Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Riparian disturbance sensitivity index for gravel-bed river morphology Bailey, Chad Elliot 2003

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

Item Metadata

Download

Media
831-ubc_2003-0446.pdf [ 10.25MB ]
Metadata
JSON: 831-1.0063629.json
JSON-LD: 831-1.0063629-ld.json
RDF/XML (Pretty): 831-1.0063629-rdf.xml
RDF/JSON: 831-1.0063629-rdf.json
Turtle: 831-1.0063629-turtle.txt
N-Triples: 831-1.0063629-rdf-ntriples.txt
Original Record: 831-1.0063629-source.json
Full Text
831-1.0063629-fulltext.txt
Citation
831-1.0063629.ris

Full Text

RIPARIAN DISTURBANCE SENSITIVITY INDEX FOR GRAVEL-BED RIVER MORPHOLOGY by CHAD ELLIOT BAILEY B.Sc, Montana State University, 2001 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Civil Engineering We, ;pt this thesis as conforming THE UNIVERSITY OF BRITISH COLUMBIA July 2003 © Chad Elliot Bailey, 2003 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my writ ten permission. Department of Clv'< 1 fie/ > 'tlc< r . The University of British Columbia Vancouver, Canada Date / ( , f £0Q 5 DE-6 (2/88) ABSTRACT This thesis is an investigation of the sensitivity of alluvial single-thread gravel-bed rivers with noncohesive banks to riparian disturbances. A riparian disturbance is any natural or anthropogenic disturbance that causes a significant removal of large coniferous trees from the riparian corridor and potentially leads to decreased bank strength and planform morphology towards a braided system. A theoretical riparian disturbance sensitivity index is formulated and tested in terms of its validity in estimating the sensitivity of the rivers by means of change in width from a historic "natural" condition to a present day disturbed condition. The riparian disturbance sensitivity index is tested through the use of case studies at the reach level collected within the province of British Columbia. A methodology is suggested for collecting the necessary data to apply towards the sensitivity index. The methodology includes reach longitudinal and cross-sectional surveys, grain size analysis, discharge analysis through watershed delineation and regionalization techniques, and historic aerial photo analysis. Fourteen out of fifteen data sets used in this study are satisfactory data sets that fit the assumptions necessary to apply the index. Even though there are large uncertainties in the accuracy of the field collected data, the proposed sensitivity index appeared to perform well when tested against the case study data sets. The role of large, or "catastrophic," floods is examined with two flood related indices: flash flood magnitude index and peak/mean ratio. The results lend support towards the theory of vegetation as a control of alluvial channel morphology, in comparison to the control of large floods. TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF FIGURES x LIST OF TABLES xiii AKNOWLEDGEMENTS xiv 1 INTRODUCTION 1 1.1 Riparian Management 3 1.1.1 Buffer Strip Treatments 3 1.1.2 Fish Habitat 4 1.2 Objectives 5 1.3 Thesis Outline 6 2 THEORY 8 2.1 Regime Theory 8 2.1.1 Instability 9 2.2 Controls on Channel Morphology 11 2.2.1 Discharge 12 2.2.1.1 Dominant Discharge 13 2.2.1.1.1 Disturbance Effects on Dominant Discharge 15 2.2.1.2 Hydrologic History 20 2.2.1.3 Catastrophic Floods 20 2.2.1.3.1 Flash Flood Magnitude Index 22 2.2.1.3.2 Ratio of Peak to Mean Annual Flood 23 iv 2.2.1.4 Bank Erosion by Fluvial Entrainment 24 2.2.2 Sediment Load. 27 2.2.2.1 Surface Sources 28 2.2.3 Valley Slope 29 . 2.2.4 Bank Stability 29 2.2.4.1 Cohesive Banks 30 2.2.4.2 Non-cohesive Banks 31 2.2.4.3 Composite Banks 32 2.2.4.4 Vegetative Influences 33 2.2.4.4.1 Root Reinforcement 33 2.2.4.4.2 Large Woody Debris 36 2.2.4.4.3 Width Morphology 37 2.2.4.4.4 Scaling 40 2.2.4.5 Bank Stability Algorithm 41 2.3 Meandering-Braiding Transition 44 2.3.1 Theoretical Stability Analysis 44 2.3.2 Millar Theoretical Meandering-Braiding Transition 47 2.4 Proposed Riparian Disturbance Sensitivity Index 49 3 METHODOLGY 51 3.1 Fieldwork 51 3.1.1 Grain Size 51 3.1.1.1 Pebble Count Technique 52 3.1.1.2 Bulk-Sieve Technique 55 3.1.1.3 Pebble Count versus Bulk-Sieve Technique 58 3.1.2 Channel Geometric Survey 59 3.1.2.1 Longitudinal Profile 59 3.1.2.2 Bankfull Channel Width or Unvegetated Channel Width 60 3.2 Analytical Methods 60 3.2.1 Watershed Drainage Area 61 3.2.2 Discharge 61 v 3.2.2.1 Regional Analysis 62 3.2.3 Width 63 3.2.4 Aerial Photo Analysis 64 3.2.4.1 Scale 65 3.2.4.2 Measurements 67 3.3 Flood Indices 68 4 C A S E S T U D I E S 7 0 4.1 Bonaparte River 71 4.1.1 Grain Size A nalysis 72 4.1.2 Discharge 72 4.1.2.1 Regional Analysis 74 4.1.3 Slope 74 4.1.4 Width 75 4.2 Coldwater River • 77 4.2.1 Grain Size Analysis 79 4.2.2 Discharge • 79 4.2.2.1 Regional Analysis 79 4.2.3 Slope '. 82 4.2.4 Width 52 4.3 Deadman River 88 4.3.1 Grain Size Analysis 90 4.3.2 Discharge 90 4.3.3 Slope 93 4.3.4 Width 94 4.4 Elk River 97 4.4.1 Grain Size Analysis 99 4.4.2 Discharge 99 4.4.3 Slope 102 4.4.4 Width 103 4.5 Eve River 106 4.5.1 Grain Size Analysis 107 4.5.2 Discharge 109 vi 4.5.2.1 Regional Analysis 109 4.5.3 Slope 110 4.5.4 Width Ill 4.6 Salmon River 113 4.6.1 Grain Size Analysis 114 4.6.2 Discharge 116 4.6.3 Slope 117 4.6.4 Width 117 4.7 Tahsis River 120 4.7.1 Grain Size Analysis 121 4.7.2 Discharge 123 4.7.2.1 Regional Analysis 123 4.7.3 Slope 124 4.7.4 Width 125 4.8 Case Studies on Existing Data 127 4.8.1 Big Horn Creek 128 4.8.1.1 Grain Size Analysis 129 4.8.1.2 Discharge 131 4.8.1.2.1 Regional Analysis 131 4.8.1.3 Slope 133 4.8.1.4 Width 134 4.8.2 Narrowlake Creek 134 4.8.2.1 Grain Size Analysis 134 4.8.2.2 Discharge 136 4.8.2.2.1 Regional Analysis 137 4.8.2.3 Slope 138 4.8.2.4 Width 139 4.8.3 Slesse Creek 139 4.8.3.1 Grain Size Analysis 141 4.8.3.2 Discharge 141 4.8.3.3 Slope 143 4.8.3.4 Width 143 vii 4.8.4 West Kettle River 144 4.8.4.1 Grain Size Analysis 145 4.8.4.2 Discharge 145 4.8.4.3 Slope '.' 146 4.8.4.4 Width 146 4.9 Summary Data 147 5 CHANNEL MORPHOLOGY ANALYSIS 151 5.1 Analysis of Morphological Indices 151 5.1.1 Riparian Disturbance Sensitivity Index 152 5.1.2 Flood Indices 154 5.2 Reach by Reach Discussion 157 5.2.1 Big Horn Creek 157 5.2.2 Bonaparte River 157 5.2.3 Coldwater River 158 5.2.4 Deadman River 162 5.2.5 Elk River 162 5.2.6 Eve River 163 5.2.7 Narrowlake Creek 164 5.2.8 Salmon River 165 5.2.9 Slesse Creek 165 5.2.10 Tahsis River 166 5.2.11 West Kettle River 166 5.3 Nature of AW% — C, Relation 167 5.4 Index Sensitivity to Grain Size Selection (Bank D50) 167 5.5 Role of Floods vs. Riparian Disturbance 170 6 CONCLUSIONS AND RECOMMENDATIONS 171 6.1 Summary 171 6.2 Conclusion of RDSI (Q Results 172 6.3 Future Work 173 REFERENCES 174 APPENDIX A - ANNUAL MAXIMUM SERIES 188 viii Peak and Mean Flow Data 188 Big Horn Creek Flow Data 189 Bonaparte River Flow Data 190 Coldwater River Flow Data 190 Deadman River Flow Data 191 Elk River Flow Data 191 Eve River Flow Data 192 Narrowlake Creek Flow Data 193 Salmon River Flow Data 194 Slesse Creek Flow Data 194 Tahsis River Flow Data 195 West Kettle River Flow Data 196 APPENDIX B - AERIAL PHOTOGRAPHS 198 Calculated Photo Scales 198 Reach Aerial Photographs 199 APPENDIX C - WIDTH MEASUREMENTS 217 Big Horn Width Measurements 217 Bonaparte River Width Measurements 218 Coldwater River Width Measurements 219 Deadman River Width Measurements 222 Elk River Width Measurements 224 Eve River Width Measurements 226 Salmon River Width Measurements 227 Tahsis River Width Measurements 228 West Kettle River Width Measurements 229 L I S T O F F I G U R E S Figure 2-1 Variables Affecting Channel Morphology (Morrisaw and LaFlure 1979) 11 Figure 2-2 Moments of Submerged Particle on a Horizontal Bed at the Threshold of Movement (Carson 1971) 26 Figure 2-3 Physical effects of vegetation (Figure 3.1 in (Coppin 1990)) 35 Figure 3-1 Sample template (gravelometer) (Center 1996) 53 Figure 3-2 Grain axes, similar to Fig. 3.3 in Church (1987) 54 Figure 3-3 Aerial Photo Scale Diagram 65 Figure 4-1 Case study sampling locations 70 Figure 4-2 Bonaparte River cumulative grain size distribution 72 Figure 4-3 Bonaparte River location map and watershed delineation (921 Ashcroft and 92P Bonaparte Lake, 1:250,000) 73 Figure 4-4 Bonaparte River longitudinal profile 74 Figure 4-5 Bonaparte River channel cross-sections (top) cross-section 1, (middle) cross-section 2, and (bottom) cross-section 3 76 Figure 4-6 Coldwater River cumulative grain size distributions 80 Figure 4-7 Coldwater River location map and watershed delineation (92H Hope and 921 Ashcroft, 1:250,000) Note north is down 81 Figure 4-8 Coldwater River longitudinal profiles 83 Figure 4-9 Coldwater River reach 1 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 85 Figure 4-10 Coldwater River reach 2 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 86 Figure 4-11 Coldwater River reach 3 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 87 Figure 4-12 Deadman River cumulative grain size distributions 91 Figure 4-13 Deadman River location map and watershed delineation (921 Ashcroft and 92P Bonaparte Lake, 1:250,000) 92 Figure 4-14 Deadman River longitudinal profiles 93 Figure 4-15 Deadman River reach 1 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 95 Figure 4-16 Deadman River reach 2 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 96 Figure 4-17 Elk River cumulative grain size distributions 100 Figure 4-18 Elk River location map and watershed delineation (92F Port Alberni, 1:250,000) 101 Figure 4-19 Elk River longitudinal profiles 102 Figure 4-20 Elk River reach 1 channel cross-sections: a) section 1, b) section 2, c) section 3, andd) section 4 104 Figure 4-21 Elk River reach 2 channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 105 r Figure 4-22 Eve River cumulative grain size distributions 107 Figure 4-23 Eve River location map and watershed delineation (92L Alert Bay, 1:250,000) 108 Figure 4-24 Eve River longitudinal profile 111 Figure 4-25 Eve River channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 112 Figure 4-26 Salmon River cumulative grain size distributions 114 Figure 4-27 Salmon River location map and watershed delineation (92E Nootka Sound, 92F Port Alberni, 92K Bute Inlet and 92L Alert Bay, 1:250,000) 115 Figure 4-28 Salmon River longitudinal profile 117 Figure 4-29 Salmon River channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 4 11.9 x i Figure 4-30 Tahsis River cumulative grain size distributions 121 Figure 4-31 Tahsis River location map and watershed delineation (92E Nootka Sound and 92L Alert Bay, 1:250,000) 122 Figure 4-32 Tahsis River longitudinal profile 125 Figure 4-33 Tahsis River channel cross-sections: a) section 1, b) section 2, c) section 3, and d) section 126 Figure 4-34 Existing Case Study Locations 127 Figure 4-35 Big Horn Creek location map and watershed delineation (map 82G Fernie, 1:250,000) 130 Figure 4-36 Big Horn Creek regional analysis based on nearest watersheds 132 Figure 4-37 Bighorn Creek longitudinal profile 133 Figure 4-38 Narrowlake Creek location map and watershed delineation (93H McBride and 93G Prince George, 1:250,000) 136 Figure 4-39 Narrowlake Creek regional regression chart 138 Figure 4-40 Slesse Creek location map and watershed delineation (Mount Baker reference map, 1:100,000) 142 Figure 5-1 Percent change in width versus riparian disturbance sensitivity index (Q. .153 Figure 5-2 Percent change in width versus flash flood magnitude index 155 Figure 5-3 Percent change in width versus peak/mean discharge ratio 156 Figure 5-4 Coldwater River reach 3 (undisturbed) (1:10,000) 160 Figure 5-5 Coldwater River reach 1 (1:10,000) 161 Figure 5-6 Percent change in width versus riparian disturbance sensitivity index (Q, using bank D 5 0 169 xii L I S T O F T A B L E S Table 2-1 Friction angle values (j)'(degrees) obtained through model calibration of each vegetation class, from Millar (2000) 43 Table 4-1 Eve River regional regression table 110 Table 4-2 Tahsis River regional regression table 124 Table 4-3 Big Horn Creek regional regression table 133 Table 4-4 Discharge summary data for all case studies 148 Table 4-5 Summary data for collected case study channels (T = return period, based on Gumbel distribution and n years of record) 149 Table 4-6 Summary data for previously collected case studies (T = return period, based on Gumbel distribution and n years of record) 150 Table 5-1 C, sensitivity to grain size. The values are the estimates of for the different sediment diameters. The percent difference is the percent difference between the riffle value and the other corresponding value 168 xiii AKNOWLEDGEMENTS This project would not have been successful without the assistance and encouragement of many other people. The one person who was paramount for this project was my advisor Dr. Rob Millar. I would like to thank him for the opportunity to study under his supervision, including ideas, support, and of course, funding. Mike Miles was extremely influential in selecting research sites for the project and assisted in numerous other aspects, in addition to my personal growth as a geomorphologist, which I graciously accept. Various members of the UBC community provided support, tools, and discussion. Dr. Younes Alila sacrificed many hours to assist me in understanding the implications of hydrology. Dr. Bernard Laval reviewed the thesis and provided many useful suggestions. Dr. Marwan Hassan and Dr. Michael Church graciously allowed me to borrow some of their field equipment. For their relentless field assistance and battle of the elements (including stitches), Ted Tedford and Andrew Wu were instrumental in collecting the field data for this study. Don Ignace of the Skeetchestn Indian Band politely allowed access to some of the intensely privately owned lands bordering the Deadman River. Several officemates philosophized and provided assistance in many aspects of the study including Dr Colin Rennie, Jose Vasquez, and Dave Roche. Numerous people from local agencies and consulting provided valuable data and information including Andrew Wilson (WLAP:EX), Lynne Campo (WSC), Greg Scarborough (formerly Klohn-Crippen), Rory Tennant (BC Hydro), Ray Pillipow (WLAP:EX), and Mark Poire (SRM:EX, LandData BC). Finally, Shasta Grenier was my encouragement, support, friend, and emotional consultant when I needed it the most. xiv CHAPTER 1 1 INTRODUCTION This thesis compares the hydraulic, geometric, and geomorphologic conditions of the alluvial stream before and after a main disturbance to riparian vegetation. Natural or anthropogenic disturbances to watershed characteristics can alter channel processes and morphology. The disturbance can be any event that may destroy the health of the bank vegetation, potentially leading to a failure of the bank strength and widening and aggradation of the stream. The diversity and magnitude of morphologic responses to a riparian disturbance depend on channel type, external influences, and disturbance history (Montgomery and Buffington, 1997). The riparian environment is where vegetation plays an integral role in the stability of the channel banks by moderating erosion and contributing to the health of the stream (Abernethy and Rutherfurd 2000). Riparian zone characteristics can be distinguished by the "bank features and vegetation including those surfaces that are inundated or saturated at least annually, such as broad forested wetlands commonly adjacent to coastal-plain streams (Hupp and Osterkamp 1996)." The BC forest practice codes (FPC) describes these characteristics from the perspective of the logging industry: "Riparian areas occur next to the banks of streams, lakes and wetlands and include both the area dominated by continuous high moisture content and the adjacent upland vegetation that exerts an influence on it. Riparian ecosystems 1 contain many of the highest value non-timber resources in the natural forest (BC Ministry of Forests, 1995; Beaudry 1998)." The riparian ecosystem is the link between the aquatic and terrestrial environments. But, historically the riparian corridor of valley flats were some of the first areas of a watershed to be logged because of the ease in reaching and the size of the trees. This situation is depicted in a discussion of the history of logging in British Columbia: "[T]he early lumbermen and timber speculators sought out the best timber, so many of the early licenses and leases contained and still contain much of the finest timber in the Province. Established in the most accessible locations, they were mostly near tidewater along the southern coast and on Vancouver Island, often strategically positioned in the mouths of rivers or in valley bottoms (Gould 1975)." This discussion is focused on the interaction between the riparian vegetation and alluvial river systems. An alluvial channel is one that flows through sediments that it has previously deposited and is currently transporting (Schumm 1969; Church 1992). Typically the nominal size and characteristics of an alluvial stream where riparian vegetation has a substantial role on the nature of the stream is at the mid-basin level. Mid-basin streams are the primary regions where fluvial entrainment is the prevailing erosion mechanism (Lawler 1992). In this study, the channels of interest for this analysis are alluvial single-thread gravel-bed rivers with noncohesive banks. 2 1.1 Riparian Management In recent years riparian management has become a popular and debated topic in river regulation and restoration. It is believed that riparian zones can be managed to reduce stream bank erosion rates and to trap pollutants coming off hillslopes (Wilson et al. 1995). Riparian zones provide substantial amounts of habitat for ecologically diverse flora and fauna while protecting the quality of the vital life-support resource water from man's urban, agricultural, and industrial development (Corbett and Lynch 1985). Therefore, the riparian disturbance sensitivity index may be a valuable tool for assisting in the management of riparian areas to protect or enhance the environmental conditions of riverine landscapes. 1.1.1 Buffer Strip Treatments Buffer strip treatments in recent years have shown increased stability of stream banks as compared to historic clearcut methods (Toews and Moore 1982; Harris 1987; Jackson et al. 2001). Buffer strips have been shown to increase bank stability by maintaining masses of living roots and decrease substantial amounts of logging induced small organic debris addition to channels, but catastrophic blowdown as opposed to periodic large organic debris recruitment must be considered when designing an effective buffer strip (Toews and Moore 1982; Steinblums et al. 1984). The majority of the timber volume susceptible to blowdown is lost in a few years and is dependent upon species composition and watershed conditions (Corbett and Lynch 1985). 3 The small organic debris typically induced by riparian removal practices is less hydraulically stable and tends to not effectively contribute to the complexity of habitat structures within the channel (Ralph et al. 1994). Buffers have shown the ability to sustain pre-harvest habitat and sediment characteristics in small headwater streams (Jackson et al. 2001). According to Ralph et al. (1994) on their study of western Washington basins, intensive logging of riparian areas results in simplifying and homogenising habitat structures because of the removal of potential LWD and decreases the potential storage of excess sediment behind large woody debris (LWD) structures. Also, buffer strips appear to be valuable management tools to decrease the amount of riparian removal or damage caused by heavy grazing pressures. Therefore, since buffer strips appear to control the in channel dynamics of protected streams, such controlling factors could be beneficial for the conservation of preferential fish habitat. 1.1.2 Fish Habitat The removal of riparian vegetation potentially could deleteriously affect fisheries habitat. The clearing of streamside vegetation has an immediate effect in removing bank cover for fish habitat in terms of protection from predation (Sotir and Nunnally 1995; Kellerhals and Miles 1996). The future implications of removal may include elevated stream temperatures, potential decrease in complexity from woody debris recruitment to the channel, and elevated amounts of fine sediment from bank erosion and increased bedload movement due to loss of LWD. All of these factors could lead to decreased spawning conditions or the loss of incubating fish eggs (Roberts and Church 1986; Hartman and Scrivener 1990; Sotir and Nunnally 1995; Kellerhals and Miles 1996). 4 Also, due to wider channels, increased loose sediment depths from aggradation, and reduced water viscosity from water temperature increases could lead to increased volumes of water travelling sub-surface during low flow periods, resulting in the loss of deep water refuge or the perish of rearing fish fry (Roberts and Church 1986; Hartman and Scrivener 1990; Kellerhals and Miles 1996). Besides habitat control, the riparian environment also supplies food for fish, through the process of insect drift, and provides "allochthonous input" to supply nutrients to the aquatic ecosystem in order to increase or stabilize aquatic productivity (Moring et al. 1985; Sotir and Nunnally 1995). It appears that wise riparian management could assist in the recovery of the declining fish stocks that have occurred over the past several decades. 1.2 Objectives The objective of the study is to formulate and test an index that describes the sensitivity of the single-thread gravel-bed river with noncohesive banks to disturbance by comparing intact and disturbed channel widths. A "need exists for the assessment of channel conditions and the relative sensitivity of channels to disturbance or altered environmental conditions (Simon and Downs 1995)." Included within the test is a comparison between the results of the sensitivity index and two indices for examining the effects of large or "catastrophic" floods on channel morphology. The objective of the sensitivity index is to formulate a potential management tool for riparian management that may be summarized by the following quote: 5 "Given that direct riparian logging is now forbidden under the Forest Practices code,... probably most useful in assessing the current channel condition, and in the case of wide and unstable channel, to assess whether direct riparian logging represents a significant factor in the current channel condition (Millar 2001)." The riparian disturbance sensitivity index will be evaluated using case studies of disturbed and undisturbed channel reaches situated in the province of British Columbia. 1.3 Thesis Outline Chapter 1 is an introduction of the project. Key concepts of disturbance, morphologic response, riparian vegetation, and alluvial gravel-bed rivers are introduced. The implications of riparian management in terms of buffer strip treatment and fish habitat is discussed. The study area is introduced and objectives are presented. Chapter 2 is a review of present literature and theory. The idea of regime is introduced and discussed. The various controls on channel morphology are identified and their implications are mentioned, with a primary focus on floods and bank stability. The theoretical meandering-braiding transition is examined, leading to the formulation of the proposed riparian disturbance sensitivity index. In Chapter 3 the methodology for collecting the required data is formulated. The methodology includes two aspects, a field survey and analytical techniques. Each individual method is critiqued in order to create a reproducible process for future applications of the riparian disturbance sensitivity index theory. Chapter 4 is a collection of the seven field case study rivers, including 11 study reaches, and four previously collected study rivers. Each case study is presented with an overview and description of the study site and examined in terms of four major requirements for collected data. Each section is presented as an individual and reasonably complete study of the corresponding study reach. Chapter 5 is a presentation of the results of the study. The statistical correlation between the riparian disturbance sensitivity index and two flood related indices with a surrogate for channel morphology, percent change in width, is presented. The implications of the results are discussed. Each case study is specifically analysed for its results in terms of the three indices. The nature of the riparian disturbance sensitivity index, its sensitivity, and a comparison between floods and riparian vegetation concludes the chapter. Final conclusions and recommendations are bestowed in chapter 6. These conclusions include an overall summary of the project, specific conclusions for the riparian disturbance sensitivity index results, and proposed future work to enhance the research completed in this thesis. Supplementary information is included in the appendices including summaries of discharge calculations and available gauge data in Appendix A , calculated aerial photograph scales and reproductions of the available aerial photographs in Appendix B, and summaries of the width measurements and analysis in Appendix C. 7 CHAPTER 2 2 THEORY 2.1 Regime Theory River regime is a concept in which a river flows through its own alluvium and seeks to dynamically adjust over time (period of years or decades) to a state where its hydraulic geometry, topographic setting, imposed quantity and calibre of sediment, and flow regime are balanced without net deposition or scour occurring within the channel (Mackin 1948; Blench 1957; Yang and Song 1979; Chang 1985; Mien 1994; ASCE Task Committee on Hydraulics 1998; Millar 2000; Rosgen 2001; Church 2002). The term "regime" denotes the situation when the alluvial river channel reaches this state of dynamic or quasi-equilibrium (Richards 1982), and forms a pillar of fluvial geomorphology. One theory on dynamic adjustment of alluvial rivers is the minimum rate of energy dissipation. Simply stated, "a river will adjust itself in such a manner that the rate of energy dissipation can be minimized to regain the equilibrium condition (ASCE Task Committee on Hydraulics 1998)." Another name for these "regime channels" is "stable mobile gravel-bed rivers," where the channel is stable but still dynamically adjusting to its upstream controls (Hey and Thorne 1986). Therefore, any natural or anthropogenic change in the controlling parameters of the stream will result in an adjustment of the fluvial system in order to absorb the affect of the change or stress on the system (Mackin 1948; Rosgen 2001). 8 This leads to the situation where any instability or disturbance in the independent variables of the fluvial environment, such as flow discharge, valley slope, or potential energy, may cause an adjustment in the morphology, planform, sediment load, or hydraulics through various feedback interactions between the flow and the channel boundary (Simon 1994). According to MacVicar and Millar (1998), the principal independent variables are discharge, bed load discharge, bed material size, bank characteristics, and valley slope. The dependent variables are those that adjust to the independent variables, such as width, mean depth, and slope. For example, if bank stability increases with an increased roughness from vegetation than a slightly greater depth will be required to convey the same flow. Hence, the increased depth of water will increase the bed shear stress and potentially decrease the bed stability. This situation will probably result in a narrower deeper channel (Jackson and Haveren 1984). 2.1.1 Instability The instability, or significant morphologic change, of a channel can be a result of a natural or anthropogenic disturbance within the watershed that can change the hydrology, sediment supply, riparian vegetation, or large woody debris loading (Montgomery and Buffington 1997). This disturbance may be direct or indirect and include a change in watershed land use, river regulation, changes in flow regime and sediment supply, channel engineering such as bank armouring, diking, or channel re-alignment; changing valley slope, or succession of riparian vegetation (Simon and Downs 1995; Kellerhals and Miles 1996; ASCE Task Committee on Hydraulics 1998). The primary focus of this thesis is the succession of riparian vegetation due to changes in watershed land use, 9 which may include activities such as a fire, logging, agriculture, drought, overgrazing, and/or human urban and suburban development. Therefore, a disturbance that alters the stability of the riparian vegetation condition could result in an adjustment of the channel's characteristics. In this study, change in the channel width will be used as a measure of disturbance or change, as width is readily measured in the filed and historical values can be measured from aerial photos. The adjustment of the channel's characteristics is directly related to the change in width from an intact riparian condition to a disturbed riparian condition. It must be remembered that an adjustment in width typically will be accompanied by other morphological changes such as channel depth, roughness, bed material composition, energy slope, and channel planform (Schumm 1969; ASCE Task Committee on Hydraulics 1998). Figure 2-1 displays a rough explanation of the role of different channel characteristics and their potential interactions. For example, a change in discharge may cause a change in depth and slope, which could cause a change in the width and velocity, or vice versa. The hydraulic interactions of a natural fluvial environment are extremely complex and careful examinations must be made when studying such processes as channel instability and resulting morphology. 10 Lithology Soil Vegetation Width Discharge Depth Slope Sediment T ,oad Velocity Roughness Pattern Figure 2-lVariables Affecting Channel Morphology (Morrisaw and LaFlure 1979) 2.2 Controls on Channel Morphology As discussed above, a change in one characteristic of a channel may lead to instability within the system and result in measurable changes in one or more other characteristic of the channel. These changes may involve short time scales in terms of days and short spatial dimensions, or longer periods of time in terms of years or centuries and complete fluvial systems depending on the size and type of disturbance (Simon 1995). Not only can the time scales be variable, but there also is an opportunity for effects to propagate downstream with a temporal lag between the cause and effect (Lewin et al. 1988). These characteristics of a channel are not only independent variables of the system but may also be dependent variables with respect to other characteristics (Simon 1995). As stated in Brice (1964), Lane (1957) enounced that the major elements, or characteristics, that affect channel form, or shape, are water discharge, longitudinal slope, sediment load, resistance of bed and banks to scour by flowing water, vegetation, temperature, geology, and work of man. Church (1992) stratified these parameters into primary and secondary factors governing river morphology. He listed the primary factors as the volume and time distribution of water supplied from upstream; the volume, timing and character of sediment delivered to the channel; the nature of the materials through which the river flows; the local geological history of the riverine landscape. The secondary factors are local climate, the nature of riparian vegetation, and land use in the drainage basin and a further important factor of human modification of the channel. A similar distinction was made by Kellerhals and Miles (1996). Therefore, channel morphology appears to be a balance between the erosion forces (quantified as bed shear stress through a relationship between discharge, slope, and sediment transport capacity) and resistance (quantified by sediment grain size and vegetation) (Ferguson and Ashworth 1991). The majority of these planform characteristics will be discussed in the following sections relating to their control on channel form and morphology. 2.2.1 Discharge It is well documented in the literature that the independent variables discharge and sediment supply have the greatest influence on fluvial channel form (Hey 1975). It is believed that a particular discharge exists that is most effective in changing the geometry of rivers due to the amount of work performed by the flow in moving sediment and the frequency at which this discharge occurs. In other words, the flow that causes the most erosion and deposition (Wolman and Miller 1960; Bridge 1993; Emmett 2001). This discharge has been coined the term "dominant" discharge because of its dominance in fluvial geomorphologic processes. Looking at a section in the long term, 100 years, the bankfull flow transports the largest volume of sediment through the section (Hey 1978). In the recent study by Emmett and Wolman (2001), the effective discharge, defined as the discharge that over time transports the most bedload sediment, of gravel-bed rivers was analysed with the use of bedload sampling data on five rivers in Wyoming and Idaho. In this study, the effective discharge, in accordance with Andrew's definition, could be determined by the maximum slope value of the cumulative bedload curve with the effective discharge being the discharge located at the centroid of this slope. In order to relate to the following argument of bankfull and dominant discharges, the recurrence interval of the bankfull flows for the five studied rivers ranged from 1.5 to 1.7 years, but the effective discharge ranged to about 130 per cent of the bankfull discharge. An increase in recurrence interval by 130 per cent could represent about a doubling of the recurrence interval (Emmett and Wolman 2001). Unfortunately, bedload transport data is unavailable for the current study, which makes the use of this method inappropriate. 