Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sediment in forested and logged gullies, coastal British Columbia Millard, Thomas H. 1993

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

Item Metadata


831-ubc_1993_fall_millard_thomas.pdf [ 13.53MB ]
JSON: 831-1.0086273.json
JSON-LD: 831-1.0086273-ld.json
RDF/XML (Pretty): 831-1.0086273-rdf.xml
RDF/JSON: 831-1.0086273-rdf.json
Turtle: 831-1.0086273-turtle.txt
N-Triples: 831-1.0086273-rdf-ntriples.txt
Original Record: 831-1.0086273-source.json
Full Text

Full Text

SEDIMENT IN FORESTED AND LOGGED GULLIES, COASTAL BRITISH COLUMBIA by THOMAS HUGH MILLARD B.Sc.(Honours), The University of British Columbia, 1986  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 1993  ©  Thomas Hugh Millard, 1993  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.  (Signature)  Department of ^Geography The University of British Columbia Vancouver, Canada  Date^October 14,1993  DE-6 (2/88)  ii  ABSTRACT  This study examines sediment storage and transfers in gullies of coastal British Columbia, and how logging affects sediment storage and transfers. Both fluvial and debris flow transport of sediment occur in gullies, and the amount of fluvial transport of sediment which occurs will affect the magnitude of a subsequent debris flow. Coarse woody debris (CWD) may affect the storage and transfer of sediment in the gully channel, and logging can affect the supply and type of CWD.  To determine whether logging affects storage and transfer of sediment in gullies, sediment budgets were constructed for gullies in four treatment classes:  A. Logged, slash full, no recent debris flows: "slash-full (SF)" B. Logged, slash removed, no recent debris flows : "slash-clear (SC)" C. Logged, naturally scoured by debris flows: "torrented (T)" D. Unlogged, naturally loaded with CWD: "unlogged (U)"  Each sediment budget had input to the channel, storage in the channel, and output from the channel estimated. Significant differences between treatment types occurred, summarized below. Treatment classes grouped together in brackets did not have significant differences.  iii Budget  Greatest  Least  term Input  Torrented  Slash-full  Storage  Torrented  Output  Slash-clear (Torrented and Unlogged) Slash-full  (Unlogged and Slash-clear)  Unlogged  Slash-clear  One objective of the study was to assess the effectiveness and feasibility of cleaning slash from the gully channels. To be effective, cleaning slash must either reduce the magnitude of a debris flow in a treated gully, or else reduce the likelihood of initiation of a debris flow in the treated gully. Removal of slash will reduce the volume of a subsequent debris flow by about 15 percent, simply from the reduction in the amount of CWD. Reduction in sediment stored in the treated channel may reduce the volume of a debris flow by a further 4 percent. There is no evidence that removing slash will decrease the likelihood of initiation of a debris flow.  iv  TABLE OF CONTENTS Abstract ^  ii  Table of Contents ^  iv  List of Tables ^  vi  List of Figures ^  xiii  Acknowledgements ^  x  Chapter 1. Introduction ^  1  1.1 Thesis statement and objectives ^ 4 Chapter 2. Gully morphology and processes ^ 2.1 Gully morphology ^ 2.2 Sediment transport processes ^ 2.2.1 Headwall and sidewall processes ^ 2.2.2 Channel processes ^ Debris flow initiation ^ CWD and channel processes ^ 2.3 Summary ^  6 6 10 10 13 14 18 20  Chapter 3. Study Area ^ 3.1 Location and topography ^ 3.2 Bedrock and surficial geology ^ 3.3 Climate and hydrology ^ 3.4 Vegetation and forestry activities ^ 3.5 Description of gullies ^  21 21 24 25 29 32  Chapter 4. Study design and measurement program ^ 4.1 Study design ^ 4.2 Gully sediment budget estimates ^ 4.2.1 Input measurements ^ Surface erosion ^ Slump erosion ^ 4.2.2 Storage measurement ^ 4.2.3 Output measurement ^  39 39 42 46 46 48 48 49  Chapter 5. Results ^ 5.1 Input ^ 5.1.1 Surface erosion ^ Comparison of rainsplash pin response between gullies ^ Rainsplash pin erosion rates for individual gullies ^ Sidewall area for calculation of rainsplash erosion ^ Calculation of rainsplash input volume ^ 5.1.2 Input volume from slumping ^ 5.1.3 Total sediment input ^  53 58 58 61 66 66 69 71 73  V  5.2 Sediment storage in gully channels ^ 78 5.2.1 Errors in cross-section measurements ^ 80 5.2.2 Variation in storage change between 80 treatment groups ^ 5.2.3 Storage volume changes in channels ^ 84 87 5.3 Output ^ 5.3.1 Coarse sediment output ^ 87 Gully morphology scaling factor 88 Variation of sediment output between treatment groups ^ 91 Volume of coarse sediment output ^ 98 5.3.2 Fine sediment output ^ 100 100 Method ^ 102 Results ^ 5.3.3 Sediment transport distances ^ 112 5.3.4 Precipitation and sediment output ^ 115 115 Method ^ 118 Results ^ 125 5.4 Debris flows ^ 5.4.1 Coquitlam precipitation ^ 126 134 5.4.2 Cl debris flow ^ 136 ^ 5.4.3 C6 debris flow 139 5.5 Sediment budgets ^ 144 5.6 Summary ^ Chapter 6. Discussion and conclusion ^ 6.1 Budget accuracy ^ 6.1.1 Unlogged gullies ^ 6.1.2 Torrented gullies ^ 6.1.3 Slash-full gullies ^ 6.1.4 Slash-clear gullies ^ 6.1.5 Budget error summary ^ 6.2 Sediment storage and transfers in gullies ^ 6.2.1 Input ^ 6.2.2 Storage ^ 6.2.3 Output ^ 6.2.4 Fluvial and debris flow transport of sediment ^ 6.3 Transferability of results to larger gullies 6.4 Study results: implications for forest^management ^ 6.4.1 Slash-clearing effects on debris flows magnitude ^ 6.4.2 Slash-clearing effects on debris flow frequency ^ 6.4.3 Assessment of effectiveness of slash clearing ^ 6.5 Conclusion ^  149 149 149 150 153 153 154 155 155 158 163 164 166 168 168 170 171 172  References ^  176  Appendix  181  vi  LIST OF TABLES  Table 3.1  Gully dimensions ^  33  Table 3.2  Sidewall materials ^  35  Table 3.3  Channel materials ^  36  Table 5.1  Date of installations ^  55  Table 5.2  Rainsplash pin distribution statistics ^ 60  Table 5.3  Eroding rainsplash pin response ^  67  Table 5.4  Sidewall areas ^  68  Table 5.5  Volume of surface erosion ^  70  Table 5.6  Error in surface erosion volume ^  72  Table 5.7  Slump volumes for Coquitlam gullies ^ 73  Table 5.8  Total sediment input ^  Table 5.9  Sediment input by treatment groups ^ 77  Table 5.10  Storage volume changes in channels ^ 85  Table 5.11  Sidewall and channel factors ^  89  Table 5.12  Gully-scale-factor ^  90  Table 5.13  Scaled sediment output by time period ^ 92  Table 5.14  Volume of coarse sediment output ^  Table 5.15  Example of trap weight reconstruction ^ 102  Table 5.16  Sediment samples, gullies C3 and C5 ^ 103  Table 5.17  Fine sediment retained at trap, gullies C3 and C5 ^  108  Table 5.18  Volume of fine sediment output ^  111  Table 5.19  Comparison of C6 precipitation with Coquitlam Lake and Coquitlam River stations ^ 117  Table 5.20  Estimated C6 precipitation = coefficient * (River station) + constant ^  75  99  118  vii Table 5.21  Slope failures in GVWD basins, 1990-1991 ^ 126  Table 5.22  Maximum one-day precipitation totals and return periods, November, 1990 ^  128  Table 5.23  Antecedent temperatures and precipitation, November, 1990 ^  132  Table 5.24  Cl debris flow volume, November 23, 1990 ^ 136  Table 5.25  C6 debris flow volume, November 23, 1990 ^ 139  Table 5.26  Sediment budgets ^  141  Table 5.27  Total measurement error ^  143  Table 6.1  Old and revised input terms for sediment budgets. 156  Table 6.2  Old and revised treatment group means ^ 156  Table 6.3  Ranking of sediment budget terms by treatment group ^  175  viii  LIST OF FIGURES  Figure 2.1 Valley sideslope ^  7  Figure 2.2 Gully morphology ^  9  Figure 3.1 Location of study area ^  22  Figure 3.2 Cedar Creek study area ^  23  Figure 3.3 Annual precipitation, Coquitlam Lake and Coquitlam River stations, 1985-1992 ^ 26 Figure 3.4 Mean monthly precipitation, Coquitlam Lake and Coquitlam River stations, 1985-1992 ^ 26 Figure 3.5 Maximum daily precipitation, Coquitlam Lake station, 1924-1992 ^  28  Figure 3.6 Daily maximum precipitation and return period for Coquitlam Lake station, 1924-1992 ^ 28 Figure 3.7 Daily precipitation events greater than 80 mm, for Coquitlam Lake and Coquitlam River stations,1985-1992  30  Figure 3.8 Partial duration series, Coquitlam Lake and Coquitlam River stations, 1985-1992 ^ 31 Figure 3.9^Channel slash in Gully C6 ^  38  Figure 4.1 Conceptual model of sediment storage and transfers ^  41  Figure 4.2^Sediment input, storage, and output for a gully channel ^  44  Figure 4.3^Sediment traps ^  51  Figure 5.1 Filled sediment trap ^  56  Figure 5.2 Storm precipitation greater than 80 mm, Years 1, 2, and 3 ^  57  Figure 5.3 Histograms of rainsplash pins, Years 1, 2, and 3 ^  59  Figure 5.4^Histograms of rainsplash pins, by gully, Years 2 and 3 combined ^  62  Figure 5.5^Rainsplash pins, slope angle, and soil type ^ 65  ix  Figure 5.6 Channel margin slump in gully C2 ^ 74 Figure 5.7 Active and inactive cross-section zones ^ 79 Figure 5.8 Active zone width and area change ^ 82 ^ Figure 5.9 Cross-section area change, Years 2 and 3 ^ 83 Figure 5.10 Cumulative sediment output, unlogged, torrented, and slash-clear gullies ^ 93 Figure 5.11 Cumulative sediment output, torrented and unlogged gullies ^  96  Figure 5.12 Cumulative sediment output, torrented and slash-clear gullies ^  97  Figure 5.13 Sidewall grain-size distributions ^ 105 Figure 5.14 Sidewall and trap grain-size distributions ^ 107 Figure 5.15 Channel sediment transport distances ^ 113 Figure 5.16 Maximum storm precipitation and sediment output ^  119  Figure 5.17 Threshold precipitation required for sediment output ^  122  Figure 5.18 Fluvial and debris flow precipitation thresholds ^  124  Figure 5.19 Location of Coquitlam Basin slope failures ^ 127 Figure 5.20 Maximum precipitation, Coquitlam Lake and River stations, November, 1990 ^ 130 Figure 5.21 Debris flow causing precipitation ^ 133 Figure 5.22 Cl and C6 debris flows ^  135  Figure 5.23 Channel slash above debris flow zone, gully C6 ^  138  Figure 5.24 Budget error vs pooled error ^  145  Figure 6.1 Channel-margin storage zones in gully C5 ^ 152 ^ Figure 6.2 Sediment storage in gully C6 ^ 160 Figure 6.3 Old trees on bedrock in gully C2 ^ 162  x ACKNOWLEDGEMENTS  This thesis is the largest project I have ever undertaken; without the help of others, I would not have completed it. Many thanks to my supervisor, Dr. Michael Bovis, who was always available when I needed his help. His efforts ensured the success of the project. My second reader, and acting supervisor in my first year, was Dr. Michael Church. As always, he provided appropriate and useful guidance.  Funding was provided by the Fish/Forestry Interaction Program. I would like to thank Steve Chatwin for his organizational efforts, and Dan Hogan for all his help.  Field work for this project was usually strenuous, and frequently in unpleasant, and sometimes treacherous, weather conditions. Many thanks are due to Scott Babakaiff, Brenda Griffiths, Sue Young, Scott Davidson, John Matechuk, Marian Oden, Lars Uunila, and Craig Nistor. Not all the time was spent scrambling up and down hillsides; many interesting and useful discussions added to my knowledge and ideas for the project.  The Greater Vancouver Water District assisted in the project. I would particularly like to thank Derek Bonin for his help with project organization, as well as financial assistance for cleaning of the slash-clear gullies. Thanks also to the GVWD crew led by Roland Totsauer, and the crew of Hans Lee Timber.  xi  Brian Fast of B.C. Hydro provided data from the Coquitlam Lake and Coquitlam River stations.  Terry Cooper provided his artistic and linguistic talents when mine did not suffice; hence, I would like to thank him as well.  Finally, I would like to thank Brenda Griffiths, for all her help and support, as I thought out loud, to understand what I was doing, and why.  1  CHAPTER 1 INTRODUCTION AND OBJECTIVES  Gullies are an important component in the sediment transfer system in the mountainous coastal region of British Columbia, but their dynamics are not well understood. Hillslope failures which either enter into a gully or occur within a gully may develop into a debris flow (also known as debris torrent), which in turn may result in extensive damage to salmonid streams downstream of the gully. Sediment delivered to the gully between debris flow events may be stored or transported through the gully, which may affect subsequent debris flows. An important control of sediment storage and transport is the amount of coarse woody debris (CWD, also called large organic debris, LOD). Logging affects the amount and type of CWD in a gully, therefore, greater understanding of how CWD affects sediment storage and transport in gullies is required.  How trees and logging affect sediment storage and transfers in gullies are not well understood. Gullies combine features of both hillslopes and channels, hence their behavior can be complex. Logging in and around gullies removes the source of large trees, and at the same time increases the amount of smaller woody debris. Often cable yarding systems are located in line with gullies, to take advantage of the greater clearance between ground level and the cable. This can result in large amounts of slash (woody debris remaining after logs are removed) deposited in the gully. Increased volumes of slash  2  may be responsible for initiation of debris flows (Swanston and Swanson, 1976, Sauder and Wellburn, 1987). Sidewalls of gullies may be important sources of sediment, and are affected by both rooting strength reduction after logging, and yarding disturbance. Hence the amount of sediment delivered to a gully channel can be affected by logging, as can the movement of sediment as it interacts with the logging slash introduced into the channel.  The frequency of debris flows and other mass movements usually increases notably in steepland areas following logging. O'Loughlin (1972) reported a 2.3 fold increase in slope failures in clearcut areas (roads excluded), compared with similar unlogged terrain. Young (1992) measured slope failure rates for two areas in the Queen Charlotte Islands (QCI). Slope failures in clearcut areas increased 1.8 and 4.6 times the unlogged areas (roads excluded). Young found the increase in gully related failures to be greater than the increase in open slope failures. Rood (1984), in a much broader study, reported an overall increase in slope failures of 30 times in clearcut terrain compared to unlogged terrain in the QCI (again, roads excluded). The volume of material entering streams from debris flows increased by a factor of 69 when clearcut areas were compared with unlogged areas. The average size of debris flows entering streams more than doubled in clearcut areas, and the frequency of debris flows increased by a factor of 27 times. Debris flows accounted for 67% of the volume of sediment entering streams from clearcut areas.  3  Sediment delivered from gullies into higher order streams can have important effects on salmonid habitat. Salmonids require stable and clean gravel for spawning; once fry are hatched, they require a variety of habitats to survive both summer and winter stream conditions (Tripp and Poulin, 1986a, 1986b). A primary component of stream habitat is CWD, which provides diverse stream morphologies and cover for juvenile fish. Debris flows which originate on hillslopes, and can scour stream channels for long distances, can have severe impacts on both gravel and CWD in the stream. In some cases spawning gravel is almost completely removed with the debris flow; in others the proportion of fine sediment becomes deleterious to egg survival (Tripp and Poulin, 1986b). Debris flows reduce the amount of CWD in the stream channel (Tripp and Poulin, 1986a). Hogan (1986) found reduced pool-riffle spacings, altered pool and riffle heights, and smaller pieces of CWD in debris torrented streams. Streams subject to debris flows had poor salmonid egg and juvenile overwinter survival due to gravel scour and winter habitat loss (Tripp and Poulin, 1992). Thus debris flows are an important factor in determining channel morphology and fish habitat.  1.1 Thesis statement and objectives  Steep gullies of coastal British Columbia are important links in the sediment transfer system. Storage and transport of  4 sediment may be affected by the amount of CWD in the gully, which in turn is affected by logging. This study proposes to monitor the rates of sediment movement in gullies, and to examine how CWD and logging affect sediment storage and transfers.  The specific objectives of this study are: 1. To monitor sediment movement and storage in gullies, including logged, logged and debris torrented, and unlogged gullies. 