional to a bedrock stream. The gravel that is present in Dennis Creek’s channel appears to represent immobile lag sediment; the sediment transport relation predicts that Dennis Creek has only been able to transport sand-sized sediment a few times over the period of record, approximately one day per year, and the effective discharge is equaled or exceeded 0.04 days per year, or approximately one day per 25 years. The Wilcock and Kenworthy (2002) sediment transport equations presume a systematic relation between discharge, sediment size and fractional distribution, and sediment transport. In response to disturbances, mountain stream channels can change their bankfull dimensions, channel gradient, sediment size, or bedform type (Buffington, 2011). This bedform type factor is not well-captured by the Wilcock and Kenworthy (2002) equations. Mountain rivers often respond to a decrease in sediment supply by forming bed structures as well as by coarsening, such as during the production of an armour layer. Coarsening is accounted for in the Wilcock and Kenworthy (2002) sediment transport model but effects of structure are not explicitly modeled. East Creek provides an example of such a change over time in the magnitude and frequency of the calculated effective discharge. East Creek’s bed sediment was sampled twice, in September and December 2006. In between the two field visits a flood occurred, in November 2006, with an estimated magnitude approximating the 1 in 5 year flood event (Caulkins 2010, Blair 2010). As part of a separate experiment, colourfully painted magnetic  146  tracer stones were present in the bed material of East Creek during both field visits; comparison of photos taken in September and December show that many of the larger stones did not move, and that most bedforms in the channel (such as point bars) remained in their location after the flood. However, areas of localized scour were present after the flood, exposing a dense, compact silty basal till that underlies the channel of East Creek. Bankfull dimensions were essentially unchanged at the surveyed cross-sections (although in the lowergradient reach downstream of the studied reach, some surveyed bed elevations changed by up to 10 cm (Blair, 2010)), but sediment size changed dramatically. The proportion of the bed sediments composed of gravel >8mm in diameter decreased from 89% in September to 75% in December, and calculated D50 and D84 decreased from 44.6 mm and 82.1mm to 25 and 46.4mm respectively. These changes in sediment size fractions resulted in estimated Qeff and Qh values decreasing from 2.31 and 1.97 m3/s in September to 0.64 and 0.72 m3/s in December; the calculated frequency of the effective discharge went from 0.07 days per year to 5.9 days per year. East Creek in September was a Type II stream; with the decrease in sediment size after the flood, it was Type I stream in December. Subsequent qualitative observations suggest that the bed of East Creek coarsened substantially in the following year, returning to a condition similar to the observations made during September 2006 by September 2007. These observations are consistent with the passage of a pulse or wave of finer sediment through the channel and may represent typical seasonal variation in sediment transport (Hassan and Church, 2001). The differences in sediment size and consequent estimation of effective and half-load discharge frequency and magnitude for East Creek in September and December 2006 present an example of relatively short-duration (scale of months) temporal variation. The differences  147  in estimated bankfull, effective and half-load discharge magnitude and frequency for the stream reaches of Albert River above and below the large woody debris jam present a complementary example of spatial variation over a scale range of hundreds of meters. Taken together, these two examples represent the spatial and temporal variability of estimates of bankfull and effective discharge. This relatively high variability over relatively short distances and timescales suggests a reason why no significant effects related to measures of drainage-basin scale sediment supply were observed. The frequency and magnitude of bankfull and effective discharges in these small mountain streams are affected by the relative supply of such sediment, but the effects of drainage-basin scale sediment supply are masked by the local effects. For instance, a recent landslide a few kilometers upstream of a studied reach (such as what was observed at Coquitlam River) can result in a locally high sediment supply, and consequent aggradation and frequent, low-magnitude Qbf and Qeff, even if overall sediment supply in the drainage basin is low. This again emphasizes Beven’s (1981) contention that it is the sequence or ordering of events (floods and sediment supply events) which determines geomorphic effectiveness, not simply their magnitude. In their discussion of Buffington (2011), Hassan and Zimmermann (2011) note that natural mountain channels, subjected to constantly varying sediment supply and streamflow, oscillate over time between a state in which the channel bed is actively exchanging grains with sediment in transport and a state in which transported sediment overpasses the bed without exchanging grains with it. This study has characterized stream channels at various stages in that process, and has observed some examples of the spatial and temporal variability inherent in the process. Under such constantly changing, non-equilibrium  148  conditions it is not surprising that a concept (effective discharge) developed for stable channels in regime reveals its limitations.  149  Chapter 7: Conclusions 7.1 Frequency and Magnitude of Bankfull and Effective Discharge Small mountain streams in British Columbia show substantial variability in the magnitude and frequency of bankfull discharge. Across all the studied streams, bankfull discharge is infrequent, with a median frequency of 0.32 days per year, but there is a high degree of variability present, with the total range of observed frequencies ranging from 0.00001 to 13 days per year. This variability is influenced by hydroclimatic regime, drainage basin size and channel width and gradient. Hydroclimatic regime can further be subdivided into mechanism of flood generation (snowmelt or rainfall), unit runoff, and coefficient of variation of peak flows. In relatively dry, snowmelt-dominated streams with narrow channels and small drainage basins bankfull discharge is likely to be an infrequent and large event, while in relatively wet, rainfall-dominated, low-gradient streams with large drainage basins and wide channels bankfull discharge is likely to be a more frequent event of lower magnitude. In these small mountain streams, there is a continuum of alluvial character, with past glacial activity and present colluvial processes contributing coarse and relatively immobile sediment to channels which range from strongly bedrock-controlled to purely alluvial in character. Across this continuum of process, empirically determined values of channel width and of the ratio of reference (Q2) streampower to reference grain size of approximately 5-8m and 10,000 kg/s³ respectively demarcate a transition from generally poorly-developed to generally well-developed alluvial character. In streams with welldeveloped alluvial character, the bankfull discharge is a smaller and more frequent event than in streams with poorly-developed alluvial character.  150  Bedload effective discharge in small gravel-bed mountain streams in British Columbia is frequent, equalled or exceeded on average for 26 days per year (7.25% of the time), but the frequency and corresponding magnitude of effective discharge are also highly variable, with observed frequencies of exceedance ranging from 0.04 to 173 days per year. Effective discharge occurs with greater frequency in rainfall-dominated, Coastal watersheds, and is less frequent in snowmelt-dominated interior watersheds. Unlike bankfull discharge, the frequency of effective discharge does not vary significantly with drainage basin area, channel gradient, or other watershed-scale geomorphic parameters. Three main types of streams are present, each with a different characteristic effective discharge regime. In the Type I streams, gravel transport is a frequent process, and dominates the transport of sediment. In the Type II streams, gravel transport is a rare process but dominates when it occurs, leading to the effective discharge occurring amongst the largest flows on record. In Type III streams gravel transport is exceedingly rare and so graveltransporting flows do not transport enough sediment, averaged over time, to become the effective discharge; the effective discharge occurs in sand-transporting flows even though the channel is gravel-bedded. Distinguishing between these three regimes in the field is difficult, but they become apparent when applying the fractional sediment transport relations against the observed flow duration curve. The use of event-based classes for the calculation of effective discharge in this study was useful in detecting those streams in which the effective discharge is a rare, large event. The failure of some previous studies to detect streams with large, infrequent effective discharges might result from the use of the class-based methodology (which used a low and fixed number of classes to determine the effective discharge) rather than the absence of low-  151  frequency, high-magnitude effective discharges. Type I and Type III streams have similar frequencies and magnitudes of effective discharge even though gravel transport is frequent in the first type of stream and vanishingly rare in the third type. The use of a single effective discharge frequency or magnitude as shorthand is insufficient to adequately describe the fluvial processes occurring in such a stream because it does not indicate the type of sediment being transported. The half-load discharge provides additional information concerning the shape and skew of the effectiveness relation; the effective discharge is the mode of the effectiveness relation and the half-load discharge is the median. For the streams in this study, using the half-load and effective discharge together was sufficient to distinguish Type II streams from Type I and Type III streams, but still insufficient to distinguish between Type I and Type III streams. The magnitude of the half-load discharge, unlike the effective discharge, is correlated with the Wohl (2004) streampower/D84 alluvial character parameter as well as with the magnitude of the bankfull discharge. Half-load discharge in some streams approximated bankfull discharge, but in most of the studied streams it was a significantly smaller and more frequent event. The geomorphic meaning  of the half-load discharge  remains uncertain, especially in cases where the half-load discharge corresponds to a discharge that itself transports little sediment, such as a half-load discharge that occurs at a low point between two peaks of the effectiveness relation.  