2.2.1.1 Dominant Discharge The dominant discharge can be viewed from a geomorphologic or a hydrologic viewpoint. Based on geomorphologic properties, this single dominant flow is usually regarded as the bankfull flow of the channel (Hey 1975; Hey and Thorne 1986; Carling 1988). This discharge is the flow that is proficient in transporting the imposed sediment load through the channel and yet maintains a width, depth and slope that is in dynamic equilibrium (Bray 1975). It is believed to be applicable where streams are alluvial, 13 competent, and free to adjust their boundaries (Carling 1988). This particular flow establishes the riparian zone along side a channel as part of the biosphere supported by recent fluvial landforms and is inundated or saturated by the bankfull discharge (Hupp and Osterkamp 1996). The problem remains at relating the bankfull discharge to a representative hydrological frequency. The problem with obtaining a representative hydrological discharge in relation to a geomorphic discharge is the variable definitions of identifying the bankfull stage. Williams (1978) compiled a list of 11 different definitions of bankfull stage based on sedimentary surfaces and boundary features. This variability in characterising bankfull stage could be a reason for the variability in the following hydrologic frequencies for a representative discharge: (1) Leopold and Maddock (1953) used the mean annual discharge, or 2.33-year flood for width-discharge relations (2) Bray (1975) used the 2-year flood for Alberta rivers for width-discharge relations (3) Leopold et al. (1964) suggested a return period of 1.5 years (4) Carling (1988) suggested a recurrence interval of 0.9 years on the partial duration series (5) Church (pers. comm.) related the discharge to the 2-year flood for British Columbia rivers in correlation to Bray's (1975) findings. Based on Church's findings, it would appear that the representative discharge for the dominant discharge in this study would be the 2-year flood, which is close to the mean 14 annual flood. On the other hand, it is a simple and straightforward calculation to estimate the average flood for instantaneous peak flow, or the mean annual flood, with the use of the HYDAT 2000 data set and the British Columbia Streamflow Inventory (Environment Canada 2000, Coulson 1998). Therefore, since the mean annual flood is similar hydraulically to the 2-year flood and is readily accessible in the flow records, the mean annual flood value is the flow representing the dominant discharge for this study. This relationship is in agreement with the relationship used by van den Berg (1995) in his analysis of alluvial channel pattern. The purpose for the relationship between dominant bankfull discharge and flood frequency is because it is more practical to use a flood frequency analysis as compared to measuring a bankfull discharge (Brush 1961). 2.2.1.1.1 Disturbance Effects on Dominant Discharge As discussed earlier, the dominant discharge is directly related to a particular flood frequency and its corresponding hydrograph or peak flow. Due to potential changes in surface and subsurface flood routing and absorption, it is a concern that disturbances in a channels' watershed could affect the hydrograph and frequency linked to the dominant discharge. In other words, is it valid to assume that the dominant discharge does not change before and after a disturbance in a watershed? Numerous researchers, Harr et al. (1975), Harr and McCorison (1979), Ziemer (1981), Harr et al. (1982), Hetherington (1982), Hetherington (1987), Jones and Grant (1996), Thomas and Megahan (1998), and Beschta et al. (2000), have performed paired watershed basin studies to analyse the peak flow responses to various forest practices: 15 road building, clearfelling, and burning. It has been hypothesized that the forest practices could alter the drainage network in a watershed, consequently changing the hydrological characteristics of the watershed. These alterations could include increased routing frequency and subsurface flow interception from road construction, clogging of subsurface channel networks, loss/increase of water consumption from vegetation removal/growth, disruption of historical water balance and evapotranspiration, formation of hydrophobic surface during slash burning, and soil disturbance or compaction during yarding (Rothacher 1970; Harr et al. 1975; Harr and McCorison 1979; Harr et al. 1982; Hetherington 1982; Hetherington 1987; Jones and Grant 1996). Most experimental watersheds are usually only a few hectares in size due to the cost of instrumentation and maintenance of the watersheds. The majority of practitioners interested in watershed dynamics are concerned with watersheds that have local or regional influence, such as reservoirs, large fisheries networks, or flood control. At basin scales of regional hydrologic analysis, greater than 10 km2, the details shown in experimental watersheds are typically of secondary importance compared to the flow routing through the channel (Pitlick 1994). Usually the travel time of runoff on hill slopes becomes negligible when compared to the travel time through the channel (Kirkby 1976). Consequently, most of the scientific data is available for small watersheds while most of the interest for watershed characteristics remains with the large watersheds. Therefore, most of the watersheds used in these studies are considered small (< 100 ha) watersheds while only a few of the studies discuss large (60 - 600 km2) watersheds based 16 on drainage area. The conclusions from the paired watershed studies appear to have contrasting and controversial results. Rothacher's (1970) study of water yields in the H.J. Andrews Experimental Forest in western Oregon found an increase in the water yields during the rainy season, but observed no apparent change in the peak flows from the flood producing events. Harr et al. (1975) concluded that peak flows were increased significantly, about 45%, when new roads occupied at least 12% of small watersheds in western Oregon. Harr and McCorison (1979) stated that the size of the annual peak flow on HJA -10, a 10.2 ha headwater basin in the H.J. Andrews Experimental Forest, was reduced by 32% with an average delay in time to peak of nearly 9 hours following clearcut logging. They theorized that since the "hydrologic response of small watersheds in western Oregon to rainfall and snowmelt is rapid, increasing snow accumulation and delaying melt on a cutover watershed also could cause delayed, smaller peak flows (Harr and McCorison 1979)." In another set of data from the H.J. Andrews Experimental Forest, with the use of a 13.0 ha clearcut watershed HJA-6, a 15.4 ha shelterwood cut watershed HJA-7, and an 21.4 ha uncut watershed HJA-8, researchers concluded that neither the size nor the timing of peak flows changed significantly after logging at either watershed as compared to the uncut, control watershed (Harr et al. 1982). In another set of paired watershed studies, numerous researchers took the same set of data; 34-year records from two pairs of 60-to-101 ha basins in the H.J. Andrews Experimental Forest and 50-to-55-year records from three pairs of adjacent basins 17 ranging from 60 to 600 km . Through the use of dissimilar statistical techniques, Jones and Grant (1996) found that due to road construction with patch clear-cutting ranging from 10 to 25% of basin area, peak discharges increased by as much as 50% in small basins and 100% in large basins. While Thomas and Megahan (1998) found increases from 40-90%) in the small watersheds, but were unable to observe any effect of cutting on peak flows in one of the large basin pairs, and results were inconclusive in the other two large basin pairs. Congruently, Beschta et al. (2000) concluded that peak flow increases following treatment averaged in the range of 13-16% for 1-year recurrence interval events and 6-9% for 5-yr recurrence interval events. In addition, for the large basins, they did not observe strong evidence (p<0.05) for peak flow increases through their particular regression analysis. There are many other studies that focused on the effects of logging on watershed hydrology. One of these studies was performed at the Caspar Creek watershed in northern California. This study constituted two watersheds. The South Fork included about 424 ha, and the North Fork roughly 508 ha. The results of the regression analysis performed on the hydrological data from this area determined that the pre and post-logging relationships for the South Fork and North Fork areas were not statistically different (Ziemer 1981). The final study to mention was performed on the 9.5 km2 Carnation Creek watershed located on the west coast of Vancouver Island, British Columbia. For this situation, data were collected for the watershed during 5 years of pre-logging calibration, 6 years of logging, and at least 3 years of post-logging conditions. The results of this study were that there was no change in the peak flows of the main 18 stream after 40% of the watershed was logged (Hetherington 1982). A retrospective study completed by McFarlane (2001) of two paired watershed studies in south eastern British Columbia concluded that changes in peak flow were observed, but the reason for these changes could not be determined. According to these numerous studies, concrete evidence is not existent that show peak flows in forested watersheds change after they have been logged to some degree. Some reasons for this conclusion are as follows: (1 ) "complex nature of the effects of forest cutting and roads on streamflow (Thomas andMegahan 1998);" (2 ) "difficult to detect the effects of forest cutting and road construction from historical data in large watersheds . . . because of the lack of physical and statistical controls (Thomas and Megahan 1998);" (3 )"peak flows may be larger, smaller, or unchanged after logging, depending on what part of the hydrologic system is altered and how much (Harr et al. 1982);" (4 ) "increases in annual yield for the large watershed will be about the same size as the error in measuring streamflow in the large stream (Harr et al. 1982);" (5 ) "temporal and spatial distribution of precipitation and snowmelt inputs, the partial heterogeneity of basin characteristics, gauging stations that are not of research calibre, and the temporal and spatial variability of land uses, coupled with the fact that only a small portion of particular large basin will experience a land use practice in any given year (Beschta et al. 2000)." 19 Finally, a literature review of snow-melt dominated peak flows performed by Scherer (2001) on research east of the Cascade and Coast Range mountains within western Canada and U.S., found that changes are highly variable with no apparent direct correlation with the amount of Equivalent Clearcut Area (ECA) as proposed by the Interior Watershed Assessment Procedure (IWAP) that is used in B.C. In other words, it appears to be valid to assume that the dominant discharge, or mean annual flood, should not change due to a coniferous vegetation removal within a watershed of significant size. 2.2.1.2 Hydrologic History The dominant discharge is the most widely used variable for hydrologic control on channels, but another consideration for discharge is the hydrologic history of the drainage basin. Recent large floods (> 10 year return period) have the tendency to widen channels by lateral erosion and by stripping the vegetation from the banks, where channels tend to remain constant or narrow during times of average discharges or minor flooding (Howard et al. 1970). Therefore, the hydrologic regime of a channel should be taken into consideration when discerning the potential morphology of the stream, but it is difficult to analytically quantify this relationship with a mathematical formulation. One relationship that has been studied is the effects of catastrophic floods on channel morphology. 2.2.1.3 Catastrophic Floods Theories exist that are contrary to the idea that most geomorphic work is attained by frequent events, or the mean annual flood, as discussed previously. Typically, these so 20 called "catastrophic events", events that are infrequent and at a magnitude that exceeds the equilibrium threshold, are dependent upon particular climatic and physiographic influences or intrinsic and extrinsic factors for channel conditions (Baker 1977). As discussed by Baker (1977), the extrinsic factors of the basin physiography may include high rainfall intensities from convective thunderstorms and orographic barriers, steep drainage slopes and high relief ratios that drastically decrease hydrograph lag times, and impermeable soils to limit the infiltration rate during rainfall events. All of these factors tend to create the condition where overland flow (direct surface flow) dominates compared to interflow (vadose zone flow) and baseflow (groundwater recharge), which leans towards an increased potential to violate the Wolman-Millar principle (Baker 1977). That is, large infrequent floods can control channel morphology as opposed to smaller, more frequent (annual) floods. The proper characteristics for this process to exist and dominate a fluvial system commonly occur in arid regions, as compared to the humid and temperate regions of southern British Columbia. "Streams in areas of higher mean annual precipitation exhibit much less variation between the rare floods and the more frequent floods, such as the 2-year event, than do streams in more arid regions (Baker 1977)." Inclusive with the extrinsic factors is the flood's dynamic interactions with the intrinsic properties of the channel. Some of these properties are discussed more thoroughly in the following sections, but they may include particle size of channel and bank sediment, rock type and bedrock controls, existence of chutes and cut-offs, and vegetative influences in 21 terms of sediment cohesion and flow retardation. These interactions are complex natural conditions that require extensive knowledge of the geomorphic environment and typically are a local phenomenon as opposed to a system wide scale. There are two relationships that attempt to describe the potential for catastrophic floods to have a substantial control on the geomorphology of fluvial systems. 2.2.1.3.1 Flash Flood Magnitude Index The flash flood magnitude index (FFMI) is an attempt at quantifying the potential dominance of catastrophic floods on the work in fluvial geomorphic systems and supplies "some measure of regional magnitude-frequency relationships (Baker 1977)." The procedure was first proposed by Beard (1975). The FFMI is the standard deviation of the base 10 log-transformed annual flood peaks. For this relationship Xj is the annual maximum event (instantaneous peak), X is the mean flood event (average of yearly instantaneous peaks for available data), n is the number of gauged events, and X; and X are expressed as base 10 logarithms (Baker 1977). Obviously, this relationship may only be applied to fluvial systems that have a yearly discharge since the log of 0 is an indeterminate form. Therefore, it is suggested that this measure of flood-frequency variability could be synonymous with catastrophic flood potential, which could lead to the situation where catastrophic floods determine the (Eq. 2.1) 22 morphology of fluvial systems. Note the high values of FFMI can also result from series with very low flows in the absence of any large floods. 2.2.1.3.2 Ratio of Peak to Mean Annual Flood A study by Stevens et. al. (1975) focused on the effects of extreme flood events on the erosion and widening of river channel form. In this study, two converging rivers, the Tonoro and Guanipa rivers in northeastern Venezuela, were the focus of the study due to their differences in channel form. The Tonoro River was a wide sand-bed river and straight, while the Guanipa was a narrow and very sinuous river (Stevens et al. 1975). Not only were the river forms different but the upper Guanipa watershed is approximately 2,800 km and the confluence area of the Tonoro watershed is approximately 1,300 km . Based on a gauging station program performed by Colorado State University from June to September during the 1969 rainy season, Stevens et. al. (1975) concluded that the difference in Tonoro and Guanipa river forms immediately upstream of the confluence is due to the difference in peak-flood discharge where the Rio Tonoro is approximately 183 meters wide at bankfull stage and the Rio Guanipa is approximately 15 meters wide at bankfull stage. It is true that this discrepancy in width for the two rivers is quite significant, especially since the drainage area of the Guanipa is almost twice that of the Tonoro, but the conclusion was based primarily on the fact that the ratio of the peak to average discharge for the four months of record on the Tonoro was 47 compared to an estimate of 6 for the Guanipa. 23 During this single gauging season, the flood on the Tonoro (535 m3/s) was significantly greater than the flood on the Guanipa (105 m3/s), but this record solely represents one rainy season and four months of data. Also, there are a number of other factors that should be considered. The Tonoro has a channel slope of 0.006 and valley slope of 0.0015, riverbed sediment of 0.35-mm sand, and no floodplain, as compared to slopes 0.00055 and 0.013, riverbed sediment "probably 0.35-mm sand", and a broad, heavily vegetated flood plain for the Guanipa (Stevens et al. 1975). Therefore, the ratio of the peak-flood to mean annual flood may give some insight into the size and flashiness of floods within a system, but there are numerous other geomorphic factors that should be considered when attempting to understand the geomorphic condition of a fluvial system. In particular, the role of riparian and flood plain vegetation was not even considered. 2.2.1.4 Bank Erosion by Fluvial Entrainment The reason for discharge being such an important variable in fluvial channel form and morphology is due to the process of fluvial entrainment, where discharge and channel slope, a surrogate for energy slope, combine to produce peak levels of flow erosivity usually in mid-basin reaches (Abernethy and Rutherfurd 1998). Fluvial entrainment is the process where the lift force due to the water overcomes the resistance due to the particles submerged weight and inter-granular friction resulting in sediment removal and transport within the channel (Richards 1982; Abernethy and Rutherfurd 1998; ASCE Task Committee on Hydraulics 1998). (Figure 2-2) In other words, the shear stress imposed by the water on the sediment particle overcomes the internally derived resistive shear stress between the particle and the surrounding matrix of particles in the boundary, 24 x > T C (Eq. 2.2) where x is the shear stress per unit area due to the flow and xc is the critical or threshold shear stress required to transport the particle (Wolman and Miller 1960; Thorne 1982). The critical shear stress is defined at the point of limiting equilibrium or where the moment due to the submerged weight of the particle is balanced by the moment due to the drag force Tc = pg(s - X)dA tan <j> (Eq. 2.3) o where pg(s-l) is the submerged unit weight of the individual particles, s is specific gravity, d is the particle size, n (= nd2) is a dimensionless measure of the packing of the particles on the bed, n is number of particles on a unit area of boundary surface, and § is angle of interlock among particles (Carson 1971). (Figure 2-2) Eq. 2.3 may be rearranged in terms of dimensionless shear stress: pg(s-\)d 6 r For a particular packing arrangement, TJ ° C tan (j). (Eq. 2.5) 25 Flrao Fnrr.e Suhme.rfreH Weight Figure 2-2 Moments of Submerged Particle on a Horizontal Bed at the Threshold of Movement (Carson 1971) This process results in the removal, or erosion, of loose alluvial gravel from the bed and banks of the stream. Therefore, if the shear stress of the mean flow or turbulent vortices is sufficient, the flow has the capability to degrade the channel bed or laterally erode the noncohesive channel banks. This erosion could possibly result in degradation and/or lateral shifting and expansion of the channel and addition of sediment to the system. This may also be the case when erosion of channel banks is unbalanced compared to deposition of point-bar sediment accumulations (Nanson and Hickin 1983). 26 It is difficult to directly link the magnitude of a flood and its entrainment capabilities leading to morphological change due to complex interactions between turbulent hydraulic parameters and the properties of the bank material (Thorne and Lewin 1979). It is possible that floods of similar magnitudes could transport different amounts of sediment based on the sediment availability and the time of the flood event (Nanson 1980). Consequently, if loose sediment is available at the toe of a bank prior to a flood, than it may be entrained, transported, and deposited downstream during that flood while it may not be available during the next flood of the same magnitude. This is usually not the circumstance when a channel has non-cohesive banks and will be discussed further in section 2.2.4. 2.2.2 Sediment Load The imposed sediment load is a significant independent variable in determining the stability of an alluvial channel. An increase or decrease in the sediment load or calibre can launch a channel into disequilibrium and result in drastic morphological adjustment in the system (Church 2002). For example, if the amount or calibre of the sediment load increases then the system may not have the ability or capacity to transport the altered sediment condition. The decreased capacity will result in aggradation of sediment, a decrease in the depth of the channel, and increased slope relative to the valley slope (Maddock 1972). Without a cognate change in flow regime, the flow will accelerate due to the decreased depth and increase the erosive stress on the banks of the channel. This increased stress may be considerable enough to erode the channel and the adjacent riparian environment, which inherently increases the sediment delivery to the system and can cause downstream widening with a potential for a cyclic disturbance or feedback 27 mechanism (Jackson and Haveren 1984; Roberts and Church 1986; Powell 1987; Thorne and Osman 1988; Simon and Thorne 1996). Typically for gravel-bed rivers with noncohesive banks the accruement of sediment is from lateral erosion of the banks as opposed to scouring of the bed (Thorne and Osman 1988). If this process advances far enough then the channel may become adequately wide such that the single-thread channel evolves into a braided morphology due to the growth of fluvial bars (Brice 1964; Bridge 1993; Simon 1994; Church 2002). 2.2.2.1 Surface Sources Sediment accruing in a channel may be elevated due to surface sediment sources caused by changes in land use. Studies have shown that a land use change related to forestry practices can alter the timing and magnitude of sediment generation due to the watershed disturbance (Beschta 1978). These sediment yields may be contributed to an increase in bare ground surface caused by road creation, tree felling, and yarding techniques and increased landslides and frequency of debris torrents (Beschta 1978; Beschta 1984; Grant et al. 1984). Therefore, an increased sediment yield from surface sources could potentially contribute to the increased sediment load on a channel, which could tend to disequilibrium in the system. Typically in gravel-bed rivers with cohesionless banks, these increases in sediment to the channel proceed in increasing the active channel width through aggradation of the loose sediments (Beschta 1984). According to Roberts and Church (1986), logging activities solely on hill slopes probably do not contribute sediment wedges to stream channels if lower slopes are cleared when compared to potential bank erosion contributions. 28 \ 2.2.3 Valley Slope On short geological time scales, it appears reasonable to assume that the valley slope of a drainage basin is constant and is the uppermost gradient for the river. According to Knighton (1998), valley slope is an inherent property of a system that determines the energy loss along a system. A river will meander along a valley slope to decrease its slope in accordance to the discharge and sediment load in order to reach a regime condition. Therefore, it may seem reasonable that the steeper the valley slope than the more sinuous the channel. Brice (1964) discusses the alternative viewpoint of Lane (1957) that a steep valley slope will promote a braided planform and low valley slopes will encourage a meandering pattern. Therefore, there should be some transition where a slope is hypothetically steep enough to cause a historically meandering river to adopt a braided planform due to an increase in the channel slope from aggradation or straightening. This theoretical meandering-braiding transition will be discussed more thoroughly in section 2.3. These discussions of valley slope and channel pattern are not isolated from other variables of the fluvial system. For instance, if the bank erodibility of a system is high than its ability to meander and resulting sinuosity will be decreased and its width increased to balance the energy of the system. 2.2.4 Bank Stability The strength of the banks of an alluvial channel may have a direct influence on the potential morphology of the stream through the interaction between the systems potential energy and resulting fluvial entrainment and sediment availability and load. Numerous 29 factors of channel form are directly related to bank stability as stated by Brice (1964): regards bank erodibility as the most significant single variable in the determination of channel pattern and that bank erodibility depends mainly on the particle size of bank material and on vegetal growth along the banks. It has also been stated that the sediment texture and vegetation, which comprises the banks, is a provision that may affect the rate and/or amount of a channel transition (Church 2002). Banks that can resist fluvial erosion decrease the amount of widening that can exist in a fluvial channel. If channel banks are susceptible to erosion processes than lateral migration of the channel is a potential response to fluvial instability. The susceptibility to erosion of channel banks is heavily determined by its spatially variable geotechnical composition, including material size-distribution, cohesiveness, vegetative binding, buried timber, animal disturbances, and ground ice (Hickin and Nanson 1984; Thorne 1990). The composition of alluvial riverbanks can be categorized into three main groups: cohesive, non-cohesive, and composite (Thorne and Lewin 1979). 2.2.4.1 Cohesive Banks Cohesive banks are comprised mainly of silt and clay particles. The cohesion in these banks results from cohesive bonds between the various particles, such as physico-chemical bonds between clay particles (Thorne 1990). The chemistry of the fluid and sediment particles can greatly increase the cohesive strength and erosion resistance. Experiments have shown that undisturbed cohesive banks have a greater potential to resist erosion forces as compared to non-cohesive banks (Thorne 1982). The erosion of 30 cohesive banks usually involves cantilever failure, desiccation, weathering, mass wasting, and fissures and cracking (Simon 1989; Abernethy and Rutherfurd 1998; Millar and Quick 1998; Abernethy and Rutherfurd 2000). These mass failure events are typically due to the properties of the cohesive material where shear stress usually increases more quickly with the depth of the bank compared to the shear strength (Thorne and Tovey 1981). Unfortunately, the intricacies and variability of cohesive bank erosion are not part of the scope of this thesis. 2.2.4.2 Non-cohesive Banks Non-cohesive banks in alluvial channels are usually developed from the deposition and erosion of fluvial deposits (Thorne and Osman 1988). The main erosion process involved with non-cohesive banks is the fluvial detachment and entrainment of individual grains or shallow slips from the face of the bank by hydraulic action (Thorne 1982; Thorne and Osman 1988; Abam 1997; ASCE Task Committee on Hydraulics 1998; Knighton 1998). The forces resisting the hydraulic erosion are usually the slope-normal component of submerged weight, inter-granular friction, close packing of grains or imbrication (the overlapping of irregular and angular particles), and apparent cohesion forces (Thorne and Osman 1988; Thorne 1990; Knighton 1998). The apparent cohesion in non-cohesive banks may be due to consolidation of the bank sediment, cementing by fines, small amounts of true cohesion due to silt and clay fractions, capillary suction in the unsaturated zone, and/or the binding effects of vegetation root masses and rhizomes (Thorne 1990; Millar and Quick 1993). 31 The main difficulty in analysing the potential erodibility of a non-cohesive bank is the variability in the properties of the resistant forces in both magnitude and location along the length of a channel (ASCE Task Committee on Hydraulics 1998). 2.2.4.3 Composite Banks A composite bank is a stratified bank constructed of a lower layer of non-cohesive gravel and sands overlain by a layer of deposited cohesive silt and clay material (Thorne and Lewin 1979; Thorne and Tovey 1981). Under most circumstances, the lower layer of non-cohesive material is scoured quicker than the fine-grained material, leaving an overhanging or cantilevered bank (Hooke 1979; Pizzuto 1984; Thorne 1990; A S C E Task Committee on Hydraulics 1998). If this cantilevered bank is either surcharged by vegetation, susceptible to non-aqueous forces, or tension cracks, it may fail in a rotational or block failure. On the other hand, uppermost cantilevers are typically stable from tensile failure due to strong root networks from surface vegetation, but depending on the thickness and rooting depth of the vegetation (Hooke 1979; Thorne and Tovey 1981; Pizzuto 1984; Abam 1997). The fluvial environment has a substantial influence on the dynamics of composite bank erosion. If the fluvial activity, or flood regime, is low, a large accumulation of gravel may persist at the toe of the bank and decrease the bank angle. Alternatively, i f the fluvial activity is high than the loose sediment is rapidly entrained from the base of the bank following failure. This control of bank erosion by the accumulation of loose sediment at the toe of the bank has been coined 'basal endpoint control' (Thorne and 32 Tovey 1981). This situation could cause a multi-stage relationship between flood magnitude and bank retreat with one flood inducing failure and a second flood removing the failed material from the bank toe. Assuming that the lower non-cohesive sediment controls the geotechnical strength, and hence the height of the bank, it is possible to analyse a composite bank using the same characteristics of a purely non-cohesive bank (Pizzuto 1984). The relative importance of the separate layers will depend on the thickness and geotechnical properties of each individual layer (Thorne 1990). 2.2.4.4 Vegetative Influences The influence of vegetation on stream morphology and channel dynamics may take many different forms. The vegetative influence of most importance in this study in the root reinforcement to noncohesive channel banks induced by large coniferous riparian vegetation. Another important influence of vegetation on stream morphology is the complexity and control induced by large woody debris (LWD) residing within the channel boundaries. 2.2.4.4.1 Root Reinforcement Many researchers have discussed the effects of vegetation on the stability of a soil matrix. This soil matrix may be a stream bank or steep hill slope; anywhere mature trees exist in substantial communities. Simply stated, many researchers conclude that vegetation through a living root network has the potential to increase bank stability by decreasing the erosion rate on banks exposed to fluvial forces and increasing sediment shear strength 33 through binding and buttressing of the tree roots (Brice 1964; Smith 1976; Bowie 1982; Beschta 1984; Hickin 1984; Hickin and Nanson 1984; Powell 1987; Wilson et al. 1995; Frankenberg et al. 1996; Huang and Nanson 1997). (Figure 2-3) As stated in the ASCE Task Committee study (1998), many studies have shown that the erosion of well-vegetated banks as compared to unvegetated banks is reduced by one to two orders of magnitude. The study by Smith (1976) on the Kicking Horse River in British Columbia concluded that the erodibility of bank alluvium varied inversely and exponentially with root density. In addition, near bank velocity and direction has been shown to increase, which could lead to higher imposed shear stresses, when bank vegetation is removed leading to a decrease in velocity retarding bank roughness (Maddock 1972; Jarrett 1984; Thorne and Furbish 1995). The bank roughness retards the flow by increasing the turbulence properties of the flow that causes energy loss within the flow (Jarrett 1984). An example of the effects of vegetation on river channel planform was conducted by Nevins (1969) on the Turanganui River in New Zealand, were the channel quickly changed from a braided to a meandering pattern after a designed willow planting. In a study of disturbed watersheds in the Queen Charlotte Ranges, riparian logging practices were determined to severely disturb the stream banks, which initiated substantial channel widening and the formation of fluvial sediment wedges (Roberts and Church 1986). 