"Applied Science, Faculty of"@en . "Civil Engineering, Department of"@en . "DSpace"@en . "UBCV"@en . "Smiarowski, Alana"@en . "2010-04-30T17:37:51Z"@en . "2010"@en . "Master of Applied Science - MASc"@en . "University of British Columbia"@en . "The purpose of this study was to provide an evaluation of the performance of R2DM using a mountainous stream reach with a complex geometry and coarse substrate. R2DM is a two-dimensional sediment transport and bed morphology model which was developed at the University of British Columbia and programmed within River2D. Bed morphology changes were modelled over a series of high flow events which occurred on the Seymour River in North Vancouver, British Columbia during the winter of 2008/2009. The model was evaluated based on its ability to simulate the overall and local bed changes by comparing the modelled results to the surveys of the pre- and post-storm ground elevations. The model displayed erosion in high bed shear stress areas and deposition in low bed shear stress areas and was able to display general bed changes through the site. R2DM was also tested as a design tool looking at bank protection using various riprap orientations through the same study reach. As a design tool R2DM showed favourable results.\n In addition to evaluating the R2DM as a sediment transport and bed morphology model, the instream works through the study reach were also assessed. Instream works were installed to improve the physical habitat for fish by decreasing velocities and increasing depths through the site. Comparing the pre- and post-project works an increase in physical habitat availability was found through the two-dimensional simulations. A second evaluation of the habitat was also completed examining the morphological changes which took place at the site. Sediment transport is an important consideration in habitat improvement projects due to the impacts that the instream structures can have on the sediment transport but also the effects that sediment transport can have on the instream structures.\nRiver2D along with R2DM can provide a means of evaluating river restoration works by combining hydrodynamic, morphological and habitat simulations into one model."@en . "https://circle.library.ubc.ca/rest/handle/2429/24252?expand=metadata"@en . " The Evaluation of a Two-Dimensional Sediment Transport and Bed Morphology Model Based on the Seymour River by Alana Smiarowski BASC, Queen\u00E2\u0080\u0099s University, 2005 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in The Faculty of Graduate Studies (Civil Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2010 \u00C2\u00A9 Alana Smiarowski, 2010 ii ABSTRACT The purpose of this study was to provide an evaluation of the performance of R2DM using a mountainous stream reach with a complex geometry and coarse substrate. R2DM is a two- dimensional sediment transport and bed morphology model which was developed at the University of British Columbia and programmed within River2D. Bed morphology changes were modelled over a series of high flow events which occurred on the Seymour River in North Vancouver, British Columbia during the winter of 2008/2009. The model was evaluated based on its ability to simulate the overall and local bed changes by comparing the modelled results to the surveys of the pre- and post-storm ground elevations. The model displayed erosion in high bed shear stress areas and deposition in low bed shear stress areas and was able to display general bed changes through the site. R2DM was also tested as a design tool looking at bank protection using various riprap orientations through the same study reach. As a design tool R2DM showed favourable results. In addition to evaluating the R2DM as a sediment transport and bed morphology model, the instream works through the study reach were also assessed. Instream works were installed to improve the physical habitat for fish by decreasing velocities and increasing depths through the site. Comparing the pre- and post-project works an increase in physical habitat availability was found through the two-dimensional simulations. A second evaluation of the habitat was also completed examining the morphological changes which took place at the site. Sediment transport is an important consideration in habitat improvement projects due to the impacts that the instream structures can have on the sediment transport but also the effects that sediment transport can have on the instream structures. River2D along with R2DM can provide a means of evaluating river restoration works by combining hydrodynamic, morphological and habitat simulations into one model. iii TABLE OF CONTENTS ABSTRACT .................................................................................................................................................... ii\u00C2\u00A0 TABLE OF CONTENTS ................................................................................................................................. iii\u00C2\u00A0 LIST OF TABLES ........................................................................................................................................... v\u00C2\u00A0 LIST OF FIGURES ........................................................................................................................................ vi\u00C2\u00A0 LIST OF EQUATIONS ................................................................................................................................. viii\u00C2\u00A0 LIST OF VARIABLES .................................................................................................................................... ix\u00C2\u00A0 ACKNOWLEDGEMENTS............................................................................................................................... xi\u00C2\u00A0 DEDICATION ............................................................................................................................................... xii\u00C2\u00A0 1.\u00C2\u00A0 INTRODUCTION ................................................................................................................................... 1\u00C2\u00A0 2.\u00C2\u00A0 BACKGROUND ..................................................................................................................................... 3\u00C2\u00A0 2.1.\u00C2\u00A0 FLOW MODELLING ...................................................................................................................................... 3\u00C2\u00A0 2.1.1.\u00C2\u00A0 One-Dimensional Hydrodynamic Modelling .......................................................................... 3\u00C2\u00A0 2.1.2.\u00C2\u00A0 Two-Dimensional Hydrodynamic Modelling ......................................................................... 4\u00C2\u00A0 2.2.\u00C2\u00A0 MORPHOLOGICAL AND SEDIMENT TRANSPORT MODELLING ....................................................................... 5\u00C2\u00A0 2.2.1.\u00C2\u00A0 One-Dimensional Sediment Transport Modelling .................................................................. 6\u00C2\u00A0 2.2.2.\u00C2\u00A0 Two-Dimensional Sediment Transport Modelling .................................................................. 6\u00C2\u00A0 2.2.3.\u00C2\u00A0 Three-Dimensional Sediment Transport Modelling ................................................................ 6\u00C2\u00A0 2.3.\u00C2\u00A0 HABITAT MODELLING ................................................................................................................................. 7\u00C2\u00A0 2.3.1.\u00C2\u00A0 Habitat Suitability Curves ....................................................................................................... 7\u00C2\u00A0 2.3.2.\u00C2\u00A0 One-Dimensional Habitat Modelling (PHABSIM ) ................................................................ 8\u00C2\u00A0 2.3.3.\u00C2\u00A0 River2D Habitat .................................................................................................................... 10\u00C2\u00A0 2.4.\u00C2\u00A0 NUMERICAL MODELLING USE IN RIVER RESTORATION ............................................................................ 11\u00C2\u00A0 3.\u00C2\u00A0 RIVER2D AND R2DM ........................................................................................................................ 14\u00C2\u00A0 3.1.\u00C2\u00A0 RIVER2D HYDRODYNAMICS ..................................................................................................................... 14\u00C2\u00A0 3.2.\u00C2\u00A0 RIVER2D MORPHOLOGY (R2DM) ............................................................................................................. 15\u00C2\u00A0 4.\u00C2\u00A0 SEYMOUR RIVER STUDY SITE .......................................................................................................... 23\u00C2\u00A0 4.1.\u00C2\u00A0 WATERSHED BACKGROUND ...................................................................................................................... 23\u00C2\u00A0 4.2.\u00C2\u00A0 HYDROLOGY ............................................................................................................................................. 24\u00C2\u00A0 4.3.\u00C2\u00A0 SEYMOUR RIVER MORPHOLOGY ............................................................................................................... 26\u00C2\u00A0 iv 4.4.\u00C2\u00A0 PROJECT SITE AND HISTORY ..................................................................................................................... 27\u00C2\u00A0 5.\u00C2\u00A0 METHODOLOGY ................................................................................................................................ 31\u00C2\u00A0 5.1.\u00C2\u00A0 DESIGN FLOW EVENT ................................................................................................................................ 31\u00C2\u00A0 5.2.\u00C2\u00A0 AT SITE FIELD DATA COLLECTION............................................................................................................ 34\u00C2\u00A0 5.2.1.\u00C2\u00A0 Bed Elevation Surveys .......................................................................................................... 34\u00C2\u00A0 5.2.2.\u00C2\u00A0 Velocities and Depths/Water Surface Elevations .................................................................. 35\u00C2\u00A0 5.2.3.\u00C2\u00A0 Sediment size ........................................................................................................................ 36\u00C2\u00A0 5.3.\u00C2\u00A0 RIVER2D SET-UP AND CALIBRATION ........................................................................................................ 37\u00C2\u00A0 5.3.1.\u00C2\u00A0 Bed Files................................................................................................................................ 37\u00C2\u00A0 5.3.2.\u00C2\u00A0 Mesh Files ............................................................................................................................. 39\u00C2\u00A0 5.3.3.\u00C2\u00A0 Calibration ............................................................................................................................. 39\u00C2\u00A0 5.4.\u00C2\u00A0 RD2M ....................................................................................................................................................... 42\u00C2\u00A0 5.5.\u00C2\u00A0 HABITAT MODELLING ............................................................................................................................... 43\u00C2\u00A0 6.\u00C2\u00A0 RESULTS AND DISCUSSION ............................................................................................................... 45\u00C2\u00A0 6.1.\u00C2\u00A0 BED MORPHOLOGY CHANGES ................................................................................................................... 45\u00C2\u00A0 6.2.\u00C2\u00A0 R2DM ....................................................................................................................................................... 46\u00C2\u00A0 6.3.\u00C2\u00A0 R2DM WITH DESIGN FLOWS ..................................................................................................................... 54\u00C2\u00A0 6.4.\u00C2\u00A0 R2DM AS A DESIGN TOOL ........................................................................................................................ 61\u00C2\u00A0 6.5.\u00C2\u00A0 HABITAT ................................................................................................................................................... 68\u00C2\u00A0 6.5.1.\u00C2\u00A0 Pre- to Post- Instream Works (2001 to 2008) ........................................................................ 68\u00C2\u00A0 6.5.2.\u00C2\u00A0 Pre- to Post- Flooding Event (2008 to 2009) ........................................................................ 74\u00C2\u00A0 6.5.3.\u00C2\u00A0 Post- Flooding Event (2009 Observed versus 2009 Simulated) ............................................ 76\u00C2\u00A0 7.\u00C2\u00A0 CONCLUSIONS ................................................................................................................................... 78\u00C2\u00A0 7.1.\u00C2\u00A0 EVALUATION OF R2DM ............................................................................................................................ 78\u00C2\u00A0 7.2.\u00C2\u00A0 EVALUATION OF HABITAT IMPROVEMENTS ............................................................................................... 79\u00C2\u00A0 REFERENCES .............................................................................................................................................. 80\u00C2\u00A0 APPENDICES ............................................................................................................................................... 84\u00C2\u00A0 APPENDIX A: WILCOCK AND CROWE (2003) SEDIMENT TRANSPORT EQUATION .................................................... 84\u00C2\u00A0 APPENDIX B: PHOTOGRAPHS OF THE BED CHANGES AT SEYMOUR RIVER SITE ....................................................... 87\u00C2\u00A0 APPENDIX C: RATING CURVE DATA ........................................................................................................................ 89\u00C2\u00A0 APPENDIX D: AT SITE VELOCITY, DEPTH AND WATER SURFACE ELEVATION DATA .............................................. 93\u00C2\u00A0 APPENDIX E: GRAIN SIZE DISTRIBUTIONS ............................................................................................................... 97\u00C2\u00A0 v LIST OF TABLES Table 4-1: Estimated Flood Frequency Return Periods and Flows .............................................. 25\u00C2\u00A0 Table 5-1: Seymour River Rating Curve and Measured Discharges ............................................ 31\u00C2\u00A0 Table 5-2: Design Events .............................................................................................................. 33\u00C2\u00A0 Table 5-3: Details of Calibration Flows ........................................................................................ 40\u00C2\u00A0 Table 6-1: Summary of R2DM Combination Runs ...................................................................... 47\u00C2\u00A0 Table 6-2: Winter 2009/2010 Design Flows ................................................................................. 62\u00C2\u00A0 Table 6-3: Modelled % WUA for Coho Salmon at Selected Discharges (2001 to 2008) ............ 68\u00C2\u00A0 Table 6-4: Modelled % WUA for Steelhead Trout at Selected Discharges (2001 to 2008) ......... 69\u00C2\u00A0 Table 6-5: Modelled % WUA for Coho Salmon at Selected Discharges (2008 to 2009) ............ 74\u00C2\u00A0 Table 6-6: Modelled % WUA for Steelhead Trout at Selected Discharges (2008 to 2009) ......... 74\u00C2\u00A0 Table 6-7: Modelled % WUA for Coho Salmon at Selected Discharges (2009 Observed to 2009 Simulated) ..................................................................................................................................... 76\u00C2\u00A0 Table 6-8: Modelled % WUA for Steelhead Trout at Selected Discharges (2009 Observed to 2009 Simulated) ............................................................................................................................ 76\u00C2\u00A0 vi LIST OF FIGURES Figure 2-1: Binary versus Univariate Habitat Suitability Curves ................................................... 8\u00C2\u00A0 Figure 3-1: Flow Chart Showing Calculation of Bedload Transport Rate ................................... 17\u00C2\u00A0 Figure 3-2: Example Diagram to Illustrate Sediment Flux Movement through an Element ........ 18\u00C2\u00A0 Figure 3-3: Conceptual Model of a Gravel Bedded River during Aggradation ........................... 20\u00C2\u00A0 Figure 3-4: Conceptual Model of a Gravel Bedded River during Degradation ............................ 20\u00C2\u00A0 Figure 3-5: Secondary Flow Illustration ....................................................................................... 21\u00C2\u00A0 Figure 3-6: Change in Bed Shear Angle Due to Secondary Flow ................................................ 22\u00C2\u00A0 Figure 4-1: Seymour River Watershed Location .......................................................................... 24\u00C2\u00A0 Figure 4-2: Mean Monthly Discharges at WSC Station 08GA030 .............................................. 25\u00C2\u00A0 Figure 4-3: Flood Frequency Analysis Based on Water Survey of Canada's Gauging Site 08GA030 ....................................................................................................................................... 26\u00C2\u00A0 Figure 4-4: Project Location ......................................................................................................... 27\u00C2\u00A0 Figure 4-5: Project Features .......................................................................................................... 29\u00C2\u00A0 Figure 4-6: Location of Art Rocks ................................................................................................ 30\u00C2\u00A0 Figure 5-1: Seymour River Empirical Rating Curve .................................................................... 32\u00C2\u00A0 Figure 5-2: Hydrograph Recorded at the Grantham Place Bridge ................................................ 33\u00C2\u00A0 Figure 5-3: Bed Surveys \u00E2\u0080\u0093 Data provided by KWL ..................................................................... 35\u00C2\u00A0 Figure 5-4: Grain Size Distributions ............................................................................................. 36\u00C2\u00A0 Figure 5-5: 2008 Bed Elevation Files for a) R2DM versus b) Habitat Simulations ..................... 38\u00C2\u00A0 Figure 5-6: Observed versus Modelled Calibration Plots for Water Discharge of 3.7m3/s .......... 41\u00C2\u00A0 Figure 5-7: Observed versus Modelled Calibration Plot for Water Discharge of 25 m3/s ........... 42\u00C2\u00A0 Figure 5-8: Habitat Suitability Curves - Coho Salmon ................................................................. 43\u00C2\u00A0 Figure 5-9: Habitat Suitability Curves - Steelhead Trout ............................................................. 44\u00C2\u00A0 Figure 6-1: Changes in Bed Morphology ..................................................................................... 