Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sediment budget from morphology : Vedder River, British Columbia Martin, Yvonne Elizabeth 1991

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

Item Metadata

Download

Media
831-UBC_1991_A6_7 M38_6.pdf [ 8.04MB ]
Metadata
JSON: 831-1.0098616.json
JSON-LD: 831-1.0098616-ld.json
RDF/XML (Pretty): 831-1.0098616-rdf.xml
RDF/JSON: 831-1.0098616-rdf.json
Turtle: 831-1.0098616-turtle.txt
N-Triples: 831-1.0098616-rdf-ntriples.txt
Original Record: 831-1.0098616-source.json
Full Text
831-1.0098616-fulltext.txt
Citation
831-1.0098616.ris

Full Text

SEDIMENT BUDGET FROM MORPHOLOGY: VEDDER RIVER, BRITISH COLUMBIA by YVONNE ELIZABETH MARTIN B.A. (Hons.) The University of Western Ontario, 1989 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Geography) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1991 © Yvonne Elizabeth Martin In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^ tZ-rOq^aph The University of British Columbia Vancouver, Canada Date flGF.-ZI , /? ?/ DE-6 (2/88) ABSTRACT This study investigates the morphologic approach to sediment transport analysis and applies it to the Vedder River, British Columbia. The approach is based on the assumption that changes in channel morphology indicate sediment transport in the river. Despite the connection between these phenomena, only a few studies have examined this relation. The procedures, assumptions and limitations of the morphologic approach are discussed. It is more straightforward to construct a sediment budget for bed material than for wash material, as bed material travels relatively small distances. The Vedder River is a cobble gravel river with dyked banks. Therefore it is a good location for a study of the morphologic approach as bed material can be distinguished easily from wash material. The results of this study are important as aggradation in Vedder River has resulted in major flooding problems. Cross-section survey data were used to estimate volume changes along the Vedder River for incorporation into sediment budgets for several periods over the last decade. The construction of a sediment budget requires knowledge of at least one transport rate or transfer distance. Although the zero downstream transport assumption used in this study was found to be incorrect, it was retained as there are no transport rate measurements. Furthermore, the transport assumption is probably within the error ranges of the sediment budgets. Most of the errors in the sediment budgets were attributed to uncertainty in volume change estimates. When volume change estimates were calculated for different sets of cross-sections, the values varied i i i significantly. This indicates that there is bias in the results. It is difficult to evaluate the degree of bias without a knowledge of actual channel changes. It. was found that the uncertainty in the transport estimates at Vedder Crossing ranged from ± 8 % to ± 2 5 % . These values compare favourably wi th error analysis results of direct measurements in the Fraser River (see McLean and Church, 1989). A n analysis was performed to evaluate the cross-section density that is necessary to obtain a reasonable representation of actual channel changes. The average distance between cross-sections should be relatively smaller in reaches which have large variabil i ty in channel change patterns. It was suggested that cross-section spacing in the Vedder River should average between 250 and 300 m. The sediment budget results provide valuable information about the patterns of channel change and the magnitude of flows responsible for large amounts of deposition in the Vedder River. It was found that significant aggradation occurs during exceptional flood events. Most of the material is deposited in the several kilometers immedatiately upstream of the Vedder Canal . The morphologic approach provides a good method for evaluating the sediment transport regime of a river. The usual management time-scale ranges from several years to decades, which is coincident wi th the time-scale of this approach. Furthermore, the total field effort is less than that required for direct measurements. i V CONTENTS ABSTRACT ii LIST OF TABLES vii LIST OF FIGURES viii ACKNOWLEDGEMENTS x 1. INTRODUCTION 1 2. METHODOLOGY 4 2.1 Overview 4 2.2 The need for further development of the morphologic approach 9 2.3 Early developments of the morphologic approach 11 2.4 Generalized applications of the morphologic approach: the sediment budget approach and the generalized Neill approach 16 2.5 The sediment budget approach based on cross-section data 19 3. CHILLIWACK RIVER BASIN 22 3.1 Basin characteristics 22 3.2 Recent history of the Vedder River 24 3.3 Hydrology 30 3.3.1 Climate control over flood generation 30 3.3.2 Discharge measurements on the Chilliwack River 31 3.3.3 Hydrological phases 33 3.3.4 Flood frequency analysis 36 3.4 Sediment transport during high magnitude flood events 38 V 4. CHANGES IN CHANNEL MORPHOLOGY 43 4.1 Survey data 43 4.2 Long profile analysis 43 4.3 Specific gauge analysis 61 4.4 Total volume changes in the Vedder River 63 4.4.1 Procedures 63 4.4.2 Observed volume change 66 4.4.3 Gross volume change 69 4.5 Estimates of volume change along the river 71 4.5.1 Procedures 71 4.5.2 Observed volume changes 72 4.5.3 Gross volume changes 82 4.5.4 Stream power index 93 5. GRAVEL BUDGET OF THE VEDDER RIVER 96 5.1 Sediment analysis 96 5.1.1 Overview 9 6 5.1.2 Sampling procedures 96 5.1.3 Proportions of bed material in various size classes 98 5.2 Construction of the gravel budget 105 5.3 Interpretation of the gravel budgets 112 5.4 Error analysis of the gravel budget calculations 120 5.5 Sediment budgets for size classes 126 5.6 Longer-term channel changes and gravel transport 132 6. THE EFFECTS OF A REDUCTION IN SURVEY DENSITY ON SEDIMENT BUDGET RESULTS 136 7. CONCLUSIONS 151 REFERENCES APPENDIX 1: HYDROGRAPHS: 1958-90 APPENDIX 2: MEAN DAILY FLOWS > 250 v i i List of Tables 3.1 Vedder River gravel removal - Aug./Sept. 1976 26 3.2 Vedder River gravel removal - 1980 to 1990 28 3.3 Discharge estimates of exceptional floods 32 4.1 Cross-section survey dates 45 4.2 Condensed Reaches 73 4.3 Volumes of material removed from condensed reaches 81 4.4 Maximum floods between survey dates 91 5.1 Regression analyses for grain sizes 101 5.2 Percentages of bed material in various size classes 104 5.3 Gross volume changes (mineral volumes > 2mm) 107 5.4 Group selection for error analysis 121 5.5 Root mean square average deviations 123 5.6 Error range of best-estimate transport rates 125 5.7 Pecentage of error attributed to volume change estimates 127 5.8 Percentages of bed material in grouped size classes 129 6.1 Deviations between best-estimate volume changes and volume change estimates based on reduced survey coverage 142 6.2 Reaches with deviations > ± 5 000 m 3 146 6.3 Mean distance between cross-sections and standard deviations of cross-sectional changes in area 147 v i i i List of Figures 2.1 Components of the sediment budget 6 2.2 Simplified meander sweep process in natural rivers 13 3.1 Chilliwack River Basin 23 3.2 1976 dredging locations 27 3.3 1990 dredging locations 29 3.4 Histogram and cumulative percent departure plots 34 3.5 Flood frequency analysis 37 3.6 Bed load grain size measured near Yarrow 41 4.1 Cross-sections and condensed reaches along the Vedder River 44 4.2 Longitudinal profiles 47 4.3 Specific gauge analysis 62 4.4 Total volume changes (observed and gross) 67 4.5 The effect of condensation on the representation of volume changes for the period 1981-82 74 4.6 Reach volume changes (observed and gross) 75 4.7 Volume changes for the 1989 and 1990 floods 83 4.8 Volume changes within the period 1987-90 86 4.9 Cumulative percent departure over time for each reach 88 4.10 Cumulative percent departure for maximum floods 91 4.11 Cumulative percent departure along the river for each period 92 4.12 Power index 95 5.1 Regression lines for grain sizes 99 5.2 Combination plot of regression lines for grain sizes 103 5.3 Percentages of material greater than x mm 109 ix 5.4 Gravel budgets for the Vedder River 113 5.5 Size class sediment budgets: 1982/83 130 5.6 Longer-term gravel budgets 134 6.1 Volume change comparisons: Complete survey coverage vs. reduced survey coverage 137 ACKNOWLEDGEMENTS X I would like to thank Dr. Michael Church for his guidance, support and concern over the last two years - I truly appreciate the effort that has gone into seeing me through this project. Thanks are also due to my second reader, Dr. Olav Slaymaker. I owe many thanks to my research assistant, Shannon Sterling, for her dedication and enthusiasm towards the project. I would like to thank Peter Woods and Bill Wyngaards at the Special Projects Section, Water Management Branch, MOE for providing the original data set for this study. B.R. Schubert and Tom Dignan of the Technical Support Section provided the complete, corrected data set. I would also like to acknowledge the friendly assistance of Karl Bornemann and Barry Paterson at the Dyking Authority. I also extend my thanks to NSERC for providing me with the scholarship that made all of this possible. Special thank yous go to the graduate students in the department who were a constant source of support and encouragement. Finally, I want to express deepest gratitude to Troy Millington for staying by my side during the many late-night work sessions - thank you very much. 1 1. Introduction There necessarily exists a relation between changes in channel morphology and sediment transport rates. Despite the inherent connection between these phenomena, there have been only a few studies that investigated the possibility of using the relation to evaluate sediment transport. This suggests that the potential of the approach has not been fully recognized. This is surprising when the problems associated with methods of sediment transport prediction are taken into consideration. In the past great emphasis has been placed on formulating equations to predict transport rates. There is an attraction in reducing physical phenomena to generalized formulae. Sediment transport formulae are based on various combinations of flow and sediment characteristics. Gomez and Church (1989) assessed the performance of several sediment transport formulae critically and found that none of them could predict sediment transport adequately. In recent years, research has focussed on evaluating the importance of bed structure in determining sediment transport. The transport process depends on a wide range of flow, grain and bed structure characteristics, some of which are difficult to quantify. As understanding of the complex nature of rivers continues to develop, it is becoming increasingly clear that sediment transport formulae based on idealized hydraulic principles consistently fail to predict transport rates accurately. Furthermore, it seems inevitable that any formulae which attempt to account for the complexities of the system quickly become unmanageable themselves. Given the inaccuracy of bed load transport equations and the time and effort necessary for direct measurement, alternative methods for evaluating sediment transport should be investigated. Changes in channel configuration 2 necessarily indicate that transport has occurred. The application of the "morphologic approach" is less complicated when bed material transport can be distinguished from wash material. Bed material moves considerably smaller distances than wash material, which makes it easier to identify transfer distances and appropriate reach lengths. The methodology of the approach is relatively straightforward and lends itself more readily to an evaluation of bed material transport, provided that the relevant sediment volumes associated with channel change can be estimated. It is critical to determine the sediment transport processes for the particular river being studied so that the different components of the load can be identified. Of the relatively few studies that have investigated the morphologic approach, even fewer have provided a thorough description of the methodology. Constraints of the approach will be examined in the this study. The present study uses the morphologic approach to evaluate sediment transport and deposition rates in the Vedder River, British Columbia. The river flows across an alluvial fan and has been aggrading significantly. Associated with aggradation is increased flood hazard risk. Various flood control operations have been implemented over the last several decades; of particular significance are several large scale gravel removal operations. Knowledge of the sediment transport regime and both observed and natural rates of deposition are useful when planning flood control procedures. The Vedder River is a cobble gravel river. The banks are defined and stabilized by dykes, which eliminates off-channel deposition. The river is a good location for a study of the morphologic approach as most of the the channel changes reflect transport of bed material, which moves only relatively small distances. The gravel bed load offers an opportunity for a 3 straightforward application of the approach. Because of flooding problems, cross-section surveys of the Vedder River have been completed in a number of years over the last several decades. The cross-section data can be used to estimate morphologic change, which in turn can provide sediment transport estimates. The methodology employed in this study is directly transferable to other projects for which sufficient cross-section data are available. Because cross-section surveys require minimal time and effort, in comparison with a program of direct sediment transport measurements, the approach provides a straightforward method for evaluating the sediment transport regime, even if data do not already exist for a river. The two primary objectives of this study are: (i) to evaluate the potential of the morphologic approach of bed material transport analysis and to discuss assumptions, procedures and limitations, and (ii) to assess the patterns of sediment transport and deposition in the Vedder River. k 2. Methodology 2.1. Overview The fluvial system can be characterized as a process-response system. A process-response system is one in which the particular morphology is essentially dependent on the throughput of energy or mass. Information about the throughput can be inferred from changes in morphology over time. In the river system this represents a link between sediment transport and morphology. Before discussing the morphologic method of sediment transport analysis, it is important to make clear the distinction between the sediment quantities that are actually measured in the field, vis suspended load and bed load, and those that are important in terms of channel-forming processes. When considering the formation and adjustment of channel morphology the important distinction that must be made is between bed material and wash material. Bed material is the sediment that comprises the bed and lower banks of a river. This material, once entrained, travels only relatively small distances. Its movement can be by traction, saltation, or in suspension. Wash material is derived from catchement slopes, upper river banks and vertical accretion deposits. This material is fine and moves through the system in suspension. The size division between these two types of material occurs at about 180 M m . Bed material has an important role in determining channel geometry. Wash material has little effect on channel morphology, except when upper bank erosion and deposition occurs. From the foregoing discussion it is apparent that when one directly measures suspended load one may also be measuring a component of the bed material transport. Bed material can move close to the bed or in suspension. 5 Transport rates obtained by direct bed load measurement include only the fraction of bed material that moves very close to the bed. The morphologic method measures changes in the channel configuration; such changes reflect the movement of material that determines channel morphology i.e., the bed material and wash material which may be involved. The bed material and wash material can be distinguished on sedimentalogical criteria. Estimates based on the morphologic approach represent the transport rate of bed material, which always has an important role in determining channel configuration. The morphologic method is based on the continuity principle applied to sediment in the channel zone. The natural framework for such a study is the sediment budget. The construction of a sediment budget involves the quantification of sediment inputs, outputs and storage changes in a defined reach and is based most simply on the equation: Si - AS = S D (2.1) where Sj is sediment input, AS is change in storage and S 0 is sediment output (see figure 2.1). If two of these terms are known then the third can be calculated to within the margin of error of the known terms. If each of these variables is measured over a finite time then the equation becomes: AS Qi - (1-p) = Q 0 At (2.2) 6 Volume of sediment deposited (Df) Q w Wash material Q 0 Sediment load out of reach Q; Sediment load into reach L r Reach length Figure 2.1. Components of the sediment budget 7 where Qj is volumetric transport into a reach per unit time, p is porosity of the sediment pile, A S / A t is change in storage within a reach per unit time, Q0 is the volumetric transport out of a reach per unit time, and the time unit is the time between surveys. The A S term can be obtained from morphologic studies. If the magnitude of sediment transport is known at one section then the calculations can be extended along the channel in either direction as the Qj value of one reach is the Q0 value of the next upstream reach (Church et al., 1987). Alternatively, a distance of travel assumption (or measurement) can replace the need to know the transport rate at a reference section. The velocity of travel is the average distance that material moves during the time interval of the morphologic study. The transport rate is calculated from the equation: s e 1 Qs = (1-p) * * L t (2.3) L r At where Se is the volume of erosion in a reach, is the transfer distance and Lr is the reach length. If the transfer distance is the same as the reach length the equation becomes: s e Qs = (1-p) * (2.4) At Transport rates can be estimated at various locations using this approach. Alternatively if one transport estimate is obtained using this approach and 8 volume changes can be estimated along the river, then the sediment budget can be extended in either direction. However if a tributary with a significant amount of load enters a reach, or if any material is removed from the channel, then the volumes must be assessed and incorporated into the sediment budget. Wash load is material that travels through a reach without being redeposited after its initial entrainment. The suspended part of the bed material, as well as material derived from upper river banks and catchment slopes, may move as wash load through a reach. The distance of travel is radically different than bed material and is not easily estimated. Therefore the approach based on equation (2.3) is best restricted to estimates of bed material transport. Wash load can be estimated for a reach from equation (2.1) using the relation (Church et al., unpubl.): S w = S Q - S e = S i - S d > 0 ; e l s e S w = 0 (2.5) where is the volume of deposition in a reach; all other terms were defined previously (all variables are measured over a finite time). However, an important assumption is made here that there is negligible re-deposition of material in the same reach from which it was eroded. Otherwise the values of S e or could be inflated, resulting in a negative value of S w . If estimates of wash load are to be made then the reach lengths should be chosen to ensure that there is no within reach transfer of material, i.e., L^->Lr. But if reaches are too short (L^->Lr) then some bed material load will appear to be wash load. If these problems are to be avoided then the reach length should approximate the average step length of bed material during the time 9 resolution of the morphologic study (L^ . = L r). Step length can be estimated by analysing: (i) the distance between erosion zones and point bars (for meandering rivers), (ii) the distance between distinct deposition zones, or (iii) direct measurements based on stone tracing methods. If there is no wash load then the same constraints apply to bed material transfer in the approach based on equation (2.1), but are avoided in the approach based on equation (2.3). It is also important to recognize that any record of sediment stored and then re-entrained within a period shorter than the time resolution of the morphologic study is lost. The transport rates for S^ (bed material load) and S w represent lower bound estimates of sediment transfers. If the loss of information due to the time step of the morphologic study is approximately equally distributed along the entire length of the study reach, then important knowledge can still be gained about relative rates of sediment transport. If a study reach is in equilibrium (net volumetric change is about 0) for the time scale of the surveys, then the sediment input is equal to the sediment output. Sites of erosion and deposition must then be distinct within a reach or information is lost, and one again has only a lower bound estimate, the severity of which will depend on how much compensating erosion and deposition (erosion and deposition in the same spot) occurs. This limits the usefulness of this approach in sand bed rivers as they experience compensating scour and fill. 2.2. The Need for Further Development of the Morphologic Approach There are many problems associated with direct measurements and sediment transport formulae. Direction measurements require a considerable 10 field effort. The large spatial and temporal variabilities of bed load transport make it difficult to obtain reliable measurements (Hamamori, 1962; de Vries, 1973; Hubbell, 1987). McLean and Tassone (1987) found that individual measurements could be up to six times greater than the overall mean transport rate. Final bed load transport values obtained by direct methods can vary considerably depending upon the chosen times and locations of individual measurements. Despite a lack of reliable field data, many attempts have been made to evaluate the performance of bed load transport formulae (Johnson, 1939; Vanoni et al., 1961; White et al., 1973; Gomez and Church, 1989). What is probably the most critical evaluation to date was performed by Gomez and Church (1989). The analysis was based on laboratory and field data which were not used in the development of the formulae. However the data did represent the conditions for which the formulae were developed. None of the formulae could predict bed load transport accurately. The poor performances of the formulae were attributed to limitations of the test data and the inability of the formulae to adequately portray the physics of the bed load transport process. Despite recognition of the connection between morphological change and sediment transport, the lack of development of the morphologic approach suggests that its potential has not been fully recognized. Prior studies introduced the topic and used various derivations of it to estimate sediment transport for specific situations (see sections 2.3 and 2.4). However these studies did not provide any general guidelines for the application of the approach. McLean (1990) has made the only thorough attempt to generalize the methodology. 