152  7.2 Bankfull and Effective Discharge as Proxies for Channel-Forming Discharge In the classical hydrologic literature, following the arguments of Wolman and Miller (1960), for self-formed alluvial streams the bankfull discharge and the effective discharge approximate each other and the mean annual flood and thus are measurable equivalents of the dominant or channel-forming discharge. For most of the streams in this study, neither bankfull nor effective discharge are close approximations of the mean annual flood, and the bankfull discharge rarely equals the effective discharge. In most cases the bankfull discharge is larger and less frequent than the mean annual flood while the effective discharge is smaller and more frequent. This general statement is a simplification which omits some information; the range of variability observed in these streams is more interesting than statements about average behaviour. For Type I and Type III streams, Qbf>Qeff, and there is a correlation between frequency and magnitude of bankfull discharge and degree of incision of the stream: where Qbf>>Qeff it is because Qbf is large and infrequent. For Type II streams, the reverse is true: Qbf ≤ Qeff. Few of these Type II streams are deeply incised: Qeff ranges from 0.8 to 1.9 times Q2, and Qbf is from 0.6 to 1.5 times Q2. (The range of frequencies this represents depends on whether flow duration or annual maximum series is used and on hydroclimatic region, but in general terms 0.6Q2 corresponds to a frequent flow, exceeded 10 to 20 days per year on average, while 1.9Q2 corresponds to a rare flood exceeded once per 10 to 20 years.) The Type II streams therefore seem to be the only streams in the study that broadly meet Wolman and Miller (1960)’s expectation that Qeff and Qbf would have similar magnitudes and Emmett and Wolman (2001)’s contention that Qbf ≤ Qeff for gravel-bed streams. In the Type II streams, therefore, the measured bedload effective discharge appears to represent a reasonable approximation for the channel-forming discharge. These Type II streams, then,  153  may be stable streams that have stopped degrading as they reach dynamic equilibrium (temporarily or permanently) with their sediment supply. In the Type I streams, the gravel bed sediment moves frequently on average, but the effective discharge is smaller and more frequent (sometimes much smaller and much more frequent) than the bankfull discharge. These streams are actively degrading, and have not yet reached equilibrium with the supply of sediment, especially paraglacial sediment, present in their drainage basins. For the Type I streams the magnitude and frequency of the bankfull discharge correlate with geomorphic and hydroclimatic parameters of the drainage. The magnitude of the bankfull discharge is not completely random, so there is some geomorphic meaning to the top of bank despite the degradational conditions. In these Type I streams, the effective discharge may correspond to a channel-maintaining discharge but it does not correspond to a channel-forming discharge. The bankfull discharge may correspond to the channel-forming discharge, but the disconnect between the effective and bankfull discharge, and the relative high frequency of sediment transport suggests that the bankfull discharge should be considered a function of both the frequency and magnitude of the channel-forming discharge and the time since the last channel-forming event. A large flood forms or reforms the channel; subsequent smaller floods erode into the sediment moved by the large flood and re-establish bedforms and structure in the bed, possibly with a corresponding increase in the apparent magnitude of the bankfull discharge if the bed is lowered relative to the banks. These streams are not in an equilibrium state; they cycle back and forth between periods of mobile beds and/or aggradation during sediment inputs and periods of immobile beds and/or degradation between such inputs.  154  The Type III streams are in some ways similar to the Type I streams, but distinguished by the immobility of the gravel in the bed. These are streams that share some similarities to bedrock streams because they flow over material which they do not transport, and the material they do transport is small relative to the bed sediment. The evolution of these streams since deglaciation is uncertain; they may have formed by degrading into glaciogenic sediment until they reached some sort of equilibrium condition in which sediment inputs and outputs balanced, in which case some of the gravel in the bed represents immobile lag sediment. Alternatively, they may represent streams that under glaciated conditions or under wetter postglacial climate conditions had sufficient stream flow and energy to transport gravel, and that have lost that ability under present conditions, in which case the gravel in the bed is glaciofluvial or postglacial fluvial sediment that is presently immobile. In either case, the effective discharge in these streams is demonstrably not a channel-forming discharge since it cannot mobilize the majority of the bed sediment. It is likely a purely channel-maintaining discharge, capable of transporting only the current sediment supply and hence maintaining the channel in its present form but not changing or reforming the channel. Small mountain streams display a diversity of channel morphologies over relatively short spatial scales; for instance, two pool and riffle reaches may be separated by a step-pool reach (Hassan et al., 2005). Therefore, within a single stream, variation in the frequency and magnitude of bankfull and effective discharge should be expected, particularly for the streams for which alluvial character is poorly developed. Likewise, two stream reaches that might be classified as Type I with respect to frequency of effective discharge might be separated by a reach that is Type II or Type III in nature.  155  The three hydroclimatic regions considered in this study each displayed a characteristic pattern and amount of precipitation and associated runoff, such that hydroclimatic processes were considered to be relatively uniform across each region. Heterogeneities in bedrock geology, elevation and precipitation occurred within and between regions and these, combined with variable patterns of sediment supply to the channel in space and time, likely account for some of the variability observed within each hydroclimatic region. Overall, the patterns of effective and bankfull discharge frequency and magnitude observed in this study suggest that (except for Type II streams) neither value provides a particularly good analogue for the channel-forming discharge in small mountain streams. The concept of channel-forming discharge was developed and has the greatest relevance for large self-formed, transport-limited rivers under equilibrium conditions. Small mountain streams are transitional between self-formed streams flowing over alluvial sediment and purely bedrock or colluvial streams flowing over material they are unable to transport. They often have high thresholds for the transport of sediment, or are supply-limited, with episodic sediment supply. They are not often in equilibrium, but undergo cycles of aggradation and degradation, or cycles of relative bed mobility and immobility. This study suggests that the banks of these small mountain streams reflect this lack of equilibrium conditions; most of the streams are in a state of degradation and bankfull discharge is correspondingly high. The effective discharge in most of these streams is a frequent event which suffices to maintain the stream channel; channel-forming events capable of substantially changing or reforming the stream channel are rare, and so appear ineffective when the period of record analyzed does not include such events.  156  The concepts of bankfull and effective discharge were developed from research conducted largely on transport-limited, self-formed alluvial channels, and these concepts and the expectations arising from them have often been applied to small mountain streams without considering the appropriateness of such transposition. In the past, this has resulted in inappropriate design criteria for applications such as drainage structures or stream restoration exercises. In addition to the clarification of the theory of channel-forming discharge as it applies to these small mountain streams, it is hoped that this study will result in improvements to such designs and applications. Climate change and land-use practices such as urbanisation or logging can result in measurable changes to the magnitude and frequency of streamflow, particularly floods (Alila et al., 2009). It can be difficult, however, to move from knowledge about the projected changes in the frequency and magnitude of streamflow to expectations about changes in stream channels and sediment transported by streams. The results and methods used in the present study can be complemented by quantitative descriptions of expected changes in the frequency and magnitude of post-disturbance streamflow to generate predictions of how such changes will affect flooding and sediment transport in the studied streams or other similar streams. 7.3 Suggestions for Further Research This study has examined watersheds distributed in space, and hence assessed variability associated with spatial characteristics such as channel gradient. There is obvious potential for a complementary study that would follow changes over time in a few stream reaches over a period several years to a decade and account for the factors resulting in changes of the estimates of bankfull, effective, and half-load discharge over time. Just  157  enough data was gathered in this study (from two samples of East Creek, before and after a relatively large flood) to confirm that these changes can be significant. This study used the Wilcock and Kenworthy (2002) two-fraction, surface-based transport model to estimate sediment transport. An obvious extension to this method would be to use a surface-based transport equation with more than two fractions, such as that presented by Wilcock and Crowe (2003), with an explicit size fraction consisting of sediment of boulder size (>256mm diameter). This would help to address the questions left unresolved from the current study relating to the characterization of the effects of very large sediment present in the channel which may move only during rare events, but with a threshold effect resulting in a very strong skew to the effectiveness relation when it does move. The third area identified for further research is to extend the investigation from this study to slightly larger drainages, specifically those with areas from 100-400 km². The results from Church and Slaymaker (1989) suggest that the disequilibrium between sediment transport and sediment yield should persist up to a drainage area of 300 km². (The results from this study relating bankfull channel width to alluvial character suggests that that area might vary with hydroclimatic region). It is therefore reasonable to expect that a more obvious scale break should be observed in these larger drainages, with well-developed alluvial banks becoming established by the time drainage area exceeds 300 km² in at least some hydroclimatic regions.  