45\u00C2\u00A0 Figure 6-2: R2DM Simulation Assuming a Uniform Riprap Grain Size Distribution over the Entire Study Reach ....................................................................................................................... 48\u00C2\u00A0 Figure 6-3:R2DM Simulation Assuming a Natural Bed Surface Grain Size Distribution over the Entire Study Reach ....................................................................................................................... 49\u00C2\u00A0 vii Figure 6-4: R2DM Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank ................................................................................... 50\u00C2\u00A0 Figure 6-5: Pre- and Post-Run D50 Grain Size Distributions for Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank ................ 51\u00C2\u00A0 Figure 6-6: Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank and Applying NEA to the Riprap Areas ............................... 52\u00C2\u00A0 Figure 6-7: Simulation Assuming a Natural Bed Surface Grain Size Distribution Over the Entire Site and Applying NEA to the Riprap Areas ................................................................................ 53\u00C2\u00A0 Figure 6-8: Results of Modelled Run for all Design Flows .......................................................... 55\u00C2\u00A0 Figure 6-9: Comparison of Bed Elevations for Observed Pre- and Post-Storm and Modelled Post Storm ............................................................................................................................................. 56\u00C2\u00A0 Figure 6-10: Comparison of Observed and Simulated Changes in Bed Elevations ..................... 57\u00C2\u00A0 Figure 6-11: Comparison of Observed and Simulated 2009 Bed Elevations ............................... 58\u00C2\u00A0 Figure 6-12: Modelled Velocities for a Water Discharge of 113.4 m3/s ...................................... 59\u00C2\u00A0 Figure 6-13: Photos of the Change in the Bank along the Art Rock Location ............................. 62\u00C2\u00A0 Figure 6-14: Results of Simulated 2009/2010 Winter Storm Event Bed Elevations Showing Failed Art Rocks ........................................................................................................................... 63\u00C2\u00A0 Figure 6-15: Results of R2DM Simulation for Design B ............................................................. 65\u00C2\u00A0 Figure 6-16: Results of R2DM Simulation for Design C ............................................................. 66\u00C2\u00A0 Figure 6-17: Results of R2DM Simulation for Design D ............................................................. 67\u00C2\u00A0 Figure 6-18: Percent WUA for Coho Salmon and Steelhead Trout 2001 to 2008 ....................... 69\u00C2\u00A0 Figure 6-19: Velocity Suitability at 2 m3/s for Coho Fry ............................................................. 71\u00C2\u00A0 Figure 6-20: Modelled Velocities at a Discharge of 150 m3/s ...................................................... 72\u00C2\u00A0 Figure 6-21: Modelled Velocity Suitability for Coho Salmon Fry at a Discharge of 150 m3/s ... 73\u00C2\u00A0 Figure 6-22: Percent WUA for Coho Salmon and Steelhead Trout 2008 to 2009 ....................... 75\u00C2\u00A0 Figure 6-23: Percent WUA for Coho Salmon and Steelhead Trout 2009 Observed to 2009 Simulated ...................................................................................................................................... 77\u00C2\u00A0 viii LIST OF EQUATIONS Equation 2-1: HEC RAS Energy Equation ..................................................................................... 3\u00C2\u00A0 Equation 2-2: Manning Equation .................................................................................................... 4\u00C2\u00A0 Equation 2-3: Conservation of Water Mass .................................................................................... 4\u00C2\u00A0 Equation 2-4: Conservation of Momentum in the X-Direction ...................................................... 4\u00C2\u00A0 Equation 2-5: Conservation of Momentum in the Y-Direction ...................................................... 4\u00C2\u00A0 Equation 2-6: Exner\u00E2\u0080\u0099s 2D Sediment Continuity Equation .............................................................. 5\u00C2\u00A0 Equation 2-7: Weighted Usable Area ............................................................................................. 9\u00C2\u00A0 Equation 2-8: Suitability Index ....................................................................................................... 9\u00C2\u00A0 Equation 3-1: Bed Elevation ......................................................................................................... 18\u00C2\u00A0 Equation 3-2: Volume of Fraction i in Element with Area AE During Aggradation .................... 19\u00C2\u00A0 Equation 3-3: Volume of Fraction i in Element with Area AE During Degradation .................... 19\u00C2\u00A0 Equation 3-4: New Surface Layer Fraction .................................................................................. 19\u00C2\u00A0 Equation 3-5: Up-Winding Scheme Used in R2DM .................................................................... 21\u00C2\u00A0 Equation 5-1: Seymour River Empirical Rating Curve Equation ................................................. 32\u00C2\u00A0 Equation 5-2: Manning-Stickler ................................................................................................... 39\u00C2\u00A0 Equation 5-3: Mean Absolute Error .............................................................................................. 40\u00C2\u00A0 Equation 5-4: Mean Error ............................................................................................................. 40\u00C2\u00A0 ix LIST OF VARIABLES A = cross sectional flow area Ah = observed available habitat Ai = unit surface area a = velocity weighting coefficient C = Chezy friction coefficient ci = suitability index Di = grain size in fraction i D50 = size of 50th percentile grain size Fs = proportion of sand in surface Fsi = fraction of grain size i in the substrate fu = velocity factor fy = depth of flow factor fsb = substrate factor ft = temperature factor g = acceleration due to gravity he = energy head loss h = water depth Ls = surface layer thickness MAE = mean absolute error ME = mean error n = Manning\u00E2\u0080\u0099s roughness value P = preference index Q = flow discharge q = discharge intensity qs = volumetric rate of bedload transport per unit width So = bed slope Sf = frictional slope s = specific gravity of sediment t = time x Uh = observed used habitat UW = up-winding weighting factor u = depth-averaged velocity in the x-direction Vi = volume of fraction i in the surface layer v = depth-averaged velocity in the y-direction W*i = dimensionless transport rate of size fraction i W*r = reference value of dimensionless transport rate (=0.002) WUA = weighted usable area Y = water depth \u00DD\u0095\u00E0\u00B4\u00A4 = cross sectional average water depth z = elevation of main channel invert zb = elevation of bed \u00CE\u00BB = porosity of the bed material \u00CF\u0081 = density of water \u00CF\u0084 = shear stress \u00CF\u0084*ci = dimensionless reference shear stress for fraction i \u00DF\u00B6 = \u00CF\u0084/\u00CF\u0084ri xi ACKNOWLEDGEMENTS I offer my thanks to the faculty, staff and my fellow students at the UBC and I owe particular thanks to Dr. Rob Millar for providing me with guidance and inspiration. I would like to thank everyone at KWL and especially David Matsubara and Jason Vine for allowing me to exploit their knowledge and years of experience. Thank you to KWL and MITACS for providing funding for the project and Brett Eaton for loaning me his ADCP. I thank all of those who helped me in the field, Joel Evans, Kelley Hishon, Sarah Portelance- Blanchard, Marcel Luthi, Jane Bachman, and Hannha Chiew. I would like to especially thank Steve Kwan for always dealing with my constant pestering about the model and Caroline Gort for brainstorming with me in the field. Special thanks are owed to my parents and to Matt who have support me throughout my years of education, both emotionally and financially xii DICATION To Matt, who has always believed in me 1 1. INTRODUCTION Many elements associated with river engineering and habitat improvement efforts, such as predicting morphological changes and observing physical habitat at various flows, are difficult to measure and require many years of monitoring (Waddle, 2009). One-dimensional (1D) and two- dimensional (2D) models have been incorporated in several studies to assist in the analysis of the hydrodynamics of restoration efforts (Lacey and Millar, 2004, Khangaonkar et al., 2005). Few studies have looked at the interactions between the hydraulic, sediment transport and bed morphology changes of river restoration efforts (Carr\u00C3\u00A9 et al., 2007), and many studies and habitat models completely omit the effects of sediment transport (Hauer et al., 2007). Nevertheless bed morphology changes and sediment transport are vital components for habitat assessment and there is a need for more projects and more field work to better quantify these aspects. Since bed morphology changes and sediment transport processes are very difficult to measure and even more difficult to predict, new means of estimation are required. Vasquez et al. (2007) developed a 2D finite element river morphology model to simulate bed elevation changes called R2D-Morph. Their model was programmed within River2D (developed by the University of Alberta) and a simulated bed change based on the Exner\u00E2\u0080\u0099s equation and was validated using flume experiments. Their model was recently updated to incorporate the Wilcock and Crowe (2003) equation for gravel transport by Kwan (2009) to better represent the gravel bed streams of the Pacific Northwest. Other studies have used River2D to simulate the hydraulic affects of restoration efforts (for example: Ghanem et al., 1996, and Schwartz, 2003). This project will not only be evaluating the hydraulics of a restoration effort, but will also be considering the affects of sediment transport on the instream structures and physical habitat. Bed elevation surveys of a pre- and post-storm event collected from the Seymour River in North Vancouver, British Columbia will be used to evaluate R2DM\u00E2\u0080\u0099s ability to predict changes to bed elevations. This study will also evaluate the instream design features to see if they acted as intended and to determine if the bed morphology changes affected their performance. An overview of flow, sediment and physical habitat numerical models is discussed in Section 2 while Section 3 presents a brief description of the River2D and R2DM models. Section 4 2 describes the Seymour River field site and the hydrologic events which took place during the study. The methodology of the design flow event calculations, data collection and analysis of results are described in Section 5 as well as how the numerical model was set up and evaluated. Section 6 presents the findings of the study and discusses the usability of R2DM for this study. Finally, the conclusions of the study and recommendations for further work are offered in Section 7. 3 2. BACKGROUND 2.1. Flow Modelling Advances in numerical modelling methods and computational technology have allowed complex systems to be numerically modelled in a shorter time frame and at a lower cost than conducting a full scale field study or physical model. For open channel flow numerical modelling can be a simple, large scale model through a one-dimensional (1D) representation, a more detailed localized two-dimensional (2D) representation or a complex and detailed three-dimensional (3D) representation. 2.1.1. One-Dimensional Hydrodynamic Modelling Typically, 1D models simulate flow in the longitudinal direction while averaging the velocity and depth values over an entire cross section. 1D models are used for large scale projects for determining flood stage at various reach locations and may only be applied when changes in cross sectional geometry and hydraulic roughness are gradual (Lacey and Millar, 2001). HEC- RAS is a 1D open channel flow model developed by the U.S. Army Corps of Engineers which is able to perform steady and unsteady flow hydraulics (U.S. Army Corps of Engineers, 2008). HEC-RAS solves for water surface profiles from one cross section to the next by solving the Energy equation. Equation 2-1: HEC RAS Energy Equation \u00DC\u00BC2 \u00E0\u00B5\u0085 \u00DC\u00BB2 \u00E0\u00B5\u0085 \u00DC\u00BD2\u00DC\u00B82 2 2\u00DD\u0083 \u00E0\u00B5\u008C \u00DC\u00BC1 \u00E0\u00B5\u0085 \u00DC\u00BB1 \u00E0\u00B5\u0085 \u00DC\u00BD1\u00DC\u00B812 2\u00DD\u0083 \u00E0\u00B5\u0085 \u00DD\u0084\u00DD\u0081 where Z1 and Z2 are the elevation of the main channel inverts, Y1 and Y2 are the water depths at the cross sections, V1 and V2 are the average velocities, a1 and a2 are the velocity weighting coefficients, g is the acceleration due to gravity and he is the energy head loss. HEC-RAS also uses the Manning equation to relate cross sectional average depth, cross sectional area, flow, and bed slope. 4 Equation 2-2: Manning Equation \u00DC\u00B3 \u00E0\u00B5\u008C 1\u00DD\u008A\u00DC\u00A3\u00DD\u0095\u00E0\u00B4\u00A4 \u00E0\u00AC\u00B6 \u00E0\u00AC\u00B7\u00DC\u00B5\u00E0\u00AF\u00A2 \u00E0\u00AC\u00B5 \u00E0\u00AC\u00B6 where Q is the flow discharge, n is the Manning roughness value, A is the cross-sectional flow area, \u00DD\u0095\u00E0\u00B4\u00A4 is the cross-sectional average water depth and So is the bed slope. The Manning equation is used to give simple parameter approximations which include flow, velocity, flow depth and flow resistance. Doyle et al., (2007) successfully used HEC-RAS to estimate the bankfull discharge on Lincoln Creek in order to investigate the potential of the bankfull discharge being the channel forming discharge in river restoration projects. 2.1.2. Two-Dimensional Hydrodynamic Modelling When local details of velocity and depth distributions are considered important in a river engineering application 2D modelling is used for its accuracy in characterizing river hydraulic behaviour (Steffler and Blackburn, 2002). The majority of 2D models are depth-averaged and compute lateral variations in velocity and depth and all 2D models are based on solving the mass (continuity) and momentum equations (Equation 2-3, Equation 2-4, and Equation 2-5). The governing differential equations of the model can then be solved using finite difference, finite element or finite volume methods (Steffler and Blackburn, 2002). Equation 2-3: Conservation of Water Mass \u00DF\u00B2\u00DD\u0084 \u00DF\u00B2\u00DD\u0090 \u00E0\u00B5\u0085 \u00DF\u00B2\u00DD\u008D\u00DD\u0094 \u00DF\u00B2\u00DD\u0094 \u00E0\u00B5\u0085 \u00DF\u00B2\u00DD\u008D\u00DD\u0095 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u0085 0 where h is the flow depth, t is time, qx is the x-discharge intensity and qy is the y-discharge intensity. Equation 2-4: Conservation of Momentum in the X-Direction \u00DF\u00B2\u00DD\u008D\u00DD\u0094 \u00DF\u00B2\u00DD\u0090 \u00E0\u00B5\u0085 \u00DF\u00B2 \u00DF\u00B2\u00DD\u0094 \u00E0\u00B5\u00AB\u00DD\u0091\u00DD\u008D\u00DD\u0094\u00E0\u00B5\u00AF \u00E0\u00B5\u0085 \u00DF\u00B2 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u00AB\u00DD\u0092\u00DD\u008D\u00DD\u0094\u00E0\u00B5\u00AF \u00E0\u00B5\u0085 \u00DD\u0083 2 \u00DF\u00B2\u00DD\u00842 \u00DF\u00B2\u00DD\u0094 \u00E0\u00B5\u008C \u00DD\u0083\u00DD\u0084\u00E0\u00B5\u00AB\u00DC\u00B5\u00DD\u008B\u00DD\u0094 \u00E0\u00B5\u0086 \u00DC\u00B5\u00DD\u0082\u00DD\u0094\u00E0\u00B5\u00AF \u00E0\u00B5\u0085 1 \u00DF\u00A9 \u00E0\u00B5\u00A5\u00E0\u00B5\u00AD \u00DF\u00B2 \u00DF\u00B2\u00DD\u0094 \u00E1\u0088\u00BA\u00DD\u0084\u00DF\u00AC\u00DD\u0094\u00DD\u0094\u00E1\u0088\u00BB\u00E0\u00B5\u00B1 \u00E0\u00B5\u0085 \u00E0\u00B5\u00AD \u00DF\u00B2 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u00AB\u00DD\u0084\u00DF\u00AC\u00DD\u0094\u00DD\u0095\u00E0\u00B5\u00AF\u00E0\u00B5\u00B1\u00E0\u00B5\u00A9 Equation 2-5: Conservation of Momentum in the Y-Direction \u00DF\u00B2\u00DD\u008D\u00DD\u0095 \u00DF\u00B2\u00DD\u0090 \u00E0\u00B5\u0085 \u00DF\u00B2 \u00DF\u00B2\u00DD\u0094 \u00E1\u0089\u0080\u00DD\u0091\u00DD\u008D\u00DD\u0095\u00E1\u0089\u0081 \u00E0\u00B5\u0085 \u00DF\u00B2 \u00DF\u00B2\u00DD\u0095 \u00E1\u0089\u0080\u00DD\u0092\u00DD\u008D\u00DD\u0095\u00E1\u0089\u0081 \u00E0\u00B5\u0085 \u00DD\u0083 2 \u00DF\u00B2\u00DD\u00842 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u008C \u00DD\u0083\u00DD\u0084\u00E0\u00B5\u00AB\u00DC\u00B5\u00DD\u008B\u00DD\u0095 \u00E0\u00B5\u0086 \u00DC\u00B5\u00DD\u0082\u00DD\u0095\u00E0\u00B5\u00AF \u00E0\u00B5\u0085 1 \u00DF\u00A9 \u00E0\u00B5\u00A5\u00E0\u00B5\u00AD \u00DF\u00B2 \u00DF\u00B2\u00DD\u0094 \u00E0\u00B5\u00AB\u00DD\u0084\u00DF\u00AC\u00DD\u0095\u00DD\u0094\u00E0\u00B5\u00AF\u00E0\u00B5\u00B1 \u00E0\u00B5\u0085 \u00E0\u00B5\u00AD \u00DF\u00B2 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u00AB\u00DD\u0084\u00DF\u00AC\u00DD\u0095\u00DD\u0095\u00E0\u00B5\u00AF\u00E0\u00B5\u00B1\u00E0\u00B5\u00A9 5 where, t is time, u and v are the depth-averaged velocities in the x and y directions, related to the discharge intensity components through qx=uh and qy=vh; g is the acceleration due to gravity, h is the flow depth, \u00CF\u0081 is the density of water, Sox and Soy are the bed slopes of the x and y axes, respectively, and Sfx and Sfy are the friction slopes of each. \u00CF\u0084xx, \u00CF\u0084xy, and \u00CF\u0084yy represent horizontal stress factors. River2D is a 2D model which was developed at the University of Alberta (Steffler and Blackburn, 2002) and is a depth averaged, finite element, hydrodynamic model and was developed specifically for gravel bed rivers. A full description of River2D is detailed in Section 3. Waddle (2009) used River2D to model the depth and velocities around boulder clusters to evaluate the habitat for brown trout. Waddle (2009) found that the model successfully represented the complex hydrodynamics of the area and was able to define habitat features that would have not been possible with a 1D model. Additional 2D hydrological models which have been used to successfully model rivers for engineering and habitat assessments include RMA2, which was developed by King (1990), but is now maintained by the Army Corps of Engineers Waterways Experiment Station (Crowder and Diplas, 2000) and CCHE2D developed by the National Center for Computation Hydroscience and Engineering (Formann et al., 2007). 2.2. Morphological and Sediment Transport Modelling Sediment transport values are some of the most important yet most difficult parameters to estimate in river engineering and related areas. The prediction of these values is intricate due to turbulent flows, irregular channel geometry, which typically changes with time, and complex sediment transport patterns (Wu et al., 2000). Most sediment transport models solve the hydrodynamics of the system using the Saint-Venant equations for the conservation of mass and momentum of flow (Equation 2-3, Equation 2-4 and Equation 2-5). Sediment accumulation and deposition are modelled using the Exner equation for sediment mass continuity (Vasquez, 2005): Equation 2-6: Exner\u00E2\u0080\u0099s 2D Sediment Continuity Equation \u00E1\u0088\u00BA1 \u00E0\u00B5\u0086 \u00DF\u00A3\u00E1\u0088\u00BB\u00DF\u00B2\u00DD\u0096\u00DC\u00BE\u00DF\u00B2\u00DD\u0090 \u00E0\u00B5\u0085 \u00DF\u00B2\u00DD\u008D\u00DD\u008F\u00DD\u0094 \u00DF\u00B2\u00DD\u0094 \u00E0\u00B5\u0085 \u00DF\u00B2\u00DD\u008D\u00DD\u008F\u00DD\u0095 \u00DF\u00B2\u00DD\u0095 \u00E0\u00B5\u008C 0 6 where zb is the bed elevation; \u00CE\u00BB is the porosity of the bed material; and qsx, qsy are the components of the volumetric rate of bedload transport per unit width and are defined as qsx = qz cos \u00CE\u00B1 and qsy = qs sin \u00CE\u00B1, where \u00CE\u00B1 is the direction of bedload transport. To determine qs in Equation 2-6 a sediment transport equation is required. Various sediment transport equations are available and R2DM allows for the choice between five different equations as shown in Section 3.2. 