11 Further evaluation of the morphologic approach, which so far has shown considerable potential (Church et al., 1987; Neill, 1987; McLean, 1990), is justified. This study applies the morphologic approach to the Vedder River in British Columbia. Although the transferability of the exact methodology is subject to the availability of appropriate data, this study is placed within the context of a more general sediment budget framework. Therefore many of the issues discussed in this thesis are applicable to other studies which use the morphologic approach. The rest of this chapter is devoted to reviewing prior studies and outlining the methods adopted in this study. 2.3. Early Developments of the Morphologic Approach Popov (1962) made the first attempt to formally link the process of sediment transport with morphologic change. He recognized the problems associated with attempting to reduce sediment transport analyses to idealized hydraulic schemes. Popov believed that such a complex and variable process is best studied by analysing channel deformations, which are an expression of the transport regime. The sediment budget was introduced as a method to quantify these phenomena. He applied this approach to the River Ob in the Soviet Union to investigate various aspects of the sedimentation process. In particular, he divided channel erosion into two components; sediment derived from bank material and from the channel bottom. This corresponds approximately with the division between bed material and wash material. However his was an introductory study and there was no formalization of the methodology. 12 In a series of papers Neill (1971, 1983, 1987) discussed and developed the approach for a particular river morphology. Neill's earliest practical developments primarily dealt with the case of consistently developing meander bends. There is a relation between sediment transport and any systematic, macro-scale channel process involving morphologic change, such as meander shifting (Neill, 1987). In the case of regularly progressing meanders, the net erosion in a reach, chosen to be one-half a meander, can be specified in equation (2.3) after substituting: S e = Ae * h * L r (2.6) where Ae is the width of bank erosion, h is the bank height and L r is the reach length measured along the river (see figure 2.2). If there is a complete exchange of material, such that sediment eroded from one bend is deposited on the next downstream point bar, then the reach length is the same as the transfer length of eroded material; these two values become interchangeable and the problems of biased measurement discussed earlier are avoided. The transfer length is the correct weighted average of transport distances for various fractions of the load. But if the true average transport distance is greater than the chosen value then the estimate is negatively biased (Church et al., unpubl). The final form of Neill's equation for regularly progressing meanders is (Neill, 1987): Ae X Q s = (1-p) * h * — At 2 (2.7) 13 CROSS — SECTION Y - Y Figure 2.2. Simplified meander sweep process in natural rivers (from Neill, 1971) 14 where Q s is the transport rate averaged over the time resolution of the morphologic measurements, A t; and is the meander wavelength. When there is a complete exchange of material over timeA t, the change in storage is effectively the transport rate averaged for the period over which measurements are made. The adoption of X/2 precludes consideration of any banktop wash load deposits, which then have to be eliminated by stratigraphic assessment. N e i l l (1971) applied this approach to the Red Deer River. The sediment transport estimate was approximately 21 000 m 3 / y r . Unfortunately no measured transport rate was available for comparison. The transport estimate was compared wi th bed load estimates obtained by the Einstein and Blench regime methods. Because of the similari ty in the results, he concluded that his estimate represented the bed load transport. He believed that the wash load did not have a significant role in the flood-plain formation of the Red Deer River. He recognized that a distinction could, be made between bed load and suspendeded load by examining the stratigraphic horizon of the eroding banks and flood-plain, although no such procedure was followed in this study. In a later study Ne i l l (1983, 1987) estimated the transport rate for the bed material component of sediment load in the Tanana River. Equation (2.7) was modified to isolate the bed material load. Instead of mult iplying the planimetric area of erosion by the total bank height, it was multiplied by the height of the bank from the channel bottom to the top of the bed material deposits. N e i l l compared an estimate of bed material transport based on this method wi th direct bed load measurements in the Tanana River. The measured average annual bed load transport rate balanced very well wi th the 15 estimated bed material transport rate, implying that the assumption of a complete exchange of material is probably valid. Neill's papers demonstrated the potential to predict bed material transport by examining morphologic changes. However, in order to cover the necessary assumptions in the most conservative way, he focussed on applying the approach to one particular situation and did not place it in the context of a general framework. Hickin and Nanson (1984) investigated lateral migration rates of river bends. They were interested primarily in assessing channel migration and emphasized that a key concept in understanding this phenomenon is sediment transport. However the relation between meander progression and sediment transport was not actually examined. Hickin and Nanson measured channel migration rates for a variety of meandering rivers in Western Canada. The Neill approach for estimating sediment transport was applied in some of the same reaches by Church and Slaymaker (1989). It was assumed that bank erosion represented the incremental contribution to suspended load (most of which is probably wash material). Therefore the suspended load transport rates obtained by the Neill approach could be compared with suspended transport estimates predicted by an independent regional analysis of specific sediment yield. Of the nine rivers analysed, eight of the estimates based on Hickin and Nanson's measurements fall within the range of values predicted by Church and Slaymaker. The results of these two very different methods support each other, thus suggesting that the Neill approach provides reasonable estimates. In this study the morphologic approach was used to estimate the suspended load transport (as opposed to the bed material component which was estimated in Neill's studies). This demonstrates the flexibility of the morphologic approach. 16 2.4. Generalized Applications of the Morphologic Method: The Sediment Budget Approach and the Generalized Neill Approach The approach developed by Neill is useful if a river displays a pattern of regularly progressing meanders. However many rivers do not show such a distinct pattern of erosion or deposition and it is necessary to turn to the more generalized application of the sediment budget approach based on equations •2.1 and 2.2. McLean (1990) attempted to generalize the morphologic method of analysis, that up until this point had focussed on the case of consistently developing river bends. He applied the sediment budget approach for the period 1952-1984 to a reach of the Fraser River having a predominantly wandering pattern. The change in storage was obtained using bathymetric survey data for 1952 and 1984. The volumetric change in storage for the active channel was estimated by comparing the common areas of active channel for the two surveys. A DTM (digital terrain model) was used to replace the irregularly spaced survey points with regularly spaced interpolated points. A precision class based on survey coverage was assigned to each reach to assess the reliability of the change in storage estimate. The budget calculations were restricted to bed material (gravel) by excluding the river bank sand body. Stratigraphic measurements were made to distinguish between these two types of material. An assumption was made that the gravel transport past Mission was negligible making it possible to extend calculations upstream. Bed load is approximately equal to bed material in cobble gravel rivers, which makes it possible to compare the results of McLean's study with direct bed load measurements. The estimated gravel transport rate using the morphologic approach was about 120 000 m3/yr at 17 Agassiz, which is probably within the precision of the measured bed load transport value of 100 000 m^/yr. It is not obvious which is the more accurate result. The agreement of the two values supports the ability of both procedures to provide reasonable bed material transport estimates. Temporal and spatial variations in bed load transport rates make it difficult to obtain accurate measurements. Furthermore the accuracy of the transport measurements also depends on the reliability of the sediment rating curve when used to estimate annual sediment transport over the entire study period. McLean (1990) emphasized that the sediment budget relies on a series of calculations whereby the sediment input into one reach is the sediment output of the next downstream reach, which can result in systematic error propagation. The reliability of the transport estimate is only as good as the weakest link in the budget. Uncertainty regarding the exact volumes of gravel extraction from the Fraser River also may have introduced error into the transport estimate from the sediment budget. McLean then carried the method one step further. He compared transport estimates based on the sediment budget approach with transport estimates based on morphologic change measured from aerial photography and planimetric maps. The maps were overlaid and areas of deposition and erosion for the banks and islands were identified. Volumes of erosion were obtained by multiplying the areas of erosion by the bank height. The total volume of erosion for a reach was converted to a unit transfer rate (m /^yr/km). The average step length was then estimated so that the erosion volumes could be converted to transport rates. In the absence of regular meanders, the step length was taken to be the distance between active deposition zones. This is really just a generalization of the Neill approach. 18 The average step length per year represents the annual virtual velocity of travel and was used in conjunction with the unit transfer rate to calculate the bed material transfer rate. The average bed material transport rate between 1943 and 1971 was estimated to be 80 000 m /^yr. This value seems reasonable as it lies within the precision of the bed load measurement of 100 000 m3/yr. An approach based on bathymetry makes fewer assumptions than an approach based on planimetric mapping. In particular, the depth assumption associated with the latter method may decrease the reliability of transport estimates. McLean's study recognized the potential of the morphologic approach and placed it in a general framework. Based on the results of this study, the approach warrants further exploration. A feasibility study was undertaken by Church et al. (1987) which further demonstrated the potential of the morphologic method. The practicability of studying bed material transport on the Mackenzie River near Norman Wells using aerial photography and planimetric maps for the morphologic study was assessed. Mackenzie River is a large, sand transporting channel. Estimates of sediment transport were based on a generalization of Neill's approach (see equation 2.3). Planimetric maps were obtained based on photography flown on 3 different dates with appropriate corrections for stage differences due to different flow conditions. Morphologic change was also estimated for a later time period using the results of two bathymetric soundings. Comparison of the two methods suggested that estimates based on the bathymetric soundings may have been too high. The average step length was assumed to be the mean distance between successive riffles. The sediment transport was estimated for two periods, each having a 19 length of about a decade, and for values of morphologic change based on both erosion and deposition. In sand transporting rivers, the bed material is often moved through the system in dune trains (wavelike pulses of sediment). The successive scour and fill at a location could confound results of the morphologic approach for short-term studies. However there are distinct deposition and erosion sites in the Mackenzie River at a larger spatial scale. The time scale of the morphologic studies (about a decade) was great enough so that the confounding of the results due to the dune trains was irrelevant and large-scale morphologic changes could be evaluated. When one anomalously high value, due to unusual aggradation conditions during one of the time periods, was disregarded the average value for the three remaining estimates was approximately 5.6 x 10^  m3/yr (1.02 x 10^  tonnes/yr). Measured suspended sediment transport on the Mackenzie River at Arctic Red River is approximately 88.2 x 10^ tonnes/yr (Church et al., 1987). This suggests that bed load accounts for about 1.1% of the total load. This is slightly lower than the expected value of about 1.5% for a sand bed river, although it is reasonable. The approach appears to provide a reasonable estimate of the bed load transport rate for this large, sand transporting river. 2.5. The Sediment Budget Approach Based on Cross-section Survey Data The method for the evaluation of morphologic change that is developed in this study is based on repeated cross section surveys. The assessment of morphologic change using cross section survey data is a variation of the approach based on bathymetric survey data. The datum points used in 20 McLean's study for the two bathymetric surveys did not correspond and common points had to be interpolated. The repeated cross section surveys can be thought of as repeat bathymetry on standard lines, which makes it easy to obtain finite differences. As one moves across the survey line, measurements are made at any important changes in channel slope. The exact resolution of each cross section survey line (or what constitutes an "important" change) depends on the precision required for the study. It is evident that the locations of the breaks in slope will change each year for a particular cross section. Assuming that all changes in slope at the desired resolution are included in the survey, then the interpolation of common points does not lead to any significant loss of information. Cross section surveys are overlaid to obtain the net change in storage area by performing the following computations. The changes in elevation for the interpolated points are calculated for each survey comparison. The change in area between successive pairs of points is obtained from the equation: A z i + A z ( i + 1 ) A A ( i f i + l ) - * d ( i , i + 1) ( 2 . 8 ) where A z is change in elevation, and d is the distance between the two points. The summation of these values is the net change in storage area for a cross-section: AA = E A A ( i f i + 1 ) ( 2 . 9 ) 21 Net volumetric changes are then calculated for reaches based on the assumption that the change in area at a cross-section is representative of the distance between it and the mid-point to each adjacent cross-section. Study reaches are chosen so that deposition and erosion volumes within a reach do not cancel each other out. Once the volumetric change for each reach is determined and if at least one sediment transport rate is known or if the virtual velocity (velocity averaged over the period) of the bed material is known, calculations can be extended upstream using the sediment budget approach. The Vedder River is a cobble gravel river that lies within fixed, dyked banks. There are no significant overbank wash deposits. This is an excellent site for a test of the morphologic approach as bed material transport can be isolated from wash material transport. 22 3. Chilliwack River Basin 3.1. Basin Characteristics The Chilliwack River basin is located approximately 160 km east of Vancouver and has an area of 1230 square kilometers (see figure 3.1). The river flows through the Cascade Mountains until it emerges at Vedder Crossing. The Cascade Mountains are a steep mountain range having many peaks exceeding 2130 m. Approximately 50% of the total basin area has an elevation greater than 1130 m. Granodiorite and complex sequences of limestone, sandstone and conglomerate are the predominant types of bedrock. There are two major lakes located in the watershed - Cultus Lake and Chilliwack Lake. The largest tributaries that flow into the Chilliwack River, named in Figure 3.1, are located above Vedder Crossing. The head of the Chilliwack fan is at Vedder Crossing and has an elevation of about 25 m.a.s.l. Below this point the Chilliwack River is referred to as the Vedder River. The fan has an area of about 70 square kilometers. The river flows across its alluvial fan until it enters the Fraser River at an elevation of about 3 m.a.s.l. Alluvial fans are aggradational features that form where rivers emerge from mountain ranges and enter relatively unconfined plains or valleys. A large quantity of sediment that is transported onto the fan is deposited as there is a significant decrease in the transporting capability of the flow. Although it is located near the former limit of the continuous Cordilleran ice cover, the Chilliwack valley has been influenced by the Pleistocene glaciations. During the Fraser Glaciation a major lobe of the Cordilleran ice-sheet and numerous valley glaciers were present in the valley. The Late Wisconsinian glacial limit reached its maximum extent about 14 000 B.P.; the Chilliwack Valley is located about 50 km from the limit at its closest Figure 3.1. Chilliwack River Basin 2k point (Clague et al., 1988). During the final ice retreat of the Fraser Glaciation (12 000-11 000 B.P.) large volumes of sediment were deposited along the valley. It was at this time that the contemporary Chilliwack alluvial fan began to form. Most of the extant valley sedimentation in British Columbia took place in the several thousand years immediately following glaciation (Church and Ryder, 1972). Making the assumptions that most of the post-glacial sedimentation occurred between 10 000 and 6 000 B.P., and that the fan head had an elevation of about 43 m, McLean (1980) estimated a deposition rate of about 190 000 m3/yr over this 4 000 year period. Sedimentation probably decreased significantly from about 6 000 to 3 000 B.P. The fan has once again been undergoing aggradation in recent times, although it is believed that deposition rates are lower than they were immediately following the Fraser Glaciation. 3.2. Recent History of the Vedder River McLean (1980) reviewed the recent history of aggradation, flooding and flood control measures on the Vedder River. In about 1850 the Chilliwack River below Vedder Crossing consisted of two channels, the Chilliwack River with its mouth near Chilliwack Mountain, and Luck-A-Kuck Creek. Some of the flow was diverted into Vedder Creek when a channel shift occurred in 1873. After several episodes of further channel shifting Luck-A-Kuck Creek and the old Chilliwack River below Vedder Crossing were abandoned and Vedder Creek became the major channel on the fan. At this time both the Sumas River and the Vedder River drained into Sumas Lake before entering the Fraser River. During the spring/summer runoff peak of the Fraser River 25 water backed up into Sumas Lake and the two rivers, which resulted in severe flooding problems. This led to the acceptance in 1919 of the "Sinclair plan" to drain Sumas Lake to increase agricultural land and to reduce flooding problems by channelizing the Vedder River along its course across the Sumas Prairie and into the Fraser River. This channelized section of the river is now known as the Vedder Canal. Dykes were constructed along the canal and the Vedder River as far up as the B.C. Electric Railway Bridge near Yarrow the early 1960's (see figure 3.2). Flood protection upstream of this location consisted primarily of bank protection, gravel removal and channel clearing. In 1967, dykes were extended upstream of the railway brige to Bergman Road. Major flooding along the Vedder River in 1975 resulted in a decision to increase flood protection. Banks were raised by 1 m and gravel was removed from bar locations above water line in order to increase channel capacity at bankfull flow. A major sediment removal operation, involving the excavation of about 534 000 m 3 of material from the Vedder River, was undertaken in 1976 to increase channel capacity (see table 3.1 and figure 3.2 for dates, locations and volumes removed). Setback dykes were constructed between the Vedder Canal and Webster Road in the early 1980's to provide long-term control of the Vedder River. The Vedder River is still undergoing aggradation and it has been necessary to conduct several smaller dredging operations to maintain the ability of the channel to contain high flood flows (see table 3.2 and figure 3.3 for dates, locations and volumes removed). Extraction took place each year from 1980 to 1983, with a total volume of material removed during this Table 3.1. Vedder River Gravel Removal - Aug./Sept. 1976 (all values are m )^ Stockpile Site 1. Peach Road 2. Klein 3. Hooge 4. Hopedale 5. Army 6. Yarrow #1 7. Yarrow #2 8. Greendale 9. Contract 14 Truck Measure 12 800 128 000 99 900 24 400 73 800 76 200 35 000 63 600 20 500 Total: 534 000 Source: B.C. Ministry of the Environment Water Management Division Table 3.2. Vedder River Gravel Removal - 1980 to 1990 (all values are m )^ Bar Location 1. U/S from Giesbrecht Rd. 2. End of Peach Rd. 3. D/S from Lickman Rd. 4. U/S from Bergman Rd. 5. U/S from Wilson Rd. 6. D/S from Giesbrecht Rd. 7. End of Peach Rd. 8. End of Bergman Rd. 9. U/S from Southern R.R. Bridge 10. U/S from Southern R.R. Bridge 11. U/S from Wilson Rd. 12. D/S from Wilson Rd. 13. D/S from Wilson Rd. Removal Date Truck Measure Aug.5-Sept. 10/80 27 400 Aug. 17-Sept. 3/81 18 000 July 29-Aug.25/82 32 400 Aug.25-Aug.27/82 6 640 Aug. 12-Aug.24/82 12 900 Aug.9-Aug.29/83 . 23 000 July-Aug./90 50 000 Aug.-Sept./90 43 000 Aug.28-29/90 1 930 Aug.29-30/90 4 480 Aug.31-Sept.12/90 45 200 Sept.8-15/90 16 500 Sept. 12-19/90 25 500 Total: 307 000 Source: B.C. Ministry of the Environment Water Management Division 30 period of approximately 120 000 m6. A large dredging operation, that removed 187 000 m 3 of material from the river, took place in 1990. 