2. To assess how logging, and logging slash, affect sediment movement and storage in gullies. 3. To assess the effects and feasibility of removing  logging slash from gullies.  Gullies are selected for treatment groups based on logging history, debris flow history, and the amount of CWD or slash in the gully. To determine the effect of slash within logged gullies, one treatment group will be composed of logged gullies with all slash removed after falling and yarding of timber is completed. How slash-clearing affects sediment storage and transport has not been investigated.  A sediment budget, composed of input, storage, and output, is constructed for each gully. Sediment budgets are useful since each term within the budget defines the sediment transfer into or out of a specific morphological component of the gully. Thus, specific effects of slash or slash-clearing can be  5  determined. In addition, the sediment budget provides a separate method of checking measurement errors, since measurement error will result in an unbalanced budget.  To determine differences between treatment groups, individual sediment budget terms will be compared. Specific hypotheses to be tested will be:  1. Sediment input is equal in all gully treatment groups. 2. Sediment storage is equal in all gully treatment groups. 3. Sediment output is equal in all gully treatment groups.  Results of this study should increase our understanding of how sediment is stored and transferred in gullies, and how logging affects gully behaviour.  6 CHAPTER 2. GULLY MORPHOLOGY AND PROCESSES  Gullies are composed of distinct morphological units which store sediment; within each unit, there exists a set of processes which mobilize and transport sediment. Since this study examines the storage and transport of sediment within gullies, it is necessary to define the morphological units of a gully, and to describe the important processes which occur within each unit.  2.1 Gully morphology  Gullies in British Columbia often occur on oversteepened sideslopes of U-shaped glacial valleys. The result is a long, linear, steeply sloping channel with few, if any tributaries (Figure 2.1). Horton (1945) defined 'unbranched fingertip tributaries' as first-order channels; later researchers extended the network to include zero order drainages or hollows (Dietrich and Dunne, 1978, Dietrich et al, 1987). Zero-order basins do not have channels, by definition, but are concave in shape, concentrate subsurface water and are subject to mass movements. In most cases gullies are first-order (or sometimes higher), but in others more closely resemble zero-order basins. This variation in form may be due in part to the amount of sediment discharged to the gully, and the relative amounts of surface and subsurface flow. Gullies are usually a few hundred  7  FIGURE 2.1 Gullies on a valley sideslope  Gullies visible on photo left. The cutblock boundary is located along a very deep gully incised into bedrock. A second gully, not as deep as the other, is located just to the right of the gully on the cutblock boundary. Note that each gully has few, if any, tributary basins.  8 metres in length, and from less than 1 hectare to several hectares in area.  Gullies can be divided into distinct morphological units (Figure 2.2): headwall, sidewalls, and channel; these units are generally easy to define in the field. Gullies originate at the headwall, generally a steep triangular failure plane scalloped into the hillside. Headwalls are generally steeper than 30 degrees. Beneath the headwall, the gully is composed of a channel and adjacent sidewalls. The sidewalls and channel generally form a U, V, or trapezoidal cross-section. Sidewalls  °  range from shallow, with slopes less than 15 , to steep, with  °  slopes greater than 50 . Channels range in slope from about 10 to over 35  °  °  The channel is the longitudinal axis of the gully, and is the lowest elevation on a cross-section perpendicular to the longitudinal axis. The channel slope is defined as the slope of the longitudinal axis of the gully. Although described as a channel, there may or may not be banks, fluvial sediment per se, or frequent surface flows of water. The channel both stores and transports sediment and CWD. Sediment from the sidewalls or headwall is delivered to the channel, then the sediment may be transported variable distances down the length of the channel, depending on channel slope and morphology, CWD, and water discharge.  9  FIGURE 2.2 Gully morphology  10  The sidewalls are bounded at their upper edge by the open hillslope and at their lower edge by the channel (Figure 2.2). The sidewall slope is defined as the angle from the channel edge to the top of the sidewall, orthogonal to the gully axis. Sidewalls do not concentrate as much water as the channel. As a result, sediment movement processes which occur on sidewalls are generally different from those in the channel. Some processes, such as debris flows, are primarily a channel phenomenon, but affect the sidewalls as well.  2.2 Sediment transport processes  The type of processes which occur in a gully will depend on the bedrock, type and thickness of glacial drift or colluvium, gully plan form, water discharge, and biological effects such as tree root strength and CWD. The morphology of the gully strongly controls the location of processes. The headwall and sidewalls are primarily the location for debris slides or debris avalanches, together with minor mass movements such as ravelling or small slumps. The channel is dominated by fluvial transport of sediment and debris flows.  2.2.1 Headwall and sidewall processes  Debris slides are the most significant process to which the headwall and sidewalls are subjected. Debris slides are shallow  11  planar movements of unconsolidated and usually unsaturated sediment (Varnes, 1978). Headwall debris slide volume averages 900 m 3 , and sidewall debris slides average 380 m 3 , in the Queen Charlotte Islands (Rood, 1984). Analysis of debris slide initiation usually adopts an infinite slope model for stability analysis. This method assumes a uniform thickness of soil with a well defined shear plane, and then analyses the slope for resisting forces (shear strength) and driving forces (shear stress). The ratio of shear strength to shear stress is called the Factor of Safety (F.S); failure occurs or is imminent when the F.S. equals 1. The equation can be expressed as:  F.S. = C a + (a f z cos 2 B - A )tan 0^(shear strength) a f z cosB sinB^  (shear stress)  "where C a = apparent cohesion (kPa), a f = unit weight of soil at field moisture (kN/m 3 ), z = vertical thickness of soil mantle (m), B = slope angle (degrees), 0 = internal angle of friction (degrees), [and] A = pore-water pressure at the failure surface (kPa)" (Sidle and Swanston, 1982).  Using the infinite slope model, slope stability decreases with a decrease in apparent cohesion, an increase in slope angle, an increase in pore pressure, and a decrease in the internal angle of friction. Sidewall and headwall slopes are sensitive to failure since they are often very steep, and a gully concentrates water, resulting in high pore pressures.  12 Logging can affect slope stability in several ways. Removal of trees immediately reduces soil weight and wind stress (Brown and Sheu, 1975). In forest soils, the apparent cohesion term includes rooting strength, which can be a significant factor in shear strength. O'Loughlin (1972) reports a root strength value of 71% of shear strength for saturated till soil on a slope of 35 degrees. Once trees are cut, roots rot and lose strength. As new vegetation grows, root strength increases. Reported periods in which soils are most susceptible to failure are variable. Sidle et al (1985) report a range of from 3 to 10 years after cutting. Rollerson (1992) considers the period from 6 to 15 years most susceptible. Falling and yarding of trees may disturb both surface and subsurface soil conditions (Sauder et al, 1987). After swales (zero order basins) in a northern California watershed were logged, macro-pore (pipe) discharge increased 3.7 times over  expected  discharge (Ziemer, 1992). In  addition, logging appears to contribute to pipe collapse and increased sediment discharge from these pipes. Interruption of macro-pore networks may result in locally increased pore pressures and subsequent slope failure (Ziemer, 1992).  Other processes common on sidewalls are dry ravel, rainsplash erosion, frost heave and small slumps. Dry ravel is the movement of individual clasts or small groups of clasts, (Sauder et al, 1987). Rainsplash erosion and frost heave are common on exposed ground, but surface vegetation prevents rainsplash erosion and appears to minimize frost heave. Small slumps, of less than a few cubic metres, are similar to debris  13 slides, but may have more complex failure surfaces, from planar to rotational.  2.2.2 Channel processes  Two processes are important within a gully channel: fluvial transport of sediment and debris flows. These two processes can act in sequence, with low magnitude, frequent fluvial events punctuated by high magnitude, infrequent debris flows. The proportion of sediment delivered to the channel that is transported out of the gully by fluvial transport is an important control on debris flow magnitude.  Both zero-order basins and first-order channels can exhibit a pattern of deposition of sediment over a long period of time followed by a complete evacuation of sediment by debris flow (Dietrich and Dunne, 1978, Benda and Dunne, 1987). A debris flow usually scours the channel to bedrock, after which a new cycle of sediment and CWD recharge occurs. As sediment enters a newly scoured channel, the rate of fluvial transport of sediment is greatest, since all water discharge is surface flow. As the depth of sediment increases in the channel, the proportion of subsurface flow increases, and less fluvial transport of sediment occurs (Dietrich and Dunne, 1978; Bovis and Dagg, 1987).  The rate of sediment recharge will depend upon the supply of  14 sediment to the channel, the proportion of material too coarse to be transported by fluvial events, and the amount of sediment trapped by CWD. Sediment sources for gullies can have extremely heterogeneous grain sizes, including large boulders. Sediment supplied to the channel may be subject to selective fluvial transport of finer material, both on the surface and in void spaces between large grains (Bovis and Dagg, 1988). The result is a coarse lag deposit, with much finer sediment subsurface. Debris flow initiation  Debris flows can be defined as a gravitational movement of solids with interstitial fluid, where the relative velocities between fluid and solid are not significant (Takahashi, 1981). Debris torrents are a type of debris flow, where "rapid movement of water charged soil, rock and organic material [flows] down steep stream channels" (Swanston and Swanson, 1976). The term debris torrent is established in the logging and slope movement literature, at least in the Western Cordillera, but there is no mechanical difference between debris flows and debris torrents (Church and Miles, 1987). Accordingly, the more general term "debris flow" will be used here.  Debris flows are complex events in that their initiation and movement depend on a large number of geologic and climatic factors. Takahashi (1981) cites three main causes of debris  15 flow: 1) a landslide enters a channel, becomes saturated and turns into a debris flow, 2) a natural dam blocks a gully channel, eventually collapsing to release a debris flow, and 3) channel sediment becomes unstable and decouples from the streambed as a debris flow when sufficient water discharge occurs. Takahashi states that mobilization of channel debris is the most common mechanism of debris flow initiation, but evidence indicates that the first or second cause is probably much more common in coastal British Columbia. Rood (1990) reported the location for debris flow initiation sites in the Queen Charlotte Islands (QCI). In forested terrain, 57% originated from headwall failures, 15% from sidewall failures, 14% from open slopes adjacent to the gully, and 14% from unknown sources - unknown largely because of dense forest cover (Rood, 1992). In clear-cut terrain, 65% originated as headwall failures, 16% from sidewall failures, and 12% from locations outside the gully. A study of landslides on southwest Vancouver Island shows less than 2% of debris flows in gullies initiated in the channel (Rollerson, 1984, and 1993). A study from coastal Oregon reported debris flows originate in zero-order basins, not in first-order channels (Benda, 1990).  Not all slope failures which enter a gully necessarily produce a debris flow. Rood (1990) reports that 1 in 3.3 failures (forested terrain) and 1 in 3.2 failures (clear-cut terrain) which enter a gully produce debris flows. Rood (1990) observed a much higher ratio of debris flows to slides if only upperslope locations are considered. Failures originating from  16 hollows located at the channel head, and entering the channel at an angle less than 45 degrees, initiate debris flows; failures from hollows entering the lower portions of the channel at a 90 degree angle do not initiate debris flows (Benda and Dunne, 1987). These observations suggest that basin geometry is an important control on debris flow initiation. Other factors affecting debris flow initiation are the size of the failure and the force with which it impacts channel debris, the slope angle, and water discharge in the gully channel (Bovis and Dagg, 1992).  All debris flows require a source of water; this is most commonly provided by rainfall or snowmelt, or a combination of both. Caine (1980) defined a minimum rainfall intensity curve for shallow landslides and debris flows, using published data from many areas of the world. The threshold for slope failure is fairly well defined for periods from 1 hour to about 5 days; the minimum value for 24 hours is 100 mm. Innes (1983) developed a similar curve, limiting his data set specifically to debris flows; his minimum value is about 25 mm of rain in 24 hours. Two curves for slope failures on the Queen Charlotte Islands have been developed, for wet antecedent conditions and dry antecedent conditions. The minimum 24-hour precipitation value for wet antecedent conditions is 20 mm; for dry antecedent conditions the minimum value is 30 mm in 24 hours (Hogan and Schwab, 1991a). It should be emphasized that all these values are minima; in some cases rainfall which produced slope failures was much greater. In addition, my choice of  17 reporting 24 hour precipitations is somewhat arbitrary, since different intensities and durations of rainfall are known to initiate debris flows.  Church and Miles (1987) emphasize the importance of parameters difficult to measure: locally intense precipitation and snowmelt. The steep mountain fronts and narrow valleys of southwest British Columbia provide effective barriers for air mass movements, forcing the air aloft. Convection cells may result, with subsequent heavy precipitation, particularly at higher elevations. For example, yearly precipitation at Hollyburn Ridge (elevation 951 m) is more than double that at the nearby Point Atkinson station (elevation 9 m). Hence the sparse precipitation gauge network, usually located in valleys, may not adequately measure precipitation which actually occurs at debris flow sites. Snowmelt in response to rising temperatures and warm rain may provide significant inputs of water to soils. The amount of snow present in a basin can vary with elevation, hence estimating water input from snowmelt is very difficult, even if snow is known to be present. Church and Miles summarize local debris flows being generated by: "(1) locally concentrated rainfall, high antecedent moisture, no snowmelt...(2) widespread moderate rainfall and snowmelt...(3) heavy rain onto thawing ground with little snowmelt...(4) apparently unremarkable rain, rain on snow, or snowmelt". Although "apparently unremarkable" water inputs may produce debris flows, very remarkable events are more likely to produce debris flows.  18 CWD and channel processes  The channel is the location where CWD accumulates. Natural sources of CWD are trees or tree fragments, usually derived from windthrow or sidewall failures. Effects of  CWD  in steep  channels have not been well investigated. Energy dissipation and sediment trapping are two  CWD  functions observed  (Froehlich, 1973). In stream channels,  CWD  significantly  affects morphology. Pool and riffle spacing and size, bank features, and sediment storage are all affected by  CWD  (Hogan,  1986). Log jams, typically from debris flows, are able to trap large volumes of sediment, and control channel morphology both upstream and downstream for significant distances (Hogan and Schwab, 1991b).  