34 W i n d load ing Protect ion by g r o u n d vegeta t ion against eros ion and surface traff ic MA Butt ressing by root cy l inde A n c h o r i n g and but t ress ing bv tap-roots (b) Mechanica l effects Figure 2-3 Physical effects of vegetation (Figure 3.1 in (Coppin 1990)) Under most circumstances the roots of the vegetation reinforce the soil matrix through the tensile strength of the roots imbedded in the soil matrix creating a composite material, which is strong in compression, therefore increasing the shear strength of the soil (Thorne and Osman 1988; Gray and MacDonald 1989; Coppin and Richards 1990; Thorne 1990; Simon 1994; Morgan and Rickson 1995; Styczen and Morgan 1995; Wu 1995; ASCE Task Committee on Hydraulics 1998; Rowntree and Dollar 1999; Simon and Collison 2002). Vidal (1969) composed the term "reinforced earth" to describe the effects of roots 35 and rooting depth on the shear strength of a soil matrix. In the case of trees, this effect can extend to several metres in depth and spread (Coppin and Richards 1990). The apparent cohesion caused by the root reinforcement and imbrication of particles leads to an increase in the critical shear strength necessary for fluvial entrainment of the bank particles by corrasion, which usually requires an increase in the flow intensity to perform bank erosion (Thorne and Osman 1988; Abernethy and Rutherfurd 1996). In terms of vegetative disturbances, it was reported by O'Loughlin and Ziemer (1982) that after a tree is removed small roots decrease strength at average rates between 300 and 500 kPa per month. Therefore, 3-5 years after removal of a parent Douglas fir or cedar small roots may lose over half of the initial tensile strength (O'Loughlin 1974). 2.2.4.4.2 Large Woody Debris Along with stabilising cohesionless banks with root reinforcement, large streamside vegetation has been shown to be influential in other dynamic equilibrium mechanisms. The most important characteristics within this aspect are the natural formation of log steps by fallen riparian trees. In a 5-year study conducted by Heede (1985), log steps were shown to be surrogates for gravel bars and reduce bedload movement in small mountain streams. Once these log steps were removed, streams were thrown out of dynamic equilibrium and adjustments were initiated in terms of increased bedload and suspended load transport, which decreased the water quality, stable spawning environments, and overall habitat within the channel. The density of these logjams can be quite erratic. A study by Gregory et al. (1985) estimated 270 dams in a 11.4 km2 drainage basin with an average of 1 dam every 27 m of main channel. In comparison, the 36 spacing for log steps were noted to be about 395 m of channel in the central Oregon coast range (Marston 1982). Not only does large vegetation get incorporated into the stream dynamics as log steps, but also large woody debris (LWD) creates stream complexity and habitat. Along with creating complexity within the environment, LWD is important in stream channel dynamics to control the force of the flowing water through the dissipation of hydraulic energy (Hartman and Scrivener 1990). These placements of LWD have been found to be the most influential structure element for pool formation within small to large river systems (D'Aoust 1998). 2.2.4.4.3 Width Morphology Alluvial rivers existing in coarse material typically have an armour layer or bed-pavement protecting the channel bed from degradation, which typically results in channel widening due to less resistant banks, especially after a riparian disturbance (Bray 1987; Simon 1994). The channel widens to reduce flow energy by decreasing the hydraulic depth (pressure energy) and flow velocity (kinetic energy) by increasing the resistance by enhancement of the relative roughness of the channel (Simon 1994). In creating a wider and shallower cross-section, the channel is attempting to regain an equilibrium condition with less resistant banks. Secondary currents drive erosion mechanisms and remove material from the toe of the bank, causing undermining and collapse of the bank material (Frankenberg et al. 1996). These processes may be accelerated without the existence of bank strengthening vegetation. 37 The width of a channel section has been stated as being a direct measure of bank erodibility when considering uniform slope and discharge (Brice 1964). Brice also discussed that sinuosity is directly linked to the bank strength, where low sinuosity was associated with low and high erodibility and high sinuosity associated with intermediate bank erodibilities. The results of various studies on the width of gravel-bed rivers show that rivers with heavily vegetated banks (shrubs and trees) were narrower than rivers with sparsely vegetated or non-vegetated banks (Charlton et al. 1978; Andrews 1984; Hey and Thorne 1986; Ikeda and Izumi 1990; Millar and Quick 1993; Huang and Nanson 1997; Piegay etal. 1997). An experimental study conducted by Gran and Paola (2001) at the St. Anthony Falls Laboratory used a flume with alfalfa seeded and unseeded runs to examine the effects of riparian vegetation on gravel-bed streams. The sand bed was run to construct braided stream morphology and then the bed was seeded with alfalfa sprouts to observe the effects of the vegetation on the established flow dynamics and channel form. The results from the study were that the "vegetation reduced the number of active channels and increased bank stability, leading to lower lateral migration rates, narrower and deeper channels, and increased channel relief (Gran and Paola 2001)." Carnation Creek located on the west coast of Vancouver Island was used as an experimental watershed to analyse the effects of logging practices. It was observed that no changes occurred in any study sections during the prelogging period (1971-76), but after logging the channel width increased significantly only in the sections where the stream banks had been logged 38 (Hartman and Scrivener 1990). In a study on the sandy Merced River, channel widening and bank erosion was primarily caused by the destruction of riparian vegetation due to human trampling with an average increase in width from 1919 to 1986 of 27% (Madej et al. 1994). Madej stated that climatic or hydrologic trends could not account for the amount of bank erosion on the Merced River. A study on the North Nashwaaksis Stream in New Brunswick showed an increase in channel width by 57- 160 % on a channelized and disturbed reach with the upstream and downstream vegetated reaches remaining unchanged during a particular flood event (Bray 1987). Included within this study was a schematic diagram of the process of recovery of the disturbed section in the following steps: widening, mid-channel bar formation, bar colonisation by vegetation, and degradation of new low-flow channel (Figure 9 of (Bray 1987)). This has also been studied for the amount of erosion occurring on vegetated bends versus non-vegetated bends with the difference in erosion observed during a major flood event attributed to the presence of vegetation in the study bends (Beeson and Doyle 1995). One of the main difficulties in characterising the effects of the riparian roots on the shear strength of the particle matrix is the wide variety of root strengths due to different types, strengths, patterns, rooting depth, density, and overall geometric development (Morgan and Rickson 1995; Styczen and Morgan 1995; Wu 1995; Abernethy and Rutherfurd 2000). Therefore, the amount of soil erosion reduction motivated by riparian vegetation can drastically change in the lateral distance of the bank due to age, type, and density of the riparian vegetation. This has made it difficult to employ the effects of vegetation into bank stability analysis. 39 2.2.4.4.4 Scaling Inconsistencies in scaling exist between the effects of tree-lined banks compared to grass-lined banks (Zimmerman et al. 1967; Murgatroyd and Ternan 1983). Field studies of gravel-bed rivers by Charlton et al. (1978), Hey and Thorne (1986), and Andrews (1984) have shown that channels with grassy banks are substantially wider than channels with tree-lined banks. Contrary to these studies, there are results that show that grassy banks have narrower channels than forested or tree-lined banks (Zimmerman et al. 1967; Murgatroyd and Ternan 1983; Trimble 1997; Hession et al. 2000; Hession 2001). There appears to be a threshold width, drainage area, or discharge on a channel where grass vegetation is more influential on narrower channels and tree vegetation is more influential on wider channels in terms of increased stability (Zimmerman et al. 1967). With regard to small streams, the rooting depths of the grass sod scale closely to the depth of the channel creating a cohesive bank, resulting in the narrowest channel (Gregory and Gurnell 1988). Grass would be an important controlling factor where subaerial mechanisms, such as frost heave, desiccation, rainsplash and micro-rill development, are mitigated by a grass riparian cover on exposed banks (Lawler 1992; Abernethy and Rutherfurd 1998). As discussed in Abernethy and Rutherfurd (1998), Zimmerman et al. (1967) discovered that channels where windthrow and LWD damming are active that the dimensions of the channel are double or triple those of an equivalent non-vegetated channel in the same geologic setting and discharge regime. Also, it seems that small streams differ from rivers due to their frequency of channel forming (dominant) discharges. As discussed in section 2.2.1.1 Dominant Discharge, rivers are 40 regulated by the bankfull stage and small stream dimensions are controlled by more frequent flows at roughly a yearly occurrence (Leopold et al. 1964). Therefore, it appears that grass lined banks are more effective than tree lined banks on very small catchments. The catchments used for this study are beyond the size of catchments where grass is the more influential vegetative force. 2.2.4.5 Bank Stability Algorithm Millar and Quick (1993) developed a model for non-cohesive gravel sediments based on the bank-stability analysis derived by the United States Bureau of Reclamation (USBR). The USBR theory, which was developed by Lane (1955), is presented in a simplified form, under the theory of tractive force, in the classic text books by Chow (1959) (p. 171) and Henderson (1966) (p.419) and restated in dimensionless form by Millar (1994): specific gravity of bank sediment, D50 is median bed or bank sediment diameter (meters), t*bedc is the critical dimensionless (Shields) shear stress for a grain equivalent D5o on the channel bed, and 0 is bank angle (degrees). The friction angle § for coarse loose gravel approaches a maximum value of 40°, the angle of repose for gravel (Lane 1955). However, according to Millar and Quick (1993) (|> may be replaced by the modified friction angle ([>' to account for the bank stabilising effects of cohesive silt and clay cementing of the grains and vegetation. Other (Eq. 2.6) where ibank is reach-averaged bank shear stress (Pa), y is unit weight of water (N m"3), s is 41 researchers characterized (|)'as an expression of the frictional interaction and interlocking between particles (O'Loughlin and Ziemer 1982). A substitution of cj)'applies a potential solution to the inability for the magnitude and distribution of root reinforcement to be applied in bank stability analysis since historically the influence of vegetation has been excluded from research based on the difficulty to quantify and physically or statistically manipulate its influences (Hickin 1984; Thorne and Osman 1988; Abernethy and Rutherfurd 2000). The new modified friction angle §' is allowed to have a minimum value of approximately 40°, which indicates no or little influence from vegetation, and approach a maximum value of 90° (Millar 2000). The is a single calibration parameter that may be mostly used to quantify the stabilising effects of vegetative root networks and the reduction of near-bank velocity and shear. But, also must include the effects of other processes such as the packing and imbrication of the sediments and cementing of the gravel particles by the cohesive properties of interstitial clays and silts (Millar 2000). These complicated processes make clarifying a §' value for given situation and vegetation conditions, such as species, scale, and root density, type, and depth, extremely difficult. It has been shown by Millar and Quick (1993) and Millar (2000) that there is a relationship between (j)' and vegetation density (see Table 1), but the conditions and the scaling issues depicted in the previous section must also be a concern. 42 Andrews (1984) Vegetation Type Hey and Thorne (1986) Vegetation Type <))' (degrees) Thin Thick I II III IV Mean 37.5 51.6 39.9 43.7 48 55.6 Range 29.3-46.3 42.7-58.8 20.1-51.7 30.1-58.8 32.0-72.0 36.3-79.1 N 14 10 12 16 13 20 Table 2-1 Friction angle values ((/(degrees) obtained through model calibration of each vegetation class, from Millar (2000) As stated in Millar (1994), it has been shown by Henderson (1966) (p.414) through the work of White (1940) that the dimensionless shear stress is related to § (proportional form shown in Eq. 2.5): r bedc = k tan <f> (Eq. 2.7) Therefore, to include the stabilising effects of vegetation on the banks (j) can be replaced by the modified friction angle ())', and inserting the above equation, the bank stability algorithm may be simplified (Millar 1994). ' bank <0.048tan^\ l-—'0 sin (f> (Eq. 2.8) Finally, the remaining coefficient k has been shown to be directly related to the critical dimensionless shear stress for bank sediment and approximate a value of 0.048 in order to display agreement between calculated and observed channel dimensions (Millar 1994; Millar 2000). 43 2.3 Meandering-Braiding Transition Leopold and Wolman (1957) proposed a transition discriminating between meandering, braiding and straight channels. The theory is that a continuum of channel patterns exists with braided patterns distinguished from meandering ones by particular combinations of slope, discharge, and width-to-depth ratio (Leopold and Wolman 1957). In their relationship, the channel slope was plotted versus the bankfull discharge to acquire the following empirical function: S* = 0.06<26/~°44 (Eq. 2.9) where Q b f is the bankfull discharge and S* is the critical channel slope between meandering and straight or braiding channels. For this relationship, a channel that had a slope S<S* would tend to be meandering and S>S* would tend to be braided. A good summary of various relationships for the hydraulic controls on channel patterns is available in Bridge (1993). 2.3.1 Theoretical Stability Analysis Since the formation of the Leopold and Wolman relationship, various authors (Parker 1976; Fredsoe 1978; Struiksma and Klaassen 1988) have performed theoretical stability analysis for the purpose of exploring bed form and bar formation, meander development, or braiding in alluvial rivers (Millar 2000). These theories use depth-averaged, two-dimensional forms of the equations of motion for momentum, mass, and sediment continuity. One example of the balance equations is supplied in Parker (1976) 44 du du du — + u — + v — dt dx dy dv dv dv + u — + v — dt dx dy ' dx pd dh ?y dy pd' , d , dd . (ud) + — (vd) + — = 0, dx dy dt d_ dx d(h-d) dt + • fdqx , dq} ^ P V dx dy (Eq. 2.10) where d is the channel depth, h is the water surface height above a reference, (u,v) is the stream velocity vector, (qx,qy) is the volumetric sediment transport vector, ( T x , x y ) is the bed stress vector, p is the density of water, and Xp is the bed porosity. In these analyses, a perturbation technique is used that includes a small variable in the form of a ratio of sediment transport to water transport. These perturbations are applied to the balance equations of motion about a steady flow condition. The stability is analyzed when a perturbation may propagate and develop a fluvial instability within the system. The fluvial instability could represent the formation of bars, meanders, or braids (Parker 1976; Millar 2000). Finally, the stability analysis formulates a theoretical criterion for an approximate division between meandering and braiding states. According to Parker (1976), the approximate relation for this division is £ = C0 SW nFY = 0(1) (Eq. 2.11) 45 where C 0 is a dimensionless friction factor, ICB is a dimensionless wave number, S is the slope, W is the width (meters), Y is the depth (meters), and F is the Froude number that equals U/V(gY), where U is the mean velocity (ms"1) and g is gravitational acceleration (ms"). This criterion for the situation where steady flow forms multiple cells, or the onset of multiple channel formation, may be written in terms of the channel slope: (Eq. 2.12) where S is the transitional slope between meandering and braiding, Y is bankfull depth (meters), W is bankfull width (meters), and F is Froude number (Parker 1976). According to Millar (2000), the direct application of this criterion is difficult because the channel dimensions and Froude number at transition must be known to be able to calculate an approximate transitional slope. Therefore, the above formulation in conjunction with the Millar and Quick (1993) analytical model were used to create a function for the meandering-braiding transitional slope in terms of the relevant independent variables discharge, sediment size, and bank stability (Millar 2000). Historically, transition formulations such as Henderson (1963), Ferguson (1984), Chang (1985), and Bridge (1993) have included bank stability within the functions, but none have considered imposed vegetation strength in the independent variables. • Henderson (1966) S' = 0.64D1 UQ-°M (Eq. 2.13) • Ferguson (1984) S* =0.042Q-°A9D50009 (Eq.2.14) 46 • Chang (1985) S'*aQ-05D 0.5 (Eq.2.15) • Bridge (1993) S' =0.0049(9"0 21D, 50 0.52 (Eq. 2.16) As can be seen from these four relationships, the threshold slope appears to be dependent on the discharge and sediment caliber (resistance and capacity), but there is no analysis of bank resistance other than sediment size. 2.3.2 Millar Theoretical Meandering-Braiding Transition Millar (2000) composed a single approximate multivariate relationship represented by a threshold value S*: where D5o is median sediment diameter for the banks and bed surface (meters), <j)' is bank sediment friction angle (degrees), and Q is the bankfull discharge (mV1). For this relationship a value of S « S * implies that the channel is meandering and a value of S » S * implies that the channel is braided (Millar 2000). The Millar meandering-braiding transition was tested against 86 meandering rivers from the data sets of Andrews (1984) and (1986) and 32 braided rivers from Fahnestock (1963). The results from this test were that all of the meandering rivers plot in the correct area where S<S* and all of the braided rivers, except for three points, were in the zone where S>S* (Millar 2000). These results demonstrate that the Millar meandering-braiding transition can fairly well predict the transitional slope where a meandering river S* = 0.0002£> 5 0 0 6 V 7 5 Q ,-0.25 (Eq. 2.17) 47 may become a braiding river, based on the independent variables of sediment calibre, bank sediment strength, and bankfull discharge. In a recent draft paper by Millar (in prep), output from the same model as generated equation 2.20 is reanalysed to obtain a new dimensionless form of the meandering-braiding transition relationship. The nondimensional approach utilized is similar to the studies of Parker (1979) and Andrews (1984) represented by the following dimensionless independent variables: Q* = 2 • Q (Eq.2.18) u' = ^ (Eq.2.19) tanctf where D50 is median sediment diameter for the bed surface (metres), g is gravitational acceleration (mV1), Q is the bankfull discharge (mV1), or mean-annual flood, and s is the specific gravity of the sediment (typically estimated at 2.65 for quartz) (Millar in prep). The dimensionless variable u' depicts the ratio of the strength of the banks relative to the bed sediment strength or more correctly, the value of the critical dimensionless shear stress of the bank relative to the bed. If the sediment size of the banks is substantially different than that of the bed, u' is adjusted through the consideration of <j)' to incorporate this difference in potential strength. Therefore, a value for u' >1 characterizes a situation where the banks are more resistant to erosion than the bed sediment, and a value for u' = 1 characterizes the condition when the banks and bed are both composed of equally erodible loose sediment that is noncohesive, i.e. no vegetative 48 influences (Millar in prep). The resultant meandering-braiding criterion using dimensionless variables is the following model: S* = 0.951 Q *-°25 u' (Eq. 2.20) It should be noted that Q* contains D 5 0 2 5 in the denominator so that Q*" 0 2 5 is proportional to Q"° 2 5 D 5 o ° 6 2 5 , which is similar to Eq. 2.17 (Millar in prep). 2.4 Proposed Riparian Disturbance Sensitivity Index Realising that the channel slope Sc is the prevailing slope where the channel is forming a meandering planform, or single-thread channel, then a relationship between Sc and S* potentially could determine the availability of a meandering gravel bed river to adjust its morphology to a braided situation. The proposed relationship for this analysis is the riparian disturbance sensitivity index. The riparian disturbance sensitivity index is a ratio between the meandering-braiding relationship and the channel slope of the channel. — = S c „,< (Eq.2.21) 5* 0.0957(2 * «' This relationship may be reduced to an empirical formula based on the meandering-braiding transition formula at a completely disturbed bank strength condition where (J) is equal to 40°. This new relationship is the riparian vegetation sensitivity index for non-cohesive gravel-bed rivers: C = 10.456?*° 2 5 Sc (Eq.2.22) where Sc is the channel slope (m/m) and Q* is the dimensionless discharge. For values of C, « 1 , a channel should be insensitive to riparian disturbance. For C, =1, the post 49 disturbance channel would be at the threshold between meanders and braided. For C, >1, the channel should be extremely sensitive to riparian disturbance. Theory shows that as slope increases, a meandering river may change from a meandering to a braiding planform, or the degree of braiding will increase (Parker 1976). Through the use of case studies, particular criteria could be set to quantify the sensitivity of a stream to fluvial geomorphologic alterations resulting from riparian vegetation removal. The hypothesis is that once C, is greater than some value, then it may be assumed that the channel could begin to show sensitivity to riparian vegetation removal. It is believed that as this ratio increases the sensitivity of the channel's stability due to vegetative influences will increase. This hypothesis was tested with a case study analysis from rivers with disturbed and intact coniferous riparian vegetation in the province of British Columbia. 5 0 CHAPTER 3 3 METHODOLGY The methodology for the case study analysis involved two main data collection components. These two components were fieldwork and analytical data collection. 3.1 Fieldwork The fieldwork includes a site survey, which includes grain size sampling of stream bank and bed sediments, longitudinal survey, and cross-sectional surveys. 3.1.1 Grain Size A pebble count analysis that was first proposed by Wolman (1954) must be conducted to understand the representative sediment of the stream. For this study, the representative sediment of the stream, or reach, is assumed to be the D50 particle size. The process depicted by Wolman is a sampling method that is applicable for streams flowing through coarse material and may be waded during low flow conditions. The pebble count is performed through selecting grains by a heel-to-toe walk technique or samplings at even-spaced marks along a measuring tape (Kellerhals and Bray 1971; Bunte and Abt 2001). For this analysis, the sampling technique along a measuring tape with outstretched finger and eyes closed will be used because of the decreased variability and operator error observed with the heel-to-toe technique (Kellerhals and Bray 1971; Bunte and Abt 2001). 51 3.1.1.1 Pebble Count Technique The pebble count along a tape consists of randomly selecting 100 pebbles at evenly spaced intersections along the tape, and classifying the pebbles based on measurements of the b-axis. As discussed in Rennie (1998), Bray (1972) suggests a pebble count of 50 stones to achieve a consistent estimate of the mean size of the distribution. However, according to Fripp and Diplas (1993) a sample of 400 stones is advised for accurate estimation of the entire grain size distribution. Since an estimate of the median sediment size is the required value for this study, a sample size of 100 stones should suffice for this requirement. The sampling interval is set to larger than the b-axis of the D m a x particle, in order to avoid counting a particle twice and biasing the sample (Bunte and Abt 2001). At each sampling interval, the particle located directly below the mark on the tape is selected and measured without introducing operator bias in the particle selection. In order to measure the b-axis of the particles correctly and quickly, the grains are passed through a template, or gravelometer (see Figure 3.1), consisting of square holes corresponding to the sizes of standard 0.5 -^increment sieve sets (Bunte and Abt 2001). The template used for this analysis starts at the separation between the gravel and sands range at 2 mm and extends to a size of 128 mm, but the sample will be truncated at 8 mm due to the interest in the gravel-sized particles for the sites. This truncation will introduce a small bias in the sample, but it was assumed that this bias is negligible for gravel sized sediments. Clasts larger than the 128 mm template hole will require manual measurement of the b-axis. 52 300 H t « -t * i < t J T . — , 4 5k t > » * * ... I - • • « ;; _ > > .... : 1 i • t t t i i 1 i "-128 ' t * ^ i « i • 11 i * - t t ! t « 1 I n t<i< tit: ... .... ::: ; ; i « ; ; ; • • tr.a j - -::: < • : ::; ::::::: ' ' • : K t 4 ~ > * » < • * * , > = — ,,, 22.6 • . . ,32. 43 J . > . > fiA . . . . Figure 3-1 Sample template (gravelometer) (Center 1996) Since the majority of stream particles approximate ellipsoids, the b-axis most closely represents the nominal diameter, where the nominal diameter is defined as the diameter of a sphere with the same volume (Center 1996). (Figure 3-2) The advantages of this technique are that the square openings in the template supply data that is equivalent to sieving in the field, the techniques are reproducible and reduce bias and operator error, and it is a relatively quick method to use in the field. A conversion 53 Figure 3-2 Grain axes, similar to Fig. 3.3 in Church (1987) from phi (()>) units and millimetres can be performed by the following equation supplied by Church et al. (1987): (j) = -log 2 D(mm) = -3.32191og10 D(mm) (Eq. 3.1) The smallest hole that the particle passes will be the size class listed for that particle, with each size class containing a percentage of the total sample (Kellerhals and Bray 1971). For example, if the particle has a diameter of 25 mm it would be recorded as smaller than the 32 mm template hole. These percentages can also be directly calculated as 54 frequencies of occurrence, which correlates to the percent finer distribution used in cumulative frequency distributions (Kellerhals and Bray 1971; Bunte and Abt 2001). The b-axis measurements are transformed into a cumulative frequency distribution that can estimate roughness values for the streambed and banks, assuming that the material consisting of the bed and banks are equivalent. In a study by Jackson (2001), the median particle size in clearcut headwater streams decreased. But, the reduction was not statistically conclusive due to high variability in the pre- and post- harvest sets. Therefore, since there is no historical data available, the particle size distributions formulated from the field studies will be assumed to be representative of the historical conditions. 3.1.1.2 Bulk-Sieve Technique The bulk-sieve technique is a volumetric sampling technique. This sampling technique is used to garner a predetermined mass of sediment from the streambed. Since a mass of material is being removed, the bulk-sieve sample is a three-dimensional sampling technique. Based on analysis by Diplas and Fripp (1992), the minimum depth of a volumetric sample is greater than the available depth of most surface layers. Therefore, the two-dimensional surface sediment cannot be sampled using this technique (Diplas and Fripp 1992; Bunte and Abt 2001). In order to sample the subsurface particles volumetrically, the entire layer of surface sediment must be removed from the sample area. This is necessary so that surface or 55 armour sediments do not contaminate the sample (Bunte and Abt 2001). The correct technique for the removal is suggested so that the armour layer is removed to the underside of the largest exposed particle available in the sampling region (Church et al. 1987). A technique for estimating the required size for the bulk sample has been suggested by Church et al. (1987) to be based on the largest available stone in the distribution (assumed to be equal to D90). In this criteria, it is suggested that the mass of the largest stone should not exceed 0.1 percent of the total sample mass for particles up to 32 mm, thereafter 1 percent to 128 mm. The reason for this criterion is so that the entire sample is independent of the dimensions of the individual particles (Diplas and Sutherland 1988). Assuming that the density of the particles is similar to quartz, the density of the largest particle would be 2650 kg/m . A sphere based on the above characteristics of the measured b-axis can estimate the volume of the largest stone. This results in the mass of the largest stone in kilograms to be equivalent to 1387.5*D9o3. Applying the total mass criterion, the formula for estimating the total volumetric sampling mass is: m*SampleMass>\m.5D^{kg) (Eq. 3.2) SampleMass > 138750/J903 (kg) (Eq. 3.3) Once the required sampling mass is achieved and positioned on the sampling tarp, the sample is separated into one half-phi size classes for all pebbles greater than 32mm based on the template sampling method. These groups of larger particles are meticulously brushed to remove any residual sand and smaller particles from their surface. After they 56 have been cleaned, the separate groups are returned to a clean sampling bucket and weighed using the field scale. These weights are tabulated and the particles are returned to the sampling location. The remaining sample consisting of all particles smaller than 32 mm is subdivided until a representative sample of approximately 8 kg is achieved. This subsample is retained in plastic sampling bags and returned to the lab for segregation of the smaller particles. The entire field process for one sample took approximately three hours, but varied based on the required sample size and conditions. It was not possible to do bulk sampling during precipitation events due to the fact that an unmeasurable water weight would be added to the sample during brushing and sieving of the larger particles. This added water weight could potentially cause large discrepancies in the original sample mass and final sample mass, especially the contribution from the sample of the particles finer than 32 mm. In the lab, the sample is removed from the sampling bags and discreetly spread out on a tarp so that the sample may be exposed to the heated air and dried. The sample must be dried to remove the water weight within the smaller particles. The dry sample was then fed through a system of sieves with square holes ranging from 25.4 mm to 1.0 mm. The sieve analysis was truncated at 1 mm because the gravel-sized particles were of most importance due to the condition that all sampling locations were on gravel bed rivers. Also, it was desirable to compare the results from the bulk sampling distributions with the results of the pebble counts, which are accurate for gravel particles or stones with a 57 diameter of 2 mm (Bunte and Abt 2001). The size fractions 25.4 mm and 18.85 mm were shaken by hand to segregate the particles. The material that was able to pass through the 18.85 mm sieve was split to obtain a manageable sample of between 1 and 2 kg to use for the rest of the analysis. The sieves of 9.52 mm and 6.35 mm were mechanically shaken for 4 minutes and the rest of the sieves (4.75 mm, 3.36 mm, 2.38 mm, 2.00 mm, 1.41 mm, and 1.0 mm) were mechanically shaken for 10 minutes. The samples for all size fractions were measured to the nearest 0.1 grams with an electronic scale accurate to .01 grams. Once all of the size fractions were tabulated according to their mass and corresponding sieve size; a cumulative frequency distribution was created for the subsurface bed and bank samples similar to the distributions from the pebble counts. The results from the sieve analysis are directly comparable to the pebble counts, or grid-by-number measurements (Kellerhals and Bray 1971; Diplas and Sutherland 1988). 3.1.1.3 Pebble Count versus Bulk- S ieve Technique The reason for selecting the pebble count procedure over the usual bulk-sieve analysis is primarily due to four factors. The first factor is that the pebble counts more appropriately covers a larger area of the streambed or bank, therefore most likely representing the reach average values. The second factor is pebble counts are more practical when trying to represent the coarser material. Since this analysis is for gravel bed streams, the percentage of surface represented by fine sediments is not a necessary factor for these particular relationships. The third factor is the time and work commitment for the pebble count versus a bulk-sieve analysis. An estimated completion time for the pebble count is roughly one-half of an hour compared to a half of a day or more for the bulk analysis. Finally, the bulk-sieve analysis would be more relevant for the analysis of bed-load 58 transport computations as compared to an interest in hydraulic friction or initiation of movement as this study is concerned (Kellerhals and Bray 1971). 