3.3. Hydrology 3.3.1. Climatic Control over Flood Generation The climate in the basin is characterized by warm, dry summers and cool, moist winters. Due to this seasonal weather pattern the Chilliwack River experiences two runoff peaks. High flows occur in the autumn and winter due to significant rainfall events. During this time most of the precipitation is in the form of snow at elevations greater than 1000 m. However rain falls at higher elevations during some events and the combination of rainfall and snowmelt can lead to particularly significant floods. As a large proportion of the watershed is at high elevations a substantial snowpack is able to accumulate during the autumn and winter. Hence high flows also occur in the spring and summer as temperatures increase and snowmelt is generated. McLean (1980) identified 4 periods in the monthly flow regime of the Chilliwack River: (i) spring freshet due to snowmelt (April to June) (ii) summer recession (July to September) (iii) autumn rise resulting from an increase in precipitation and mild temperatures (October to December) and (iv) winter recession due to decreased precipitation and cooler temperatures (January to March). Although significant flood events occur during the spring freshet, it is the autumn/early winter flood events that result in the highest flows and are responsible for most of the flood damage that has taken place along the river. 31 3.3.2. Discharge Measurements on the Chilliwack River Continuous water level measurements have been made since 1968 at Vedder Crossing gauging station (WCS# 08MH001). Prior to 1968 manual observations of stage were made daily, or possibly several times a day during higher flows. The discharge rating curve may be unreliable at high flows as discharge measurements are seldom made during large flood events (Jordan, 1990). Furthermore the discharge rating curve is often unstable during large flood events as the channel configuration undergoes change during the course of an event. During several particularly large floods the water level recorder has been washed out and no reliable estimates of water level are available. This has been a problem during the recent flood events of January 4, 1984; November 10, 1989; and November 10, 1990. It is necessary to obtain an estimate of peak discharge in these events as they represent 3 of the largest events since the measurement program began on the Chilliwack River. A regression analysis was performed for maximum daily flows > 150 m /^s at the Chilliwack River station at Slesse Creek (WSC# 08MH103) and the corresponding maximum daily flows on the Chilliwack River at Vedder Crossing, when the latter were available. The regression equation was then used to estimate the maximum daily discharge for the 3 missing values (see table 3.3). The estimates are probably not very accurate because the data used in the regression analysis are subject to the many limitations encountered when estimating high flow values: the r^ value for the regression equation is only 0.48 and the standard error of estimate is 54.2 m /^s. However the estimated values are probably reasonable to use when assessing the relative magnitude of these floods in relation to other floods. Table 3.3. Discharge Estimates of Exceptional Floods Regression Equation: y=1.83x +12.5 (all discharges are m3/s) Date Discharge - Chilliwack Estimated Discharge -River at Slesse Creek Chilliwack River at Vedder Crossing Jan.4, 1984 278 524 Nov.10, 1989 345 647 Nov.ll, 1990 415 776 33 The hydrographs for 1958 to 1990 are presented in appendix 1 and are referred to in later relevant sections. Attention is now focussed on analysing the general flow regime of the Chilliwack River. 3.3.3. Hydrological Phases In order to obtain a better understanding of long-term flow variations on the Chilliwack River histograms and graphs of cumulative percent departure from the mean were constructed for both the annual mean discharge series and the annual maximum daily discharge series (see figure 3.4ab). Cumulative percent departure plots illustrate whether a value for a particular year is above or below the mean value of a series. Therefore the increase and decrease between two points, not their absolute position, should be examined. Runs of persistently positive or negative slopes are of particular interest in this analysis. The records are discontinuous as no data are available for a number of years in each series. However as successive values are independent of each other, these gaps do not matter in a statistical sense. Several periods can be identified for the annual mean discharge series. There is no trend in the data for the period 1912-25, while values for the period 1926-30 are persistently low. There is an overall trend showing above average values of annual mean discharge from 1952-76. However the values are generally below average for the period 1977-90. Examination of the annual maximum daily discharge series shows very different trends than those described above. The period 1912-30 is characterized by a repeating pattern whereby a much above average value is followed by 2 or 3 years of values close to or slightly below the mean. This pattern is illustrated very clearly on the histogram. The data for the period 120 0) \ 100 34 ANNUAL MEAN DISCHARGE 80-u &3 60 Q 40-2 20 1 (ml n nTTTTTTnTfI111111111 u 111111111 f T T i r V t m t i f n t m T T T n t f T r T i f r n n m" 1910 1930 1950 1970 1990 1920 1940 1960 1980 YEAR 100 — 1 00 I I I I I I I I I I I I I I I I I M I I I I I II M I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I II I I I I I I I I I I I I I I 1910 1930 1950 1970 1990 1920 1940 1960 1980 YEAR Figure 3.4. (a) Histogram and cumulative percent departure plot: annual mean discharge series 35 ANNUAL MAXIMUM DAILY DISCHARGE 800 T <700-•^600-1920 1940 1960 1980 YEAR 250 —250 'i i I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 M 1 1 u 1 1 1 1 1 M 1 1 1 1 1 1 1 1 N 1 1 1 1 1 1 1 1 1 1 , 1 1 1 1 1 1 , 1 1 1 1 1 1 1 M i 1 1910 1930 1950 1970 1990 1920 1940 1960 1980 YEAR (b) Histogram and cumulative percent departure plot: annual maximum daily discharge series 36 1952-1974 show a fairly consistent series of below average flows. Once again a pattern similar to the one described for 1912-1930 is found in the data for the period 1975-1990. Between each significant flood event there is a period of 3 or 4 years of slightly below average flows. Comparison of the two series shows that there is no direct correlation between them. The existence of a below or above average value for the annual mean daily discharge does not imply that the same is true for the annual maximum daily discharge. Despite an overall decrease in the annual mean discharge for the most recent period, the analysis of the maximum daily discharge shows that there was still a particularly significant individual flood event about every 4 or 5 years. It is these large flood events that are responsible for moving large volumes of sediment through the river system. 3.3.4. Flood Frequency Analysis Three periods of hydrological activity were identified for the Chilliwack basin on the basis of the cumulative percent departure plots: (i) 1911-1930, (ii) 1952-1974, and (iii) 1975- 1990. It was decided to construct flood frequency curves for the two most recent periods as the present study concerns the latter half of the century. The hydrological behaviour is sufficiently unique in each period to warrant individual analyses. Within each of these periods the snowmelt and rainstorm events are analysed separately due to the differing nature of these two types of flood events (Jordan, 1990). Daily mean discharge values were used for the analyses. The results are presented in figure 3.5. The rainfall floods are greater in magnitude than the snowmelt floods for both hydrological time periods, an observation made for many coastal rivers in British Columbia (Jordan, 1990). 800 700 600 500 ui o O 400 CO - — i —\--— z --n - n - . . i n 1 ( i n n - - -— - - it: —1- _ _ _ — ~\- ££ £: z z -— - . . . z £ £ z £ £ I RAINFALL FLOODS (1975-90) O RAINFALL FLOODS — (1952-74) • SNOWMELT FLOODS --(1952-74) : : x SNOWMELT R.00DS (1975-90) : : _ — _ - | - _ - - - - - — — —1-— — f _ _ — -• n -• —1- - - - Z". - : - : - -T - l -— -—h i _ z - -Nov. 10. 1989: — : £ — — - -- z | : — z Z - £ — — • — — - :££- ££j 5£ - £ ~-_ ;£ £ - - ,_ '_- - £ I z _ £ £ • --- -* _i _ £ _ z _ £ - _ £ £ — — — — — — — — — — — — ££ £ £ £ £ ^[ ££ z — -- -_ - — • £ ££ z £ £ 3 c 1 £ Z —C z £ £ : £ z} £ £ £ — £ z : zz z £ zz : £ ££ -— ; ::zz — - -— y - £ - £ z z £ £ — £ £ £ zz £ £ • ->-< — - - — - - - z\ £ --— - £ £ - — £ - £ £ L#- — _ _ i o - — — — I 1— >(»<>- _ — i X — X - z i z 2 z £ z -£ — — — — — • .1 III —« V -* _ -*- I« -•x - X - £ -_ -_ z = z - -— o - f — — — - — - - zz • ' , < • v X j — -- - — - - - — — - - - — - - - — » < » • I —j- — - £ -— - - - £ £ £ : -— - - — - - — £ £ I'-— X. 1 - — • -- - _ — x -x- K—> — — — -— £ £ £ _ £ - z £ _ 12 z £ — — z :_zz £ ll _ = = T - - ~z £ -300 200 100. 1.001 1.01 1.1 1.5 2 5 10 20 50 RECURRENCE INTERVAL IN YEARS 100 200 500 1000 Figure 3.5. Flood frequency analysis 1^ 38 For the period 1952-74 the largest rainfall flood had a discharge of 450 m 3/y r whereas the largest snowmelt flood had a discharge of 363 m3/yr. However the difference is much greater for the period 1975-90. The largest rainfall flood had an estimated discharge of 776 m3/s while the largest snowmelt flood had a discharge of only 280 m3/s. Significant snowmelt often occurs during periods of rainfall thus contributing further to the flow magnitude. It is also interesting to observe that while the rainfall floods are much greater for the period 1975-90 than for the period 1952-74, the snowmelt floods are considerably smaller. This may indicate that temperatures were characteristically warmer in the more recent period, resulting in less snowpack accumulation and therefore lower magnitude snowmelt floods. The larger rainfall floods are possibly a result of higher magnitude rainfall events or greater immediate runoff during autumn storms due to warmer temperatures. These high magnitude floods have the ability to transport vast quantities of sediment and to change the channel configuration significantly. Therefore it seems reasonable to expect an increase in the quantities of sediment transported through the Chilliwack River and deposited in the Vedder River during the period 1975-90. 3.4. Sediment Transport During High Magnitude Flood Events The morphologic processes occurring upstream of Vedder Crossing determine the size and amount of sediment that is transported onto the fan (McLean, 1990). McLean identified and assessed the 3 major sources of sediment supply to the Vedder River: (i) bank erosion (ii) tributary creeks and (iii) landslides in terrace deposits. The greatest amount of bank erosion occurs between Liumchen Creek and Vedder Crossing as bank sediments in this 39 reach are unstable during major floods. Upstream of this location the bed and bank material is sufficiently large to resist erosion during most floods. Based on field observation and aerial photography, McLean identified Slesse Creek as the tributary which supplies the greatest amount of sediment to the Chilliwack River. The occurrence of landslides is limited temporally and spatially. However when landslides do occur they can be an important source of sediment supply to the river. During the largest exceptional floods extensive bank erosion occurs upstream of Vedder Crossing. Large quantities of sediment are available for transport to the Vedder River during these events. McLean emphasized that the quantity of sediment transported past Vedder Crossing during major floods is determined by the transport capacity of the river. As the large floods are responsible for most of the sediment transport into the Vedder River, sedimentation is primarily a transport-limited process. Church et al. (1989) analysed records of suspended load for 63 stations in British Columbia, one of which was the Chilliwack River. Analysis of the data showed that suspended sediment transport is highly episodic, with most of it occurring during major runoff peaks. Most of the rivers in the province, excluding those in coastal regions and in the Coast and Cascade mountain ranges, experience an annual sediment transport peak in spring due to the snowmelt freshet; between 65% and 90% of suspended sediment transport occurs during this time. The Chilliwack River basin experiences two major peaks of sediment transport. Most of the suspended sediment is transported during the spring/summer snowmelt period and the autumn/early winter rainfall period. Church et al. (1989) found that the December 1975 flood on ko the Chilliwack River had a greater sediment yield than any other entire year between 1966 and 1975. The movement of bed load (material which moves close to the bed) is expected to show even greater temporal variability as shear stresses do not exceed the critical threshold necessary for transport very much of the time. However once flow conditions exceed this threshold, there is a great increase in sediment transport as the larger material can be entrained. Therefore the movement of the bed material, which includes material moving close to the bed as well as in suspension, is probably even more highly episodic than the movement of suspended sediment. Therefore one may conclude safely that high magnitude floods are responsible for most of the aggradation in the Vedder River. There is no overbank fine material deposited along the Vedder River. Up to 30% of fines may be incorporated into the bed material as interstitial fill. Sand values in the Vedder River bed material are in the range 14-22% (see section 5.1.3), but most of the bed material in the Vedder River is greater than 2 mm. Therefore the fine and gravel fractions of the bed material probably display contrasting transport behaviours. The larger material travels shorter distances than the fine material as it requires larger shear stresses to be entrained and transported. The fine material travels greater distances during floods and fills in the voids of large material when it is deposited; it is, in effect, wash material. McLean (1980) analysed the change in median bed load grain size with discharge for the Vedder River (see figure 3.6). He found that there was a large increase in the D50 grain size between flows of 225 and 250 m3/s; above this discharge range the bed load consisted primarily of gravel. As the largest 50.0 20.0 c o 6 -a c o -a v m 10.0 5.C 2.0 1.0 0.5 .10 -_ : ; D l s c h a r ? e C o n v e r s i on f t 3 / s ffl^Zs. » i 2 000 57 4 000 113 J l 6 000 170 8 000 227 10 000 283 \ • 12 000 340 -• / 1 • • • • • • / • • • * - • / • 1 1 Note : A l l d a t a d e r 1 v e d • , f r o m W . S . C . mea-i i s u r e m e n t s I 1971-75 1 4000 6000 8000 10,000 12.000 Dischorse ( f t 3 / s ) 3.6. Bed load grain size measured near Yarrow (from McLean, 1980) 42 proportion of bed material is gravel-sized, activity levels in the river increase dramatically when flows become capable of transporting larger sediment. Since this study is concerned with detectable changes in volume which depend on relatively high activity levels, a threshold discharge for significant bed activity of 250 m3/s was chosen (the upper limit of McLean's transitional zone). A list of all mean daily flows which exceeded this threshold criterion during the study period 1958-90 is given in appendix 2. Both the magnitude and duration of a flood are important factors in determining the quantity of sediment that is transported. The hydrograph data reveal that several extended events occurred during the study period. In these cases a significant flood was followed by a temporary decrease in discharge, after which flows soon increased once again. The initial peak flow breaks up the coarse surface layer. Providing that the time lapse is short, or that flows remain at reasonably high levels during the intervening period, the surface may not stabilize before the second peak flow. These events may be particularly effective in eroding sediment upstream of Vedder Crossing and depositing it in the Vedder River. 43 4. Changes in Channel Morphology 4.1. Survey Data Surveys of the Vedder River were completed by the Water Management Branch of the Ministry of the Environment for most years during the period 1981-1991. A complete survey of the river consists of 49 cross-sections between the entrance to the Vedder Canal and Vedder Crossing (see figure 4.1; condensed reaches, which are introduced in section 4.5, are also illustrated). Cross-section density varies considerably along the length of the river; survey line spacing is significantly greater in the upper parts of the Vedder River. Complete surveys were conducted for the following years: 1981, 1982, 1983, 1984, 1987 and 1990. In addition, partial surveys of the river, which consist of cross-sections 12-28, are available for 1989, 1990 (this survey was conducted about 7 months prior to the complete survey conducted in this same year) and 1991. The dates of all surveys are given in table 4.1. Cross-section data were re-calculated by the Technical Support Section of the MOE with a common coordinated plot reference point added on the left bank to make direct comparisons possible. The data for 1981-84 were obtained from stadia measurements. Total station direct measurements were used to obtain the data for 1987 through 1991. The data for the more recent period are probably more accurate because of the improvement in the survey technique. 4.2. Long Profile Analysis In order to analyse the pattern of aggradation on the Vedder River long profiles were constructed for a series of years from 1981 to 1991. Mean bed levels were calculated from cross-section survey data using the equation: Figure 4.1. Cross-sections and condensed reaches along the Vedder River Table 4.1. Cross-section Survey Dates Survey Designation Complete or Partial Survey Dates-Regular Survey Dates Julian Day 1981 C April 27/May 28 117-148 1982 C April 26/May 7 116-127 1983 C Sept. 12/17 255-260 1984 C Feb.27/March 6 58-66 1987 C June 3/5, July 20/23 154-156 201-204 1989 P May 8/10 201-204 1990a P March 12/13 71-72 1990b C Oct. 16/20 289-293 1991 P Feb. 13/14 44-45 46 Z = Z {zi * Ax) (4.1) where ZJ is the mean bed elevation for the horizontal increment A x, and x is the bank-to-bank width. Long profiles were also constructed for 1958, 1963, 1975 and 1976 using mean bed level values reported by McLean (1980) in his M.A.Sc. thesis. He obtained mean bed levels using cross-section survey data and the equation given above. The exact positions of the earlier cross-sections could not be established as the Vedder River decreased in length by about 160 m between 1958 and 1975. However the designated positions are probably sufficiently accurate so that the results of the long profile analysis are not adversely affected. Flood discharges are discussed in the following section; the reader is advised to refer to appendices 1 and 2 as necessary. A comparison of the long profiles for 1958 and 1963 is presented in figure 4.2a. Aggradation occurred along most of the river except for a 1.5 km reach immediately upstream of the railway bridge that experienced degradation. During this time a series of channel maintenance procedures, including channelization, bank protection and channel clearing, were initiated at this location (McLean, 1980). It is likely that the degradation was the result of this channel work. There were 6 floods which exceeded the threshold for significant bed material transport (250 m3/s). One of these floods was an extended event, but the maximum discharge during this period did not exceed 300 m3/s. Despite the lack of exceptional floods, some notable aggradation occurred. 30 -i 25 H 20 H o £ 15 <! > 10H 5 H Railway Bridge 0-) 1 1 1 1 r 0 2000 Vedder Canal Figure 4.2a. Long profiles : 1958-1963 1958 1963 1 1 1 1 1 i i i | i i 4000 6000 8000 DISTANCE (m) Vedder Crossing 30 - i 2 5 -20 o £ 15 < > H ioH o Railway Bridge 1963 1975 1 r 0 I i i i | \ i i | r 2000 4000 6000 DISTANCE (m) Vedder Canal -i i | r 8000 Vedder Crossing Figure 4.2b. Long profiles: 1963-1975 00 30-i 25 H 20 H o S 15 < > 9 10H 5-0 Railway Bridge 1975 1981 i r 0 2000 i | i i i | r 4000 6000 DISTANCE (m) Vedder Canal 1 r 8000 Vedder Crossing Figure 4.2c. Long profiles: 1975-1981 -c-U3 30 -i 0 • i r 2000 Vedder Canal Figure 4.2d. Long profiles: 1981-1984: 1981 1984 i I i i 1 1 1 1 1 1 1 1 4000 6000 8000 DISTANCE (m) Vedder Crossing 30 25 H 20 o S 15 H 10H Railway Bridge 5H 0 i i i | r 0 2000 Vedder Canal Figure 4.2e. Long profiles: 1984-90 1984 1990 1 i i i | i i i 1 1 1 4000 6000 8000 DISTANCE (m) Vedder Crossing 30 25 H 20 H 55 o £ 15 •J IOH 5H 0 Railway Bridge 1 I T" 2000 0 Vedder Canal Figure 4.2f. Long profiles: 1958-1990 1958 1990 | I 1 I | I I I | I l 4000 6000 8000 DISTANCE (m) Vedder Crossing 30 -i 25 H 20 H 55 o g 15H 10 5-4 0 Railway Bridge — i 1 1 1 r 0 2000 1975 1976 1 1 1 " 1 — 4000 6000 DISTANCE (m) Vedder Canal 1 1 1 8000 Vedder Crossing Figure 4.2g. Long profiles: 1975-1976 Vedder Canal Figure 4.2h. Long profiles: 1983-1984 1983 1984 "i I i i i 1 i i 1 1 1 1 4000 6000 8000 DISTANCE (m) Vedder Crossing 30-i 25 H 20 H 55 O £ 15 ioH 5 H 0 Railway Bridge 1989 1990a i i 1 1 1 r 0 2000 Vedder Canal Figure 4.2i. Long profiles: 1989-1990a -i | i r 4000 I i i i i i 1 6000 8000 DISTANCE (m) Vedder Crossing Ul Ul 30 -i 25 -\ 20 -\ 55 o £ 15 w w IOH o Railway Bridge i i i 1 r 0 2000 Vedder Canal Figure 4.2j. Long profiles: 1990b-91 1 | i i 1 1 — 4000 6000 DISTANCE (m) 1990b 1991 n i | 1 1 8000 Vedder Crossing 57 Comparison of the long profiles for 1963 and 1975 shows significant aggradation between the canal entrance and the railway bridge (see figure 4.2b). There were 16 floods that exceeded the threshold discharge during this time, of which 4 were extended events. Maximum mean daily discharge exceeded 300 m /^s during 8 of the 16 floods. Degradation occurred between the railway bridge and Giesbrecht Road (previously Ford Road), probably a result of channelization that occurred at this location between 1964 and 1968. Material probably was transported out of this reach and deposited where the channel diverges downstream (McLean, 1980). Examination of the 1975 and 1981 long profiles reveals a significant lowering of the bed along most of the river length (see figure 4.2c). A large flood control operation, involving the removal of approximately 534 000 m 3 of gravel, took place in the summer of 1976. Gravel was removed from a number of locations between the entrance to the Vedder Canal and Peach Road. Most parts of the channel between these points were either directly or indirectly influenced by the dredging. The 1975 survey was completed prior to gravel removal and therefore the effects of dredging are inherent in this comparison. It is likely that the river was actually aggrading during this period. There were 7 floods which had discharges greater than the threshold criterion. Particularly large floods, which had discharges greater than 500 m /^s occurred in December 1975 and December 1980. Moreover, the 1975 flood was an extended event which lasted about one week. Significant aggradation occurred as the result of this flood. However the large amounts of material removed from the channel offset the deposition that occurred between 1975 and 1981, resulting in an overall lowering of mean bed levels. 