CWD  in gullies may act in different ways from  stream channels. Since gully walls are steep and narrow, tall trees falling into a gully will probably be suspended over the channel on at least one end, unless the tree happens to fall parallel to the channel axis. Smaller fragments, such as broken trunks, stumps, or branches, are more likely to lie across the channel, an effective position for trapping sediment. In addition to their role in supplying  CWD,  also act as barriers to sediment or  CWD  standing trees may  movement.  Logging will reduce the long-term supply of large  CWD  available  for a gully, but at the same time will dramatically increase the amount of smaller CWD introduced to the gully. Slash measured in three headwater channels in western Oregon (average size, 70 hectares) averaged 0.4 m 3 /m of channel (Froehlich,  19 1973). Total CWD in these headwater channels increased 2.5 fold after falling and yarding; fine CWD (0.3-10 cm diameter) increased 4.5 fold. Clearly, logging increases the total volume of CWD, and in particular, increases the amount of smaller CWD. Although these results are for only one area, similar results can be expected in forests of coastal British Columbia, since tree species and logging methods are similar.  Forestry management literature often recommends minimizing the introduction of slash to gully channels, or removal of slash after yarding (Froehlich, 1973; Chatwin, 1991; British Columbia Ministry of Forests, 1992). Channel slash has been cited as a cause of debris flows, primarily from debris jams which fail during high flows and initiate a debris flow (Swanston and Swanson, 1976; Krag et al, 1986; Swanston and Howes, 1991). Only Swanston and Swanson (1976) observed debris flows in which CWD contributed to the initiation of the event; however, how slash affected the initiation was not stated.  2.3 Summary  Gullies act as both a source and conduit for sediment transport and are an important link in the sediment transfer system. The headwall and sidewalls are source areas for sediment delivered to the channel. Sediment introduced to the channel may be stored temporarily, and can be subject to frequent fluvial events and less frequent debris flows. CWD affects storage of  20 sediment, and hence may influence the magnitude and frequency of debris flows. Logging changes the type and volume of CWD in the channel, thus it may also affect the timing and volume of debris flows. Hillslope failures and debris flows usually occur during times of high rainfall or snowmelt (or combination), but no simple relationship between debris flow occurrence and precipitation is observed.  21 CHAPTER 3. STUDY AREA  3.1 Location and topography  The study area is located in Coast Mountains approximately 30 km northeast of Vancouver (Figure 3.1), within Coquitlam Watershed. The Watershed is managed by the Greater Vancouver Water District to supply water to metropolitan Vancouver. Coquitlam Basin is a long, narrow valley trending north-south; Cedar Creek is a major tributary on the east side of the basin, and trends NE-SW. The study site is on the NW side of Cedar Creek, on a ridge which ranges in elevation from about 600 m to more than 1200 m (Figure 3.2). The Branch 200 road is located along the bottom of the hillslope; Branch 230 is a switchback spur road, and reaches mid-slope positions.  Bedrock is close to the surface on the steep middle and upper portions of the hillslope. Numerous gullies are incised into the hillslope. Lower portions of the hillslope have colluvial fan or cone shaped debris deposits emanating from many of these gullies. At the apex of the fan, the gully is usually well expressed whereas lower portions of the fan may show little, if any, expression of a gully or channel. In most cases, a clearly defined channel does not extend from the hillslope to the channel of Cedar Creek. The colluvial fans are therefore influent flow zones in that water discharge is often not  22  FIGURE 3.1 Location of study area  Cl  Gully  ^ Legend  ^Cutblock Logged, 19_ L 89  Contours (200 foot interval) Stream (surface flow) Stream (subsurface flow) FIGURE 3,2 Cedar Creek study area  Roads 0  ^  metres^1000  24 evident in lower sections of a gully, but is visible well above its fan.  3.2 Bedrock and surficial geology  Bedrock in Cedar Creek is mostly gabbro, with some quartz diorite areas and small sulfide exposures (Roddick, 1965). Bedrock is generally massive and resistant to erosion; however, within recently torrented gullies bands of finely fractured bedrock trending NE-SW are evident. Faults visible on aerial photographs show a similar trend.  Basal till overlies bedrock in much of the study area. The till is consolidated and dense; the Branch 230 roadcut shows almost vertical till exposures over 5 m high. In other areas bedrock is exposed at the surface, so depths of till are not uniform. Grain size range is considerable within the till, as expected. Many boulders, some with intermediate diameter > 1 m, are visible within the roadcut exposures.  Colluvial material often overlies the bedrock or till. The source of this material is either bedrock or till delivered from upslope, or else in situ weathering of the till. Depths are typically less than 2 m, except in the fans at the base of gullies, where depths may be much greater.  25 3.3 Climate and hydrology  The south Coast Mountains have a cool temperate climate, but are strongly affected by orography. In Vancouver, near sealevel, mean annual precipitation is about 1400 mm. On the North Shore Mountains, 3500 mm of precipitation is recorded annually at elevations of 1200 m. Low elevations within the mountains may receive comparable precipitation totals, as a result of topographic confinement of air. In winter the freezing level is often between sea-level and the mountains tops. Snow falls about 80 days of the year at the crest of North Shore Mountains, and in Vancouver about 15 days per year (Hay and Oke, 1973; Wright and Trenholm, 1969). These ranges in precipitation totals, temperature and snowfall are similar to those experienced in nearby Coquitlam Basin.  A weather station has operated at the Coquitlam Lake Dam since 1924. The station operated as an Atmospheric Environment Service station until 1982. Since then, B.C. Hydro has operated the station, collecting hourly information, although the data are incomplete until 1985. The mean annual precipitation for this station from 1924 to 1992 was 3490 mm. B.C. Hydro also established a second weather station within the valley, on Coquitlam River above Coquitlam Lake (Figure 3.1). It has been operating since 1984, with essentially complete data since 1985. The total annual precipitation from both stations for the period 1985-1992 is shown in Figure 3.3. Monthly mean precipitation is shown in Figure 3.4. The Coquitlam River  26  FIGURE 3.3 Annual precipitation Coquitalm Basin stations, 1985-1992 4000 3500 3000  -  -  Z 25002000 H 0  tat  1500 1000 500 0  -  ^ 1985 1986 1987 1988 1989 1990 1991 1992 AVG YEAR Lake station River station Elev. = 161 m Elev. = 280 m  go  AL  A:  FIGURE 3.4 Mean monthly precipitation Coquitlam Basin stations, 1985-1992 600 500 400 300 200 100  -  -  -  -  I14  •  •  I  •  • •• .A JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC MONTH Lake station V V ■■• River station 14  Elev. 161 m  04  ^Elev 260 m  27 station has the greater total precipitation in all years, as well as consistently higher monthly means. This may be related to elevation, since the River station is at an elevation of 280 m, compared to the Lake station at 161 m. Alternatively, constriction of air as the valley narrows in its headwaters may account for the greater precipitation at the River station.  Since gully sediment transport is associated mainly with short periods of intense precipitation, the frequency of large precipitation events is of interest. The maximum daily precipitation for Coquitlam Lake station for 1924-1992 is shown in Figure 3.5. Figure 3.6 shows the return period and magnitude relationship for these data. The largest event occurred in 1989; the second largest in 1990, the year this study began. The maximum event in 1991 is the same as the mean for the period, 135 mm, and the maximum daily precipitation for 1992 is 116, lower than the mean annual maximum. Therefore, 1990 was an exceptional year, and 1991 and 1992 were average and below average, respectively.  Large precipitation events may occur more than once a year; so partial-duration series (De Ploey et al, 1991) have been constructed for both B.C. Hydro stations for the period of 1985-1992. The minimum value chosen for the partial duration series is based on a 24-hour maximum precipitation of about 90 mm, which can be identified as a threshold for sediment transport (Section 5.3.4). Since the partial-duration series are constructed using daily totals, rather than 24 hour maxima,  28  FIGURE 3.5 Maximum daily precipitation Coquitlam Lake station, 1924-1992  I^I^I  1920  1930  1940  1950  1960 YEAR  1970^1980^1990  2000  FIGURE 3.6 Maximum daily precipitation and return period Coquitlam Lake station, 1924-1992  1  ^  ^ 10 RETURN PERIOD (years)  100  29 a correction factor of 1.13 is applied to the maximum 24 hour precipitation (Linsley et al. 1975), resulting in a total daily value of about 80 mm. Results for the 2 stations, grouped by year, are shown in Figure 3.7. Figure 3.8 shows the partial duration series for both stations, based on the same data sets as Figure 3.7.  Results are similar for the two stations, although the Coquitlam River station has more frequent large events. On average, 3.1 significant events (greater than 80 mm) in a year can be expected at the Lake station, and 5.7 significant events at the River station. The 1989 and 1990 events at the Lake station exceed all other events at both stations.  Large events can occur at any time of the year, but autumn and winter account for the majority of large events. November has the greatest number of significant events (44% of Lake station events, 24% of River station events), and October to February has 81% of the events at both stations combined.  3.4 Vegetation and forestry activities  The study area lies within the Submontane Wetter Maritime Coastal Western Hemlock Biogeoclimatic zone (GVWD Forest cover maps, Slaymaker et al, 1992). Dominant tree species within the study area are western hemlock (Tsuga heterophylla), western redcedar (Thuja plicata), grand fir (Abies grandis) and Douglas  • Figure 3.7 Daily precipitation events greater than 80 mm, 1985-1992  30  3.7a Coquitlam Lake station  3.7b Coquitlam River station 280 260 — 240 220  O 200 •  180  I-4  160  fal  -  140 •  120 l0O  El= —^  80 1985  ••••.  1986^1987^1988^1989 YEAR  1990^1991^1992  ^  31  FIGURE 3.8 Partial duration series for daily precipitation Coquitlam Basin stations, 1985-1992 ^280 ^ 260^ 240 0 2202001801601401 2010080  0  t^I^i^l^t^t^t^1^t^1^till 10 1^ RETURN PERIOD (years)  X Lake station^+ River station  I  t^t^1^Itti 100  32 fir (Pseudotsuga menziesii). Higher elevations have mountain hemlock (Tsuga mertensiana).  Logging in the study area began in 1977. Figure 3.2 indicates dates when individual patches were logged. High-lead yarding was used in all of the earlier logging. In the most recently logged area, where gullies C10 and C11 are located, a Wyssen skyline system was used. All areas are replanted within two years of logging. Regeneration is well established in almost all areas, except the most recently logged areas, and in some gullies which have had slope failures or debris flows.  3.5 Description of gullies  The locations of gullies are shown in Figure 3.2. Gullies vary in length, primarily dependent upon how close to the top of the ridge they begin. Since gully headwalls are located in forested terrain, the location of the exact start of each gully is not usually observed on aerial photographs. Some gullies were surveyed to the top of their headwalls, but this was not possible in all gullies. As a result, the highest reaches of a gully may not be accurately represented on the map.  Gullies were surveyed for channel and sidewall slope, sidewall length, channel material, and sidewall vegetation cover. Table 3.1 summarizes the spatial dimensions of each gully. Where possible, channel length is the surveyed distance from the  33 Table 3.1 Gully dimensions  Gully^Channel Mean^Mean^Mean^Total and^Length^Channel Sidewall Sidewall ^Area 2 Type 1^Slope^Length^Slope (n)  (degrees)  (m)^(degrees)  (ha)  C1-U^330s  28.5  7.4^16.4  0.30  C2-U^350e  27.2  9.6^26.4  0.59  C3-T^900m  19.9  10.5^23.8  1.63  C5-T^500m  24.1  17.0^38.1  1.13  C6-SF,T^260e  32.0  7.2^28.1  0.30  C4-SF^140s  29.3  12.0^19.5  0.29  C8-SF^195s  33.7  6.4^20.8  0.21  C10-SC^250m  30.5  6.4^28.9  0.24  CU-SC^500m  29.9  7.6^31.6  0.64  1 U: unlogged; T: torrented; SF: slash-full; and SC:  slash-  clear. 2 Gully area approximate, measured by average width (top sidewall to top sidewall) and channel length. s Surveyed distance. e Surveyed distance, plus estimated distance. m Distance measured from Figure 3.2.  34 sediment trap to the headwall of the gully. Some gullies extend almost to the top of the ridge and their headwalls are visible in aerial photographs. Channel lengths for these gullies are measured from Figure 3.2. If total channel length was not surveyed, and location of the gully headwall is not visible in aerial photographs, an estimate of total length, based on surveyed distance, is used. Gully C3 has two gullies join together; channel length for C3 is the combined length. The torrented gullies C3 and C5 are the largest gullies selected; the slash-full and slash-clear gullies are generally the smallest. The largest gullies tend to have the lowest channel slopes, but sidewall slope is not associated with gully size. Sediment traps for gullies C3 and C5 are both located on the fan; this is one reason why the larger gullies have lower channel slopes.  Tables 3.2 and 3.3 summarize the sidewall and channel materials in each gully. Gully cross-profiles were surveyed at 10-30 m distances along each longitudinal profile. At each station, the sidewall and channel materials were noted. Materials noted were classified as bedrock, bare sediment, vegetation, and slash. Sidewall sediment is generally very heterogeneous, with grain sizes ranging from clay to boulders. Channel sediment is typically cobbles and boulders, with some finer sediment. Proportions of each material at each station were estimated at 100%, 85%, 70%, 50%, 30%, 15%, or 0%. The amount of sidewall or channel covered by each material is the cumulative proportion estimated at each station.  35 Table 3.2 Sidewall materials  Gully  Bedrock^Sediment  Vegetation  Slash 1  Type  (%)^(%)  Cl-U  0^0  100  0  C2-U  0^2  98  0  C3-T  0^19  71  10  C5-T 2  0^58  42  0  44^33  15  8  C4-SF  0^6  30  64  C6-SF  0^16  0  84  C8-SF  0^10  12  78  C10-SC  3^0  3  94  C11-SC  0^2  33  65  C6-T  (%)  (%)  1 Slash includes naturally introduced CWD. 2 Vegetation includes seeded grasses and legumes.  36 Table 3.3 Channel materials.  Gully  Bedrock^Sediment  Vegetation  Slashl  Type  (96)  Cl-U  0  89  0  11  C2-U  52  42  0  6  C3-T  0  97  0  3  C5-T  0  100  0  0  C6-T  90  5  5  0  C4-SF  0  0  0  100  C6-SF  33  2  0  65  C8-SF  0  8  2  90  C10-SC 2  78  22  0  0  C11-SC?  0  83  0  17  (%)  (%)  (%)  1 Slash includes naturally introduced CWD. 2 Surveyed in May, 1992; slash clearing occurred the previous autumn.  37 Sidewall materials vary according to gully group. The unlogged gullies have almost completely vegetated sidewalls. Slash-full and slash-clear sidewalls are mostly covered in slash, with some vegetation and smaller bare areas. Torrented gullies vary in the nature of their sidewall materials. C3 was logged in 1977, and since then, growth of planted conifers has vegetated most of the sidewalls. C5 is partly covered in seeded grasses, but also has bare soil. Gully C6 has mostly bedrock and bare soil, but C6 is located higher on the hillslope than either C3 or C5; consequently bedrock is closer to the surface. The surveyed portions of C3 and C5 are in colluvial fans, where bedrock is not close to the surface.  Channel materials are similar in all gullies except for slashfull gullies. Unlogged, torrented, and slash-clear channels are either bedrock, sediment, or a combination of both. The gully C10 channel was primarily bedrock when surveyed, however after slash-clearing in the previous autumn, the channel was mostly sediment. Slash-full channels are clearly dominated by slash (Figure 3.9), although some bedrock or sediment may be present.  