2.2.1. One-Dimensional Sediment Transport Modelling One-dimensional (1D) sediment transport models simulate sediment transport in the stream-wise longitudinal direction and do not solve the governing equations over the full cross section. These models are considered limited in their abilities and are most applicable in the evaluation of reach scale, long-term processes in channels which have limited hydraulic complexity and minimal variation in channel geometry (Waddle et al., 2000; Formann et al., 2007). 1D models provide reasonable prediction of water surface elevations but have the tendency to under estimate cross sectional velocities by 10 to 20% (Johnson, 2008) which affects their ability to estimate shear stress and therefore sediment transport. However, since 1D models have low data and computational power requirements and because they are comparatively simple to use; these models are still commonly used in industry where the detail of 2D or 3D models is not required. For example, Formann et al., (2007) successfully used a 1D model to simulate long term sedimentation event at the reach scale in the alpine regions of Austria. 2.2.2. Two-Dimensional Sediment Transport Modelling Two-dimensional (2D) sediment models use depth averaged continuity and Navier-Stokes equations along with the sediment mass balance equation and methods of finite difference, finite element or finite volume to compute values in both the streamwise and transverse velocity components. Due to the grid structures of 2D models, such as R2DM, CCHE2D and MIKE 21, these models are able to represent more complex reaches than 1D models (Papanicolaou et al., 2008). 2D models can provide spatially varied information about water depth and bed elevation therefore predicting sediment transport with more accuracy than 1D models. 2.2.3. Three-Dimensional Sediment Transport Modelling Three-dimensional (3D) models, such as CCHE3D, Delft3D and MIKE 3, are used in scenarios where 2D models cannot correctly represent the physics. Areas where 3D models may be required include scour around piers and near hydraulic structures where velocity vectors are 7 more complex (Papanicolaou et al., 2008). Like 2D models, 3D models solve the continuity and Navier-Stokes equations, along with the sediment mass balance equation through the methods of finite difference, finite element or finite-volume; however 3D models introduce variations in velocity in the vertical direction in addition to the streamwise and transverse directions (Wu, 2008). To obtain the detailed results of a 3D model there is a high demand on data collection and computational time, therefore 2D sediment transport models are preferred for simulation of the lateral and longitudinal velocity components which are critical to the evaluation of sediment transport (Johnson, 2008). There have been discrepancies between hydrodynamic/sediment transport model predictions compared to field measurements for many reasons. The major cause is the oversimplification of the problem by using the inappropriate type of model, e.g. using a 1D model when the detail of a 2D model is required, incorrect data, insufficient data or the limitations of the sediment transport equations. At the time of their study, Papanicolaou et al., (2008) believed that there was no sediment transport model which can reliably describe the two-phased phenomenon of sediment transportation and flow. 2.3. Habitat Modelling 2.3.1. Habitat Suitability Curves Habitat suitability curves (HSC) are used to represent the complex and discontinuous relationship between physical variables and species preferences in mathematical terms (Ghanem et al., 1996). The original HSCs were based on a binary definition where optimal areas were defined by a value of 1 and unsuitable areas were defined by a value of 0 (Figure 2-1) (Waddel, 2001). A major disadvantage with the binary method was that if defined habitat as being optimal or unsuitable and did not account for regions which would be considered tolerable by a species (Bovee, 1978). 8 Figure 2-1: Binary versus Univariate Habitat Suitability Curves Waters (1976) was the first to introduce a univariate distribution by using a weighting factor for habitat preference rather than just 0 and 1 (Figure 2-1). He argued that there is a wider range of conditions that a species may find habitable and that there is only a narrow range which that species would select as being optimal (Bovee, 1978). This led away from the binary definition of habitat since it was too restricting. Typically, HSC are defined for the depth, velocity and channel index (based on substrate or cover) as these values are responsible for the primary and secondary productivity and fish population (Waddle, 2001). 2.3.2. One-Dimensional Habitat Modelling (PHABSIM ) The Physical Habitat Simulation (PHABSIM) software is a one-dimensional program used to calculate the potential physical micro-habitat in a river reach. PHABSIM is one of the most widely used assessment model for species\u00E2\u0080\u0099 instream flow requirements in North America (Ghanem et al., 1996). The model was developed as a major component of the Instream Flow Incremental Methodology (IFIM) introduced by the Instream Flow and Aquatic Systems Group (IFG), US Fish and Wildlife Service in 1981 (Ghanem et al., 1996; Waddle, 2001). PHABSIM is calibrated based on hydraulic parameters: water depth and velocity for measured flows. It then estimates the flow conditions and physical habitat for other discharges. PHABSIM 0.0 0.2 0.4 0.6 0.8 1.0 0 1 2 3 4 5 6 Su it ab ili ty Habitat\u00C2\u00A0Parameter\u00C2\u00A0(depth,\u00C2\u00A0velocity,\u00C2\u00A0substrate) Univariate Binary 9 uses basic hydraulic simulation models (Manning\u00E2\u0080\u0099s n equation) to determine the depths and velocities at cross sections through a reach for unknown discharges and evaluates the habitat based on suitability curves (Waddle, 2001). The software multiplies the surface area for a section of stream by univariate suitability curve values to determine the Weighted Usable Area (WUA). Since WUA is not a direct measure of available habitat, it is best used to indicate relative changes in habitat between different flows allowing for a WUA rating curve to be developed (Loranger and Kenner, 2004; Waddle, 2001). WUA is dependent on discharge and species and is calculated using the following equation: Equation 2-7: Weighted Usable Area \u00DC\u00B9\u00DC\u00B7\u00DC\u00A3\u00DC\u00B3,\u00DC\u00B5 \u00E0\u00B5\u008C\u00E0\u00B7\u008D\u00E0\u00B5\u00AB\u00DC\u00BF\u00DD\u0085,\u00DD\u008F \u00C2\u00B7 \u00DC\u00A3\u00DD\u0085\u00E0\u00B5\u00AF \u00DD\u008A \u00DD\u0085\u00E0\u00B5\u008C1 Equation 2-8: Suitability Index ci,s = fu\u00C2\u00B7fy\u00C2\u00B7fc\u00C2\u00B7fsb\u00C2\u00B7ft Where ci,s is the suitability index, Ai is the unit surface area, fu is the velocity factor, fy is the depth of flow factor, fsb is the substrate factor and ft is the temperature factor (Lacey and Millar, 2001). The velocity, depth, substrate and temperature factors are extrapolated from defined habitat preference curves for the reach. PHABSIM has been used to determine the habitat for many different indicator species at a wide range of life stages including modelling various fish and macroinvertebrates habitats (Gard, 2005; Loranger and Kenner, 2004; Stewart et al., 2005). In general, the model has performed well. Studies have been completed that compare the results of a PHABSIM simulation with actual values measured in the field. Gore et al., (1998) used PHABSIM and direct field observations to evaluate the improvements to benthic macroinvertebrates habitat due to the placement of artificial riffles in Holly Fork - a low-order tributary to West Sandy River in West Tennessee. They found a strong correlation between the observed habitat in the field and the modelled habitat in PHABSIM and concluded that PHABSIM is a useful aid in planning and demonstrating the value of restoration structures. Shuler and Nehring, (1993) performed a similar analysis on the Rio Grand and South Platte Rivers in Colorado using adult and juvenile brown 10 trout as their indicator species; again determining the effectiveness of PHABSIM as an evaluation tool. Other studies by Moir et al., (2005) and Spence and Hickley, (2000) have found strong correlations between the modelled and the field results therefore agreeing that PHABSIM is a useful tool for evaluating stream enhancement activities and is suitable in cases where the physical habitat limits the indicator species. 2.3.3. River2D Habitat The fish habitat component of River2D is based on the same Weighted Usable Area (WUA) concept as is used in PHABSIM. As with PHABSIM suitability curves are used in River2D based on depth, velocity and channel index to predict habitat locations. Unlike PHABSIM which uses a number of cross sections to view the stream, River2D uses a continuum and WUA is evaluated at every point in the defined domain and is calculated as an aggregate of the product of a composite suitability index (Ghanem et al., 1996). In River2D, the calculations points are the computational nodes of the finite element mesh. The suitability index of each parameter is evaluated by linear interpolation from the appropriate preference curve supplied by the user. The depths and velocities are taken directly from the hydrodynamic component of the model which also incorporates supercritical and subcritical flow conditions and wet-dry simulation capabilities (Steffler and Blackburn, 2002; Garc\u00C3\u00ADa de Jal\u00C3\u00B3n, and Gort\u00C3\u00A1zar, 2007). Ghanem et al., (1996) used both PHABSIM and River2D to simulate real fish habitat along a reach of Waterton River in Alberta. Their findings showed that both PHABSIM and River2D preformed relatively well for predicting velocities, however River2D was able to simulate a low- velocity zone which was missed by PHABSIM. Since low-velocity zones are very important to fish habitat, River2D was considered a better option for modelling habitat. Loranger and Kenner (2004) also compared the modelling abilities of PHABSIM and River2D by looking at the habitat of brown trout in Rapid Creek, South Dakota. The two models produced very similar results in some reaches yet very different results in other reaches. The differences were accredited to the inability of PHABSIM to simulate low-velocity zones through pools where as in River2D the 2D modelling of the pool habitat was better represented therefore showing larger WUAs. Loranger and Kenner (2004) recommend that River2D be used in studies where spatial analysis of WUA is of importance. 11 Lacey and Millar (2004) used River2D to assess river enchantment works on the Chilliwack River, British Columbia. Coho salmon and steelhead trout habitat was modelled to assess the impacts of instream large woody debris and groyne habitat structures and a side channel of the Chilliwack River. Habitat was well represented by River2D and the improvement efforts resulted in an increase in WUA of 150 to 210 percent. Lacey and Millar (2004) found River2D to be a useful tool in assessing the performance of instream restoration efforts. 2.4. Numerical Modelling Use in River Restoration With the advancement of computers and processing power the use and practicality of numerical models has increased in many fields of engineering and natural sciences. For restoration and river management projects this has meant moving away from the more common natural channel design approach (regime theory) and trial-and-error based designs towards two-dimensional (2D) and three-dimensional (3D) modelling (Schwartz, 2003; Crowder and Diplas, 2000). Two- dimensional hydrodynamic and habitat models, such as River2D, SRH-2D and CASiMiR, have been used in past restoration projects as either design, planning or evaluation tools (e.g. Gore et al., 1998; Schwartz, 2003; Lacey and Millar, 2004; Garc\u00C3\u00ADa de Jal\u00C3\u00B3n, and Gort\u00C3\u00A1zar, 2007; Mouton et al., 2007; Klumpp, 2008; Waddle, 2009). In these projects, the authors found that the models provided a reasonable simulation of the complex flow patterns and evaluated potential physical habitat areas. One common theme among these projects is that the numerical evaluation of the restoration efforts have been completed assuming a static, stable bed. Very few habitat monitoring programs have incorporated bed morphological considerations into their studies and most projects have completely omitted the effects of sediment transport (Hauer et al., 2007). In reality, streams are constantly changing and adjusting as sediment moves though them. The resulting changes in bed morphology are important to fish habitat as a healthy habitat is typically dependant on the geomorphic structures which result from sediment transport - such as pool-riffle systems which plays a vital role in the life cycle of fish (Thompson 2002a, 2002b). With instream restoration projects engineers and planners are generally trying to mimic what happens through nature; however, the complexities of varying geometry and varying bed size materials need to be taken into account when implementing instream habitat structures. Flow and sediment dynamics are 12 very complex and difficult to predict, therefore creating designs which incorporate them can be very tricky and many times the designs are trial-and-error based (Carr\u00C3\u00A9 et al. 2007). Some studies have included sediment transport and bed morphology changes into their work. Klumpp (2008) attempted to predict future geomorphologic changes in the Upper Methow River basin as part of the criteria when evaluating potential restoration sites through the basin. Klumpp\u00E2\u0080\u0099s (2008) documented historic morphological changes and used their professional experience and knowledge of past projects to aid in choosing prospective sites to incorporate instream habitat improvements. Considerations included bed stability and channel migration. Carr\u00C3\u00A9 et al. (2007) looked at a shorter sediment time scale and completed a field study in the Nicolet River, Quebec looking at how the installation of paired deflectors affected the flow of sediment. Their method included monitoring the progression of scour holes through ground surveys, tracer rocks and sediment traps. Carr\u00C3\u00A9 et al (2007) concluded that the deflectors did not move sediment as the design standards had indicated they would. Stream deflectors increase water velocity, and in turn shear stress, causing local scouring of the bed creating a pool and the accumulation of sediment downstream of the pool, forming a riffle. Carr\u00C3\u00A9 et al. (2007) found that many deflectors are installed on a trial-and-error basis, and that their performance change from site-to-site. Carr\u00C3\u00A9 et al. (2007) believe that had if they used a numerical model to represent the complex flows around the deflectors they could have tested different scenarios for deflector angles and lengths to produce the desired location and size of scour prior to the installation of the deflects in the field. Other studies investigating the impact of sediment transport on instream structures and the impact of the instream structures on sediment transport have based their modelling on 2D hydrodynamic models. The models were used to predict the velocities and depths through the site, these results were then used to calculate the shear stresses and therefore the potential for sediment transport. Lacey and Millar, (2001) evaluated 11 instream structures along a side channel of the Chilliwack River, British Columbia. Their evaluation was completed using River2D to establish the velocities and shear stresses of a bankfull event. The results of their numerical modelling were compared with field surveys of a pre- and post-bankfull event and showed favourable results. Hauer et al., (2007) incorporated a stability analysis into their modelling of nase (Chondrostoma naus) spawning grounds in the Slum River, Austria. Based on 13 their study, Hauer et al., (2007) believe that morphodynamic considerations need be made in restorations projects when artificially restoring spawning habitats. River2D along with the R2DM module will provide a means of evaluating river restoration works by combining hydrodynamic, morphological and habitat simulations into one model. River2D has already proven to be a valuable tool in determining the post installation benefits of restoration efforts and offers a useful predictive tool in the design of instream structures Lacey and Millar, (2001). This study will evaluate if R2DM will also be an effective tool for predictions of sediment transport and bed morphology changes. 14 3. RIVER2D AND R2DM 3.1. River2D Hydrodynamics River2D is a two-dimensional, depth-average, finite element model which was developed to model natural gravel bed streams and rivers and has special features to accommodate supercritical/subcritical flow transitions, variable wetted area and ice covers. Inputs for the model include bed topography, roughness and transverse eddy viscosity distributions, boundary conditions, initial flow conditions as well as a discrete mesh or grid to capture the flow variations. River2D is based on the Saint-Venant equations for the conservation of mass and momentum (defined in Equation 2-3, Equation 2-4, and Equation 2-5) which are used along with the Petrov-Galerkin implicit method to solve depth and discharge intensities in the lateral (x) and longitudinal (y) directions (Steffler and Blackburn, 2002). The dependent variables solved for are the water depth, h, and discharge intensities qx and qy. There are three basic assumptions made by River2D: 1. There is a hydrostatic pressure distribution\u00E2\u0080\u0093 this tends to limit the accuracy in areas of steep slopes and rapid bed slope changes and in general any bed feature less than 10 depths horizontally (i.e. dune bedforms) and slopes in the direction of flow which exceed 10% will not be modelled accurately. 2. The distribution of horizontal velocities over the depth are constant, specifically, information on secondary flows and circulations is not available. 3. Coriolis and wind forces are assumed negligible, this may not hold true for very large water bodies where these factors may be significant, but is valid for rivers and streams. The density of the finite element meshes used in River2D is defined by the user. Typically a higher mesh density leads to a more accurate model simulation, however, if a mesh is too dense this can lead to very long, unrealistic computational times. The challenge to the user is to distribute nodes in such a way that the most accurate solution is obtained for a particular purpose while maintaining a mesh density within the computational power of the computer being used for the simulations (Steffler and Blackburn, 2002). 