58 Comparison of long profiles for 1981 and 1984 shows that only small amounts of aggradation occurred during this time (see figure 4.2d). Discharges exceeded the threshold for significant transport only four times during this 3 year period. Considerable quantities of material probably were deposited during the large flood of January 1984 (estimated discharge was 524 m3/s). However further dredging operations were undertaken in the summers of 1980 to 1983 inclusive, which limited the observed aggradation in the river. Examination of 1984 and October 1990 long profiles shows that the mean bed level increased at most locations, despite the removal of about 137 000 m 3 of sediment in the summer of 1990 (see figure 4.2e). During this time there were eleven floods that were capable of transporting considerable quantities of bed material; four were extended events. The exceptional flood of November 10 1989 (estimated discharge was 647 m 3 / s) probably was responsible for most of the aggradation in the river. However the gravel removal operations nearly offset the net deposition that occurred between 1984 and 1990, and net increases in bed level were minimal. It is useful to examine the long profiles for the first available survey in 1958 and the most recent complete survey of the river in October 1990 (see figure 4.2f). The section of the river extending from the canal entrance to the railway bridge is the only reach that experienced net aggradation. Considering the time span of this period, the deposition volumes are quite low. Most other locations experienced a lowering of mean bed levels. This suggests that the dredging operations have been successful in limiting long-term aggradation on the Vedder River. 59 It is possible to examine the patterns of aggradation for the large floods of 1975, 1984, 1989 and 1990 as re-surveys of the river were completed soon after the occurrence of these events. Unfortunately only partial surveys are available to examine the recent floods of 1989 and 1990. The long profile analyses for these 4 major floods were not affected by the major gravel removal operations of 1976 and 1990 as dredging did not occur between the dates used in these comparisons. An extended flood with a peak mean daily discharge of 530 m /^s occurred in December 1975 (see figure 4.2g). Although significant volumes of material were deposited immediately upstream of the railway bridge, there was significant degradation downstream of the bridge (see figure 4.2a). This contrasts with the pattern of channel change between 1963 and 1975. The change can probably be attributed to the channelization in the years preceding this flood. Due to the narrowing of the channel, there was a significant increase in velocities and shear stresses upstream of the railway bridge. The channel diverges downstream and sediment was deposited due to the decrease in the transporting capability of the flow. The channel maintenance procedures led to degradation and associated downstream aggradation which subsequently resulted in an overall decrease in slope. The river probably was in a state of disequilibrium and was still adjusting to its new configuration during the 1975 flood, which may account for the anomalous channel change pattern. The effects of the January 1984 flood can be examined by comparing the 1983 and 1984 long profiles (see figure 4.2h). The flood had a discharge of about 524 m /^s and was responsible for deposition in the reach immediately downstream of the railway bridge. Despite discharge identical to that of the 60 1976 flood, there was deposition where there had been erosion. However, bed levels did not change significantly upstream of the railway bridge. It is surprising that there were only relatively small amounts of deposition during this flood. A comparison of the partial long profiles based on the 1989 and March 1990 surveys provides some information about the nature of aggradation during the large flood of November 1989 (estimated discharge was 647 m3/s) (see figure 4.2i) The mean change in bed level for this flood, based on the available survey data, was +0.3 m. The maximum increase was 0.56 m. The mean bed level increased at all surveyed cross- sections. The lowest amount of aggradation occurred along a several hundred meter length of channel upstream of the railway bridge. Material was transported past this location as it is relatively narrow. Greater aggradation occurred both upstream and downstream of this reach. The pattern of deposition for the 1990 flood (estimated discharge was 776 m3/s) was similar to that described for the 1989 flood (see figure 4.2j). The mean bed level change was +0.26 m and the maximum increase was 0.52 m. The November 1989 and November 1990 floods were responsible for significantly larger quantities of deposition than the 1984 flood. It is difficult to make a comparison with the 1975 flood as the river was probably adjusting to new conditions at that time. These discharges are all estimates so it is not possible to make conclusive statements about the magnitude of flood which is necessary for various degrees of deposition. But it appears that some threshold exists between the discharge of the 1984 flood and the two floods of 1989 and 1990, as the aggradation during the latter two events is considerably higher. 61 4.3. Specific Gauge Analysis A specific gauge record has been compiled by McLean to examine the trends of aggradation on the Vedder River. A specific gauge record is constructed by obtaining the stage (water level) for a chosen discharge for each year. As a river aggrades and the bed elevation increases, the water level for a given discharges rises, providing there is no corresponding change in channel width. The data were obtained from stage measurements taken at the B.C. Electric Railway Bridge from 1954 to 1988. Figure 4.3 shows an increasing trend in stage for the three chosen discharges over the study period 1954-1988. This indicates that there is persistent deposition in the river over time. The large decreases in stage evident in the early 1960's and 1970's are a result of large gravel removal operations that occurred at those times. The average increase in bed level at this location (not including the years of significant gravel removal) is about 7 cm/yr. Unfortunately aggradation depth for the 1975 flood could not be evaluated as the results were affected by the 1976 gravel removal operation. The maximum annual discharges in 1980 and 1984 were respectively 533 m /^s and 524 m /^s. However despite these large floods, the aggradation depths at the railway bridge for these periods were not significantly greater than for other periods. Greater amounts of material are deposited immediately downstream of the railway bridge because of the significant channel widening. Therefore increases in bed levels at the railway bridge may be greater than at other locations. Examination of mean bed levels used in the construction of the long profiles shows that there is considerable variation in the values as channel morphology changes along the length of the river. 2.0 - j — i — i — i — i — | — i — i — i — i — | — i — i — i — i — \ — i — i — i — i — | — i — i — i — i — | — i — i — i — i — | — i — i — i — i — | — i — i — i — i — | 1950 1955 1960 1965 1970 1975 1980 1985 1990 YEAR Figure 4.3. Specific gauge analysis OS KJ 63 4.4. Total Volume Changes in the Vedder River 4.4.1. Procedures An analysis of volume changes along the Vedder River provides useful information about the pattern of deposition over time. The cross-section survey data can be directly compared for any combination of years as the elevations are given in terms of the geodetic datum and horizontal measurements are tied into a control point. The various surveys completed over the last decade did not occur on the same dates; most of the surveys were conducted between the months of March and October. Examination of the hydrograph data shows that most of the floods which could transport significant amounts of bed material did not occur between these months. Therefore it is reasonable to refer to results based on the cross-section data as annual values. The results for the 1984-87, 1987-90 and 1987-89 survey comparisons are, unless otherwise indicated, annual estimates that were calculated by dividing the net channel change for each of these periods by the number of years that each period spans. An assumption is made here that there is an even rate of change over each of these periods. This assumption is probably not valid and this issue is explored more fully later in this section. For some of the survey comparisons only data for cross-sections 12 to 28 are available. In these cases the total volume changes for this 2230 m reach were calculated. It should be kept in mind that these results are not directly comparable to the results for which complete survey coverage was available. A FORTRAN program was written to obtain the net change in area for each cross-section for all years of comparison. The survey data were used to interpolate the elevation at 1 m increments along the channel width. The 6k difference in elevation was calculated for each of these 1 m increments. The summation of these values provides an estimate of the net change in area for a given cross-section: n-1 A Z i + A z ( i + 1 ) AA = L ( 4 . 2 ) i = 1 2 where is the difference in elevation at a location between two years, z{+\ is the difference in elevation at the next 1 m increment and n is channel width. There are two possible approaches for delineating reaches. In one case the survey lines represent the mid-points of the reaches. The change in volume for a reach is obtained from the equation: AV = ( A A i * ( 4 . 3 ) where Dj is the summation of half the distance to the downstream cross-section and half the distance to the upstream cross-section. However this approach requires an approximation at the end points as there is no obvious distance that can be assumed to be represented by the change in area downstream of cross-section 1 and upstream of cross-section 49. The second approach, which is the one used in this study, is to calculate volume changes between each pair of cross-sections. The net volume change between each pair of cross-sections was obtained by multiplying the mean value of their change in area by the distance between them: 65 + A A ( i + D AV = * L ( i f i + 1 ) - ( 4 . 4 ) where AA^ is the change in area at cross-section i, A +1 is the change in area at the next upstream cross-section and i + i ) is the distance between the two cross-sections. This approach assumes no approximation at the end points. In this chapter, volume changes are calculated in terms of bulk quantities. Therefore the volume changes reported in this chapter are bulk changes and do not take into consideration the porosity of the material. Estimates of aggradation rates in the Vedder River were made by calculating the total net change in volume between the canal entrance and Vedder Crossing for each year. The net volume change for Vedder River was calculated by performing a summation of the volume changes between each pair of cross- sections: 48 V ( l _ 4 g ) = L ( A V i ) ( 4 . 5 ) i = 1 where A Vj is the change in volume for the 48 reaches delineated by pairs of cross-sections. This analysis examines only the observed changes in volume between cross-sections 1 and 49. At this point no conclusions about the amount of sediment transported past Vedder Crossing and into the Vedder River, or the amount of sediment transported into the Vedder Canal are made. These 66 issues are dealt with in the examination of the sediment budget in sections 5.2 and 5.3. It was mentioned previously that significant volumes of sediment were removed from the channel during a series of gravel removal operations in the early 1980's and 1990. It is worthwhile to examine net volume changes using two different approaches. In the first approach the effects of dredging, which are included in the data, make it possible to assess the actual change in conditions that occurred on the river. However a second approach is to examine "natural" rates of aggradation and this requires the elimination of the effects of dredging from the data. 4.4.2 Observed Volume Change Net volume changes along the Vedder River were calculated using the first approach and the results are presented in figure 4.4 (gross volume changes, which are discussed in section 4.4.3, are also illustrated in this graph). Large amounts of material were removed from the channel during the gravel removal operations and this affects the results quite significantly in some cases (see table 4.3 in section 4.5.2 for quantities removed). Therefore it is not possible to make any conclusions at this point about the importance of various floods in determining channel modification. It is readily apparent that despite dredging there was a net increase in volume on the Vedder River for each year over the last decade. The least channel change occurred between 1981 and 1982. About 1 200 m 3 of material were deposited during this period. The highest net deposition occurred between the survey dates for 1983 and 1984. There are minimal differences in deposition for the remaining Figure 4.4 Total volume changes (observed and gross) ON 1 68 three complete survey comparisons; their net changes in volume range from 24 400 m 3 to 34 300 m 3 . The value of 24 400 m3/yr for the period 1987-90 appears to be low when the high magnitude flood of 1989 is taken into consideration (estimated discharge was 647 m3/s). Two factors are probably responsible for the apparently low value of aggradation during this time. A flood control operation, involving the removal of approximately 187 000 m 3 of material, occurred in the summer of 1990 and significantly decreased the observed deposition between 1987 and 1990. Furthermore the assumption that there was equal aggradation between 1987 and 1990 decreases the apparent impact of the 1989 flood. Before further examination of this assumption it is necessary to first eliminate the effects of dredging from the data. The partial survey results provide some indication of the temporal pattern of channel change for various periods during the latter part of the decade, although the results are restricted to three reaches. Of the four partial survey comparisons, gravel removal occurred only between the 1990a and 1990b surveys. There was relatively low activity between cross-sections 12 and 28 from 1987 to 1989. The net channel change was about +3540 m 3 . During this time there was only one flood that was capable of transporting a notable quantity of sediment. The results for 1989-90a and 1990b-91 data reveal that large amounts of sediment were deposited in the river during these periods; the respective aggradation volumes were 91 900 m 3 and 74 700 m 3 . These large quantities of deposition are the result of the two major floods of November 1989 and November 1990. Between the surveys in March 1990 and October 1990 there was one large flood which exceeded the threshold discharge for significant transport. However the large sediment removal 69 operation in 1990, which removed about 187 000 m d of material from river, occurred during this time. The observed volume change was -54 800 m 3 . 4.4.3. Gross Volume Change In order to assess the "natural" rate of aggradation on the Vedder River the volumes of sediment removed during each dredging operation were added to the observed net volume changes for the appropriate survey comparisons. It is assumed here that if the material had not been removed from the channel, it would have remained somewhere between the canal entrance and Vedder Crossing. For the comparisons based on partial surveys an assumption is made that any gravel removed between cross-sections 12 and 28 would otherwise have remained in this same reach. The adjusted results show a very different pattern of aggradation on the Vedder River over time (refer to figure 4.4). The periods experiencing the highest rates of natural aggradation are 1982-83 and 1987-90, which contrasts with the observed values. Therefore it can be concluded that the dredging operations significantly decreased the observed volume changes during these two periods and were successful in limiting aggradation on the river. Of the complete surveys, the periods 1981-82 and 1984-87 experienced the lowest annual deposition rates. Material eroded in the Chilliwack River upstream of Vedder Crossing during significant floods is the primary source of bed material supply to the Vedder River. Between the 1981 and 1982 surveys there were no floods greater than the threshold for significant bed material transport, which explains the very low deposition rate. Despite the fact that there were no flows greater than this threshold, there was a net 70 deposition of 19 200 m0*. There evidently is some bed activity at relatively low discharges. Annual net deposition in the river was about 28 100 m 3 between 1984 and 1987. Seven floods exceeded the threshold criterion in this period. Two of the floods were extended events. However none of these floods was greater than 350 m3/s. The fact that there was minimal deposition during this time suggests that the occurrence of an extended flood does not necessarily result in high deposition volumes. Significant, but not exceptional, aggradation occurred during the period 1983-84. About 52 900 m 3 of material were deposited in the Vedder River. Only one flood exceeded the threshold criterion. It occurred in January 1984 and had a discharge of 524 m3/s. The greatest aggradation occurred during the periods 1982-83 and 1987-90. Annual deposition rates were 109 000 m3/yr and 86 600 m3/yr respectively. Between 1982 and 1983 only three floods had discharges greater than the threshold for significant bed load transport, and the largest had a discharge of about 313 m3/s. The high deposition rates cannot be explained by the frequency, magnitude or duration of floods. There were four significant floods between the 1987 and 1990 survey dates. However it seems likely that the particularly large flood of November 1989 (estimated discharge was 776 m3/s) was responsible for the high aggradation rate. Although dredging is now accounted for in the 1987-90 value the non-weighted averaging used to obtain an annual estimate of change diminishes the real importance of the 1989 flood. Complete survey data are not available for any date between 1987 and 1990 but a partial survey is available for 1989, making it possible to assess changes within this 71 time period along a 2 230 m reach of the river. It is estimated that net deposition was about 3 540 m3/yr between cross-sections 12 and 28 over the period 1987-89. Between the 1989 survey and the October 1990 survey approximately 133 000 m 3 of net deposition occurred in this same reach. The latter value represents 95% of the total deposition that occurred between 1987 and 1990. Furthermore the net aggradation rate between cross-sections 12 and 28 of .133 000 m3/yr is itself significantly greater than the three-year aggradation between cross-sections 1 and 49 of 86 600 m3/yr. The latter value is based on the assumption of an even rate of change. This suggests that the aggradation between the canal entrance and Vedder Crossing was substantially less than the estimated value during the first two years of the period 1987-90, whereas the deposition for 1989-90 was many times greater than the results indicate. The 1990a-90b partial survey comparison was affected by the 1990 gravel removal operation. Gross deposition was about 40 800 m 3 during this period. This value seems quite high, given that there was only one flood during this time which exceeded the threshold discharge for significant bed material transport. 4.5. Estimates of Volume Change along the Channel 4.5.1. Procedures The previous section analysed total deposition along the Vedder River. It is also important to obtain an understanding of the pattern of deposition over time and space. In order to do this it is necessary to analyse net volume changes for individual reaches for the various periods. 72 Volume changes were initially calculated for the 48 reaches that are delineated by each pair of cross-sections. It was decided to condense the data into 10 reaches to simplify the analysis. These reaches were chosen so that they are approximately the same length; the mean reach length is about 800 m (see figure 4.1 and table 4.2). In chapter 2 it was stated that reach length should approximate the average transfer length of bed material for good resolution. The transfer length can be estimated by measuring the distances between deposition sites. The average distances between bars in the Vedder River are 450 m for reaches 1-3, 500 m for reaches 4-6, and 530 m for reaches 7-10. Therefore assuming that two floods capable of transporting significant amounts of bed material occur each year, the chosen average reach length of 800 m may reasonably represent the transfer distance. In order to facilitate the identification of distinct zones of channel change, care was taken to ensure that each reach had a reasonably characteristic morphology. An example of the effect on the data of condensation is presented in figure 4.5. Although information about change at a specific location is lost, the overall pattern of change along the river remains the same. The effect of this condensing procedure was checked for every case to ensure that no significant loss of information occurred about the pattern of change (trend and general magnitude). It appears that the condensed reaches were well chosen as the overall patterns of change along the river are preserved. 4.5.2. Observed Volume Changes Observed net changes in volume for years of complete survey coverage are presented in figure 4.6a-e (gross volume changes, which are discussed in section 4.5.3, are also included in this diagram). Volumes of material Table 4.2. Condensed Reaches Condensed Reach Reach Limits (x-section #) Location: m from x-section 1 Length (m) 1 1-7 0-855 855 2 7-12 855-1545 690 3 12-18 1545-2390 845 4 18-23 2390-3040 650 5 23-28 3040-3775 735 6 28-34 3775-4670 895 7 34-38 4670-5610 940 8 38-41 5610-6375 765 9 41-45 6375-7310 935 10 45-49 7310-8175 865 Mean Value: 817.5 Figure 4.5. The effect of condensation on the representation of volume changes for the period 1981-82 8000 <^ 6000 CO S 4000 H O 2000 < a o u § -2000 3 O -4000 > -6000 -8000 c I 30 —i i i i i i i i i •  i i i i i i m l i i i i . ; . i i i i : i i i . i i i i . i i i i i i i 0 5 10 15 20 25 30 35 40 45 CROSS-SECTION (a) individual reaches 12 u >> \ CO & o — <-> -o 3 § U g 3 o > a a; 8 -4 -12 -i -i 1 1 - — — l r i i — —i 1 ••• i -' "*i 1 1 1 2 3 4 5 6 7 8 9 10 CONDENSED REACH (b) condensed reaches Figure 4.6. Reach volume changes (observed and gross) \ CO a, W O u w 3 H W TJ C o • o SZ 30-20-10-0 - 1 0 --20 - 30 --40 OBSERVED GROSS 1 ML V7^\ V?. 1 m 4 5 6 7 8 CONDENSED REACH 9 10 (a) 1981-82 Ul CD OO bO 1 CO CO N E T VOLUME CHANGE ( m 3 / y r ) (Thousands) o 5=0 o CO CO o CO S3 i. 9^ 60 50 40 30' 20 ^ c ffi 8 1 0 U =5 O 3 H W 0 - 1 0 -20 -30 -40 m ~i r 4 5 6 7 8 CONDENSED REACH 1 0 OBSERVED (c) 1983-84 CD CO r CO NET VOLUME CHANGE (m3/yr) (Thousands) I I -p^  C*J o o I ho o CM 8 ^ n cn as 00 CD O O J I I L _ r o O O I I 4=^  O !