38 Figure 3.9^Channel slash in Gully C6  Vertical view of the gully. Stadia rod visible in center bottom of photo; numbers on rod are in decimeters. Photo scale changes drastically from top to bottom.  39 CHAPTER 4. STUDY DESIGN AND MEASUREMENT PROGRAM  4.1 Study design  The most general objective of this study is to understand how sediment and CWD are stored and mobilized in gullies. The more specific objective of this study is to understand how logging slash affects the storage and mobilization of sediment. Four treatment groups were created to assess these effects (Bovis, 1989) :  A. Logged, slash full, no recent debris flows: "slash-full (SF)" B. Logged, slash removed, no recent debris flows : "slash-clear (SC)" C. Logged, naturally scoured by debris flows: "torrented (T)" D. Unlogged, naturally loaded with CWD: "unlogged (U)"  Treatment groups are defined primarily by the CWD or slash conditions in the channel. Comparison of Group A with Group D will show how adding slash to a gully affects sediment storage and transport, relative to the natural behaviour of an unlogged gully. Comparison of Group A with Group B will show how the absence of slash affects sediment storage and transport, and will allow an assessment of the role of intentional CWD removal following harvest, as a factor in sediment management in logged areas. Gullies in both Group B and Group C do not have much  40 slash or CWD in their channels; comparison of these two groups will show how similar slash-cleared gullies are to torrented gullies.  Secondary differences between treatment groups concern variations in logging and debris flow history. Groups A and B are almost completely equivalent, with the removal of slash from the channel being the most important difference. Some secondary effects on sediment movement may occur from the placement of slash on sidewalls of Group B gullies, and from differences in yarding methods. Group C has been logged, like Groups A and B, but in addition has had a debris flow occur since logging. This implies changes to the amount of sediment stored in the channel, as well as sidewall conditions. Group D is not logged: the sidewalls as well as the channel are in their natural states. For these reasons, differences in sidewall and channel sediment and vegetation conditions, not related to slash or CWD, may affect behaviour of the treatment classes.  Within this experimental framework, the main focus of study is the monitoring of sediment mobilization and transport, as well as an assessment of which changes have resulted from treatment effects, particularly the presence or absence of logging slash in the channel of the gully. Sidewalls are treated separately from channels, since sidewalls in different treatment groups may respond differently, but not as a result of the primary treatment effect.  41 4.2 Gully sediment budget estimates  A sediment budget defines the input, change in storage, and output for an individual sediment storage element. If the gully channel is defined as the storage element, a sediment budget will define the amount of sediment entering the channel, the change in channel storage, and the output of sediment from the channel. The quantification of these terms for each gully permits testing of the specific hypothesis, stated in Chapter 1, to determine differences in treatment groups.  "A sediment budget for a drainage basin is a quantitative statement of the rates of production, transport, and discharge of detritus." (Dietrich et al, 1982). Storage elements within the drainage basin are identified, and the rates of input and output resulting from transport processes are measured. The sediment budget equation can be expressed as (Roberts and Church, 1986): I -()S = 0 where I is the volume of sediment input, CSS is the volume change in storage (an increase is positive), and 0 is the output. All terms are for an identical period of time. Each term refers to a specific storage element, with the output from one element becoming the input to the next element downstream. In this study, three storage elements can be identified: sidewalls, headwall and channel. Figure 4.1 is a conceptual model of storage elements and sediment transfers in a gully.  42.  Figure 4.1 Conceptual model of sediment storage and transfers in gullies  Bedrock  Till  Debris slides Debris flows  Frost action  Channel and Bank deposits  Storage sites  Transfer processes  43 Although all sediment transfers should be identified, a practical sediment budget will focus on the largest sediment transfers. A complete sediment budget includes dissolved material transfers (Dietrich and Dunne, 1978), but dissolution of sediment is probably not important in the short term. This is especially true of intrusive rock, typical of the study site. Consequently sediment can be considered unchanged, given the three year period reported in this study. Similarly, creep will be considered minor over the period considered.  Since the channel is the primary element of interest, the sediment budget will be constructed with reference to the channel as storage element (Figure 4.2). Input of sediment to the channel is from sidewalls of the study reach; erosion pins monitor sediment input from sidewall sources. Input from the channel upstream is not monitored, but will be shown to be minor. Change in storage is defined as an increase or decrease in channel sediment within the study reach; storage change is monitored using cross-sections. Output is defined as the volume of sediment transported past the bottom of the study reach. Coarse sediment is trapped at the bottom of the study reach.  If all major storage sites and fluxes are identified and accurately measured, the budget should balance. In practise, each budget term has error associated with it. The error for each term can be estimated, and the pooled error is calculated (the square root of the sum of all terms squared). If the budget balances within the pooled error, the budget can be  44  FIGURE 4.2 Sediment input, storage, and output for a gully channel  Upstream Input  Open slope  Erosion pins ---  Open slope Measured Cross-section  Trap Deposition bar /Channel Output  45 considered to be an adequate representation of sediment movement and storage within the gully. If the budget imbalance is greater than the pooled error, either the model does not sufficiently represent the actual storage elements and transport processes, or else measurement procedures are not sufficiently accurate.  Since construction of a sediment budget requires a balancing of the input, storage, and output terms, the length of the gully, above the sediment trap, which contributes to these terms must be estimated. Sediment input to the channel is subject to transport in high rainfall or snowmelt events (or both). If all other factors remain equal, the distance that sediment is transported in the channel should be an inverse function of the calibre of sediment introduced to the channel; thus an accurate model of sediment input into a gully would have input zones which vary with sediment size. Without further detailed study, the length of gully which contributes sediment must be somewhat arbitrarily set.  In general, the cross-sections monitor storage changes upstream of the trap for distances of 50 to 100 m. The upper limit of the cross-sections often coincides with a change in the gully morphology or vegetation, and provides a convenient break to define the study section of the gully, as well as the upper limit for sediment input from the sidewalls. If the upper cross-section is not a significant location, the greatest length of gully study section is set at 100 m.  46 4.2.1 Input measurements  Input of sediment is defined as eroded sidewall sediment which enters the gully channel. Although sediment from the upstream channel may enter the channel study reach, this term could not be measured without disturbing the sediment regime of the study reach. If input from the channel above the study reach is an important term, then the storage and output terms will total more than the sidewall input. It will be shown that since input is not under-estimated, input from the channel above the study reach may be ignored. Two main types of sidewall erosion are recognized: slumps and surface erosion. These are combined to give a total input of sediment to the channel. Surface erosion  Surface erosion is a result of rainsplash erosion, ravelling, or frost heave, and can occur over broad areas of exposed mineral soil on gully sidewalls. The widespread occurrence of surface erosion requires a monitoring program which samples the depth of sidewall retreat at several locations, then calculates total surface erosion volume as the product of average retreat and the total area of the eroding sidewall. The amount of surface erosion is monitored with erosion pins, and the amount of eroding area is estimated during gully surveys.  Erosion pins are metal spikes 250 mm long and 6 mm thick,  47  hammered into the ground perpendicular to the surface. Approximately 50 mm of pin is left exposed, and each pin is identified with a numbered plastic tag. Pins are placed in groups of 8 to 16 individuals, spaced approximately 0.5 m apart, with two rows extending across the sidewall slope. Pins usually are located on the lower portions of the sidewall slope, near the channel, so as to prevent excessive disturbance of bare gully sideslopes. In some cases these sites are the only accessible locations.  Pin exposure above the ground is always measured on the downhill side of the pin. Repeated measurement of the exposed portion of the downhill side of the pins establishes the amount of erosion or deposition over the specified measurement period. If a pin has fallen out of the ground, the previous measurement of the pin length is used to calculate the minimum amount of erosion required to cause the pin to fall.  In some sets of pins, several may be found lying on the ground at one measurement time. If these pins exhibit a clustered distribution, the erosion of these pins is treated as a slump, rather than erosion typical of the entire sidewall. As a result, the erosion pins are separated into two groups: slump pins and the rainsplash pins. The slump pin volume is estimated by multiplying the slump area by slump depth, as determined by the minimum amount of erosion required to cause the pin to fall. Individual pins which have fallen out are treated as anomalies and ignored for calculation of erosion, unless there  48 is clear evidence of significant erosion at those sites. Some possible reasons for pins falling out are animal disturbance, or falling wood or rocks dislodging the pin. Slump erosion  Slumps in Coquitlam gullies are generally small shallow failures which frequently occur near the channel margins. Individual slumps are identified during monitoring of erosion pins and cross-sections. Each slump is measured for width, depth, and length, and a volume total is calculated. Most slumps are less than 1 m 3 in volume.  4.2.2 Storage measurement  Storage of sediment in a gully channel is monitored at measured cross-sections. Cross-sections were installed during the summer and fall of 1991, with 3 to 5 cross-sections in each gully, usually less than 10 m apart. Cross-sections consist of a metal stake mounted on each bank, perpendicular to the channel axis. A fiber measuring tape is tied across the tops of the stakes, and the height of the tape above the ground at regular stations is measured. Repeated survey of each station reveals changes in the ground elevation. Each change in elevation is multiplied by half the distance between the two adjacent measurement stations. Total area change for a cross-section is the sum of  49 individual station area changes. If the cross-section stakes are located high on the gully sidewalls, some stations are deleted from the summation, since these stations do not represent changes in storage within the channel.  To calculate the change in sediment volume in the gully, the areal storage change is multiplied by a representative length of gully channel. Each gully is surveyed, and the position of the cross-sections noted. If the channel is consistent in slope and sediment size, the length of gully represented by a specific cross-section is half of the length between the two adjacent cross-sections. Cross-sections at the top or bottom of the gully have a representative length equal to the distance between the cross-section and the adjacent cross-section. Some gullies have discrete storage areas, separated by lengths of bedrock channel. In these cases, the length of channel represented by each cross-section is simply the total length of the storage zone sampled by the individual cross-section. Once each cross-section has a storage volume change calculated, the volume change for the entire gully is then the sum of the individual storage changes.  4.2.3 Output measurement  Coarse sediment output is monitored using traps positioned across the channel, at the lower end of the gully study reach. Traps are of two types, but are similar in function. Both  50 screen and culvert traps create a pool of water in the gully channel in which coarse sediment is able to settle (Figure 4.3). Sediment which remains suspended in water as it passes through the pool will be lost. Hence an estimate is required of the amount of fine sediment carried past the trap. Section 5.3.2 contains details of how fine sediment output is estimated.  Screen traps use a geotextile material, Amoco Siltstop, as a porous dam to create a local impoundment of water (Figure 4.3a). The Siltstop is 1 m in height, and is supported by reinforcing bars driven into the ground approximately 0.5 m deep. Approximately 20 cm of the bottom of the screen material is anchored by large rocks. Siltstop is a fairly permeable material, and it was originally expected that all water delivered to the trap would pass through the screen, trapping the sediment behind. In practise the permeability of the screen rapidly decreases as fine sediment and algae clog the pores. Rock spillways are constructed at the screen edges to permit passage of water. The best location for a screen trap is between large boulders, which prevent erosion of the channel banks as water is discharged around the sides of the screen.  Culvert traps are simply the basin excavated on the uphill, or ditch side of the logging roads above a culvert (Figure 4.3b). The basins are excavated for the purpose of trapping sediment to prevent clogging of the culvert. Culvert traps have greater  51 FIGURE 4.3^Geotextile screen and culvert sediment traps  a) Screen trap  b) Culvert trap  52  capacity to trap sediment than the screen traps, however, there can be additional input of sediment from ditch sources.  Both types of traps use short sections of reinforcing rod to monitor the amount of sediment deposited. "Deposition" bars are hammered into the trap basin vertically, with between 0.5 m and 1 m of bar exposed. Between three and twelve bars are installed in a trap, depending on the size of trap. Repeated measurements of the height of the exposed section of each rod records the depth of sediment deposited.  The locations of the deposition bars and the boundaries of the trap basin are mapped using triangulation. The depositional area corresponding to each bar is calculated using Theissen polygons. The volume of sediment deposited in each polygon is the product of the depth of deposition and the area of the polygon. Total trap deposition volume is simply the sum of the polygon volumes.  53  CHAPTER 5 RESULTS  Installations were monitored beginning in the summer of 1990, and continued to May, 1993. Partial gully monitoring systems were installed in the first year and by December 1991, the complete network was established. Slash-clearing of gullies occurred in September and October of 1991. Slash-clearing occurred a year later than planned, due to a logging fatality and subsequent logging moratorium. Since the autumn and winter periods are the time of greatest storms and sediment movement, yearly periods run from spring to spring. Data were collected for three years:  Year 1: July, 1990 to May 22, 1991 Year 2: May 23, 1991 to May 27, 1992 Year 3: May 28, 1992 to May 26, 1993  The study design called for 3 replicates in each treatment group (unlogged, logged and slash-full, logged and slash-clear, and logged and torrented). Site features, debris flows, and logistic problems resulted in most groups having less than three replicates. Only two appropriate old-growth gullies existed at the study site at the start of the experiment. Originally, a third gully (C7) was located in the unlogged area, but the channel had been subject to recent mass movements, and consequently, was not loaded with CWD. The November 23, 1990 storm resulted in a debris flow in gully Cl,  54  which left only one unlogged gully. Similarly, three slash-full gullies were monitored at the beginning of the period, but C6 torrented at the same time as Cl, resulting in only two slashfull gullies over most of the study period. Removal of slash from gullies was expensive, and was limited to two gullies. The study began with three torrented gullies, C3, C5, and C9. Gully C9 had torrented before logging occurred in 1990-1991; yarding in 1991 resulted in increased disturbance and slash loads to C9. As a result, C9 had to be abandoned, and C6 replaced it as the third torrented gully.  