3.1.2 Channel Geometric Survey The case study sites are based on a reach length of river. The reach length was determined by the conditions of the river and sampling time. In most cases, the reach length was between 10 and 20 channel widths long and incorporated at least four riffles and three pools, or three riffle pool sequences with the survey beginning and ending at a riffle section. A reach of 10 to 20 channel widths is a satisfactory length to associate stream morphology to channel processes and response and habitat characteristics (Montgomery and Buffington 1997). According to Knighton (1998), the average spacing of a riffle (shallows) and pools (deeps) is a distance between 5 to 7 times the channel width. The surveys were conducted with a rod and level for the sites located in the interior of British Columbia and the surveys were conducted with a total station for the sites located on Vancouver Island. The total station deployed for the surveys was the Leica TC705 and a Leica GPR111 circular prism. The TC705 has a standard measurement accuracy of 2mm+2ppm and range of between 1800 m - 3500 m, depending on the weather and visibility, and an internal data collection system. 3.1.2.1 Longitudinal Profile Beginning at the most upstream riffle in the reach, a longitudinal profile was completed along the length of the thalweg for the entire study reach. An elevation point was taken at intervals of roughly 10 meters in an attempt to include all existing breaks in slope and prominent features in the channel. Using the elevation measurements acquired from the survey along the thalweg, a contour of the bed surface was created by plotting the 59 elevation points against their horizontal position in the downstream direction. In order to acquire an estimate of the average channel slope for the reach, a simple linear regression was used on the data. The slope of the regression equation is an estimate of the average channel slope value for the reach. 3.1.2.2 Bankfull Channel Width or Unvegetated Channel Width The bankfull channel width is the width of the channel positioned at the top of the banks before entering the floodplain. It is the width of the channel that is formed by the frequency and work of the dominant discharge (see section 2.2.1.1). In the field, this width is usually recognized by the non-existence or scarcity of vegetative growth occurring at the sides of the channel. In order to compare the measured channel width with the historic photo width (section 3.2.5.2), the unvegetated channel width was the value used to calculate the average width of the channel. In some instances, such as situations with large prevalent unvegetated bars, the unvegetated channel width is drastically different than the main channel bankfull width. While using the rod and level, the unvegetated channel width was estimated by using a survey tape at increments of 30 m. The width was directly measured using a calculation of the distance with the Pythagorean theorem and the x and y co-ordinates of the end points with the total station survey data. 3.2 Analytical Methods After the field visits and data analysis were completed, a particular set of analytical methods was necessary to gather the remaining data for the case study analysis. These 60 analytical methods included watershed delineation, regional hydrology analysis, and historic aerial photo analysis. 3.2.1 Watershed Drainage Area The drainage area for a watershed consists of all of the land surface area that could potentially produce runoff from a precipitation event to the watershed outlet. The boundaries of the watershed must be delineated in order to show the break points or ridges that distinguish adjacent watersheds. The delineations for the sampled watersheds were completed by locating the sampled location on 1:250,000 topographical maps of British Columbia. Realising the end of the reach is the watershed outlet, the necessary line was constructed perpendicular to the elevation contour lines for land that would drain to the outlet in order to form the boundary for the watershed. Once the watershed boundaries were constructed, each watershed was scanned into AutoCAD as an image file. The boundary was identically recreated with a polygon and the area of the polygon was calculated with the measurement tool and a ratio between the AutoCAD scale and the true map scale. 3.2.2 Discharge A necessary data component for the sensitivity analysis is a reasonable estimate of the mean annual flood for the watershed (Q2.33). For this analysis, as stated in section 2.2.1.1, the mean annual flood is assumed to be equivalent to the bankfull discharge. In all cases, the necessary watershed is either situated upstream or downstream from an existing gauge or is located on an ungauged system. This leads to the use of a regional analysis to obtain an estimate of the mean annual flood for the watershed. 61 3.2.2.1 Regional Analysis The regional analysis for this study is the scaling and regionalization technique proposed by Eaton et al. (2002) specifically designed for flood flows in British Columbia. In this study, regionalization of the flood data shows that the simple scaling appears to apply to the province of British Columbia, since the scaling is not a function of the flood return period. Using a scaling and krigging method, Eaton et al. constructed a simple relationship to estimate the discharge for any location in the province: Qmaf =kmafAfJ5 (Eq.3.4) where Qm af is the mean annual flood, km af is the krigging factor for the mean annual flood and the specific location of interest, and Ad is the drainage area corresponding to the specific location of interest (Eaton et al. 2002). Therefore, an estimate of the mean annual flood for any watershed in British Columbia may be calculated with the use of the drainage area and relevant krigging factor. This seems reasonable because a hydrologic study completed by Pitlick (1994) on mountainous regions concluded that drainage area was the variable most highly correlated with the mean annual flood. The krigging factor may be extracted from the appropriate map of k values available in Eaton et al. or back calculated in two ways depending on the available gauging locations within the region of the watershed, which is the method used in this analysis. For a map of the krigging factors see Figure 8 of Eaton et al. (Eaton et al. 2002). If there is a gauge located on the same stream or river, than the krigging factor for the watershed may be back calculated by substituting the know drainage area and mean 62 annual flood for the gauge, acquired from the HYDAT database (2000), into the above equation and solving for the corresponding km af value. This value may then be applied with the estimated drainage area to calculate a value for the mean annual flood at the desired location. On the other hand, if there is no gauge situated upstream or downstream from the required location, than a regional analysis must be performed in order to obtain a value for the krigging factor to be used in the analysis. Focusing on the gauged streams in the corresponding hydrologic region, a watershed area to the power of 0.75 versus mean-annual flood discharge relationship should be created. Keeping in mind the aspect and proximity of the gauged streams used to create the relationship, an estimate of the krigging factor can be established with a slope estimate of the linear regression formed from the above mentioned regional relationship. If a linear regression appears inappropriate for an estimate, then an average value of the krigging factors is obtained. Once again, the krigging factor can be applied to the known drainage area of the watershed to calculate an estimate of the mean annual flood. It should be noted that the importance of this analysis is to obtain at a minimum a correct order of magnitude flood value for the dominant discharge. 3.2.3 Width A main focus of this analysis is the morphology of a river from an equilibrium state to a disturbed state and the change in width between these two states. It is an ideal situation where on-site historical and present day reach average widths are available. Under most circumstances this is not the case and aerial photos are the available source for geometrical measurements. Therefore, the representative width for the stream can be 63 calculated by averaging the widths of the reach taken over a certain increment in the stream length. For present day width values, a reach-averaged width may be surveyed in the field. 3.2.4 Aerial Photo Analysis In order to measure the effects of vegetation removal on stream morphology, it is suggested that an aerial photo analysis be used to visualize the changes in a stream's width before and after a disturbance to the riparian zone (Neill and Yaremko 1989). Since bankfull stream channel width is not dependent upon stage height, such as size of gravel bars or islands; it is an accessible planimetric measurement (Beaudry and Gottesfeld 1999). The majority of the photos used in this study were collected at the Air Photo Warehouse affiliated with the Base Mapping and Geomatic Services Branch of the Ministry of Sustainable Resource Management in Victoria, British Columbia. The two oldest photos, which are photographs of the Elk River in 1930, were laser copies obtained from the National Air Photo Library of the Centre for Topographic Information - Natural Resources Canada in Ottawa, Ontario. The majority of these photographs were the standard 23cm X 23cm black and white panchromatic print that is taken directly below the aircraft in a vertical alignment (Mollard and Janes 1984). A selection of the newer photos used were standard colour panchromatic editions. The vertical photographs were used because of the ability to calculate a scale and horizontal measurements directly on the photograph. The techniques used in this analysis are comparable to the technique discussed in the Roberts and Church (1986) study of disturbed watersheds of the Queen Charlotte Ranges, British Columbia and techniques outlined in Grant et al. (1984). 64 3.2.4.1 Scale The most direct way to calculate the scale of an aerial photo is by comparing a known distance between two points on the ground with the same two points on the photograph (Mollard and Janes 1984). When a discernible distance is not known on the ground, as is the case in the majority of the photos including the predisturbed photos, a simple geometric relation must be used to calculate the scale. Through the use of similar triangles, a ratio of the triangle bases is equal to the ratio of the triangle heights (Mollard and Janes 1984). a - (^airphoto dist.) /(focal length) = (Eq. 3.5) A - (^ground dist.) i/(flying height) F i g u r e 3-3 A e r i a l P h o t o S c a l e D i a g r a m 65 Therefore, the scale of the photograph may be determined by calculating the ratio of the focal length of the camera and the flying height of the airplane (Miles pers. comm.). The record of the camera focal length and airplane flying height may be obtained from the film record for each particular photographic roll number. In some cases, the aerial photo print is not a contact print, consequently the print is not the same size as the negative and the focal length must be adjusted to match this discrepancy. The enlargement ratio of the print size to negative size must be multiplied by the camera focal length to calculate the adjusted focal length for the print. print size(mm) m -i x\ — * f(m) = fadjusted (m) (Eq. 3.6) negative size(mm) The scale of a particular point or region on the photograph may be calculated more accurately by including the relief elevation (FT) for that point. The average elevation for the reach of interest was found for this study from 1:50,000 topographical maps. This elevation is subtracted from the flying height to get the elevation of the camera above the point of interest (Miles pers. comm.). scale = (Eq. 3.7) H(m)-H'(m) Once the scale has been calculated, the aerial photos may be used directly to take the necessary measurements for the study. Due to unknown error in the flying height, the photo scale was calculated to the nearest lm on the photo equals 100 m on the ground or 1:100 scale. 66 3.2.4.2 Measurements The main objectives of the aerial photo analysis is to obtain an estimate of the river's width prior to riparian disturbances. In order to obtain reasonably accurate measurements from the aerial photographs, a STM electronic digital caliper with a measuring range of 0-150mm, reading value of 0.01 mm, and accuracy of 0.03 mm was used to take the measurements. This was necessary because an error of 0.1 mm in the measurement would result in a width error of 3.5 m for a 1:35,000 aerial photo. This level of accuracy required a visualisation tool to assist in seeing the details on the photo. To accomplish this, a 2x magnifying glass or one lens of an Alan Gordon Enterprises pocket stereoscope was used in conjunction with the digital caliper. The procedure was initiated by orienting the photos of the same reach in a similar manner and locating the field surveyed reach on the photo. To obtain an average width value for the reach, bankfull width, or unvegetated width, measurements were taken at 50 m intervals starting upstream and working towards the downstream termination for the reach. The width was measured perpendicular to the flow at each location. Therefore, the number of measurements taken for each photo was dependent upon the length of the reach. These measurements were then averaged to obtain a predisturbed estimate for the width of the river. This procedure was also completed on the most recent aerial photo available for the location in order to compare the field-surveyed value for the width with an aerial estimate from the photograph. The predisturbed estimate was compared to the reach survey measurement to approximate the change in width from the historic condition to the present day condition. 67 %AW = Obs-Width(^)-Historic W i d t h ( m ) n o Q % Historic Width(w) Using the first and last width measurements as reference points, the down valley length was measured between these two points along with an estimate of the length of the channel between these two points. These two measurements were performed on both photos for redundancy purposes, but the values from the most current photo were used to estimate a value for the reach sinuosity. 3.3 Flood Indices Large or "catastrophic" floods may control the geomorphic condition of some fluvial systems. One possible index of potential effects of flood regime of the system is with Beard's (Baker 1977) flash flood magnitude index (FFMI) (Section 2.2.1.3.1). This index was calculated for each of the case studies based on the data available in the Environment of Canada's Hydat and the British Columbia Streamflow Inventory (Coulson and Obedkoff 1998; Environment Canada 2000). For the case studies where specific gauging data was not available, a similar regional analysis was used that was performed for estimating the mean annual flood for the location. But, an average FFMI was calculated based on the data of each regional gauge FFMI. Before the FFMI was calculated for each gauge, the discharge value was configured based on the drainage area of the site of study according to the Eaton et. al. (2002) method that was used for the mean annual flood calculations. 68 Similar to the FFMI, it has been suggested that historic "large" floods may control the geomorphic condition of fluvial systems (section 2.2.1.3.2) and the influence of these floods may be indexed by the ratio of the peak discharge to the mean annual discharge (Stevens et al. 1975). In order to examine any relationship between width change and the occurrence of large floods, the ratio of instantaneous peak discharge and the mean annual flood was calculated for the case studies for which gauge data is available (Coulson and Obedkoff 1998; Environment Canada 2000). Since this relationship appears to be highly related to each particular drainage basin, the regional analysis approach was not used to estimate the peak floods for the sites where no gauge data is available. 69 CHAPTER 4 4 CASE STUDIES The goal of the study is to formulate and test a parameter that describes the sensitivity of the single-thread gravel-bed river with noncohesive banks to disturbance by comparing undisturbed and disturbed channel widths. The sensitivity index is evaluated using case studies of disturbed and channel reaches with intact riparian vegetation situated in the province of British Columbia (Figure 4-1). The data sets for the following case studies are collected during a field site research campaign of the summer of 2002 and analytical analysis during the following fall. The purpose of these case studies is to give some insight into the potential function of the sensitivity index for use in analysing disturbed or potentially disturbed alluvial gravel-bed river reaches within British Columbia. Scale / Echel le 0 100 200 T a h s i s Rivei Figure 4-1 Case study sampling locations The 12 case studies are briefly described below. 70 4.1 Bonaparte River The Bonaparte River is a tributary to the Thompson River and its confluence is near Ashcroft, British Columbia. The Bonaparte River watershed covers roughly 5300 km2 of the Fraser (Interior) Plateau in central B.C. consisting of the Arrowstone Hills, Thompson Plateau and Green Timber Plateau (Figure 4-3). Consequently, the Bonaparte River drains many of the plateau lakes, including Bonaparte Lake. It is located in hydrologic zone 12A Southern Interior of the British Columbia streamflow inventory and consists of several biogeoclimatic zones, including the Interior Douglas Fir, Mountain Hemlock, Bunchgrass, and Montane Spruce zones. The temperatures in the region can exceed 35°C in the summer and the average winter daytime temperature is -5°C. The reach used for this study is situated just north of the village of Cache Creek where Highway 97 crosses the river, which is about 18 km northeast of the confluence with the Thompson River. The majority of the valley land use includes irrigated agriculture, cattle grazing, and forestry. Due to these land uses, substantial amounts of the riparian vegetation has been cleared or damaged. Potentially due to these disturbances, it has been concluded that the distribution of unvegetated gravel bars has substantially increased in the period between 1950 and 1995 based on an aerial photo analysis (M. Miles and Associates 1997a). According to the report by M. Miles and Ass., the most valuable restoration strategy for this region of the Bonaparte River would be to re-establish the riparian vegetation. This situation discloses the Bonaparte River to be an important candidate for a riparian vegetation sensitivity analysis as supported by this study. 71 4.1.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-2. The size distributions of the riffle and bank sediments are very similar, and D50 for the bed and bank surface material are both approximately 0.02 m. B o n a p a r t e R ive r S u r f a c e G r a i n Size D i s t r i b u t i o n s 1 10 100 1000 D (mm) Figure 4-2 Bonaparte River cumulative grain size distribution 4.1.2 Discharge The drainage area for the study reach on the Bonaparte River is approximately 5070 km2 at an average elevation of 475.5 m (Figure 4-3). There are two streamflow inventory gauges available for the Bonaparte River: Bonaparte River below Cache Creek with a drainage area of 5020 km2 and Bonaparte River near Bridge Lake with a drainage area of 650 km . There is a minor discrepancy between the reach drainage area and the gauge 72 area since the reach outlet is located upstream from the gauge situated just downstream of Cache Creek. Despite this minor error either in the streamflow inventory or the watershed delineation of this study, the gauge below Cache Creek is the data set that was used for the analysis. Figure 4-3 Bonaparte River location map and watershed delineation (921 Ashcroft and 92P Bonaparte Lake, 1:250,000) 73 4.1.2.1 Regional Analysis The flow regime of the Bonaparte River below Cache Creek is a snowmelt-dominated flood with a spring freshet that usually occurs between the second half of April to the first half of July. A substantial flood occurred on the system in 1990 with an estimated 50-year event and an instantaneous peak discharge of 88.7 m3/s, which could have caused significant bank erosion along the disturbed section of river (M. Miles and Associates 1997a). Based on the hydrologic analysis, the mean annual flood (Q) for the study reach is approximately 31.8 m3/s. 4.1.3 Slope The first half of the hydraulic survey on the Bonaparte River consisted of a longitudinal profile using a rod and level where a point was measured at 10 m increments for a total survey length of 300m. The average channel slope estimated from the longitudinal profile is approximately 0.0029 m/m for the reach (Figure 4-4). Bonaparte River Longi tudinal Channel Profile * 7.5 J 1 1 1 • , , 1 0 50 100 150 200 250 300 350 Relative Distance Downstream (m) Figure 4-4 Bonaparte River longitudinal profile 74 4.1.4 Width The second half of the hydraulic survey included cross-sectional profiles taken on three out of four of the riffles dissected by the longitudinal profile. The third riffle of the reach was not surveyed because substantial amounts of riprap had been placed on the right bank (looking downstream) to prevent the river from potentially eroding an access road, fence line, and small corral. With the use of the remaining cross-sections, the average unvegetated bankfull width was estimated to be 29 m (Figure 4-5). Comparatively, the photograph analysis measurements concluded a bankfull width of 13 m in 1959. The measured width from the 1959 photograph of 13 m is considered the historic or undisturbed width for the Bonaparte Paver. These two widths suggest a change in width over the time period from 1959 to 2002 of 127 % (Appendix B Figure BI & Appendix C). 75 4.2 Coldwater River The Coldwater River starts with its headwaters on the North side of the Coquihalla Summit Recreation Area of Highway 5 (Coquihalla Highway) at an elevation of approximately 1120 m and 45 km north of Hope. It flows north, roughly parallel to highway 5, through Merritt, British Columbia, where it is a tributary of the Nicola River just downstream from Nicola Lake. The Coldwater watershed constitutes roughly 936 km2 of the Nicola Forest on the east side of the Cascade Mountains in hydrologic zone 12B Southern Interior of the British Columbia Streamflow Inventory. At the headwaters of the Coldwater River, the watershed is comprised of the Engelmann Spruce-Subalpine Fir and Montane Spruce biogeoclimatic zones, but it quickly traverses to the dry Interior Douglas-fir zone. As the river flows north it enters the Ponderosa Pine zone and finally concludes around Merritt in the hot and dry Bunchgrass zone. There were three reaches of the Coldwater River used for this study and they are: • Reach 1 begins at kilometre 27 from the mouth of the Coldwater River and extends downstream along the old highway. • Reach 2 begins at kilometre 70.5 just downstream from a split in the channel and upstream of the cross-channel right of way • Reach 3 begins at kilometre 76 where the river bends near the old access road off of the Coquihalla highway 77 Reach 1 of the analysis is situated in the valley flat of the Coldwater River and is surrounded by agricultural and cattle grazing land uses (Figure 4-7). These land uses have resulted in extensive floodplain clearing, grazed woody riparian vegetation and trampled stream banks, which has led to destabilized and eroding banks (M. Miles and Associates 1997b). This reach is potentially a good candidate for the sensitivity of a gravel-bed river to riparian disturbances. The second reach is located upstream of Vought Creek and is accessible through a gate in the fence line paralleling the Coquihalla highway. The entire reach is upstream of the right of way, so should not be affected by anthropological straightening and bank hardening occurring at this location. The 1938 fire, which occurred in the upper watershed, potentially could have affected the riparian vegetation and could be the reason for the slight instabilities and dual channels in this section of the river. The riparian vegetation appeared to be intact and healthy along this section of the river, but it could be still recovering from the logging and fire that occurred earlier in the last century. The third and final reach is located approximately 5 km upstream from reach 2, just downstream from the upper Coquihalla highway crossing. Similar to reach 2, reach 3 is located within an unaltered section of river and may have been disturbed by the 1938 fire. On the other hand, reach 3 does not appear to show any signs of instability or widening that could be caused by a riparian disturbance. Therefore, a good contrast is presented between the unstable condition of reach 1 and the intact and stable condition of reach 3. 78 4.2.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-6. The size distributions of the riffle sediments are coarser than the bank sediments for all three reaches. The D50 for the bed sediments are 0.065 m, 0.037 m, and 0.03 m, as compared to the bank surface materials that are approximately 0.033 m, 0.017m, and 0.017m for reach 1, 2, and 3, respectively. 4.2.2 Discharge The major hydro logic floods on the Coldwater System appear to be the major floods due to snowmelt in the spring with minor rain on snow events occurring in the early winter. The drainage area for the study reaches 1, 2, and 3 on the Coldwater River are 9 9 9 approximately 735 km , 91.5 km , and 81.6 km at average elevations of 716.3 m, 1036.3 m, and 1082 m (Figure 4-7). There are two streamflow inventory gauges from hydrologic zone 12B Southern Interior available for the Coldwater River: Coldwater River at Merritt (8LG010) with a drainage area of 914 km and Coldwater River near Brookmere (8LG048) with a drainage area of 311 km . 4.2.2.1 Regional Analysis It seems appropriate to relate reach 1 with the discharge relationship from the gauge of the Coldwater River at Merritt because of their similarity in drainage area and down valley location and orientation. Similarly, it seems appropriate to relate the discharge at reach 2 and reach 3 to the gauge of the Coldwater River near Brookmere for the same reasons. Therefore, using the methodology for transferring gauged data to an ungauged 79 location on the same system discussed previously, the mean annual discharges reaches 1, 2, and 3 of the Coldwater River are 77 m3/s, 31 m3/s, and 28 m3/s. C o l d w a t e r R i v e r R e a c h 1 S u r f a c e G r a i n S i z e D i s t r i b u t i o n s • B a n k Riffle 1 10 100 1000 D ( m m ) C o l d w a t e r R i v e r R e a c h 2 S u r f a c e G r a i n S i z e D i s t r i b u t i o n Bank Riffle 1000 D (m m ) C o l d w a t e r R i v e r R e a c h 3 S u r f a c e G r a i n S i z e D i s t r i b u t i o n - B a n k •Riffle 1 10 100 1000 D (m m ) Figure 4-6 Coldwater River cumulative grain size distributions Figure 4-7 Coldwater River location map and watershed delineation (92H Hope and 921 Ashcroft, 1:250,000) Note north is down. 81 4.2.3 Slope The first half of the rod and level hydraulic survey on the three individual reaches of the Coldwater River consisted of a longitudinal profile, where a point was measured at 10 m increments for total survey lengths of 520 m, 350 m, and 300 m for reaches 1,2, and 3, respectively. The average channel slopes estimated from the longitudinal profiles are approximately 0.0049 m/m, 0.0033 m/m, and 0.0047 for the three reaches (Figure 4-8). Based on these estimations there appears to be a substantial change in slope between reach 2 and reach 3. This situation potentially could be due to the relatively short distance that the reaches cover. According to M. Miles and Ass. (1997b), the valley slope for the Coldwater River watershed upstream of Vought Creek was estimated to be 0.01 m/m from a longitudinal profile. Due to the fact that the current analysis is based on the reach scale, the channel slopes from the measurements will be used in the analysis. 4.2.4 Width Reach 1 of the Coldwater River survey included the measurements of four cross-sections. The average width from these four riffle cross-sections was 62 m (Figure 4-9). This value is very similar to the value obtained from the 2000 aerial photograph analysis of 64 m. The predisturbed value is assumed to be 26 m according to the measurements from the 1953 aerial photograph. Therefore, the percent change in the width from the 1953 aerial photo observation and the 2002 reach survey is roughly 135% (Appendix B Figures B3 and B4 & Appendix C). 82 Coldwater River Reach 1 Longitudinal Channel Profile y = -0.0049X + 8.23 Relative Distance Downstream (m) Coldwater River Reach 2 Longitudianl Channel Profile y = -0.0033x + 9.02 5.5 \ 0 50 100 150 200 250 300 350 400 450 Relative Distance Downstream (m) Coldwater River Reach 3 Longitudinal Channel Profile y = -0.0047X + 9.30 ui 4> > 6 5 ra * 5.5 -I , . , , , , , 0 50 100 150 200 250 300 350 400 450 Relative Distance Downstream (m) Figure 4-8 Coldwater River longitudinal profiles Similarly, the survey of reach 2 of the Coldwater survey included four riffle cross-sections. The average width that was calculated from this survey was 34 m (Figure 4-10). The values obtained from the 2000 and 1953 aerial photographs were 34 m and 28 m. Between the 2002 measurements and the 1953 photographic measurements, the percent change in width is approximately 23% (Appendix B Figures B5 and B6. Due to the close proximity of riffle locations along the profile of reach 3 of the Coldwater River survey, five riffle cross-sections were measured. The average value obtained from these five measurements was 16 m (Figure 4-11), as compared to the slightly larger value of 18 m obtained from the 2000 aerial photo measurements. The historic value from the 1953 photo was 15 m, which is similar to the other two values (Appendix B Figures B7 and B8 & Appendix C). These differences are possibly due to minor measurement errors in the aerial photo analysis or survey. In order to be consistent in the methods, the comparison in widths was between the historic photo and current observation. The percent change in width from the 1953 photograph measurement and the 2002 observation was 4%. 84 c q o "> — , S 5 O CD O o> — O m II »9 fl o • p* ? fl ii (ui) } u 6 j 3 H (ui) mBiaH a 7 3 fl -o « fl - f l 03 - S « .o S- 4> U *< 5 ^ >- 5 V o g "S "3 2s • A TT O <U '-{3 i - u S v CJD * <u _ <? 3 to •& o £ O m II (ui) ju6jaH (ui) mBjeH 87 e.) 2 1.75 I 1 5 —•—Cross-section ~ ~ m bankfull 0.25 I 0+-0 5 10 15 20 Width (m) Figure 4-11 Coldwater River reach 3 channel cross-section e.) section 5 4.3 Deadman River The Deadman River watershed is roughly a 1486 km2 watershed situated directly east of the Bonaparte River watershed within the Thompson and Fraser Plateau areas of the Interior Physiographic Region (M. Miles and Associates 1995). The Deadman River discharges at its confluence on the north side of the Thompson River, approximately 4 kilometres west of Kamloops Lake. The watershed consists mainly of three biogeoclimatic zones. The valley bottom and valley walls lie within the Bunchgrass and Ponderosa Pine zones with the majority of the remaining land in the watershed lying within the Interior Douglas-fir zone. The land use within the region persists mainly of agriculture, grazing, and limited forestry practices. The valley bottoms are irrigated in the summer to grow hay and are used for cattle grazing pastures during the winter 88 months. Due to these activities, large portions of the valley flat have been cleared of the large trees and woody riparian vegetation. According to a report completed by M. Miles and Ass. (1995), 43 of the studied 64 kilometres of the Deadman River are unvegetated or poorly vegetated. Also, 16 of the 20 kilometres that were classified as actively eroding belong to the category of unvegetated stream banks. In conjunction with the eroding stream banks, a substantial flood occurred in the Deadman River system on June 14, 1990 with an instantaneous peak flow of 64.4 m3/s. It has been reported that this event was approximately a 30-year flood (M. Miles and Associates 1995). On the other hand, this same report states that the laterally unstable morphology on the Deadman River predates this substantial flood that occurred in 1990. This leads to the analysis of applying the sensitivity criteria to a river that has been cleared of substantial amounts of riparian vegetation and displays a tendency towards a laterally unstable morphology. Two reaches were studied on the Deadman River. Both of the reaches are situated within the Skeetchestn Indian Reserve located in the lower river valley. These two reaches were selected primarily due to access problems within the privately owned land of the Deadman Valley. Don Ignace of the Skeetchestn Indian Band graciously allowed access to these sections of the river. The first reach terminates about 11 kilometres upstream from the Deadman River with the Thompson River. The second reach terminates roughly 2.5 kilometres upstream from the end of reach 2 (Figure 4-13). Unfortunately, while in the field the downstream reach was labelled as reach 2 and upstream reach was labelled 89 as reach 1. This is contrary to the other situations within the study where more than one reach was studied on a single river or stream. In order to be consistent with the infield survey and sediment sampling, the numbering of the reaches were not changed to be compatible with the remainder of the study. These reaches were situated approximately 1 km downstream from an eroding river bend and riparian vegetation restoration project. It is unknown what sort if any affects this work may have had on the river downstream from its location. 