_ cn CD o o -1 I 1_ 4 5 6 7 8 CONDENSED REACH OBSERVED GROSS (e) 1987-90 80 removed from each reach during the various gravel removal operations are given in table 4.3. The results for 1981-82 show that volume changes along the river were minimal. The greatest observed changes were in reaches 3 and 8; about 10 300 m 3 of material were eroded from reach 3 and about 11 400 m 3 of material were deposited in reach 8. About 18 000 m 3 of material were dredged from reach 7 in the summer of 1982. Natural aggradation in reach 7 was considerably greater than the observed channel change. The observed volume changes for the ten reaches varied considerably during the period 1982-83, ranging from a positive value of 26 400 m 3 for reach 3 to a negative value of 29 800 m 3 for reach 6. During this period large amounts of material were dredged from reaches 3, 5, 6 and 7. The natural volume changes for all of these reaches were probably significantly higher than the observed values. The pattern of change found in the 1983-84 data contrasts with that found in the 1982-83 data. Whereas there was great variation in the values for 1982-83, the values for 1983-84 are generally lower and show less variation. Reaches 4, 5 and 10 experienced the least change; these are relatively narrow reaches and sediment is probably transported quickly through them. Only small amounts of material were deposited in most reaches between 1984 and 1987. This was a period of relative inactivity in the river. It is surprising that the results for 1987-90 do not show a considerably greater amount and extent of aggradation, given the high magnitude of the November 1989 flood. These values are partly explained by the fact that a major gravel removal operation, involving the excavation of about 187 000 m 3 of sediment, occurred in the summer of 1990. Large volumes of material 81 Table 4.3 Volumes of Material Removed from Condensed Reaches (all values are m d truck measure) Reach 1981-82 1982-83 1983-84 1984-87 1987-90 1 - - - 25 500 2 - - - - 16 600 3 - 12 900 - - 45 300 4 - - - - 6 420 5 - 6 640 - - 42 900 6 - 23 000 -7 18 100 32 400 -8 - - -9 - - - - 50 100 10 Total: 18 100 74 900 187 000 82 were removed from reaches 1-5 and 9. Furthermore, the assumption of an even rate of change between 1987 and 1990 means that a large proportion of the deposition that occurred during the 1989 flood is attributed to the period 1987-89. The effects of the November 1989 and 1990 floods can be analysed and compared as a partial survey of the river, covering reaches 3, 4 and 5, was completed soon after each of these floods (see figure 4.7ab). Both floods were responsible for significant amounts of aggradation. The greatest deposition occurred in reach 3, which is located immediately downstream of the railway bridge: about 57 000 m 3 of sediment were deposited during the 1989 flood and about 40 000 m 3 during the 1990 flood. There was net aggradation in reach 4, although the values were considerably lower. 4.5.3 Gross Volume Changes The previous section examined observed changes along the Vedder River. It is also crucial to analyse natural rates and patterns of change as this information is important when planning future flood control procedures. These results are also of particular interest as they may better represent the volume changes that would have occurred if gravel had not been removed from the river. An assumption is made in this analysis that sediment removed from the river during gravel removal operations would have remained within the same reach until the date of the next survey, if dredging had not occurred. This is the best assumption that can be made even though dredging actually does affect patterns of sediment transport. There remains some question as to whether deposition would have occurred in a similar way without the dredging. Figure 4.7. Volume changes for the 1989 and 1990 floods 60 CONDENSED REACH (a) 1989 flood 60 £ 50 CONDENSED REACH (b) 1990 flood 84 The volumes of material removed from reaches during the various dredging operations were added to the appropriate net volume change estimates (refer to figure 4.6a-e). The data for 1981-82, 1982-83 and 1987-90 include the effects of gravel removal; the data for 1983-84 and 1984-87 do not need adjusting for this analysis as there was no dredging during these periods. Examination of the graphs reveals trends similar to those found in the results of total deposition between cross-sections 1 and 49 (see section 4.4). Activity levels were generally low during the period 1981-82 as no significant floods occurred. Despite the lack of major floods, there was some notable degradation in reaches 3 and 9 and some aggradation in reaches 7 and 8. The results for 1982-83 suggest that this was a period of considerable aggradation in the river, although dredging removed large volumes of the deposited material from the channel. The gross deposition volumes in reaches 3 and 7 are particularly high; 39 400 m 3 and 44 100 m 3 respectively. A large amount of sediment was eroded upstream of Vedder Crossing and supplied to the Vedder River. This seems somewhat surprising as the largest flood had a discharge of only 313 m3/s. There were no extended floods during this time. Eight of the ten reaches aggraded between the 1983 and 1984. The deposition volumes are not particularly high; the maximum value was 13 700 m 3 . These volumes seen somewhat lower than might be expected, when the large flood of January 1984 is taken into consideration. Volume changes were relatively minor between 1984 and 1987; a maximum deposition of 8 000 m 3 occurred in reach 9. Although 7 floods had discharges greater than the threshold for significant bed material transport, the maximum discharge was only 349 m3/s. Two of these floods were 85 extended events, which once again suggests that moderate floods of long duration do not result in particularly high aggradation rates. This analysis shows that gross deposition was considerably greater than the observed volumes of deposition between 1987 and 1990 because of the removal of large quantities of sediment from the river during the 1990 dredging operation. It should be emphasized that because of the assumption of an even rate of change over this period, the analysis overestimates the amount of deposition between 1987-89 and underestimates the value for 1989-90. In order to gain insight into the time distribution of channel change within the period 1987-90 for individual reaches, estimates of natural volume change for the partial survey comparisons of 1987-89 and 1989-90 were examined (see figure 4.8ab). About 96% of the net deposition in both reaches 3 and 5 occurred between 1989 and 1990, which suggests that there may be some degree of consistency in the time distribution of change for the 10 reaches. However the value drops to 74% for reach 4. Therefore it cannot be assumed that the distribution of change over the period remains the same between the canal entrance and Vedder Crossing. Morphological characteristics change along the river, which affects the relative deposition and erosion in different reaches. If the transfer length is less than the reach length, then what is in the channel from the last flood affects the subsequent behaviour. However most of the channel change between 1987 and 1990 undoubtedly occurred in the final year of this period; volume change estimates for 1989-90 may be nearly 3 times greater than the estimates based on the 1987 and 1990 surveys. Between the 1987 and 1989 survey dates only one flood was capable of transporting significant quantities of sediment. Furthermore this flood had a discharge of only 250 m3/s. Although the Figure 4.8. Volume changes within the period 1987-90 \ CO u w I W 80 70 60 50 ~ 40 •o | 30 o fc, 20 10 0 -10 -20 V77A 2 3 4 5 6 7 8 9 CONDENSED REACH 10 (a) 1987-89 (b) 1989-90 87 discharge of several floods exceeded the threshold for significant transport between 1989 and 1990, the November 1989 flood undoubtedly was responsible for most of the aggradation that occurred during this time. In order to better observe temporal and spatial patterns of natural sedimentation on the Vedder River plots of cumulative percent departure from the mean were constructed for the estimates of natural volume change. Using this approach one can assess whether a particular year experienced greater or lesser natural net volume change relative to other years, in each reach. An above or below average value of net volume change does not necessarily imply either aggradation or degradation; this changes depending upon whether the mean value of volume change is positive or negative. Cumulative percent departure plots are presented in figure 4.9. These results support the findings of the initial analysis of volume changes over time. A majority of the reaches show below average change between 1981 and 1982. This is explained by the fact that the largest flood during this time had a discharge of only 142 m3/s. There is no trend in the data for the period 1982-83; some reaches show above average values of net change while others show below average values. The values for most of the reaches are below average for 1983-84 and 1984-87. It is surprising that the changes for 1983-84 were not greater as the second largest flood of the period 1981-90 occurred during this time. The below average changes between 1984 and 1987 are probably explained by the fact that there were no particularly large floods during this time. Nine of the ten reaches had above average gross aggradation rates for the period 1987-90, which is a result of the high magnitude flood in November 1989. This contrasts with the data for 1982-83, the other period during which large channel changes occurred. Whereas 88 Figure 4.9. Cumulative percent departure over time for each reach (continued on next page) UJ CC =3 I— CC < UJ O I— < <_> 1981- 1982- 1983-1982 1983 1984 -i \ 1 1 1984- 1985- 1986- 1987- 1988- 1989-1985 1986 1987 1988 1989 1990 PERIOD 89 -200--400--600-200-i REACH 6 REACH 7 REACH 8 REACH 9 REACH 10 i i 1 1 1 1 1 1981- 1982- 1983- 1984- 1985- 1986- 1987- 1988- 1989-1982 1983 1984 1985 1986 1987 1988 1989 1990 PERIOD 90 nearly all of the changes were positive for 1987-90, they were both positive and negative for 1982-83. For comparison a cumulative percent departure plot was constructed for the annual maximum mean daily discharge series (see table 4.4 and figure 4.10). Examination of this graph suggests that there is not a direct relation between peak floods and deposition. The floods of January 1984 and October 1990 are the only ones which show discharge values above the average annual peak flood. Despite this, channel changes during the period 1983-84 were below average in most cases. The flood analysis shows that the magnitudes of the peak floods in 1987-88 and 1988-89 were below average. The low values of net channel change between 1981-82 and 1984-87 are explained by the relatively low magnitude of peak floods during these periods. The Vedder River was very active during the period 1982-83, experiencing high volumes of both erosion and deposition. However total deposition in the river was considerably greater than total erosion, which means that large quantities of sediment were transported past Vedder Crossing. The maximum discharge was slightly below average in comparison to other annual peak floods. The flood analysis does not provide an explanation of the high activity levels during this period. Examination of figure 4.11 reveals whether a reach had a natural change greater or lesser than the average along the river, for each survey comparison. Reach 3 appears to be an important depositional zone as it has above average values for each period. Channel width increases at reach 3 and flow divergence results in the deposition of large quantities of sediment. The data for reaches 8 and 10 show a trend of below average values. Reach 10 is relatively straight, narrow and steep, and most of the material transported 91 Table 4.4 Maximum Floods Between Survey Dates Period Maximum Flood Between Survey Dates Cm3/s) * 1981-82 142 1982-83 313 1983-84 524 1984-85 273 1985-86 344 1986-87 349 1987-88 250 1988-89 222 1989-90 647 Average: 340 * when no survey was completed for a year (1985, 1986, and 1988), an approximate survey date was chosen based on the surrounding survey dates; the exact date does not matter as surveys are conducted during low flow periods 1981- 1982- 1983- 1984- 1985- 1986- 1987- 1988- 1989-1982 1983 1984 1985 1986 1987 1988 1989 1990 PERIOD Figure 4.10. Cumulative percent departure for maximum floods Figure 4.11. Cumulative percent departure along the river for each period 1000 - i UJ cn ZD h-< UJ O I— O 4 5 6 7 C O N D E N S E D R E A C H 8 10 93 past Vedder Crossing can travel quickly through this reach. Consequently there is minimal sediment accumulation at this location. Channel width increases at reach 8 so it is not immediately apparent why this reach experiences below average values of natural change. The natural change estimates for reach 9 are below average for 1981-82 and 1982-83 but it appears that this location has become a major depositional zone since that time. There are no obvious trends in the data for the remaining reaches. 4.5.4. Stream Power Index In order to understand the patterns of deposition along the river an index of the stream power was calculated for each reach. Stream power is obtained from the equation: u = S d v (4.6) where S is slope, d is depth and v is velocity. Since it is known that: Q vT = d V (4.7) where Q is discharge, then the equation for stream power can be written in the form: S Q a = (4.8) w Because all calculations are linear, for different floods: 0)1 CO' Q 2 (4.9) An index of stream power can be obtained from the equation: S P . I . = (4.10) w Average slope and width were calculated for each reach using the survey data. Figure 4.12 displays the index of stream power for each reach. The ability of the river to transport material decreases as one moves downstream of Vedder Crossing. Local minima of stream power are found in reaches 7 and 8. The stream power increases in reaches 4, 5 and 6, primarily because of the reduction in channel width. The lowest values of stream power are found in reaches 1, 2 and 3. These data support some of the observations based on the cumulative percentage departure plots. Reach 3 was found to be a depositional zone and Reach 10 a transport zone. However the volume change for reach 8 was below average, although the stream power index is quite low relative to other reaches. 0.0 2 5 6 REACH 8 9 10 Figure 4.12. Power index UD U l 96 5. Gravel Budget of the Vedder River 5.1. Sediment Analysis 5.1.1. Overview The morphologic approach to sediment transport analysis has the potential to provide a relatively straightforward method for estimating transport rates of size classes (Church et al., 1987). However previous studies have not focussed on this aspect. The procedure is as follows. The size distribution of bed material is determined along the entire study reach. The values of net volume change for each reach are then adjusted for each size class, according to the proportion of bed material in that category. If one sediment transport rate is known for each size class (or can be reasonably estimated) then the sediment budget calculations can be extended along the river. 5.1.2. Sampling Procedures Fifteen samples were taken between the entrance to the Vedder Canal and Vedder Crossing. Time, access and site suitability constraints limited the number of samples collected. The fifteen samples provide a reasonable characterization of the bed material along the Vedder River (refer to figure 4.1 for sample sites). The average inter-site spacing is about 600 m. There is at least one sample in each condensed reach, with the exception of reach 2. Although it was initially thought that there was no gravel in the Vedder Canal, field observation revealed that there are significant amounts of gravel there. This observation becomes critical when estimating a transport rate for the sediment budget calculations. Four samples were taken from the canal to assess the size distribution of material that is transported past cross section 1. 97 The results of the grain size distribution analysis for the canal are given in the discussion about the transport assumption in section 5.2 Sampling criteria are given in Church et al. (1987) to ensure that all size classes in the population are adequately represented in the sample. It was found that the largest clast should not account for more than 0.1% of the total sample weight. Because of the large sizes encountered (the D m a x values range from 95 mm at the canal entrance to 260 mm at Vedder Crossing), a more relaxed criterion was adopted in this study. A 300 kg sample was taken at sites where the b-axis of the largest individual clast did not exceed 128 mm. When D m a x is 128 mm, a 300 kg sample represents a 1.0 % sampling criterion. A 500 kg sample was taken if the b-axis of the largest clast exceeded 128 mm. For a D m a x of 128 mm, a 500 kg sample represents a 0.6 % sampling criterion. The criterion for a 500 kg sample with a D m a x value of 260 mm is about 5%. These are nonetheless very large samples. In order to assess the composition of bed material subsurface samples were taken. Because it is essential to obtain samples that adequately represent the material involved in active channel change, samples were collected from bar locations that were situated along the main flow path of the river at the time of deposition. The surface was first cleared to the depth of the deepest-lying exposed grain. Coarse material (> 32 mm) was sorted into size classes using hand templates and rocker sieves and weighed in the field. Splits of about 10 kg were returned to the laboratory for analysis of the -32 mm fraction. 98 5.1.3. Proportions of Bed Material in Various Size Classes The proportions of material greater than various grain sizes were calculated for each of the samples. Preliminary results showed that the data for two samples, from site 2 and site 6, did not fit the overall trend in the data (these anomalies are shown in figure 5.1; this figure is discussed in more detail in the following paragraphs). Because of access problems in the vicinity of site 2, the sample was taken from a small side bar that probably is not representative of the active bed material in the main channel. It is not clear why the texture of the material at site 6 is so much finer than for the surrounding sites. As the primary objective of this sediment analysis is to describe general trends in the data, the anomalous data for these sites were excluded from further analyses. The percentages of material greater than each chosen grain size were obtained from the results for the 13 remaining sites on the Vedder River. Linear and exponential regression analyses were performed on the data for each grain size. The overall fit and the values of these two methods are similar for the smaller grain sizes but the overall fit of the linear model is considerably better for the larger grain sizes. The linear regression model was adopted for all grain sizes. The individual data points and regression lines for each grain size are presented in Figure 5.1; the regression equations, r^ values and standard errors of estimate are presented in table 5.1. Despite the low r^ values for the smaller grain sizes, the regression lines do appear to follow the main trends in the data fairly well. The standard errors of estimate for the data sets are fairly consistent, which shows that there is little variation in the reliability of the regression lines. 99 Figure 5.1. Regression lines for grain sizes (continued on next page) " denotes anomalous values 100 Table 5.1 Regression Analyses for Grain Sizes % Greater than x mm Regression Equation rz S.E.E. 2 y = O.OOllx + 77.5 0.30 4.2 4 y = 0.0014x + 70.5 0.36 4.7 8 y = 0.0018x + 60.8 0.43 5.2 16 y 0.0024x + 48.3 0.52 5.9 32 y = 0.0030x + 32.6 0.65 5.5 64 y — 0.0038x + 9.19 0.74 5.7 128 y 0.0034x - 9.41 0.70 3.4 256 y — 0.0071x - 52.4 * * * too few points to obtain a value 102 The initial 128 mm regression line does not fit the otherwise very strong trend showing a gradual increase in slope sensitivity as grain size increases. Examination of the data revealed that the percentages of material greater than 128 mm for each of the two most upstream sites are lower than expected. The D m a x is very large so the sample size may be inadequate to characterize the largest sizes properly. The results for these two sites significantly influenced the regression analysis Therefore these data points were excluded in the regression analysis of the 128 mm data. The resulting best-fit lines for all sizes are presented together in figure 5.2. There is a clear trend showing an increase in slope as grain size increases. The results indicate that there is no material greater than 256 mm downstream of 7 735 m and greater than 128 mm downstream of 2 730 m. The gentle slopes of the best-fit lines for the smaller grain sizes indicate that the relative proportion of finer material remains about the same along the length of the Vedder River. The greater slopes of the best-fit lines for the larger grain sizes suggest that the size and abundance of the largest material decrease rapidly as one moves downstream from Vedder Crossing. Using the regression equations, percentages of material greater than and less than 2 mm, and the percentages of material in each size class were calculated for each reach (see table 5.2). The proportion of gravel-sized material (> 2 mm) decreases as one moves downstream. Gravels comprise about 87% of the bed material in the reach directly below Vedder Crossing and about 75% in the reach immediately upstream of the canal. When the sediment < 2 mm comprises less than 25-30 % of the bed material, it is considered to be interstitial fill (see section 3.4). The proportions of material less than 2 mm along the Vedder River are below this Reach: lOO-i 9 0 -8 0 -< 7 0 -— 6 0 -W "— < 5 0 -W PS O 4 0 -3 0 -20 -10-0 -6 8 10 -i | i | i I i j i | i | i | i | i | 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 DISTANCE FROM X-SECTION 1 (m) Figure 5.2. Combination plot of regression lines for grain sizes Table 5.2. Percentage of Bed Material in Various Size Classes (Based on Regression Equations) (all values are %) Reach <2mm >2mm 2-4mm 4-8mm 8-16mm 16-32mm 32-64mm 64-128mm 128- >256mm 256mm 1 22.0 78.0 6.89 9.48 12.2 15.5 23.1 10.8 0 0 2 21.2 78.8 6.67 9.15 11.8 15.0 22.4 13.8 0 0 3 20.3 79.7 6.45 8.83 11.3 14.6 21.8 16.7 0 0 4 19.5 80.5 6.23 8.52 10.9 14.1 21.2 19.4 0.256 0 5 18.7 81.3 6.03 8.22 10.5 13.7 20.6 19.9 2.33 0 6 17.8 82.2 5.80 7.88 9.97 13.3 19.9 20.2 5.14 0 7 16.7 83.3 5.54 7.50 9.42 12.7 19.2 20.6 8.31 0 8 15.8 84.2 5.29 7.14 8.91 12.2 18.5 20.9 11.2 0 9 14.8 85.2 5.05 6.78 8.41 11.7 17.8 21.3 14.2 0 10 13.8 86.2 4.79 6.40 7.87 11.2 17.0 21.6 14.4 2.91 105 criterion in all cases (see table 5.2). Church et al. (1989) estimated that on average about 50 000 m3/yr of suspended sediment are transported past Vedder Crossing. Given the low proportions of fine material in the bed and the high magnitude of the suspended load transport estimate, it seems reasonable to assume that most of the material < 2 mm is interstitial fill. This indicates that fine material in the bed is transported as throughput; it moves considerably greater distances during floods than the larger material. Even though the nature of the transport assumption used in this study meant that the sediment budget had to be constructed for gravel material, it is the erosion and deposition of material > 2 mm that determines most of the morphological changes in the Vedder River. The percentages of bed material in the size classes < 64 mm do not change very much between Vedder Crossing and the Vedder Canal; the maximum variation between reaches for these size classes is about 5%. In contrast, there is nearly 10% drop in 64-128 mm material between reach 4 and reach 1. The fraction of material in the size class 128-256 mm decreases from 14% in reach 10 to 0% in reach 3, while there is no material greater than 256 mm downstream of reach 10. The rapid decrease in the amount of large material found in the bed is typical of alluvial fans (Blissenbach, 1952). Sediments are coarse on the steep upper slopes and decrease in size on the gentler lower slopes as the transporting capability of the flow decreases. 