Table 5.1 shows the installation dates of sediment traps, cross-sections, and erosion pins in each gully. Sediment traps did not necessarily operate continuously after installation, since large events often either filled or destroyed them (Figure 5.1). However, cross-sections and erosion pins operated continuously after their installation.  Sediment movement can be expected to be related to the number and size of large precipitation events. Figure 5.2 shows all storms with daily precipitation greater than 80 mm in each year. Year 1 clearly had much larger events than Years 2 and 3. Year 3 had very few large events, and Year 2 had several large events, particularly at the Coquitlam River station. Therefore, if sediment movement is driven by the number and size of precipitation events, Year 1 should have the greatest sediment movement, and Year 3 the least. Since sediment traps were  55  Table 5.1 Date of installation (day/month/year)  Gully  Sediment  Cross-  Erosion  Trap  Sections  Pins  Cl-U  17/7/90  none  nonel  C2-U  17/7/90  13/6/91  2/8/91  C3-T  26/7/90  12/6/91  26/7/90  C4-SF  19/7/90  13/6/912  8/8/91  C5-T  19/7/90  12/6/91  26/7/90  C6-SF,T  23/7/90  7/8/91  26/7/90  C8-SF  6/10/90  none2  7/8/91  C10-SC  2/12/91  2/12/91  none 3  Cll-SC  2/11/91  2/11/91  none 3  1) Cl monitoring discontinued after debris flow 23/11/90. 2) C4 and C8: channel slash prevents effective cross-section monitoring. 3) C10 and C11: sidewall slash prevents installation of erosion pins.  56  Figure 5.1 Filled sediment trap, gully C3  Sediment trap at gully C3 filled after November 9-11, 1990 storm. Volume of sediment retained in trap was 0.58 m 3 . Screen has been pulled away from center support. Stadia rod (1.1 m) in left-center of photo.  57 Figure 5.2 Storm precipitation greater than 80 mm, Years 1, 2, and 3 5.2a Coquitlam Lake station 260 — E 240 220  -  0 200 -H 180 4--) Q, 160 -H 0 140a) P4 120 ›-, 100 -H 0 80 60  0  ^ ^ ^ ^ 1 2 3 4 Year  5.2b Coquitlam River station  260 — 0 240 220 0 200 -H 0 180 4-) 160-H 140 C-14 120 >1 100 -H 8060  0  = ^ ^ ^ 1 2 Year  4  58  monitored frequently in Years 1 and 2, sediment output response to individual storms will be examined (Section 5.3.4).  5.1 Input  5.1.1 Surface erosion  Rainsplash pins are compared to examine their response in each measurement year. Histograms of all rainsplash pins for each year are shown in Figure 5.3. Distribution statistics for the rainsplash pins are given in Table 5.2. Pins generally registered a small net deposition in Year 1, despite the largest storms having occurred in this year. Little difference is observed in pin responses between Years 2 and 3; both Year 2 and Year 3 have greater rates of erosion than Year 1, the year with the greatest storms. Erosion pins can act as barriers to downward movement of sediment, and hence the first year may represent a period during which the slope equilibrated to the presence of the pins. Pins are probably in equilibrium with the slope in the second year after installation. Since the distributions of pin erosion tend to be negatively skewed, median values will be used for determining differences in rainsplash pin response.  Figure 5.3 Histograms of rainsplash pins, Years 1, 2, and 3 c^,---  30  5.3b Year 2  -125-105 -85 -65 -45 -25 -5 15 35 55 75  Net change (mm/yr; erosion is negative)  59  60  Table 5.2 Rainsplash pin distribution statistics  Year 1  Year 2  Year 3  77  121  109  n mean (mm) 1  0.4  median (mm) 1  1  -12.3  -8.3  -5  -4  standard dev. (mm)  25.1  25.3  26.2  skewness  -0.4  -1.1  -1.2  kurtosis  0.7  1.7  4.5  number eroding 2^35^71^69 percent eroding^45.5^58.7^63.3 mean erosion (mm) 2^-19.2^-26.8^-20.4 median erosion (mm) 2 -12^-18^-13  1 Positive values are deposition, negative values are erosion. 2 Number of pins eroding is those pins which have negative values. The mean erosion and median erosion are based on negative values only.  61 Comparison of rainsplash pin response between gullies  Rainsplash pin response varies widely between gullies. Figure 5.4 shows the distribution of net change for each gully. In Figure 5.4, net change is combined for Year 2 and Year 3, when all erosion pin sets were operational.  All sets of pins were installed in bare soil areas, except for most of the C2 pins. C2 pins installed in vegetated slopes show an average change of +2 mm, that is, net deposition. No spatial pattern of deposition results exists on these sidewalls. Since no clear evidence of actual deposition exists at these sites, this result is considered to represent measurement error. Toy (1983) reported a measurement error of 0.5 mm from erosion pins on bare ground. The greater value in this study may be a result of the vegetation surface, which compresses a variable amount when the pin is measured, or else a result of checking pins for frost heave. Frost heave partially ejected some pins from the slope after cold weather periods. These pins could be easily pushed back into the hole which was created when the pin was ejected. All pins were checked for frost heave by pushing slightly on them, which, if the ground is soft can increase the amount of pin inserted into the ground.  All pins installed in slash-full and torrented gullies are located on bare ground. Rainsplash pins in slash-full and torrented gullies show similar responses. The modal class for all gullies is 0 to -20 mm. Some gullies have negatively skewed  62 Figure 5.4 Histograms of rainsplash pins, by gully, Years 2 and 3 60 50  -  40  -  (C2-8  30 20 10  I  In = 48  -  -  -  kN -  130  -  110  -  90  -  70 -50  N -30 -10 10 30 50 70  60 50 40 30 20 10 0  -130-110 -90 -70 -50 -30 -10 10 30 50 70 NET CHANGE (mm/yr; Erosion is negative)  63 Figure 5.4 continued  I ( I I I E i  I f  i c  64 distributions (C3-T, C4-SF, and C6-T); the other gullies (C5-T, C8-SF) have positively skewed distributions. There is no evident separation between distributions of slash-full and torrented gullies. Since bare soil areas may have several origins (debris torrent, sidewall failure, logging disturbance), the presence of some bare soil does not necessarily correlate with treatment type.  Since all pin installations are in bare ground (except for the C2 pins), rainsplash pin response at different sites should be similar. Differences in response may be caused by either slope angle or soil type. Pins are installed in either a loose, illuviated B horizon of colluvial origin, a more compact, unaltered colluvial C horizon, or else a basal till deposit.  Figure 5.5 shows median rainsplash pin response, separated by soil type and slope angle. C horizon pins tend to be located on steeper slopes, but the range in erosion rates is similar. The greatest erosion rate is for the set of B horizon pins on the steepest slope, but this group is composed of only 4 pins. If these pins are excluded, no trend in the B horizon pins is evident. The C horizon pins appear to have a trend, but slopes less than 45 degrees would have significant deposition if the data were extrapolated to include less steep slopes. Either a trend for C horizon pins is valid only for slopes steeper than 45 degrees, or else the trend is spurious.  65  Figure 5.5 Rainsplash pins Slope angle and soil type 10  C C C  0  B  C  B B  T C  B B  C C C  B  B  -50  -60  -  B  I^ I^ I^ I^ I  20  ^  30^40^50^60 70 Slope angle (degrees) B= B Horizon, C= C Horizon, T= Till  80  66 Rainsplash pin erosion rate for individual gullies.  The median rainsplash pin response is positive (i.e. deposition) in two gullies in each of Year 2 and Year 3. Since some volume of sediment is almost certainly eroded from sidewalls through surface erosion, the median rainsplash pin response is not an adequate measure of erosion. To calculate the volume of sediment input from surface erosion, only the rainsplash pins which show a negative (eroding) net yearly change will be used. The total bare area is adjusted by the proportion of rainsplash pins that are eroding. Table 5.3 shows the median erosion rate (only negative values used), the total number of pins in the gully, the number of pins with negative values, and the percentage of pins which register net erosion. Sidewall areas for calculation of rainsplash erosion  To calculate rainsplash input volume, the study reach sidewall area which is eroding must first be estimated. Total study reach sidewall area was measured during survey of the gully, and at the same time the amount of bare area was estimated. The bare area total for each gully is multiplied by the fractional proportion of rainsplash pins showing net erosion (Table 5.3) to obtain an eroding area for each gully. During surveying, the bare area was visually estimated as a percentage of the sidewall during the survey of the gully. Each survey segment (10 - 30 m) has the amounts of bare area and vegetation types  67  Table 5.3 Eroding rainsplash pin response.  Gully^Median^Eroding^Total^Fraction Erosion^Pins^Pins^Eroding (mm) Year 1 C3-T  -16.0  7  16  0.44  C5-T  -4.0  15  27  0.56  -23.5  12  30  0.40  C2-U  -1.0  2  23  0.09  C3-T  -5.0  7  14  0.50  C5-T  -18.5  20  27  0.74  C6-T  -29.0  27  30  0.90  C4-SF  -13.5  10  16  0.62  C8-SF  -7.0  4  10  0.40  C2-U  -5.0  14  24  0.58  C3-T  -16.5  6  14  0.43  C5-T  -14.0  6  15  0.40  C6-T  -21.0  26  30  0.87  C4-SF  -15.5  12  16  0.75  C8-SF  -7.0  5  10  0.50  C6-SF  Year 2  Year 3  68  present estimated (0, 15, 30, 50, 70, 85, 100%). Error is estimated at 10%, about half the range between classes. Results are shown in Table 5.4.  Gully C5 is partly vegetated in grasses and legumes. A factor of 0.5 has been applied to C5 total sidewall area to account for the presence of these grasses.  Table 5.4 Sidewall areas  Gully  ^  Total^Bare^Percent Sidewall^Sidewall^Bare  Area (m 2 )^Area (m 2 )^Area  C2-U  2040  15  1  C3-T  1910  90  5  C5-T  3120  1560  50  C6-T Y2  1130  450  40  C4-SF  2570  210  8  C6-SF Y1  1010  0  0  C8-SF  1760  140  8  C10-SC  670  0  0  C11-SC  1500  0  0  Gully type clearly has an effect on the amount of bare area. Neither slash-clear gully has any bare area. Average bare  69  area for slash-full gullies is 5%. Gully C3 has only 5% bare area, compared with 40% (C6-T) and 50% (C5-T) for the other torrented gullies. Vegetation growth on the gully sidewalls since the debris flow event in C3 has resulted in a similar amount of bare area as in the slash-full gullies. In terms of its sidewall erosion, gully C3 is more appropriately classified as a slash-full gully. Therefore, a distinction must be made between gullies which have torrented recently and those which torrented many years ago. Calculation of rainsplash input volume  The bare area in each gully (Table 5.4) is multiplied by the percentage of pins showing erosion (Table 5.3) to obtain the eroding area. The eroding area is then multiplied by the median rainsplash pin erosion depth (Table 5.3) to calculate an input volume of sediment for each gully. Table 5.5 shows the total bare area, the eroding area in each gully, the median rainsplash erosion rate, and the estimated volume of sediment eroded from the sidewalls of each gully. Torrented gullies C5 and C6 have the greatest input, and slash-full gullies have the next greatest input. The volume figures in Table 5.5 will be used to calculate total sediment input to each gully.  Error in the rainsplash erosion volume is from two sources: error in measurement of the pin, and error in estimating the area of the bare sidewall. Maximum error would occur when both  70  Table 5.5 Volume of surface erosion  Gully  Bare^Percent^Eroding  Median  Erosion  Area^Pins^Area  Erosion  Volume  (m2)^Eroding^(m2)  Rate (mm)  (m3)  Year 1 C5-T  1560^56^874  -4.0  -3.49  C3-T  91^44^40  -16.0  -0.64  C6-SF  29^48^14  -23.5  -0.33  C2-U  17^9^1  -1.0  -0.00  C5-T  1560^74^1160  -18.5  -21.4  C6-T  450^90^403  -29.0  -11.7  C3-T  90^50^46  -22.0  -1.00  C4-SF  210^62^130  -13.5  -1.76  C8-SF  141^40^56  -7.0  -0.39  C2-U  17^58^10  -5.0  -0.05  C5-T  1560^40^625  -14.0  -8.74  C6-T  450^87^390  -21.0  -8.15  C3-T  90^43^40  -16.5  -0.64  C4-SF  210^75^158  -15.5  -2.44  C8-SF  140^50^70  -7.0  -0.49  Year 2  Year 3  71  errors are in the same direction: that is, when both the area and the erosion are overestimated or underestimated. Error in estimating area is 10%, and error in erosion pin measurement is +/- 2 mm. The error volume is the difference between the erosion volume (Table 5.5) and the volume obtained when both area and erosion rate have been underestimated. Table 5.6 shows the estimated errors. Largest errors are associated with the largest eroding areas.  5.1.2 Input volume from slumping  Slumps observed in the Coquitlam gullies were generally less than 0.5 m deep and a few square metres in area. Slumps were inferred when groups of erosion pins had moved from their original locations, or from the appearance of a fresh scar on the sidewall. Calculation of slump volume in pin array areas uses the pin array to determine the areal extent of the slump, with depth assumed to be 0.2 m unless the slump scar indicates greater depth. For slumps outside of the pin arrays, length, width and depth of the slump were obtained from the scar edges. Table 5.7 combines the data to show the total input of sediment from slumping each year. Errors in slump measurements are estimated at 0.1 m for length and width, and 0.05 m for depth. Volume error estimates are included as bracketed figures.  Input of sediment from slumps is limited to the torrented gullies, apart from one slump in gully C2, and a minor one in  72  Table 5.6 Error in surface erosion volume  Gully Eroding^Erosion^Minimum^Erosion^Error Area Less Rate less Volume^Volume^Volume 10 %^(m 2 )  2 mm  (m3)  (m3)  (m3)  Year 1 C5-T  788  -2.0  -1.58  -3.49  -1.91  C3-T  36  -14.0  -0.50  -0.64  -0.14  C6-SF  13  -21.5  -0.27  -0.33  -0.06  -0.0  -0.00  -0.00  -0.00  Year 2 C2-U  1.7  -17.2  -21.4  -4.22  -11.7  -1.90  C5-T  1040  -16.5  C6-T  363  -27.0  -9.80  C3-T  41  -20.0  -0.82  -1.00  -0.18  C4-SF  117  -11.5  -1.35  -1.76  -0.41  C8-SF  51  -5.0  -0.25  -0.39  -0.14  C2-U  9  -3.0  -0.03  -0.05  -0.02  C5-T  562  -12.0  -6.74  -8.74  -2.00  C6-T  349  -19.0  -6.64  -8.15  -1.51  C3-T  35  -14.5  -0.51  -0.64  -0.13  C4-SF  142  -13.5  -1.91  -2.44  -0.53  C8-SF  63  -5.0  -0.32  -0.49  -0.18  Year 3  73  gully C4. The C2 slump occurred along the channel bank, and although the input is to the channel, the location of the slump is not strictly the sidewall (Figure 5.6). Of the torrented gullies, C5 clearly has the most active sidewalls in terms of slumping. It is worth noting that C5 sidewalls are the longest and steepest sidewalls of all the gullies.  Table 5.7 Slump volumes for Coquitlam gullies  Gully  Year 1  Year 2^Year 3  (m 3 )  (m3)  (m3)  C2-U  -1.00(0.34)  -0.03(0.02)  -0.17(0.08)  C3-T  -0.48(0.18)  -0.12(0.05)  -0.28(0.12)  C5-T  -2.55(0.62)  -2.70(0.93)  -0.07(0.04)  -0.45(0.17)  -0.54(0.23)  C6-T  0  C4-SF  0  0  C8-SF  0  0  0  C10-SC  n/a  0  0  CIA-SC  n/a  0  0  -0.07(0.04)  5.1.3 Total sediment input  The total input of sediment to each gully is the sum of the surface erosion input and slump inputs. Table 5.8 shows total input to each gully, and to allow comparison of gullies on an  74 Figure 5.6 Channel margin slump in gully C2  Large log in upper left of photo defines the left bank of the channel. Smaller log in center of photo had broken in 1990, exposing sediment which has slumped into channel. Stadia rod is 1.1 m long.  75  Table 5.8 Gully  Total sediment input Rainsplash^Slump^Total  Sidewall  Total  Input^Input^Input  Area  Normalized  (m2)  Input (mm)  (m 3 )  (m3)  (m3)  Year 1 C2 -U  n/a  1.00  n/a  2040  C5-T  3.49  2.45  5.94  3120  1.90  C3-T  0.64  0.50  1.14  1910  0.60  C4-SF  n/a  n/a  n/a  2570  C6-SF  0.27  0  0.27  1010  C8-SF  n/a  0  n/a  1760  C2-U  0.0  0.03  0.03  2040  0.02  C5-T  21.4  4.22  24.1  3120  7.71  C6-T  11.7  0.45  12.1  1130  10.7  C3-T  1.00  0.12  1.12  1910  0.59  C4-SF  1.76  0.00  1.76  2570  0.69  C8-SF  0.39  0.00  0.39  1760  0.22  C10-SC 1  0  0  0  670  0  C11-SC 1  0  0  0  1500  0  0.27  Year 2  76  Table 5.8 Continued Gully  Rainsplash  Slump  Total  Sidewall  Total  Input  Input  Input  Area  Normalized  (m 3 )  (m3)  (m3)  2 (m )  Input (mm)  Year 3 C2-U  0.05  0.17  0.22  2040  0.11  C5-T  8.74  0.07  8.80  3120  2.82  C6-T  8.15  0.54  8.69  1130  7.69  C3-T  0.64  0.28  0.92  1910  0.48  C4-SF  2.44  0.07  2.51  2570  0.98  C8-SF  0.49  0.00  0.49  1760  0.28  C10-SC 1  0  0  0  670  0  C11-SC  0  0  0  1500  0  1) Gullies C10 and C11 do not have erosion pins; zero input is based on complete vegetation and logging slash cover of the sidewalls, and no observed slumps.  