4.3.1 Grain S ize Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-12. The size distributions of the riffle sediments are obviously coarser than the bank sediments for reach 1, but similar for reach 2. The D50 for the bed sediments are 0.03 m for reach 1 and 0.035 m for reach 2, as compared to the bank surface materials that are approximately 0.05 m and 0.04m for reach 1 and 2, respectively. 4.3.2 Discharge The Deadman River watershed exists within hydrologic zone 12B, Southern Interior, of the British Columbia streamflow inventory. One gauge is located on the Deadman River, Deadman River above Criss Creek (8LF027), and it has a drainage area of 870 kmz and a mean annual flood of 15 m3/s. Unfortunately, this gauge is located upstream from the two study reaches and the confluence with Criss Creek, but it is the only available stream flow data for the system. Therefore, this gauge was used to translate the gauged mean 90 annual flood to the mean annual floods for the two downstream reaches. Using the drainage areas of the two reaches, 1432 km for reach 2 and 1412 km for reach 1, the mean annual discharges are both approximately 22 m3/s (Figure 4-13). It seems reasonable that the discharge for the two reaches is comparably similar due to the small differences in their drainage areas. D e a d m a n R i v e r R e a c h 1 S u r f a c e G r a i n S i z e D i s t r i b u t i o n • B a n k •Riffle 1 000 D (m m ) D e a d m a n R i v e r R e a c h 2 S u r f a c e G r a i n S i z e D i s t r i b u t i o n • B a n k •Riffle 1 10 100 1000 D ( m m ) Figure 4-12 Deadman River cumulative grain size distributions 91 Figure 4-13 Deadman River location map and watershed delineation (921 Ashcroft and 92P Bonaparte Lake, 1:250,000) 92 4.3.3 Slope The longitudinal survey of reach 1 constituted 310 m of channel length with an average channel slope of 0.0027 m/m (Figure 4-14). The longitudinal survey of reach 2 involved 390 m of channel with an average channel slope of 0.0047 m/m. The discrepancies between the two closely situated reaches and their channel slopes is thought to be due to the relatively small extent of channel surveyed for each reach as compared to the length of the valley. Based on this assumption, the estimated values for the channel slopes are assumed to be valid and different. Deadman River Reach 1 Longitudinal Channel Profile y = -0.0027X + 8.53 & 6 J . , . , • • , 1 0 50 100 150 200 250 300 350 400 Relative Distance Downst ream (m) Figure 4-14 Deadman River longitudinal profiles 93 4.3.4 Width The cross-sectional surveys for reach 1 and reach 2 included four and five cross-sections, respectively. The average unvegetated bankfull width for the two reaches was 30 m for reach 1 (Figure 4-15) and 37 m for reach 2 (Figure 4-17). The historic widths according to the 1959 aerial photograph were 17 m for reach 1 and 14 m for reach 2 (Appendix B Figures B9). Therefore, the calculated change in widths for reach 1 was 76% and for reach 2 was 164%. The change in widths appear fairly reasonable because it has been reported that the Deadman River channel has widened substantially with average widths of 9-10 m in 1949 and average widths of 30-45 m in 1992 (M. Miles and Associates 1995). 94 •8 O co (ui) ;qB!3H (ui) m6|aH d _o w _ • c o i. u 5 u a -g « 1-1 f*r JS c u o 53 « fa S fa <» <u ,—, 8 5 5 5 8 .2 a *0 « cs «o <w /-«> Q -o o a> v— fa U 3D > » Si- « ? 3 2 g O CD f l E. (ui) jqBjaH (ui) ;MB|3H 95 (ui) ) u 6 j 3 H •a (iu) )MB!3H a • 8 "•*•> ii T f • fl 5 0 © !*> • « O tm O U Q> U /—s O ^ fl -fl fl u o es CD y ii 2 <s fl o u fl et s SJ CO Q -= fl s E. o O CO o (ui) }u6jaH (ui) }MB|3H 96 e.) — • — Cross-section ^ — B a n k f u l l 0 5 10 15 20 25 30 35 40 45 50 Width (m) Figure 4-16 Deadman River reach 2 channel cross-section e) section 5 4.4 Elk River The Elk River is located within the confines of the Strathcona Provincial Park on north central Vancouver Island. The approximately 260 km watershed is located within hydrologic zone 17, Vancouver Island of the streamflow inventory and contains 3 of the 14 biogeoclimatic zones identified within the province. These zones include the Coastal Western Hemlock Zone, generally in the valley bottoms up to 900 m, a Mountain Hemlock Zone from 900 m up to 1500 m, and a limited amount of Alpine Tundra Zone above this elevation (Meidinger and Pojar 1991). The terminus of the Elk River is Upper Campbell Lake primarily situated outside the park boundaries. On June 24, 1946 a portion of Mt. Colonel Foster broke away and deposited in the lake below, which is now called Landslide Lake. This large and violent deposition of material within the lake displaced a large volume of water that travelled down the Elk River valley. This instantaneous flood eroded vegetation and material up to 8 kilometres downstream from the lake and moved sizeable quantities of sediment up to 13.5 kilometres downstream (M. Miles and Associates 1998). Along with the naturally occurring landslide, there are numerous anthropogenic disturbances that have transpired within the watershed. Downstream of the confluence of Drum Creek and the Elk River, widespread logging of the valley flat transpired in the 1940's (M. Miles and Associates 1998). These activities are fairly apparent when viewing the aerial photos from 1930 and 1957. In 1957-58, the Strathcona Dam and the Heber River and Crest Creek diversions were completed by the BC Power Commission (predecessor of BC Hydro) in order to create the Upper Campbell Lake Reservoir and increase the storage for the Campbell River hydro-electric system (M. Miles and Associates 1998; Nagtegaal et al. 2000). Therefore, there was an intensive riparian disturbance that occurred in the 1940's and solely persisted until 1957 when increased flows were added to the system. Due to the complicated effects of a landslide induced flood wave and induced diversion flows, two reaches were studied on the Elk River (Figure 4-18). The first reach is located approximately 7 kilometres upstream from the confluence with the Upper Campbell reservoir. It is located within the historically logged valley section and it exists just downstream of the initial intersection of highway 28 and the cross-valley logging road. During the construction of the highway, the upstream channel was realigned in order to bypass the need for the construction of two highway bridges. Therefore, the channel 98 realignment and downstream movement of the landslide-induced sediment could contribute increased sediment loads to this reach of the river. The second reach is located approximately 8-9 kilometres downstream of Landslide Lake. Hiking approximately 2-3 km upstream on the Elk River Trail accesses this reach where the river diverges from the course of highway 28. This reach is situated on the downstream end of the landslide influences and is outside of the logged portion of the watershed. 4.4.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-17. The size distributions of the riffle sediments are coarser than the bank sediments for reach 1 and reach 2. The D50 for the bank surface materials are 0.025 m for reach 1 and 0.015 m for reach 2, as compared to the bed sediment that are approximately 0.05 m and 0.062 m for reach 1 and 2, respectively. 4.4.2 Discharge According to a report completed by M. Miles & Ass. (1998), BC Hydro collects data from a gauging station that is not part of the Water Survey of Canada (WSC) records. Unfortunately, the data from this gauge was not obtainable at the time of the study. A single gauge does exist on the Elk River with 9 years of record available in the HYDAT 2000 data set. The gauge is the Elk River above Campbell Lake (8HD018) with a drainage area of 132 km , which includes the drainage area for the Crest Creek diversion, and a mean annual flood of 123 m /s (Campo pers. comm.). This gauge is located just 99 upstream from the location of reach 1 within the confining bedrock canyon. Using this short data set to estimate the mean annual floods for the two reaches, the discharge values are 112 m /s for reach 1 and 61 m /s for reach 2. In order to acquire the correct discharge for reach 1, the drainage area for the Crest Creek diversion of 17 km must be included within the reach 1 's drainage area. This adjustment results in a discharge of 124 m /s for -y reach 1 of the Elk River with a drainage area of 134 km" (Figure 4-18). E l k R i v e r R e a c h 1 S u r f a c e G r a i n S i z e D i s t r i b u t i o n s • B a n k Riffle D (m m ) Elk River Reach 2 Surface Grain Size Distributions • B a n k Riff le D (mm) Figure 4-17 Elk River cumulative grain size distributions 100 50°00 Reach 2 outlet Reach 1 outlet I 5 k n Figure 4-18 Elk River location map and watershed delineation (92F Port Alberni, 1:250,000) 101 4.4.3 Slope The longitudinal profiles for both study reaches of the Elk River were performed using the total station and prism. The reach lengths for the two reaches were 473 m for reach 1 and 317 m for reach 2. The average channel slopes that were estimated for reach 1 and reach 2 are 0.0063 m/m and 0.0104 m/m, respectively (Figure 4-19). These values seem reasonable because the reach within the valley flat section of the river downstream from the Drum Creek confluence is shallower than the reach located in the steep valley wall section of the river. For comparison purposes, the average channel slope reported by M. Miles and Ass. (1998) was 0.008 m/m, which falls midway between the two values obtained from the reach surveys. Elk River Reach 1 Longitudinal Channel Profile y = -0.0063X + 4.30 R2 = 0.90 Relative Distance Downstream (m) Elk River Reach 2 Longitudinal Channel Profile y _ _0 oio4x + R2 = 0.90 5 i 0 50 100 150 200 250 300 350 400 450 500 Relative Distance Downstream (m) Figure 4-19 Elk River longitudinal profiles 102 4.4.4 Width The cross-sectional surveys completed for the Elk River system were conducted using the total station. There were four cross-sections surveyed for both reach 1 and reach 2 (Figures 4-20 and 4-21). Based on the survey data, the average widths for the two reaches were 88 m for the heavily logged reach 1 and 27 m for reach 2. The estimated change in width according to the 1930 photographs from the National Air Photo Library (NAPL) in Ottawa was approximately a 274% change for reach 1 and a 49% change for reach 2 (Appendix B Figures BI 1 and B13 & Appendix C). 103 (UJ) I U B I O H (ui) luBjeH 104 u CP in CN in T - in o C N o (LU) WB|9H c o o cu in C N in T - in o esi o (LU) 1U&8H t l (ui) } u 6 ! 8 H a O TT a o u CU • u eu ~ CU fl fl 03 - f l o -a A 03 fl w CU c/> U 03 C U I . U eu £ fl a. I T3 • CU s-s _ fl O 4.5 Eve River The Eve River is the major tributary to the Adam River located directly north of Schoen Lake Provincial Park. The Adam River discharges approximately midway between Telegraph Cove and Kelsey Bay into Johnstone Strait on the north central coast of Vancouver Island. The Eve River watershed consists of approximately 260 km2 of drainage area mainly comprised of the Coastal Western Hemlock biogeoclimatic zone in the lower elevations and the Mountain Hemlock zone forming the higher elevation areas. The Eve River and surrounding watersheds have experienced heavy logging activities from the 1960's until the present. These logging activities have included the removal of the vast majority of valley flat vegetation, including the riparian zone up to the banks. This removal of streamside vegetation could have caused the unstable condition and could have assisted in the formation of substantial amounts of exposed gravel bars apparent in the current photo of the Eve River. The reach that was studied for this case study is located directly downstream from the junction of the logging road situated on the eastside of the river and highway 19 (Figure 4-23). Access to this reach was gained by parking at the campsite established at an abandoned landing or spur road. The objective of the study of the Eve River was to observe whether or not the instabilities in this section of the system are caused by the logging activities or are inherent properties of this reach of the river. The Eve River appears to be a stable single-channel system upstream of the confluence with Kunnum Creek. A reach upstream of this confluence could not be studied due to poor weather 106 conditions that resulted in time and accessibility constraints. Therefore, the goal of the study was to obtain the necessary characteristics of the reach of the Eve River that appears to be in a disturbed and unstable condition. 4.5.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-12. The size distributions of the riffle sediments are obviously coarser than the bank sediments for reach 1. The D50 for the bed sediments are 0.015 m for reach 1, as compared to the bank surface materials that are approximately 0.034 m. This procedure was performed following a substantial rainstorm that left the majority of the river floodplain inundated with floodwaters. Due to the unstable and braided nature of the reach, it was difficult to find a bank that represented the dominant bank for the bankfull limitations on the channel. Therefore, it is possible that the bank sample was performed on gravel deposits that does not ideally represent the historic bank sediment characteristics. Eve River Surface Grain Size Distributions 1 10 100 1000 D (mm) Figure 4-22 Eve River cumulative grain size distributions 107 Figure 4-23 Eve River location map and watershed delineation (92L Alert Bay, 1:250,000) 108 4.5.2 Discharge The Eve River watershed is located within hydrologic zone 17 Vancouver Island of the B.C. streamflow inventory and the watershed of the study reach has a drainage area of approximately 146 km (Figure 4-23). Unfortunately, there are no gauges located on either the Eve River or Adam River systems. Therefore, a regional analysis was required to obtain the mean annual instantaneous discharge for the study reach of the Eve River. 4.5.2.1 Regional Analysis The gauges used in the regional analysis were available gauges in hydrologic zone 17 that were similar in drainage area, location, and drainage aspect as the Eve River watershed. The three gauges were the Kokish River below Bonanza River (8HF003), Salmon River above Memekay River (8HD007), and Tsitika River below Catherine Creek (8HF004). All three of these watersheds appear to have a north west drainage compared to the north east drainage pattern for the Eve River and all three of their drainage areas (269-437 kmz) are larger than the drainage area for the Eve River reach. On the other hand, all three of the gauged watersheds are located on the north coast of Vancouver Island within a radius of approximately 50 kilometres and terminate in Johnstone Strait, as does the Eve River and Adam River system. Since these three watersheds are the only gauged watersheds available within the physiographic region and have reasonably similar characteristics, it was assumed that their hydrologic history would closely represent the hydrologic characteristics of the Eve River. 109 Using an average of three krigging coefficients for the gauged sites, an estimate of 3.6 is the estimate for the k coefficient for the Eve River (Table 4-1). This led to an instantaneous mean annual flood for the Eve River reach to be estimated at 152 m3/s. Gauged River Location Q (m3/s) Area (km2) k (calculated) Kokish R. Below Bonanza 131 269 2.0 Salmon R. Above Memekay 325 437 3.4 Tsitika R. Below Catherine Cr. 450 359 5.5 Average k 3.6 Table 4-1 Eve River regional regression table 4.5.3 Slope The longitudinal profile of the Eve River reach performed with the total station included 436 m of channel length and three riffle-pool sequences. The average slope computed from the profile was 0.0107 m/m (Figure 4-24). 110 Eve River L o n g i t u d i n a l C h a n n e l Prof i le 1 0 0 1 -i-9 9 4 -I 1 , , , 1 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 Relative Distance Downstream (m) Figure 4-24 Eve River longitudinal profile 4.5.4 Width The total station was used to transect a cross-section at the four separate riffles dissected with the longitudinal survey of the Eve River reach. These four cross-sections gave an average width of 69 m (Figure 4-25). An aerial photo analysis was performed on photos from 1953, 1981, and 1996, which supplied measurements for the average widths of the reach to be 48 m, 97 m, and 106 m, respectively (Appendix B Figures B15, B16, and B17 & Appendix C). After looking at the aerial photos, the field surveyed reach is too short to represent the width of the section. The surveyed reach appears to have widened and contracted between the three photos. Therefore, a longer section was used for the photo analysis and the 1953 and 1996 photos will be used for the width measurements. The estimated change in width from the predisturbed 1953 condition to the present disturbed condition (1996) is 123%. I l l 4.6 Salmon River The Salmon River watershed is approximately a 1300 km2 watershed that drains into the Johnstone Strait near Sayward, B.C on the north eastern coast of Vancouver Island (Figure 4-27). The majority of the Salmon River watershed consists of the Coastal Western Hemlock biogeoclimatic zone with higher elevations comprised of the Mountain Hemlock zone and acutely isolated areas of Alpine Tundra at the extreme elevations of the Vancouver Island Ranges and a small portion of the Strathcona Provincial Park. The major tributaries to the Salmon River system are Grilse Creek, Memekay River, and White River. The primary historic land use of the Salmon River valley was logging, which began sometime prior to the earliest available aerial photograph from 1946 (Appendix B Figure B18). As may be seen from this photograph, large portions of the valley flat and riparian stands were completely deforested. This potentially could have led to the apparent quantities of unvegetated coarse fluvial sediments and the eroding riparian areas exhibited by the historic photo (M. Miles and Associates 1998). The reach surveyed for the Salmon River is located downstream of the confluence with the Memekay River. The reach is situated along a historically deforested region and contains present day unvegetated riparian areas due to agricultural and grazing land uses. It was attempted to access a single-thread sinuous reach immediately upstream of the confluence with the White River. This reach was not studied due to the nature of the riverbanks. The banks consisted of a vertical cohesive clay layer with a thickness of approximately 3 m. Therefore, this reach of the Salmon River does not meet the criteria of the theory due to the controlling cohesiveness of the banks and potentially is a relatively thin single-thread channel due to these high cohesive banks. 4.6.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-26. The size distributions of the riffle sediments are coarser than the bank sediments for reach 1. The D50 for the bed sediments are 0.027 m for reach 1, as compared to the bank surface materials that are approximately 0.042 m. Salmon River Sur face Grain Size D is t r ibu t ions 1 10 100 1000 D (mm) F i g u r e 4-26 S a l m o n R i v e r c u m u l a t i v e g r a i n s i z e d i s t r i b u t i o n s 114 Figure 4-27 Salmon River location map and watershed delineation (92E Nootka Sound, 92F Port Alberni, 92K Bute Inlet and 92L Alert Bay, 1:250,000) 115 4.6.2 Discharge In 1956, a pipeline was constructed to divert flow from the Salmon River into Brewster Lake and subsequently into Campbell Lake Reservoir. The first water was removed from the Salmon River to be diverted to the Campbell River system in 1958 (British Columbia Power Commission 1958). According to the Campbell River Water Use Plan (WUP) from 2002, the average inflow to the Campbell River system from the Salmon River diversion is 11 m3/s, based on historical daily inflows from 1963-1999 (BC Hydro 2002). There are two gauges in operation on the Salmon River according to hydrologic zone 17 Vancouver Island of the B.C. Streamflow Inventory. The two gauges are the Salmon River near Sayward (8HD006) and the Salmon River above Memekay River (8HD007). Since the study reach is downstream of the confluence between the Memekay and Salmon rivers, the gauge near Sayward was used for the discharge analysis. As mentioned above, this gauge is downstream from the diversion into Brewster Lake. The uncorrected peak mean annual flow (Q) for the Salmon River at Sayward is 873 m3/s. If this flow were corrected for the diversion, the mean annual flow would approximately be 862 m3/s. These two values for the mean annual flow at Sayward lead to an estimate of Q for the study reach (Figure 4-27) to be 603 m /s and 596 m /s based on the corrected value, or a difference of approximately 1.2%. Due to the uncertainty of the diversion estimate and small difference in discharge values, the uncorrected value of 603 m /s was the value used for the analysis. 116 4.6.3 Slope A total length of 1075 m was surveyed on the Salmon River using the total station and prism. According to the longitudinal profile, the estimate for the channel slope was 0.0018 m/m (Figure 4-28). S a l m o n R i v e r L o n g i t u d i n a l C h a n n e l P r o f i l e 991 990 -I , , , , 1 0 200 400 600 800 1000 1200 R e l a t i v e D i s t a n c e D o w n s t r e a m (m ) Figure 4-28 Salmon River longitudinal profile 4.6.4 Width Four cross-sections were surveyed with the total station on the Salmon River reach (Figure 4-29). According to these cross-sections, the average width for the Salmon River is 195 m. The earliest available aerial photo for the Salmon River was taken in 1946 (Appendix B Figure B18). Consequently, no aerial photo exists that predates the logging disturbances along the study reach of the Salmon River. According to this photograph, the average width along the Salmon River after logging but before the small diversion was initiated was approximately 138 m. This leads to a change in width from 1946 to 2002 of approximately 41%. It is expected that the change from an unlogged condition 117 (pre-1946) to the current condition would be greater because cleared riparian areas appear to eroding in the 1946 photo (M. Miles and Associates 1998). Unfortunately, these impacts are not quantifiable so that the current approximation of 41% must be used for the case study. 118 4.7 Tahsis River The Tahsis River is located on the central west coast of Vancouver Island at the head of the Tahsis Inlet and the forestry community of Tahsis. The mouth of the river and the village may be reached by driving 70 km northwest of Gold River on a gravel road. The river drains an approximate 76.5 km2 area of the southwest aspect of the Vancouver Island Ranges and the Coastal Western Hemlock biogeoclimatic zone (Figure 4-31). Since the village of Tahsis was a floating logging camp in the 1940's and was established as a permanent village and mill town in the 1950's, logging impacts have occurred in the river valley upstream from the village. According to Kellerhals and Miles (1996), the earliest available aerial photo from 1956 slightly precedes valley bottom clearcut logging. According to this report, the channel used to be single thread, or anastamosing, and laterally stable. But, by 1980 the river had morphed into a multi-thread pattern with eroding banks, wider unvegetated channel width, and numerous unvegetated gravel bars (Kellerhals and Miles 1996). Initially the reach located in the upper valley flat where extensive channel widening has occurred was the desired reach of study. Due to the deactivation of the road upstream from the Tahsis village dump, this section of the river was inaccessible to be surveyed. Therefore, the reach that was studied during the 2002 field trip was the reach just upstream from the dump. This reach appears to have been disturbed during the logging in the 1950's, but appears to be slightly confined by the valley walls and a rock canyon just downstream from the end of the surveyed reach. 120 4.7.1 Grain Size Analysis Pebble counts were performed on selected riffles and sections of the banks. The grain size distribution curves are shown in Figure 4-30. The size distributions of the riffle sediments are coarser than the bank sediments for reach 1. The D50 for the bed sediments are 0.018 m for reach 1, as compared to the bank surface materials that are approximately 0.041 m. Tahsis River Surface Grain Size Distributions D ( m m ) Figure 4-30 Tahsis River cumulative grain size distributions 121 1 km Figure 4-31 Tahsis River location map and watershed delineation (92E Nootka Sound and 92L Alert Bay, 1:250,000) 122 4.7.2 Discharge The Tahsis River is an ungauged river located in hydrologic zone 17 Vancouver Island of the B.C. streamflow inventory. A regional analysis was necessary to calculate the mean annual flood for the study reach on the Tahsis River, since no gauge data is available for the system. 4.7.2.1 Regional Analysis The Tahsis River is located on the west coast of Vancouver Island. In order to get an estimate of the mean annual flood coefficient for the Eaton et. al. (2002) method, several watersheds located on the central west coast of the island were used for a regression analysis. The two gauges that were used for the regression were the Gold River below Ucona River (8HC001) and the Zeballos River near Zeballos (8HE006). These two rivers discharge into the Pacific Ocean on the west coast of the island and have southern drainage patterns similar to the Tahsis River. The Gold River is slightly inland as compared to the Zeballos and Tahsis Rivers, which might affect the hydrologic similarity between the watershed characteristics. The Zeballos River has a greater flood flow regime than the Gold River, but since the flood regime is unknown for the Tahsis River it would be hard to assume that the precipitation intensity for the Tahsis watershed is identical to the Zeballos watershed. This was the reason for using the two most similar gauged watersheds in the region as compared to relating the Tahsis River directly to the measured characteristics of the Zeballos River. 123 The result from the averaging between the two measured data points is a k coefficient of approximately 10.2 (Table 4-2). According to the map of k values for the mean annual flood in Eaton et al. (2002), the range for k values for this region of the province is from 4.001 to greater than 5.900. Since this is an uncapped range, it is hard to compare the results of the back calculation for the two watersheds that are beyond the 5.900 range. Therefore, it must be assumed that the regression formulation from the two watersheds that lie on either side of the study watershed in a east-west fashion are the best representation for an estimate of the watershed dynamics for the Tahsis River watershed. Using the k coefficient of 10.2, the mean annual discharge for the study reach drainage area of 51.6 km (Figure 4-31) gave a mean annual flood estimation of approximately 196 m3/s. Gauged River Location Q (m7s) Area (km2) k (calculated) Gold R. Below Ucona R 1519 1010 8.5 Zeballos R. Near Zeballos 584 181 11.8 Average k 10.2 Table 4-2 Tahsis River regional regression table 4.7.3 Slope The total station longitudinal profile for the Tahsis River covered a channel length of approximately 611m and three riffle-pool sequences. Unfortunately, a small portion of the first pool was not surveyed due to a substantial quantity of spawning sockeye salmon 124 inhabiting the pool. The channel slope estimated from this profile was 0.0039 m/m (Figure 4-32). T a h s i s R i v e r L o n g i t u d i n a l C h a n n e l P ro f i l e 9 9 9 . 5 9 9 8 . 5 s o Ui o > « -9 9 6 . 5 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0 Relative Distance Downstream (m) 6 0 0 7 0 0 Figure 4-32 Tahsis River longitudinal profile 4.7.4 Width As mentioned in section 4.7.3, the longitudinal profile surveyed three riffle-pool sequences or four riffles and three pools. The average channel width measured from the four cross-sections was 60 m (Figure 4-33). The disturbed width measured from the 1954 aerial photo was 29 m, which leads to an observed change in width of nearly 106% (Appendix B Figure B20 & Appendix C). 125 o ? 1 2 g O OD II (ui) J M B I B H ( u i ) J M B j O H a o u V en « cn a o « cn i cn S O O fa u fa cu a « a .2 £ c o cu ca fa *> M) - D fa * H o s u. CO O m II in O 00 II ( u i ) )u6|3|-| ( u i ) m B j a H 126 4.8 Case Studies on Existing Data The following case studies are individual river reaches obtained from pre-existing studies and theses (Figure 4-38). Supplementary data was analysed and collected for certain parameters in order to fit the requirements for application of the sensitivity index. Therefore, certain onsite characteristics are missing from the data sets depending on the methods that were used by the original investigators. Figure 4-34 Existing Case Study Locations 127 4.8.1 Big Horn Creek Big Horn Creek is located approximately 35 km south east of Fernie, British Columbia. It is a watershed of approximately 134 km situated in the south east corner of British Columbia within the MacDonald Range of the Rocky Mountain physiographic region and a tributary to the Wigwam River (Klohn-Crippen Consultants LTD. 1998) (Figure 4-35). Included within this region are four of the fourteen biogeoclimatic zones of British Columbia. These four zones include the Interior Cedar-Hemlock zone, Engelman Spruce zone, Ponderosa Pine zone, and at extreme elevations the Alpine Tundra zone. A few disturbances have occurred in the Big Horn Creek watershed during recent history. A large fire occurred downstream of reach 7 in the mid-1930's, which burned both valley walls and most of the valley flat vegetation (M. Miles and Associates 2001). The fire appears to have been intense due to little residual wood apparent in the 1948 photos and no indication of any salvage logging. Upstream of reach 8, logging has taken place due to various beetle epidemics. The headwater area was sanitary logged due to a spruce bark beetle epidemic in the mid-1960's through to the late-1970's. Also, sections of Lodgepole Pine were harvested throughout the late 1970's and early 1980's due to a pine beetle epidemic (M. Miles and Associates 2001). The reach of the Big Horn Creek that is of interest is reach 4. According to the report on channel stability, the fire destroyed approximately 100% of the hillside and riparian vegetation in the late 1930's (M. Miles and Associates 2001). It has been observed that the loose compaction of the valley sediments and potential loss of riparian root structures have resulted in low structural integrity of the stream banks (Klohn-Crippen Consultants 128 LTD. 1998). This reach currently is a severely aggraded reach with elevated bars, unstable stream banks, and multiple braided channels (Klohn-Crippen Consultants LTD. 1998). The unstable morphology of reach 4 may have been caused by the low gradient sedimentation basin confined by the bedrock geometry of the valley, the riparian disturbances that potentially affected the channel bank characteristics, or a combination of the two controlling factors. Therefore, reach 4 of Big Horn Creek appears to be a good candidate to apply the sensitivity analysis. 4.8.1.1 Grain Size Analysis A pebble count distribution for Reach 4 was available in the Bighorn Creek Integrated Watershed Restoration Project Report, Appendix III (Klohn-Crippen Consultants LTD. 1998). The specific techniques performed by Klohn-Crippen for this pebble count were not included within the report. Fortunately, personal communication with Greg Scarborough, formerly with Klohn-Crippen, supplied the details of the sampling procedure. The techniques used for this particular sample are similar to the techniques outlined in section 3.1.1.1. At reach 4 on Big Horn Creek, one sample was gathered at a riffle by blindly extending an index finger after each step. This technique was performed for 100 paces throughout the riffle, or a total sample of 100 stones. The cumulative grain size distribution available in the report shows an estimate of the median grain size D 5 0 for the bed surface to be on the order of 35 mm (Klohn-Crippen Consultants LTD. 1998). Since pebble counts were not performed on the bank surface of Bighorn Creek, the assumption must be made that the surface size distribution for the bed 129 and bank surface are equal. Therefore, a D50 value for the bank surface is assumed to be 0.035 m. Figure 4-35 Big Horn Creek location map and watershed delineation (map 82G Fernie, 1:250,000) 130 4.8.1.2 Discharge Big Horn Creek is expected to be a typical mountain watershed with peak flows occurring in the late spring and summer because of snowmelt dominated process (Klohn-Crippen Consultants LTD. 1998). Unfortunately, Big Horn Creek is an ungauged system and therefore does not have any official streamflow data. A regional analysis of surrounding gauged watersheds was required to estimate a mean annual discharge for the system. 4.8.1.2.1 Regional Analysis A regional analysis was performed on Big Horn Creek using the British Columbia Streamflow Inventory (BCSI) and HYDAT to obtain an estimate of the mean annual flood for the watershed (Coulson and Obedkoff 1998; Environment Canada 2000). According to the BCSI, Big Horn Creek is situated in hydrologic zone 14, Rocky Mountain Trench, of British Columbia. Using the available gauged watersheds' reported mean annual instantaneous floods and watershed areas, a graphical relationship was constructed for the hydrologic region within the relative vicinity of Big Horn Creek (closer than about 250 km). The three watersheds of extra importance due to their close proximity to Big Horn Creek are Cabin Creek, Couldrey Creek, and Howell Creek. All three of these watersheds are located within about 25 km of the Big Horn Creek watershed, but they are situated on an eastern aspect compared to the relative western aspect of the Big Horn Creek watershed. These relationships can be found in the Figure 4-36. 