158  Bibliography Ackers P 1972. River regime: research and application. J. Inst. Wat. Engineers, 26, 257-281. Alila Y, A Mtiraoui 2002. Implications of heterogeneous flood frequency distributions  on  traditional  stream  discharge  prediction  techniques,  Hydrological Processes 16, 5, 1065-1084. Alila Y, PK Kuras, M Schnorbus, R Hudson 2009. Forests and floods: a new paradigm sheds light on age-old controversies. Water Resour. Res. 45, W08416, doi:10.1029/2008WR007207 Andrews ED 1980. Effective and bankfull discharges of streams in the Yampa River Basin, Colorado and Wyoming. J. Hydrology 46, 311-330 Andrews ED 1984. Bed material entrainment and hydraulic geometry of gravel-bed rivers in Colorado. Geol. Soc. Amer. Bull., 95, 371-378 Andrews ED, JM Nankervis 1995. Effective discharge and the design of channel maintenance flows for gravel-bed rivers. Natural and Anthropogenic Influences in Fluvial Geomorphology: Geophysical Monograph 89, American Geophys. Union. Ashmore PE, TJ Day 1988. Effective discharge for suspended sediment transport in streams of the Saskatchewan River Basin, Water Res. Res. 24, 6, 864–870 Baker VR 1977. Stream-channel response to floods with examples from central Texas. Geol Soc. Amer. Bull 88, 1057–1071. Baker VR 1988. Overview. In Baker VR, RC Kochel, PC Patton (eds.) Flood geomorphology. John Wiley and Sons. 528 pp.  159  Baker VR 1998. Catastrophism and uniformitarianism: logical roots and current relevance in geology. In Blundell DJ, AC Scott (eds) Lyell: the Past is the Key to the Present. Geol. Soc., London, Spec. Pub. 143, 171-182. Baker VR, G Pickup 1987. Flood geomorphology of the Katherine Gorge, northern territory, Australia. Geol Soc. Amer. Bull 98, 635–646. Barry JJ, JM Buffington, P Goodwin, JG King, WW Emmett 2008. Performance of bedload transport equations relative to geomorphic significance: predicting effective discharge and its transport rate. J. Hydraulic Eng 134, 5, 601-615 BC Ministry of Forests and Ministry of Environment, 2001. Forest road engineering guidebook, second edition. Victoria, BC, Queen's Printer. Beckers J, Y Alila, A Mtiraoui 2002. On the validity of the British Columbia Forest Practices Code Guidelines for stream culvert discharge design. Can. J. of For. Res., 32 684-692 Benson MA, DM Thomas 1966. A definition of dominant discharge. Bull. Of the Int. Assoc. Of Sci. Hydrol., 11, 76-80. Bernedo CE 2009. Some findings on effective discharge determination for gravel bed rivers. World Environmental and Water Resources Congress 2009: Great Rivers. ASCE Conf. Proc. 342, 315 Beven K, 1981. The effect of ordering on the geomorphic effectiveness of hydrologic events. Erosion and Sediment Transport in Pacific Rim Steeplands, IAHS Publication 132.  160  Biedenharn DS, CR Thorne, PJ Soar, RD Hey, CC Watson, 2000. A practical guide to effective discharge calculation. In: Watson, C.C. et al. (Ed.), Demonstration Erosion Control: Design Manual. Engineer Research and Development Center, US Army Corps of Engineers, Vicksburg, MS, pp. 239–274. Biedenharn DS, CR Thorne, PJ Soar, RD Hey, CC Watson, 2001, Effective Discharge Calculation Guide, International Journal of Sediment Research, 16 (4): 445-459. Blair TJ, 2010. Stage discharge estimation using a 1D river hydraulic model and spatially-variable roughness. Unpublished MASc thesis, University of British Columbia. Brardinoni F, MA Hassan 2006. Glacial erosion, evolution of river long profiles, and the organization of process domains in mountain drainage basins of coastal British Columbia. J. Geophys. Res. 111, F01013, 12pp. Bravo-Espinosa M, WR Osterkamp, VL Lopes 2003. Bedload transport in alluvial channels. J. Hydraul. Eng 129, 10, 783-795 Bray DI, 1972. Generalized regime-type analysis of Alberta rivers. Unpublished PhD thesis, University of Alberta. Buffington JM, 2011. River channel changes over human timescales. In: Gravel bed rivers 7. In press. Bull WB 1988. Floods; degradation and aggradation. in Flood Geomorphology, Baker VR, RC Kochel, PC Patton (eds.)  161  Bunte K, SR Abt, 2001. Sampling surface and subsurface particle-size distributions in wadable gravel-and cobble-bed streams for analyses in sediment transport, hydraulics, and streambed monitoring. Gen. Tech. Rep. RMRS-GTR-74. Fort Collins,CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. Bunte K, SR Abt, JP Potyondy, SR Ryan, 2004. Measurement of coarse gravel and cobble transport using portable bedload traps. J. Hydraul. Eng., 130, 879–893. Carling P, 1988. The concept of dominant discharge applied to two gravel-bed streams in relation to channel stability thresholds. Earth Surf. Proc. and Land. 13, 355–367. Castro JM, PL Jackson 2001. Bankfull discharge recurrence intervals and regional hydraulic geometry relations: patterns in the Pacific Northwest, USA. J American Water Resources Assoc. 37, 5, 1249-1262 Caulkins JL 2010. Sediment transport and channel morphology within a small, forested stream: East Creek, British Columbia. Unpublished PhD thesis, University of British Columbia. Chow VT 1950. Discussion of: Annual floods and the partial-duration flood series. Trans. Am. Geophys. Union 31, 6, 939-941. Church M, 2002. Geomorphic thresholds in riverine environments. Freshwater Biology 47, 541-557 Church M, DG McLean, JF Wolcott 1987. River bed gravels: sampling and analysis. In: Thorne CR, JC Bathurst, RD Hey (eds.) Sediment transport in gravel-bed rivers, John Wiley and Sons. pp. 43-88.  162  Church M, JM Ryder 1972. Paraglacial sedimentation: a consideration of fluvial processes conditioned by glaciation. Geol. Soc. Amer. Bull. 83, 10, 30593072 Church M, O Slaymaker 1989. Disequilibrium of Holocene sediment yield in glaciated British Columbia. Nature 337, 2, 452-454. Clague JJ, TS James 2002. History and isostatic effects of the last ice sheet in southern British Columbia. Quaternary Science Rev. 1, 1-3, 71-87. Collins BD, DR Montgomery 2011. The legacy of Pleistocene glaciation and the organization of lowland alluvial process domains in the Puget Sound region. Geomorphology 126, 1-2, 174-185. Costa JE, JE OConnor 1995. Geomorphically effective floods. Natural and Anthropogenic  Influences  in  Fluvial  Geomorphology:  Geophysical  Monograph 89. American Geophysical Union. Crowder DW, HV Knapp, 2005. Effective discharge recurrence intervals of Illinois streams. Geomorphology 64, 167–184 Cui Y, G Parker, TE Lisle, J Gott, ME Hansler-Ball, JE Pizzuto, NE Allmendinger, JM Reed, 2003. Sediment pulses in mountain rivers: 1. Experiments, Water Resour. Res., 39, 9, 1239-1250 Curran JH, EE Wohl 2003. Large woody debris and flow resistance in step-pool channels, Cascade Range, Washington. Geomorphology 51, 141–157 Dodov B, E Fofoula-Georgiou 2005. Fluvial processes and streamflow variability: Interplay in the scale-frequency continuum and implications for scaling. Water Resources Research, 41 (5), 18pp.  163  Dunne T, LB Leopold 1978. Water in Environmental Planning. W.H. Freeman and Col, San Francisco Dury GH, 1961. Bankfull discharge: an example of its statistical relations. Bulletin International Association of Scientific Hydrology VI, 3 48. Dury GH, 1973. Magnitude-frequency analysis and channel morphology. In: Morisawa, M. (Ed.), Fluvial Geomorphology.Suny, Binghamton, pp. 91–121. Dury GH, 1976. Discharge prediction, present and former, from channel dimensions. Journal of Hydrology, 30, 219–245. Dury GH, 1977. Underfit streams: retrospect, perspect, and prospect. In: Gregory KJ (ed.) River channel changes, John Wiley, New York, pp.81–293. Dutnell, RC 2000. Development of bankfull discharge and channel geometry relations for natural channel design in Oklahoma using a fluvial geomorphic approach. Unpublished M.Sc. Thesis, University of Oklahoma. Eaton B, M Church, D Ham 2002. Scaling and regionalization of flood flows in British Columbia, Canada. Hydrological Processes, 16, 3245-3263. Einstein HA, 1950. The bed load function for sediment transport in open channel flows, Technical Bulletin 1026, U.S. Department of Agriculture, Soil Conservation Service, Washington, DC. Elliott LP, BW Brook, 2007. Revisiting Chamberlain: multiple working hypotheses for the 21st century. BioScience 57, 608-614. Emmett WW, 1980. Bedload sampling in rivers. Proceedings, International Symposium on River Sedimentation, Beijing 24–28 March. Guanghu Press: Beijing; 991–1014.  164  Emmett WW, MG Wolman 2001. Effective discharge and gravel-bed rivers. Earth Surf. Proc. Land. 26, 1369-1380 Finnegan NJ, G Roe, DR Montgomery, B Hallett 2005. Controls on the channel width of rivers: implications for modeling fluvial incision of bedrock. Geology 33, 3, 229-232 Gilbert GK 1877. Report on the geology of the Henry Mountains. US Geological Survey, Rocky Mountain Region, Washington DC. Gilbert GK 1917. Hydraulic mining debris in the Sierra Nevada. US Geological Survey Professional Paper 105. Goodwin P, 2004. Analytical solutions for estimating effective discharge. J. Hydraulic Eng. 130, 8, 729-738. Graham DJ, SP Rice, I Reid 2005. A transferable method for the automated grain sizing of river gravels. Water Res. Research 41, 12 pp. Grant, GE, FJ Swanson, MG Wolman 1990. Pattern and origin of stepped-bed morphology in high-gradient streams, western Cascades, Oregon. Geol. Soc. Amer. Bull., 102, 340-352. Gupta A 1983. High-magnitude floods and stream channel response. In: Collinson JD, J Lewin (eds.) Modern and Ancient Fluvial Systems. Oxford, Blackwell. 219-227. Gupta VK, EC Waymire 1990. Multiscaling properties of spatial rainfall and river flow distribution. J. Geophys. Res. 95, 3 1999–2009.  165  Hardy T, P Panja, D Mathias 2005. WinXSPRO, A Channel Cross Section Analyzer, User's Manual, Version 3.0 Gen. Tech. Rep. RMRS-GTR-147. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. Harris RR 1999. Defining reference conditions for restoration of riparian plant communities: examples from California, USA. Env. Mgmt 24, 1, 55-63. Harvey AM 1969. Channel capacity and adjustment of streams to hydrologic regime. J. of Hydrology 9, 82–98. Haschenburger JK, SP Rice 2004. Changes in woody debris and bed material texture in a gravel-bed channel. Geomorphology 60, 241-267 Hassan MA, AE Zimmermann 2011. Channel response and recovery to changes in sediment supply: Discussion of Channel change in mountain rivers over human time scales by JM Buffington. Gravel Bed Rivers 7, in press. Hassan MA, M Church 2000. Experiments on surface structure and partial sediment transport on a gravel bed. Water Res Research 36, 1885-1895. Hassan MA, M Church 2001. Sensitivity of bed load transport in Harris Creek: Seasonal and spatial variation over a cobble-gravel bar. Water Res Research 37, 813-826 Hassan MA, M Church, TE Lisle, F Brardinoni, L Benda, GE Grant, 2005. Sediment transport and channel morphology of small forested streams.  