77  equal basis, the total sediment volume is normalized by dividing by the total sidewall area in each study reach.  Sediment input for each treatment type is summarized in Table 5.9. Gully C3 is included with the slash-full gullies since the amount of bare area, and input, are similar to those in gullies C4 and C8.  Table 5.9 Sediment input by treatment groups  Treatment^n^Mean^Variance (mm)  ^(mm2  )  Unlogged  2  0.065  0.004  Slash-full  7  0.546  0.067  Torrented  5  6.18  13.7  Slash-clear  4  0  0  Clear differences exist between treatment types. Torrented gullies have the greatest input of sediment, as would be expected of gullies which have the greatest areas of bare sidewalls. Slash-full gullies produce the next greatest amount of sediment; this is also reflected in the amount of bare area, since slash-full gullies have the greatest amount of bare area next to the torrented gullies. The unlogged gully and the slash-clear gullies are similar. Since almost all the input into the unlogged gully came from a bank slump, the true amount of sidewall erosion approaches the amount of sidewall erosion  78  exhibited by the slash-clear gullies, that is, measurably zero.  5.2 Sediment storage in gully channels  Changes in sediment storage within gullies are the result of processes which affect the movement of sediment within the gully channel. During the study period major fluvial transport and debris flow events occurred. Unfortunately, cross-sections were not in place in Year 1 to measure changes when the largest events occurred. The storms of November, 1990 affected channels to an unknown degree. Cross-section monitoring during Year 2 and Year 3 showed moderate changes in most gullies.  Cross-sections are separated into two zones. The active zone is the middle area of the cross-section, where fluvial transport of sediment occurs. The inactive zones are at the ends of the cross-section, above the channel area. Figure 5.7 shows the active and inactive zones in a typical cross-section. Only changes within the active zone are considered, unless a major event caused clear change in the inactive zone. Since both Year 2 and Year 3 did not have major storms, no significant changes occurred in the inactive zone of any cross-section.  79  Figure 5.7 Active and inactive cross-section zones 5 TNACTIVE^ACTIVE 4.8 4.6  4  >C  -  ACTUAL CHANGE  A  3.8 3.6  0^1^2^3^4^5^6^7^8 Station (m)  —I— OCT 19/91 X JAN 11/92  80 5.2.1 Errors in cross-section measurement  Error in cross-section measurements is a significant problem. The error in an individual elevation measurement has been equated to the D 90 in stream channel surveys (Hogan, 1992). In some of the gully cross-sections, this standard may result in an estimated error of decimetres. Since this error is so large as to hide most instances of real change, an alternative method of separating real change from error is applied.  If error in an individual measurement is assumed to be random, then the probability of an error in either direction (either apparent deposition or apparent erosion) should be 0.50. Given a station with change in one direction, the probability of two adjacent stations recording change in the same direction is 0.25, assuming no actual change has occurred. This standard will be adopted to indicate whether or not real change has occurred: if three adjacent stations all show change in one direction, then the change is taken as real. As further confirmation, the cross-section profiles are examined to determine whether the changes make "geomorphic" sense. Figure 5.7 also shows actual change and measurement error.  5.2.2 Variation in storage change between treatment groups  All cross-sections were in place by November, 1991, and were monitored regularly through to May 1993. Changes in cross-  81  section area are summarized for Years 2 and 3. Year 2 changes refer to the period October, 1991 to May, 1992. Most change occurs during the autumn and winter seasons, when storm events are most frequent. Change in cross-sectional area is used since deposition or erosion tends to occur in discrete areas of a cross-section; as a result, no normalization based on crosssection width is necessary. Figure 5.8 shows there is no relationship between active zone width and amount of area change. The greatest area changes are recorded for active zones of 2-3 m width, not for cross-sections of greater width.  Figure 5.9 shows the amount of area change at each crosssection in each gully for Years 2 and 3. Year 3 shows a small but consistent pattern of deposition in almost all crosssections. Year 3 results for all gullies are not significantly different (ANOVA test, probability of equal means = 0.87).  Year 2 has a much greater range of changes, with greater amounts of both erosion and deposition, and changes appear to be related to treatment type. An ANOVA test shows significant difference in Year 2 (probability of equal means = 0.01). Torrented gullies C3 and C5 show small to moderate amounts of deposition in almost all cross-sections. Conversely, the slashclear gullies show a moderate to large amount of erosion in all but one of the cross-sections (C10, cross-section XS-3). By the end of Year 2, almost all of the channel of C10 had eroded to bedrock, which was close to the surface before treatment (i.e. slash removal). Cross-section 3 in C10 had a boulder roll into  82  Figure 5.8 Active zone width and area change  0.3 0.2  -  0.1(NI 0 gtn -0.1  XX X^ + 4‘x X +^ ^X^ x^*x X  x  X  X  X ++  -  4  -0.2  -  P -0.3  -  v 0  -0.4-0.5-0.6 0  1  ^ ^ ^ ^ ^ 2 3 4 5 6  Active zone width (m)  + YEAR 2 X YEAR 3  83  Figure 5.9 Cross-section area change, Years 2 and 3 5.9a Year 2  0.4 0.3 — 0.2  =  N  0.1  =  0^0 m 0 It 4 -0.1 0 0 -0.2 m It P -0.3 0 4) w -0.4  XS-3 ME  =  -0.5 -0.6  lir  C2 - U  C3 - T  C5 - T^C6 - T C10 - SC Cll - SC Gully  5.9b Year 3  0.4 0.3 (Ni^  0.2  0.1 W 0)^0 A (1:1 4 -0.1 0 W  m  =  -0.2  M P -0.3 0 -4-) -0.4  -0.5 -0.6  C2 - U  C3 - T  C5 - T C6 - T C10 - SCC11 - SC Gully  84 the cross-section, which caused sediment deposition around it. This effect is local, and bedrock is exposed within 2 m either side of the cross-section. Deposition shown at cross-section 3 is anomalous for C10 as a whole. The unlogged gully, C2, has a wide range of change for Year 2, as does C6, the most recently torrented gully. Overall average response for these gullies falls between the responses of the slash-clear gullies and the torrented gullies C3 and C5.  5.2.3 Storage volume changes in channels  The change in volume of channel sediment is used in the calculation of the sediment budget. To calculate the change in volume of sediment stored in a gully, the change of volume of sediment represented by all cross-sections in the gully is summed. Change in volume represented by a cross-section is the product of the areal change for the cross-section and the length of channel represented by that cross-section. In most cases, the length of the represented channel is half the distance between adjacent cross-sections. In some cases the represented channel length is less, if bedrock exposure or other evidence indicates lack of erosion or deposition. Table 5.10 shows the volume change for each cross-section and the total volume change for each gully. Error in volume estimates depends upon how representative the cross-section change is for the channel length, in addition to the measurement error for the cross-section. Since storage change between cross-sections  85  Table 5.10 Storage volume changes in channels  Gully & X-section  Channel Length (m)  Area Storage Change Year 2^Year 3 (m2)^(m2)  Volume Change Year 2^Year 3 (1113)  (m3)  C2-2  4.0  -0.03^0.03  -0.11  0.13  C2-3  6.1  -0.34^0.02  -2.09  0.12  C2-4  8.5  -0.03^0.07  -0.27  0.59  C2-5  11.5  0.20^0.03  2.29  0.38  -0.18  1.20  C2 Unlogqed  Total  C3-1  10.5  0.15^-0.06  1.55  -0.68  C3-2  11.9  0.04^0.07  0.45  0.83  C3-3  7.2  0.05^-0.06  0.39  -0.45  C3-4  5.8  0.16^0.14  0.94  0.81  C3-5  6.6  -0.03^0.01  -0.18  0.07  3.20  0.58  C3 Torrented  Total  C5-2  7.4  0.21^0.00  1.60  0.00  C5-3  7.4  0.06^0.09  0.44  0.68  C5-4  9.8  0.03^0.11  0.29  1.10  C5-5  14.8  0.23^1.6  0.05  3.45  18.1  0.03^0.00  0.63  0.00  6.40  2.60  0.78 C5-6 C5 Torrented  Total  86 Table 5.10 Continued  Gully &^Channel X-section^Length  Area Storage Change^Volume Change Year 2^Year 3^Year 2^Year 3 (m2)  (m2)  C6-1^7.0  -0.18  0.00  -1.27  0.00  C6-2^5.0  -0.23  0.09  -1.14  0.46  C6-3^4.0  0.01  0.05  0.05  0.20  -2.40  0.66  C4 Slash-full^Total 1  1.76  2.44  C8 Slash-full^Total l  0.39  0.49  (m)  C6 Torrented^Total  (m3)  (m3)  C10-1^17.0  -0.32  0.00  -5.40  0.00  C10-2^7.8  -0.50  0.07  -3.90  0.55  C10-3^4.0  0.11  0.00  0.46  0.00  C10-4^10.4  -0.51  0.00  -5.32  0.00  C10 Slash-clear^Total  -14.2  0.55  C11-1^15.1  -0.13  0.04  -1.99  0.42  C11-2^7.4  -0.23  -0.06  -1.72  -0.43  C11-3^9.8  -0.24  0.04  -2.37  0.44  C11-4^16.7  -0.32  0.05  -5.39  0.88  C11 Slash-clear^Total  1) Gullies C4 and C8, storage equal to input.  -11.5  1.30  87  is unknown, error may be considerable, and is estimated at 50 percent.  Slash accumulation in C4 and C8 prevented effective crosssection monitoring. Since no sediment output was measured in these gullies (Section 5.3), storage is assumed to be equal to input.  5.3 Output  There are two types of sediment output: coarse sediment measured at the sediment traps, and fine sediment output. Fine sediment output is carried past the traps in suspension. Despite this shortcoming, the sediment traps provide an effective measure of gully response to storms, and consideration of the role of individual storms is possible. Data for the transport distances of sediment are presented, since transport distance is an important factor for determining how much of the gully contributes to sediment output.  5.3.1 Coarse sediment output  Sediment traps were installed in July and August, 1990 at gullies Cl, C2, C3, C4, C5, C6, and C8. The C11 trap was installed November 2, 1991, and C10 trap was installed December 2, 1991. Some traps were subject to events which filled or destroyed them, consequently not all traps record all output.  88  In particular, storm events in November, 1990 filled or destroyed sediment traps in gullies C1, C2, C3, and C6. Gully morphology scaling factor  Comparison of sediment output between gullies must take account of gully size, since output volume is expected to be proportional to gully size. Sediment yield studies usually express yield as an equivalent depth of sediment throughout the basin (Schumm, 1977). However, in a set of gullies of equal basin area, variation in morphological dimensions between gullies may strongly influence the areas actually producing sediment. To compare sediment output, gully morphological parameters which might affect sediment supply and susceptibility to movement are considered. Two gully facets are used: sidewalls and the channel. Sidewalls which are longer or steeper should deliver more sediment, if all other factors are equal. Similarly, longer or steeper gully channels are also more likely to deliver sediment to the trap. A scale factor for sediment output combines these two facets: Sidewall factor = sidewall length * sin (sidewall slope) Channel factor = channel length * sin (channel slope)  Table 5.11 shows the gully dimensions. Each gully's sidewall and channel factors are scaled against the average of all gully sidewall or channel factors. The gully-scale-factor is then the average of these two factors. Table 5.12 shows the scaled  89  Table 5.11 Sidewall and channel factors  Gully Channel Average Channel Average Average Sidewall Length Channel Factor^Sidewall Sidewall Factor Slope^Length^Slope (deg.)  (rn)  (m)  (deg.)  Cl-U  136  28.7  65.3  7.4  16.3  2.08  C2-U  104  27.2  47.5  9.6  26.4  4.27  C3-T  190  19.9  46.3  17.0  24.0  4.27  C4-SF  140  29.3  68.5  12.0  19.5  4.00  C5-T  136  24.1  77.5  17.0  38.1  C6-SF,T  126  32.0  66.7  7.2  28.1  3.39  C8-SF  118  33.7  65.4  6.4  20.8  2.27  C10-SC  53  30.5  26.9  6.4  29.2  3.12  C11-SC  86  28.6  41.1  7.2  29.9  3.59  Mean  56.2  10.5  4.16  90  Table 5.12 Gully-scale-factors  Gully  ^  Scaled^Scaled^Gully  Channel^Sidewall^Scale Factor^Factor^Factor Rank  Cl-U  1.16  0.50  0.83  7  C2-U  0.85  1.02  0.94  4  C3-T  0.82  1.03  0.92  5  C4-SF  1.22  0.96  1.09  2  C5-T  1.38  2.52  1.95  1  C6-SF,T  1.19  0.81  1.00  3  C8-SF  1.17  0.55  0.86  6  C10-SC  0.48  0.75  0.61  9  C11-SC  0.73  0.86  0.80  8  Mean  1.00  1.00  1.00  91  sidewall and channel factors, and the average gully-scalefactor. The average gully-scale-factor serves to normalize sediment output between gullies. Variation in sediment output between treatment groups.  Sediment output is calculated for each gully for a set of periods. Not all gullies have data from the same set of periods, since traps were not always working, and the slashclear gullies were not instrumented until autumn 1991. Table 5.13 lists the start and end dates for each period, and the scaled (normalized) sediment output for each gully.  Slash-full gullies clearly have very little sediment output. Gully C4 does not produce any measurable sediment, and gully C8 produces measurable sediment only twice, out of twelve measurement periods. C8 output during these periods is just slightly greater than that attributable to error.  The other three types of gullies (unlogged, torrented, and slash-cleared), all produce significant but variable amounts of sediment. Since each period has different sized storm events, sediment output variation between periods is to be expected. Figure 5.10 shows the cumulative sediment output for gullies C2, C3, C5, C6, C10 and C11. Gaps in the lines indicate periods when the traps were not functioning. Steep line slopes indicate high rates of sediment output. Since gullies do not report for  Table 5.13^Scaled sediment output by time period  Period ^  Date (d/m/y Start End  1  2  3  4  5  6  7  8/90  1/11/90  18/11/90  6/91  9/8/91  29/8/91  5/9/91  2/1/91  8/8/91  28/8/91  4/9/91  0.11  0.02  0.18  1.22  0.47  0.00  0.00  0.02  0.00  0.48  0.00  0.25  0.00  0.41  0.15  0.26  0.55  0.16  0.44  0.21  0.18  0.03  0.12  0.41  0.76  0.28  0.12  0.05  0.14  0.03  0.18  31/10/90 17/11/90  8  9  10  11  12  17/11/91 22/12/91 12/1/92 26/1/92 19/11/92  16/11/91 21/12/91  11/1/92  25/1/92 26/5/92 25/5/93  Gully Cl-U  0.03  0.13  C2-U  0.04  0.03  C3-T  0.12  0.67  C5-T  1.22  3.89  5.89  C6-T*  0.16  C4-SF  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  C8-SF  0.01  0.00  0.00  0.00  0.00  0.00  0.00  0.02  0.00  0.00  0.04  0.00  C10-SC  0.03  0.59  0.65  2.46  Cu-SC  3.39  0.29  0.42  0.20  * Gully C6 operated as a slash-full gully for first 2 periods. Data is suspect and is not used.  93  FIGURE 5.10 Cumulative sediment output U, T, and SC gullies 14 12  -  0^1^2^3^4^5^6^7^8^9^10 Time period (not equal lengths)  + ^ C2-U^--X-- C3-T^---0--- C5-T - C6-T^Y^ C10-SC^C11-SC  11  12  I  94  all periods, comparison of the cumulative total is not always appropriate.  Immediately evident from Figure 5.10 is the very large output recorded from gully C5 in periods 2 and 3, representing the November 11 and November 23, 1990 storms, respectively. Rates of sediment output from gully C5 in other periods are not excessively large; thus, the scale factor for this gully is appropriate. Results from other traps in these two periods are sketchy, since almost all traps filled in either the first or the second of the two periods; however, the balance of evidence suggests that sediment output from gully C5 was anomalous during the two November 1990 events.  Slash-clear gullies C10 and C11 show steep rises in Figure 5.10 during periods 9 - 12. These two gullies produce more sediment than all others, except gully C5, despite the fact they were not monitored in Year 1 when the largest storms occurred.  