15 Projects have used River2D successfully to model hydrodynamics for either design or assessment. Schwartz (2003) used River2D to predict high-flow habitat of low-gradient Midwest streams in order to assess various restoration efforts through the region. River2D was able to successfully predict the hydraulic recirculation zones and different locations through various morphologies. A field test of River2D was conducted by Waddle (2009) to determine the accuracy of the model to predict near boulder habitat. Simulated depths and velocities were compared to 204 sample locations and the bulk of differences between modelled and observed values were found to fall within a likely error of measurement, therefore showing that River2D could be used as a tool for near boulder habitat prediction. 3.2. River2D Morphology (R2DM) River2D-MOR was originally developed by Vasquez (2005) as part of River2D to simulate sediment transport and moveable beds. The original model: \u00EF\u0082\u00B7 solved the Exner\u00E2\u0080\u0099s equation for sediment continuity and neglected grain sorting as shown in Equation 2-6; \u00EF\u0082\u00B7 was based on the finite element method and unstructured meshes (triangular elements); was valid for both subcritical and supercritical flows; \u00EF\u0082\u00B7 was capable of simulating scour and deposition caused by forced inflow of sediment; took into account the effects of secondary flow and transverse slope in the bedload direction; \u00EF\u0082\u00B7 was capable of simulating dam-break surges over initially wet or dry channels; and \u00EF\u0082\u00B7 was capable of simulating scour and deposition in meandering rivers. The bedload transport rate, which is a function of the flow hydraulics and sediment properties, may be solved in River2D-MOR using one of four sediment transport equations: Meyer Peter Muller, Engelund-Hansen, Van Rijn or through an empirical formula. River2D-MOR was tested using four flume experiments (Vasquez et al., 2007). The tests included bed aggradation due to sediment overload, bed degradation by sediment supply shut off, knickpoint migration, and bar formation in a variable-width channel. River2D-MOR successfully demonstrated its capability to simulate bed level changes for the centrelines all of these flume tests. However, some limitations of Vazquez\u00E2\u0080\u0099s (2005) River2D-MOR model include limited bedload transport equations that only model uniform grains and are unable to model mixed 16 sediment sizes, there was no user interface, the model was unable to simulate moving bedforms and was unable to correctly calculate secondary flow correction for natural river simulations (Kwan, 2009). Kwan (2009) addressed these main limitations of River2D-MOR to create a new version of the model called R2DM. To address the need to simulate the transportation of mixed sediment sizes Kwan (2009) incorporated the Wilcock and Crowe (2003) transport algorithm. A description of the Wilcock and Crowe (2003) equation can be found in Appendix A. A flow chart showing a summary of how R2DM calculates the sediment transport and bed changes is shown on Figure 3-1. 17 Figure 3-1: Flow Chart Showing Calculation of Bedload Transport Rate 18 Once the total transport rate per unit width, qbT, is calculated it is substituted into the bedload transportation equation to compute the bed elevation in a given time step. The bed elevation is updated each time step using the following equation: Equation 3-1: Bed Elevation \u00DD\u0096\u00DC\u00BE\u00DD\u008A\u00DD\u0081\u00DD\u0093 \u00E0\u00B5\u008C \u00DD\u0096\u00DD\u008B\u00DC\u00BE\u00DD\u0088\u00DD\u0080 \u00E0\u00B5\u0085 \u00CE\u0094\u00DD\u0096\u00DC\u00BE where zbnew is the new bed elevation, zbold is the original bed elevation and \u00CE\u0094zb is the change in bed elevation, completed using the Exner equation. R2DM also recalculates the surface and subsurface grain distributions each time step according to the flow of each grain size fraction in and out of an element using the gravel transport algorithm based on a constant active layer thickness. Figure 3-2 shows how sediment may enter and leave an element. Figure 3-2: Example Diagram to Illustrate Sediment Flux Movement through an Element The total sediment transport Q12i, is calculated by multiplying the sediment transport per unit width, q12i, by the length between points 1 and 2. During aggradation the net volume of sediment into an element is positive and a new fraction of grain size i in the surface layer of the element can be determined. Figure 3-3 shows the subsurface, surface and active layer during aggradation. 19 Equation 3-2: Volume of Fraction i in Element with Area AE During Aggradation \u00DC\u00B8\u00DD\u0085 \u00E0\u00B5\u008C \u00E1\u0089\u0080\u00DC\u00B31,2\u00DD\u0085 \u00E0\u00B5\u0085 \u00DC\u00B32,3\u00DD\u0085 \u00E0\u00B5\u0085 \u00DC\u00B33,1\u00DD\u0085\u00E1\u0089\u0081 \u00DD\u0080\u00DD\u0090 \u00E0\u00B5\u0085 \u00E1\u0088\u00BA\u00DC\u00AE\u00DD\u008F \u00E0\u00B5\u0086 \u00DD\u0080\u00DD\u0096\u00E1\u0088\u00BB \u00C2\u00B7 \u00E1\u0088\u00BA1 \u00E0\u00B5\u0086 \u00DF\u00A3\u00E1\u0088\u00BB\u00DC\u00A8\u00DD\u0085\u00DC\u00A3\u00DC\u00A7 where Vi is the volume of fraction i in the surface layer, Fi is the surface layer fraction i, Ls is the surface layer thickness and Q1,2i is the volume of sediment in fraction i entering or leaving the element through side 1,2 of the element per unit time. When sediment leaves the element degradation occurs and the new volume fraction of i is calculated using the following equation: Equation 3-3: Volume of Fraction i in Element with Area AE During Degradation \u00DC\u00B8\u00DD\u0085 \u00E0\u00B5\u008C \u00E1\u0088\u00BA1 \u00E0\u00B5\u0086 \u00DF\u00A3\u00E1\u0088\u00BB\u00DC\u00A3\u00DC\u00A7\u00E1\u0088\u00BE\u00E1\u0088\u00BA\u00DC\u00AE\u00DD\u008F \u00E0\u00B5\u0086 \u00DD\u0080\u00DD\u0096\u00E1\u0088\u00BB \u00C2\u00B7 \u00DC\u00A8\u00DD\u0085 \u00E0\u00B5\u0085 \u00DD\u0080\u00DD\u0096 \u00C2\u00B7 \u00DC\u00A8\u00DD\u008F\u00DD\u0085\u00E1\u0088\u00BF where Fsi is the fraction of grain size i in the substrate. Since sediment is leaving the element and therefore the bed elevation is decreasing and since the active layer is assumed to be constant sediment from the substrate is now considered to be part of the active surface layer. Figure 3-4 shows the subsurface, surface and active layer during degradation. The new surface layer fraction at time step t + \u00CE\u0094t is: Equation 3-4: New Surface Layer Fraction \u00DC\u00A8\u00DD\u008A\u00DD\u008F\u00DD\u0081\u00DD\u0093 \u00E0\u00B5\u008C \u00DC\u00B8\u00DD\u0085\u00DC\u00B8\u00DD\u0090\u00DD\u008B\u00DD\u0090\u00DC\u00BD\u00DD\u0088 \u00E0\u00B5\u008C \u00DC\u00B8\u00DD\u0085 \u00DC\u00AE\u00DD\u008F\u00DC\u00A3\u00E1\u0088\u00BA1 \u00E0\u00B5\u0086 \u00DF\u00A3\u00E1\u0088\u00BB 20 Figure 3-3: Conceptual Model of a Gravel Bedded River during Aggradation Note: The bed load mixes with the surface layer to form a new gravel size distribution. The surface layer thickness is assumed to be a constant thickness Ls. Source (Kwan, 2009) Figure 3-4: Conceptual Model of a Gravel Bedded River during Degradation Note: The surface mixes with the subsurface layer to form a new gravel size distribution. The surface layer thickness is assumed to be a constant thickness Ls. Source (Kwan, 2009) The introduction of moveable bedforms into R2DM was created by using an up-winding factor. The up-winding factor calculates the sediment flux in each triangular element by taking into account the flux flowing in and out from the neighbouring elements allowing for the bed elevation to be changed accordingly. 21 Equation 3-5: Up-Winding Scheme Used in R2DM \u00DD\u008D\u00DD\u008F \u00E0\u00B5\u008C \u00E1\u0088\u00BA1 \u00E0\u00B5\u0086 \u00DC\u00B7\u00DC\u00B9\u00E1\u0088\u00BB \u00C2\u00B7 \u00DD\u008D\u00DC\u00A6\u00DC\u00B5 \u00E0\u00B5\u0085 \u00DC\u00B7\u00DC\u00B9 \u00C2\u00B7 \u00DD\u008D\u00DC\u00B7\u00DC\u00B5 where UW is the up-winding weighting factor and is a value between 0 and 1, qs is the sediment flux, qDS is the sediment flux from the downstream direction and qUS is the sediment flux from the upstream direction. To stabilize the secondary flow calculations Kwan (2009) developed a method which uses the streamlines generated by the Cumulative Discharge function of River2D to compute a radius of curvature at each node. Secondary flow occurs in around bends in natural channels due to centripetal forces which cause the upper layers of the water column to be forced to the outer bank. In order to maintain mass conservation the lower layers of the water column are therefore forced to the inner bank of the bend. These forces in flow result in erosion along the outer bank and deposition along the inner bank of a bend as shown on Figure 3-5. Figure 3-5: Secondary Flow Illustration Secondary flow causes the direction of the bed shear stress to deviate from the direction of the mean depth averaged flow velocity as shown on Figure 3-6. 22 Figure 3-6: Change in Bed Shear Angle Due to Secondary Flow 2D models are not able to simulate this 3D secondary flow affect; therefore R2DM uses the Cumulative Discharge function of River2D to aid in the calculation of the radius of curvature and therefore to calculate the effects of secondary flow. 23 4. SEYMOUR RIVER STUDY SITE A study reach on the Seymour River in North Vancouver was selected to evaluate R2DM. This site was selected since it was a gravel bed river which underwent morphological changes due to a series of high flow events and the bed topography of the site was recorded both before and after the storm events. A water level gauging station was also located upstream of the site allowing for the size of the storm events to be determined. The combination of these factors provided the opportunity to evaluate R2DM based on real field data rather than flume data as it had previously been tested against. 4.1. Watershed Background The Seymour River watershed is located in British Columbia, Canada. Its headwaters are in the Coastal Mountains, its mouth is in North Vancouver it discharges into the Burrard Inlet. Elevations in the basin range from 1,727 m at Mount Cathedral to sea level at Burrard Inlet. The total watershed area is 188 km2, with 67% of the basin area located upstream of the dam as shown on Figure 4-1. The channel is relatively straight and entrained with a high degree of coupling in the headwaters which decreases in the downstream direction. Most of the urban development is located at the downstream end of the basin within the alluvial fan. The watershed is home to one of the main drinking water supplies for the Greater Vancouver Area. The Seymour Falls Dam, completed in 1961, is located 17 km upstream of the mouth and impounds Seymour Lake which only serves to retain drinking water and not to contain any specified peak flow event. Just downstream of the Seymour Falls Dam is a hatchery maintained by the Seymour Salmonid Society in cooperation with Metro Vancouver and the Department of Fisheries and Oceans (DFO). This hatchery supplements the natural salmon spawning and rearing in the Seymour River. 24 Figure 4-1: Seymour River Watershed Location 4.2. Hydrology Water Survey of Canada (WSC) currently maintains a gauging station on the Seymour River located close to the project site, 08GA030 \u00E2\u0080\u0093 Seymour River near North Vancouver, which has 80 years of recorded peak daily flows. The flow regime of the Seymour catchment is a mixed regime meaning that peak flows are caused by both rainfall and rain-on-snow events. The highest peak flow events in the Seymour catchment tend to happen during high rainfall events between October and January. The monthly mean discharges based on the entire record of data from the WSC gauge are shown on Figure 4-2. 25 Figure 4-2: Mean Monthly Discharges at WSC Station 08GA030 Daily peak flows from the WSC station were plotted using the Weibull plotting position formula (Wiessman and Lewis, 2003) to determine a flood frequency curve summarized in Table 4-1 and shown on Figure 4-3. The highest instantaneous flow on record occurred on October 31, 1981 with a magnitude of 650 m3/s. Table 4-1: Estimated Flood Frequency Return Periods and Flows Return Period (yr) Q (m3/s) 20 345 10 304 5 248 2 186 Mean Annual 113 26 Figure 4-3: Flood Frequency Analysis Based on Water Survey of Canada's Gauging Site 08GA030 4.3. Seymour River Morphology In the Seymour River watershed there is a bedrock incised canyon approximately 4.5 km upstream of the Burrard Inlet. This canyon separates the downstream urban reach of the river from the relatively natural upper reach of the river (Lian and Hickin, 1993). Downstream of the bedrock canyon the stream enters into more urban development. This portion of the river has been confined and constrained. Portions of the river banks have been hardened with bank protection works and the development in the floodplain has resulted in local changes in topography leading to the loss of an active floodplain. Due to the regulation of the Seymour River beginning in 1928 with the completion of the Seymour Dam, sediment transport from the upper reach of the watershed has been removed. All of the sediment transport since the construction of the dam has been the result of contributions 27 from the tributary streams and floodplain deposits in the lower 17 km. Moving from upstream to downstream there are fewer gravel bars exposed at low water levels, this indicates that less sediment transport is occurring at the bottom reaches of the channel. 4.4. Project Site and History The project site is located on the alluvial fan approximately 1.4 km upstream from the outlet of the Seymour River and between Mt Seymour Parkway and Dollarton Highway in North Vancouver, British Columbia. A water level gauging site was installed on the Grantham Place Bridge located 650 m upstream of the project site as shown on Figure 4-4. Figure 4-4: Project Location Contaminated soil was eroding into the river from the right downstream bank of the project site. Kerr Wood Leidal (KWL), an environmental consulting firm in Burnaby, British Columbia, was engaged by the District of North Vancouver to design bank protection works to minimize the erosion of the bank and to remove the contaminated soil. To protect the bank KWL used riprap and five riprap bendway spurs. In order to complete the work along the right downstream bank 28 the river had to be diverted away from the bank. The design was to divert water through a temporary diversion channel through an existing gravel bar along the left downstream bank. During the processing of the application for these instream works the Department of Fisheries and Oceans (DFO) approached KWL and suggested redesigning the temporary diversion channel into a side channel for fish habitat in order to increase the total physical habitat through the reach. The final design at the site included five bendway spurs and a 200 m riprap revetment along the right downstream bank, a side channel with pools, riffles and a long run through the existing gravel bar along the left downstream bank, a backwater channel downstream of the side channel and localized boulder clusters located throughout the site as shown on Figure 4-5. The habitat features of the side channel were not picked for any specific species but were chosen to generally decrease velocities and increase water depths. A series of five art rocks were also placed on the site. The District of North Vancouver (DNV) commissioned an artist to produce the art work that would represent erosion and natural river processes. The rock furthest away from the water\u00E2\u0080\u0099s edge was very clean cut and squared, while the rock closest to the water\u00E2\u0080\u0099s edge was very rounded and roughed up as if time and erosion had caused this. The intermediate rocks in the middle represented the progression of the erosion and rounding of the rock. A 10 m stretch or riprap was placed along the side channel to help protect the bank where the art rocks were situated. The location of the art rocks and a photograph are shown on Figure 4-6. 29 Figure 4-5: Project Features; a) Site Pre-Instream Works and b) Site Post-Instream Works Note: the post-instream works air photo was taken after the bed morphological changes of the winter 2008/2009 season. 30 Figure 4-6: Location of Art Rocks The project was completed in September 2008. At this time a ground elevation survey was completed to detail the project. During the winter of 2008/2009 a series of flood events, including an event close to the mean annual flow, passed through the site causing the topography to change. The bed changes which occurred at the Seymour River site made it a excellent candidate to beta test R2DM. 31 5. METHODOLOGY 5.1. Design Flow Event A water level gauging station was installed by KWL on the Seymour River at the Grantham Place Bridge on August 15 2008 as shown on Figure 4-4. The gauging station recorded water depths every five minutes, providing a detailed record of water levels through the storm events. An empirical rating curve was developed at the gauging site by measuring flow discharges at various water depths. Discharge measurements were made along the same cross section using a Swoffer Velocity Meter and Top Setting Wading Rod at shallower depths and an Acoustic Doppler Current Profiler (ADCP) at deeper depths. These measurements and their associated discharges are summarized in Table 5-1 and shown on Figure 5-1 along with the empirical rating curve. Appendix C contains the field data collected to determine the rating curve at the gauging site. Table 5-1: Seymour River Rating Curve and Measured Discharges Date Water Depth (gauge) y (m) Measured Discharge Q (m3/s) % Difference from Empirical Rating Curve 17 May 2009 0.87 11.5 4.1 3 June 2009 0.94 12.3 -13.5 8 June 2009 0.75 6.21 -13.5 10 June 2009 0.72 6.51 3.6 7 Dec 2009 0.61 4.55 17.0 32 Figure 5-1: Seymour River Empirical Rating Curve The rating curve equation relating water depth to discharge at the gauging site is defined in Equation 5-1. Equation 5-1: Seymour River Empirical Rating Curve Equation Q = 16.8y3.02 Where Q is the discharge in m3/s and y is the water depth in m. All flow measurements were undertaken for discharges less than 13 m3/s, and therefore there is considerable uncertainty when extrapolating to higher flows. The rating curve and water level gauge information was used to estimate the high-water flow events which occurred between the 2008 and 2009 ground elevation surveys. A time series plot of the water level gauge data are shown on Figure 5-2. 33 Figure 5-2: Hydrograph Recorded at the Grantham Place Bridge The events indicated on Figure 5-2 where chosen as the design flows since these flows produced velocities which were able to initiate sediment movement on the site. The details of the events are summarized in Table 5-2. The Water Survey of Canada (WSC) values of discharges at the 08GA030 gauge are also shown on Table 5-2 for comparison. Table 5-2: Design Events Event Date Peak Water Level (m) Peak Discharge (m3/s) Duration (hrs) WSC Gauge Discharge (m3/s) A Oct 4 2008 1.55 63.4 6 40E B Oct 7 2008 1.60 69.9 3 22E C Oct 17 2008 1.58 67.3 8 58.2 D Nov 8 2008 1.88 113.4 20 137 E Nov 12 2008 1.83 104.5 6 90.4 Note: E \u00E2\u0080\u0093 indicates estimated values at the WSC gauge The discharge values at the WSC gauge were not corrected for drainage area and represent daily average discharges unlike the KWL gauge which represents peak discharges. These differences result in the WSC values being lower in magnitude. For events A and B WSC could not provide a definite value and therefore have provided estimated values which are much lower than the 34 KWL gauge values. For event D, the largest event used in the study, there is a 17% difference between the WSC and KWL gauges. It is unclear why there is a difference, however, since the rating curve was developed at the KWL gauging site and since the other flow events appear to be reasonable the flows based on the empirical rating curve developed for the project were used rather than the WSC values. 