5.2. Construction of the Gravel Budget In order to construct the sediment budgets for the Vedder River it was initially assumed that the gravel transport rate into the Vedder Canal is zero 106 i.e., that no gravel is transported out of reach 1 (the downstream limit of this reach is located at the canal entrance). Gravel budgets were constructed for periods of complete survey coverage; 1981-82, 1982-83, 1983-84, 1984-87 and 1987-90. Net changes in volume for the latter two periods were divided by the three years that each period spans to obtain annual estimates. Gross volume change estimates of section 4.5 were used in the calculations. Since the assumed transport rate applies only to gravel-sized material, each estimate of net volume change was multiplied by the proportion of bed material > 2 mm in that particular reach. Furthermore, as transport values are generally given as mineral volumes per unit time, each gravel volume change estimate was multiplied by (1-p), where p is the porosity of the material. The porosity of bed material in gravel bed rivers falls in the range 0.3 to 0.4; the value 0.35 has been adopted in this study. The adjusted volume change estimates were then used in the construction of the sediment budget. All gross and observed volume changes, dredging volumes and transport rates reported in this chapter are given in terms of mineral volumes of material > 2 mm, unless stated otherwise (see table 5.3 for gross mineral volume changes > 2 mm). Initial gravel budget computations resulted in negative transport rates for two periods, 1981-82 and 1984-87. The outputs were negative for six reaches in 1981-82 and for two reaches in 1984-87. This implies that gravel was transported in an upstream direction, which clearly is not the case. The negative transport rate of -764 m3/yr in the 1984-87 data is probably within the precision of the volume change estimates and does not require further examination. Examination of the 1981-82 gravel budget reveals several negative transport rates, the largest having a value of about -5880 m3/yr. 107 Table 5.3. Gross Volume Changes (mineral volumes > 2mm) (all values in m'-Vyr) Reach 1981-82 1982-83 1983-84 1984-87 1987-90 1 -29.2 10 800 2 040 -764 8 630 2 -524 4 310 3 430 429 6 210 3 -5 320 20 400 4 790 2 210 12 900 4 1 410 791 -1 980 2 480 3 760 5 -545 6 710 713 1 360 7 550 6 1 290 -3 670 6 930 3 010 2 620 7 11 300 23 900 3 300 -634 -3 550 8 6 240 1 340 1 960 1 770 -12 100 9 -3 210 -10 900 7 560 4 440 18 300 10 3.40 3 350 -502 842 1 130 108 Several factors, acting either alone or in combination, may be responsible for the breakdown of the budget. If the volumes of material dredged from the channel during the period 1981-82 were underestimated, the deviation between the values used in this analysis and the correct values could be large enough to result in negative transport rates. It is also possible that the estimates of net volume change do not provide an accurate representation of true channel changes during this time. These two problems may also exist within the data for years not showing negative transport rates. The fact that there are significant amounts of gravel in the Vedder Canal suggests that the assumption of a negligible gravel transport rate past cross-section 1 is incorrect. Bed material samples were collected at four locations in the Vedder Canal; 2 800, 2 200, 1 500 and 700 m downstream of cross-section 1. These sample locations are referred to as sites 1, 2, 3, and 4 respectively. Site 3 is located immediately upstream of the bridge at Keith Wilson Road. The data for this site were not included in the final results as the high proportions of large material do not fit the overall trend in the data. Gravel between 44 and 64 mm was found at this location, which suggests that some large material can be transported a considerable distance in the canal. However the largest material found at site 4 was between 22.6 and 32 mm. The results of the size distribution analysis for the canal material are incorporated into figure 5.3 which shows percentages of material greater than x mm along the Vedder River. Large amounts of gravel are found in the canal; about 68% of the bed material at site 4 is greater than 2 mm. The proportions of gravel decline quite rapidly downstream of this location. Approximately 27% and 5% of the bed material is greater than 2 mm at sites 2 and 1. The gravel transport past site 1 (2 800 m downstream of the canal 100 - i Entrance to Vedder Canal Vedder Crossing 2 mm 4 mm 128 mm 256 mm T - ° - I — f — r * T — f — i — i — P -\ r | •3000 -1500 0 1500 3000 i i I I I | r 4500 6000 7500 DISTANCE FROM X-SECTION 1 (m) positive numbers are in the upstream direction i—1 9000 Figure 5.3. Percentages of material greater than x mm 110 entrance) appears to be negligible. This would be a better transport assumption than the one used in this study. The use of this transport assumption would require extension of survey coverage so that volume change estimates could be made as far downstream as this location. However survey data are not available for the canal. There appears to be minimal aggradation in the canal, which suggests that the gravel transport past cross-section 1 is low; it may be in the order of 103 m3/yr. Several assumptions were made to calculate an approximate average transport rate into the canal. It was assumed that the canal slope was about zero immediately after construction (which will produce positive bias in the calculation) and that since that time aggradation has been occurring in the form of a wedge between the canal entrance and site 1 (2 800 m downstream from the canal entrance). Based on the results of the sediment analysis it was assumed that 50% of material in the canal is greater than 2 mm. A present-day canal slope of about 0.001 and a canal width of 100 m were also assumed. Given these assumptions it was estimated that about 1.96 x 10^  m 3 (bulk volume) of material has been deposited in the canal since its construction in 1919. This translates into an average annual transport rate at the canal entrance of about 1.82 x 103 m3/yr. This result supports the contention that transport rates into the canal are in the order of a thousand m3/yr. If it be assumed that the material transported out of reach 1 and into the canal is deposited approximately evenly along its length, then for any transport rate at cross-section 1 the corresponding deposition in the canal can be calculated. Assuming an average canal width of 100 m, a gravel transport rate into the canal of 1 820 m3/yr, and that most of the gravel is deposited between the canal entrance and site 1 (a length of about 2 800 m) the annual 111 depth of aggradation in the canal is 0.65 cm. This represents 6.5 cm of aggradation during the 10 year study period and 47 cm of aggradation since the construction of the canal in 1919. These aggradation depths should be detectable. Given the uncertainty and magnitudes of the volume change estimates, the assumption of negligible transport into the canal probably did not significantly affect the final results. An error analysis of the data is performed in section 5.3, which supports this contention. It is nevertheless clear that the zero transport assumption is incorrect. This incorrect assumption is probably at least partly responsible for the negative transport rates for 1981-82. If the transport rate out of reach 1 is changed to the value of the largest negative transport rate then the gravel budget balances, would imply that 5 880 m 3 of gravel were transported into the canal between the 1981 and 1982 surveys. Although this adjustment is necessary to make the budget balance, the incorrect transport assumption may or may not have been the actual cause of failure. It is possible that poor representation of channel change was responsible for the negative transport rates; this issue is explored in more detail in chapter 6. Furthermore, there is no reason to believe that 1981-82 was the only period to experience gravel transport into the canal; particularly as there were no large floods during this time. It seems likely that gravel-sized material was transported into the canal during most time periods. Therefore all calculated transport rates are lower-bound estimates, and each estimate should be increased by the gravel transport rate into the canal for that time period. However the correct gravel transport rates are not known so this adjustment cannot be performed. 112 The gravel budgets are presented in figure 5.4a-e. The output for reach 1 was adjusted for 1981-82 and 1984-87 so that all inputs and outputs are greater than or equal to 0 m3/yr. Even though the.assumed gravel transport rate probably is not valid in most cases, the results of the gravel budgets illustrate the overall patterns of bed material transport along the Vedder River. The relative differences in transport rates for different locations are maintained as all inputs and outputs are underestimated by exactly the same amount. The adjustments relative to the calculated inputs at Vedder Crossing are 36% for 1981-82 and 5% for 1984-87. The adjustment performed for the former period is more significant than for the latter period. Furthermore, if actual gravel transport rates into the canal are within the precision of the volume change estimates then the calculated transport rates may not be adversely affected by the incorrect assumption. 5.3. Interpretation of the Gravel Budgets Sediment removal operations occurred between the dates of the 1981-82, 1982-83 and 1987-90 surveys. For reaches from which material was removed, the differences between the within-channel inputs and outputs do not represent the observed channel changes; these values are the summations of observed gravel volume changes and volumes of gravel removal. The differences between transport rates into and out of a reach reflect the natural channel changes that would have occurred had dredging not taken place. These are the results that identify zones of deposition and erosion along the Vedder River. The 1981-82 gravel budget indicates that 16 500 m 3 of gravel were transported past Vedder Crossing and into the Vedder River. This value is 113 Figure 5.4. Gravel budgets for the Vedder River E-« CO K Q O 2 (X •< "Z & <° E-s Q W CO 80 n 70 60-50-40-30-20-Sediment Budget ) 2mm — - Error Range Reduced Survey Coverage 1000 2000 3000 4000 5000 6000 7000 8000 9000 DISTANCE FROM X-SECTION 1 (m) (a) 1981-82 DISTANCE FROM X-SECTION 1 (m) (b) 1982-83 114 Sediment Budget ) 2mm Error Range Reduced Survey Coverage 1000 2000 3000 4000 5000 6000 7000 DISTANCE FROM X-SECTION 1 (m) 1 1 1 8000 9000 (c) 1983-84 Sediment Budget ) 2mm — - Error Range Reduced Survey Coverage r ^ r i — i — i — I — r 1000 2000 3000 4000 i 1 r 5000 6000 7000 8000 9000 DISTANCE FROM X-SECTION 1 (m) (d) 1984-87 115 (e) 1987-90 116 not very large, which indicates that there was minimal activity in the Chilliwack River upstream of Vedder Crossing. There were three distinct transport zones during this period. The most upstream transport zone comprises reaches 9 and 10. Transport rates were approximately constant; the average value is about 17 600 m3/yr. Downstream of this location transport rates declined fairly rapidly, which indicates that reaches 7 and 8 comprised a depositional zone during the period. About 9 800 m 3 of gravel were dredged from reach 7 in the summer of 1981, accounting for nearly all of the difference between the gravel input and output values for this reach. Downstream of this location transport rates maintained an approximately constant value of 1 100 m3/yr. About 5 320 m 3 of gravel were eroded from reach 3, which resulted in a slight increase in downstream transport rates. A gravel transport rate of 5 880 m3/yr at cross-section 1 had to be assumed for the budget to balance. The overall magnitudes of transport rates show that 1981-82, a period with no significant floods, was one of relative inactivity. The 1982-83 gravel budget reveals that transport rates were significantly greater during this period. About 57 000 m 3 of gravel were transported past Vedder Crossing and into the Vedder River. This appears to be a high transport rate, given the lack of exceptional floods. Of the three floods greater than the threshold for significant bed load transport, the largest had a discharge of 313 m3/s. It is not possible to account for the large amount of gravel transported past Vedder Crossing on the basis of the discharge data. The transport rates into and out of reach 10 were constant, which indicates minimal channel change. Indeed, transport rates through reach 10 were approximately constant for all gravel budget periods, although the magnitude of the value does change. About 10 900 m 3 of gravel were 117 eroded from reach 9 during this period. Transport rates remained approximately through reach 8. Although output was significantly lower than input for reach 7, the actual deposition volume was only 6 330 m 3 . About 17 500 m 3 of gravel were dredged from this reach. Transport rates remained approximately constant through reaches 4, 5, and 6. This is a relatively narrow section of the river that probably experiences high flow velocities during floods. Material is transported relatively quickly through this reach. Transport rates decreased through reach 3; the input was 35 500 m 3 and the output was 15 100 m 3 . The observed change in volume of gravel for this reach was about 13 700 m 3 and about 6 680 m 3 of gravel were removed during dredging. Reaches 1 and 2 comprised a depositional zone as indicated by decreasing transport rates. Gravel removal operations did not occur between the dates of the 1983 and 1987 surveys. The results for 1983-84 show a fairly constant decrease in transport rates along the Vedder River, which contrasts with the two previous gravel budgets. The 28 200 m 3 of gravel transported past Vedder Crossing and through reach 10 were deposited approximately evenly along the length of the river. The large flood of January 1984 was the only one which exceeded the threshold for significant transport (estimated discharge was 524 m3/s). Transport rates decreased from a value of 28 700 m3/yr into reach 9 to a value of 8 990 m3/yr out of reach 6. Transport rates through reaches 5 and 4 had a constant value of about 9 100 m 3 . These reaches also showed constant transport rates for the 1981-82 and 1982-83 gravel budgets, which suggests that this is a zone of minimal change. The 10 300 m 3 of gravel transported out of reach 4 were deposited approximately evenly between reach 3 and the Vedder Canal. 118 The gravel budget results for 1984-87 show that about 15 900 m d of gravel were transported past Vedder Crossing. This indicates that minimal erosion occurred upstream of Vedder Crossing. Seven floods exceeded the threshold discharge for significant bed material transport, although none of these was exceptional in magnitude. There was a progressive decrease in transport rates along the Vedder River; this indicates that deposition rates were about equal for all reaches. Significant amounts of material were deposited in reaches 4, 5, and 6 relative to the other reaches, a result in contrast with the other gravel budgets. There was also some significant aggradation in reach 3. However channel changes in reaches 1 and 2 were small. An adjusted gravel transport rate of 764 m3/yr was used to balance the gravel budget. However this value is probably within the precision of the volume change estimates. In 1987-90, 45 300 m 3 of gravel were transported into the Vedder River. It should be kept in mind that most of the channel change for this period occurred in 1989-90. The volume change estimates and transport rates were probably considerably lower for 1987-89 and up to three times higher for 1989-90 than the reported values. Between the dates of the 1987 and 1990 surveys there were four floods which exceeded the threshold for significant bed material transport. Given the exponential increase in bed load transport with increasing discharge, it is reasonable to assume that the exceptional flood of November 1989 (estimated discharge was 776 m 3 / s) was responsible for most of the aggradation between 1987 and 1990. Reach 9 was a major depositional zone during this period, although the actual channel change of 9 020 m 3 was significantly lower than it would have been if significant amounts of gravel had not been dredged from this reach. Downstream of this location transport 119 rates increased as a result of the 15 700 m^ of combined erosion in reaches 7 and 8. Transport rates declined at a fairly constant rate between reaches 5 and 1. Large volumes of material were dredged from these 5 reaches, which decreased the natural deposition rates to the considerably lower observed values. The gravel budget results show that there is considerable variation in the magnitude of gravel transport for the five periods. The channel change behaviour of a reach does not remain constant over time; it can occasionally deviate quite dramatically from a prior trend of aggradation or degradation. Nonetheless, it is possible to make some generalizations about the transport regime along the Vedder River: (i) reaches 1, 2 and 3 comprise a major depositional zone; (ii) reaches 4, 5, and 6 are a transport zone and show minimal channel change; (iii) transport behaviour varies considerably from year to year for reaches 7, 8 and 9, and (iv) reach 10 is a transport zone. The magnitude of morphologic change and gravel transport rates are dependent to some degree on flood behaviour during the given period. Patterns of deposition along the river are more obvious in this analysis than they were in the cumulative departure analyses. The stream power index was re-examined for each reach for comparison with the general transport regime given in the preceding paragraph (see figure 4.15 for stream power indices). The descriptions of transport for reaches 1, 2, 3, 4, 5, 6, and 10 are supported by the values of the stream power indices. This suggests that the stream power index used in this study provides a good method for quickly assessing the transport potential in a reach. The values of the stream power index for reaches 7, 8 and 9 are intermediate in magnitude relative to 120 the other reaches. Perhaps this can partly explain why these reaches show no distinct trends in their transport behaviour. 5.4. Error Analysis of the Gravel Budget Calculations An error analysis was performed in order to evaluate the uncertainty in the gravel budget calculations. Error in volume change estimates and grain size multipliers are the two major sources of uncertainty in the data. Dredging volumes were obtained from direct measurements of truck loads of excavated material. It is assumed that the errors associated with these values, because they were obtained by direct methods, are negligible in relation to other errors. However there is some uncertainty in these values associated with measurement procedures. Moreover, there may have been some uncontrolled dredging operations in the Vedder River. If these were not accounted for in the calculations then bias would be introduced to the results. Deviations between volume change estimates for two different sets of data were assessed for each period; in one case only odd-numbered cross-sections were used and in a second case only even-numbered cross-sections were used. It was assumed that the difference between the two volume change estimates for a reach represents the maximum error range. Based on these deviations and average cross-section spacings, four distinct groups were identified for analysis of volume change errors (see table 5.4). Reaches 1, 2, 4 and 5 exhibit minimal deviations and have a mean cross-section spacing of about 140 m. Despite a similar average cross-section spacing of 145 m, reaches 3 and 6 show considerably greater deviations. The deviations for reaches 7, 8 and 9 are very large, probably as a result of the low cross-section density. The average cross-section spacing in these reaches is about 241 m. 121 Table 5.4. Group Selection for Error Analysis:  Absolute Deviations Between Volume Change Estimates (all values are md) 1981-82  Group 1 (n = 20) 1 (143m)* 1 260 2 (138m) 3 460 4 (130m) 2 880 5 (147m) 2 200 Group 2 (n=10) 3 (141m) 5 570 6 (149m) 12 800 Group 3 (n=15) 7 (235m) 5 200 8 (255m) 16 000 9 (234m) 11 100 Group 4 (n = 5) 10 (216m) 8 190 1982-83 1983-84 1 320 6 900 4 530 2 540 5 960 7 560 9 050 4 990 18 000 3 730 59 400 4 290 16 200 43 300 8 640 23 100 3 940 2 280 7 210 1 460 1984-87 1987-90 6 370 11 000 11 000 2 280 5 730 138 1 020 9 680 392 4 180 14 500 16 600 1 950 41 300 33 800 6 120 3 920 29 300 1 810 9 240 * these values are reach lengths 122 Despite a mean cross-section spacing of 216 m, the deviations for reach 10 are relatively low. The root mean square averages of the deviations for the four groups were calculated (see table 5.5). These values were derived from the data in table 5.4. It is assumed that the error ranges for volume changes based on complete survey coverage (twice the number of cross-sections) are about one-half the magnitude of these deviations. Uncertainties in annual volume change estimates for reaches 1, 2, 4, 5, and 10 are about ± 1 100 m 3 . The error range is about ± 2 650 m 3 for reaches 3 and 6 and ± 3 350 m 3 for reaches 7, 8 and 9. Several sources of error are associated wi th the grain size multipliers; uncertainty regarding both the representativeness of sediment samples and the assumption that the grain size distribution along the river did not change over the study period. The standard error of estimate of the regression line for percentage of material > 2 m m is about 4%. Based on this information, an error of about ± 5% was chosen for the grain size multipliers. It is likely that the effects of some errors cancel others out in the gravel budget calculations. If there are compensating errors, an additive calculation of error propagation may exaggerate the level of uncertainty in the results (Taylor, 1982). The method for calculating the uncertainty of a function when there are compensating measurement errors w i l l now be discussed. The transport rate into a reach is obtained from the equation: n G n = L ( V ^ i + D i f i M l - p ) (5.1) i = 1 123 Table 5.5. Root Mean Square Average Deviations (all values are m3) Average Absolute  Deviations) Group 1 4 340 Estimated Average  Absolute Deviations  for Complete Survey  Coverage 2 170 Error Range + 1 100 Group 2 10 500 5 250 ± 2 650 Group 3 13 400 Group 4 4 960 6 700 2 480 ± 3 350 ± 1 250 124 where Vj is the volume change estimate for a reach, fj is the fraction of material > 2 mm, Dj is the volume of material dredged from the channel and n is the reach number for which the gravel input is being calculated. The uncertainty in the transport rate is calculated from the equation: n 3G n 2 9G n 2 E n = L SVi + 5f i (5.2) i=1 9Vi dti where 6VJ is the uncertainty of the volume change estimate and 6 fj is the uncertainty of the grain size multiplier. The probable range of each transport rate is: G b e s t e s t i m a t e ± E n (5.3) The results of the error analysis for the five gravel budgets are presented in figure 5.4 (refer to section 5.3) and table 5.6. The magnitudes of the error limits are approximately the same for all years. Because of the greater activity in the channel, in terms of morphologic change and gravel removal operations, the errors are slightly greater for the periods 1982-83 and 1987-90. The overall similarity of the values suggests that the error analysis is not very sensitive to the magnitudes of channel change and dredging volumes. This is because errors, in the end, depend on absolute errors in the volume change and grain size estimates. The degree of uncertainty in transport rates increases in an upstream direction as there is error propagation in the cumulative calculations. The error range of the transport 125 Table 5.6. Error Range of Best-Estimate Transport Rates (all values in md) Reach 1981-82 1982-83 1983-84 1984-87 1987-90 Mean 1 ± 5 5 8 ± 8 9 1 ± 5 7 3 ± 5 6 0 ± 6 8 3 ± 6 5 3 2 ± 7 9 3 ± 1 086 ± 8 3 2 ± 7 9 3 ± 9 3 0 ±887 3 ± 1 620 ± 2 000 ± 1 630 ± 1 590 ± 1 760 ± 1 720 4 ± 1 720 ± 2 080 ± 1 740 ± 1 700 ± 1 860 ± 1 820 5 ± 1 820 ± 2 180 ± 1 830 ± 1 800 ± 2 000 ± 1 930 6 ± 2 310 ± 2 870 ± 2 350 ± 2 300 ± 2 460 ± 2 460 7 ± 2 990 ± 3 580 ± 2 980 ± 2 930 ± 3 060 ± 3 110 8 ± 3 530 ± 4 020 ± 3 500 ± 3 460 ± 3 640 ± 3 630 9 ± 3 990 ± 4 470 ± 3 990 ± 3 930 ± 4 150 ± 4 110 10 ± 4 050 ± 4 530 ± 4 040 ± 3 990 ± 4 210 ± 4 170 126 rate into reach 10 is about 5 to 7 times greater than the error range of the transport rate into reach 1. However the degree of error is fairly low in comparison with the absolute and relative magnitudes of the best-estimate transport rates. The error ranges are considerably lower than the transport assumption adjustment of 5 880 m3/yr for 1981-82. This suggests that bias in the volume change estimates and grain size multipliers do not explain the negative transport rates found in the original gravel budget for this period. The transport assumption adjustment for 1984-87 is within the calculated error range. This indicates that the adjustment for this period was not significant. In order to evaluate the proportional errors associated with volume change estimates and grain size multipliers, the sum of the respective squared error components were compared (see table 5.7). For the periods 1981-82, 1983-84 and 1984-87 nearly all of the error in these terms is attributable to uncertainty in the volume change estimates. The relative contributions of errors in the grain size multipliers to total uncertainty are significantly greater for the periods 1982-83 and 1987-90, primarily because of the larger absolute magnitudes of volume changes. 