To examine the differences between treatment groups, individual gullies are compared over the same set of periods. The unlogged and slash-clear gullies cannot be compared directly, since there are only two common measurement periods. Unlogged gully C2 is compared with torrented gullies C3 and C5 using sediment output from periods 1,2,4,5,6,7,8,9 and 12. Gully C6 is included for periods 7,8,9, and 12. Period 3 is not included since both gully C2 and gully C3 traps had filled by that time. Cumulative sediment outputs for the comparison periods are  95  shown in Figure 5.11.  No difference is apparent between the unlogged gully, C2, and torrented gullies. The cumulative sediment output of C2 is between the cumulative output of C3 and C5, and gully C6 shows a similar rate of sediment output in periods 7,8,9 and 12. Gully C2 shows a high rate of output in periods 6 and 7; this coincides with relocation of the trap to the road culvert pool, and hence these outputs may be affected by roadcut sediment. If this is the case, gully C2 output would be less than in the torrented gullies. Since C2 is the only unlogged gully (except for the short Cl record), C2 must be assumed to be representative of unlogged gullies. Sediment outputs from gully Cl in the first two periods are similar to gully C2 outputs (Table 5.13), which supports the assumption of C2 being representative of unlogged gullies.  Torrented gullies and slash-clear gullies are compared for periods 9-12 (Figure 5.12). Torrented and slash-clear gullies have different responses. Slash-clear gullies have similar totals, although patterns of sediment output vary. Average slash-clear sediment output is 5.0 times greater than the average torrented sediment output. The slash-clear output is actually greater than shown in Figure 5.12, since both C10 and C11 traps lost sediment; therefore, the increase in slash-clear output of 5.0 times over the torrented total is regarded as a minimum. If the torrented data and the slash-clear data are  96  FIGURE 5.11 Cumulative sediment output Unlogged and torrented gullies 8 7  -  Gr  3 0  2  -  -  X  -  1  ^X  —  ---  ^ ^ 0^1^2^4^5^6^7^8 9 12 Time period (not equal lengths) C2-U X C3-T 0 C5-T ^ C6-T  •^ 97  FIGURE 5.12 Cumulative sediment output T and SC gullies 4.5 4  -  -P  O 3.5 a, 41 3-P  O  o w  2.5-  - -1 '0^2 a) a) 73  a) 1 . 5 -  r-f  n:1  can^1-  0.5  -  1 0 ^ i 8^9^10^11 Time period (not equal lengths)  +^C3-T^ X ---  ---  C5-T^---0--- C6-T  Y C11 SC C10 SC ^ -  -  12  98  pooled, a Student's t-test rejects a null hypothesis of equal means (2-tail probability of equal means = 0.03). Hence there is a significant difference between slash-clear and torrented gully responses.  These results show significant differences between most treatment groups. Sediment output is greatest in the slashclear gullies. These gullies have a sediment output at least 5.0 times greater than the torrented gullies. Torrented gullies and the one unlogged gully appear to have similar sediment outputs, whereas slash-full gullies have very little or no sediment output. Volume of coarse sediment output  As noted above, not all sediment output could be measured since some events exceeded the trap storage volume, or else destroyed the trap. Therefore, the volumes recorded represent minima in many cases. Error is estimated in two ways. If sediment has not been lost, then the error is the product of trap area and deposition bar measurement error (0.005 m). If sediment has been lost, then the amount lost is estimated as the product of either 0.2 m or 0.3 m depth (based on depths of sediment in traps which did not fill) and the trap area. Sediment output is presented in Table 5.14. The sediment output volumes in Table 5.14 are used in Section 5.4 to calculate the sediment budget.  99  Table 5.14 Volume of coarse sediment output  Gully and  ^ Sediment Output (m 3 ) ^  Treatment  --- Year 1 ---^--- Year 2 --Total  --- Year 3 --  Error  Total  Error  Total  Error  Cl-U  0.14 1  C2-U  0.40 2  0.60  1.50 2  0.25  0.18  0.02  C3-T  0.73 2  2.0  1.18 2  0.68  0.30  0.02  C5-T  25.5  3.7  3.23  0.19  1.66  0.19  2.27 3  0.04  0.20  0.04  0.00  0.00  0.00  0.00  0.00  0.01  0.02  0.01  0.00  0.01  C10-SC  0.774,2 0.98  1.61  0.01  Cl1-SC  3.185,2 2.4  0.16  0.40  C6-T C4-SF  0.01  C6-SF  00.95 1  C8-SF  0.00  1) Torrented November 23,^1990. 2) Sediment trap filled or destroyed at some time; volume is a minimum volume. 3) C6 gully is slash-full for Year 1, and torrented for Year 2. 4) C10 trap installed December 2, 1991. 5) C11 trap installed November 2, 1991.  100  5.3.2 Fine sediment output  The sediment input from various sources cannot be directly compared with sediment trapped at the lower end of the gully. Sources of sediment, particularly from the sidewalls, can contain significant amounts of fine material which are carried in suspension past the trap. In order to balance the sediment budget, the output of both coarse and fine sediment must first be accounted for. Fine sediment output is estimated by comparing the size distribution of the source material with the size distribution of the trapped material. Method  Sediment delivered to the channel can do one of three things: it can remain in the channel, it can be carried to the trap and stored there, or it can be carried past the trap. Generally, distance travelled is associated with sediment size: larger particles are stored in the channel, intermediate sizes are stored at the trap, and fine particles are carried past the trap. Since fine material is separated from the intermediate size material at the trap, the relative proportions of fine and intermediate material in the trap and sidewall sediments should reflect the amount of lost material.  Sidewall and trap sediment are both sampled. The sediment is separated into coarse ( >16 mm), intermediate (8 - 16 mm) and  101  fine fractions ( <8 mm). All fine and intermediate material is assumed to reach the trap, and the samples are truncated to include only the intermediate and fine fractions. A unit amount of sidewall sediment yields a fraction, X, as trap sediment, and a fraction 1 - X of fine sediment output. Since the intermediate sizes (8 - 16 mm) are assumed to reach the trap, but are highly unlikely to be carried past the trap, the weight (but not the proportion) of this class is the same in both the sidewall and trap samples. The fraction of the truncated sidewall sample retained at the trap for each size class is: Trapi X = Trapi  * Sidewall 8  16  Tra p 8_16  Where: Trapi X = trap sediment of sizei, measured as a percent of truncated sidewall sample. Trapi = trap sediment of sizei, measured as a percent of truncated trap sample. Sidewall8 -16 = sidewall sediment of 8-16 mm, measured as percent of truncated sidewall sample. Trap8_16 = trap sediment of 8-16 mm, measured as percent of truncated trap sample.  Table 5.15 shows an example of the calculation procedure. In this example, a unit amount of truncated sidewall material would result in 65 % of the sediment being retained at the trap. The fraction of material carried past the trap in suspension, measured as a fraction of the total sidewall sample (all grain sizes), is then:  102  Fine Loss % = (1 - Total Trap % 1 ) * (Sidewall sub-16^weight) 100^(Total Sidewall weight) 1 Total Trap as measured as a percent of the truncated Sidewall  Table 5.15 Example of trap weight reconstruction Size Class^Truncated^Truncated^Truncated Sidewall^Trap^Trap (mm)^Sample, %^Sample, %^% Truncated Sidewall  8 - 16  29  44  29  4 - 8  29  44  29  2 - 4  14  5  3  1 - 2  14  4  2  0.5 - 1  7  3  2  < 0.5  7  1  1  Total  100  100  65 Results  Several sidewall samples and a single trap sample were collected for two gullies, C3 and C5 (Table 5.16). The sidewall samples are from widely spaced locations, and C5 Sidewall 1 is from the roadcut. Away from the road, larger samples were difficult to obtain, so smaller samples were collected.  103  Table 5.16 Sediment samples, gullies C3 and C5 ^Gully C3  ^ Gully C5 ^  Sample^Weight^Maximum^Weight Maximum (kg)^Representative^(kg)^Representative Size Fraction l^Size Fraction l (mm)  (mm)  Trap^139^32 - 45  195  32 - 45  Sidewall 1^10.4^11 - 16  131  32 - 45  Sidewall 2^6.39^11 - 16  6.44  11 - 16  Sidewall 3^4.36^8 - 11  9.08  8 - 16  Sidewall 4^3.69^8 - 11  11.9  16 - 22  Sidewall 5  12.2  16 - 22  1 Representative sample defined as n => 100 stones in a size class (Church et al, 1987).  104  The sidewall grain-size distributions for each gully are shown in Figures 5.13a and 5.13b. Each gully has a composite average sidewall sample created. Since most sidewall samples are not large enough to characterize the distribution of material larger than 16 mm, the larger samples are used to determine the fraction of material greater than 16 mm. For gully C5, sidewall sample 1 is used; for gully C3, the trap sample is used, since none of the sidewall samples are sufficiently large. When truncated for a maximum size at 16 mm, most samples have all size classes adequately represented. Three samples are not representative of the 11 - 16 mm size class since they have fewer than 100 stones in the size class, but since they are generally close (n = 67, 87, and 94), they will be used. The composited sidewall sample uses all sidewall samples within a gully to determine the average for all size classes smaller than 16 mm. The average sidewall grain-size distribution and the trap distribution are plotted in Figure 5.14. The grain size distribution of sediment in trap C6 is also shown. Once the composite sidewall sample has been constructed, it is compared to the trap sample. Each is truncated at 16 mm, and the amount of sediment retained and lost at the trap is calculated. Table 5.17 shows the results for each gully.  105  FIGURE 5.13a C3 sidewall sediment Cumulative distributions 100 90  ,X  80 70 60 50  J=1  40 30 20 10 •^ 0  ^ ^ 0.01^0.1^1 100 10 Grain size (mm)  106  FIGURE 5.13b C5 sidewall sediment Cumulative distributions 100 90 80  4=^  70 60 50 40 30 20 10  +' == -  --+  0 ^ ^ 0.01^0.1^1 10 100 Grain size (mm)  ^  ^—  107  FIGURE 5.14 Sidewall and trap sediment Cumulative distributions 100 90 80 70 60  ,*  50  ,?/  40 30 20  is,  10 0  0.01^0.1^1 Grain size (mm)  1  —  C5 SIDE^C5 TRAP C3 TRAP ^1^ C6 TRAP  -  *  1 0^100  C3 SIDE  108  Table 5.17 Fine sediment retained at trap, gullies C3 and C5  Gully C3 ^  ^  Size Class^Composite Composite C3 Trap C3 Trap C3 Trap Sidewall Side-Trunc.^Trunc.^Trunc. (mm)  91 - 128  %  %  %  %  (41  7.6  9.5  64 - 91  10.1  12.6  45 - 64  9.3  11.6  32 - 45  9.6  11.9  22 - 32  9.2  11.5  16 - 22  11.5  14.4  11 - 16  4.0  9.3  9.6  33.8  9.3  8.0 - 11  3.6  8.3  7.3  25.5  8.3  5.6^- 8.0  3.1  7.2  5.1  17.8  5.3  4.0 - 5.6  2.4  5.6  2.5  8.7  2.6  2.8^- 4.0  2.5  5.9  1.8  6.3  1.9  2.0 - 2.8  2.2  5.2  1.0  3.5  1.0  1.0 - 2.0  4.4  10.2  0.76  2.7  0.80  0.5 - 1.0  4.5  10.6  0.22  0.77  0.23  0.25^- 0.5  4.0  9.5  0.08  0.27  0.08  0.12^- 0.25  3.8  8.8  0.05  0.19  0.06  0.06^-^0.12  2.2  5.3  0.05  0.17  0.05  Pan  6.0  14.0  0.09  0.33  0.10  Total Total, <16 mm  100 42.7  100 100  28.5  100  29.7  109  Table 5.17 Continued Gully C5 ^  ^  Size Class^Composite Composite C5 Trap C5 Trap C5 Trap Sidewall Side-Trunc ^Trunc.^Trunc. (mm)  91 - 128  %  %  %  %  %1  7.7  22.2  64 - 91  10.1  19.4  45 - 64  4.6  15.3  32 - 45  7.7  11.5  22 - 32  4.5  5.2  16 - 22  5.8  5.2  11 - 16  5.4  9.0  3.9  18.3  9.0  8.0 - 11  5.1  8.6  3.9  18.3  8.6  5.6^- 8.0  4.8  8.1  2.9  13.4  6.5  4.0 - 5.6  4.6  7.7  2.9  13.4  6.5  2.8^-^4.0  3.8  6.5  1.7  8.0  3.9  2.0^-^2.8  3.5  5.9  1.7  8.0  3.9  1.0 - 2.0  6.6  11.1  2.1  10.0  4.8  0.5^-^1.0  5.9  9.8  1.1  5.3  2.6  0.25 -^0.5  4.7  7.9  0.45  2.1  1.0  0.12^- 0.25  4.7  7.9  0.25  1.2  0.57  0.06 - 0.12  2.9  4.9  0.23  1.1  0.52  Pan  7.5  12.6  0.18  0.85  0.41  Total Total, <16 mm  100 59.6  100 100  1 Percent of Truncated Sidewall.  21.3  100  48.1  110  The amount of fine sediment loss from each gully is calculated as: Fine sediment loss, gully C3: (1 - 29.7/100) * (42.7/100) = 30% Fine sediment loss, gully C5: (1 - 48.1/100) * (59.6/100) = 31%  The product of the fine sediment loss and sediment input (Section 5.1) is the total fine sediment output. Torrented and unlogged gullies are assigned a fine sediment loss of 30%. In slash-full gullies fine sediment loss is zero, since no sediment transport is observed. Fine sediment loss from gullies C10 and C11 in Year 2 is attributed to erosion of the channel material, which has an estimated 20% fine sediment. Channel sediment in slash-clear gullies had become armoured by Year 3, and in addition, storms in Year 3 were very moderate. Given these conditions, fine sediment output in Year 3 for slashclear gullies is assumed to be zero.  The near equivalence of Gullies C3 and C5 in terms of fine sediment loss is almost certainly fortuitous. The variability of sidewall grain size distributions within one gully, let alone two, suggests that the amount of sediment lost is much more variable than these two fine sediment loss terms indicate. Given the variability of sidewall grain size distributions, error in fine sediment loss could be significant. For all gullies, an error value of 50% is chosen. Table 5.18 shows the input, fine sediment output, and error for gullies in Years 2 and 3. The fine sediment output and error will be used to  111  Table 5.18 Volume of fine sediment output  Gully  ^  Input Fine Sediment Fine Sediment Fine Sediment  Volume^Loss^Output Volume^Error Volume (m 3 )  (%)  (m3)  (m3)  Year 2 C2-U  0.03  30  0.01  0.00  C5-T  24.1  30  7.23  3.62  C6-T  12.1  30  3.63  1.82  0.34  0.17  C3-T  1.12  30  C4-SF  1.76  0  0  0  C8-SF  0.39  0  0  0  C10-SC  14.  211  20  2.84  1.42  Cll-SC  11.  511  20  2.30  1.15  C2-U  0.22  30  0.07  0.03  C5-T  8.80  30  2.64  1.32  C6-T  8.69  30  2.61  1.30  C3-T  0.92  30  0.28  0.14  C4-SF  2.51  0  0  0  C8-SF  0.49  0  0  0  C10-SC  0  20  0  0  Cul-SC  0  20  0  0  Year 3  1 Input for Gullies C10 and C11 is erosion of gully channel.  112  complete a budget for each gully (Section 5.5). Since fine sediment output depends upon input, gully C3 is grouped with the slash-full gullies. Fine sediment output is dependent upon the input volume; therefore, the torrented gullies have the greatest loss of fine sediment. Fine sediment outputs in Year 2 are generally greater than those in Year 3.  5.3.1 Sediment transport distances  Sediment transported along the channel may be from two sources: the channel bed material, and sediment introduced from sidewall input. Channel bed material and introduced sediment of the same calibre may respond differently to the same flow conditions, since channel sediment fabric may resist transport of sediment. Sediment transport distances were monitored for both channel bed sediment and sediment introduced to the channel. Channel bed sediment transport was monitored in Year 2, and introduced sediment was monitored in Year 3.  Bed sediments, ranging in size from pebbles to small boulders, were painted across two cross-sections in each of C2, C5 and C6. The stones were painted on the upper surface only, so as to not disturb them. Stones were painted October 3, 1991; transport distances were measured on November 16, 1991, after a storm which delivered 93 mm of precipitation in 24 hours. Figure 5.15 shows the transport distances and the intermediate axis of each stone.  113  Figure 5.15 Channel sediment transport distances  + C2 ^ C5 — C6  114 Almost all the data shown in Figure 5.15 are from C6: 50 stones registered movement in C6, C2 had only 2 stones move, and C5 had only 8 stones move. Stones are easily transported in C6, since the channel is steep (32 0 ), and almost entirely bedrock. Stones up to 50 mm in intermediate diameter were found in the trap, over 60 m from their original position. In contrast, in C2, with a channel slope of 27 ° , only two stones moved, just 1 metre. Gully C5 showed sediment transport distances of up to 8 metres. Since the number of stones which were originally painted, and the number of stones which remained at the start point after the storm were not counted, it is not known what recovery rate was obtained.  In September of Year 3 painted stones were introduced to the channel. Gullies C2, C3, C5, C6, C10 and C11 had stones placed in two cross-sections. The stones were of two sizes: 11-16 mm, and 23-32 mm; 15 stones of each size were placed in each crosssection. Positions of stones were noted subsequently in November 1992 and May 1993. The largest storm of Year 3 was 94 mm in December 1992.  