166  Helie JF, DL Peters, KR Tattrie, JJ Gibson 2003. Review and synthesis of potential hydrologic impacts of mountain pine beetle and related harvesting activities in British Columbia. Canadian Forest Service, Mountain Pine Beetle Initiative Working Paper 2005-23. Water and Climate Impacts Research Center, Environment Canada, Ottawa. Hey RD 1997. Channel Response and Channel Forming Discharge. Final Report. US Army Corps of Engineers, R&D6871-EN-01, European Research Office, London, United Kingdom, 108 pp. Howard AD, 1980. Thresholds in river regimes. In: Coates, DR, JD Vitek (eds) Thresholds in Geomorphology. pp. 227 258. Inglis CC 1949. The behavior and control of rivers and canals. Research Publication, vol. 13. Central Waterpower Irrigation and Navigation Research Station, Poona, India, pp. 79–91. Jakob M, P Jordan 2001. Design flood estimates in mountain streams – the need for a geomorphic approach. Can. J. Civ. Eng. 28: 425–439 Jarrett RD 1984: Hydraulics of high-gradient streams. Journal of Hydraulic Engineering, 110: 1519-1539. Jarrett RD 1990. Hydrologic and hydraulic research in mountain rivers. Water Resources Bulletin, 26, 3419-429. Paper no. 89106. Johnson PA, TM Heil 1996. Uncertainty in estimating bankfull conditions. Journal of the American Water Resources Association 32, 6, 1283–1291. Kennedy BA 1972. Bankfull discharge and meander forms. Area, 4, 209-212.  167  Kennedy RG 1895. Prevention of silting in irrigation canals. Inst. Civil Eng. Proc. 119, pp. 281-290. Kidson R, KS Richards 2005. Flood frequency analysis: assumptions and alternatives. Progress in Physical Geography 29, 3, 392-410 Kilpatrick FA, HH Barnes Jr. 1964. Channel geometry of piedmont streams as related to frequency of floods. Physiographic and Hydraulic Studies of Rivers, USGS Professional Paper 422-E. 11pp. Klemes V, 1974.Some problems in pure and stochastic hydrology. In Proc. of a Symposium on Statistical Hydrology, Misc. Publ. 1275, pp. 2-15, US Dept. of Agriculture. Klemes V, 2000. Tall tales about tails of hydrological distributions I. J Hydrol. Eng. 5, 3, 227-232 Kochel RC, 1988. Geomorphic impact of large floods: review and new perspectives on magnitude and frequency. In: Baker VR, RC Kochel, PC Patton (Eds.), Flood Geomorphology. Wiley, New York, pp. 169–187. Lacey G, 1933. Stable channels in alluvium. Inst. Civil Eng. Proc., 229, 1, 259-384. Lacey G, 1939. Regime flow in incoherent alluvium. Central Board Irr. (India), Publication 20, Simla. Langbein WB, 1949. Annual floods and the partial-duration series. Trans Am Geophys Union 30, 6, 879-881. Lenzi MA, L Mao, F Comiti, 2006. Effective discharge for sediment transport in a mountain river: computational approaches and geomorphic effectiveness. Journal of Hydrology 326, 257-276  168  Leopold LB, 1992. Sediment size that determines channel morphology. In: Billi P, RD Hey, CR Thorne, P Tacconi. (Eds.), Dynamics of Gravel-bed Rivers. Wiley, Chichester, pp. 297–307. Leopold LB, GM Wolman, JP Miller 1964. Fluvial Processes in Geomorphology. W.H. Freeman and Co.: San Francisco. Leopold LB, HE Skibitzke 1967. Observations on unmeasured rivers. Geografiska Annaler 49A, 247-255. Leopold LB, T Maddock 1953. The hydraulic geometry of stream channels and some physiographic implications. US Geol. Surv. Prof. Paper 252, 57p. Marcus WA, K Roberts, L Harvey, G Tackman 1992. An evaluation of methods for estimating Manning’s n in small mountain streams. Mountain Research and Development 12, 13, 227– 239. Marriott SB, J Alexander 1999 (eds.) Floodplains: interdisciplinary approaches. Geol. Soc. Spec. Pub. 163, Bath, United Kingdom McCandless TL, RA Everett 2002. Maryland stream survey: bankfull discharge and channel characteristics in the Piedmont hydrologic region. U.S. Fish and Wildlife Service, Annapolis, MD. CBFO-S02-02. Melone AM, 1985. Flood producing mechanisms in coastal British Columbia, Can. Water Resour. Jour 10, 3, 47–64. Merritts DJ, KR Vincent, EE Wohl 1994. Long river profiles, tectonism, and eustasy: a guide to interpreting fluvial terraces. J. Geophys. Res. 99, 7, 14031-14050. Millar RG 2005. Theoretical regime equations for mobile gravel-bed rivers with stable banks. Geomorphology 64, 207-220  169  Miller AJ, 1990. Flood hydrology and geomorphic effectiveness in the central Appalachians. Earth Surf. Proc. Land. 15, 119-134 Miller JP, 1958. High mountain streams: effects of geology on channel characteristics and bed material. State Bureau of Mines and Mineral Resources/New Mexico Institute of Mining and Technology Memoir, vol. 4. 57 pp. Montgomery DR, JM Buffington 1997. Channel-reach morphology in mountain drainage basins. Geological Society of America Bulletin 109, 5, 596– 611. Nanson RA, GC Nanson, H Qing-Huang 2010. The hydraulic geometry of narrow and deep channels: evidence for flow optimization and controlled peatland growth. Geomorphology 117, 143-154. Nash DB 1994. Effective sediment-transporting discharge from magnitude-frequency analysis. J. Geol. 102, 79–95. Navratil O, M-B Albert, E Herouin, J-M Gresillon (2006). Determination of bankfull discharge magnitude and frequency: comparison of methods on 16 gravel-bed river reaches. Earth Surf Proc Landforms, 31, 1345-1363. Newson M, 1980. The geomorphic effectiveness o floods – a contribution stimulated by two recent events in Mid-Wales. Earth Surf. Proc. Land. 5, 1-16. Nixon M, 1959, A study on the bank-full discharges of rivers in England and Wales, Proc. Inst. Civil Eng., 12, 157-174. Nolan, KM, TE Lisle, HM Kelsey, 1987. Bankfull discharge and sediment transport in northwestern California. Erosion and sedimentation in the Pacific Rim. IAHS Publication 165, 439–449.  170  Nordin CF Jr.1985. The sediment loads of rivers. In Rodda JC (ed.) Facets of hydrology. John Wiley and Sons, New York, 183–204. Nordin, CF Jr., JP Beverage, 1965. ‘Sediment transport in the Rio Grande New Mexico.’ US Geological Survey Professional Paper 462-F, 35. Parker G, PF Wilcock, C Paola, WE Dietrich, J Pitlick, 2007. Physical basis for quasi-universal relations describing hydraulic geometry of single-thread gravel bed rivers. J Geophys Res 112, F04005, 21 pp. Petit F, A Pauquet 1997. Bankfull discharge recurrence interval in gravel-bed rivers. Earth Surface Processes and Landforms 22, 7, 685–693. Phillips JD 2002. Geomorphic effects of flash flooding in a forested headwater basin. J. Hydrol. 269, 236-250. Pickup G, WA Reiger, 1979. A conceptual model of the relation between channel characteristics and discharge. Earth Surf. Proc. Land. 4, 37– 42. Pickup G, Warner RF. 1976. Effects of hydrologic regime on magnitude and frequency of dominant discharge. Journal of Hydrology 29: 51–75. Pilon PJ, KD Harvey 1994. Consolidated frequency analysis (CFA) version 3.1 reference manual. Environment Canada, Ottawa ON. Pitlick J, 1988, Response of coarse bed rivers to large floods in Colorado and California [Ph.D. dissert.] Colorado State University, 137 p. Pitlick J, 1992. Flow resistance under conditions of intense gravel transport. Water Resources Research 28, 891–903. Radecki-Pawlik A, 2002. Bankfull discharge in mountain streams: Theory and practice. Earth Surf. Proc. Land. 27 (2): 115–123.  171  Rao DV, 1981. Return period for mean annual hydrological event. J Hydraulics Div, Proc Amer Civ Eng, 107, 366-369. Richards KS, 1982. Rivers: form and process in alluvial channels. Methuen: London. Riedel JL, RA Haugerud, JJ Clague, 2007. Geomorphology of a Cordilleran ice sheet drainage network through breached divides in the North Cascade Mountains of Washington and British Columbia. Geomorphology 91, 1-17 Riley SJ, 1972. A comparison of morphometric measures of bankfull. J. Hydrol. 17, 23–31. Roberts CR, 1989. Flood frequency and urban-induced channel change: some British examples. In: Floods: Hydrological, Sedimentological and Geomorphological Implications. John Wiley & Sons: New York, p 57-82. Robison EG, RL Beschta 1990. Characteristics of coarse woody debris for several coastal streams of southwest Alaska. Can J. Fish Aquatic Sci, 47: 1684-1693 Rosgen D, 1996. Applied river morphology. Wildland Hydrology, Pagosa Springs, Colorado. Schaffernak F, 1922. Neue Grundlagen fur die Berechung der Qeschisbefuhrung in Flusslaufen. Franz Deuticke, Leipzig,Germany. Schmudde TH, 1968. Floodplain. In Fairbridge, RW (ed) The Encyclopedia of Geomorphology, Reinhold, New York, 359-362. Schumm SA 1960.The shape of alluvial channels in relation to sediment type, U.S. Geol. Surv. Prof Pap. 352-B, 30 pp. Schumm SA 1985. Patterns of alluvial rivers. Ann. Rev. Earth Planet. Sci., 13, 5-27.  172  Schumm SA, 1979. Geomorphic thresholds: the concept and its applications. Trans. Inst. British Geographers, new series. 4, 4, 485-515. Schmidt FL, 1996. Statistical significance testing and cumulative knowledge in psychology: Implications of training on researchers. Psychol. Methods 1, 115129. Shields A. 1936. Application of similarity principles and turbulence research to bedload movement. Ott WP and Uchelechen JC (trans.) Mitteilungen der Preussischen Versuchsanstalt für Wasserbau und Schiffbau, 43 pp Sichingabula HM, 1999.Magnitude-frequency characteristics of effective discharge for suspended sediment transport, Fraser River, British Columbia, Canada. Hydrol. Process. 13, 1361-1380 Soar PJ, CR Thorne 2001. Channel restoration design for meandering rivers. US Army Corps of Engineers Coastal and Hydraulics Laboratory, Publication ERDC/CHL CR-01-1. Washington, DC, 454 pp. Takeuchi K, 1984. Annual maximum series and partial-duration series: analysis of Langbein’s formula and Chow’s discussion. J. Hydrol, 68, 275-284. Timber-Fish-Wildlife (TFW) Ambient Monitoring Program Manual. 1994. Edited by D Schuett-Hames, A Pleus, L Bullchild, S Hall. Northwest Indian Fisheries Commission, Olympia WA Torizzo M, J Pitlick 2004. Magnitude-frequency of bedload transport in mountain streams in Colorado. Journal of Hydrology 290, 137–151 Turowski JM, N Hovius, A Wilson, M-J Horng 2008. Hydraulic geometry, river sediment and the definition of bedrock channels. Geomorphology 99, 26-38  173  US Army Corps of Engineers, 2008. HEC-RAS River Analysis System. Hydraulic Reference Manual CPD-69 Version 4.0. US Environmental Protection Agency, 2005. Watershed Assessment of River Stability  &  Sediment  Supply  (WARSSS).  http://www.epa.gov/WARSSS/sedsource/bankfull.htm accessed September 4, 2007 USDA Forest Service, 1995. A guide to field identification of bankfull stage in the western United States. Stream Systems Technology Centre. Video: online at http://www.stream.fs.fed.us/publications/videos.html accessed September 4, 2007.. Vogel RM, JR Stedinger, RP Hooper, 2003. Discharge indices for water quality loads. Water Res. Research 39, 10, 1293-1302 Waylen, P, MK Woo, 1982. Prediction of annual floods generated by mixed processes. Water Resour. Res. 18, 1283–1286. Waylen, P, MK Woo, 1984. Regionalization and prediction of floods in the Fraser River catchment, B.C. Water Res. Bull. 20, 941–949. Werrity A, 1997. Short-term changes in channel stability. In: Thorne CR, RD Hey, MD Newson (eds.) Applied fluvial geology for river engineering and management. John Wiley and Sons Ltd. pp.47-65 Whelan F, 2000. Analysis of land use/land cover and the frequency of bankfull flow in selected salmon habitat recovery streams in the Pacific Northwest using GIS. Unpublished PhD thesis, Oregon State University.  