5.5. Sediment Budgets for Size Classes The sediment budget provides a method for estimating transport rates of different size classes. Using results of the sediment size analysis, the estimates of volume change were adjusted for each size category. The size classes chosen for this study are 2-8 mm, 8-32 mm, 32-128 mm and 128-512 mm. These are consistent categories as each one covers a 2^ phi range interval. These categories also reflect natural separations in the size 127 Table 5.7. Percentage of Error Attributed to Volume Change Estimates  (Based on Squared Component of Error Propogation Equation) (all values in %) Reach 1981-82 1982-83 1983-84 1984-87 1987-90 1 100 39 95 99 67 2 100 53 91 100 73 3 96 63 94 99 82 4 96 66 94 98 83 5 96 67 95 98 80 6 98 63 94 98 86 7 95 66 96 99 91 8 95 73 97 99 89 9 96 76 96 99 89 10 96 77 96 99 89 128 distribution of the samples. The average proportions of bed material in the four size classes were calculated for all reaches (see table 5.8). The fraction of material greater than 128 mm decreases rapidly downstream of Vedder Crossing. About 17% of the bed material exceeds 128 mm in reach 10. There is no material greater than this size downstream of reach 4. Even though the largest quantity of bed material is between 32-128 mm for all study reaches, the proportion of material in this size class first increases and then decreases downstream of Vedder Crossing. As the fractions of bed material in the larger size categories decrease, the fractions of material in the 2-8 mm and 8-32 mm size classes increase. It was assumed that the proportions of material in each size class for each study reach remained the same over the study period 1981-1990. Because of this assumption it is necessary to examine a set of size-adjusted sediment budgets for only one period as the relative transport rates for the four size classes remain the same for each sediment budget. Size class sediment budgets were calculated for the 1982-83 data. The volume change estimates for the ten condensed reaches were multiplied by the fraction of material in each size class. Sediment budgets for the four size classes were then constructed. The results are presented in figure 5.5. As grain size category increases, there is a progressive increase in its fractional contribution to sediment transport up to 128 mm. The highest proportion of material transported in the Vedder River is between 32 and 128 mm. It also follows that differentials between reach input and output values are also greatest for this size category. Using the results of the size class sediment budgets and sediment size distribution analysis, the nature of sediment supply to the Vedder River was 129 Table 5.8. Percentages of Bed Material in Grouped Size Classes (all values in %) Reach 2-8 mm 8-32 mm 32-128 mm 128+ mm 1 16.4 27.7 33.9 0 2 15.8 26.8 36.2 0 3 15.3 25.9 38.5 0 4 14.7 25.0 40.5 0.255 5 14.3 24.2 40.5 2.33 6 13.7 23.2 40.2 5.14 7 13.0 22.1 39.8 8.31 8 12.4 21.1 39.4 11.2 9 11.8 20.1 39.0 14.2 10 11.2 19.1 38.6 17.3 SEDIMENT TRANSPORT ( m 3 / y r ) (THOUSANDS) 131 assessed. The sediment input into reach 10 represents the amount of sediment transported past Vedder Crossing. Although there is probably some error in these values as a result of the incorrect transport assumption, the degree of uncertainty is probably small relative to the magnitude of the input values. The highest proportion of sediment transported into the Vedder River was between 32 and 128 mm. The transport rate at Vedder Crossing for this size class was about 27 000 m3/yr. The transport rate for material greater than 128 mm was only 1 420 m3/yr. In order to assess the change in character of sediment at Vedder Crossing, a sample was taken from the reach immediately upstream of Vedder Crossing. The fraction of material < 4 mm (sand and fine gravel) is the same in both cases; about 16% of the bed material is smaller than 4 mm. Despite the similarity of fine material, the size distribution of the large material is very different for these two locations. The size of the coarsest material is much greater upstream of Vedder Crossing. The D 9 5 value is 323 mm for the upstream site and 236 mm for the downstream site. The D g 4 and D 5 0 values are also larger upstream of Vedder Crossing. These results suggest that the finer fractions of bed material are transported past Vedder Crossing. The decrease in the size of the largest material indicates that a large proportion of the coarse gravels transported along the Chilliwack River are deposited upstream of Vedder Crossing because of the decrease in the transporting capability of the flow at this location. 132 5.6. Longer-term Channel Changes and Gravel Transport In order to analyse long-term aggradation on the Vedder River, gravel budgets for gravel-sized material were constructed for the periods 1971-75 and 1976-81. McLean (1980) calculated the changes in area for cross-sections surveyed by the Water Survey of Canada in 1971 and 1975. These values were then multiplied by the appropriate distances to obtain volume change estimates for the ten condensed reaches. The survey coverage for the 1971-75 data was very limited. The mean values of cross-section spacing for these periods were 1 200 m and 1 100 m respectively. Therefore these data provide only a general indication of morphologic changes and gravel transport rates. There is a large degree of uncertainty regarding the accuracy of the results. The cross-section spacing was considerably lower for the 1975-1976 surveys; the average distance between survey lines was 320 m. The extent of survey coverage limited the calculation of volume changes to reaches 4 through 8. The construction of a sediment budget requires a knowledge of channel changes as far downstream as the canal entrance (reach 1). Cross-sectional changes in area between 1976 and 1981 were obtained by directly overlaying plots and using a digitizing routine. No control points were available to tie in the two surveys with each other. Therefore there is probably some error in area change estimates. The 1976 cross-section data were obtained from results of a survey conducted by the Ministry of the Environment dyking authorities after the 1976 gravel removal operation. Direct comparisons of the plots were possible as the same lines were surveyed in both cases. Data for cross-sections 40 to 49 are not available for 1976. This necessitates an assumption of net changes in area of 0 m 3 for these cross-sections. This assumption may be reasonable for cross-sections 45 to 49 133 as this reach consistently showed minimal change throughout the period 1981-90. There is a larger degree of uncertainty regarding the appropriateness of this assumption for reach 9. There was one minor sediment removal operation between 1976 and 1981. About 27 400 m 3 of material were dredged from reach 7 in 1980 (refer to table 3.2). The gravel budget results are presented in figure 5.6. The results for 1971-75 give a transport rate at Vedder Crossing of about 24 400 m3/yr. This value may be somewhat low as there were five floods which had discharges greater than 250 m3/s. Comparison with the gravel budgets constructed in section 5.2 suggests that the results for this period may be reasonable. The pattern shows a progressive decrease in transport rates along the length of the river. This same pattern is also found in the 1983-84 and 1984-87 results. Nonetheless there is probably a wide margin of error in these results because of the poor survey coverage. The 1976-81 gravel budget results show that this was period of relative inactivity on the Vedder River. There were five floods with discharges greater than the threshold for significant bed material transport. Two of these floods had high discharge values of 447 m3/s and 533 m3/s. The transport rate of 12 200 m3/yr at Vedder Crossing seems rather low when the occurrences of these 2 large floods are taken into consideration. A volume change of 0 m3/yr was assumed for reaches 9 and 10. It is possible that there actually were positive changes in these reaches, which means that the transport rates in this part of the river are underestimates. Furthermore the assumption of an even rate of change may result in an underestimation of the transport rate for the years of the two high magnitude floods and an overestimation of the transport rate for the other years. Unfortunately this issue cannot be Figure 5.6. Longer-term gravel budgets — 80-u 70 H « Q 50 H (—1O 2 2 KI Q H CO 1000 "I 1 1 1 1 r 2000 3000 4000 5000 6000 DISTANCE FROM X-SECTION 1 (m) l 1 1 1 1 7000 8000 9000 (a) 1971-75 T ^ I I I I I I I 1 1 1 1 1 1 1 1 0 1000 2000 3000 4000 5000 6000 7000 8000 DISTANCE FROM X-SECTION 1 (m) (b) 1976.-81 analysed any further as no survey data are available for any dates between 1976 and 1981. 136 6. The Effects of a Reduction in Survey Density on Volume Change and Gravel Budget Results It is a major project to conduct a complete survey of the Vedder River, which presently consists of 49 cross-sections. It would reduce the time and effort required to survey the river if the total number of survey lines could be reduced without significantly affecting volume change estimates. However cross-section density (number of cross-sections per channel length) must remain great enough to ensure adequate representation of channel modification. The required spacing between survey lines varies along a river according to the variability of channel change. Cross-section densities can be relatively low in reaches which exhibit a relatively uniform morphologic change. For any chosen spacing, it is critical that the nature of change at each survey line be characteristic of the channel length it represents. This chapter investigates whether volume change estimates for the condensed reaches are affected by a considerable decrease in the number of survey lines used in calculations. Volume changes for the 10 reaches were re-calculated using only half of the survey lines; in one case only odd-numbered cross-sections were used, and in a second case only even-numbered cross-sections were used. Survey coverage can be reduced in reaches for which the new estimates consistently show minimal deviation from the original results. The reduced results are plotted against the results based on complete survey coverage in figure 6.1a-j. The vertical distance between each datum point and the 1:1 line represents the deviation between the re-calculated estimate and the original estimate. The means and the standard deviations of these differences are given in table 6.1. The mean value indicates if the new estimates have an overall positive or negative bias. Because of the 137 Figure 6.1. Volume change comparisons: complete survey coverage vs. reduced survey coverge NET VOLUME CHANGE (1000 m 3/yr) COMPLETE SET OF X-SECTIONS (a) 1981-82: full vs. odd -20 -10 0 10 20 NET VOLUME CHANGE (1000 m / y r ) COMPLETE SET OF X-SECTIONS (b) 1981-82: full vs. even U I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -40 -20 0 20 40 NET VOLUME CHANGE (1000 m 3/yr) COMPLETE SET OF X-SECTIONS NET VOLUME CHANGE (1000 m 3 / y r ) COMPLETE SET OF X-SECTIONS (e) 1983-84: full vs. odd 30-20-| o o co o jz; -^2 w o 1<H 2 co < I K X O 25 s > o > 25 oH - I O H -20 Us I V i '" I ' i 20 ' - 1 0 ~> r~ 0 - i — i — i — i — i — i — i i i i 10 20 NET VOLUME CHANGE" (1000 m3/yr) COMPLETE SET OF X-SECTIONS 30 (f) 1983-84: full vs.even COMPLETE SET OF X-SECTIONS (g) 1984-87: full vs. odd NET VOLUME CHANGE (1000 m 3/yr) COMPLETE SET OF X-SECTIONS (h) 1984-87: full vs.even COMPLETE SET OF X-SECTIONS 142 Table 6.1. Deviations Between Best-estimate Volume Changes and Volume  Change Estimates Based on Reduced Survey Density * Full/Odd indicates deviations between volume changes based on full survey coverage and odd-numbered survey lines * Full/Even indicates deviations between volume changes based on full survey coverage and even-numbered survey lines (all values are m3) Survey Data 1981- 82 Full/Odd Full/Even 1982- 83 Full/Odd Full/Even 1983- 84 Full/Odd Full/Even 1984- 87 Full/Odd Full/Even 1987-90 Full/Odd Full/Even Mean Value -989 2 310 9 370 -266 4 080 -3 540 3 660 7 800 4 630 4 200 Standard Deviation 4 360 3 840 18400 8 590 7 200 7 100 6 730 9 470 18 700 23 900 143 considerable variation in cross-section spacing, some patterns of changes in area along the river may result in biases for the two subsets which have the same direction. The standard deviation is a measure of the spread of the data points. For the purpose of simplifying the terminology used in this discussion, the volume change estimates based on complete survey coverage will be referred to as expected values. Although these values are the best estimates that are available, there is undoubtedly some degree of error associated with them. The results for 1981-82 reveal that estimates of volume change based on odd-numbered cross-sections fluctuate about the correct values. There are no exceptional outliers in this data set. The mean deviation is about -989 m 3 ; the bias is probably within the precision of the estimates based on complete survey coverage. This suggests that there is negligible overall bias in these data. The standard deviation of the differences is approximately 4 360 m 3 . The estimates based on even-numbered cross-sections have a slight positive bias. The variability of these data points is similar to that found in the previous data set. In both of these cases, the estimates for reach 8 deviate from the correct values by the greatest amount. The 1981-82 results are a useful benchmark for comparison of the subsequent results. The volume change estimates based on odd-numbered cross-sections for 1982-83 have a large positive bias and variability. However most of the estimates are reasonably close to the correct values, with the exception of the datum point for reach 6; this estimate exceeds the correct value by 62 200 m 3 . The next largest deviation is 16 300 m 3 for reach 9. There is no overall bias in the results based on even-numbered cross-sections, although there is 144 notable spread in the data. The value for reach 6 overestimates channel change by only 2 840 m 3 in this case, whereas the deviation for reach 9 increases to 20 200 m 3 . The revised volume change estimates for 1983-84 in reaches 7 and 8 distort the true nature of morphologic change along this part of the river. The data points for these reaches have a large positive deviation for values based on odd-numbered survey lines and a large negative deviation for values based on even-numbered survey lines. The mean values and standard deviations indicate that there is a large bias and spread in these data sets. These statistics reflect the strong influence of the data for reaches 7 and 8; the estimates for the other reaches are fairly close to the correct values. There is a strong positive bias in the volume change estimates for both sets of revised data in 1984-87. Reaches 8 and 9 display the greatest discrepancy between estimates based on odd-numbered cross-sections and the correct values; -10 700 m 3 for reach 8 and 13 700 m 3 for reach 9. The estimates for all reaches, with the exception of reaches 7 and 8, have a positive bias. The estimates based on even-numbered cross-sections have a very strong tendency to overestimate channel change - the average bias is about 7 800 m 3 . The estimates for reaches 4, 6, 8 and 9 overestimate the correct value by 13 000 m 3 , 18 200 m 3 , 23 100 m 3 and 17 600 m 3 respectively. The results for 1987-90 based on odd-numbered survey lines have an overall positive bias. The standard deviation has a very high value of 18 700 m 3 , which indicates that the data points fluctuate significantly about their mean value. Despite the overall positive bias, the estimate for reach 8 shows the greatest deviation at -48 600 m 3 . The data for the even-numbered survey 145 lines also display a mean positive bias. However two of the three largest deviations are negative The values for reaches 7 and 8 are respectively -30 600 m 3 and -42 400 m 3 . The volume change for reach 9 exceeds the correct value by 45 700 m 3 . In order to determine the locations where the number of survey lines can be reduced, the reaches which have either positive or negative deviations exceeding 5 000 m 3 were compiled (see table 6.2). The same overall results were found when a criterion of 10 000 m 3 was used. When the criterion was decreased below 5 000 m 3 , it became difficult to detect any patterns in the data. Examination of this table reveals that the results for reaches 8 and 9 are most affected by the reduction in survey coverage. There is a considerable discrepancy between the revised estimates and the correct values in 8 of the 10 cases. The estimates for reaches 3, 6, and 7 are also notably altered by a reduction in the number of survey lines. The deviations of re-calculated estimates exceed ± 5 000 m 3 in about half of the cases. The values for reaches 1, 2, 4, 5 and 10 are least affected by the reduction in survey coverage. The deviation for these reaches exceeds ± 5 000 m 3 in less than half of the cases. It is possible that an examination of cross-section density may reveal information regarding losses of accuracy in the results. If the original cross-section density for a reach is low then there may already be significant variation in the values of cross-sectional area change. Any reduction in survey coverage can then significantly modify the values of volume change since it is doubtful that any one survey line can provide adequate representation of a very large channel length. Inspection of table 6.3 shows that there is no obvious correlation between cross-section spacing and the Table 6.2. Reaches with Deviations > ± 5 000 m 3 Period Full/Odd Full/Even 1981- 82 6 8 9 6 8 9 10 1982- 83 3 6 9 3 7 9 1983- 84 7 8 7 8 1984- 87 2 3 4 8 9 1 3 4 6 8 1987-90 1 2 3 4 6 1 2 3 4 7 8 9 10 7 8 9 j 147 Table 6.3. Mean Distance Between Cross-sections and Standard Deviations of  Cross-sectional changes in Area Reach Mean Distance Between Standard Deviation Cross-sections (m) of Cross-sectional Changes in Area (m )^ 1 143 14.1 2 138 ' 14.4 * * 3 141 23.0 4 130 12.2 5 147 11.8 * * 6 149 20.5 235 44.0 * *g 255 41.3 234 22.8 10 216 9.38 ** denotes reaches which have deviations > ± 5 000 for more than 50% of the cases (refer back to table 6.2) 148 reaches experiencing the greatest losses in information. Although the mean distance between survey lines is relatively high for reaches 7, 8 and 9, it is not exceptional for reaches 3 and 6. Regression analysis between mean cross-section spacing and variability in the values of cross-sectional area change gives an r^ value of 0.46, which suggests that there is no clear relation between these variables. Therefore it appears that cross-section density cannot alone account for the differences between the revised and the original data sets. It is possible that variability in channel change, and not cross-section density, may determine the nature of the revised estimates. The data for reaches 3, 6, 7, 8 and 9 have the highest standard deviations (refer to table 6.3); these are the same reaches for which volume change estimates are most affected by a reduction in survey coverage. This suggests that the variability of channel change within a reach is a major factor in determining the accuracy of estimates based on reduced survey coverage. Gravel budgets were constructed using data based on complete survey coverage for reaches 3, 6, 7, 8, and 9 and data based on reduced survey coverage for reaches 1, 2, 4, 5 and 10. The results for the revised and original gravel budgets, and the error ranges, are presented together in figure 5.4a-e. The reduction in the number of survey lines does not change the characteristics of the 1981-82 gravel budget; the overall pattern and magnitudes of transport rates are maintained. A maximum deviation between transport rates of 2 640 m 3 is found at the downstream boundary of reach 10. The overall pattern and approximate magnitudes of transport rates for 1982-83 are not significantly modified by the reduced survey coverage. 