131 Watersheds within nearest Proximity of Big Horn Creek 3 0 2 5 2 0 ro 1 5 a 1 0 5 0 Watershed Area (km ) • C a b i n • C o u l d r e y A H o w e l l - B i g H o r n 1 0 0 1 5 0 2 0 0 Figure 4-36 Big Horn Creek regional analysis based on nearest watersheds As can be seen from Figure 4-36, a reasonable estimate for the mean annual flood for Big Horn Creek is about 20 m /s. This value was compared with the mean annual flood according to Figure 4.5 of the Bighorn Creek Integrated Watershed Restoration Project Report (Klohn-Crippen Consultants LTD. 1998). This figure presents a discharge of about 18 m3/s for a flood with a two-year return period. In comparison, the three watersheds mentioned earlier along with the western aspect watersheds of Line Creek and Grave Creek were used to calculate an average k value of 0.52 to be applied to the Eaton et al. (2002) method (Table 4-3). Using this method, the mean annual flood for Big Horn Creek is estimated at 19 m3/s. In order to be consistent with the rest of the analysis, this value will be used for a value of Q. 132 Gauged River Location Q (m3/s) _ Area (km ) k (calculated) Cabin Cr. Near mouth 24 93.3 0.80 Couldrey Cr. In lot 9380 18 118 0.50 Howell Cr. Above Cabin Cr. 20 145 0.48 Grave Cr. At mouth 8 75.8 0.31 Line Cr. At mouth 21 138 0.52 Average k ~ 0.52 Table 4-3 Big Horn Creek regional regression table 4.8.1.3 Slope The channel slope for the desired reach was required to apply the index of bank vegetation sensitivity model. The majority of reach 4 on Big Horn Creek was surveyed at numerous cross-sections and other interspersed points. The data from this survey was used to produce a longitudinal profile of the channel thalweg through reach 4 (Figure 4-37). The average slope of the channel thalweg through reach 4 was estimated to be • 0.0064 m/m. B i g h o r n C r e e k R e a c h 4 L o n g i t u d i n a l C h a n n e l P r o f i l e 1 1 50 -I , , , 1 1 0 1000 2000 3000 4000 5000 Rela t ive D is tance D o w n s t r e a m (m) Figure 4-37 Bighorn Creek longitudinal profile 133 4.8.1.4 Width An estimate of the historic width of the stream prior to affects of the disturbance was obtained using historical aerial photographs from 1962 (Appendix B Figure B22). Using this available resource, measurements of cross-sectional width were taken at discreet intervals of approximately 50 m throughout the length of the reach. A reach-averaged width of Big Horn Creek was calculated with these measurements to be on the order of 18 meters (Appendix C). Therefore, the percent change in width for Big Horn Creek from 1962 to 2000 is approximately 114%. 4.8.2 Narrowlake Creek Narrowlake Creek is a demonstration watershed located 80 km southeast of Prince George, British Columbia on the west side of the Cariboo Mountains and the Interior Plateau region. According to Wilson et. al. (2001), Narrowlake Creek is a fifth order watershed in the Sub-boreal Spruce biogeoclimatic zone. It flows from Narrow Lake and is a tributary to the Willow River (Figure 4-38). Mostly during the 1960s, 35% of the watershed's area was logged, including logging to the banks on 80% of the fish bearing reaches (Wilson, 2001). Therefore, since substantial data has been collected on this disturbed creek, it is a worthy case study to be applied to the sensitivity analysis. 4.8.2.1 Grain Size Analysis An employee of the Ministry of Water, Land, and Air Protection, Ray Pillipow, determined the median grain size for Narrowlake Creek. According to Pillipow, samples were taken for both the channel banks and channel bed (Pillipow pers. comm.). The technique used to collect the stones was a blind walk with the stone located under the toe 134 of the boot measured at intervals of two paces. The bed samples were selected from waded long runs and riffles of the river, but no pool sites. A downstream zigzag pattern was taken from bank to bank along several sections of the same reach, while the bank material was selected from exposed portions at the outside of bends (Pillipow pers. comm.). The value that was reported in the literature for the substrate, or streambed, D 5 0 is 0.058 m (Wilson, 2001). Information was provided that broke the sampling on treatment reach 3 (TR3) into upper, middle, and lower sections. Averaging these three values for the entire reach resulted in an estimate of the D50 for the bed to be 0.063 m and for the banks to be 0.052 m (Wilson pers. comm.). The value for the bank surface grain calibre is the value used in the following stability analysis. 135 1 kn Figure 4-38 Narrowlake Creek location map and watershed delineation (93H McBride and 93G Prince George, 1:250,000) 4.8.2.2 Discharge The Narrowlake Creek watershed is an ungauged watershed located within hydrologic zone 12A, Southern Interior, of BCSI. According to the published literature, the bankfull discharge for Narrowlake Creek at treatment reach 3 is 31.1 m3/s (Wilson 2001). Unfortunately, the methods used to obtain this discharge value were not described in 136 Wilson (2001). Therefore, a regional analysis was used to estimate the mean annual flood for Narrowlake Creek. 4.8.2.2.1 Regional Analysis The watershed drainage area for the entire watershed is 210 km2 based on the watershed delineation method discussed in section 3.2.1. According to Andrew Wilson (pers. comm.), the drainage area for the watershed of treatment reach 3 is slightly less at 185 km . As stated earlier, Narrowlake Creek is a tributary to the Willow River, this has a gauge at Willow River above Hay Creek (8KD006) (Figure 4-38). The drainage area and mean annual discharge at the gauge are 2810 km2 and 243 m3/s. If this data is used to calculate the k coefficient of 0.63, the resulting mean annual flood for Narrowlake Creek was approximately 32 m3/s. Even though Narrowlake Creek is a tributary to the Willow River, the hydrologic characteristics for the Willow River watershed may not represent the characteristics for Narrowlake Creek. Since Narrowlake Creek drains a lake (Narrow Lake) and it has a relatively small drainage area as compared to the Willow River, a larger regional analysis was used instead of just the gauge data from the Willow River. Similar to Narrowlake Creek, a relatively close watershed within hydrologic zone 12A McKinley Creek drains a lake in an eastern drainage pattern, but has a slightly larger drainage area of 426 km . Another gauged stream within the same region, approximate location, comparable drainage area of 539 km , and similar drainage pattern, is Moffat Creek. The final two gauges used for the regional analysis are gauges located on streams with similar drainage patterns and location, but they do not discharge from headwater or mid-drainage lakes. These two streams are Chuchinka Creek and the Little Swift River. 137 Therefore, the gauges and locations used for this analysis are the Willow River above Hay Creek (8KD006), McKinley Creek below outlet of McKinley Lake (8KH020), Moffat Creek near Horsefly (8kH019), Chuchinka Creek near the mouth (7EE009), and Little Swift River at the mouth (8KE024). The results from the regression analysis are a k coefficient of 0.59 and a discharge for TR3 of Narrowlake Creek of 30 m3/s (Figure 4-40). Auspiciously, this result is fairly similar to the value of 31.1 m3/s available in the literature. Figure 4-39 Narrowlake Creek regional regression chart 4.8.2.3 Slope The value that was reported for the channel slope of Narrowlake Creek was 0.007 m/m (Wilson 2001). According to personal communication with Andrew Wilson of the Ministry of WLAP, the section of disturbed banks on Narrowlake Creek is limited in its length (Wilson pers. comm.). 138 4.8.2.4 Width The following information is a summary from Wilson (2001). Wilson analysed six sets of aerial photos from 1946 to 1997 and collected measurements of bankfull width, channel length and channel area. Through this analysis, Wilson concluded that the channel had substantially widened from the pre-logging state in 1946 to the post-logging state in 1997. Also, the morphology of Narrowlake Creek had changed from a meandering channel to a braided channel. The two photos of most interest are the historic state in 1946 to the most current state available in the 1997 photo. The values for the widths measured by Wilson have been reported to be a pre-logging channel width of 29 m (1946) and a post-logging channel width of 58 m (1997) (Wilson 2001). These two values result in a change in width from the undisturbed to the present day disturbed condition of 100%, or a twofold change in width. 4.8.3 Slesse Creek Slesse Creek is a mountain stream with approximately 100 km2 of its total drainage area of 172 km2 located south of the Canadian-U.S. border in the pristine Mt. Baker Wilderness Area (Millar 2000) (Figure 4-40). It is a fourth-order stream and joins the Chilliwack River system 19 km upstream from Vedder Crossing in British Columbia (MacVicar 1999). The reach of Slesse Creek used in this study will be Reach D, which is an alluvial system with a wide floodplain, minimal bedrock influences, and no major inflows from tributaries (MacVicar 1999). 139 The following is a short description of the forest activities in the watershed available in MacVicar (1999) and summarized from Babikaiff and Ass. (1997). The majority of the • low elevation timber, including the riparian zone, was cut and an extensive road network was established on the watershed hill slopes from 1936 to 1956. It was typical to have a narrow buffer strip and numerous landings for cross-stream yarding. After 1956 and until 1973, logging progressed in an upstream and upslope fashion to the Canadian-U.S. border. This logging included removing the majority of the low elevation timber from Reach F to the border and mid to high elevation timber up to Reach F. Due to many factors, such as a the entire 60% of the upper watershed included in a pristine wilderness area and a high level of natural sediment supply, it is believed that the forest harvesting impacts on water sediment supply to Reach D to be limited. The major disturbance caused by the logging within the Slesse Creek watershed has been a removal and disruption of riparian vegetation. This disturbance to the riparian ecosystem has led to a substantial decrease in the bank stability enhanced by the root strength of the mature forest vegetation (MacVicar 1999). It has been identified that the major source of material situated within the channel boundary has been due to extensive bank erosion (Babikaiff and Associates Geoscience 1997). Therefore, it is believed that the planform transformation of Slesse Creek from a sinuous single-thread channel to a wide unstable braided channel is due to reduced bank strength following the riparian logging practices (Millar 2000; Millar 2001). It has been calibrated with the use of the Millar and Quick (1993) hydraulic geometry model that the historic d)' value is approximately 70°, while the current value is assumed to be 40° due to the logging and existence of sparse deciduous trees and shrubs constituting the riparian vegetation. 140 4.8.3.1 Grain Size Analysis The following is the sediment sampling techniques described by MacVicar (1999) for Slesse Creek. The bed particle distribution was measured using the standard pebble count technique described by Wolman (1954). Therefore, a sample of 100 stones was collected using the heel - to - toe technique in a section of the stream that is shallow enough to allow the surveyor to wade the section. The pebble count was taken in the single thread dike section of the channel with an estimated D50 of 0.133m. Based on a comparison of measured bed and bank particles, it was concluded that the D50 is equal to the Dsobank for Slesse Creek. 4.8.3.2 Discharge According to the streamflow inventory, the mean annual flood for Slesse Creek is 92 m /s. This value is for a drainage area of 166 km for where Slesse Creek discharges near Vedder Crossing (gauge 8MH056), but this is approximately 3.5 km downstream from the reach that was studied by MacVicar (1999) (Figure 4-40). In this study, MacVicar used a cumulative departure analysis to notice that there are two trends in the hydrologic flow regime on Slesse Creek. The interest for this study is to obtain the long term mean flow between the historical situation and the present situation, therefore the entire record from 1960 to 1995 will be used to establish the mean annual flood. 141 Figure 4-40 Slesse Creek location map and watershed delineation (Mount Baker reference map, 1:100,000) 142 Using the method discussed in section 3.2 for the drainage area and discharge, the gauged discharge was adjusted for the drainage area at the study reach. The drainage area for the watershed with an outlet at the study reach is approximately 157 km as compared to the 166 km2 for the gauge. This drainage area difference leads to a mean annual flood for the study reach to be approximately 88 m /s. 4.8.3.3 Slope The slope of the channel was surveyed using a longitudinal profile of the thalweg. The elevations were collected at repeating stream forms, such as riffle crests, in order to establish the average energy gradient of the channel. The surveyed channel slope from the 1997 dike section is 0.021 m/m. 4.8.3.4 Width In 1997, MacVicar surveyed the dike section of Slesse Creek using a minimum of 5 cross-sections at pool, riffle, and glide sections to establish an average bankfull width of 41m. Assuming the width of Slesse Creek has been anthropogenically altered due to the establishment of the dikes, the observations from the aerial photos are used for the width measurements. The techniques used by MacVicar for the aerial photo analysis to establish the reach-averaged width was to use a minimum of five measurements equally spaced along the reach. He believed that this technique would provide a representative distribution of the width throughout the reach length. The bankfull width measurements reported from the aerial photo analysis of MacVicar (1999) are 28 m, 21 m, and 145 m in 143 1936, 1973, and 1993, respectively. Therefore, the value of 28 m was used for the historical or predisturbed width, and the value of 145 m was used for the current or disturbed width. These widths suggest a change in width of approximately 418%. 4.8.4 West Kettle River The following discussion is a summary of information available in D'Aoust (1998). The West Kettle River is a 6th order stream with a watershed drainage area of roughly 1870 km2. The watershed consists of four biogeoclimatic zones including the upper elevations in the Sub-alpine and Montane Spruce zones and lower elevations in the Kettle Dry Mild Interior Douglas Fir and Interior Cedar-Hemlock zones. The land use in the watershed is variable with the lower watershed downstream of Carmi mostly privately owned and upstream of Carmi the river is primarily bounded by crown lands. The private lands located in the valley bottom have had an adverse affect on the riparian vegetation conditions. Typically trees have been removed up to the channel banks for grazing and forestry purposes. The majority of the riparian corridor is intact within the crown land property. The reach of the West Kettle River isolated for this study is located roughly midway between the communities of Westbridge and Beaverdell, which is approximately 20 km downstream from the town of Carmi. Since this reach is located within the disturbed portion of the watershed, it is a candidate for the potential effect of riparian disturbance influences on channel morphology. 144 4.8.4.1 Grain Size Analysis According to D'Aoust (1998), the substrate for the West Kettle River was sampled by the pebble count. It is assumed that the pebble count performed for this evaluation is similar to the pebble count technique discussed in section 3.1.1.1, except it was reported that only 50 randomly selected pebbles were used instead of 100. The values reported for the D50 and D90 of the substrate are 0.0775 m and 0.155 m, respectively. It must be assumed that the substrate on the bed and banks of the channel are similar in their calibre and frequency in order to assume that the D50bank is equal to the D5o of the substrate, or 0.0775 m. 4.8.4.2 Discharge The watershed drainage area to reach of interest on the West Kettle River is approximately 1520 km (D'Aoust 1998). A stream flow gauging station is located on the West Kettle River below Carmi Creek in hydrologic zone 12A Southern Interior. As reported in the British Columbia streamflow inventory, this gauge has an estimated drainage area of 1170 km" and a mean annual flood of 104 m /s. Since this gauge is located on the same stream and within a reasonable distance of the study reach, approximately 22 km upstream, the method of Eaton et. al. (2002) was used to calculate the mean annual flood (Q). Based on the gauged data, the krigging factor (k) for this location is 0.52. This leads to an estimate of Q for the study reach to be 127 m3/s. The value for the bankfull flow available in D'Aoust is 116 m3/s based on the adjustment method to translate a gauged discharge to an ungauged discharge stated in Harris (1986). 145 It is believed that the value of 127 m3/s for Q is more representative of the actual situation based on the variability included within the Eaton et al. (2002) technique. 4.8.4.3 Slope The following information is available in D'Aoust (1998). A longitudinal profile was surveyed along the centreline of the bankfull channel using stadia measurements with a rod and level. The slope reported for the study reach on the West Kettle River is 0.0028 m/m. After looking at the historic aerial photograph of the West Kettle River, it appears that the river has not changed very much in the last fifty years. Therefore the sinuosity of the river in 1951 is likely to be similar to the sinuosity of the river when the slope measurements were taken in 1997. 4.8.4.4 Width The bankfull width for West Kettle River was surveyed around an artificially constructed large woody debris (LWD) structure. According to D'Aoust, between 3 to 5 cross-sections were surveyed at each structure with additional sections surveyed upstream or downstream from the centreline section located at the structure. With the use of these cross-sections, it was estimated that the bankfull width of the West Kettle River in 1997 was approximately 44 m. The historic width for the West Kettle River was available on a 1:31400 black and white aerial photograph taken in 1951 (Appendix B Figure B24). While using the reach map 146 (figure 5-7) in D'Aoust, the average channel width of the river was measured on the photograph to be approximately 38 m (Appendix C). The technique for this measurement was identical to the method used for the current fieldwork and is discussed in detail in section 3.1.2.2. These two widths lead to a change in width from the period of 1951 -1997 to be approximately 15%. 4.9 Summary Data The following Tables 4-4, 4-5, and 4-6 are summary tables of the major physical and hydrological data collected for the preceding case studies. Table 4-4 is a brief summary of the discharge data for each case study. Table 4-5 is summary data for the collected case studies including channel characteristics and hydrological analysis. Similarly, Table 4-6 is summary data for the previously collected four case studies. 147 River Reach # Area (km2) Qmaf(mJ/s) Method Bonaparte 1 5069 32 Single gauge Coldwater 1 735 78 Single gauge Coldwater 2 91.5 30 Single gauge Coldwater 3 81.6 27 Single gauge Deadman 1 1412 . 22 Single gauge Deadman 2 1432 22 Single gauge Elk 1 134 124 Single gauge Elk 2 52 61 Single gauge Eve 1 146 152 Regional average Salmon 1 733 603 Single gauge Tahsis 1 51.6 196 Regional average Big Horn Creek 4 117 19 Regional average Narrowlake Creek TR3 185 30 Regional regression Slesse Creek D 158 89 Single gauge West Kettle 1 1520 127 Single gauge Table 4-4 Discharge summary data for all case studies Peak T, I-I CM CO CM •* o o A o o A o> 03 N/A CO N/A P/M ratio 2.80 1.59 2.15 2.17 4.24 4.33 1.51 1.50 N/A 1.94 N/A FFMI 0.29 0.13 0.15 0.15 0.28 0.28 0.15 0.15 0.16 0.15 0.17 Kf 0.57 0.59 0.45 0.71 0.28 0.56 1.01 1.22 2.29 0.48 0.79 1 a to E CN r o LL CO 00 CN CN CN CN CN •* CN VO CN I/O r o O VO VO OS % Change W t -CN uo r o r o CN VO r~ "3-VO r~ CN OS •sf r o CN •SI-v o o acteristics Historic E r o v o CN 00 CN «1 r~ r o CN 00 m •* OO r o Os CN Channel Char; £> o E OS CN CN VO r o v o o r o r o 00 00 r~ CN VO o OS © VO Sinuosity, - r> -. - - - r o CN r o 0.0029 0.0049 0.0033 0.0047 0.0027 0.0047 0.0063 0.0104 0.0107 0.0018 0.0039 3 1 o Q E 0.016 0.016 0.0085 0.005 0.03 0.011 0.016 0.017 0.005 0.015 0.009 •O o d E 0.02 0.033 0.017 0.017 0.03 0.035 0.025 0.015 0.015 0.027 0.018 s" o Q E 0.013 0.012 0.015 0.013 0.035 0.027 0.033 0.045 0.02 0.018 0.03 a> E •c o Q E 0.021 0.065 0.037 0.03 0.05 0.04 0.05 0.062 0.034 0.042 0.041 Reach # Bonaparte 1 Coldwater 1 Coldwater 2 Coldwater 3 Deadman 1 Deadman 2 Elk 1 Elk 2 Eve 1 Salmon 1 Tahsis 1 o c j cu fa cn fa « a -a a « C O '& 3 £i 'fa CA -a "3 s s a o ns v CA S S • f i 'fa cj a C J fa "3 a a « cj >< •n a V so « C J T5 V •** C J <u "© CJ fa cS CM « i? es s a s in • 3 a H 149 03 u OH N/A V/N m T f m P/M ratio N/A N/A 1.94 1.44 FFMI 0.16 0.31 0.15 0.11 0.80 1.67 0.68 0.34 i a 6 O s o m 00 S O f N % Change W Tf o o 00 Tf Wl Historic W, m S O O s C N 00 f N 00 m i» o OH O s m S O Tf 00 _o O Tf Tf Sinuosity, P - r n u C O 0.0064 0.007 0.021 0.0028 1 E Q 0.052 0.133 e o Q 0.035 0.063 0.133 0.0775 Reach # Big Horn 4 Narrowlake TR3 Slesse D West Kettle o t. es <u >> fl fl fl s £ "E cu S s U fl o -o CU S3 Cs o • HH s-CU a A i . 9 CU CU T3 S cu Vi « u CU u "3 Vi S © u a . i_ « es -a E B s C/5 s© 4 ^ Si 150 CHAPTER 5 5 CHANNEL MORPHOLOGY ANALYSIS 5.1 Analysis of Morphological Indices Data from the case studies was used to assess whether the proposed riparian disturbance sensitivity index (RDSI) C, or flood-related indices, FFMI or Peak/Mean (P/M) ratio, explain the changes in channel width and morphology. The measured changes in width, expressed as AW%, are plotted for the three indices (Figures 5-1, 5-2, and 5-3), with the data stratified into disturbed and intact reaches - depending upon the current condition of the riparian zone. The data are presented in Tables 4-5 and 4-6. It is realized that error exists for the AW% values shown in Figures 5-1, 5-2, and 5-3. The error remains difficult to quantify for two reasons. First, the calculation for AW% represents the percent difference between two average values of measured width. These two measurements in themselves have their own standard error (Appendix C). It is impossible to calculate a standard error for the difference between the two measurements because the individual width measurements do not directly correlate in location from the historic aerial photograph to the present day reach survey. In order to obtain some grasp of the potential error for the AW% values, the range in standard errors for the individual average measurements was 3% - 21% with an average of 11% error. Therefore, it is 151 qualitatively known that errors potentially exist in the width measurements and the resulting AW%, but the precise extent of this error is not known. Regardless, more precise knowledge of this error would not change the general relation demonstrated in Figure 5-1. 5.1.1 Riparian Disturbance Sensitivity Index The data collected and analysed for the case studies was necessary to calculate the proposed riparian disturbance sensitivity index defined by Eq. 2.22. An estimated sensitivity index was determined based on the data from each reach discussed in chapter 4. According to the calculations and the estimates for the change in width of the channel from an intact to disturbed condition (see Eq. 3.8), a graphical relationship between the sensitivity index and percent change in width was created (Figure 5-1). The linear relationship suggests that the riparian disturbance sensitivity index can reasonably estimate the potential morphology of an alluvial gravel-bed river to an established coniferous riparian disturbance (Figure 5-1). Based on a least squares linear regression analysis, the relationship for the disturbed channels, excluding the Eve River, between the sensitivity index and change in width is strongly correlated (r = 0.85, p < 0.0005). It should be noticed that the data point representing Slesse Creek could potentially be an outlier influencing the regression statistics. When Slesse Creek is excluded from the regression, AW% for the disturbed channels still strongly correlates with C, (r2 = 0.57, p = 0.01). There is no apparent physical reason for removing Slesse 152 Creek from the rest of the data set, and therefore, it is assumed that the initial relationship represents the best-fit relationship for the data set of eleven disturbed river case studies (Figure 5-1). AW% = 272 r - 49.9 r2 = 0.85 p < 0.0005 5 Figure 5-1 Percent change in width versus riparian disturbance sensitivity index (Q. Points corresponding to Slesse Creek and Eve River are indicated. Solid line represents linear regression. Two things to point out with the observed relationship between AW% and C, are the significance of the intercept and the location of data points representing intact riparian reaches compared to the linear regression relationship. The intercept for this relationship is approximately at a value of 0.18 for C,. This suggests that for any value of £, less than or equal to about 0.18, the change in width of the channel would be essentially zero even for a heavily disturbed riparian zone. That is, the riparian vegetation is not significantly 153 influencing channel morphology for very low values of C,. Unfortunately, no value of C, less than or equal to 0.18 was obtained for any of the disturbed case studies. Also, none of the disturbed case studies displayed a negligible or zero change in width, probably obscured by noise and measurement error. Therefore, no physical data is available to promote the significance of the intercept for the relationship, but it is an interesting attribute to discuss. It is difficult to test the significance of the location of the intact case studies, but it is reassuring that they are located below the relative trend of the disturbed data points. It is reassuring because the intact reaches should have relatively little or no change in width because their riparian vegetation is presently intact and able to be a controlling factor on the lateral morphology of the channels. This relationship positively contributes to the theory that the riparian vegetation is the primary controlling factor for the lateral stability of alluvial gravel-bed rivers with noncohesive banks in southern British Columbia, including northern Vancouver Island. 5.1.2 Flood Indices To test whether there is any relation between large or "catastrophic" floods and change in width, AW% is plotted against FFMI and P/M ratio (Figures 5-2 and 5-3). For both flood-based indices there is no statistical correlation for AW% (Figures 5-2, 5-3). The apparent weak relation between AW% and FFMI (r = 0.24, p = 0.11) is due to one single 154 point (Slesse Creek). When Slesse Creek is removed there is no relationship (r2 = 0.02, p = 0.68). AW% = 714 FFMI + 1.83 r 2 = 0.23 p = 0.12 0.35 Figure 5-2 Percent change in width versus flash flood magnitude index. The point corresponding to Slesse Creek is indicated. The case studies from the more arid climatic zones of the province (Bonaparte River and Deadman River) show some of the most variable results based on the FFMI value (Table 4-2). But, the high FFMI values for these two watersheds does not reflect the magnitude of their morphologic changes, especially for the two separate reaches of the Deadman River. 155 450 400 350 300 g 250 I 200 150 100 50 • • Disturbed • Intact • • • n • , — • — 0 J 0.00 1.00 2.00 3.00 P/M Ratio AW% = -16.1 P/M+ 196 r2 = 0.02 p = 0.73 4.00 5.00 Figure 5-3 Percent change in width versus peak/mean discharge ratio Results indicate that the ratio of the instantaneous peak flood divided by mean annual flood does not explain the observed change in channel width and there is no significant correlation between AW% and the P/M ratio (r2 = 0.02, p = 0.73). Based on this conclusion, the large or "catastrophic" flood does not appear to be a necessary factor that accounts for the lateral geomorphology of rivers examined in this study. Apart from the details of the regressed relationship, the Eve River case study appears to be an outlier compared to the other disturbed case study results. This case study has potential reason(s) why it appears to be an outlier to the relationship represented by the 156 remaining case studies in Figure 5-1. These reasons are discussed in the following section 5.2.6. 5.2 Reach by Reach Discussion The following sections are a description of the sensitivity index (Q results for each individual case study. 5.2.1 Big Horn Creek Big Horn Creek was a case study where the majority of the data is supplied by external sources [see (Klohn-Crippen Consultants LTD. 1998)] and produced under the former Watershed Restoration Program. Using this data, the estimated C, was 0.80 for an increase in width of approximately 110%. The result from this case study closely resembles the result of the Tahsis River site and appears to correspond with the relationships of the remaining data (Figure 5-1). FFMI for this site is 0.16, which indicates low annual variability in flood magnitude for this watershed. Being an ungauged system, a peak flow analysis was not performed for this case study. 5.2.2 Bonaparte River The Bonaparte River represents a system that has been impacted by grazing and forestry in recent history. It has shown an increase in width of approximately 130% with and estimated C, value of 0.57. This result appears to coincide with the relationship of Figure 157 5-1. It should be mentioned that the system has experienced two abnormally large floods according to the discharge record (see Appendix A Figure A2) from the gauge located downstream from the study site. The substantial increase in channel width appears to have occurred between 1995 (year of last aerial photograph) and 2002 (field survey). This width increase may have occurred during a large flood in 1999, which had a magnitude of approximately twice the mean annual flood, or a return period of approximately 60 years (Table 4-2, Appendix B Figures BI & B2, and Appendix C). 5.2.3 Coldwater River The Coldwater River provides good support for the influence of riparian vegetation on channel morphology. A valuable comparison exists between reaches 1 and 3 of the Coldwater River (Figure 5-4 and 5-5) because of the drastically divergent disturbance histories between the two reaches. As mentioned previously, the uppermost reach (Reach 3) of the Coldwater River experienced a fire in 1938, but currently has substantial coniferous tress which create an intact riparian environment. This is contrasted by the riparian conditions of Reach 1, which has been deforested because of heavy logging and grazing since 1953. Reach 3 has a C, of approximately 0.71 but its riparian environment is intact, and a minimal change in width (less than 5%) between 1953 - 2002. Reach 1 has a C, value of approximately 0.59 and its increase in width is 135%. This drastically different morphology for a relatively similar sensitivity index is believed to be due to differences in condition of riparian vegetation between the two reaches. 158 Reach 2 of the Coldwater River represents an intermediate reach between the two reaches of primary interest. The channel slope (0.0033 m/m) for this reach is substantially less than the channel slopes for the other two reaches (0.0049 m/m and 0.0047 m/m) resulting in a reduced L, of 0.45. The change it width for this reach was estimated to be 23%, but it was classified as an intact reach because of its current riparian condition even though it appears to agree reasonably well with the disturbed data set. One other thing to mention is that the impacts resulting from the construction and persistence of the Coquihalla highway are not quantifiable. The locations of the study reaches were located in order to minimize any potential effects; therefore it was assumed that any controls due to the road were negligible. The flood related indices do not appear to represent the changes experienced on the Coldwater system (FFMI = 0.13, 0.15, 0.15; P/M = 1.59, 2.15, 2.17; Rl = 27,49,49) (Table 4-2). The upper two reaches experienced substantial floods according to the discharge record, approximately 1 in 50 year events, but the non-existent lateral geomorphology of these reaches does not show any major effects from this event. 159 5.2.4 Deadman River The results for the two Deadman River reaches (Appendix B Figures B9 & BIO) are consistent with estimations of Reach 1 has a C, of 0.28 and an increase in width of 76%, while reach 2 has a C, of 0.56 and an increase in width of 160%. As mentioned in section 4.3.3, the main disparity between the two reaches appears to be the channel slopes. The greater channel slope of reach 2 results in an enhanced sensitivity of this reach to riparian influences. Since both of these reaches were analysed using the same gauge, and no major inflows exist between the two reaches, both of the reaches have the same FFMI (0.28) for drastically different changes in width. Similarly, the peak/mean ratios are very similar (4.2 & 4.3), with the largest flood return period of > 100 years occurring in 1990. Therefore, the Deadman River is a fairly variable system with a large flood in recent record, but according to this study the vegetation is the most influential factor in controlling the change in width and morphology of the two separate reaches. 