5.2. At Site Field Data Collection There were four major components to the field data collection: bed elevation surveys, collection of velocities and depths, water surface elevations and sediment grain sizes. The bed elevation surveys and the grain size distributions were used as information to set up the River2D and R2DM simulations. The velocity, depth and water surface elevation data were used to calibrate and validate the River2D model for two separate flow events. 5.2.1. Bed Elevation Surveys A total of three bed elevation surveys were completed: \u00EF\u0082\u00B7 2001 \u00E2\u0080\u0093 pre-project; \u00EF\u0082\u00B7 2008 \u00E2\u0080\u0093 immediately after the instream works were competed; and \u00EF\u0082\u00B7 2009 \u00E2\u0080\u0093 after a series of high flow events which resulted in bed changes through the site. These surveys were completed by a crew from KWL using a total station. Information through the main channel was collected based on cross sections spaced approximately 10 m apart and through the side channel based on cross sections spaced approximately 3 m apart. Detailed and accurate topographic information is essential in 2D modelling (Waddle et al., 2000), therefore major breaklines including bank tops and toes, thalweg, details of the bendway spurs, and other irregularities which would cause hydraulic changes were highlighted in the topographic data collection. The bed elevation surveys are presented on Figure 5-3. 35 Figure 5-3: Bed Surveys \u00E2\u0080\u0093 Data provided by KWL; a) 2001 Pre-Project Bed Elevations, b) 2008 Pre-Storm Bed Elevations and c) 2009 Post-Storm Bed Elevations 5.2.2. Velocities and Depths/Water Surface Elevations In order to calibrate the River2D model velocities and depths were collected at six cross section locations and select random points through the site. Cross section locations where chosen based on different hydraulic characteristics, i.e. upstream and downstream of the site, through a pool and a riffle. Additional points were taken around the bendway spurs and rock clusters to help characterize the flow patterns through these areas. Velocities were measured using a Swoffer Velocity Meter with a Top Setting Wading Rod. To determine a depth averaged velocity at a point a series of velocities were taken through the water column at 10 cm intervals with a minimum of two readings where possible. Point velocities were taken as the average velocity over a 10 second reading window. Depths were determined with the Top Setting Wading Rod. 36 Water surface elevations along the banks were taken using TCR705 Leica total station. Appendix D contains the raw data collected in the field. 5.2.3. Sediment size Two methods were used to determine the grain size distributions at the site: surface sampling for the larger bed surface grains and volumetric bulk sampling for the smaller bedload grain sizes. The grain size distributions used for the R2DM model are shown on Figure 5-4. Figure 5-4: Grain Size Distributions Wolman (1954) pebble counts were completed to collect information regarding the surface grain size distribution through the study site. This method included collecting 100 pebbles at random through a defined region. Pebble counts were completed along the left-downstream bank, along the gravel bar separating the main and side channel, and along the bed of the side channel. A total of six Wolman pebble counts were combined to represent the bed surface. The complete Wolman pebble counts for the surface grain size distributions can be found in Appendix E. An accumulation of bedload material occurred in a low velocity zone at the upstream end of the backwater channel on the site. Material from this area was collected, dried, weighed and sorted. Larger material (>22.6 mm) was sieved and weighed in the field while smaller material (<22.6 mm) was brought back to the lab and sieved using a shaker. The size fractions were then 37 weighed to create a grain size distribution for the bedload material. The bedload material grain size distribution data can be found in Appendix E. The grain size distribution for the riprap used to construct the bendway spurs and the riprap revetment was determined from KWL, (2004) but is represented in the R2DM model as non- erodible areas. A subsurface grain size distribution was assigned to the site by assuming an armour ratio of 4. For mountainous streams in the Pacific Northwest an armour ratio of 4 is considered reasonable (Hassan et al., 2006). 5.3. River2D Set-Up and Calibration R2DM relies on the hydrodynamic calculations preformed in River2D and then uses information regarding the sediment grain size distributions on the site to determine the bed morphological changes. A River2D model was set-up and calibrated based on information gathered at the site. The calibrated River2D model was then used in R2DM to model sediment transport and bed morphology changes. River2D uses a bed file, which contains the topographic information and the roughness, ks, of the site; a mesh file is then set-up and the density of the finite element triangulation is set. These files are then brought into River2D where the hydraulic simulations are completed. 5.3.1. Bed Files The bed files used in River2D and R2DM were based on the bed surveys completed by KWL. A digital elevation model (DEM) was created based on the raw survey data points by importing them into the surface mapping program Surfer TM. To interpolate between data points and create bed contours the Kriging option was used since it has been found to provide the best representation of bed topography for lower density sampling (Formann et al., 2007). The interpolated points were then exported at a 1 x 1 m2 grid for use in River2D. Steffler and Blackburn (2002) recommend using breaklines in the bed program to capture the detail of the site and to insure that hydraulic properties are correctly represented; however, the presence of breaklines in R2DM created an unstable model often causing the simulation to crash. Breaklines were therefore not used, but additional information nodes were added to the 1 x 1 m2 grid bed files to correctly represent the site. For R2DM simpler bed files were used that did not contain the details of the rock clusters and the art rocks located on the left downstream bank. 38 These details were omitted as they were found to produce errors and cause the model to crash during the sediment transport simulations. For the habitat simulations rock clusters where included since they provide important low-velocity features which are critical for the weighted usable area habitat calculations. The difference between the 2008 bed file used for sediment transport simulations versus habitat simulations are shown on Figure 5-5. Figure 5-5: 2008 Bed Elevation Files for a) R2DM versus b) Habitat Simulations 39 5.3.2. Mesh Files A mesh of 4 m2 was used for the R2DM simulations to minimize computation time and errors. The bias parameter value was set to 0.95 allowing for a coarser mesh to be used while still picking up the important instream structures such has the bendway spurs and the pools of the side channel which have an influence on the hydraulics of the site. For the habitat simulations a denser mesh was required so that meso-scale topographic features such as the low-velocity zones created at the downstream end of boulders would be included in the calculation of suitable habitat. Crowder and Diplas, (2000) found that to properly model and represent boulder clusters for local habitat a mesh density around the boulders needed to be between (1.45y)2 to (0.20y)2, where y is the average flow depth, in the vicinity of the boulders. For the study reach a mesh density from 0.1 to 6 m2 would be required for high flows and for low flows 0.03 to 1.5 m2. A uniform mesh density of 2 m2 was used for the entire site with a denser mesh of 0.5 m2 around the boulders. For all simulations additional mesh refinement was required to obtain a mesh quality index of at least 0.42. According to Steffler and Blackburn (2002) an acceptable range of mesh quality index values may be in the order of 0.1 to 0.5. 5.3.3. Calibration Calibration of River2D models is typically completed by comparing modelled water surface elevations or velocities and depths to values measured in the field. The bed roughness can then be altered to better represent the field measurements. R2DM calculates the roughness of the bed surface based on the Manning-Strickler formulation. Equation 5-2: Manning-Stickler ks = C90D90 where D90 is the size of the surface material such that 90% is finer and C90 is a dimensionless constant for a given simulation (typically 2 to 3.5) (Kwan, 2009). Therefore rather than calibrating the model based on the changes to the ks values C90 values were used. Based on the D90 of the natural bed surface equalling 0.21 m and a C90 of 3.1 obtained from Bray, (1980) a ks value of 0.65 was applied to the site. 40 Two separate flow events where used to check the calibration of the River2D model. The details of these events are summarized in Table 5-3. Table 5-3: Details of Calibration Flows Event Date Data Collected Discharge (m3/s) A Oct 3 2009 Water surface elevations Velocities, Depths 3.7 B Oct 19 2009 Water surface elevations 25 Only water surface elevations were measured for the Oct 19 event since the water level was too high to safely enter the stream and too wide to use the ADCP that was used at the gauging station. The results of the calibrations are shown on Figure 5-6 for a discharge of 3.7 m3/s and Figure 5-7 for a discharge of 25 m3/s. On each of these plots a 1:1 relationship line is shown representing where the data points would fall if the relationship between the modeled and measured values were identical. The mean absolute error (MAE) and mean error (ME) were also calculated for each flow and were calculated given the following equations (velocity forms presented): Equation 5-3: Mean Absolute Error \u00DC\u00AF\u00DC\u00A3\u00DC\u00A7 \u00E0\u00B5\u008C 100\u00DD\u008A \u00E0\u00B7\u008D\u00E0\u00B8\u00AC \u00DD\u0091\u00E0\u00AF\u00A0\u00E0\u00AF\u00A2\u00E0\u00AF\u0097 \u00E0\u00B5\u0086 \u00DD\u0091\u00E0\u00AF\u00A2\u00E0\u00AF\u0095\u00E0\u00AF\u00A6 \u00DD\u0091\u00E0\u00AF\u00A2\u00E0\u00AF\u0095\u00E0\u00AF\u00A6 \u00E0\u00B8\u00AC Equation 5-4: Mean Error \u00DC\u00AF\u00DC\u00A7 \u00E0\u00B5\u008C 100\u00DD\u008A \u00E0\u00B7\u008D\u00E0\u00B5\u00AC \u00DD\u0091\u00E0\u00AF\u00A0\u00E0\u00AF\u00A2\u00E0\u00AF\u0097 \u00E0\u00B5\u0086 \u00DD\u0091\u00E0\u00AF\u00A2\u00E0\u00AF\u0095\u00E0\u00AF\u00A6 \u00DD\u0091\u00E0\u00AF\u00A2\u00E0\u00AF\u0095\u00E0\u00AF\u00A6 \u00E0\u00B5\u00B0 Where umod is the modelled velocity, uobs is the observed velocity and n is the number of samples. 41 Figure 5-6: Observed versus Modelled Calibration Plots for Water Discharge of 3.7m3/s; a) Water Surface Elevation, b) Velocity, c) Unit Discharge, and d) Water Depth 42 Figure 5-7: Observed versus Modelled Calibration Plot for Water Discharge of 25 m3/s Water Surface Elevations The MAE and ME for the water surface elevations and unit discharge indicate that River2D predicts the water surface elevations and unit discharges well. The errors associated with the depths and velocities are much higher. The misrepresentation of depth and velocity may be related to the large grain sizes observed at the site. When collecting point measurements the instruments may have been placed on a large rock that was not represented in the model bed elevation, therefore providing incorrect depths and velocities associated with that measurement. 5.4. RD2M R2DM was run using the Wilcock and Crowe (2003) sediment transport equation. Grain size distributions for the feed (bedload), surface and subsurface were defined as shown on Figure 5-4. The active layer was defined as 2\u00C2\u00B7D90 of the surface being 0.51 m. The C90 factor was set to 3.1 which based on Equation 5-2 and the surface D90 results in a ks value of 0.65. The unwinding factor was set to 0 since moveable bedforms should not be occurring through this study reach. The secondary flow correction was utilized by setting the number of nodes between points to 10 nodes. To help with the stability of the simulation the bed elevation boundary conditions were set to fixed elevations for both the inflow and outflow boundaries. 43 A feed of 0 m2/s was set as the sediment supply entering at the upstream end of the simulation. The construction of the Seymour Falls Dam, which cuts off the sediment supply from the upper reaches, and the urbanization through the lower reaches which has lead to bank hardening, have caused the loss of sediment transport to the site. An input of 0 m2/s was therefore selected since the Seymour River is a relatively inactive river with respect to sediment transport and lateral instability on the alluvial fan (KWL, 2004). 5.5. Habitat Modelling The target species chosen for the habitat simulations were Coho Salmon fry and spawners and Steelhead Trout juveniles and adults. These species were picked based on recommendation from the Seymour River Hatchery since they are native to Seymour River and would best represent the diversity in habitat needs (Walls, 2009). Habitat suitability curves for velocity, depth and substrate used in the River2D model were created by Bovee (1978) and are illustrated on Figure 5-8 and Figure 5-9. Figure 5-8: Habitat Suitability Curves - Coho Salmon 44 Figure 5-9: Habitat Suitability Curves - Steelhead Trout To assess the changes in available habitat three different cases were modelled. A baseline of avalaible habitat was determined by modelling the pre-construction case using the 2001 bed geometry. This case was used to determine what habitat was avalaible before any instream strucutures were put in place. The post-construction case, the 2008 bed geometry highlights how the instream strucures changed the avalibale physical habitat. Finally a post-storm, the 2009 bed geometry, was also modelled to show how sediment movement and natural bed changes changed the habitat through the reach. Since the design intent of the instream structures was to decrease velocities and increase pool depth the effects of substrate suitabilities were not used for the WUA calculations. To do this an optimal substrate suitability was set for each case and for each species therefore based the WUA calcuations purely on the velocity and depth changes. 45 6. RESULTS AND DISCUSSION 6.1. Bed Morphology Changes The site was surveyed in February 2009 to help identify the morphological changes which took place due to the storm events. The results of the bed surveys and the changes in bed topography are summarized on Figure 6-1. Photographs of the site and the bed changes can be found in Appendix B. Figure 6-1: Changes in Bed Morphology; a) Observed 2008 Bed Elevations, b) Observed 2009 Bed Elevations and c) Observed Change in Bed Elevations Throughout the site there are some minor changes which have resulted in the channel bed either aggrading or degrading. There were minimal changes around the bendway spurs along the right downstream bank. Around the third bendway spur from the upstream end of the site there was 46 some sediment accumulation in a pool upstream of the spur. The rest of the bendway spurs are generally the same. The pools in the side channel have had some sediment accumulation resulting in slightly smaller, shallower pools. There has also been a widening through the lower reach of the side channel increasing the channel width from approximately 4 m to 6 m. The most notable change is the redirection of the side channel at its outlet. The side channel was built to outlet to the backwater channel, however, after the storm events which occurred the side channel now outlets to the main channel as shown on Figure 6-1 around Northing 1700 m. The change of direction is the result of bedload sediment accumulation at the original mouth of the side channel which led to the formation of a natural dam. The natural dam then forced the side channel to be redirected towards the main flow in the channel. The backwater channel has also had some sediment accumulation resulting in a higher elevation through this section. Along the banks of the site, especially along the right downstream bank at a Northing of 1700 m and an Easting of 8420 m there are some artificial changes in bank. In this section there appears to be an increase in bed elevation of approximately 1 m, however, in reality this was not caused by sediment moving through the site, but is actually a result of how the survey points were interpolated. Survey points were much denser through the active part of the channel to help capture real bed change with a point spacing of 2.5 m, but were scarcer outside of the active bed with a point spacing of 14 m. 6.2. R2DM The purpose of this study was to evaluate R2DM\u00E2\u0080\u0099s capability to simulate bed changes which occurred due to sediment transport over a series of high flow events on the Seymour River. R2DM gives various options and settings to aid in the simulation of different river scenarios, both natural and man-made. Different bed roughness, ks, values may be linked to different grain size distributions therefore allowing sediment sizes to be spatially distributed through the model. Being able to vary grain size distributions allows for a wide range of river stages to be modelled, such as pool-riffle systems or riprap lined banks. Non-erodible areas (NEA) can be applied within the model to represent materials which cannot be eroded, for example bed rock or concrete structures. For this study various combinations of R2DM options were first tested in an attempt to try and determine the optimal combination to increase the chances of mimicking the 47 actual bed changes. Different combinations involved spatially varying the grain size distributions to represent the riprap versus the natural bed surface areas and the use of non-erodible areas (NEA). Each combination was initially modelled using only the largest flow event shown in Table 5-2 with a discharge of 113.4 m3/s and duration of 20 hours. Once the best combination was determined, all five design flows were applied in the model to try and represent the bed changes. The different combinations of options which were modelled are as follows: Table 6-1: Summary of R2DM Combination Runs Run Bed Roughness, ks (m) / D50 (mm) Riprap Roughness, ks (m) / D50 (mm) NEA Description 1 Set to riprap = 2.98 / 660 Set to riprap = 2.98 / 660 None 2 Set to natural bed = 0.65 / 74 Set to natural bed = 0.65 / 74 None 3 Set to natural bed = 0.65 / 74 Set to riprap = 2.98 / 660 None 4 Set to natural bed = 0.65 / 74 Set to riprap = 2.98 / 660 Riprap areas 5 Set to natural bed = 0.65 / 74 Set to natural bed = 0.65 / 74 Riprap areas Run\u00C2\u00A01\u00C2\u00A0 First, the entire study site was set to having a bed roughness and grain size distribution of the riprap. The riprap was designed for a 200 year return period event and a discharge of 808 m3/s. As the discharge used in the R2DM model is smaller than the riprap design flow, sediment should not move during the simulation. The final bed elevation and bed elevation difference between the 2008 bed survey and model are shown on Figure 6-2. 