149 However there is a positive bias in the revised data that becomes progressively greater in the upstream direction; this is a direct result of the inflated volume changes for reaches 1, 2, 4, 5, and 10. Each overestimation is carried through the data as calculations are extended upstream. Therefore the maximum difference between transport rates is at the most upstream site. The new transport rate past Vedder Crossing exceeds the expected value by about 6 630 m 3 . Comparison of the gravel budgets for 1983-84 reveals that the reduction in the number of survey lines does not significantly alter the results. The relative differences among the transport rates and their approximate magnitudes are maintained. The maximum difference between transport rates at a location is 1 850 m3/yr. Overestimation of deposition volumes for the data based on partial survey coverage for 1984-87 results in inflated transport rates along the river. The deposition in reach 4 is overestimated by 6 800 m 3 and is primarily responsible for the large positive bias of the new transport rates. The re-calculated transport rate past Vedder Crossing exceeds the original value by 12 000 m 3 . This example illustrates the impact that one very exaggerated estimate of volume change can have on the gravel budget results. Despite variation in the magnitudes of transport rates for the period 1987-90, the overall pattern of change along the river is preserved. Decreases in the accuracy of re-calculated volume change estimates result in inflated transport rates. The discrepancy between transport rates is greatest at Vedder Crossing; the transport rate based on the decreased number of survey lines exceeds the correct value by 16 600 m3/yr. 150 It can be concluded that the relative differences among transport rates along the river are not significantly affected by the chosen modification of survey coverage. If it is not possible to obtain a representative portrayal of channel change for a reach using the decreased number of survey lines, then systematic error propagation occurs along the length of the river. Depending upon the severity of this problem, transport estimates can be modified quite dramatically. This phenomenon is found in the 1984-87 and 1987-90 gravel budgets, although it is most pronounced in the former case. Losses in accuracy for the periods 1981-82, 1982-83, and 1983-84 are within the error range of the gravel budget results. In order to determine if it is desirable to reduce survey coverage for a particular study, the potential loss in precision of the gravel budget results must be weighed against the decreased time and effort involved in surveying the Vedder River. The average cross-section spacing for reaches 1, 2,- 4, 5 and 10 is about 150 m. When only half of the survey lines were used in calculations, the loss in information was minimal. The average distance between cross-sections for reaches 7, 8 and 9 is about 240 m. However a decrease in cross-section density for these reaches resulted in considerably different volume change estimates. Based on this information, it is suggested that cross-section spacing should be between 250 and 300 m. The only exceptions to this occur in reaches 3 and 6. Despite relatively low average cross-section spacings of 141 m and 149 m, the volume change estimates were altered quite significantly when cross-section densities were reduced. 151 7. Conclusions There were two major components in this study: (i) the exploration of methodological issues related to the morphologic approach and (ii) the investigation of gravel transport and aggradation in the Vedder River. Although many of the recommendations made in this thesis are specific to the Vedder River, several key methodological issues were introduced. An attempt is made to generalize the conclusions so that they are of direct relevance to the application of this approach to other projects. The morphologic method provided a useful set of data for the evaluation of gravel transport and aggradation in the Vedder River. The assumptions, procedures and limitations of the morphologic approach to sediment transport analysis were investigated. Of particular interest is the transport assumption used in the sediment budget calculations. The construction of a sediment budget requires knowledge or the assumption of at least one transport rate or transport distance. If available, direct measurements can provide estimates of sediment transport rates. However in most cases a transport assumption must be made as measurements are not available for many rivers. It was initially thought that gravel transport into the Vedder Canal was negligible. However field investigation revealed that this was not the case, but there are no estimates of the actual transport rate into the canal. As no better transport assumption could be made, the zero transport assumption was retained, except when an adjustment was necessary. The amount of gravel transported into the canal appears typically to be in the order of one thousand m3/yr. The mean error ranges of the inputs into each reach varied between ± 6 5 3 m3/yr to ± 4 170 m3/yr. Therefore the error introduced as a result of the incorrect transport 152 assumption is quite small relative to errors in the volume change estimates and grain size multipliers. Even if the transport assumption is incorrect, as in this study, the relative transport rates along the river remain the same. The differences between transport rates for different periods are also preserved if there is reason to believe that the transport rate at the chosen location does not vary significantly over time. In the case of the Vedder River, the transport rates into the canal were probably minimal relative to the overall magnitudes of transport rates. This means that differences between transport rates over time and space were preserved. Furthermore, gravel transport rates into the canal were probably within the error range of the gravel budget results. Nonetheless, it is preferable to obtain the best possible estimates of absolute transport rates. Therefore, when designing a project, a study reach should be chosen which includes at least one location where a reasonable transport assumption can be made. This study provided recommendations for cross-section spacing on the Vedder River. Although the results of the analysis were restricted to the Vedder River, the methods used to obtain them can be followed for other projects. Two procedures for evaluating optimal cross-section densities are now discussed. Cross-sections can be surveyed at a relatively high density for several years. Volume change estimates are calculated and compared with estimates based on a reduced number of survey lines. Cross-section density is reduced in reaches which experience only minimal losses in accuracy when survey coverage is decreased. An alternative approach, which is less time-consuming but not so exact, is to compare the variations in cross-sectional area change for all reaches. Survey coverage is reduced in reaches which show the lowest variation. It was recommended that the average distance 153 between cross-sections along the Vedder River should be about 250-300 m. It would also be useful to have several additional cross-sections in the canal. As canal dimensions are regular, patterns of change are probably uniform. Therefore it would be necessary to put in only several cross-sections. The zero transport assumption could be made at the bridge at Site 1 (2 800 m downstream of the canal entrance) as there is negligible gravel downstream of this location. This would provide a better transport assumption when constructing future gravel budgets. There are several factors to consider when choosing the study reaches to be used in the construction of the sediment budget. If estimates of wash load are made then reach length should be no greater than the transfer length of the material so that there are no within reach transfers of material. Furthermore, the reach length should be no less than the transfer length or some of the bed material load is assigned to throughput. For optimum results, the length of reaches should approximate the average transfer distance of bed material during the time resolution of the morphologic study. This can be estimated by examining the distance between significant deposition sites. Erosion and deposition sites should be distinct within each reach for each period. It facilitates the identification of distinct zones of channel change if reaches are chosen so that they have reasonably uniform morphologies. In the present study, the average reach length used in the construction of the gravel budget was about 800 m. It is difficult to estimate the degree of uncertainty in the gravel budget results. There are errors associated with the volume change estimates and grain size multipliers. In this study errors in volume change estimates were responsible for most of the uncertainty in the final results. The results of 154 chapter 7 showed that varying degrees of bias were introduced into the results when different subsets of cross-sections were used in the calculations of volume change estimates. It was assumed that the results based on the full set of cross-sections provided the best representation of correct channel changes. However, it necessarily follows that there is also bias within these estimates. Error ranges of the volume change estimates for the ten reaches were determined by comparing the deviation between estimates based on two sets of survey lines. As each data set consisted of only half of the cross-sections, the maximum deviation for volume change estimates based on the complete set of cross-sections was obtained by dividing these results by two. It may be possible to obtain a better estimate of the error range by conducting a very detailed and careful survey of the river and calculating volume changes. These results could then be compared with the estimates calculated in this analysis to determine the error range. The uncertainty in the gravel budget results for the Vedder River is quite low compared to the relative differences in transport rates along the river and between study periods. McLean and Church (1986) assessed the uncertainty in direct measurements of sediment transport for the Fraser River. Annual loads, which were estimated from rating curves, were specified within ±20% for a one standard error confidence interval or within ±40% for a two standard error confidence interval. The uncertainty in the transport estimates at Vedder Crossing for the present study range from ±8% to ±25%. Leaving aside the possibility for bias, this indicates that the probable error associated with the morphologic approach is approximately the same magnitude as errors associated with direct measurements. 155 Sediment budgets were calculated for various size classes; this procedure was not performed in previous studies of the morphologic approach. Because the transport assumption in this study required a knowledge of the percentage of material greater than 2 mm, grain size distribution data were already available for this part of the analysis. The construction of size class sediment budgets provides a straightforward method for assessing the transport and deposition behaviours of various grain sizes. The application of the morphologic approach to the Vedder River reveals patterns of bed material transport and deposition over time and space. Hydrological analysis showed that in recent times a large flood has occurred about every three or four years. It is these floods that were responsible for most of the aggradation. Because of the variability in flood activity, annual transport rates varied significantly from one year to the next. Furthermore even among these significant floods, there appeared to be a large exponential increase in bed material transport with increasing discharge. The particularly exceptional floods of November 1989 and 1990 deposited large amounts of sediment in the river. During the period 1982-83 the river was very active, in terms of both deposition and erosion. Yet no exceptional floods occurred during this time, which makes it difficult to account for the high activity rates. It is possible that dredging volumes were overestimated or that cross-sections did not adequately represent true channel changes. Comparisons of long profiles showed that dredging volumes have reduced net deposition. Of particular interest was the comparison of the 1958 and 1990 long profiles which illustrated that gravel removal operations have been successful in limiting long-term aggradation. 156 Some generalizations can be made regarding patterns of aggradation along the Vedder River: (i) reaches 1, 2 and 3 are depositional zones, (ii) reaches 4, 5, 6 and 10 are transporting zones, and (iii) the behaviour of reaches 7, 8 and 9 varies considerably from one year to the next. This information is useful when planning flood control procedures. Because the ability of the river to contain flood flows decreases as bed levels rise, gravel removal should be focussed in reaches 1, 2 and 3. Depending upon the relative amounts of aggradation and degradation in reaches 7, 8 and 9 over time, some gravel removal may also be necessary to minimize flood hazard risks. Most of the sediment deposited in the Vedder River is supplied from bank erosion upstream of Vedder Crossing. Bank protection at the appropriate locations can provide a preventive measure against flooding hazards. The major limitations of this approach are the difficulties associated with: (i) obtaining or estimating a reasonable transport rate at a specific location, (ii) obtaining the best estimates of volume change, and (iii) evaluating the degree of uncertainty in the results. Further work should focus on examining these three issues. The accuracy of the sediment budget results depends upon obtaining the best known or assumed transport rate, and volume change estimates. Future studies should focus on identifying channel characteristics and locations for which it is possible to estimate transport rates reasonably. It is important that the best possible estimates of volume change be obtained within the constraints imposed by available time and effort. Volume changes can be estimated from cross-section survey data, bathymetric data, or aerial photography. Results based on these different methods should be compared to evaluate and contrast their abilities to 157 estimate volume change. If the results based on aerial photography are found to be reasonably accurate, then the potential for application of this approach to a wide variety of rivers is very great. Investigations should also attempt to provide guidelines for obtaining the most accurate volume change estimates for each method. This would involve investigating the effects on the results of varying cross-section density and scales of aerial photography. The morphologic approach provides a good method for evaluating the sediment transport regime of a river. The cross-section surveys provided data for analyses of long profiles, volume change estimates and gravel transport rates. The time scale of the morphologic approach is several years to several decades. This is coincident with the usual management time scale. The total field effort is very- much less than that required for direct measurements. Therefore the morphologic approach is a cost-effective method for evaluating sediment transport. This study has demonstrated the great potential of the morphologic approach and has shown it to be worthy of further investigation. 158 REFERENCES Armstrong, J. (1959) Surficial geology of the Sumas map area, Paper 59-9, Geological Survey of Canada, 27 pp. Blissenbach, E. (1952) Relation of surface angle distribution to particle size distribution on alluvial fans. Journ. Sed. Petrology, 22, 25-28. Church, M. and Ryder, J.M. (1972) Paraglacial sedimentation of fluvial processes conditioned by glaciation. Geol. Soc. Am. Bull., 83, 3059-3071. Church, M., Miles, M. and Rood, K. (1986) Sediment transfer along Mackenzie River: a feasibility study. Environment Canada, Inland Waters Directorate, Water Resources Branch, Sediment Survey Section, Report. Church, M., McLean, D.G., and Neill, CR. (1987) Sediment transport estimates form changing river morphology. Unpublished report. Church, M.A., McLean, D.G., and Wolcott, J.F. (1987) River bed gravels: sampling analysis. In Thorne, C , Bathurst, J.C., and Hey, R.D. (eds) Sediment Transport in Gravel-bed Rivers, John Wiley and Sons, Chichester, U.K., 43-88 Church M., Kellerhals, R., and Day T.J. (1989) Regional Clastic Yield in British Columbia. Can. Journ. Earth Sci., Vol. 26, 31-45. Church, M. and Slaymaker, O. (1989) Disequilibrium of Holocene sediment yield in glaciated British Columbia. Nature, Vol. 337, 452-454. Clague J., Saunders, I.R., and Roberts, M.C. (1988) Ice-free conditions in southwestern British Columbia at 16 000 years BP. Can. Journ. Earth Sci., Vol. 25, 938-941. de Vries, M. (1973) On measuring discharge and sediment transport in rivers. Int.Assoc.Hydraul.Reseearch, Seminar on Hydraulics of Alluvial Streams, New Delhi, 1-9. 159 Gomez, B. and Church, M. (1989) An assessment of bed load transport formulae for gravel bed rivers, Water Resources Research, Vol. 25, 1161-1186. Hamamori, A. (1962) A theoretical investigation on the fluctuations of bed load transport, Delft Hydraul. Lab. Rep. R4, Delft Hydraul. Lab, 14 pp. and 7 figs. Hickin, E.J. and Nanson, G.C. (1984) Lateral migration rates of river bends. Journ. ofHydr. Eng., 110, ASCE, 1557-1567. Hubbell, D. (1987) Bed load sampling and analysis. In Thorne, C , Bathurst, J.C., and Hey, R.D. (eds) Sediment Transport in Gravel-bed Rivers, John Wiley and Sons, Chichester, U.K., 89-118 Johnson, J.W. (1939) Discussion of laboratory investigation of flume traction and transportation, by Y.L. Chang, Trans.Am.Soc.Civ.Eng., 104, 1247-1313. Jordan, P. (1990) Hydrology of the November, 1989 Chilliwack River flood, and some observations on the impacts of forest management. Report prepared for B.C. Ministry of Forests, Chilliwack District, 28 pp. McLean, D.G. (1980) Flood control and sediment transport study of the Vedder River, M. A.Sc. thesis, University of British Columbia, Dept. of Civil Engineering, 185 pp. McLean, D.G. and Tassone, B. (1987) Discussion of bed load sampling and analysis by David Hubbell. In Thorne, C , Bathurst, J.C., and Hey, R.D. (eds) Sediment Transport in Gravel-bed Rivers, John Wiley and Sons, Chichester, U.K., 109-113 McLean, D.G. and Church, M. (1986) A re-examination of sediment transport observations in the lower Fraser River. University of British Columbia, Dept. of Geography, Fraser River Progress Report No. 2, 56 pp. McLean, D.G. (1990) Channel instability on lower Fraser Fraser River, Ph.D. thesis, University of British Columbia, Interdisciplinary Hydrology Program, 275 pp. 160 Nanson, G.C. and Hickin, E.J. (1986) A statistical analysis of bank erosion and channel migration in western Canada. Geol. Soc. Am. Bull., 97, 497-504. Neill, CR. (1971) River bed transport related to meander migration rates. Journ. of Waterways, Harbours and Coastal Engineering Division, 97, ASCE, 783-786. Neill, CR. (1983) Bank erosion vs. bed load transport in a gravel river, in River Meandering, ASCE Sp. Publ., Proc. of Conference Rivers '83, 204-211. Neill, CR. (1987) Sediment balance considerations linking long-term transport and channel processes. In Thorne, C , Bathurst, J . C , and Hey, R.D. (eds) Sediment Transport in Gravel-bed Rivers, John Wiley and Sons, Chichester, UK, 225-240. Popov, I.V. (1962) A sediment balance of river reaches and its use for the characteristics of the channel process. Soviet Hydrology, 3, 249-266. Popov, I.V. (1962) Application of morphological analysis to the evalutaion of the general channel deformations of the River Ob. Soviet Hydrology, 3, 267-324. Vanoni, V.A., Brooks, N.H., and Kennedy, J.F. (1961) Lecture notes on sediment transportation and channel stability, Rep. KH-R-1, W.M. Keck Lab. of Hydraul. and Water Resources., Calif. Inst. Technol., Pasadena, 121 pp. White, W.R., Milli, H., and Crabbe, A.D. (1973) Sediment transport; an appraisal of available methods. Rep. 119, U.K. Hydraulics Res. Stat., Wallingford, England. Taylor, J.R. (1982) An Introduction to Error Analysis, 70-74. Mill Valley: University Science Books, 271 pp. APPENDIX 1 HYDROGRAPHS: 1958-1990 1960 500 H \ 4 0 0 -6 0 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 166 600-n 500 H .400 O 300 « 200 100 " i— r 30 i 1 i 1 i ' i i i i i i — i — i — i — i — i — i — i — i — r 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 600 -i 1967 ~I i 1 i i i i i i — i — i — i — i — i — i — i — i — i — i — r 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 169 600 - i 1972 500 H \ 4 0 0 -g 300-05 200-100H — I — i — | — i — | — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r ~ 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 600 - i 1973 500 H \ 4 0 0 -g 300-co 200-100 H — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i i i i I 1 i 1 i 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 170 600 -i 500 .400 g 300 H co 200 100 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 600-, 1975 " i — i — r 30 60 "i 1 i i i i i i i — i — i — i — i — i — i — i — i — i — r 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 1976 500 H \ 4 0 0 -CO S 8 300 H 0-\ i i i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — | — 0 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 173 600 - i 1980 500 H .400-g 300-m 200 H Q IOOH -i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i i i i i |— 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 600-1981 500-.400-g 300-"> i 1 i 1 i i i i i — i — i — i — i — i — i — i — i — i — i — i — r 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 174 175 176 177 600 500 H .400H g 300 cq 200 100 H 1988 - i — i — i — i — i — i — i — i — i — i — i — i — i — I — i I i l i l i 1 i l 0 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 700 600 H 500 H • 400H o 300 H C O 200 H 100 H 1989 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 r 0 30 60 90 120 150 180 210 240 270 300 330 360 JULIAN DAY 178 APPENDIX 2 MEAN DAILY FLOWS > 250 m3/s n = the number of independent flood events greater than 250 m3/s * indicates a temporary decrease in flow levels between 2 periods which had discharges > 250 m3/s (the two peak floods are dependent events) m = the maximum mean daily discharge during a flood event which lasts more than 1 day Year Julian Day Discharge (m3/s) 1958 283 249 (n=l) 1959 119 282 m (n = 3) 120 274 328 253 1960 (n = 0) 1961 15 264 (n = 2) 154 261 155 254 156 249 157 254 158 249 * * 169 262 m 170 252 1962 3 263 (n=l) 1963 330 368 (n=l) Year Julian Day Discharge (m3/s) 1964 162 251 m (n=l) 163 249 1965 (n = 0) 1966 350 328 m (n=l) * 352 283 1967 170 258 (n = 2) 171 294 m 172 292 173 286 174 249 304 340 m 1968 19 265 (n = 3) 20 289 21 276 24 323 m 154 436 m 155 271 261 249 1969 144 271 (n=l) 1970 (n = 0) 1971 30 269 (n = 2) 31 357 m 133 249 Year Ju l ian Day Discharge (m 3/s) 1972 (n = 2) 142 252 143 253 * 149 244 150 326 151 348 152 314 153 262 * 158 261 159 297 160 309 161 360 m 162 348 163 300 164 249 194 292 195 311 m 1973 (n = 0) 1974 163 257 ( n = l ) 164 300 165 331 166 357 167 340 168 343 169 357 170 362 m 171 345 172 314 173 303 174 276 175 247 Year Julian Day Discharge (m3/s) 1975 153 265 (n = 3) * * 156 266 m 189 251 336 467 337 530 m 338 345 339 269 * * 342 261 343 297 1976 171 269 (n=l) 1977 18 289 (n = 2) 336 247 1978 311 249 (n=l) 1979 348 311 (n = l) * * 351 382 352 447 m 353 348 354 288 1980 361 533 m (n=l) 362 450 363 265 1981 (n = 0) 1982 169 280 m (n=l) 170 270 171 250 Year Julian Day Discharge (m3/s) 1983 10 287 (n = 2) 193 275 194 266 195 313 m 1984 4 524 m (n = 2) 5 337 6 277 181 256 1985 144 273 m (n = 2) 145 256 * * 155 * 250 •* 158 257 300 320 m * 306 256 ,307 258 1986 55 313 (n = 3) 56 344 m 151 253 152 256 m 327 349 m 328 292 1987 132 299 (n=l) 1988 134 250 (n=l) Year Julian Day Discharge (m3/s) 1989 313 629 (n = 2) 314 647 m 315 337 343 363 1990 277 314 (n = 2) 314 776 AWARDS NSERC 1967 Science and Engineering Postgraduate Scholarship, 1989-present University Gold Medal (Geography), 1989 National Council for Geographic Education Undergraduate Award, 1989 C e r t i f i c a t e of Merit for Academic Excellence, 1389 Royal Canadian Geographical Society Research Award, 1988 NSERC Undergraduate Student Research Award, 1988 The Wilhelmina and J . Gordon Mcintosh P r i z e , 1986-87, 1987-88 University of Western Ontario Senate Scholarship, 1987-88 Nomination for Faculty Scholarship, 1387-88 Social Science Students' Council Essay Award, 1987-88 University of Western Ontario 4-Year Continuing Scholarship, 1985-89 Dean's Honour L i s t , 1985-89 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098616/manifest

Comment

Related Items