5.2.5 Elk River Reach 1 of the Elk River has a sensitivity index of 1.01, which indicates high sensitivity to a riparian disturbance. This reach was heavily disturbed from logging sometime in the 1940's and was potentially influenced by an increased influx of sediment from the 1946 landslide event (see Appendix B Figures BI 1 & B12). An increase in width of approximately 270% is a consistent response for this river to the induced changes of the riparian environment. 162 The intact Reach 2 of the Elk River has a value of C, of approximately 1.22. The riparian zone remains intact and has not been impacted by logging or forestry (see Appendix B Figures B13 & B14). Observed channel widening of approximately 50% may be a consequence of the flood wave and sediment input resulting from the 1946 landslide into the headwater lake at the top of the system (M. Miles and Associates 1998). In addition, steep valley walls on both sides of the channel parallel the reach. These valley walls may be confining the reach and not allowing it to adjust its planform morphology as is necessary for a free forming alluvial system. Flood events do not appear to be an effective determinant for the morphology of the Elk system (Table 4-2). 5.2.6 Eve River The transition slope of the Eve River with no vegetative influences (<t>'= 40°) was estimated at 0.0030, see Eq. 2.16. When compared to the slope of the channel of 0.0107, the channel slope is significantly greater than the transition slope, which predicts the channel to have a transitional (wandering) or braided river morphology (see Millar (2000)). This result appears to be consistent with the morphology displayed in the 1953 photograph of the unlogged Eve watershed (see Appendix B Figure BI 5). It appears that the Eve River reach surveyed for this study is historically a wandering river, which means that the assumptions necessary for this analysis do not hold. The theory presented by Millar & Quick (1993) is only valid for single-thread channels. Therefore, the Eve River is beyond the scope of this analysis. 163 The riparian disturbance sensitivity index for Eve River reach takes a value of approximately 2.3, which suggests that the Eve River would be extremely sensitive to bank strength removal. The instabilities of the Eve River reach appear to be the natural morphology, historically a wandering gravel-bed river channel. In conjunction with the single-thread channel requirement, the flow capacity of the current channel appears to be much less than the mean annual flood. During the site visit in September of 2002, a rainfall event caused inundation of the floodplain and activation of numerous side channels, which figures into the invalid assumption of the mean annual flood estimation and a single-thread channel. After the justification to disregard the Eve River, it appears that the sensitivity index predicts the morphological result of riparian removal from alluvial gravel-bed rivers with noncohesive banks. 5.2.7 Narrowlake Creek Data for the Narrowlake Creek case study was collected and supplied by Andrew Wilson (Ministry of Water, Land, and Air Protection). The study reach is heavily impacted by past logging, including removal of riparian vegetation, see Figure 2 of Wilson (2001). The calculated value for C, of 0.68, is consistent with an increase in width of approximately 100%. No value was available for the peak/mean ratio, but the regional estimate for the FFMI (0.15) indicates low variability in the annual flood magnitudes. A gauge is not available for the Narrowlake Creek system, therefore no peak analysis was performed for this case study. 164 5.2.8 Salmon River The Salmon River is an interesting case because of the lack of historical data. As mentioned in section 4.6.4, logging practices in the Salmon River watershed existed prior to the first aerial photograph taken in 1946 (Appendix B Figure B18). Therefore, the results from this case study are speculative because the increase in width (41%) may not represent the true morphology of the system from an intact state. The calculated value for C. of 0.48 suggests a system that is moderately sensitive to riparian disturbances. According to the regression, the AW% for this reach could be approximately 80%. Once again, the estimated values for the FFMI (0.15) and the P/M ratio (1.94), with an estimated return period of 40 years, do not appear to be major contributors to the morphology of the Salmon River as compared to the results of the sensitivity index. Unfortunately, the results and conclusions for the Salmon River are fairly speculative. 5.2.9 Slesse Creek Data used in analysis of Slesse Creek was collected by MacVicar (1999). As discussed in section 5.1, Slesse Creek appears to have a substantial influence on the regressed relationship because of its abnormally high increase in width (420%) for the analysed data set. Besides this fact, the morphology of Slesse Creek appears to be justified by the relationship because of its large C, value (1.67). Based on a C, value of 1.67, Slesse Creek is a highly sensitive system to riparian destruction and its present day condition reflects this situation. The P/M ratio for Slesse Creek (1.94) and estimated return period of 45 years is very similar to that of the Salmon River, but the estimated FFMI (0.31) is substantially greater, although low compared to rivers in arid and semi-arid regions. This 165 may suggest that the variability and not the ratio of the sizes of the floods on the Slesse system may have some influence on the morphology of the system. 5.2.10 Tahsis River The estimate of <^  for the Tahsis River was 0.79 for an increase in width of 106%. These values appear to conform reasonably well to the remaining points in the disturbed data set. One concern is the intensive logging impacts that occurred within the system upstream of the study site (Kellerhals and Miles 1996). The increased lateral morphology and resulting additional sediment within the system could be a contributing factor to the lateral morphology of the study reach. Unfortunately, no means exist to separate out these potential effects from the induced changes caused by riparian destruction. But, based on the relationship of this study, the riparian structure was a controlling factor on the morphology of the Tahsis River, as opposed to the regional flood variability (FFMI = 0.17). 5.2.11 West Kettle River Similar to three other case studies, the principal components of this case study's data are attributed to an exterior investigator. Stephane D'Aoust collected the data for his thesis topic of large woody debris performance (D'Aoust 1998). The West Kettle River was the least sensitive river out of the disturbed case studies both theoretically and physically. The insensitivity is represented by C, of 0.34 and increase in width of roughly 15%. The insensitivity of the West Kettle River may be attributed to its larger bed sediment and shallower channel slope when compared to the bulk of the other disturbed case studies. 166 Similarly, the magnitudes of the FFMI (0.11) and P/M ratio (1.44) appear to be show that variability in the magnitude of annual floods is low. The estimated return period for the peak flood is approximately 15 years. 5.3 Nature of A W % — C Relation The AW% — C, relation formulated in this study is not intended to be a universal predictive tool for estimating the change in width or morphology of induced riparian disturbances. The design of the riparian disturbance sensitivity index C, is to assist in explaining the influence of riparian vegetation on the structure of gravel-bed rivers. The data used to formulate the relation (Figure 5-1) are from similar biogeoclimatic zones and all of the data was collected within the boundaries of southern British Columbia. Therefore, this particular relationship is significantly dependent upon the vegetation that inhabits these particular regions. The slope of the linear relationship of AW% versus t, is a function of the influence of the historic pristine vegetation. Most likely, the greater the antecedent influence of the riparian vegetation on the structure of the fluvial system, the steeper the slope for AW% — C, relation for disturbed reaches. 5.4 Index Sensitivity to Grain Size Selection (Bank D50) The grain size for the bed and bank pebble counts and the bar and bank bulk samples were estimated. Since this data was readily available for the measured case studies, this data was applied to see how sensitive the riparian disturbance sensitivity index C, was to 167 the grain size. The following table shows the estimation for C, based on the different choices for the representative D 5 0 o f the different sampling procedures and locations. Stream Reach # D riffle Dbar % Diff. Dbank % Diff. D sub bank % Diff Bonaparte 1 0.57 0.77 35 0.59 3.5 0.67 18 Coldwater 1 0.59 1.7 188 0.9 53 1.42 141 Coldwater 2 0.45 0.79 76 0.73 62 1.13 151 Coldwater 3 0.71 1.21 70 1.02 44 2.19 209 Deadman 1 0.28 , 0.35 25 0.39 39 0.39 39 Deadman 2 0.56 0.72 29 0.61 8.9 1.25 123 Elk 1 1.01 1.31 30 1.56 55 2.06 104 Elk 2 1.22 1.49 22 2.96 143 2.74 125 Eve 1 2.29 3.19 39 3.82 67 7.6 231 Salmon 1 0.48 0.81 69 0.63 32 0.91 90 Tahsis 1 0.79 0.96 22 1.33 68 2.04 158 Table 5-1 C, sensitivity to grain size. The values are the estimates of C, for the different sediment diameters. The percent difference is the percent difference between the riffle value and the other corresponding value. The sensitivity analyses displays the influence that grain size has on the morphological sensitivity o f the alluvial gravel-bed river with cohesionless banks. From a physical perspective, a larger grain size describes a more resistant material to fluvial entrainment and erosion. Therefore, generally a river with a larger mean particle size fabricating the banks w i l l be less sensitive to riparian disturbances with consistent valley slopes and discharges. It would seem more physically accurate to chose the bank surface grain size value because this value most closely represents the resistance properties of the potentially weakened banks caused by riparian disturbances. O n the other hand, the 168 theory is based on the assumption that the material composing the bed and banks is consistent. Based on a comparison of the regressions of the bed D50 (Figure 5-1, r2 = 0.85, p < 0.0005) and the bank D 5 0 (r 2 = 0.63, p = 0.011), excluding the Eve, the relationship for C, is largely more statistically significant when the bed D 5 0 is the sediment diameter used in the relationship. The values for the bar and bank subsurface are not analyzed because the represent the sediment below the surface of the bed and the banks. These values were obtained through the bulk sieve analysis and are more useful for sediment transport studies (section 3.1.1.3). Figure 5-6 Percent change in width versus riparian disturbance sensitivity index (Q, using bank D50. Points corresponding to Slesse Creek and Eve River are indicated. 169 5.5 Role of Floods vs. Riparian Disturbance Typically, substantial changes in channel width habitually occur during large, infrequent flood events. The question is whether or not these floods are the cause of the widening. According to the results of this study (section 5.1), since the variability and size of these large floods do not significantly correlate with change in width, they do not appear to be the primary cause of thechannel widening. The role of the floods potentially could trigger an already unstable condition caused by a decrease in strength of the banks from riparian vegetation removal. A potential situation exists where the destabilized banks already exist prior to a peak flow event, with the peak flow event being the agent for widening and the riparian disturbance the cause. This conclusion should be tested for varying discharge regimes and inherent vegetation communities, in order to examine potentially different vegetation influences, and variability and magnitude of flood events. 170 C H A P T E R 6 6 CONCLUSIONS AND RECOMMENDATIONS 6.1 Summary Fifteen case studies from British Columbia were examined to investigate the sensitivity o f alluvial single-thread gravel-bed rivers with noncohesive banks to riparian disturbances. Eleven out of twelve of the disturbed case studies were used to investigate the validity of a proposed theoretical riparian disturbance sensitivity index (RDSI) ( Q in explaining the role of riparian vegetation in channel change in width and morphology. Data for the case studies was collected using a methodology incorporating field and analytical techniques. The field technique consisted of a reach survey including channel sediment sampling, longitudinal profile survey, and riffle cross-sectional surveys. The analytical techniques included a historical aerial photo analysis and hydrological analysis. The aerial photo analysis provided estimates of historical pristine, or intact riparian, channel widths. The hydrological analysis, including watershed delineation, regional discharge analysis, flood variability, and peak flood magnitude, provided a mean annual flood, flash flood magnitude index (FFMI) , peak/mean ratio (P/M), and peak flood return period for each reach. The potential correlation of the R D S I , F F M I , and P / M with percent change in width ( A W % ) was used to analyse the influencing factors in the morphology of gravel-bed rivers with disturbed riparian vegetation. 171 6.2 Conclusion of RDSI (Q Results The riparian disturbance sensitivity index appears to explain the sensitivity of alluvial gravel-bed rivers in this study to morphological changes caused by riparian disturbance. The riparian disturbance sensitivity index C. statistically correlates very significantly with AW%. In contrast, the flood-based indices displayed no statistically significant correlation to the change in width for the presented case studies. Potentially the large floods are triggers for channel morphology, but it is not evident that they represent a primary cause of the channel widening. While flooding may be an agent for widening, it is not clear that such widening would have occurred in the absence of riparian disturbance. Based on the methodology presented in this thesis, the sensitivity index could be used as a screening tool to determine the sensitivity of a particular river or stream to riparian logging (Millar 2001). In other words, the sensitivity index could potentially test whether riparian removal caused an existing unstable river reach. Typically, the establishment of vegetation is a more economically viable and an ecologically preferred option for stream restoration as compared to bank armouring or more classic engineering solutions for unstable, eroding channel reaches (Abernethy and Rutherfurd 2000). 172 6.3 Future Work It is recommended that future work be completed for the complete test of the riparian disturbance sensitivity index. This thesis represented 15 data sets from the province of British Columbia that were applied to test the validity of the riparian disturbance sensitivity index. It appears that only fourteen of these data sets are valid since one of the data sets appears to be unrepresentative of the alluvial single-thread gravel-bed river for which the index was formulated. Therefore, it seems reasonable that more data sets from various conditions, disturbances, and locations to incorporate varying flood regimes and vegetation should be used to further test the performance of the sensitivity index in predicting the possible morphology of a single-thread gravel-bed river with noncohesive banks to riparian disturbances and the role of riparian disturbance versus floods. 173 REFERENCES Abam, T. K. S. (1997). "Aspects of Alluvial River Bank Recession: Some Examples from the Niger Delta." Environmental Geology 31(3/4): 211-220. Abernethy, B. and I. D. Rutherfurd (1996). Vegetation and Bank Stability in Relation to Changing Channel Scale. First National Conference on Stream Management in Australia, Merrijig, Australia. Abernethy, B. and I. D. Rutherfurd (1998). "Where Along a River's Length Will Vegetation most Effectively Stabilise Stream Banks?" Geomorphology 23: 55-75. Abernethy, B. and I. D. Rutherfurd (2000). "The Effect of Riparian Tree Roots on the Mass-stability of Riverbanks." Earth Surface Processes and Landforms 25: 921-937. Andrews, E. D. (1984). "Bed-Material Entrainment and Hydraulic Geometry of Gravel-Bed Rivers in Colorado." Geological Society of America Bulletin 95: 371-378. ASCE Task Committee on Hydraulics, B. M., and Modelling of River Width Adjustment (1998). "River Width Adjustment I: Processes and Mechanisms." Journal of Hydraulic Engineering 124(9): 881-902. Babikaiff and Associates Geoscience (1997). Slesse Creek: Channel Assessment and Prescription Development (Draft), Prepared for Steelhead Society Habitat Restoration Corp.: 72. Baker, V. R. (1977). "Stream-channel Response to Floods, with Examples from Central Texas." Geological Society of America Bulletin 8 8 : 1057-1071. BC Hydro (2002). Campbell River Water Use Plan Update. 2003. Beard, L. R. (1975). Generalized Evaluation of Flash-flood Potential, Texas University Center for Research in Water Resources: 27. Beaudry, P. G. (1998). Effects of Forest Harvesting on Streamflow and Sediment Concentrations of Small Streams in Central British Columbia. CWRA 51 st Annual Conference Proceedings, Mountains to Sea: Human Interaction with the Hydrologic Cycle, Victoria, B.C. Beaudry, P. G. and A. Gottesfeld (1999). Causes of Stream Channel Change in B.C.'s Central Interior. Confronting Uncertainty: Managing Change in Water Resources and the Environment, Richmond, B.C., Canadian Water Resources Association. Beeson, C. E. and P. F. Doyle (1995). "Comparison of Bank Erosion at Vegetated and Non-vegetated Channel Bends." Water Resources Bulletin 31(6): 983-990. 174 Beschta, R. L. (1978). "Long-Term Patterns of Sediment Production Following Road Construction and Logging in the Oregon Coast Range." Water Resources Research U 1011-1016. Beschta, R. L. (1984). River Channel Responses to Accelerated Mass Soil Erosion. Symposium on the Effects of Forest Land Use on Erosion and Slope Stability, Honolulu, Hawaii, Environment and Policy Institute, East-West Center. Beschta, R. L., M. R. Pyles, A. E. Skaugset and C. G. Surfleet (2000). "Peakflow Responses to Forest Practices in the Western Cascades of Oregon, USA." Journal of Hydrology 233: 102-120. Blench, T. (1957). Regime Behaviour of Canals and Rivers. London, Butterworths Scientific Publications. Bowie, A. J. (1982). "Investigations of Vegetation for Stabilizing Eroding Streambanks." Transactions of the A.S.A.E. 25f6): 1601-1606,1611. Bray, D. I. (1972). Generalised Regime-type Analysis of Alberta Rivers. Edmonton, Alberta, University of Alberta. Bray, D. I. (1975). "Representative Discharges for Gravel-Bed Rivers in Alberta, Canada." Journal of Hydrology 27: 143-153. Bray, D. I. (1987). "A Study of Channel Changes in a Reach of the North Nashwaaksis Stream, New Brunswick, Canada." Earth Surface Processes and Landforms 12: 151-165. Brice, J. C. (1964). Channel Patterns and Terraces of the Loup Rivers in Nebraska. Washington, United States Geological Survey: D1-D40. Bridge, J. S. (1993). The Interaction Between Channel Geometry, Water Flow, Sediment Transport and Deposition in Braided Rivers. Braided Rivers. J. L. Best and C. S. Bristow, Geological Society Special Publication. 75: 13-71. British Columbia Power Commission (1958). Annual Structures Inspection Report on the Salmon River Diversion. Burnaby, British Columbia, BC Hydro. Brush, L. M. (1961). Drainage Basins, Channels, and Flow Characteristics of Selected Streams in Central Pennsylvania. Washington, U.S. Geological Survey. Bunte, K. and S. R. 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, U.S.D.A. Forest Service, Rocky Mountain Research Station, RMRS-GTR-74: 428 p. Campo, L. (pers. comm.). Elk River Gauge Drainage Area, Environment Canada, Water Survey Branch. 175 Carling, P. (1988). "The Concept of Dominant Discharge Applied to Two Gravel-Bed Streams in Relation to Channel Stability Thresholds." Earth Surface Processes and Landforms 13: 355-367. Carson, M. A. (1971). The Mechanics of Erosion. London, Pion Limited. Center, S. S. T. (1996). Gravelometers: Gravel Templates for Pebble Counting in Gravel-Bed Streams. Stream. Apr i l : 1-4. Chang, H. H. (1985). "River Morphology and Thresholds." Journal of Hydraulic Engineering 111(3): 503-519. Charlton, F. G., P. M. Brown and R. W. Benson (1978). The Hydraulic Geometry of Some Gravel Rivers in Britain. Wallingford, Hydraulics Research Station. Chow, V. T. (1959). Open-channel Hydraulics. Boston, Mass., McGraw Hill. Church, M. (1992). Channel Morphology and Typology. The Rivers Handbook. P. Calow and G. E. Petts. Oxford, Blackwell Science: 126-143. Church, M. (2002). "Geomorphic Thresholds in Riverine Landscapes." Freshwater Biology 47(4): 541-557. Church, M. (pers. comm.). Bankfull Discharge Return Period. Church, M. A., D. G. McLean and J. F. Wolcott (1987). River Bed Gravels: Sampling and Analysis. Sediment Transport in Gravel-bed Rivers. C. R. Thorne, J. C. Bathurst and R. D. Hey, John Wiley & Sons Ltd.: 43-88. Coppin, N. J. and I. G. Richards, Eds. (1990). Use of Vegetation in Civil Engineering. London, Butterworths. Corbett, E. S. and J. A. Lynch (1985). Management of Streamside Zones on Municipal Watersheds. Riparian Ecosystems and their Management: Reconciling Conflicting Uses, First North American Riparian Conference, Tucson, Arizona, USDA Forest Service. Coulson, C. H. and W. Obedkoff (1998). British Columbia Streamflow Inventory, Water Inventory Section, Resources Inventory Branch, Ministry of Environment Lands and Parks. D'Aoust (1998). Large Woody Debris Fish Habitat Structure Performance and Ballasting Requirements. Civil Engineering. Vancouver, B.C., University of British Columbia: 175. Diplas, P. and J. B. Fripp (1992). "Properties of Various Sediment Sampling Procedures." Journal of Hydraulic Engineering 118(7): 955-970. 176 Diplas, P. and A. J. Sutherland (1988). "Sampling Techniques for Gravel Sized Sediments." Journal of Hydraulic Engineering 114(5): 484-501. Eaton, B., M. Church and D. Ham (2002). "Scaling and Regionalization of Flood Flows in British Columbia, Canada." Hydrological Processes 16(16): 3245-3263. Emmett, W. W. and G. M. Wolman (2001). "Effective Discharge and Gravel-Bed Rivers." Earth Surface Processes and Landforms 26: 1369-1380. Environment Canada (2000). HYDAT. 2003. Fahnestock, R. K. (1963). "Morphology and Hydrology of a Glacial Stream - White River, Mount Rainier, Washington." U.S. Geological Survey Professional Paper 422-A: 1-70. Ferguson, R. and P. Ashworth (1991). "Slope-Induced Changes in Channel Character Along a Gravel-Bed Stream: The Alt Dubhaig, Scotland." Earth Surface Processes and Landforms 16: 65-82. Ferguson, R. I. (1984). The Threshold Between Meandering and Braiding. 1st International Conference on Hydraulic Design in Water Resources Engineering: Channels and Channel Control Structures, University of Southampton, Springer-Verlag. Frankenberg, J., W. K. Tennant and J. W. Tilleard (1996). Mechanisms of Stream Bank Erosion. First National Conference on Stream Management in Australia, Merry ig, Australia. Fredsoe, J. (1978). "Meandering and Braiding of Rivers." Journal of Fluid Mechanics 84(4): 609-624. Fripp, J. B. and P. Diplas (1993). "Surface Sampling in Gravel Streams." Journal of Hydraulic Engineering 119(4): 473-490. Gould, E. (1975). Logging : British Columbia's logging history. Saanichton, British Columbia, Hancock House Publishers Ltd. Gran, K. and C. Paola (2001). "Riparian Vegetation Controls on Braided Stream Dynamics." Water Resources Research 37(12): 3275-3283. Grant, G. E., M. J. Crozier and F. J. Swanson (1984). An Approach to Evaluating Off-site Effects of Timber Harvest Activities on Channel Morphology. Symposium on the Effects of Forest Land Use on Erosion and Slope Stability, Honolulu, Hawaii, Environment and Policy Institute, East-West Center. Gray, D. H. and A. MacDonald (1989). The Role of Vegetation in River Bank Erosion. Hydraulic Engineering, Proceedings of the ASCE Conference. M. A. Ports. New York: 218-223. 177 Gregory, K. J. and A. M. Gurnell (1988). Vegetation and River Channel Form and Process. Biogeomorphologv. H. A. Viles. Oxford, UK, Basil Blackwell Ltd.: 11-42. Gregory, K. J., A. M. Gurnell and C. T. Hill (1985). "The Permanence of Debris Dams Related to River Channel Processes." Hydrological Sciences Journal 30: 371-381. Harr, R. D., W. C. Harper and J. T. Krygier (1975). "Changes in Storm Hydrographs After Road Building and Clear-cutting in the Oregon Coast Range." Water Resources Research 11(3): 436-444. Harr, R. D., A. Levno and R. Mersereau (1982). "Streamflow Changes After Logging 130-year-old Douglas Fir in Two Small Watersheds." Water Resources Research 18(3) 637-644. Harr, R. D. and F. M. McCorison (1979). "Initial Effects of Clearcut Logging on Size and Timing of Peak Flows in a Small Watershed in Western Oregon." Water Resources Research 15(1): 90-94. Harris, C. D. (1987). Hydrology and Logging in the Carnation Creek Watershed - What Have We Learned? Proceeding of the Workshop: Applying 15 Years of Carnation Creek Results, Nanaimo, B.C., Carnation Creek Steering Committee. Harris, J. D. (1986). Design Flood Estimation for Medium and Large Watersheds. Chapter 8 of MTC Drainage Manual, Ontario Ministry of Transportation and Communications: H4-7. Hartman, G. F. and J. C. Scrivener (1990). Impacts of Forestry Practices on a Coastal Stream Ecosystem, Carnation Creek, British Columbia. Nanaimo, British Columbia, Department of Fisheries and Oceans, Biological Sciences Branch, Pacific Biological Station: 53-56,96-105,126-127. Heede, B. H. (1985). Interactions Between Streamside Vegetation and Stream Dynamics. Riparian Ecosystems and their Management: Reconciling Conflicting Uses, First North American Riparian Conference, Tucson, Arizona, USDA Forest Service. Henderson, F. M. (1963). "Stability of Alluvial Channels." Transactions of the A.S.C.E. 128(Part I): 657-686. Henderson, F. M. (1966). Open Channel Flow. New York, Macmillan Pub. Co. Hession, W. C. (2001). Riparian Forest and Urban Hydrology Influences on Stream Channel Morphology: Implications for Restoration. World Water and Environmental Resources Congress, Orlando, FL. Hession, W. C, et al. (2000). "Ecological Benefits of Riparian Reforestation in Urban Watersheds: Study Design and Preliminary Results." Environmental Monitoring and Assessment 63: 211-222. 178 Hetherington, E. D. (1982). A First Look at Logging Effects on the Hydrologic Regime of Carnation Creek Experimental Watershed. Proceedings of the Carnation Creek Workshop, A 10 Year Review, Malaspina College, Nanaimo, B.C. Hetherington, E. D. (1987). Hydrology and Logging in the Carnation Creek Watershed -What Have We Learned? Proceedings of the Workshop: Applying 15 Years of Carnation Creek Results, Nanaimo, B.C., Carnation Creek Steering Committee. Hey, R. D. (1975). Design Discharge for Natural Channels. Science. Technology and Environmental Management. R. D. Hey and T. D. Davies. Westmead, England, Saxon House, D.C. Heath Ltd.: 73-88. Hey, R. D. (1978). "Determinate Hydraulic Geometry of River Channels." Journal of the Hydraulics Division. Proceedings of the A.S.C.E. 104(HY6): 869-885. Hey, R. D. and C. R. Thorne (1986). "Stable Channels with Mobile Gravel Beds." Journal of Hydraulic Engineering 112(8): 671-689. Hickin, E. J. (1984). "Vegetation and River Channel Dynamics." Canadian Geographer 28(2): 111-126. Hickin, E. J. and G. C. Nanson (1984). "Lateral Migration Rates of River Bends." Journal of Hydraulic Engineering 110(11): 1157-1567. Hooke, J. M. (1979). "An Analysis of the Processes of River Bank Erosion." Journal of Hydrology 42: 39-62. Howard, A. D., M. E. Keetch and C. L. Vincent (1970). "Topological and Geometrical Properties of Braided Streams." Water Resources Research 6(6): 1674-1688. Huang, H. Q. and G. C. Nanson (1997). "Vegetation and Channel Variation; a Case Study of Four Small Streams in Southeastern Australia." Geomorphology 18: 237-249. Hupp, C. R. and W. R. Osterkamp (1996). "Riparian Vegetation and Fluvial Geomorphic Processes." Geomorphology 14: 277-295. Ikeda, S. and N. Izumi (1990). "Width and Depth of Self-Formed Straight Gravel Rivers With Bank Vegetation." Water Resources Research 26(10): 2353-2364. Jackson, C. R., C. A. Sturm and J. M. Ward (2001). "Timber Harvest Impacts on Small Headwater Stream Channels in the Coast Ranges of Washington." Journal of the American Water Resources Association 37(6): 1533-1549. Jackson, W. L. and B. P. Haveren (1984). "Design for a Stable Channel in Coarse Alluvium for Riparian Zone Restoration." Water Resources Bulletin 20(5): 695-703. 179 Jarrett, R. D. (1984). "Hydraulics of High-Gradient Streams." Journal of Hydraulic Engineering 110(11): 1519-1539. Jones, J. A. and G. E. Grant (1996). "Peak Flow Responses to Clear-cutting and Roads in Small and Large Basins, Western Cascades, Oregon." Water Resources Research 32(4): 959-974. Julien, P. Y. (1994). Theoretically-Derived Regime River Width. 1994 Conference of Hydraulic Engineering, Buffalo, New York, American Society of Civil Engineers. Kellerhals, R. and D. I. Bray (1971). "Sampling Procedures for Coarse Fluvial Sediments." Journal of Hydraulic Engineering 97(8): 1165-1180. Kellerhals, R. and M. J. Miles (1996). Fluvial Geomorphology and Fish Habitat: Implications for River Restoration. Proceedings of the Second IAHR Symposium on Habitats Hydraulics. Kirkby, M. J. (1976). "Tests of the Random Network Model and its Application to Basin Hydrology." Earth Surface Processes 1: 197-212. Klohn-Crippen Consultants LTD. (1998). Bighorn Creek Integrated Watershed Restoration Project Report, Prepared for Ministry of Environment, Lands and Parks. Knighton, D. (1998). Fluvial Forms & Processes. London, Arnold. Lane, E. W. (1955). "Design of Stable Channels." Transactions of the American Society of Civil Engineers 120: 1234-1260. Lane, E. W. (1957). A Study of the Shape of Channels Formed by Natural Streams Flowing in Erodible Material. Omaha, Nebraska, U.S. Army Corps of Engineers, Missouri River Div.: 106. Lawler, D. H. (1992). Process Dominance in Bank Erosion. Lowland Floodplain Rivers. P. A. Carling and G. E. Petts: 117-143. Leopold, L. B. and T. Maddock (1953). The Hydraulic Geometry of Stream Channels and some Physiographic Implications, U.S. Geological Survey. Leopold, L. B. and M. G. Wolman (1957). River Channel Pattern: Braided, Meandering and Straight. Physiographic and Hydraulic Studies of Rivers, United States Geological Survey: 85 pg. Leopold, L. B., M. G. Wolman and J. P. Miller (1964). Fluvial Processes in Geomorphology. San Francisco, W.H. Freeman and Co. 180 Lewin, J., M. G. Mackin and M. D. Newson (1988). Regime Theory and Environmental Change - Irreconcilable Concepts? International Conference on River Regime, Wallingford, England, John Wiley & Sons, Ltd. M. Miles and Associates (1995). Deadman River Channel Stability Analysis, prepared for the federal Department of Fisheries and Oceans (DFO) and the provincial Ministry of Environment, lands and Parks (MOELP): 36. M. Miles and Associates (1997a). Bonaparte River Restoration Project Hydrotechnical Analyses, prepared for the Steelhead Society of BC: 10. M. Miles and Associates (1997b). Coldwater River Hydrotechnical Hazards Assessment, prepared for the Trans Mountain Pipeline Company Ltd.: 24. M. Miles and Associates (1998). Preliminary Assessment of the Effects of the Crest Creek and Heber River Diversions on Channel Morphology, prepared for the BC Hydro and Power Authority Co. Ltd.: 26. M. Miles and Associates (2001). Historical Analysis of Channel Stability on Bighorn Creek, prepared for Tembec Industries Inc.: 8. Mackin, J. H. (1948). "Concept of the Graded River." Bulletin of the Geological Society of America 59: 463-512. MacVicar, B. J. (1999). Application of a Rational Model of Stream Equilibrium for Predicting Channel Adjustments. Civil Engineering. Vancouver, B.C., University of British Columbia: 203. MacVicar, B. J. and R. G. Millar (1998). Abilities, Limits, and Potential in Stream Restoration. CWRA 51st Annual Conference Proceedings, Mountains to Sea: Human Interaction with the Hydrologic Cycle, Victoria, B.C. Maddock, T. (1972). Hydrologic Behaviour of Stream Channels. North American Wildlife and Natural Resources Conference, Washington, D.C. Madej, M. A., W. E. Weaver and D. K. Hagans (1994). "Analysis of Bank Erosion on the Merced River, Yosemite Valley, Yosemite National Park, California, USA." Environmental Management 18(2): 235-250. Marston, R. A. (1982). "The Geomorphic Significance of Log Steps in Forest Streams." Annals Association of American Geographers 72: 99-108. McFarlane, B. E. (2001). Retrospective Analysis of the Effects of Harvesting on Peak Flow in Southeastern British Columbia. Watershed Assessment in the Southern Interior of British Columbia, Penticton, B. C , Canada, B.C. Ministry of Forests Research Program. 