48 Figure 6-2: R2DM Simulation Assuming a Uniform Riprap Grain Size Distribution over the Entire Study Reach; a) Post-Simulation Bed Elevations and b) Change in Bed Elevations The results of this simulation show that no bed changes occurred when the entire site is set to the grain size equivalent to that of the riprap. These results are expected since the riprap was sized for a much larger event and implies that R2DM recognizes that this grain size is non-moveable at this flow. Run\u00C2\u00A02\u00C2\u00A0 To test how R2DM would model the more moveable surface material the entire site was set to have a grain size distribution equal to that of the bed surface. The final bed elevation and bed elevation differences between the 2008 bed survey and model are shown on Figure 6-3. 49 Figure 6-3:R2DM Simulation Assuming a Natural Bed Surface Grain Size Distribution over the Entire Study Reach; a) Post-Simulation Bed Elevations and b) Change in Bed Elevations By modelling the entire surface as the more moveable grain size distribution, erosion occurred along the right downstream bank where the non-erodible riprap is located. The riprap bendway spurs were eroded away and as a result the deep pools just downstream of the spurs were filled in. This created a smooth bank, a result which was not seen at the actual project site. Run\u00C2\u00A03\u00C2\u00A0 For the next combination both the riprap and the bed material were simulated with their corresponding grain sizes. R2DM allows the grain sizes to be spatially distributed through the model by linking a distribution to a ks value. A R2DM model was set up by defining the bed 50 surface with a ks of 0.65, the riprap regions with a ks of 2.98 and the corresponding grain size distributions. The results of this simulation are shown on Figure 6-4. Figure 6-4: R2DM Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank; a) Post-Simulation Bed Elevations and b) Change in Bed Elevations In this run the spurs along the right downstream bank were eroded. This is concerning since the riprap and spurs were modelled as the non-erodible riprap yet this combination of model inputs of varying grain sizes allowed those spurs to erode. At the interface between the two different grain sizes the model is calculating a new, previously undefined grain size distribution. R2DM calculates the grain size distribution at an element based on its surrounding elements and the sediment sizes moving in and out of that element. The method in which R2DM calculates the change in grain size distribution caused a fining of the grain sizes through the riprap areas 51 making those regions more erodible in the model. Figure 6-5 shows how in the pre-model run the D50 grain size along the right downstream bank ranging from 0.4 to 0.6 m where as in the post run this same region ranges from 0.08 to 0.1 m. The reduction in the D50 is most likely contributing to the erosion of the spurs and riprap revetment. Figure 6-5: Pre- and Post-Run D50 Grain Size Distributions for Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank; a) Initial D50 Surface Distribution and b) Final D50 Surface Distribution Run\u00C2\u00A04\u00C2\u00A0 To eliminate the erosion of the riprap, non-erodible areas (NEA) were applied to these zones. In R2DM, applying NEA to nodes means that the elevation of those nodes cannot go below the original elevation set in the bed file. NEA areas still allow for sediment to be deposited with in these zones and allow for the grain size distributions to be updated, but the minimum elevation is 52 set. Figure 6-6 shows the results of a R2DM simulation using varying ks values and grain size distributions and with NEA being applied to the riprap zones. Figure 6-6: Simulation Assuming a Natural Bed Surface Grain Size Distribution with Riprap along the Right Downstream Bank and Applying NEA to the Riprap Areas; a) Post-Simulation Bed Elevations and b) Change in Bed Elevations In these results, sediment is being deposited on the downstream side of the spurs raising the bed elevation resulting in larger spurs. R2DM is depositing sediment into low-velocity zones created at the downstream end of the spurs and raising the bed elevation. 53 Run\u00C2\u00A05\u00C2\u00A0 Finally a simulation was set up assuming a uniform grain size distribution of the natural bed and defining the riprap areas with NEAs. The final bed elevation and bed elevation change results for this simulation are found on Figure 6-7. Figure 6-7: Simulation Assuming a Natural Bed Surface Grain Size Distribution Over the Entire Site and Applying NEA to the Riprap Areas; a) Post-Simulation Bed Elevations and b) Change in Bed Elevations These results are not very different from the previous combination assuming varying grain sizes with NEA for the riprap, however there does not appear to be as much sediment accumulation behind the spurs with this simulation. This combination of bed grain size distribution and NEA from run 5 is the simplest way to setup the R2DM simulation therefore minimizing the 54 simulation time, and it appears to provide the best results of bed morphology changes to compare with the actual 2009 results. 6.3. R2DM with Design Flows To simulate the bed changes which occurred between 2008 and 2009 five different storms were simulated. The simulations were completed by using the 2008 bed survey information as the initial bed elevation condition and then running the first 6 hour storm event at a discharge of 63.4 m3/s. The resulting bed elevation from this simulation was then used as the initial bed elevation for the next design flow. This process was continued so that all of the storm events were modelled in sequence to one another as they occurred in nature but omitting the smaller events in between which were not considered large enough to move sediment on the site. The final modelled bed elevation and bed changes are shown on Figure 6-8, the actual bed elevations from the 2008 and 2009 bed survey and the modelled bed elevations are shown on Figure 6-9 and the actual changes in bed elevation (2009 bed elevation survey minus the 2008 bed elevation survey) and the modelled change in bed elevation (modelled bed elevation minus the 2008 bed elevation survey) are shown on Figure 6-10. To empathize the similarities and differences between the actual bed changes and the modelled bed changes an overlay contour plot showing the 2009 bed survey and modelled elevations are shown on Figure 6-11. 55 Figure 6-8: Results of Modelled Run for all Design Flows; a) Final simulated Bed Elevations and b) Difference between the Inital and Final Bed Elevations 56 Figure 6-9: Comparison of Bed Elevations for Observed Pre- and Post-Storm and Modelled Post Storm; a) Observed 2008 Bed Elevations, b) Observed 2009 Bed Elevations and c) Simulated 2009 Bed Elevations 57 Figure 6-10: Comparison of Observed and Simulated Changes in Bed Elevations; a) Observed Bed Elevation Changes and b) Simulated Bed Elevation Changes 58 Figure 6-11: Comparison of Observed and Simulated 2009 Bed Elevations; a) Overlap of Observed and Modelled 2009 Bed Elevations and b) Difference in Observed and Simulated 2009 Bed Elevations The above figures provide an idea of how R2DM represents sediment transport and bed morphology changes. R2DM is predicting both bed aggradation and erosion through the site. Areas outside of the set NEA areas which have higher velocities, such as just upstream of the spur separating the main and side channel as shown on Figure 6-12 are typically being eroded while areas of lower velocities, such as the zones directly behind the bendway spurs also shown on Figure 6-12 are resulting in sediment deposition. 59 Figure 6-12: Modelled Velocities for a Water Discharge of 113.4 m3/s; a) Modelled Velocities and b) Corresponding Bed Elevation Changes R2DM is predicting an accumulation of sediment at the downstream end of the bendway spurs where the actual bed change plots do not indicate this accumulation. Low-velocity zones are created at the downstream end of the bendway spurs as per their design. River2D and R2DM model the 2D hydrodynamics of these low-velocity zones accurately and R2DM deposits sediment through these zones since the shear stress is lower. However, in reality there is a downward force, (in the z-direction due to the over topping of the spurs), which is not modelled by a 2D simulation. This downward force acts to scour out grains creating holes (Kuhnle et al., 2002). Since R2DM is a 2D and not a 3D simulation one of the limitations of the model is that it is programmed for modelling general bed changes and not local scour. In the simulation sediment is being deposited in the low-velocity zones therefore resulting in bed accumulation 60 directly downstream of the bendway spur, to overcome this issue a 3D model would have to be used. The berm separating the side channel from the main channel shows some erosion in the R2DM model. The upstream end of the berm has eroded approximately 20 m; looking at the actual bed change results there is only minimal erosion of the upstream end of the berm, approximately 1.5 m. At the project site the berm has a few larger rocks at the top of the berm. The sizes of these rocks fit within the grain size distribution assigned to the surface; however the rocks interact in way that creates a stronger point than if they were sitting alone on the surface. R2DM is not capable of modelling individual grain interactions and structures; therefore this area was eroded in the simulation. Around the fourth bendway spur from the upstream end of the site the model is predicting the formation of a bar extending from the right downstream bank across to the center of the main channel. Upstream of this bar there is some local erosion, it appears that the model is taking the sediment from this erosion and depositing it just downstream therefore forming the sediment bar. Figure 6-9 and Figure 6-10 do not show this sediment bar as being part of the actual bed changes which occurred on the site. Around the tip of the fourth bendway spur there are high velocities and a drop in elevation of approximately 0.5 m. The combination of these elements results in an area of high bed shear therefore causing the sediment to be scoured. Lower velocities just downstream of this high velocity zone cause the model to dump the scoured sediment therefore forming the bar in the model. R2DM was able to simulate a change in direction of the outlet of the side channel as shown on Figure 6-9. In 2008 the outlet of the side channel went towards the backwater channel; however in 2009 the outlet had changed orientation and now outlets to the main channel. The R2DM simulation showed the starting of this redirection, however, it is not to the full extent of the actual 2009 survey. Also, the R2DM simulation indicates that the backwater channel underwent some sediment accumulation which is similar to what is shown on the 2009 survey. Channel widening was seen through the side channel close to the outlet approximately at 1740 m N on Figure 6-11b). The R2DM simulation of the site also predicted that some channel widening through this section would occur. In Figure 6-11 it is seen that R2DM did not exactly represent 61 the changes as seen through the 2009 survey, but the fact that R2DM did model these changes indicates that it is able to model general bed changes which are occurring through the site. 6.4. R2DM as a Design Tool During the winter of 2009/2010 another series of storm events passed through the Seymour River site. These storm events caused further channel widening through the side channel and the loss of an art rock placed on the left downstream bank. The bed changes which occurred were not surveyed as they were with the previous year\u00E2\u0080\u0099s storm events however, visual inspection was used to show the alterations through the site. Using the position of the art rocks to gauge changes in the bed a series of R2DM simulations were completed to see if different riprap configurations would have been able to protect the bank and therefore the art rocks. The art rocks were placed on the gravel bar located on the left downstream bank of the site as shown on Figure 4-6. With the storms during the winter of 2008/2009 the side channel started to widen putting the art rock closest to the water\u00E2\u0080\u0099s edge in danger of falling into the water. During the storms of 2009/2010 the side channel widened further and resulted in the first art rock falling into the channel and putting the second rock in danger. Photos of the art rock positions in 2008 and 2010 are shown on Figure 6-13. 62 Figure 6-13: Photos of the Change in the Bank along the Art Rock Location; a) 2008 Art Rock Location and b) 2010 Art Rock Locations There were three large storm events during the 2009/2010 winter season and are summarized in Table 6-2. Table 6-2: Winter 2009/2010 Design Flows Date Water Elevation (m) Duration (hrs) Discharge (m3/s) 31 Oct 2009 2.5 2.5 258.8 16 Nov 2009 2.5 9.0 275.0 21 Dec 2009 2.3 6.5 200.5 R2DM was used to model the bed changes due to these storm flows in the same manner as the 2008/2009 winter storm events discussed in section 6.3. Figure 6-14 shows the positioning of the art rocks when they were first installed in 2008 and their final position after the storm events of the winter 2009/2010 based on R2DM modelling. 63 Figure 6-14: Results of Simulated 2009/2010 Winter Storm Event Bed Elevations Showing Failed Art Rocks; a) Observed 2008 Bed Elevations and b) Simulated 2010 Bed Elevations The location of the 4 m elevation contour is the indication if the art rocks are considered safe or if the bank has failed. In the 2008 plot the 4 m contour line is located on the left of the first art rock indicating that the art rock is still located on the bank. In the 2010 simulation, the 4 m contour is located on the right side of the first art rock, indicating that the art rock is no longer located on the bank but is now part of the river. 64 When the site was designed riprap was simulated by creating a non-erodible zone upstream of the art rocks along the left downstream bank of the side channel to protect the bank and minimize the chance of bank erosion. A total length of 10 m of riprap was installed upstream of the art rocks as shown on Figure 6-14. Since the channel did widen and some of the art rocks were lost to the river, the 10 m stretch of riprap was either not long enough or was not positioned correctly to protect the bank. To test R2DM as a design tool different riprap lengths and configurations were modelled to see if it would have been possible to better protect the art rocks. Different riprap designs which were considered for R2DM modelling included: \u00EF\u0082\u00B7 Design A: the actual riprap design totalling 10 m; \u00EF\u0082\u00B7 Design B: an additional 10 m of riprap extended in the downstream direction for a total of 20 m of riprap; \u00EF\u0082\u00B7 Design C: an additional 20 m of riprap extended in the downstream direction for a total of 30 m of riprap; and \u00EF\u0082\u00B7 Design D: an additional 2 m of riprap for a total length of 12 m with the location shifted further downstream. The results of the R2DM simulations for the above riprap designs are shown on Figure 6-15, Figure 6-16, and Figure 6-17. Each of these plots shows the varying riprap lengths and orientations, the art rock locations and the modelled bed contours. 65 Figure 6-15: Results of R2DM Simulation for Design B; a) Bed Elevations with As-Built Riprap Bank Protection and b) Bed Elevations for the Hypothetical Design B with 20 m of Riprap For Design B all of the art rocks are safe. However, the 4 m elevation contour is located very close to the first art rock therefore putting it in danger of falling in during the next large event. 66 Figure 6-16: Results of R2DM Simulation for Design C; a) Bed Elevation As-Built Riprap Bank Protection and b) Bed Elevations for the Hypothetical Design C with 30 m of Riprap The art rocks are safe for Design C. However using 30 m of riprap is not cost effective. 67 Figure 6-17: Results of R2DM Simulation for Design D; a) Bed Elevation As-Built Riprap Bank Protection and b) Bed Elevations for the Hypothetical Design D with 10 m of Riprap Shifted Downstream For Design D only an additional 2 m of riprap was used but the orientation of the riprap was moved further dowsntream. By moving the riprap the art rocks were saved during the same storm event which caused the first rock to fail at the actual site. Even through all of the above designs do protect the art rocks under these flow conditions, the most optimal design tested was Design D showing that a change in positioning and an additional 2 m of riprap could have secured the art rocks. 68 More configurations and larger storm events could also be tested with R2DM to develop a practical, stable design. By looking at these four designs R2DM showed how it could be used to test, optimize and improve on instream river works. Sediment transport and bed changes are very difficult to predict due to the complexity of the problem, lack of information and limitations of the sediment transport theory. R2DM is a step in the right direction using 2D modelling and the Wilcock and Crowe (2003) sediment transport equation to generally show the possible changes that may occur through a site where instream works have been installed. 6.5. Habitat 6.5.1. Pre- to Post- Instream Works (2001 to 2008) The Bovee (1978) preference curves for Coho Salmon fry and spawners and Steelhead Trout adults and juveniles were used with River2D to determine how the weighted useable area (WUA) changed from pre-instream works in 2001 to post-instream works in 2008. The main goal of the instream features were for velocity and depth, therefore the other physical parameters, such as cover, temperature and substrate preferences were assumed to be optimum to simplify the results. Table 6-3, Table 6-4, and summarize the WUA of the target species computed by River2D for the discharges investigated. Table 6-3: Modelled % WUA for Coho Salmon at Selected Discharges (2001 to 2008) Q m3/s Fry Spawners 2001 2008 % change 2001 2008 % change 2 5.0 10.0 50.0 5.0 3.5 -42.9 5 3.9 7.0 44.3 4.5 4.3 -4.7 25 1.5 1.8 16.7 4.0 3.1 -29 150 0.8 1.3 38.5 0.9 1.3 30.8 69 Table 6-4: Modelled % WUA for Steelhead Trout at Selected Discharges (2001 to 2008) Q m3/s Adults Juveniles 2001 2008 % change 2001 2008 % change 2 1.8 1.2 -50 14.6 16.5 11.5 5 5.8 5.6 -3.6 16.4 17.4 5.7 25 7.1 9.6 26.0 11.4 10.0 -14 150 1.6 3.0 46.7 3.1 5.1 39.2 Figure 6-18: Percent WUA for Coho Salmon and Steelhead Trout 2001 to 2008; a) WUA Results for Coho Salmon and b) WUA Results for Steelhead Trout The main goal of the instream works was to increase the area of low velocity zones (LVZ) to provide refuge for fish from the faster flows. Instream works included deep pools, boulder clusters, and a backwater channel. As the discharge increases through the site due to higher flow events velocities also increase. Fish must find refuge behind boulder clusters and in other LVZ to avoid being swept downstream. The instream structures installed at this site created LVZ on their downstream side, however at lower flows these structures are not as important since the velocities in the river are not as fast and fish can use more of the channel. 