174  Whipple KX, 2004. Bedrock rivers and the geomorphology of active orogens. Annual Reviews of Earth and Planetary Science, 32, p. 151–185. Whiting PJ, JF Stamm, DB Moog, RL Orndorff 1999. Sediment-transporting flows in headwater streams. Geol. Soc. Amer. Bull., 111, 3, 450-466. Whiting PJ, WE Dietrich, 1990. Boundary shear stress and roughness over mobile alluvial beds. Journal of Hydraulic Engineering 116, 1495–1511. Wilcock  PR, ST Kenworthy 2002. A two-fraction model for the transport of  sand/gravel mixtures. Water Res. Research, 38, 1194-1205 Wilcock, PR, JC Crowe 2003. Surface-based transport model for mixed-size sediment. J. Hydraulic Eng. 129, 2, 120-128 Wilford DJ, ME Sakals, JL Innes, RC Sidle, WA Bergerud 2004. Recognition of debris flow, debris flood and flood hazard through watershed morphometrics. Landslides 1,1, 61-66. Williams GP 1978. Bankfull discharge of rivers. Water Resources Research, 14(6): 1141-1154. Wohl EE 2004. Limits of downstream hydraulic geometry. Geology 32, 897-900. Wolman MG 1954. A new method of sampling coarse river-bed material. Transactions of the American Geophysical Union. Wolman MG 1955. The natural channel of Brandywine Creek, Pennsylvania. US Geological Survey Professional Paper 271. Wolman MG, JP Miller 1960. Magnitude and frequency of forces in geomorphic processes, J. Geology, 68, 54-74.  175  Wolman MG, LB Leopold 1957. River flood plains: some observations on their formation. US Geological Survey, US Government Printing Office Washington. Wolman MG, R Gerson 1978. Relative scales of time and effectiveness of climate in watershed geomorphology. Earth Surface Processes 3, 189-208. Woodyer KD, 1968. Bankfull frequency in rivers. J. Hydrol. 6, 114–142.  176  Appendix A: Hydrologic Data from East Creek, Malcolm Knapp Research Forest Almost all of the streams included in this study had hydrometric data provided by Water Survey of Canada. The one exception is East Creek, located in UBC’s Malcolm Knapp Research Forest. A streamflow gauge has been maintained for approximately 30 years in East Creek by Dr. Michael Feller of the UBC Forestry Department. Dr. Feller provided copies of the original drum recordings of streamflow to Andre Zimmermann and Joshua Calkins, both at that time PhD students of the UBC Geography Department, and they digitized the data and shared their results for this study. The stream gauge location in the East Creek drainage basin is several hundred meters upstream of the stream reach surveyed in this study; an appropriate scaling factor based on the relative ratios of the two drainage areas raised to the 0.8 power was used to transform the observed discharges into values representative of the surveyed reach. As this data is not otherwise publicly available, it is summarized here for the benefit of other researchers (Tables A-1 and A-2). There is one notable error in the data presented here: out of the approximately 5492 days of record, approximately 2100 days had low flow which was below a (scaled) threshold of 0.11 m3/s and which was not recorded. In other words, it is only known that the discharge was less than 0.11 m3/s on those days. There are also, however, approximately another 2400 days for which discharges less than 0.11 m3/s were recorded. On the whole it would be reasonable to expect the distribution of discharge frequency and magnitude for unobserved days below the threshold to follow the corresponding distribution of the observed days, but to be conservative all the unobserved days have been assigned a discharge of 0.10 m3/s in Table A-2. Because the threshold for initiation of sediment transport in the studied reach is  177  approximately 0.30 m3/s, the assigning of this specific discharge value to the unobserved days does not affect the determination of the effective discharge for East Creek.  Table A-1 East Creek Peak Discharges by Magnitude and Year  Peak Flow Data by year Water Year Estimated Peak Flow (m3/s) 1971 3.28 1972 3.53 1973 2.86 1974 1.86 1975 3.53 1976 1.90 1977 1.95  Water Year Estimated Peak Flow (m3/s) 1989 5.38 1990 2.54 1991 2.72 1992 2.01 1993 2.04 1994 2.51 1995 2.38  178  Table A-2: East Creek Daily Discharge Classes by Frequency Discharge Class (m3/s) 0-0.01 0.01-0.02 0.02-0.03 0.03-0.04 0.04-0.05 0.05-0.06 0.06-0.07 0.07-0.08 0.08-0.09 0.09-0.1 0.1-0.11 0.11-0.12 0.12-0.13 0.13-0.14 0.14-0.15 0.15-0.16 0.16-0.17 0.17-0.18 0.18-0.19 0.19-0.2 0.2-0.21 0.21-0.22 0.22-0.23 0.23-0.24 0.24-0.25 0.25-0.26 0.26-0.27 0.27-0.28 0.28-0.29 0.29-0.3 0.3-0.31  Number of Days 580 349 265 220 190 173 126 122 114 109 2246 62 61 68 53 49 37 47 35 36 28 26 31 24 19 27 17 19 15 23 20  Average Discharge (m3/s) 0.004 0.015 0.025 0.035 0.045 0.055 0.065 0.075 0.085 0.095 0.100 0.115 0.125 0.135 0.145 0.155 0.165 0.174 0.185 0.196 0.205 0.215 0.225 0.235 0.244 0.255 0.264 0.275 0.284 0.296 0.305  Discharge Class (m3/s) 0.31-0.32 0.32-0.33 0.33-0.34 0.34-0.35 0.35-0.36 0.36-0.37 0.37-0.38 0.38-0.39 0.39-0.4 0.4-0.41 0.41-0.42 0.42-0.43 0.43-0.44 0.44-0.45 0.45-0.46 0.46-0.47 0.47-0.48 0.48-0.49 0.49-0.5 0.5-0.51 0.51-0.52 0.52-0.53 0.53-0.54 0.54-0.55 0.55-0.56 0.56-0.57 0.57-0.58 0.58-0.59 0.59-0.6 0.6-0.61 0.61-0.62  Number of Days 15 13 11 6 10 6 8 10 10 11 12 3 6 7 5 6 3 6 8 5 4 3 3 6 3 5 2 4 1 5 4  Average Discharge (m3/s) 0.315 0.325 0.334 0.343 0.356 0.364 0.376 0.383 0.394 0.406 0.415 0.426 0.435 0.444 0.456 0.465 0.475 0.487 0.495 0.505 0.512 0.524 0.533 0.543 0.558 0.567 0.576 0.588 0.594 0.602 0.618  Discharge Class (m3/s) 0.62-0.63 0.63-0.64 0.64-0.65 0.65-0.66 0.66-0.67 0.68-0.69 0.69-0.7 0.7-0.71 0.71-0.72 0.72-0.73 0.73-0.74 0.74-0.75 0.76-0.77 0.77-0.78 0.78-0.79 0.81-0.82 0.83-0.84 0.84-0.85 0.85-0.86 0.87-0.88 0.88-0.89 0.89-0.9 0.9-0.91 0.91-0.92 0.93-0.94 0.96-0.97 0.97-0.98 0.98-0.99 0.99-1 1.01-1.02 1.03-1.04  Number of Days 5 6 3 2 2 5 2 4 2 2 3 2 1 2 1 2 2 1 1 1 2 2 1 3 1 1 1 1 2 1 2  Average Discharge (m3/s) 0.625 0.635 0.646 0.653 0.667 0.684 0.693 0.703 0.716 0.725 0.733 0.743 0.765 0.773 0.781 0.813 0.839 0.842 0.855 0.873 0.884 0.891 0.906 0.914 0.931 0.970 0.971 0.980 0.996 1.014 1.031  Discharge Class (m3/s) 1.04-1.05 1.06-1.07 1.09-1.1 1.1-1.11 1.12-1.13 1.2-1.21 1.23-1.24 1.24-1.25 1.25-1.26 1.26-1.27 1.27-1.28 1.33-1.34 1.38-1.39 1.39-1.4 1.41-1.42 1.45-1.46 1.48-1.49 1.58-1.59 1.63-1.64 1.64-1.65 1.78-1.79 1.85-1.86 1.92-1.93 1.93-1.94 1.95-1.96 1.96-1.97 2.04-2.05 2.08-2.09 2.15-2.16 2.23-2.24 2.28-2.29 2.3-2.31  Number of Days 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  Average Discharge (m3/s) 1.042 1.065 1.092 1.109 1.123 1.207 1.235 1.245 1.254 1.268 1.272 1.335 1.382 1.394 1.416 1.456 1.483 1.582 1.635 1.644 1.785 1.852 1.923 1.932 1.952 1.968 2.044 2.088 2.155 2.237 2.288 2.306  179  Appendix B: Examples of Channel Conditions and Sediment Size Distributions The discussion of the streams in this study in Chapters 3 and 4 is complemented in this appendix by photographs, graphs and tables illustrating channel conditions (including in-channel sediment) and apparent banks, and presenting information related to the size distribution of gravel-fraction (>8mm) sediment in the channel, based on analysis of the pooled distribution of individual sediment photos. A representative cross-section is also presented for each stream. The photographs of in-channel sediment are presented as illustrative examples, and are not necessarily the photos most characteristic of each stream. For some streams, such as Norish Creek and Hidden Creek, large boulders are visible in the stream bed in the oblique channel photos which are larger than the dimensions of the frames used for the photographs of the sediment. The photographed sediment is generally typical of the mobile fraction of sediment in the channel. The sediment size distributions presented are truncated; they are only the distribution of the gravel-sized fraction (clasts >8mm) of sediment present in the channel. These sediment size distributions should be considered along with the information on proportion of sand-size sediment present in the channel (Fs) from Table B-1 when considering the total sediment size distribution for the streams.  180  Cabin Creek channel  Cabin Creek sediment  Percentile  mm  Phi  Psi  5  15.40  -3.95  3.95  16  28.21  -4.82  4.82  1308  25  40.70  -5.35  5.35  1307.5  50  77.33  -6.27  6.27  75  130.99  -7.03  7.03  84  176.00  -7.46  7.46  95  237.37  -7.89  7.89  100  284.22  -8.15  8.15  Cabin Creek grain size percentiles  1309  Elevation (m)  1308.5  1307 1306.5 1306 1305.5 1305 1304.5 0  10  20 Distance (m)  Cabin Creek representative cross-section  30  40  Hosmer Creek channel  Hosmer Creek sediment 1122  Percentile  mm  Phi  Psi  5  10.90  -3.45  3.45  16  15.57  -3.96  3.96  25  19.92  -4.32  4.32  50  38.76  -5.28  5.28  75  77.85  -6.28  6.28  1119.5  84  113.00  -6.82  6.82  1119  95  275.14  -8.09  8.09  100  401.20  -8.65  Hosmer Creek grain size percentiles  8.65  Elevation (m)  1121.5 1121 1120.5 1120  0  5  10 Distance (m)  Hosmer Creek representative cross-section  15  20  Albert River above jam channel sediment  Percentile  mm  Phi  Psi  5  9.29  -3.22  3.22  16  12.11  -3.60  3.60  25  14.60  -3.87  3.87  50  22.14  -4.47  4.47  75  33.72  -5.08  5.08  84  40.71  -5.35  5.35  95  58.08  -5.86  5.86  100  153.41  -7.26  7.26  1308.5 1308 Elevaion (m)  Albert River channel above jam  1307.5 1307 1306.5 1306  Albert River above jam grain size percentiles  0  10  20  30  40  Distance (m)  Albert River above jam representative cross section  50  60  Albert River channel below jam  Albert River below jam channel sediment 1303.5  Percentile  mm  Phi  Psi 1303  5  10.45  -3.38  3.38  16  16.87  -4.08  4.08  25  22.81  -4.51  4.51  50  41.53  -5.38  5.38  75  72.08  -6.17  6.17  1300.5  84  91.73  -6.52  6.52  1300  95  150.33  -7.23  7.23  Elevation (m)  1302.5 1302 1301.5 1301  0  5  10  15  20  25  Distance (m)  100  215.75  -7.75  Albert River below jam grain size percentiles  7.75 Albert River below jam representative cross section  30  35  Sullivan Creek channel  Sullivan Creek sediment  Percentile  mm  Phi  Psi  5  11.61  -3.54  3.54  16  18.00  -4.17  4.17  25  23.69  -4.57  4.57  50  40.66  -5.35  5.35  75  73.02  -6.19  6.19  84  99.40  -6.64  6.64  95  148.86  -7.22  7.22  884  100  259.74  -8.02  Sullivan Creek grain size percentiles  8.02  Elevation (m)  883.5 883 882.5 882 881.5 881 0.000  5.000  10.000 Distance (m)  Sullivan Creek representative cross-section  15.000  20.000  Arrow Creek channel  Arrow Creek sediment 769  Percentile  mm  Phi  Psi  5  11.62  -3.54  3.54  16  19.06  -4.25  4.25  25  26.01  -4.70  4.70  50  51.74  -5.69  5.69  75  88.63  -6.47  6.47  767  84  110.39  -6.79  6.79  766.5  95  162.56  -7.34  7.34  100  285.97  Arrow Creek grain size percentiles  -8.16  8.16  Elevation (m)  768.5 768 767.5  0  5  10  15  Distance (m)  Arrow Creek representative cross-section  20  25  Split Creek sediment  Percentile  mm  Phi  Psi  5  10.27  -3.36  3.36  16  13.78  -3.78  3.78  25  16.92  -4.08  4.08  50  28.33  -4.82  4.82  75  55.13  -5.78  5.78  84  78.95  -6.30  6.30  95  174.84  -7.45  7.45  100  260.19  Split Creek grain size percentiles  -8.02  8.02  Elevation (m)  Split Creek channel  1003 1002.5 1002 1001.5 1001 1000.5 