181 Meidinger, D. and J. Pojar (1991). Special Report Series 6, Ecosystems of British Columbia, BC Ministry of Forests: 330. Miles, M. (pers. comm.). Air Photo Analysis. Millar, R. (2001). Riparian Logging and Channel Stability. Watershed Assessment in the Southern Interior of British Columbia, Penticton, B. C , Canada, B.C. Ministry of Forests Research Program. Millar, R. G. (1994). An Optimization Model for the Development and Response of Alluvial River Channels. Civil Engineering. Vancouver, British Columbia, University of British Columbia: 320. Millar, R. G. (2000). "Influence of Bank Vegetation on Alluvial Channel Patterns." Water Resources Research 36(4): 1109-1118. Millar, R. G. (in prep). "Theoretical Regime Equations for Mobile Gravel-bed Rivers with Stable Banks.". Millar, R. G. and M. C. Quick (1993). "Effect of Bank Stability on Geometry of Gravel Rivers." Journal of Hydraulic Engineering 119(12): 1343-1363. Millar, R. G. and M. C. Quick (1998). "Stable Width and Depth of Gravel-Bed Rivers with Cohesive Banks." Journal of Hydraulic Engineering 124(10): 1005-1013. Mollard, J. D. and J. R. Janes (1984). Airphoto Interpretation and the Canadian Landscape. Hull, Quebec, Canadian Government Publishing Centre. Montgomery, D. R. and J. M. Buffington (1997). "Channel-Reach Morphology in Mountain Drainage Basins." Geological Society of America Bulletin 109(5): 596-611. Morgan, R. P. C. and R. J. Rickson (1995). Water Erosion Control. Slope Stabilization and Erosion Control. R. P. C. Morgan and R. J. Rickson. London, E & FN Spon. Moring, J. R., G. C. Garman and D. M. Mullen (1985). The Value of Riparian Zones for Protecting Aquatic Systems: General Concerns and Recent Studies in Maine. Riparian Ecosystems and their Management: Reconciling Conflicting Uses, First North American Riparian Conference, Tucson, Arizona, USDA Forest Service. Morisawa, M. and E. LaFlure (1979). Hydraulic Geometry, Stream Equilibrium and Urbanization. Adjustments of the Fluvial System. D. D. a. W. Rhodes, G.P., Kendall/Hunt Publishing Company: 333-350. Murgatroyd, A. L. and J. L. Ternan (1983). "The Impact of Afforestation on Stream Bank Erosion and Channel Form." Earth Surface Processes and Landforms 8: 357-369. 182 Nagtegaal, D. A., B. Riddell, S. Lehmann, D. Ewart and B. Adkins (2000). Assessment of Campbell/Quinsam Chinook Salmon (Oncorhynchus tshawytscha), Fisheries and Oceans Science. Nanson, G. C. (1980). "Point Bar and Floodplain Formation of the Meandering Beatton River, Northeastern British Columbia, Canada." Sedimentology 27: 3-29. Nanson, G. C. and E. J. Hickin (1983). "Channel Migration and Incision on the Beatton River." Journal of Hydraulic Engineering 109(3): 327-337. Neill, C. R. and E. K. Yaremko (1989). Identifying Causes and Predicting Effects of Bank Erosion. National Conference of Hydraulic Engineering, New Orleans, Louisiana. Nevins, T. H. F. (1969). "River Training - the Single-thread Channel." New Zealand Engineering December: 367-373. O'Loughlin, C. and R. R. Ziemer (1982). The Importance of Root Strength and Deterioration Rates upon Edaphic Stability in Steepland Forests. .U.F.R.O. Workshop P.I.07-00 Ecology of Subalpine Zones, Oregon State University, Corvallis, Or, Forest Research Laboratory, Oregon State University. O'Loughlin, C. L. (1974). "A Study of Tree Root Strength Deterioration Following Clearfelling." Canadian Journal of Forest Research 4: 107-113. Parker, G. (1976). "On the Cause and Characteristic Scales of Meandering and Braiding in Rivers." Journal of Fluid Mechanics 76(3): 457-480. Parker, G. (1979). "Self-formed Rivers with Equilibrium Banks and Mobile Bed." Journal of the Hydraulic Division 105(9): 1185-1201. Piegay, H., M. Cuaz, E. Javelle and P. Mandier (1997). "Bank Erosion Management Based on Geomorphological, Ecological and Economic Criteria on the Galaure River, France." Regulated Rivers: Research & Management 13: 433-448. Pillipow, R. (pers. comm.). Narrowlake Creek Sediment Information. Pitlick, J. (1994). "Relation between Peak Flows, Precipitation, and Physiography for Five Mountainous Regions in the Western USA." Journal of Hydrology 158: 219-240. Pizzuto, J. E. (1984). "Bank Erodibility of Shallow Sandbed Streams." Earth Surface Processes and Landforms 9: 113-124. Powell, L. H. (1987). Hydrology and Logging in the Carnation Creek Watershed - What Have We Learned? Proceedings of the Workshop: Applying 15 Years of Carnation Creek Results, Nanaimo, B.C., Carnation Creek Steering Committee. 183 Ralph, S. C , G. C. Poole, L. L. Conquest and R. J. Naiman (1994). "Stream Channel Morphology and Woody Debris in Logged and Unlogged Basins of Western Washington." Canadian Journal of Fisheries and Aquatic Sciences 51: 37-51. Rennie, C. (1998). Bed Mobility of Gravel Rivers: Mobilization (Scour) Depth of Chum Salmon Redds, and Equilibrium Bedload Transport. Civil Engineering. Vancouver, B.C., University of British Columbia: 169. Richards, K. (1982). Rivers, Form and Process in Alluvial Channels. London, Methuen & Co. Ltd. Roberts, R. G. and M. Church (1986). "The Sediment Budget in Severely Disturbed Watersheds, Queen Charlotte Ranges, British Columbia." Canadian Journal of Forestry Research 16: 1092-1106. Rosgen, D. L. (2001). A Stream Channel Stability Assessment Methodology. Seventh Federal Interagency Sedimentation Conference, Reno, Nevada, FISC Conference. Rothacher, J. (1970). "Increases in Water Yield Following Clear-cut Logging in the Pacific Northwest." Water Resources Research 6(2): 653-658. Rowntree, K. M. and E. S. J. Dollar (1999). "Vegetation Controls on Channel Stability in the Bell River, Eastern Cape, South Africa." Earth Surface Processes and Landforms 24: 127-134. Scherer, R. (2001). Effects of Changes in Forest Cover on Streamflow: A Literature Review. Watershed Assessment in the Southern Interior of British Columbia, Penticton, B. C , Canada, B.C. Ministry of Forests Research Program. Schumm, S. A. (1969). "River Metamorphosis." Journal of the Hydraulics Division, Proceedings of the A.S.C.E. 95(HY1): 255-273. Simon, A. (1989). "A Model of Channel Response in Disturbed Alluvial Channels." Earth Surface Processes and Landforms 14(11-26). Simon, A. (1994). Width Adjustment: Relative Dominance in Unstable Alluvial Streams. 1994 Conference of Hydraulic Engineering, Buffalo, New York, American Society of Civil Engineers. Simon, A. (1995). "Adjustment and Recovery of Unstable Alluvial Channels: Identification and Approaches for Engineering Management." Earth Surface Processes and Landforms 20: 611-628. Simon, A. and A. J. C. Collison (2002). "Quantifying the Mechanical and Hydrologic Effects of Riparian Vegetation on Streambank Stability." Earth Surface Processes and Landforms 27: 527-546. 184 Simon, A. and P. W. Downs (1995). "An Interdisciplinary Approach to Evaluation of Potential Instability in Alluvial Channels." Geomorphology 12: 215-232. Simon, A. and C. R. Thorne (1996). "Channel Adjustment of an Unstable Coarse-Grained Alluvial Stream: Opposing Trends of Boundary and Critical Shear Stress, and the Applicability of Extremal Hypothesis." Earth Surface Processes and Landforms 21: 155-180. Smith, D. G. (1976). "Effect of Vegetation on Lateral Migration of Anastomosed Channels of Glacier Meltwater River." Geological Society of America Bulletin 87: 857-860. Sotir, R. B. and N. R. Nunnally (1995). Soil Bioengineering for Stream Restoration. First International Conference of Water Resources Engineering, San Antonio, Texas, American Society of Civil Engineers. Steinblums, I. J., H. A. Froehlich and J. K. Lyons (1984). "Designing Stable Buffer Strips for Stream Protection." Journal of Forestry 82: 49-52. Stevens, M. A., D. B. Simons and E. V. Richardson (1975). "Nonequilibrium River Form." Journal of the Hydraulics Division. Proceedings of the ASCE 101(HY5): 557-566. Struiksma, N. and G. J. Klaassen (1988). On the Threshold Between Meandering and Braiding. International Conference on River Regime, Wallingford, England, John Wiley & Sons, Ltd. Styczen, M. E. and R. P. C. Morgan (1995). Engineering Properties of Vegetation. Slope Stabilization and Erosion Control. R. P. C. Morgan and R. J. Rickson. London, E & FN Spon. Thomas, R. B. and W. F. Megahan (1998). "Peak Flow Responses to Clear-cutting and Roads in Small and Large Basins, Western Cascades, Oregon: A Second Opinion." Water Resources Research 34(12): 3393-3403. Thorne, C. R. (1982). Processes and Mechanisms of River Bank Erosion. Gravel-bed Rivers. R. D. Hey, Bathurst, J.C. and Thorne, C.R., John Wiley & Sons: 227-271. Thorne, C. R. (1990). Effects of Vegetation on Riverbank Erosion and Stability. Vegetation and Erosion: Processes and Environments. J. B. Thornes. Chichester, England, John Wiley & Sons: 125-160. Thorne, C. R. and J. Lewin (1979). Bank Processes, Bed Material Movement and Planform Development in a Meandering River. Adjustments of the Fluvial System. D. D. Rhodes and G. P. Williams. Dubuque, Iowa, Kendall/Hunt Publishing Co.: 117-137. 185 Thome, C. R. and A. M. Osman (1988). The Influence of Bank Stability on Regime Geometry of Natural Channels. International Conference on River Regime, Wallingford, England, John Wiley & Sons, Ltd. Thorne, C. R. and A. M. Osman (1988). "Riverbank Stability Analysis II: Applications." Journal of Hydraulic Engineering 114(2): 151-172. Thorne, C. R. and N. K. Tovey (1981). "Stability of Composite River Banks." Earth Surface Processes and Landforms 6: 469-484. Thorne, S. D. and D. J. Furbish (1995). "Influences of Coarse Bank Roughness on Flow within a Sharply Curved River Bend." Geomorphology 12: 241-257. Toews, D. A. and M. K. Moore (1982). A First Look at Logging Effects on the Hydrologic Regime of Carnation Creek Experimental Watershed. Proceedings of the Carnation Creek Workshop, A 10 Year Review, Malaspina College, Nanaimo, B.C. Trimble, S. W. (1997). "Stream Channel Erosion and Change Resulting from Riparian Forests." Geology 25(5): 467-469. van den Berg, J. H. (1995). "Prediction of Alluvial Channel Pattern of Perennial Rivers." Geomorphology 12: 259-279. Vidal, H. (1969). "The Principal of Reinforced Earth." Highway Research Record 282: 1-16. White, C. M. (1940). "Equilibrium of Grains on the Bed of a Stream." Proceedings of the Royal Society of London. Series A 174: 322-338. Williams, G. P. (1978). "Bank-Full Discharge of Rivers." Water Resources Research 14(6): 1141-1154. Wilson, A. (2001). Predicting Channel Change in Narrowlake Creek in the Central Interior: A Tool for Watershed Protection and Restoration. Watershed Restoration Technical Bulletin - Streamlines. 5: 8-12. Wilson, A. (pers. comm.). Narrowlake Creek Information. Wilson, A., B. Nesbit and R. Pillipow (2001). Riparian Restoration at the Narrowlake Creek Demonstration Watershed. Watershed Restoration Technical Bulletin - Streamline. 6: 24-29. Wilson, C , P. Hairsine and I. Rutherfurd (1995). An Australian Perspective on the Roles of Riparian Vegetation in River Restoration. First International Conference of Water Resources Engineering, San Antonio, Texas, American Society of Civil Engineers. 186 Wolman, M. G. (1954). "A Method of Sampling Coarse River-Bed Material." Transactions of the American Geophysical Union 35(6): 951-956. Wolman, M. G. and J. P. Miller (1960). "Magnitude and Frequency of Forces in Geomorphic Processes." Journal of Geology 68: 54-74. Wu, T. H. (1995). Slope Stabilization. Slope Stabilization and Erosion Control. R. P. C. Morgan and R. J. Rickson. London, E & FN Spon. Yang, C. T. and C. C. S. Song (1979). Dynamic Adjustments of Alluvial Channels. Adjustments of the Fluvial System. D. D. Rhodes and G. P. Williams, Kendall/Hunt Publishing Company: 55-67. Ziemer, R. R. (1981). "Storm Flow Response to Road Building and Partial Cutting In Small Streams of Northern California." Water Resources Research 17(4): 907-917. Zimmerman, R. C , J. C. Goodlett and G. H. Comer (1967). "The Influence of Vegetation on Channel Form of Small Streams." International Association of Scientific Hydrology: Symposium River Morphology Publication 75: 255-275. 187 APPENDIX A - ANNUAL MAXIMUM SERIES Appendix A is a collection of the discharge data used in this study. The Table A l is a summary of the regional hydrological analyses. The charts on the proceeding pages are the non-translated annual peak discharges for the gauge data available in the BCSI and HYDAT data sets (Coulson and Obedkoff 1998; Environment Canada 2000). Table A2 is the gauge information for the regional analysis. Peak and Mean Flow Data Stream Reach # Peak Q (mJ/s) Mean Q (mVs) Big Horn 4 N/A N/A Bonaparte 1 89 32 Coldwater 1 123 77 Coldwater 2 66 31 Coldwater 3 61 28 Deadman 1 93 22 Deadman 2 94 22 Elk 1 188 124 Elk 2 92 61 Eve 1 N/A N/A Narrowlake TR3 N/A N/A Salmon 1 1170 603 Slesse D 171 88 Tahsis 1 N/A N/A West Kettle 1 183 127 Table A l Summary of discharge data translated from hydrological analysis. The values were estimated from gauge data using Eaton et. al. (2002) technique 188 Big Horn Creek Flow Data Cabin Creek near the mouth Instantaneous Peak Flow Record 45 40 — 35 30 o 25 OJ <5 20 j= S 15 Q 10 Q = 24rriVs Year Grave Creek at the mouth Instantaneous Peak Row Record O C \ j T r c D e O O C \ ] T r c D O O O C \ l T t < 0 0 0 r ^ r ^ l " — h - N - O O O O C O C 0 0 0 0 3 0 ) 0 ) 0 > 0 > 0 ) 0 ) C n O ) 0 ) 0 ] 0 ) 0 ) 0 ) C T G ] 0 3 0 ) 0 ) 0 ) Year Couldrey Creek in Lot 9380 Instantaneous Peak Row Record 40 J30 .25)1 )20 | 15 i 1 ° 5 Q = 18m7s ^ d> ^ J> # # J> J> # & J-Year 45 J2 35 I 3 0 o 25 o 15 I* 5 Howell Creek above Cabin Creek Instantaneous Peak How Record U=ZJ rifts 0 0 f f l O ^ « P l \ t i n t 0 N 0 3 0 ) O T - f M O ^ t l f ) t D r ^ h - O O C O C O G O O O O O O O O O C O O O O O O O O O O ) CTCTOCTCTCTCTOOIOOCTCTCTCTCTCTCTO*) Yea-Line Creek at the mouth Instantaneous Peak Flow Record 50 S>30 5 o £ 20 Q * 21 m3/s 0 ) 0 1 0 ) 0 ) 0 ) 0 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 0 ) 0 ) 189 Bonaparte River Flow Data B o n a p a r t e R i v e r b e l o w C a c h e C r e e k I n s t a n t a n e o u s P e a k F l o w R e c o r d 100 90 ; W n Q = 32 m3/s I'l — ' 1 .jUIJLJLrL . . -rt- - ! - | - : - i - - AM I-i-i l l ] i l l l n n l l C n A u ) u > 0 > O ) O ) 0 ) O ) 0 > O > 0 > O ) 0 > O ) 0 > O > 0 > O ) O ) Year Coldwater River Flow Data 190 Deadman River Flow Data Deadman River above Cr iss Creek Instantaneous Peak Flow Record Q 2 0 10 0 j I i Q i 15 m 3 / s in JlJU c n c n o > < n o ) 0 > c n o > a > c n c n o ) 0 ) C Q c n c n o ) Year Elk River Flow Data Elk R iver a b o v e C a m p b e l l Lake Instantaneous Peak Flow R e c o r d 2 0 0 1 8 0 1 6 0 J . 1 2 0 o> E> 1 0 0 ro o 80 W 5 6 0 4 0 2 0 Q = 124 m 3 / s 1 9 9 2 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 8 1 9 9 9 2 0 0 0 Year Eve River Flow Data Kokish River below Bonanza River Instantaneous Peak Flow Record 180 160 _ 140 "g 120 IT 100 ca ra 80 0 n Q = 131 m3/s i I — 8. i 1 i i S i i — : — I i I cn o C N ' C O ~i co r-- CO CD O l O t O ( 0 < D C O ( D < 0 ( D C D < O C O r * -c n c n c r i c n c D C D o o o o c n o Year Salmon River above Memekay River Instantaneous Peak Flow Record Tsitika River below Catherine Creek Instantaneous Peak Flow Record 1200 „ 1000 "g 800 0 -1 Q = 450 m3/s - - - - - - - -- - - - 4T- - - - - —j|J—m-m h - o ^ T - c o m r ^ O T - c o m r ^ o ) r*- h - O O C O O O O O O O C T ^ C D C D C r j C T ) 0 ) 0 0 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 0 ) 0 ) Year Narrowlake Creek Flow Data Chuchinka Creek near the mouth Instantaneous Peak Flow Record 90 80 70 „ ^ 6 0 E IT so jS 40 u £ 30 a 20 10 0 0*53rrr7s Year McKinley Creek below McKfnley Lake Instantaneous Peak Flow Record To 4 0 "s. a 30 Q = 25 m3/s | llill 1 1 1 I 1 i i Moffat Creek near Horsefly Instantaneous Peak Row Record Little Swift River at the mouth Instantaneous Peak Flow Record 70 60 _ 50 £• 40 co oo a u> CT> o) m 0)010)010)0 VUllow River above Hay Creek Instantaneous Peak Row Record O f O ( 0 0 > C N l f ) C O T - T f h - 0 C O C O C O C D r - h - r ^ - C O C O O O O ) C J ) O ) O > O ) G ) C J ) 0 ) 0 ) 0 ) O ) 0 ) OJ 0> Ol Year 193 Salmon River Flow Data S a l m o n R i v e r n e a r S a y w a r d I n s t a n t a n e o u s P e a k F l o w R e c o r d Slesse Creek Flow Data 250 200 §. 150 100 Slesse Creek Instantaneous Peak Flow Record Y e a r Tahsis River Flow Data Gold River below Ucona River Instantaneous Peak Flow Record Year Zebal los River near Zeballos West Kettle River Flow Data West Kett le River below Carmin Creek Ins tan taneous Peak Flow Record 160 n 140 «• 120 " i 100 -H CD O) 80 nl •5 60 5 40 20 Q = 104 m3/s Year Case Study Gauge Location Gauge # Qmaf(m3/s) Area (km2) Est. k Mean k Bonaparte 1 Bonaparte below Cache Creek 08LF002 32 5020 0.054 N/A Coldwater 1 Coldwater at Merritt 08LG010 92 914 0.553 N/A Coldwater 2, 3 Coldwater near Brookmere 08LG048 75 311 1.01 N/A Deadman 1, 2 Deadman above Criss Creek 08LF027 15 870 0.094 N/A Elk 1,2 Elk above Campbell L 08HD018 123 132 3.16 N/A Eve 1 Kokish R. below Bonanza R. 08HF003 131 269 1.97 Eve 1 Salmon R. above Memekay R. 08HD007 325 437 3.4 Eve 1 Tsitika R. below Catherine Cr. 08HF004 450 359 5.46 3.6 Salmon 1 Salmon near Sayward 08HD006 873 1200 4.28 N/A Tahsis 1 Gold R. below Ucona 08HC001 1519 1010 8.48 Tahsis 1 Zeballos R. near Zeballos 08HE006 584 181 11.84 10.2 Big Horn 4 Cabin Cr. near mouth 08NP004 24 93.3 0.78 Big Horn 4 Couldrey Cr. lot 9380 08NP002 18 118 0.5 Big Horn 4 Howell above Cabin 08NP003 20 145 0.48 Big Horn 4 Grave Cr. at mouth 08NK019 8 75.8 0.31 Big Horn 4 Line Cr. at mouth 08NK022 21 138 0.52 0.52 Narrowlake TR3 Willow R. above Hay 08KD006 243 2810 0.63 Narrowlake TR3 McKinley Cr. below McKinley Lake 08KH020 25 426 0.27 Narrowlake TR3 Moffat Cr. near Horsefly 08KH019 25 539 0.22 Narrowlake TR3 Chuchinka Cr. near mouth 07EE009 53 311 0.72 Narrowlake TR3 Little Swift R. 08KE024 30 122 0.82 0.59 (reg.) Slesse D Slesse near Vedder Crossing 08MH103 92 166 1.99 N/A West Kettle 1 W. K. R. below Carmi Cr. 08NN022 104 1170 0.52 N/A Table A2 Discharge Summary Data (mean k = value estimated from regional analysis) 197 t/5 X PH o o H o s OH 2 I CQ X hH Q W PH PH c« u to o JS PH u Q. CO X O CO o ra ^ u Q. CO E " CD O X *: "o „ £. x: g .SH _>» CD LL. X Ot lc C _>* (D U_ X CO < u. g "O* CO < or E E Q. CO CD E S E CO w O) CD CD N Z CO O = O t CL Z o o o o o o o o o o o o o o o o o o o O O o o o o O o o o o o o o o o o o o o o o o o o o o o O o o o o O o IO o Tf *— CO CM CM *— in CO CO o r- o o Tf CM r- o o CO CO CM o CD Tf CO to 0 0 CO CO m r-- ro CO CM r-~ CO co CO o 0 0 r*. o Ot Tf CM CD cn CO cn CO CM CO CM Tf CN CO T~ CM ' LO CO CO I O o m O o ^ Tf cn o CO CO CO CM CO _ OJ Tf CM LO O) r— O) Tf CO CO o CO IO o cn CO CO CO o LO LO CO o CO T— T— LO T— CM CO CO O o l-~ r-- CO .— cn Tf o CO .— CM T— CN o LO r- CO o f- o cn Tf CM CO cn O CO r- T— o CD .— CO CO LO CO CD CO LO r- CO CO CN CO co CO cn 0 0 r— o CO Tf CM CD T— cn CO CD CO CM Tf CM CO T _ CN T _ m CO CO CM CN I O m CO CO co CO o o Ot O) o o o ^ CO CO CO co CO co CD CO m cn m m CO CO CO CO CM CM CM CN cn cn T— CJ) Tf Tf Tf CO CO f~ , J *— T— T— T— CO CO CO CO o o m in GO 0 0 CO m Tf Tf Tf m m CO CM CN •* Tf r- t- o o o o LO LO CM CN CO CO CO CM T _ ' T _ o o o O o o o o o o o o o o o o o o in in m m m m in LO o o o CO CO m to o o LO m m m m m LO m m m r~ r- r- CM CM CM o o o m in CO CO Tf LO LO CO CO co CO CM CM CM CO Tf Tf Tf Tf TT Tf CN CN CO CO CO CO CM o o CN Tf o Tf o o o CO CO CO o in CO o o CO CO Tf CD oo CM CD o Tf CO CO T- CO CO CO CO CO CO CO h~ CO CO CO CD CN Tf CO CO cn CO CO in o> cn Tf Tf O) Tf Tf cn Tf o> •f Tf cn CM CD CO co cn CM o> CO CO o CM o CM o o o cn h~ cn o m cn o LO O) Tf Tf T— 0 0 CM o Tf CO CO m CO CO CO CO CO CO m in in CO TT m CO Tf r*- h~ in m Tf m 0 0 CD IO o o o o O o o o o o o o o o o o o o o O o o o o o o o o o o o O o o o o o o CM o o o o o o o O o o o o o o o m o o o m o LO o o o in O m o cn LO o o o m o o o o o o o CM cn o cn o o o o o o cn cn cn o Tf cn o m CO Tf CO CO cn r*- o CO CM 1 CN T— CN CM CM *- T ~ T— CM T~ CM CM CM , — T - CN CN m ci Tf m Tf cn in o T- CM m co co o co o CM CO r-CM f-CM CM LO m m 0 0 CO CO CD CO CO CO m CO CM CO cn m CO cn CO m cn LO o m r~ O) CO cn CO cn CO LO r- cn CO o CO o CM o CO in CO LO CM o CM o CM o LO r— r- o CM o CN • CO CO o o CO ci o CO O o CO ci o CO ci O CO b o CO d co o d d CM O d d CO m d d o d ci ci ci d CO CO T^  Tf | Tf I CD • "> , O CO CM O ) CO cn LO co s n o CO CO Q o in ° CO i-CN •— co Tf oo cn Tf <-Tf co o o <=> r-O co ° 2 CQ ^ 5 o CO CD cn O oo S « cn O CM O O CO CQ CO CO CO CO o o o ct m ci co ° co Tf O • -co CO o o o _ -> in o co ° O Tf O Tf O Tf O Is- O I"- O IS-m ^ m *~ IYI T~ o O o O o O CO CQ CO CQ CO CQ O in O CQ CM 5 o CO CQ CO Tf o m £ °> cn cn o S CQ CM Q O O Tf CQ CQ < Tf o m S °> o ° co CQ CM o O O Tf CQ CQ < O CM m O) o co o CO o CO o cn m f- o m r~ o CO , CO ^_ CO m CM Tf ^ CO o CO cn in o m o m o in o LO in CO cn in CO cn 0 0 m cn Tf at in LO CD o cn cn cn o cn o cn o cn o cn ai cn at cn cn cn cn cn cn cn at at cn cn cn > CM T~ T" O t^  CO f~ o CN r- co O) CO Tf CM o at at CO cn o CM o in CO Tf in m CD CO o Tf r~ cn T— T— CO T— LO CO CM m co CO Tf Tf Tf T— CO CO Tf m ja T - T - CM T _ n Tf CM in at ,— CO CO o CJ) T- o o CO CO CM CO O T- T- CN m cn m O cn at LO cn Tf CO CO CO CO T- CQ CN CQ LO T— .— O O O O O O O o o CQ CQ CO ca ca ca CO CO ca o X o CQ 5 o O CO E T> CO CD a CD > LU O E co CO CD 198 Reach Aerial Photographs The straight white lines represent reach borders. If no lines are present, than the edges of the photographs represent the borders. Figure BI Bonaparte River 1959 BC2588:47 (approx. scale 1:8050) Figure B2 Bonaparte River 1995 BCC95019:008 (approx. scale 1:9200) 199 Figure B 3 Coldwater River reach 1 1953 BC1745:92 (approx. scale 1:12200) The second white line going downstream (left - right) is the reach end for the field reach, but during measuring the third line was the last measurement point. The scale for the 1953 photo was adjusted based on congruent measurements on the 1953 photo and 2000 photo. The new adjusted scale for the 1953 photo is identical to the 2000 photo. 200 Figure B4 Coldwater River reach 1 2000 BCC00036:170 (approx. scale 1:12200) 201 Figure B5 Coldwater River reach 2 1953 BC1746:6 (approx. scale 1:8600) The scale between the 1953 photo and the 2000 photo were determined to be congruent contrary to the calculation shown in the table. 202 Figure B6 Coldwater River reach 2 2000 BCC00033:77 (approx. scale 1:8600) 203 Figure B7 Coldwater River reach 3 1953 BC1746:8 (approx. scale 1:11000) The scale between the 1953 photo and the 2000 photo were determined to be congruent contrary to the calculation shown in the table. Figure B8 Coldwater River reach 3 2000 BCC00033:19 (approx. scale 1:11000) 204 Figure B9 Deadman River 1953 BC2652:82 (approx. scale 1:16000) -reach 1 upstream (left reach) 205 206 Figure B l l Elk River reach 1 1930 A4059:39 (approx. scale 1: 10000) Figure B12 Elk River reach 1 1995 BCB95046:110 (approx. scale 1:9350) Figure B14 Elk River reach 2 1995 BCB95047:128 (approx. scale 1:9200) 208 Figure B15 Eve River 1953 BC1711:41 (approx. scale 1: 16000) Note the "wandering" character of the pristine river. Figure B17 Eve River 1996 BCB96094:30 (approx. scale 1:16000) Figure B18 Salmon River 1946 BC252:13 (approx. scale 1:21200) 212 Figure B20 Tahsis River 1954 BC1856:45 (approx. scale 1:9780) Figure B22 Big Horn Creek reach 4 1962 BC4087:107 (approx. scale 1:37500) 215 Figure B23 Big Horn Creek reach 4 2000 BCC00084:130 (approx. scale 1:37000) Figure B24 West Kettle River 1951 BC1302:55 (approx. scale 1:31400) 216 APPENDIX C - WIDTH MEASUREMENTS Big Horn Width Measurements Air Photo Air Photo scale = 15000 scale = 18500 1962 2000 W(mm) W(m) W(mm) W(m) 0.65 9.8 2.12 39.2 0.53 8.0 1.97 36.4 0.46 6.9 1.95 36.1 0.5 7.5 2.15 39.8 1.01 15.2 4.36 80.7 0.54 8.1 3.48 64.4 1.99 29.9 4.96 91.8 1.21 18.2 1.72 31.8 0.85 12.8 3.07 56.8 0.96 14.4 3.53 65.3 1.29 19.4 2.18 40.3 0.87 13.1 1.34 24.8 1.62 24.3 1.04 19.2 1.03 15.5 2.79 51.6 0.86 12.9 1.58 29.2 1.7 25.5 1.32 24.4 0.61 9.2 1.87 34.6 1 15.0 1.78 32.9 0.83 12.5 1.38 25.5 2.38 35.7 2.77 51.2 1.5 22.5 3.91 72.3 1.9 28.5 1.2 22.2 1.29 19.4 2.2 40.7 1.54 23.1 2.63 48.7 1.42 21.3 1.74 32.2 1.25 18.8 0.94 17.4 0.9 13.5 0.79 14.6 2.26 33.9 0.84 15.5 0.76 11.4 0.83 15.4 0.8 12.0 1.82 33.7 3.2 48.0 0.85 15.7 average = 18.2 average = 38.9 1962 Mean Standard Error 18.2 1.7 2000 Mean Standard Error 38.9 3.6 1959 L (mm) L (m) channel 205.6 3084.0 valley 171.1 2566.5 sinuosity 1.2 1995 L (mm) L (m) channel 161.6 2989.6 valley 143.85 2661.2 sinuosity 1.1 217 2 CD CU CD V) O LO CO . • . • CO r— r» CM CM CM II CD O) CO CD > CO o in un oo od o CM LU CO c "a ro 5 aj co 2 55 ^ co co ^ co T-- I Tf to o o o o o CO II a) co o CM OT O T - CO in OT _ M ^ (N -cf OJ CD g ^ T - C M T - C M T - T -II CU CT) CO in E CT) f= OT o m ^ (O LO (D CO T — CD CO O 6 ^ d r ^ a> > co in E S O) c to o o) Sco s , CM — 2 > CO o C co OT O Tf CD CM LU CO ro £ 0 CO 2 CO a CU a cu k. s so es cu o o CL O O T — to II 0) CO o to > CU r es a es ca o CQ £ 00 LO CO ^ CO CO in co o o co co OT m . OT iz. ;= CD T- K UJ UJ E co CM . m co ^ o OT OT CO o o o II CD CJ) 5 CD > CO - p - N in —' CM CM co E in c: M O co e, • CD >. c ^ £ ro ro > o co o c TJ- CO o CO T -• LU CO ro £ CD CO OD 218 £ o o in o o _j / r\J O °> ^  ^ CD CN m T3 CO > O II CO co <u > CO . oo o M- CD l - ~ CD CO q o cj co CD T-UJ TJ CO C T3 ro c co ro 5 55 o o CO CO £ o» Tt N; cDNcnioinco^rco csi oi co c\ii-:a)'<i:r^cot-'co g CO N r— NSMIOUICMWID O F CO CO O c 0 0 ° o S o CD CO a s fa s « -a fa fa «J 2 "o U o II o a> -<= co •. o < CO or II CO O tf)^<0<OCM'-CO'<t_S0 ) g E u i n q o i o i o i m T - ^ T - J , CM ^ ^ " ^ ^ CO CO ^ CO CO T— Q S > CO co = CO CO o o o CO c o i— o M o a) CO . c CO LO a. o in < c j i n q q r o ^ ^ i i q N T - i o n t q o i D C i i N i n ^ " -r^  ^ c s i c o c s i c n c M u S c d o i c d i ^ c b W r-^  co ^ T J - C \ I C \ I C N C \ I C \ I C O * - C \ I T I / T - C \ I C \ 1 T - C \ I C \ I C \ | C N iMnin iO ' -coo iun-ENT-r-(<)(M(MSON CO CO CO ^* CD LO O) LO co O T-^  ^ II CO CO > ~ ° LO CO P CO CO Ifl r ^ ^ o> £. j^- o *" _ ! CO LO CO = in CO CO ^ •* CO ^ co LU T3 CO C T3 co ca 2 co O) CD CO LO LU "P ca CO c co co 219 C N P O L O L O O o ; n CD O) in m c\i > co CN CN •sr 33 LL. T3 CD £ CD CO _Q O II CU C D CD i CD > m co CM C D •<fr CO CO -p ^ ^ _ J T J - CM CM O O CM LU CO C X I CO C © ro 1> CO o o CM g S T - c q s o N i O ' c f ^COOr^CNCNOJcNTf > C M C O C O T J - C O C M ' > S - C O o co r-~ o • ? w ro o § L O CM fc CM T -o I" o £ - C CD < CD o p L O L O - . c o c D ^ r - - g, Q ^ _ : , _ : CM ~; _: ~; ir CM CM CM ro > o co o 3 C 'co CM •st r--CM co o o o CM LU "E CO ro S d) CD S CO o o CM h-£r^r~--r^LOCT)Tt-<};CO CD ^ N N o i s c o c v i c x i N r-^  > T - C O C D C M T - C M C M C M CN r~- CD _ Csi -p r-- L O fc -<r co co o II o £ - C CD Q . O to m E N ^ to o> _• — • co <9 » II _ CD ^ f- CJ) > CD co ^ 30 2 c l~- o S. CN CM CD <= ID > CD o co o C 'co co L O O ) co r~-CD "tf L O CN LU CM o CD a: CD C XI ro £ (1) CD S CO o E 9 w ^ h o "—' 00 Cf) CJ) O CO CM £ — - - -CO 32 CD T3 OJ c! OJ (0 - Q O OJ O ) CO cu > CO t m to ' io co —I Tf CM CO CO CM r«-cd T-* LU 73 i CO co c 2 W o o to CO C C M C M T - C O S C O N -^cbcxDT^incoTrcM > r r C M T - r t \ | r T-o E 0 0 f^ -O r Ol o £ N LO CM , CM i -O -4—* o sz a. n 0) CO o to o E CM ~> CO CD CO N Lf) CM O) CO Tf N O O T-^  O CU CD CO L_ CU > CO £ ro ro > o to o 3 C CO cd CD LO LU co c "a ro £ a ro S CO o o LO CD EfCMOTl^CMCO^LO ^CD CD OO CN m ^ Tf LO -~- O CO E co! N CO CD —I CM CM O -*—' o II a> SZ CO Q. O v. W CO _sz o CO cu on £2 E oo oo lo g co cn CM CO IT) o o II a> o> co <u > ro CO C CO CM Lo g r- cm O) 3 K CQ 2 1 5 ro ro > o CO o 3 C co CO TT co r-LO T~ LU •o ro c "2 ro £ CU CO S CO 221 CN O O CN {= CO lO ^ CO LO > CN CN CD LO CD CD CN r~- CD CO CN T3 0 CD CO -O o II 0 ro CD > ro 00 T-00 h~ CD CM CN £ CD LO ' CD CO —I CN CN 3 O CM LO O L U ro CD C cu ro ^ CO o o 00 p O 00 CO CD CO LO CO CO *— CO Cji CD O) N CO LO Ol > T- T- CM T - T - C M T - T -o o o CM CO CD '. o CO II 0 ro o co O ^ CO •<- 00 CM _ O T- T-II 0 CM CD a) CO 00 co > ro ro > o co o c co co co co h-o o o CM LU ro C T3 ro £ 0 ro S CO o o o CD P CD CO CM CO > CM t-O Tt CO O) N o o CD K -p- CO CD •S 1 ^ CD ^ CD TJ-—I CO CO II 0 ro o CO CO to CO I ^ CO E ^ oo d CD o 00 CM r< o CO II 0 CD ro 0 > ro cn in CO C CD LO p CO 00 s ^ ^ , CN CN £ ro 2 > CO o 3 C co LO CD CO CD CM CO ^ LU TD s ro C TD ro £ 0 cc ^ CO CM O O CM E LO O CO * — T f C\i CO > CM CO CO °- I- °-CO * N T f CO = S LL CO CU CO CO 1_ CU > CO - CO CO £ CM CO CM 00 —I T f CO CO CO CO LO CO T f CO LU CO c -a ro £ S CO o o 00 I'-ll o o ® sz ro A" 8 CT CM CO CD S CO s > CM CM T-CM CO T f CO CO T- T-T—• T f T f 00 CD CM CM T f CM CM O O O CM CO E io LO d) ^ T— T— O CO CO CM T f CO 00 T f LO CM T~* II CD CO ro i_ cu > ro o E £ co o E. — CM CO CM CM £ IT £ ro £ > o CO o 3 C to T- CO CM CO CO CM CM LU CO C "a ro £ a J5 S CO o o sz 0-o o o CO II 0) ro o CO CM _c o ro CD , oo CO O T f CM CO O LO T f CO CM O CO LO LO CM T f T f CO LO CO l T- LO T-E co CM ^ CD CD T - S CO m u/ 00 CD CO *R 00 o d d CO II CD CO CO 1_ CD > CO •poo o o co —I CO CM g | CO 00 £ c o N CD CO = SZ CO o > CO o 3 c co O T-o o LU "2 ro c T3 ro £ a? ro 223 CNJ o o CM £ CO ^ O CD > r- co ro l>! 10 => CM ^ ro 5! co X J cu £ cu CO O cu 2 cu > co -p ro CD S. CD O CO CD —I in r-r-r-^  csi co •<-o o co II a) 8 f£rocococpincocNc?r^ J5 £ CD co 92 E m cn CM CM cD^rmco^gjO^co ll ro co ro g £ g> ^ °> " ^ iri S 2 ™ cMCNi^r--^? J > CO ro § ro t co CM ™ CM O O CM CO O C LU "2 CO T J CZ ns CO CM CM CD CM iri CO O O O co co o o JZ 0. r- T • iri O T- CO CM CM iri cii S CM co s co m h- 0)~ — rocOCOCOCO £ CO CM ro ro CO CM CO CM T CO CM (D CO N ^ S S O ) CM CM CM CM x-' CM II CU O) ca i_ cu > ca -p- co o c in iri s E fi in ro 3 ^ co o c co LU "2 CD T J CZ ro CD •>- o iri iri co lO LU T J CO T J CZ ro 55 a a u s « TS S-•2 3 o o o o CM II 0) s o ro cu OH cz o o co oo g CM ^  •q- ^ o o o t o c M i f o * o c i ics i^cNiO ' r^^r 'cs ico CM CMCMCMT-T-CMCOTJ-CM ,_ CO •<-co E ' - ° ! o * T : ( i 1 ' ! i n o N v : O) ^  ^  O ,J ,r_T-T-Ocj.rJ.rJCM o i~^  o CO CM in c co >> cu CO o C co r-co r-co csi CM LU "2 ca T J CZ ro to 224 C M O O C M (O (M lO O) N o od co co C O C M C M C M C M T J .32 §5 LL. CO .Q O 3 O » 3 o o C M r--co o » • : 8 11 CD o> CD > CO f= in co CT) co xj- o i f cd iri C M ir i > C O C O C O C O C M C O i n co CO i- 1 CM CO C M co in r- co 10 m co co 01 N co co in CD O) 2 cu > CO CO (O O) * T - o • * o ri eo N iri cd to o i " C M C M C M ^ C O C M C M C O m i s M O CM N N S C J CO o o o •>-' o CU 2 CU co • p CM O) £ N O) — I C O C M o> e 1^  ,x 5 > £ M CD ^ T- Ol — I CO C M l>. LO _ E a co g CO O) 0 0 r-~' 2 > = CO o c 'co 10 o Z J CO C M co" 1-: C M LU CO c -a w £ cu w 2 co l>- C M O -vf C O LU CO cu co 2 55 3 g LU "2 CO TJ 1= CO 55 S > If! t J; ? i>. o C M II 0 } 8 o CO cu cc C O 0 0 T J - C O o 0 0 iri r-i cri o i C M co ,_ CO , -2 E N 0 5 ~ 0 0 o §3 3 «> 0 0 r-11 cu a) 2 cu CO t T- o in e o> £ m 2 > CO o =1 co 5 co a> o g ill CO C TJ o iS 2 co 225 C M O O C M £ CO 00 O 00 CO t— r--. o o' > OO S IO CD CD o o CO T-OJ CO CD 5 T3 CD CD ii £ CO O cu C O CO cu > CO CM Q Q CM LU CO Cl) CO 2 CO ^ to CO E •* co - O CO I CO o o o o CO E <* ^ CD > CO T— CM co r*» CM O O i -CM i - T- T-lO CO CO CO ^ T5 § O) CO CD CO ^ ^ CO Tf CO to P CO CO $ £ m o o) £ r- oo , CO CD w Tf O T-II ° 5 1 s .n CQ 0_ CO OJ CD CO CM CO N CD LO CO 1- CO O CM O) O CM CO CD N CD LO CO Tf N CM TfCO O) CO cu > CO 5 « 2 > CO o c 00 I C O loj ill " D u. CO ro £ « ro 2 CO I— OJ E CM CD Tf CO —I CD 00 o o o CO o S P~ CM 0 5 . 0 0 0 0 0 C P C ? C M ^ C O C M TT r ' x g T - c o N ^ S i c o s c o r - o> > O J ° C O C D N J 5 2 C O ° C O O ) X-TT p Tf- f\j 0 0 E T- co IT" co cb OJ IO CO _ CM CM co — in ™ Tf Tf 0 5 00 W II CD in I- SJ tO CM fO CD Tf" eg > CO C CM OO 3 co Tf , IO LO CD CO o c CO s m CM T- LO I— 00 OJ o LU CO c "a <" £ CD CO 2 55 a C U a C U u a Vi es C U T3 la C U & C U o o o to II _ CD I s x: to D. E IO Tf CO N T f i - T f l N l N t f O m co c\i od io iri o in d cd N iri i— > m Tf co mcMcoiococMcoinTf . ^ in in 8 o £ i co in m E m co n . Tf •r- ^ C O C M OO OJ 00 CO OJ to Tf in co Tf co r--CO T™" CM CO T-" T"" T- Tf CM Tf Tf CO II CD C O CD > CO OJ o co E m c . . 3 1 S to o> T- , CO Tf CO S = CO £ ro o ro > 3 o •£ CO in to in to Tf Tf LU "2 CO "a c CO CO 226 Salmon River Width Measurements channel valley Air Photo Air Photo scale = 21200 scale = 16800 1946 1991 W(mm) W(m) W(mm) W(m) 5.7 120.8 7.55 126.8 5.6 118.7 7.61 127.8 3.5 74.2 9.29 156.1 6 127.2 13.28 223.1 9 190.8 14.29 240.1 11.5 243.8 16.5 277.2 12.2 258.6 11.44 192.2 6.2 131.4 14.21 238.7 7.1 150.5 17.35 291.5 6.6 139.9 10.28 172.7 5.5 116.6 14.1 236.9 5.2 110.2 8.47 142.3 7 148.4 6.8 114.2 8 169.6 10.22 171.7 4.2 89.0 10.92 183.5 5.9 125.1 11.14 187.2 5.55 117.7 6.41 107.7 4.3 91.2 12.15 204.1 7 148.4 8.91 149.7 4.1 86.9 9.4 157.9 average = 138.0 average = 185.1 1946 1991 L (mm) L(m) L (mm) 51.72 1096.5 channel 61 40.1 850.1 valley 50 Field Observed 2002 W(m) 149.3 218.8 251.9 159.1 average = 194.8 L(m) 1024.8 840.0 sinuosity 1.3 sinuosity 1.2 1946 1991 2002 Mean 137.96 Mean 185.07 Mean 178.63 Standard Error 10.74 Standard Error 11.80 Standard Error 35.26 227 CN O O CM £ CO CO CO Tf CO i f CM Tf CQ LO OJ > o> TJ io co in co cn in co cri CM I CD CO . Q O CD O) 2 S CO I CM I LU "S co co £ CD CD 2 55 •pCD CO ^ CM OO —' CO Tf O O O CJ) E (— TfTfcOCMOl— CMCM rf LO'cgcdds^do) co TfcocoininTi-coTf m O) O) CM CO . CM CO r-- m O) CO Tf % SI CL 3 E Tf E 3 I-co CM O) tn CO Tf CO . O CO CNJ CO *sj #vj m r-co co T-CM CO II CD CD 2 $. CO CD c c CO - C o CD = cn o 3 CO CD ro 2 55 o o 8 in II 0) 8 CO E « ' CM CM | CM E Tf CM o) co co m oo Tf oo CM co S 3 o o O O) T-co in Tf I— 1- oo in T- i— o d n CD O) CO 8 CO 3 co co ~ O) O) T-—i in Tf c oo 1 ° CO o c CO lo> 5 in CO i LU •J3 CO ro £ <D CO :> co d cu a CU u s CA es CU TS L . CU Vi JS es H o o CD CO O II "5 CD £ 8 j=Tf in cj)cptpoqpT-coc3) • i— CM co o) r- cn to m T- CD co CM T-T-T-COTfTfCMCM S E . r- io „, 3 m L O in cn £ > o O ~ rS o O T- T- O CD O) CO s CO -p CO Tf £,iri s —I CO S s § o) S E ® 0) s CO o c O) CM 00 CO 00 CO CM 8 LU "5 CO ro £ CD eg 2 55 228 West Kettle River Width Measurements Air Photo scale = 31400 1951 1951 Mean 38.31 W (mm) W (m) Standard Error 2.95 0.5 15.7 0.9 28.3 ~ 1 31.4 1.7 53.4 1.6 50.2 1.3 40.8 1.1 34.5 0.95 29.8 1.1 34.5 0.9 28.3 1.8 56.5 1.5 47.1 1.45 45.5 1.5 47.1 1 31.4 average = 38.3 1951 L (mm) L (m) channel 69.9 2194.9 valley 55' 1727.0 sinuosity 1.3 229 

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.831.1-0063629/manifest

Comment

Related Items