70 Figure 6-18 indicates that Coho Salmon prefer lower flows while the Steelhead Trout benefit more during medium size events. The instream structures perform best for Coho Salmon at lower flows by further increasing areas of lower velocity; however at high flows the instream structures do not provide refuge for Coho Salmon. Steelhead Trout get most of their benefits during medium sized events. The percent difference in WUA ranges from +50% to -50% depending on the target species and the flow (Table 6-3 and Table 6-4). The greatest increase in WUA was for Coho fry at a low flow. For the Coho fry the increase in WUA at the lowest flow is the result of the Coho fry\u00E2\u0080\u0099s preference for very low velocities and shallower depths. Figure 6-19 shows the Coho Salmon fry velocity suitability as modelled by River2D at the lower discharge of 2 m3/s. The instream structures and the changes to the reach act to decrease the velocities at all of the tested flows. However due to the narrow range for velocity suitable for Coho Salmon fry at the larger flows there is not much of an increase in WUA. Figure 6-20 shows the modelled velocities for 2001 and 2008 geometries at a flow of 150 m3/s. The velocities shown on the 2008 figure are lower. The average velocity through the site has been reduced. However, when you look at the velocity suitability of the Coho Salmon fry at the same discharge for both the 2001 and 2008 geometries as shown on Figure 6-21 there is little difference between the two. This is due to the narrow range of velocity suitability of the Coho Salmon fry. Even though the instream works of the 2008 geometry decrease the velocities through the site, there is not enough of a reduction to satisfy the needs of the Coho Salmon fry. 71 Figure 6-19: Velocity Suitability at 2 m3/s for Coho Fry; a) 2001 Pre-Instream Structures and b) 2008 Post Instream Structures 72 Figure 6-20: Modelled Velocities at a Discharge of 150 m3/s; a) Observed 2001 Topography and b) Observed 2008 Topography 73 Figure 6-21: Modelled Velocity Suitability for Coho Salmon Fry at a Discharge of 150 m3/s; a) Observed 2001 Topography and b) Observed 2008 Topography In general the instream works installed at the Seymour River site were designed to decrease velocites and increase depths which they did managed to do this at various river discharges. However, there is not a huge increase in the WUA for the site for any of the chosen indicator species. Since the instream works were designed without a specific species target the balance between depth and velocity suitabilities may have been missed. More dramatic habitat improvements could have been obtained if a specific target species was identifiied before the instream works were designed. 74 6.5.2. Pre- to Post- Flooding Event (2008 to 2009) A habitat assessment was also completed on post-flooding, 2009 topography. The same discharges were selected as in the section above and the WUA are summarized in Table 6-5, Table 6-6 and on Figure 6-22. Table 6-5: Modelled % WUA for Coho Salmon at Selected Discharges (2008 to 2009) Q m3/s Fry Spawners 2008 2009 % change 2008 2009 % change 2 10 8.5 -17.6 3.5 4.1 14.6 5 7.0 7.2 2.8 4.3 4.2 -2.4 25 1.8 2.6 30.8 3.1 2.4 -29.2 150 1.3 1.5 13.3 1.3 0.7 -85.7 Table 6-6: Modelled % WUA for Steelhead Trout at Selected Discharges (2008 to 2009) Q m3/s Adults Juveniles 2008 2009 % change 2008 2009 % change 2 1.2 1.1 -9.1 16.5 18.4 10.3 5 5.6 6.3 11.1 17.4 19.7 11.7 25 9.6 10.0 4.0 10.0 11.2 10.7 150 3.0 2.9 -3.4 5.1 4.2 -21.4 75 Figure 6-22: Percent WUA for Coho Salmon and Steelhead Trout 2008 to 2009; a) WUA Results for Coho Salmon and b) WUA Results for Steelhead Trout The differences in WUA results between the 2008 and 2009 topographies are due to the channel bed shifting. These changes alter the velocities and depths though the site. The final bed elevations from the flooding events which occurred during 2008/2009 generally acted to improve the WUA for all of the indicator species through the site. No major changes in habitat occurred except for the Coho Salmon spawners which had a loss of 86% of WUA at a discharge of 150 m3/s. This change is a result of the redirection of the outlet of the side channel. The redirection of the side channel caused the loss of some of the side channel length. The original side channel design provided areas of suitable depth for the Coho Salmon spawners. With the loss of this length there was also a loss in area of suitable depth, and therefore a decrease in WUA. By modelling not only the changes in habitat pre- and post-instream structures but also the changes in habitat due to sediment transport allows a deeper evaluation of the site design. Little change in WUA due to sediment transport were seen at the Seymour River site with these flows, but the possibility of sediment transport affecting the instream structures can be great. Changes in the reach may cause the instream structures to become non-functional. Sediment may fill in pools or erosion may destroy riffles. By looking at how sediment transport and bed morphology changes interact with instream habitat structures more efficient, longer lasting designs may be created. 76 6.5.3. Post- Flooding Event (2009 Observed versus 2009 Simulated) A habitat assessment was also completed to compare the differences between the observed 2009 topography to the simulated 2009 topography based on the R2DM simulation. The same discharges were selected as in the above sections and the WUA are summarized in Table 6-5, Table 6-6 and on Figure 6-22. Table 6-7: Modelled % WUA for Coho Salmon at Selected Discharges (2009 Observed to 2009 Simulated) Q m3/s Fry Spawners 2009 Observed 2009 Simulated % change 2009 Observed 2009 Simulated % change 2 10.6\u00C2\u00A0 5.8\u00C2\u00A0 \u00E2\u0080\u009082.8\u00C2\u00A0 4.7\u00C2\u00A0 3.5\u00C2\u00A0 \u00E2\u0080\u009034.3\u00C2\u00A0 5 7.4\u00C2\u00A0 4.3\u00C2\u00A0 \u00E2\u0080\u009072.1\u00C2\u00A0 5.5\u00C2\u00A0 4.5\u00C2\u00A0 \u00E2\u0080\u009022.2\u00C2\u00A0 25 2.5\u00C2\u00A0 1.7\u00C2\u00A0 \u00E2\u0080\u009047.1\u00C2\u00A0 3.3\u00C2\u00A0 1.3\u00C2\u00A0 \u00E2\u0080\u0090153.8\u00C2\u00A0 150 1.5\u00C2\u00A0 1.2\u00C2\u00A0 \u00E2\u0080\u009025.0\u00C2\u00A0 1.6\u00C2\u00A0 1.1\u00C2\u00A0 \u00E2\u0080\u009045.5\u00C2\u00A0 Table 6-8: Modelled % WUA for Steelhead Trout at Selected Discharges (2009 Observed to 2009 Simulated) Q m3/s Adults Juveniles 2009 Observed 2009 Simulated % change 2009 Observed 2009 Simulated % change 2 1.1\u00C2\u00A0 1.2\u00C2\u00A0 8.3\u00C2\u00A0 21.9\u00C2\u00A0 13.6\u00C2\u00A0 \u00E2\u0080\u009061.0\u00C2\u00A0 5 5.9\u00C2\u00A0 4.4\u00C2\u00A0 \u00E2\u0080\u009034.1\u00C2\u00A0 23.6\u00C2\u00A0 14.3\u00C2\u00A0 \u00E2\u0080\u009065.0\u00C2\u00A0 25 8.3\u00C2\u00A0 2.7\u00C2\u00A0 \u00E2\u0080\u0090207.4\u00C2\u00A0 12.2\u00C2\u00A0 5.7\u00C2\u00A0 \u00E2\u0080\u0090114.0\u00C2\u00A0 150 2.8\u00C2\u00A0 1.7\u00C2\u00A0 \u00E2\u0080\u009064.7\u00C2\u00A0 5.9\u00C2\u00A0 4.3\u00C2\u00A0 \u00E2\u0080\u009037.2\u00C2\u00A0 77 Figure 6-23: Percent WUA for Coho Salmon and Steelhead Trout 2009 Observed to 2009 Simulated; a) WUA Results for Coho Salmon and b) for Steelhead Trout The R2DM simulated sediment transport and bed morphology changes resulted in different levels of erosion and therefore depths and velocities compared to the observed 2009 changes. Since the habitat simulations are heavily based on predictions of depth and velocity, the differences between the observed and simulated 2009 topographies resulted in discrepancies in the habitat predictions of up to a loss of 207% in WUA. These results indicate that great care must be taken when performing R2DM simulations to aid in the predication of habitat since errors in depths and velocities can have a large impact on the simulated results. 78 7. CONCLUSIONS 7.1. Evaluation of R2DM R2DM is able to model sediment movement and changes in bed topography. In general it simulates deposition in areas of lower shear stress and erosion in areas of higher shear stress. The full extent of bed changes which occurred at the Seymour River site were not predicted by R2DM, but the model was able to simulate broad changes and patterns which did occur. That being said, the general changes predicted by the model are still very useful and provide a large amount of insight into the processes occurring at the site. Sediment transport and bed changes are very complex processes. Generally sparse field data and limitations of the sediment transport theory often lead to large differences between prediction and actual outcome. The interactions between grains and the formation of bed structures is an important factor in sediment transport, but is a very difficult element to predict and is a short coming in most sediment transport theory. It is not surprising that these interactions were not fully modelled by R2DM. R2DM shows potential as a possible design tool. For example, the model was able to predict the widening of the side channel and the danger that this caused for the art rocks located on the bank. As a design tool the model showed different resulting bed topographies based on different riprap designs. While there remains uncertainty with any prediction the model still provides a tool for the useful evaluation of the various designs. Further development of the R2DM model is still required. The spatial distribution of grain sizes and the interaction between nodes of different grain size distributions needs to be further tested. Most of the original beta testing of R2DM and R2D-Morph were completed using flume tests which all started with a uniform bed in terms of grain size distributions and elevations, therefore the test of the model was to see if various bed changes would occur. When modelling real gravel rivers there is no such thing as a uniform bed, therefore the ability to spatially distribute grain sizes distributions would be ideal. Modelling the difference between a pool and riffle not only based on geometry, but also by grain size distributions could produce a much more accurate representation of the site. 79 7.2. Evaluation of Habitat Improvements The amount of physical habitat which was available at the site did increase due to the installation of the instream structures. The addition of bendway spurs, a side channel, local boulder clusters and a backwater channel resulted in an increase in areas with lower velocities and deeper depths for the various discharges tested. Greater improvements to the habitat based on a specific species may have been seen if a target species was picked before the instream structures were designed. However, since a target species was not picked the improvements to habitat are still great. By modelling not only the changes in habitat pre- and post-instream structures, but also the changes in habitat due to sediment transport events, a deeper evaluation of the site design was seen. Due to the bed changes which occurred over the winter of 2008/2009 the WUA through the site was actually increased. This is a positive result to sediment transport; however it is possible that the sediment transport through the site could have reduced the available habitat. Changes may have caused the instream structures to become non-functional. It is possible that sediment may fill in pools or erosion may destroy riffles. By looking at how sediment transport and bed morphology changes interact with instream habitat structures more efficient, longer lasting designs may be created. 80 REFERENCES Addley, C., K. Clipperton, T. Hardy, and A. Locke. 2003. South Saskatchewan River basin, Alberta, Canada - Fish habitat suitability criteria (HSC) curves. Alberta Fish and Wildlife Division, Alberta Sustainable Resource development. Edmonton, Alberta. 63 pp. ISBN-0-7785- 359-4. Bovee, K.D. 1978. Probability-of-use criteria for the family salmoniae. U.S.D.I., Fish and Wildlife Service, Office of Biological Services. Instream Flow Information Paper No.4. Bovee, K.D., B.L. Lamb, J.M. Bartholow, C.B. Stalnaker, J. Taylor, and J. Henriksen. 1998. Stream Habitat Analysis Using the Instream Flow Incremental Methodology. U.S. Department of the Interior, U.S. geological Survey Biological Resource Division. Information and Technology Report: USGS/BRD/ITR-1998-0004. Bray, D.I. 1980. Evaluation of effective boundary roughness for gravel-bed rivers. Canadian Journal of Civil Engineering 7: 392-397. Carr\u00C3\u00A9, D.M., P.M. Biron and S.J. Gaskin. 2007. Flow dynamics and bedload sediment transport around paired deflectors for fish habitat enhancement: a field study in the Nicolet River. Canadian Journal of Civil Engineering, 34: 761-769. Crowder, D.W. and P. Diplas. 2000. Using two-dimensional hydrodynamic models at scales of ecological importance. Journal of Hydrology, 230: 172-191. Doyle, M.W, D. Shields, K.F. Boyd, P.B. Skidmore and D. Dominick. 2007. Channel- Forming Discharge Selection in River Restoration Design. Journal of Hydraulic Engineering, 133(7): 831-837. Formann, E., H.M. Habersack and St. Schober. 2007. Morphdynamic river processes and techniques for assessment of channel evolution in Alpine gravel bed rivers. Geomorphology, 90: 340-355. Garc\u00C3\u00ADa de Jal\u00C3\u00B3n, D. and J. Gort\u00C3\u00A1zar. 2007. Evaluation of instream habitat enhancement option using fish habitat simulations: case-studies in the river Pas (Spain). Aquatic Ecology, 41: 461-474. Gard, M. 2005. Variability in flow-habitat relationships as a function of transect number for PHABSIM modelling. River Research and Applications, 21: 1013-1019. Ghanem, A., P. Steffler, F. Hicks and C. Katopodis. 1996. Two-dimensional hydraulic simulation of physical habitat conditions in flowing streams. Regulated Rivers: Research & Management, 12: 185-200. Gore, J.A., D.J. Crawford and D.S. Addison. 1998. An analysis of artificial riffles and enhancement of benthic community diversity by physical habitat simulation (PHABSIM) and direct observation. Regulated Rivers: Research & Management, 14: 69-77. 81 Hassan, M. A., R. Egozi, and G. Parker. 2006. Experiments on the effect of hydrograph characteristics on vertical grain sorting in gravel bed rivers. Water Resources Research, 42, W09408; doi:10.1029\u00C3\u00A92005WR004707. Hauer, C., G. Unfer, S. Schmutz and H. Habersack. 2007. The importance of morphodynamic processes at riffles used as spawning grounds during the incubation time of nase (Chondrostoma nasus). Hydrobiologia; 579: 15-27. Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology, 61: 65-71. Johnson, D. 2008. MASC Thesis: The application of a two-dimensional sediment transport model in a Cumberland Plateau mountainous stream reach with complex morphology and coarse substrate. University of Tennessee, Knoxville. Kerr Wood Leidal, (KWL). 2004. Design Brief, Construction of Seymour River Bank Protection Works. Prepared for: Seymour River I.R. No. 2, Squamish Nation. King, I.P. 1990. Program Documentation RMA-2V. Two Dimensional Finite Element Model for Flow in Estuaries and Streams. Version 4.3. Resources Management Associates., Lafayette, CA. Khangaonkar, T, S. Breithaupt, and F. Kristanovich. 2005. Restoration of hydrodynamic and hydrologic processes in the Chinook River estuary, Washington \u00E2\u0080\u0093 Feasibility assessment. Estuarine and Coastal Modeling, 735-751. Klumpp, C.C. 2008. Using two dimensional hydraulic models for restoration of Upper Columbia basin streams. World Environmental and Water Resources Congress 2008 Ahumpa\u00E2\u0080\u0099a. Kuhnle, R.A., C.V. Alonso and F.D. Shields, Jr. 2002. Local Scour Associated with Angled Bank Spurs. Journal of Hydraulic Engineering. 128(12): 1087-1093. Kwan, S. 2009. MASC Thesis: A two dimensional hydrodynamic river morphology and gravel transport model. University of British Columbia, Vancouver. Lacey, R.W.J. and R.G. Millar. 2004. Reach scale hydraulic assessment of instream slamonid habitat restoration. Journal of the American Water Resources Association, 40(6): 1631- 1644. Lacey, R.W.J. and R.G. Millar. 2001. Application of a 2-dimensional hydrodynamic model for the assessment and design of instream channel restoration works. Watershed Restoration Management Report No. 9. Watershed Restoration, Ministry of Water, Land and Air Protection and Ministry of Forests, British Columbia, Canada. Lian, O.B., and Hickin, E.J. 1993. \u00E2\u0080\u009CLate Pleistocene stratigraphy and chronology of lower Seymour Valley southwestern British Columbia\u00E2\u0080\u009D, Canadian Journal of Earth Sciences, vol. 30: 841-850. 82 Loranger, J. and S. Kenner. 2004. Comparison of one- and two-dimensional hydraulic habitat models for simulation of trout stream habitat. Critical Transitions in Water and Environmental Resources Management, pp 1-10. Moir, H.J., C.N. Gibbins, C. Soulsby and A.F. Youngson. 2005. PHABSIM modelling of Atlantic salmon spawning habitat in an upland stream: testing the influence of habitat suitability indices on the model output. Rivers Research and Applications, 21: 1021-1034. Mouton, A.M., M. Schneider, J. Depestele, P.L.M. Geothals and N. De Pauw. 2007. Fish habitat modelling as a tool for river management. Ecological Engineering, 29: 305-315. Papanicolaou, A.N., M. Elhakeem, G. Krallis, S. Prakash and J. Edinger. 2008. Sediment Transport Modeling Review \u00E2\u0080\u0093 Current and Future Developments. Journal of Hydraulic Engineering. 134(1): 1-14. Raleigh, R.F., L.D. Zuckerman, and P.C. Nelson. 1986. Habitat suitability index models and instream flow suitability curves: brown trout. Fish and Wildlife Service, U.S. Department of the Interior, Biological Report 82(10.124). Schwartz, J.S. 2003. Use of a 2D hydrodynamic model for stream restoration design of high-flow habitat in low-gradient Midwest streams. Pages 242-251; In M. Clar, D. Carpenter, J. Garcie, and L. Slate (editors). ASCE Proceedings of the Symposium on the Protection and Restoration of Urban and Rural Streams; Philadelphia, PA. June 23-25, 2003. Shuler, S.W. and R.B. Nehring. 1993. Using the physical habitat simulation model to evaluate a stream habitat enhancement project. Rivers, 4: 175-193. Spence, R. and P. Hickley. 2000. The use of PHABSIM in the management of water resources and fisheries in England and Wales. Ecological Engineering, 16: 153-158. Steffler, P.M. and J. Blackburn. 2002. River2D: Two-dimensional depth-averaged model of river hydrodynamics and fish habitats, User\u00E2\u0080\u0099s Manual. University of Alberta. Stewart, G., R. Anderson and E. Wohl. 2005. Two-dimensional modelling of habitat suitability as a function of discharge on two Colorado rivers. River Research and Applications, 21: 1061-1074. Thompson, D.M. 2002a. Long-term effect of instream habitat-improvement structures on channel morphology along the Blacklege and Salmon rivers, Connecticut, USA. Environmental Management, 29(1): 250-265. Thompson, D.M. 2002b. Channel-bed scour with high versus low deflectors. Journal of Hydraulic Engineering, ASCE, 128(6): 640-643. Vasquez, J. (2005). PhD Thesis: Two dimensional finite element river morphology model. University of British Columbia. Vasquez, J., R.G. Millar and P.M. Steffler. 2007. Two-dimensional finite element river morphology model. Canadian Journal of Civil Engineering 34: 726-760. 83 Viessman, Warren and Gary Lewis. Introduction to Hydrology. Upper Saddle River, NJ: Prentice Hall, 2003. Wilcock, P.R. and J.C. Crowe. 2003. Surface-based transport model for mixed-sized sediment. Journal of Hydraulic Engineering, 129(2): 120-128. U.S. Army Corps of Engineers. 