1000 999.5 999 998.5 998 997.5 0  10  20  30 Distance (m)  Split Creek representative cross-section  40  50  60  Redfish Creek channel  Redfish Creek sediment 577  Percentile  mm  Phi  Psi  5  13.09  -3.71  3.71  16  26.03  -4.70  4.70  25  38.70  -5.27  5.27  50  80.34  -6.33  6.33  75  125.52  -6.97  6.97  575  84  163.19  -7.35  7.35  574.5  95  237.94  -7.88  7.88  100  347.75  -8.44  Redfish Creek grain size percentiles  8.44  Elevation (m)  576.5 576 575.5  0  5  10  15 Distance (m)  Redfish Creek representative cross-section  20  25  30  Hidden Creek channel Percentile  mm  Hidden Creek sediment Phi  919  Psi  918.5  10.50  -3.39  3.39  16  44.91  -5.49  5.49  25  74.57  -6.22  6.22  50  173.16  -7.44  7.44  75  292.15  -8.19  8.19  84  368.31  -8.52  8.52  95  446.87  -8.80  8.80  100  493.42  -8.95  8.95  Hidden Creek grain size percentiles  918 Elevation (m)  5  917.5 917 916.5 916 915.5 0  5  10  15  20  Distance (m)  Hidden Creek representative cross-section  25  30  35  Anderson Creek channel near Nelson Percentile  mm  Phi  Anderson Creek (near Nelson) sediment 750  Psi  749.5  10.96  -3.45  3.45  16  17.70  -4.15  4.15  25  25.77  -4.69  4.69  50  53.18  -5.73  5.73  75  98.08  -6.62  6.62  747  84  124.70  -6.96  6.96  746.5  95  176.76  -7.46  7.46  100  195.90  -7.61  7.61  Anderson Creek near Nelson grain size percentiles  Elevation (m)  5  749 748.5 748 747.5  0  5  10  15  Distance (m)  Anderson Creek near Nelson representative cross-section  20  Fell Creek channel near Nelson Percentile  mm  Fell Creek (near Nelson) sediment Phi  797  Psi  796.5  11.08  -3.47  3.47  16  18.79  -4.23  4.23  25  26.12  -4.71  4.71  50  59.66  -5.90  5.90  75  111.94  -6.81  6.81  84  138.89  -7.12  7.12  95  215.28  -7.75  7.75  796 Elevation (m)  5  795.5 795 794.5 794  100  266.24  -8.06  Fell Creek near Nelson grain size percentiles  8.06  793.5 793 0  2  4  6 Distance (m)  Fell Creek representative cross-section  8  10  12  Deer Creek sediment photo  Percentile  mm  Phi  Psi  5  12.26  -3.62  3.62  16  21.77  -4.44  4.44  25  30.14  -4.91  4.91  50  56.21  -5.81  5.81  75  94.31  -6.56  6.56  84  124.17  -6.96  6.96  95  278.21  -8.11  8.11  100  300.12  -8.23  8.23  702.2 702 701.8 Elevation (m)  Deer Creek channel  701.6 701.4 701.2 701 700.8 700.6 700.4 700.2 0  5  10  15 Distance (m)  Deer Creek grain size percentiles  Deer Creek representative cross-section  20  25  30  Coldstream Creek channel  Coldstream Creek sediment photo 597.5  Percentile  mm  Phi  Psi  5  10.47  -3.39  3.39  16  14.19  -3.83  3.83  25  17.96  -4.17  4.17  50  31.89  -4.99  4.99  75  60.40  -5.92  5.92  595  84  74.77  -6.22  6.22  594.5  95  127.12  -6.99  6.99  100  155.07  -7.28  Coldstream Creek grain size percentiles  7.28  Elevation (m)  597 596.5 596 595.5  0  5  10  15 Distance (m)  Coldstream Creek representative cross-section  20  25  30  Daves Creek channel Percentile  mm  Daves Creek sediment photo Phi  1087.5  Psi  1087  5  12.36  -3.63  3.63  16  21.80  -4.45  4.45  25  30.80  -4.94  4.94  50  60.40  -5.92  5.92  75  126.68  -6.99  6.99  1084.5  84  156.75  -7.29  7.29  1084  95  251.88  -7.98  7.98  100  259.48  Daves Creek grain size percentiles  -8.02  8.02  Elevation (m)  1086.5 1086 1085.5 1085  0  2  4  6  8  Distance (m)  Daves Creek representative cross-section  10  12  14  Dennis Creek channel Percentile  mm  Dennis Creek sediment photo Phi  1796  Psi  1795.5  14.28  -3.84  3.84  16  40.98  -5.36  5.36  25  63.84  -6.00  6.00  50  119.99  -6.91  6.91  75  171.40  -7.42  7.42  1793.5  84  202.58  -7.66  7.66  1793  95  307.29  -8.26  8.26  100  313.31  -8.29  8.29  Dennis Creek grain size percentiles  Elevation(m)  5  1795 1794.5 1794  0  5  10 Distance (m)  Dennis Creek representative cross-section  15  20  Two Forty One Creek channel  Two Forty One Creek sediment photo 1596.4  Percentile  mm  Phi  Psi  5  11.49  -3.52  3.52  16  18.55  -4.21  4.21  25  26.97  -4.75  4.75  50  53.73  -5.75  5.75  75  105.63  -6.72  6.72  1595.4  84  132.84  -7.05  7.05  1595.2  95  207.44  -7.69  7.69  100  244.37  -7.93  Two Forty One Creek grain size percentiles  7.93  Elevation (m)  1596.2 1596 1595.8 1595.6  0  5  10 Distance (m)  Two Forty One Creek representative cross-section  15  20  Two Forty Creek channel  Two Forty Creek sediment photo 1625  Percentile  mm  Phi  Psi  5  11.22  -3.49  3.49  16  20.20  -4.34  4.34  25  31.94  -5.00  5.00  50  80.93  -6.34  6.34  75  156.35  -7.29  7.29  1623  84  166.55  -7.38  7.38  1622.5  95  231.41  -7.85  7.85  100  238.93  -7.90  7.90  Elevation (m)  1624.5 1624 1623.5  0  2  4  6 Distance (m)  Two Forty Creek grain size percentiles  Two Forty Creek representative cross-section  8  10  12  Corning Creek channel  Corning Creek sediment photo  Percentile  mm  Phi  Psi  5  11.90  -3.57  3.57  16  20.74  -4.37  4.37  25  29.68  -4.89  4.89  50  61.97  -5.95  5.95  75  107.32  -6.75  6.75  84  160.58  -7.33  7.33  95  271.95  -8.05  8.05  100  414.37  -8.69  8.69  407.5  Elevation (m)  407 406.5 406 405.5 405 404.5  Corning Creek grain size percentiles  0  5  10  15  Distance (m)  Corning Creek representative cross-section  20  25  Greata Creek sediment photo  Percentile  mm  Phi  Psi  5  10.14  -3.34  3.34  16  13.74  -3.78  3.78  25  16.62  -4.05  4.05  50  25.59  -4.68  4.68  75  39.40  -5.30  5.30  84  54.08  -5.76  5.76  95  106.18  -6.72  6.72  941.5 941 940.5 Elevation (m)  Greata Creek channel  940 939.5 939 938.5 938 937.5  100  139.41  -7.12  Greata Creek grain size percentiles  7.12  937 0.00  5.00  10.00  15.00  Distance (m)  Greata Creek representative cross-section  20.00  25.00  Beak Creek sediment photo  Percentile  mm  Phi  Psi  5  11.21  -3.49  3.49  16  17.55  -4.13  4.13  25  24.35  -4.61  4.61  50  52.38  -5.71  5.71  75  112.53  -6.81  6.81  84  138.32  -7.11  7.11  95  193.91  -7.59  7.59  100  237.88  -7.89  7.89  Beak Creek grain size percentiles  1018 1017.8 1017.6 Elevation (m)  Beak Creek channel  1017.4 1017.2 1017 1016.8 1016.6 1016.4 1016.2 1016 0  5  10 Distance (m)  Beak Creek representative cross-section  15  20  Camp Creek sediment photo  Percentile  mm  Phi  Psi  5  12.27  -3.62  3.62  16  22.15  -4.47  4.47  25  32.73  -5.03  5.03  50  63.58  -5.99  5.99  75  110.72  -6.79  6.79  84  138.06  -7.11  7.11  95  204.36  -7.67  7.67  982.4 982.2 982 Elevation(m)  Camp Creek channel  981.8 981.6 981.4 981.2 981 980.8  100  256.07  Camp Creek grain size percentiles  -8.00  8.00  980.6 0  5  10 Distance (m)  Camp Creek representative cross-section  15  20  Pennask Creek sediment photo  Percentile  mm  Phi  Psi  5  10.33  -3.37  3.37  16  14.17  -3.82  3.82  25  17.58  -4.14  4.14  50  27.65  -4.79  4.79  75  42.82  -5.42  5.42  84  50.64  -5.66  5.66  95  71.21  -6.15  6.15  100  170.26  -7.41  7.41  1437.6 1437.4 1437.2 Elevation (m)  Pennask Creek channel  1437 1436.8 1436.6 1436.4 1436.2 1436 1435.8 1435.6 0  5  10 Distance (m)  Pennask Creek grain size percentiles  Pennask Creek representative cross-section  15  20  Guichon Creek channel Percentile  mm  Guichon Creek sediment photo Phi  1169.4  Psi  1169.2  5  9.94  -3.31  3.31  16  13.16  -3.72  3.72  25  16.76  -4.07  4.07  50  39.51  -5.30  5.30  75  122.49  -6.94  6.94  1168.2  84  134.61  -7.07  7.07  1168  95  218.78  -7.73  7.73  100  281.83  -8.14  Guichon Creek grain size percentiles  8.14  Elevation (m)  1169 1168.8 1168.6 1168.4  0  2  4  6  8 Distance (m)  Guichon Creek representative cross-section  10  12  14  16  Ambusten Creek sediment photo  Percentile  mm  Phi  Psi  5  3.13  -1.65  1.65  16  5.72  -2.52  2.52  25  7.82  -2.97  2.97  50  14.78  -3.89  3.89  75  25.34  -4.66  4.66  84  32.33  -5.01  5.01  95  73.05  -6.19  6.19  100  115.05  -6.85  6.85  1060 1059.5 1059 Elevation (m)  Ambusten Creek channel  1058.5 1058 1057.5 1057 1056.5 1056 1055.5 0  5  10  15 Distance (m)  Ambusten Creek grain size percentiles  Ambusten Creek representative cross-section  20  25  30  Anderson Creek at Hat Creek sediment photo  Percentile  mm  Phi  Psi  5  10.67  -3.42  3.42  16  16.73  -4.06  4.06  25  22.02  -4.46  4.46  50  54.01  -5.75  5.75  75  136.57  -7.09  7.09  84  186.71  -7.54  7.54  95  262.96  -8.01  8.01  1179.5 1179 1178.5 Elevation (m)  Anderson Creek at Hat Creek channel  1178 1177.5 1177 1176.5 1176 1175.5  100  299.14  -8.22  Anderson Creek at Hat Creek grain size percentiles  8.22  1175 0  5  10  15  20  25  Distance (m)  Anderson Creek at Hat Creek representative cross-section  30  35  Harris Creek channel  Harris Creek sediment photo 595.5  Percentile  mm  Phi  Psi  5  12.95  -3.69  3.69  16  22.99  -4.52  4.52  25  30.49  -4.93  4.93  50  55.01  -5.78  5.78  75  90.05  -6.49  6.49  593  84  105.58  -6.72  6.72  592.5  95  140.50  -7.13  7.13  100  155.27  Harris Creek grain size percentiles  -7.28  7.28  Elevation (m)  595 594.5 594 593.5  0  5  10  15 Distance (m)  Harris Creek representative cross-section  20  25  30  Coquihalla River channel Percentile  mm  Coquihalla River sediment photo Phi  798.5  Psi  798  5  12.67  -3.66  3.66  16  23.19  -4.54  4.54  25  33.35  -5.06  5.06  50  62.77  -5.97  5.97  75  113.44  -6.83  6.83  795.5  84  135.54  -7.08  7.08  795  95  199.38  -7.64  7.64  100  376.31  -8.56  Coquihalla River grain size percentiles  8.56  Elevation (m)  797.5 797 796.5 796  0  5  10  15  20 Distance (m)  Coquihalla River representative cross-section  25  30  35  40  Norrish Creek sediment photo  Percentile  mm  Phi  Psi  5  12.01  -3.59  3.59  16  21.87  -4.45  4.45  25  31.42  -4.97  4.97  50  69.70  -6.12  6.12  75  133.77  -7.06  7.06  84  185.96  -7.54  7.54  95  262.67  -8.04  8.04  274 273 272 Elevation (m)  Norrish Creek channel  271 270 269 268 267 266  100  295.02  -8.20  8.20  Norrish Creek grain size percentiles (not corrected for very large lag stones)  265 0  10  20  30 Distance (m)  Norrish Creek representative cross-section  40  50  60  Kanaka Creek channel Percentile  mm  Kanaka Creek sediment photo Phi  42.5  Psi  42  5  14.18  -3.83  3.83  16  30.02  -4.91  4.91  25  44.69  -5.48  5.48  50  88.99  -6.48  6.48  75  154.84  -7.27  7.27  39.5  84  195.93  -7.61  7.61  39  95  255.63  -8.00  8.00  100  291.07  -8.19  Kanaka Creek grain size percentiles  8.19  Elevation (m)  41.5 41 40.5 40  0  5  10  15  20  Distance (m)  Kanaka Creek representative cross-section  25  30  35  Coquitlam River channel Percentile  mm  Coquitlam River sediment photo Phi  270.5  Psi  270  5  10.66  -3.41  3.41  16  15.68  -3.97  3.97  25  19.29  -4.27  4.27  50  28.64  -4.84  4.84  75  40.20  -5.33  5.33  267.5  84  46.27  -5.53  5.53  267  95  63.53  -5.99  5.99  100  74.74  -6.22  Coquitlam River grain size percentiles  6.22  Elevation (m)  269.5 269 268.5 268  0  10  20  30  Distance (m)  Coquitlam River representative cross-section  40  50  Noons Creek sediment photo  Percentile  mm  Phi  Psi  5  11.09  -3.47  3.47  16  16.59  -4.05  4.05  25  21.47  -4.42  4.42  50  38.99  -5.29  5.29  75  70.04  -6.13  6.13  84  88.18  -6.46  6.46  95  136.84  -7.10  7.10  334 333.5 333 Elevation (m)  Noons Creek channel  332.5 332 331.5 331 330.5 330  100  187.62  Noons Creek grain size percentiles  -7.55  7.55  329.5 0  5  10  15 Distance (m)  Noons Creek representative cross-section  20  25  30  Upper Carnation Creek channel  Upper Carnation Creek sediment photo  Percentile  mm  Phi  Psi  5  10.19  -3.35  3.35  16  14.86  -3.89  3.89  25  19.60  -4.29  4.29  50  40.24  -5.33  5.33  75  106.47  -6.73  6.73  84  150.03  -7.23  7.23  139.6  95  404.91  -8.64  8.64  139.4  100  506.15  -8.98  8.98  141  Elevation (m)  140.8 140.6 140.4 140.2 140 139.8  0  Upper Carnation Creek grain size percentiles  2  4  6  8  10  Distance (m)  Upper Carnation Creek representative cross-section  12  14  16  Lower Carnation Creek channel  Lower Carnation Creek sediment photo 7.5  Percentile  mm  Phi  Psi  5  9.34  -3.22  3.22  16  12.03  -3.59  3.59  25  14.51  -3.86  3.86  50  23.29  -4.54  4.54  75  36.98  -5.21  5.21  5  84  48.26  -5.59  5.59  4.5  95  81.95  -6.36  6.36  100  147.35  -7.20  Lower Carnation Creek grain size percentiles  7.20  Elevation (m)  7 6.5 6 5.5  0  5  10  15  Distance (m)  Lower Carnation Creek representative cross-section  20  25  East Creek channel Percentile  mm  East Creek sediment photo Phi  136.6  Psi  136.5  5  11.54  -3.53  3.53  16  18.74  -4.23  4.23  25  25.04  -4.65  4.65  50  44.61  -5.48  5.48  75  71.28  -6.16  6.16  136  84  82.08  -6.36  6.36  135.9  95  104.30  -6.70  6.70  100  143.02  East Creek grain size percentiles  -7.16  7.16  Elevation (m)  136.4 136.3 136.2 136.1  0  1  2  3  4  Distance (m)  East Creek representative cross-section  5  6  7  Appendix C: Example of Bankfull Discharge Calculation This appendix presents a more detailed example of the calculation of bankfull discharge in the studied streams. Presentation of the calculations for every cross-section of every surveyed reach would be tedious; for the example presented here, Greata Creek, the overview map and location of cross-sections and channel banks are shown, and the associated surveyed dimensions and calculations used to determine the range of bankfull discharges for two of the twelve crosssections are given in full. This method therefore gives two of the calculations summarized in Figure 4-1 of Chapter 4. The two cross-sections chosen as examples are numbers 7 and 10, where the ranges of bankfull indicators present in the channel respectively create a narrow and wide range for the potential value of the bankfull discharge; as indicated in Figure 4-1, the overall value of bankfull discharge determined for the reach is that value which best satisfies (i.e. falls within) the greatest number of individual cross-section ranges. The overview map of the channel, showing the locations of the surveyed points and banks on each cross-section, the digital elevation model generated from the surveyed points, and the resampled straight-line cross-sections perpendicular to the direction of flow, is shown in Figure C-1. The channel gradient for the reach was estimated by comparing the elevation difference of the left bank, right bank, and thalweg over their respective lengths long the reach. The resultant slope gradients (m/m) were computed as 0.0217 (thalweg), 0.0237 (left bank slope) and 0.0259 (right bank slope); the mean of the three slope gradients was 0.0238 m/m, very close to the gradient of the left bank. This mean gradient was used as the assumed water surface gradient for simulating the stage-discharge relationship.  215  The analysis of cross-section 7 is presented first. The resampled cross-section is shown in 216  The analysis of cross-section 7 is presented first. The resampled cross-section is shown in Figure C-2. The range of field-observed indicators of bankfull corresponds to stages from 0.59 to 0.61m above the thalweg. The cross-section is analysed in WinXSPro using the Jarrett method to estimate Manning’s n and the output is presented in Table C-1 with the stage-discharge relationship presented in Figure C-3. The bankfull indicators and the stage-discharge relationship strongly indicate that the bankfull discharge is in the range of 1.27 to 1.35 m³/s for this cross-section.  941.5 Elevation (m)  940.5 939.5 938.5 937.5 936.5 0.00  5.00  10.00  15.00  20.00  25.00  Distance (m)  Figure C-2. Cross-section 7 for Greata Creek.  The analysis of cross-section 10 is presented deliberately for contrast with cross-section 7, as cross-section 10 is one for which the range of stages from bankfull indicators and the stagedischarge curve is wide. The resampled cross-section is shown in Figure C-4. The range of stages corresponding to the field observations of left and right banks is from 0.37 to 0.80m respectively. The output from WinXSPro’s analysis of this cross-section is presented in Table C-2 with the stage-discharge relationship shown in Figure C-5. Unlike cross-section 7, there is not a perceptible discontinuity in the stage-discharge relationship. The field-observed indicators of bankfull stage correspond to a range of potential bankfull discharges ranging from  217  0.33 to 2.67 m³/s, with the best field estimate of bankfull stage (0.58m above thalweg) corresponding to a discharge of 1.15 m³/s. The visible geometric breaks in the cross-section on Figure C-4 at stages of 1.0 to 1.36m correspond to terraces rather than active banks – the corresponding range of discharges is from 4.65 to 9.91 m³/s, which correspond to flood frequencies of >>Q500.  1 0.9 0.8  Stage (m)  0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0  0.5  1  1.5  2  2.5  3  3.5  Discharge (cms)  Figure C-3. Greata Creek cross-section 7 stage-discharge curve  941 Elevation (m)  940 939 938 937 936 0.00  2.00  4.00  6.00  8.00  10.00 12.00 Distance (m)  14.00  16.00  18.00  20.00  Figure C-4. Cross-section 10 for Greata Creek  218  For completeness’ sake, Table C-3 presents the results of the CFA analysis of peak flow for Greata Creek. The log-Pearson Type III, Wakeby, and nonparametric distributions best fitted the observed peak flows; the Greata Creek peak flows appear to follow a mixed distribution overall, with an apparent break at around Q10. The reach-averaged bankfull discharge value of 1.13 m³/s corresponds to an event with a return period of about 6 years. 1.2  1  Stage (m)  0.8  0.6  0.4  0.2  0 0  1  2  3  4  5  Discharge (m3/s)  Figure C-5 Greata Creek cross-section 10 stage-discharge curve  219  Table C-1: WinXSPro output for Greata Creek, cross-section 7  STAGE (m)  AREA (sq m)  PERIM (m)  WIDTH (m)  R (m)  DHYD (m)  SLOPE (m/m)  n  VAVG (m/s)  Q cms  SHEAR (N/sq m)  0.01  0  0.12  0.12  0  0  0.0238  0.183  0.02  0  1.12  0.02  0  0.45  0.44  0.01  0.01  0.0238  0.172  0.03  0  1.63  0.03  0.01  0.83  0.82  0.01  0.01  0.0238  0.16  0.05  0  2.64  0.04  0.02  1.21  1.19  0.02  0.02  0.0238  0.151  0.06  0  3.75  0.05  0.03  1.43  1.41  0.02  0.02  0.0238  0.143  0.09  0  5.31  0.06  0.05  1.58  1.55  0.03  0.03  0.0238  0.136  0.11  0.01  7.01  0.07  0.06  1.72  1.69  0.04  0.04  0.0238  0.132  0.13  0.01  8.62  0.08  0.08  1.82  1.79  0.04  0.05  0.0238  0.128  0.15  0.01  10.36  0.09  0.1  1.92  1.88  0.05  0.05  0.0238  0.125  0.17  0.02  12.05  0.1  0.12  1.98  1.94  0.06  0.06  0.0238  0.122  0.19  0.02  13.93  0.11  0.14  2.02  1.97  0.07  0.07  0.0238  0.12  0.21  0.03  15.93  0.12  0.16  2.06  2  0.08  0.08  0.0238  0.117  0.24  0.04  17.89  0.13  0.18  2.09  2.03  0.08  0.09  0.0238  0.116  0.26  0.05  19.82  0.14  0.2  2.13  2.06  0.09  0.1  0.0238  0.114  0.28  0.05  21.72  0.15  0.22  2.17  2.09  0.1  0.1  0.0238  0.112  0.3  0.06  23.58  0.16  0.24  2.2  2.12  0.11  0.11  0.0238  0.111  0.32  0.08  25.42  0.17  0.26  2.24  2.15  0.12  0.12  0.0238  0.11  0.33  0.09  27.22  0.18  0.28  2.28  2.18  0.12  0.13  0.0238  0.109  0.35  0.1  29  0.19  0.3  2.31  2.21  0.13  0.14  0.0238  0.108  0.37  0.11  30.76  0.2  0.33  2.35  2.25  0.14  0.15  0.0238  0.107  0.39  0.13  32.49  0.21  0.35  2.39  2.28  0.15  0.15  0.0238  0.106  0.4  0.14  34.19  0.22  0.37  2.42  2.31  0.15  0.16  0.0238  0.105  0.42  0.16  35.88  0.23  0.4  2.46  2.34  0.16  0.17  0.0238  0.104  0.44  0.17  37.54  0.24  0.42  2.5  2.37  0.17  0.18  0.0238  0.104  0.45  0.19  39.18  0.25  0.44  2.53  2.4  0.17  0.18  0.0238  0.103  0.47  0.21  40.8  0.26  0.47  2.57  2.43  0.18  0.19  0.0238  0.102  0.48  0.23  42.41  220  Table C-1: WinXSPro output for Greata Creek, cross-section 7  STAGE (m)  AREA (sq m)  PERIM (m)  WIDTH (m)  R (m)  DHYD (m)  SLOPE (m/m)  n  VAVG (m/s)  Q cms  SHEAR (N/sq m)  0.27  0.49  2.6  2.46  0.19  0.2  0.0238  0.102  0.5  0.24  43.99  0.28  0.52  2.64  2.49  0.2  0.21  0.0238  0.101  0.51  0.26  45.56  0.29  0.54  2.68  2.52  0.2  0.21  0.0238  0.101  0.53  0.29  47.12  0.3  0.57  2.71  2.55  0.21  0.22  0.0238  0.1  0.54  0.31  48.65  0.31  0.59  2.75  2.58  0.22  0.23  0.0238  0.1  0.56  0.33  50.18  0.32  0.62  2.79  2.61  0.22  0.24  0.0238  0.099  0.57  0.35  51.68  0.33  0.64  2.82  2.64  0.23  0.24  0.0238  0.099  0.58  0.38  53.18  0.34  0.67  2.86  2.67  0.23  0.25  0.0238  0.098  0.6  0.4  54.66  0.35  0.7  2.9  2.7  0.24  0.26  0.0238  0.098  0.61  0.42  56.13  0.36  0.72  2.93  2.73  0.25  0.26  0.0238  0.097  0.62  0.45  57.58  0.37  0.75  2.97  2.76  0.25  0.27  0.0238  0.097  0.64  0.48  59.03  0.38  0.78  3.01  2.79  0.26  0.28  0.0238  0.097  0.65  0.51  60.46  0.39  0.81  3.04  2.82  0.27  0.29  0.0238  0.096  0.66  0.53  61.88  0.4  0.84  3.08  2.86  0.27  0.29  0.0238  0.096  0.67  0.56  63.29  0.41  0.86  3.12  2.89  0.28  0.3  0.0238  0.096  0.69  0.59  64.69  0.42  0.89  3.15  2.92  0.28  0.31  0.0238  0.095  0.7  0.62  66.08  0.43  0.92  3.19  2.95  0.29  0.31  0.0238  0.095  0.71  0.65  67.46  0.44  0.95  3.23  2.98  0.3  0.32  0.0238  0.095  0.72  0.69  68.83  0.45  0.98  3.26  3.01  0.3  0.33  0.0238  0.094  0.73  0.72  70.19  0.46  1.01  3.3  3.04  0.31  0.33  0.0238  0.094  0.75  0.75  71.54  0.47  1.04  3.34  3.07  0.31  0.34  0.0238  0.094  0.76  0.79  72.89  0.48  1.07  3.37  3.1  0.32  0.35  0.0238  0.094  0.77  0.82  74.23  0.49  1.1  3.41  3.13  0.32  0.35  0.0238  0.093  0.78  0.86  75.56  0.5  1.14  3.45  3.16  0.33  0.36  0.0238  0.093  0.79  0.9  76.88  0.51  1.17  3.48  3.19  0.34  0.37  0.0238  0.093  0.8  0.94  78.19  0.52  1.2  3.52  3.22  0.34  0.37  0.0238  0.093  0.81  0.98  79.5  221  Table C-1: WinXSPro output for Greata Creek, cross-section 7  STAGE (m)  AREA (sq m)  PERIM (m)  WIDTH (m)  R (m)  DHYD (m)  SLOPE (m/m)  n  VAVG (m/s)  Q cms  SHEAR (N/sq m)  0.53  1.23  3.56  3.26  0.35  0.38  0.0238  0.092  0.82  1.01  80.73  0.54  1.26  3.6  3.29  0.35  0.38  0.0238  0.092  0.83  1.05  81.95  0.55  1.3  3.64  3.32  0.36  0.39  0.0238  0.092  0.84  1.1  83.17  0.56  1.33  3.68  3.36  0.36  0.4  0.0238  0.092  0.85  1.14  84.39  0.57  1.36  3.72  3.39  0.37  0.4  0.0238  0.091  0.86  1.18  85.6  0.58  1.4  3.76  3.43  0.37  0.41  0.0238  0.091  0.88  1.22  86.8  0.59  1.43  3.8  3.46  0.38  0.41  0.0238  0.091  0.89  1.27  88  0.6  1.47  3.84  3.49  0.38  0.42  0.0238  0.091  0.9  1.31  89.2  0.61  1.5  3.97  3.62  0.38  0.41  0.0238  0.091  0.89  1.33  88.28  0.62  1.54  4.17  3.82  0.37  0.4  0.0238  0.091  0.87  1.34  86.11  0.63  1.58  4.37  4.02  0.36  0.39  0.0238  0.092  0.85  1.35  84.23  0.64  1.62  4.58  4.22  0.35  0.38  0.0238  0.092  0.84  1.36  82.48  0.65  1.66  4.79  4.43  0.35  0.38  