2008. HEC-RAS River Analysis System, Hydraulic Reference Manual, Version 4.0. Institute for Water Resources. Waters, B.F. 1976. A methodology for evaluating the effects of different streams flows on salmonid habitat. Pages 254-266 in J.F. Orsborn and C.H. Allman (eds.). Instream flow needs. American Fisheries Society, Bethesda, MD. Waddle, T. 2001. PHABSIM for Windows, User\u00E2\u0080\u0099s Manual and Exercises. Midcontinent Ecological Science Center. Open File Report 01-340. Waddle, T., P. Steffler, A. Ghanem, C. Katopodis and A. Locke. 2000. Comparison of one- and two-dimensional open channel flow models for a small habitat stream. Rivers, 3: 205- 220. Waddle, T. 2009. Field evaluation of a two-dimensional hydrodynamic model near boulders for habitat calculation. River Research and Applications. Walls, E. August 4 2009. Email correspondence. Seymour River Fish Hatchery Wolman, M.G. 1954. A method of sampling coarse river-bed material. Transactions, American Geophysical Union 35 (6): 951-956. Wu, W., W. Rodi and T. Wenka. 2000. 3D Numerical Modeling of Flow and Sediment Transport in Open Channels. Journal of Hydraulic Engineering 126(1): 4-15. Wu, W. 2004. Depth-Averaged Two-Dimensional Numerical Modeling of Unsteady Flow and No uniform Sediment Transport in Open Channels. Journal of Hydraulic Engineering, 130(10): 1013-1024. Wu, Weimings. (2008). Computation River Dynamics. Taylor & Francis- Balkema, Leiden, The Netherlands. 84 APPENDICES Appendix A: Wilcock and Crowe (2003) Sediment Transport Equation The Wilcock and Crowe (2003) sediment transport equation is a surface-based transport model for mixed sand/gravel sediments. The model is cable of predicting transient conditions, incorporates a hiding function and uses the full size distribution of the bed surface to calculate sediment transport. The transport model was developed from coupled observations of flow, transport and bed surface grain sizes in 48 laboratory flume experiments. Five different sand and gravel sediment mixtures were used over nine to ten various flows to produce different sediment transport rates. The runs produced transport rates which differed over four orders of magnitude with a minimum transport rant of 1.8x10-5 kg/m.s and a maximum rate of 1.2x10-1 kg/m.s. Small samples of sediment were removed from the experiments for sieving and calculation of fraction transport rates. The fractional transport rate, qbi, was calculated using the following formula: \u00EF\u0080\u00A8 \u00EF\u0080\u00A9 bibi qpq \u00EF\u0080\u00BD where pi is the proportion of each fraction i in transport and qb is the total transport rate. The transport model was then developed using the similarity collapse over the fractional transport rate. The form of the similarity collapse is: \u00EF\u0080\u00A8 \u00EF\u0080\u00A9 3 * * 1 uF gqsW i bi i \u00EF\u0080\u00AD\u00EF\u0080\u00BD where s is the ratio of sediment to water density, g is gravity, qbi is the volumetric transport rate per unit width of size fraction i, Fi is the proportion of size i on the bed surface and u* is the shear velocity defined as 5.0 * \u00EF\u0083\u00BA\u00EF\u0083\u00BB \u00EF\u0083\u00B9\u00EF\u0083\u00AA\u00EF\u0083\u00AB \u00EF\u0083\u00A9\u00EF\u0080\u00BD \u00EF\u0081\u00B2\u00EF\u0081\u00B4u where \u00CF\u0081 is the density of water. A reference shear stress, \u00CF\u0084ri, was defined by a value of \u00CF\u0084 when W*i=0.002. Plotting the experimental data by the calculated 85 W*i as a function of \u00CF\u0084/ \u00CF\u0084ri a clear trend was detected. A plot of the experimental data and the sediment transport model function are shown on the Figure A-1. Figure A-1: similarity collapse of all fraction transport observations The transport function is as follows: 35.1 35.1 894.0114 002.0 5.4 5.0 5.7 * \u00EF\u0082\u00B3 \u00EF\u0080\u00BC \u00EF\u0083\u00AF\u00EF\u0083\u00AE \u00EF\u0083\u00AF\u00EF\u0083\u00AD \u00EF\u0083\u00AC \u00EF\u0083\u00B7\u00EF\u0083\u00B7\u00EF\u0083\u00B8 \u00EF\u0083\u00B6\u00EF\u0083\u00A7\u00EF\u0083\u00A7\u00EF\u0083\u00A8 \u00EF\u0083\u00A6 \u00EF\u0080\u00AD\u00EF\u0080\u00BD \u00EF\u0081\u00A6 \u00EF\u0081\u00A6 \u00EF\u0081\u00A6 \u00EF\u0081\u00A6 for for Wi where ri\u00EF\u0081\u00B4\u00EF\u0081\u00B4\u00EF\u0081\u00A6 \u00EF\u0080\u00BD The hiding function used in the transport model takes into account the fact that while smaller grains are essentially more mobile than large grains on a mixed grain size bed these smaller 86 particles may be trapped in deep pockets between the larger grains. In gravel beds this may cause equal mobility, which is when the small grains can move just as easily as the large one. The hiding function used by the model is as follows: b s i rs ri D D \u00EF\u0083\u00B7\u00EF\u0083\u00B7\u00EF\u0083\u00B8 \u00EF\u0083\u00B6 \u00EF\u0083\u00A7\u00EF\u0083\u00A7\u00EF\u0083\u00A8 \u00EF\u0083\u00A6\u00EF\u0080\u00BD 5050\u00EF\u0081\u00B4 \u00EF\u0081\u00B4 where Di is the grain size of fraction i, Ds50 is the median grain size of the bed surface and b is defined as: \u00EF\u0083\u00B7\u00EF\u0083\u00B7\u00EF\u0083\u00B8 \u00EF\u0083\u00B6 \u00EF\u0083\u00A7\u00EF\u0083\u00A7\u00EF\u0083\u00A8 \u00EF\u0083\u00A6 \u00EF\u0080\u00AD\u00EF\u0080\u00AB \u00EF\u0080\u00BD sm i D D b 5.1exp1 67.0 where Dsm is the mean grain size of the bed surface. 87 Appendix B: Photographs of the Bed Changes at Seymour River Site Figure B-1: Changes to the Outlet of the Side Channel Photos looking in the upstream direction at the outlet of the side channel. From these photos it is possible to see how the outlet of the side channel redirected from outletting to the backwater channel to outletting to the main channel. 88 Figure B-2: Changes to the Left Bank of the Side Channel Looking in the upstream direction at the position of the art rocks on the left downstream bank. In 2008 there was approximately 2.5 m of bank between the first art rock and the channel bank; in 2009 the first art rock is sitting at the top of the bank. 89 Appendix C: Rating Curve Data Table C-1: 17 May 2009 Discharge Measurement Date: 17 May 2009 Q = 11.5 m3/s WL = 0.87 m instrument ADCP Notes: StationID Depth Mean Velocity Discharge (m) (m/s) (m3/s) LDB 0 0 0 0 1 -0.8 0.06 0.05 2 -0.8 0.053 0.09 3 -0.8 0.166 0.22 4 -0.8 0.186 0.37 5 -0.91 0.433 0.77 6 -0.86 0.513 1.21 7 -0.88 0.398 1.56 8 -0.68 0.113 1.63 9 -0.67 0.122 1.72 10 -0.58 0.234 1.85 11 -0.59 0.241 1.99 12 -1.25 0.011 2.01 13 -1.25 0.031 2.05 14 -1.3 0.393 2.56 15 -0.69 0.46 2.87 16 -0.7 0.494 3.22 17 -1.5 0.312 3.92 19 -0.72 0.524 4.49 20 -0.8 0.69 5.04 21 -0.76 0.579 5.48 22 -0.7 0.498 5.83 23 -0.62 0.471 6.12 24 -0.86 0.835 6.84 25 -0.91 0.925 7.68 26 -0.8 0.372 7.98 27 -0.79 0.811 8.94 29 -0.88 0.701 10.17 31 -0.78 0.694 10.99 32 -0.78 0.436 11.33 33 -0.45 0 11.33 34 -0.58 0.199 11.50 RDB 36 0 0 11.50 90 Table C-2: 3 June 2009 Discharge Measurement Date: 3 June 2009 Q = 11.48 m3/s WL = 0.94 m instrument ADCP Notes: StationID Depth Mean Velocity Discharge (m) (m/s) (m3/s) LDB 0 0 0 0 2 0.40 0.22 0.10 4 0.50 0.29 0.27 6 0.76 0.60 0.99 8 0.73 0.60 1.71 10 0.74 0.20 1.78 12 0.74 0.21 1.87 14 1.28 0.03 1.87 16 1.28 0.36 2.13 18 0.74 0.45 2.73 22 0.74 0.41 3.23 24 0.78 0.80 4.51 26 0.75 0.46 4.93 28 0.81 1.10 7.35 30 0.78 1.30 10.73 32 0.55 0.57 11.38 34 0.35 0.18 11.45 36 0.2 0.16 11.48 RDB 36.5 0 0 11.48 91 Table C-3: 8 June 2009 Discharge Measurement Date: 8 June 2009 Q = 6.21 m3/s WL = 0.75 m instrument ADCP Notes: StationID Depth Mean Velocity Discharge (m) (m/s) (m3/s) LDB 1 0 0 0 2 0.75 0.01 0.01 3 0.75 0.09 0.07 4 0.75 0.06 0.11 5 0.75 0.19 0.26 6 0.76 0.51 0.64 7 0.69 0.52 1.00 8 0.75 0.35 1.26 9 0.59 0.08 1.31 10 0.57 0.04 1.33 11 0.49 0.19 1.43 12 1.10 0.14 1.58 13 1.10 0.09 1.68 14 0.90 0.00 1.68 15 0.75 -0.02 1.67 16 0.66 0.11 1.74 17 1.10 0.27 2.04 18 0.75 0.22 2.20 19 0.63 0.20 2.33 20 0.71 0.25 2.51 21 0.67 0.22 2.66 22 0.74 0.50 3.02 23 0.65 0.42 3.30 24 0.65 0.32 3.51 25 0.65 0.31 3.71 26 0.65 0.29 3.90 27 0.65 0.67 4.33 28 0.65 0.60 4.72 29 0.76 0.54 5.13 30 0.77 0.58 5.58 31 0.74 0.63 6.04 32 0.74 0.22 6.21 RDB 33 0 0 6.21 92 Table C-4: 10 June 2009 Discharge Measurement Date: 10 June 2009 Q = 6.51 m3/s WL = 0.72 m instrument Swoffer Notes: StationID Depth Mean Velocity Discharge (m) (m/s) (m3/s) RDB 0 0 0 0 1 0.07 0.02 0.00 3 0.51 0.50 0.51 5 0.80 0.52 1.34 7 0.52 0.78 2.16 9 0.62 0.64 2.95 11 0.72 0.53 3.71 13 0.73 0.36 4.24 15 0.37 0.41 4.54 17 0.26 0.53 4.82 19 0.25 0.02 4.83 21 0.38 0.47 5.18 23 0.30 0.43 5.44 25 0.32 0.27 5.61 27 0.73 0.52 6.37 29 0.77 0.09 6.51 31 0.31 0.00 6.51 33 0.15 0.00 6.51 LDB 34 0 0 6.51 93 Appendix D: At Site Velocity, Depth and Water Surface Elevation Data Table D-1: October 3 2009 Field Data \u00E2\u0080\u0093 Velocity and Depth 3 Oct 2009 Q = 3.27 m3/s Description Northing Easting Velocity Depth (m) (m) (m/s) (m) inlet LDB 8470.4 1877.6 0.00 0.00 inlet 8469.4 1878.0 0.05 0.14 inlet 8468.8 1878.2 0.69 0.20 inlet 8468.3 1878.4 0.77 0.25 inlet 8467.9 1878.6 0.69 0.33 inlet 8467.2 1878.9 0.27 0.22 inlet RDB 8467.0 1879.0 0.00 0.00 pool1 RDB 8447.5 1822.8 0.00 0.00 pool1 8448.3 1822.8 0.02 0.17 pool1 8449.3 1822.8 0.07 0.54 pool1 8450.3 1822.8 0.11 0.98 pool1 8451.3 1822.8 0.05 0.96 pool1 8452.3 1822.9 0.02 0.77 pool1 8453.3 1822.9 0.00 0.36 pool1 LDB 8455.0 1822.9 0.00 0.00 run1 RDB 8453.9 1795.3 0.00 0.00 run1 8456.0 1795.6 0.13 0.26 run1 8457.0 1795.7 0.34 0.32 run1 8458.0 1795.8 0.38 0.34 run1 8459.0 1795.9 0.06 0.24 run1 LDB 8460.4 1796.1 0.00 0.00 spur1US 8448.3 1900.8 0.06 0.56 spur1US 8452.1 1897.9 0.05 0.42 spur1tip 8453.4 1897.1 0.43 0.77 spur1ds 8452.3 1894.9 0.20 0.87 spur2US 8435.4 1848.2 0.36 0.86 spur2DS 8433.0 1846.6 0.27 0.47 spur3US 8419.3 1801.2 0.09 0.46 spur3tip 8423.1 1800.4 0.60 0.76 spur3DS 8417.9 1789.2 0.73 0.74 spur3rocksRDB 8426.5 1801.1 0.38 0.62 spur3rocksLDB 8427.9 1798.6 0.51 0.59 DSspur4 8419.5 1732.8 0.66 0.62 DSspur4 8426.4 1714.7 1.01 0.54 Spur5tip 8432.9 1701.9 1.01 0.31 USLDB 8475.6 1888.6 0.00 0.00 US 8475.4 1888.7 0.06 0.16 US 8474.3 1889.7 0.00 0.00 94 Continued 3 Oct 2009 Q = 3.27 m3/s Description Northing Easting Velocity Depth (m) (m) (m/s) (m) US 8473.1 1890.6 0.20 0.33 US 8472.0 1891.6 0.29 0.53 US 8471.2 1892.2 0.26 0.53 US 8470.5 1892.9 0.32 0.43 US 8470.1 1893.2 0.25 0.45 US 8468.9 1894.2 0.20 0.42 US 8467.8 1895.1 0.19 0.46 US 8466.6 1896.1 0.17 0.70 US 8465.5 1897.0 0.30 0.65 US 8464.3 1898.0 0.44 0.58 US 8463.2 1899.0 0.37 0.45 US 8462.0 1899.9 0.14 0.55 US 8460.9 1900.9 0.13 0.40 US 8459.7 1901.8 0.19 0.44 US 8458.6 1902.8 0.10 0.45 US 8457.4 1903.8 0.23 0.56 US 8456.3 1904.7 0.20 0.57 US 8455.1 1905.7 0.14 0.44 US 8454.0 1906.6 0.15 0.59 US 8452.8 1907.6 0.34 0.61 US 8451.7 1908.6 0.23 0.70 US 8450.5 1909.5 0.02 0.22 USRBD 8449.4 1910.5 0.00 0.00 DSLDB 8493.9 1626.4 0.00 0.00 DS 8491.8 1626.0 0.00 0.24 DS 8489.9 1625.6 0.02 0.28 DS 8487.9 1625.3 0.12 0.35 DS 8485.9 1624.9 0.03 0.33 DS 8484.0 1624.5 0.01 0.17 DS 8482.0 1624.1 0.05 0.20 DS 8480.1 1623.7 0.13 0.21 DS 8478.1 1623.4 0.52 0.28 DS 8476.1 1623.0 0.77 0.42 DS 8474.2 1622.6 0.59 0.46 DS 8472.2 1622.2 0.38 0.42 DS 8470.3 1621.8 0.71 0.38 DS 8468.3 1621.5 0.61 0.44 DS 8466.3 1621.1 0.58 0.29 DS 8464.4 1620.7 0.34 0.70 DS 8462.4 1620.3 0.31 0.34 DS 8460.4 1619.9 0.00 0.31 DSRDB 8458.0 1619.5 0.00 0.00 95 Table D-2: October 3 2009 Field Data \u00E2\u0080\u0093 Water Surface Elevation 3 Oct 2009 Q = 3.27 m3/s Description Northing Easting Water Surface Elevation (m) (m) (m) water level 1767.0 8441.4 3.93 water level 1775.9 8442.0 4.01 water level 1819.0 8444.4 4.10 water level 1809.7 8444.4 4.12 water level 1759.3 8445.0 3.89 water level 1789.5 8445.2 4.08 water level 1748.9 8445.8 3.77 water level 1828.8 8446.0 4.12 water level 1799.4 8446.3 4.11 water level 1739.7 8446.4 3.62 water level 1836.1 8447.0 4.14 water level 1713.2 8447.8 2.98 water level 1777.3 8447.8 4.16 x-sec 2, West bank 1819.2 8448.2 3.75 u/s x-section, West bank 1909.9 8448.8 4.20 water level 1720.7 8449.0 3.29 water level 1729.8 8449.2 3.50 water level 1845.6 8449.8 4.13 water level 1855.9 8450.6 4.14 water level 1707.0 8451.7 2.93 water level 1863.3 8454.1 4.13 x-sec 3, West bank 1794.9 8454.9 3.40 x-sec 2, East bank 1819.4 8455.7 3.74 water level 1873.5 8458.3 4.16 x-sec 3, East bank 1795.7 8461.4 3.40 d/s x-section, West bank 1619.0 8461.5 2.43 water level 1789.1 8462.4 3.39 water level 1694.8 8462.4 2.91 water level 1708.4 8462.4 3.08 water level 1723.9 8462.5 3.12 d/s x-section, West bank 1619.1 8462.9 2.43 water level 1783.0 8463.8 3.31 water level 1734.2 8464.1 3.14 water level 1680.0 8464.9 2.76 water level 1881.2 8466.7 4.18 water level 1765.6 8466.7 3.30 x-sec 1, West bank 1878.7 8466.9 4.14 water level 1751.5 8468.6 3.30 x-sec 1, East bank 1877.3 8470.3 4.13 water level 1671.1 8470.8 2.71 u/s x-section, East bank 1889.2 8474.5 4.19 water level 1657.6 8482.4 2.41 d/s x-section, East bank 1625.6 8492.6 2.41 96 Table D-3: October 19 2009 Field Data \u00E2\u0080\u0093 Water Surface Elevation 19 Oct 2009 Q = 25 m3/s Northing Easting Water Surface Elevation (m) (m) (m/s) 8432.09 1684.03 0.00 8430.19 1689.87 3.29 8429.08 1694.24 3.28 8432.63 1698.94 3.29 8426.04 1704.15 3.54 8423.83 1710.38 3.52 8421.00 1721.88 3.55 8418.39 1730.52 3.65 8416.92 1735.39 3.72 8415.70 1740.89 3.86 8414.31 1743.74 4.24 8413.03 1753.81 4.29 8412.45 1767.59 4.30 8411.92 1777.33 4.33 8412.56 1790.24 4.36 8413.01 1795.63 4.37 8413.48 1799.01 4.50 8415.16 1815.22 4.49 8420.46 1829.22 4.50 8426.15 1840.87 4.49 8427.81 1848.62 4.57 8434.21 1867.02 4.57 8440.28 1884.94 4.59 8443.33 1894.28 4.58 8446.73 1896.02 4.68 8446.75 1906.21 4.71 8464.16 1847.40 4.53 8460.99 1833.11 4.46 8458.89 1824.01 4.43 8461.59 1813.20 4.39 8462.01 1803.48 4.07 8464.76 1786.97 4.08 8467.57 1768.92 4.04 8468.23 1747.55 3.91 8465.14 1731.62 3.65 8465.50 1726.84 3.57 8473.73 1718.78 2.94 8481.28 1673.12 2.94 8486.04 1650.43 2.89 97 Appendix E: Grain Size Distributions Table E-1: Wolmen Pebble Counts for the Seymour River 22 May 09 09 June 09 17 June 09 16 July 09 # Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) 1 180 45 100 1000 16 64 2 300 2 90 180 200 2000 22.6 16 8 128 3 128 22.6 200 2000 128 16 8 90 4 169 64 90 900 180 90 8 32 5 45 45 550 5500 64 90 8 90 6 128 300 180 1800 64 64 16 32 7 128 16 190 1900 90 64 22.6 8 8 180 90 40 400 90 45 22.6 128 9 300 22.6 50 500 90 32 16 45 10 300 16 3 30 180 45 22.6 45 11 180 90 5 50 200 45 64 22.6 12 16 22.6 6 60 8 45 64 180 13 180 45 20 200 64 45 64 128 14 180 32 2 20 64 64 45 64 15 180 64 500 5000 45 64 45 45 16 200 45 50 500 4 128 45 45 17 64 128 170 1700 45 90 45 250 18 180 45 5 50 8 45 32 45 19 128 11 240 2400 90 22 32 180 20 128 5.6 30 300 64 128 32 64 21 45 45 260 2600 5.6 90 300 128 22 128 64 290 2900 300 64 180 90 23 180 11 2 20 90 45 45 45 24 22.6 2 2 20 45 64 32 128 25 180 128 80 800 180 32 45 64 26 128 90 80 800 32 64 64 32 27 90 180 30 300 45 45 64 250 28 45 64 80 800 90 45 32 32 29 128 64 25 250 32 22.6 45 90 30 180 16 20 200 45 45 45 128 31 45 90 40 400 32 90 45 64 32 22.6 32 500 5000 250 11 45 45 33 22.6 300 2 20 64 22.6 32 32 34 32 64 4 40 22.6 90 64 128 35 5.6 16 10 100 64 90 64 80 36 32 90 80 800 32 90 45 45 37 200 32 150 1500 32 90 45 90 38 128 128 170 1700 32 128 64 300 39 8 45 10 100 45 64 64 11 40 45 4 3 30 128 32 45 128 22 May 09 09 June 09 17 June 09 16 July 09 98 # Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) 41 16 64 260 2600 5.6 32 64 180 42 90 8 30 300 90 32 64 45 43 90 128 140 1400 16 22.6 45 45 44 16 4 120 1200 128 64 16 180 45 22.6 90 200 2000 16 64 32 32 46 250 11 10 100 90 90 16 64 47 128 2.8 270 2700 45 90 22.6 500 48 128 2 60 600 64 16 22.6 250 49 45 90 140 1400 22.6 32 45 90 50 16 2 100 1000 180 32 11 128 51 128 180 2 20 45 90 5.6 64 52 180 16 3 30 22.6 32 32 200 53 250 300 260 2600 45 32 32 64 54 90 45 20 200 64 22.6 32 32 55 180 45 300 3000 180 64 11 300 56 45 32 100 1000 32 64 8 180 57 90 90 50 500 16 22.6 8 180 58 16 64 50 500 11 22.6 8 128 59 250 180 30 300 90 22.6 5.6 90 60 128 350 40 400 16 64 5.6 45 61 180 22.6 40 400 32 128 5.6 90 62 128 180 4 40 45 180 4 16 63 90 32 10 100 64 45 2.8 350 64 90 128 60 600 128 32 4 64 65 180 16 20 200 22.6 22.6 16 90 66 64 64 40 400 128 22.6 16 16 67 64 11 80 800 128 128 22.6 45 68 180 32 30 300 90 128 22.6 45 69 90 45 260 2600 5.6 64 22.6 90 70 90 4 20 200 90 128 22.6 45 71 90 16 13 130 32 22.6 32 350 72 180 64 30 300 16 22.6 32 180 73 128 32 8 80 180 32 32 64 74 180 2.8 40 400 64 90 32 11 75 64 22.6 10 100 32 90 16 45 76 64 45 30 300 90 64 8 16 77 16 32 15 150 2.8 64 5.6 22.6 78 22.6 2 2 20 32 32 4 200 79 250 128 70 700 128 32 11 90 80 45 128 150 1500 45 32 11 64 81 45 180 20 200 90 22.6 16 64 82 180 45 6 60 90 22.6 16 45 83 128 5.6 50 500 8 16 16 200 84 180 11 30 300 90 16 16 250 85 11 8 15 150 64 90 22.6 200 86 22.6 2 50 500 45 90 16 200 22 May 09 09 June 09 17 June 09 16 July 09 99 # Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) Size (mm) 87 128 64 20 200 45 128 22.6 32 88 45 45 80 800 16 128 22.6 90 89 128 250 90 900 8 200 32 180 90 180 11 30 300 64 300 32 32 91 45 22.6 8 80 16 500 32 45 92 180 5.6 350 3500 32 22.6 32 180 93 32 11 150 1500 200 22.6 32 45 94 180 300 10 100 45 475 45 64 95 64 16 10 100 32 45 45 32 96 32 22.6 2 20 32 32 45 64 97 128 2 10 100 64 16 45 128 98 128 11 4 40 64 22.6 45 45 99 64 90 25 250 16 45 64 32 100 180 4 2 20 128 200 64 8 Table E-2: Bedload Bulk Sample In\u00C2\u00A0Field\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 Count\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 size\u00C2\u00A0 (mm)\u00C2\u00A0 count\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 22.6\u00C2\u00A0 1\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 32\u00C2\u00A0 105\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 45\u00C2\u00A0 75\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 64\u00C2\u00A0 34\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 90\u00C2\u00A0 10\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 128\u00C2\u00A0 7\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 In\u00C2\u00A0Lab\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 Sieve\u00C2\u00A0 1\u00C2\u00A0 2 3 4 Total \u00C2\u00A0 \u00C2\u00A0 size\u00C2\u00A0 (mm)\u00C2\u00A0 weight\u00C2\u00A0 (g)\u00C2\u00A0 weight\u00C2\u00A0 (g) weight\u00C2\u00A0 (g) weight\u00C2\u00A0 (g) weight\u00C2\u00A0 (g) weight\u00C2\u00A0 (kg)\u00C2\u00A0 \u00C2\u00A0 <1\u00C2\u00A0 140.00\u00C2\u00A0 165.17 246.03 256.24 807.44 0.81\u00C2\u00A0 \u00C2\u00A0 1\u00C2\u00A0 212.54\u00C2\u00A0 160.54 301.34 249.48 923.90 0.92\u00C2\u00A0 \u00C2\u00A0 1.41\u00C2\u00A0 268.66\u00C2\u00A0 227.29 319.01 265.66 1080.62 1.08\u00C2\u00A0 \u00C2\u00A0 2\u00C2\u00A0 212.90\u00C2\u00A0 185.24 228.45 180.00 806.59 0.81\u00C2\u00A0 \u00C2\u00A0 2.88\u00C2\u00A0 446.40\u00C2\u00A0 406.96 428.68 350.67 1632.71 1.63\u00C2\u00A0 \u00C2\u00A0 3.36\u00C2\u00A0 458.58\u00C2\u00A0 481.38 442.49 348.86 1731.31 1.73\u00C2\u00A0 \u00C2\u00A0 4.75\u00C2\u00A0 33.39\u00C2\u00A0 43.19 31.74 25.18 133.50 3.13\u00C2\u00A0 \u00C2\u00A0 6.35\u00C2\u00A0 4400.00\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 4400.00 4.40\u00C2\u00A0 \u00C2\u00A0 9.52\u00C2\u00A0 12000.00\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 12000.00 12.00\u00C2\u00A0 \u00C2\u00A0 18.85\u00C2\u00A0 6000.00\u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 \u00C2\u00A0 6000.00 6.00\u00C2\u00A0 "@en . "Thesis/Dissertation"@en . "2010-11"@en . "10.14288/1.0062463"@en . "eng"@en . "Civil Engineering"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "Attribution-NonCommercial-NoDerivatives 4.0 International"@en . "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en . "Graduate"@en . "The evaluation of a two-dimensional sediment transport and bed morphology model based on the Seymour River"@en . "Text"@en . "http://hdl.handle.net/2429/24252"@en .