0.0238  0.092  0.83  1.37  81  0.66  1.71  4.9  4.54  0.35  0.38  0.0238  0.092  0.83  1.41  81.24  0.67  1.75  5.02  4.65  0.35  0.38  0.0238  0.092  0.83  1.46  81.52  0.68  1.8  5.13  4.77  0.35  0.38  0.0238  0.092  0.83  1.5  81.84  0.69  1.85  5.25  4.88  0.35  0.38  0.0238  0.092  0.84  1.55  82.19  0.7  1.9  5.36  4.99  0.35  0.38  0.0238  0.092  0.84  1.59  82.58  0.71  1.95  5.49  5.12  0.36  0.38  0.0238  0.092  0.84  1.64  82.83  0.72  2  5.59  5.21  0.36  0.38  0.0238  0.092  0.85  1.69  83.52  0.73  2.05  5.7  5.33  0.36  0.39  0.0238  0.092  0.85  1.75  83.97  0.74  2.11  5.82  5.44  0.36  0.39  0.0238  0.092  0.86  1.8  84.45  0.75  2.16  5.93  5.55  0.36  0.39  0.0238  0.092  0.86  1.86  84.95  0.76  2.22  6.05  5.67  0.37  0.39  0.0238  0.091  0.86  1.92  85.48  0.77  2.27  6.17  5.78  0.37  0.39  0.0238  0.091  0.87  1.98  86.03  0.78  2.33  6.28  5.9  0.37  0.4  0.0238  0.091  0.87  2.04  86.61  222  Table C-1: WinXSPro output for Greata Creek, cross-section 7  STAGE (m)  AREA (sq m)  PERIM (m)  WIDTH (m)  R (m)  DHYD (m)  SLOPE (m/m)  n  VAVG (m/s)  Q cms  SHEAR (N/sq m)  0.79  2.39  6.4  6.01  0.37  0.4  0.0238  0.091  0.88  2.1  87.2  0.8  2.45  6.51  6.13  0.38  0.4  0.0238  0.091  0.88  2.17  87.82  0.81  2.51  6.63  6.24  0.38  0.4  0.0238  0.091  0.89  2.23  88.45  0.82  2.58  6.75  6.35  0.38  0.41  0.0238  0.091  0.89  2.3  89.1  0.83  2.64  6.83  6.44  0.39  0.41  0.0238  0.091  0.9  2.38  90.16  0.84  2.7  6.91  6.51  0.39  0.42  0.0238  0.091  0.91  2.47  91.35  0.85  2.77  6.98  6.58  0.4  0.42  0.0238  0.09  0.92  2.56  92.53  0.86  2.84  7.06  6.65  0.4  0.43  0.0238  0.09  0.93  2.64  93.72  0.87  2.9  7.11  6.7  0.41  0.43  0.0238  0.09  0.94  2.74  95.19  0.88  2.97  7.17  6.75  0.41  0.44  0.0238  0.09  0.96  2.84  96.67  0.89  3.04  7.22  6.8  0.42  0.45  0.0238  0.089  0.97  2.94  98.15  0.9  3.1  7.27  6.84  0.43  0.45  0.0238  0.089  0.98  3.05  99.62  0.91  3.17  7.32  6.89  0.43  0.46  0.0238  0.089  0.99  3.15  101.08  0.92  3.24  7.38  6.94  0.44  0.47  0.0238  0.089  1  3.26  102.54  0.93  3.31  7.43  6.99  0.45  0.47  0.0238  0.089  1.02  3.37  103.99  0.94  3.38  7.48  7.04  0.45  0.48  0.0238  0.088  1.03  3.48  105.44  0.95  3.45  7.53  7.08  0.46  0.49  0.0238  0.088  1.04  3.59  106.88  0.96  3.52  7.59  7.13  0.46  0.49  0.0238  0.088  1.05  3.7  108.31  0.97  3.59  7.64  7.18  0.47  0.5  0.0238  0.088  1.06  3.82  109.75  0.98  3.67  7.69  7.23  0.48  0.51  0.0238  0.088  1.07  3.94  111.17  0.99  3.74  7.75  7.27  0.48  0.51  0.0238  0.088  1.09  4.06  112.59  1  3.81  7.8  7.32  0.49  0.52  0.0238  0.087  1.1  4.18  114.01  223  Table C-2: WinXSPro output for Greata Creek, cross-section 10. STAGE  AREA  PERIM  WIDTH  R  DHYD  SLOPE  (m)  (sq m)  (m)  (m)  (m)  (m)  (m/m)  n  VAVG  Q  SHEAR  (m/s)  cms  (N/sq m)  0.01  0  0.23  0.23  0  0  0.0238  0.183  0.02  0  1.14  0.02  0  0.47  0.46  0.01  0.01  0.0238  0.164  0.04  0  2.26  0.03  0.01  0.63  0.62  0.02  0.02  0.0238  0.151  0.06  0  3.69  0.04  0.02  0.78  0.77  0.02  0.02  0.0238  0.144  0.08  0  5.05  0.05  0.03  0.91  0.89  0.03  0.03  0.0238  0.138  0.1  0  6.45  0.06  0.03  1.03  1.01  0.03  0.03  0.0238  0.134  0.12  0  7.85  0.07  0.05  1.12  1.09  0.04  0.04  0.0238  0.13  0.14  0.01  9.44  0.08  0.06  1.2  1.17  0.05  0.05  0.0238  0.127  0.16  0.01  11.03  0.09  0.07  1.27  1.24  0.05  0.06  0.0238  0.124  0.18  0.01  12.56  0.1  0.08  1.35  1.32  0.06  0.06  0.0238  0.122  0.19  0.02  14.04  0.11  0.09  1.43  1.39  0.07  0.07  0.0238  0.12  0.21  0.02  15.48  0.12  0.11  1.51  1.46  0.07  0.07  0.0238  0.119  0.22  0.02  16.88  0.13  0.12  1.58  1.54  0.08  0.08  0.0238  0.117  0.24  0.03  18.26  0.14  0.14  1.66  1.61  0.08  0.09  0.0238  0.116  0.25  0.04  19.61  0.15  0.16  1.74  1.69  0.09  0.09  0.0238  0.115  0.27  0.04  20.94  0.16  0.17  1.82  1.76  0.1  0.1  0.0238  0.113  0.28  0.05  22.25  0.17  0.19  1.89  1.83  0.1  0.1  0.0238  0.112  0.3  0.06  23.55  0.18  0.21  1.97  1.91  0.11  0.11  0.0238  0.112  0.31  0.06  24.83  0.19  0.23  2.05  1.98  0.11  0.12  0.0238  0.111  0.32  0.07  26.1  0.2  0.25  2.12  2.05  0.12  0.12  0.0238  0.11  0.34  0.08  27.44  0.21  0.27  2.19  2.11  0.12  0.13  0.0238  0.109  0.35  0.09  28.76  0.22  0.29  2.26  2.18  0.13  0.13  0.0238  0.108  0.36  0.11  30.07  0.23  0.31  2.33  2.24  0.13  0.14  0.0238  0.107  0.38  0.12  31.36  0.24  0.34  2.4  2.31  0.14  0.15  0.0238  0.107  0.39  0.13  32.64  0.25  0.36  2.47  2.37  0.15  0.15  0.0238  0.106  0.4  0.14  33.91  0.26  0.38  2.54  2.44  0.15  0.16  0.0238  0.105  0.41  0.16  35.16  224  Table C-2: WinXSPro output for Greata Creek, cross-section 10. STAGE  AREA  PERIM  WIDTH  R  DHYD  SLOPE  (m)  (sq m)  (m)  (m)  (m)  (m)  (m/m)  n  VAVG  Q  SHEAR  (m/s)  cms  (N/sq m)  0.27  0.41  2.62  2.5  0.16  0.16  0.0238  0.105  0.43  0.17  36.41  0.28  0.43  2.69  2.57  0.16  0.17  0.0238  0.104  0.44  0.19  37.65  0.29  0.46  2.76  2.64  0.17  0.17  0.0238  0.104  0.45  0.21  38.88  0.3  0.49  2.83  2.7  0.17  0.18  0.0238  0.103  0.46  0.22  40.1  0.31  0.51  2.9  2.77  0.18  0.19  0.0238  0.103  0.47  0.24  41.31  0.32  0.54  2.97  2.83  0.18  0.19  0.0238  0.102  0.48  0.26  42.52  0.33  0.57  3.04  2.9  0.19  0.2  0.0238  0.102  0.5  0.28  43.72  0.34  0.6  3.11  2.96  0.19  0.2  0.0238  0.101  0.51  0.3  44.92  0.35  0.63  3.18  3.03  0.2  0.21  0.0238  0.101  0.52  0.33  46.11  0.36  0.66  3.25  3.09  0.2  0.21  0.0238  0.101  0.53  0.35  47.29  0.37  0.69  3.32  3.16  0.21  0.22  0.0238  0.1  0.54  0.37  48.47  0.38  0.72  3.4  3.23  0.21  0.22  0.0238  0.1  0.55  0.4  49.61  0.39  0.75  3.49  3.31  0.22  0.23  0.0238  0.1  0.56  0.42  50.51  0.4  0.79  3.57  3.39  0.22  0.23  0.0238  0.099  0.57  0.45  51.46  0.41  0.82  3.66  3.48  0.22  0.24  0.0238  0.099  0.58  0.47  52.42  0.42  0.86  3.75  3.56  0.23  0.24  0.0238  0.099  0.58  0.5  53.36  0.43  0.89  3.84  3.64  0.23  0.25  0.0238  0.098  0.59  0.53  54.31  0.44  0.93  3.93  3.73  0.24  0.25  0.0238  0.098  0.6  0.56  55.27  0.45  0.97  4.02  3.81  0.24  0.25  0.0238  0.098  0.61  0.59  56.23  0.46  1.01  4.1  3.9  0.25  0.26  0.0238  0.098  0.62  0.62  57.2  0.47  1.05  4.19  3.98  0.25  0.26  0.0238  0.097  0.63  0.66  58.18  0.48  1.09  4.28  4.06  0.25  0.27  0.0238  0.097  0.64  0.69  59.16  0.49  1.13  4.37  4.15  0.26  0.27  0.0238  0.097  0.65  0.73  60.15  0.5  1.17  4.46  4.23  0.26  0.28  0.0238  0.097  0.65  0.76  61.14  0.51  1.21  4.55  4.32  0.27  0.28  0.0238  0.096  0.66  0.8  62.13  0.52  1.25  4.6  4.36  0.27  0.29  0.0238  0.096  0.68  0.85  63.61  225  Table C-2: WinXSPro output for Greata Creek, cross-section 10. STAGE  AREA  PERIM  WIDTH  R  DHYD  SLOPE  (m)  (sq m)  (m)  (m)  (m)  (m)  (m/m)  n  VAVG  Q  SHEAR  (m/s)  cms  (N/sq m)  0.53  1.3  4.65  4.41  0.28  0.29  0.0238  0.096  0.69  0.89  65.12  0.54  1.34  4.7  4.45  0.29  0.3  0.0238  0.095  0.7  0.94  66.63  0.55  1.39  4.75  4.5  0.29  0.31  0.0238  0.095  0.72  0.99  68.12  0.56  1.43  4.8  4.54  0.3  0.32  0.0238  0.095  0.73  1.04  69.61  0.57  1.48  4.85  4.59  0.3  0.32  0.0238  0.094  0.74  1.1  71.08  0.58  1.52  4.9  4.63  0.31  0.33  0.0238  0.094  0.75  1.15  72.55  0.59  1.57  4.95  4.67  0.32  0.34  0.0238  0.094  0.77  1.2  74.01  0.6  1.62  5  4.72  0.32  0.34  0.0238  0.093  0.78  1.26  75.46  0.61  1.66  5.05  4.76  0.33  0.35  0.0238  0.093  0.79  1.32  76.9  0.62  1.71  5.1  4.81  0.34  0.36  0.0238  0.093  0.8  1.38  78.34  0.63  1.76  5.15  4.85  0.34  0.36  0.0238  0.093  0.82  1.44  79.76  0.64  1.81  5.2  4.9  0.35  0.37  0.0238  0.092  0.83  1.5  81.18  0.65  1.86  5.25  4.94  0.35  0.38  0.0238  0.092  0.84  1.56  82.59  0.66  1.91  5.3  4.98  0.36  0.38  0.0238  0.092  0.85  1.62  84  0.67  1.96  5.35  5.03  0.37  0.39  0.0238  0.092  0.86  1.69  85.4  0.68  2.01  5.4  5.07  0.37  0.4  0.0238  0.091  0.87  1.76  86.79  0.69  2.06  5.44  5.12  0.38  0.4  0.0238  0.091  0.89  1.82  88.17  0.7  2.11  5.49  5.16  0.38  0.41  0.0238  0.091  0.9  1.89  89.55  0.71  2.16  5.54  5.21  0.39  0.42  0.0238  0.091  0.91  1.97  90.92  0.72  2.21  5.59  5.25  0.4  0.42  0.0238  0.09  0.92  2.04  92.29  0.73  2.27  5.65  5.3  0.4  0.43  0.0238  0.09  0.93  2.11  93.61  0.74  2.32  5.7  5.34  0.41  0.43  0.0238  0.09  0.94  2.19  94.91  0.75  2.37  5.75  5.39  0.41  0.44  0.0238  0.09  0.95  2.26  96.21  0.76  2.43  5.8  5.44  0.42  0.45  0.0238  0.09  0.96  2.34  97.5  0.77  2.48  5.86  5.49  0.42  0.45  0.0238  0.089  0.97  2.42  98.79  0.78  2.54  5.91  5.53  0.43  0.46  0.0238  0.089  0.98  2.5  100.08  226  Table C-2: WinXSPro output for Greata Creek, cross-section 10. STAGE  AREA  PERIM  WIDTH  R  DHYD  SLOPE  (m)  (sq m)  (m)  (m)  (m)  (m)  (m/m)  n  VAVG  Q  SHEAR  (m/s)  cms  (N/sq m)  0.79  2.59  5.96  5.58  0.43  0.46  0.0238  0.089  1  2.58  101.36  0.8  2.65  6.02  5.63  0.44  0.47  0.0238  0.089  1.01  2.66  102.64  0.81  2.7  6.07  5.68  0.45  0.48  0.0238  0.089  1.02  2.75  103.92  0.82  2.76  6.12  5.72  0.45  0.48  0.0238  0.089  1.03  2.83  105.19  0.83  2.82  6.17  5.77  0.46  0.49  0.0238  0.088  1.04  2.92  106.45  0.84  2.88  6.23  5.82  0.46  0.49  0.0238  0.088  1.05  3.01  107.72  0.85  2.93  6.28  5.87  0.47  0.5  0.0238  0.088  1.06  3.1  108.97  0.86  2.99  6.33  5.91  0.47  0.51  0.0238  0.088  1.07  3.19  110.23  0.87  3.05  6.39  5.96  0.48  0.51  0.0238  0.088  1.08  3.29  111.48  0.88  3.11  6.44  6.01  0.48  0.52  0.0238  0.088  1.09  3.38  112.73  0.89  3.17  6.49  6.06  0.49  0.52  0.0238  0.087  1.1  3.48  113.98  0.9  3.23  6.54  6.1  0.49  0.53  0.0238  0.087  1.11  3.58  115.22  0.91  3.29  6.6  6.15  0.5  0.54  0.0238  0.087  1.12  3.68  116.46  0.92  3.35  6.65  6.2  0.5  0.54  0.0238  0.087  1.13  3.78  117.69  0.93  3.42  6.7  6.25  0.51  0.55  0.0238  0.087  1.14  3.88  118.93  0.94  3.48  6.75  6.3  0.52  0.55  0.0238  0.087  1.15  3.99  120.16  0.95  3.54  6.81  6.34  0.52  0.56  0.0238  0.086  1.16  4.09  121.39  0.96  3.61  6.86  6.39  0.53  0.56  0.0238  0.086  1.17  4.2  122.61  0.97  3.67  6.91  6.44  0.53  0.57  0.0238  0.086  1.18  4.31  123.83  0.98  3.73  6.97  6.49  0.54  0.58  0.0238  0.086  1.18  4.42  125.05  0.99  3.8  7.02  6.53  0.54  0.58  0.0238  0.086  1.19  4.54  126.27  1  3.86  7.07  6.58  0.55  0.59  0.0238  0.086  1.2  4.65  127.48  227  Table C-3. Results from CFA runs fitting distributions to observed record of instantaneous peak flows for Greata Creek. Fitted Distribution  Q2  Q5  Q10  Q20  Q50  Q100  Q200  Q500  Comments Tolerably good fit but ignores effect of mixed distribution (overestimates above Q50)  Generalized extreme value  0.520  1.00  1.39  1.83  2.50  3.09  3.78  4.85  Three-parameter lognormal  0.450  1.02  1.58  2.27  3.40  4.46  5.72  7.73  Poor fit overall  log-Pearson type III  0.500  1.03  1.44  1.86  2.43  2.87  3.31  3.90  Relatively good fit  Wakeby  0.500  1.05  1.47  1.88  2.43  2.84  3.24  3.78  As good as or better than LP3  Weibull  0.520  1.08  1.48  1.86  2.34  2.69  3.04  3.49  OK, but not the best fit overall.  Nonparametric  0.590  1.08  1.28  2.31  2.49  2.56  2.59  2.63  Captures effect of mixed distributions at Q>Q20.  228  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072555/manifest

Comment

Related Items