Results are similar for both periods, so May results will be discussed. Recovery rates of the painted stones are low: an average of 45% for the large stones, and 24% for the small stones. One problem is that the painted stones are often hidden by larger clasts. Of the stones recovered, 74% had not moved. Transport distances for 22 large stones ranged from 1 to 15 metres. Only gully C6 had large stones transported more than 10  115  metres. Transport distances for 11 small stones ranged from 1 to 17 metres; again, only C6 had distances greater than 10 metres recorded. With the exception of C6, there does not seem to be any difference between gullies. Given the low recovery rates, it is not certain whether the transport distances recorded represent average results.  The distance sediment is transported in the channel was monitored for both channel bed material and introduced sediment. Although conducted at different times, both types of sediment were monitored during periods when the largest magnitude storm was either 93 or 94 mm of precipitation in 24 hours. With the exception of gully C6, no sediment transport distances greater than 10 metres were measured. Since recovery rates were low, or not monitored, transport distances greater than 10 metres may have occurred, but were not observed. Gully C6 had transport distances greater than 60 metres, a consequence of its steep bedrock channel. No differences in channel bed sediment and introduced sediment transport distances are observed.  5.3.4 Precipitation and sediment output Method  The approach in this investigation is to determine whether sediment output from gullies is related to storm intensity.  116 Stage data from the Coquitlam River station indicate an average lag of 2 hours from the start of precipitation to the start of hydrograph rise (10 storms, drainage area 54.7 km 2 ). Since gullies have much smaller drainage basins (about four orders of magnitude less), the lag time should be much shorter. Therefore, there should be a strong relationship between precipitation intensity, stormflow, and fluvial sediment transport.  To examine the relationship between storms and sediment output, the maximum 3, 6, 12, 24, 48, and 72 hour precipitation totals are determined for individual storms. The storm record between sediment output measurements is examined, and any sediment output is associated with the largest storm for the period. In almost all cases, only one significant storm occurred in any given period. The gullies used in this procedure are C2, C3, C5, and C6. Other gullies did not have records of sufficient length to warrant their use.  The short record of precipitation at the study site (C6 station: November, 1991 - May, 1992) requires extension through correlation with data from the B.0 Hydro Coquitlam Lake or River stations. B.C. Hydro station data from each storm for the period that the C6 Station was operating are compared using maximum 3, 6, 12, 24, 48, and 72 hour precipitations. Results of the correlation analysis are given in Table 5.19.  117  Table 5.19 Comparison of C6 precipitation with Coquitlam Lake and River stations  Time  ^  Period  ^  Lake^DF^River^DF Station^Station (R 2 )  (R2)  3  Hrs  0.261  13  0.284  15  6  Hrs  0.600  13  0.617  15  12 Hrs  0.328  13  0.581  15  24 Hrs  0.552  13  0.594  15  48 Hrs  0.265  9  0.620  13  72 Hrs  0.238  8  0.538  12  The Lake station shows little correlation with C6 station; however, the River station shows a consistently moderate correlation, except for the 3 hour summation, which is poorly correlated. In all cases, the River station is better correlated with C6 station than the Lake station. Thus, to extend the C6 record, the River station data will be used. Regression equations used to convert River station data to C6 data are given in Table 5.20.  118  Table 5.20 Values for (est. C6) = coeff(River station) + constant  Time^River Station  Constant  Standard Error  Period^Coefficient  (mm)  C6 Estimate (mm)  3 hour^0.463  11.9  6.11  6 hour^0.810  11.4  7.91  12 hour^0.775  21.2  14.1  24 hour^0.817  26.9  22.5  48 hour^0.902  35.6  31.4  72 hour^1.01  38.9  47.2 Results  Each storm precipitation period  (3,^6,^12, 24,^48,^and 72 hour)  is compared with scaled sediment output for each gully. Figure 5.16 presents graphs of maximum precipitation intensity and sediment output for the 3 to 72 hour periods. The two outlying data points from gully C5 are the November 11 and November 23, 1990 storms.  There is no strong relationship between precipitation intensity of any time period and the amount of sediment output. Both simple regression and multiple regression (of 3 to 24 hour precipitation with antecedent precipitation for periods up to 72 hours) yield poor results (best simple linear regression R2  119  Figure 5.16 Maximum storm precipitation and sediment output 2  P n 1.8 a E 1.6 O  P  .  L,  /I  C5 3.9  C5 5.9  1.47  Z 1.2 -  0  W Z 1-4^1 A W 0.8 w A 0.6W 0.4 -  41-  >< ><  >< GI ><  0 w 0.2 0  0  0  0  ><  0^><  -0ioal5^10^15^20^25^30^35  40-  -0-  0  40  ^  45  MAXIMUM 3 HOUR PRECIPITATION (mm) 2  1.8 1.6 1.4 1.2 1  0.8 0.6 0.4  0.2 ^  10^20^30^40^50^60 MAXIMUM 6 HOUR PRECIPITATION (mm)  P n  Pi  70  ^  80  2 1.8  P 1.6 D  0 P  1.4 1.2  1 H A W 0.8 w  A 0.6  W u w  0.4 0.2  0  20^40^60^80^100^120^140 MAXIMUM 12 HOUR PRECIPITATION (mm) +  C2 x C3  D C5 - C6  ^  160  ▪ 120 Figure 5.16 continued 2 1.81.6  C5 3.9  -  1.47 1.2  C5 5.9  -  1 0.8 0.6  -  0.4  -  0.2  -  0  0  50^100^150^200 MAXIMUM 24 HOUR PRECIPITATION  250  C5 3.9^C5. 5.91111.-  1  0.8  -  0.6  -  0.4  -  ^  ).<  0.2 0  0  a  Owl  50^100^150^200^250 MAXIMUM 48 HOUR PRECIPITATION (mm)  300  2 a 1.8 1.6 0 1.4 P  1.2 1  W 0.8 N4 o  0.6  0.4 0.2 0  0  ^  50^100^150^200^250 MAXIMUM 72 HOUR PRECIPITATION (mm) ^ C6 C2 x C3 D C5  300  ^  350  121  is 0.47 for the 48 hour period; the best multiple regression R 2 is 0.55 for 3 hour maximum precipitation with 12 hour antecedent precipitation).  The data in Figure 5.16 do not show a strong trend of increasing sediment output with increasing storm precipitation; however, examination of the data shows that there is a threshold storm intensity which must be exceeded to transport sediment. The threshold is exhibited in each time period, and in almost all cases, the exceedence of the threshold is associated with sediment output. The definition of a threshold for each time period results in a storm intensity curve for the 3 to 72 hour period, which defines the minimum amount of precipitation required to transport sediment (Figure 5.17).  The threshold value for the 72 hour period is only slightly greater than the value for the 48 hour period. This suggests that the amount of precipitation in a 72 hour period is not important for sediment transport in gullies; therefore, storm intensities of less than 2 days duration appear to determine whether sediment transport will occur.  No simple explanation exists for the scatter of data points in Figure 5.16 which exceed the threshold precipitation values for sediment transport. Any relationship which exists can be confounded by precipitation gauge location relative to the study area, rain on snow events, antecedent soil moisture, or sediment supply variation. The storms for the period of C6  • 122  Figure 5.17^Threshold precipitation required for fluvial sediment transport 120  100  6 O  -  80  rts  4)  •  60 -  w  P  O  40  -H  -d  20  -  0 0  1^i^1^1^ ^ 1 ^ 1^ 70 10^20^30^40^50 60 Time (hours)  80  123  station and B.C. Hydro station comparisons are not large; thus, the correlation for storms used in this data set may not apply to storms of greater magnitude. The storms of November, 1990 appear to exhibit strong localized precipitation; very large, localized events are probably not well characterized by the River station data. In addition, the study site is about 600 m above the River station, so snow may fall (or melt) at the site when rain is recorded at the River station. Event sequences may affect the amount of sediment available for transport. During Year 2 and Year 3, the channel margins appeared to accumulate sediment in some of the gullies (C3, C5); the moderate storm events of these years may not transport as much sediment as is delivered to the channel margins.  The precipitation threshold for fluvial transport of sediment in Coquitlam gullies is compared to the threshold for shallow landslides and debris flows developed by Caine (1980) (Figure 5.18). For periods from 6-24 hours, the fluvial transport threshold is almost parallel to Caine's threshold, and shifted just slightly lower. This is a remarkable amount of similarity between fluvial processes and mass movement processes. The almost parallel lines suggest that hillslope geomorphic response to precipitation intensity over a range of durations is similar, whether the response is of a fluvial or a mass movement nature. However, since the threshold values are so similar, it suggests that the Coquitlam gullies would require a greater threshold precipitation for initiation of landslides or debris flows than the threshold determined by Caine.  124  Figure 5.18 Fluvial and debris flow precipitation thresholds 10  1  I^I^11T111^I^I^I^11111  1  10^  Time (hours) —1-- Fluvial transport X Mass transport Mass movement curve from Caine (1980)  100  125  5.4 Debris flows  During autumn and winter of 1990-1991 several large storms occurred in the North Shore Mountains. The storms of November 9-11, November 23, December 3, 1990, and April 3, 1991 caused many slope failures in the GVWD basins (Table 5.21; Thurber, 1991). The storm of November 23 caused the largest number of failures, most of them in Coquitlam Basin. All Coquitlam failures occurred in the southern half of the basin, mostly in two clusters. On the west shore of Coquitlam Lake, opposite the mouth of Cedar Creek, were eight failures, and along the length of Cedar Creek were nine failures, centered around the study site (Figure 5.19). Hence the study site was one of the two areas showing the highest frequencies of failures over the entire North Shore area. Unfortunately, no precipitation records were available from the C6 precipitation gauge located at the study site during this period. Hourly precipitation from the two B.C. Hydro stations are examined instead.  The occurrence of debris flows in some of the monitored gullies provides an opportunity to compare low intensity, frequent sediment transport events with high intensity, infrequent events. Location, surficial materials and geomorphic features are described for the two monitored gullies which torrented. Volumes of sediment incorporated into debris flows provide a comparison with the two-year sediment budgets based on low intensity processes.  126  Table 5.21 Dates and numbers of slope failures in GVWD basins in 1990-1991.  Date  Basin  ^  Capilano^Seymour^Coquitlam  9 - 11/11/90  2  3  3  23/11/90  1  3  19  3/12/90  0  1  1  3/4/91  1  1  0  5.4.1 Coquitlam precipitation  Storm intensities for the two November storms varied widely over southwest British Columbia. Table 5.22 shows the 1 day precipitation total and return period for various stations in southwest British Columbia for each storm.  The Coquitlam stations have the greatest precipitation values for each storm period. The greater precipitation amounts may simply be a result of a general increase of precipitation in mountainous areas, compared to low-elevation areas such as the Fraser Valley. Church and Miles (1987) note the Coquitlam Lake record often registers the highest precipitation in the Fraser Lowland region.  127  FIGURE 5.19 Location of Coquitlam Basin slope failures I Coquitlam River Climate Station 0 -/)  Watershed Boundary  a  N  x  x  C)k  0  ce66  x Coquitlam Lake  A Climate Station  x  Dam 0 A 5)  A Nov. 9-11 slides X Nov. 23 slides 0^  CD  km Source: Thurber, 1991  5  128  Table 5.22 Maximum one day precipitation and return periods, southwest British Columbia stations l  Station  ^  November 9 - 11  ^  November 22 - 23  Precip. Return Per.  Precip. Return Per.  (mm)  (years)  (mm)  (years)  Abbotsford A  80  9  40  1  Chilliwack  99  10  42  1  Hope A  173  > 100  116  15  North Vancouver  51  1  40  1  Squamish  164  n/a  124  n/a  Vancouver A  35  1  21  1  Whistler  72  15  52  2  Coquitlam Lake  226  252 25  156  52  Coquitlam River  144  2-3 3  189  93  1 Data from Environment Canada, 1991a and 1991b, except Coquitlam Stations. 2 From Maximum Daily Precipitation Series, Figure 4.6 3 From Partial Duration Series, Figure 4.8  129  Although the greatest number of slope failures occurred on November 23, the greatest amount of precipitation was recorded at the Coquitlam Lake station on November 9. Figure 5.20 shows the maximum precipitation for 1 hour to 72 hour periods at both the Lake and River stations. In each period, the Lake station during the November 9-11 storm recorded the greatest precipitation. If local slope failures were simply and positively related to precipitation recorded at either of the two stations, then most slope failures should have occurred during the November 9-11 storm.  Several confounding factors may prevent a simple correlation between the amount of precipitation recorded at one of the gauge locations and the number of slope failures which occurred. The gauge may not accurately reflect the amount or type of precipitation at the study site, or snowmelt may occur. Alternatively, the severe storm of November 9-11 may have fully saturated slopes, thereby increasing the chance of slope failures in subsequent events.  Severe storm precipitation is known to occur within cells of convective uplift producing increased precipitation (Church and Miles, 1987). The linear alignment of slope failure locations along Cedar Creek and on the west side of Coquitlam Lake suggests topographic confinement of the storm front with resulting convective cells. As a result, precipitation at the study site may have been much greater than that recorded at either of the two climate stations.  • 130  FIGURE 5.20 Maximum precipitation Coquitlam stations; November, 1990 400 3500  •  300  -  GZ  O 2500  _9  200  -  -H 4  c) 0 s-4 •  150100  -  -H  (t)•  50-  0  I^I^I^I^I^I^1  0  10^20^30^40^50 Time (hours)  ^  60  ^  70  80  131 Snowpack prior to a storm may cause increased delivery of water to the soil as warm rain and wind melts the snow, with possible effects on slope failure occurrences. Table 5.23 shows antecedent precipitation and temperatures leading up to the day with the greatest precipitation in each storm. Snowpacks may have been present before both storms. In the case of the November 9-11 storm, the Lake station antecedent temperatures suggest rain; however the River station temperatures indicate some precipitation may have been in the form of snow. The Lake station is at an elevation of 161 m, the River station at 280 m. Higher elevations may have held shallow snowpacks prior to the warm rain of November 8 and 9. The November 23 storm almost certainly had an antecedent snowpack. Temperatures at both stations were close to zero, and temperatures showed a diurnal pattern, suggesting solar radiation heating rather than warmer air mass invasion. Significant precipitation fell on November 19 and 21, totalling 82 mm and 85 mm at the Lake and River stations, respectively. A rapid rise in temperature was observed from 1400 hrs (0.7  °  C) to 2100 hrs (7.0  °  C) November  22, coincident with the onset of precipitation. Snowmelt probably delivered about 80 mm of water during the November 23 storm. If all the snow melted in one day, and the same amount of precipitation fell at the study site as fell at the River station (probably a minimum), total water input would be about 270 mm. The total water input, whether snowmelt is included or not, is greater than the threshold intensity defined by Caine (1980) (Figure 5.21).  132  Table 5.23 Antecedent precipitation and temperature for November, 1990 Storms  Day  Coquitlam Lake Station  Coquitlam River Station  Min Temp^Max Temp^Prec.  Min Temp^Max Temp Prec.  (0 C )^(0C)^(mm)  (0C)^(Cc)^(mm)  Nov. 5  1.1  5.7  5  -0.5  4.6  6  Nov. 6  1.9  6.1  38  0.4  3.7  27  Nov. 7  1.4  6.5  33  0.4  5.5  29  Nov. 8  2.9  9.2  93  0.4  6.1  65  Nov. 9  6.9  11.1  226  5.5  12.1  144  Nov. 19  0.6  2.9  26  0.1  2.2  43  Nov. 20  -0.3  4.0  0  -0.2  1.0  0  Nov. 21  -0.2  1.1  56  0.1  0.4  42  Nov. 22  2.9  8.4  33  0.4  7.3  47  Nov. 23  8.5  10.2  156  8.2  10.6  189  133  Figure 5.21 Coquitlam precipitation compared with Caine's curve  1  I^!III  10 Time (hours)  I^I^i^i^I^II  100  134  5.4.2 Cl debris flow  The C1 debris flow of November 23, 1990 originated as a debris slide about 380 m upslope of Branch 230 (Figure 5.22). Although on an open slope, the failure was in direct line with gully C1. The failure was roughly rectangular in plan-form, and soil depths ranged from 0.3 to 2 metres. The debris slide failure surface was the bedrock-till interface, and sloped between 29  °  °  and 32 . Piping was evident at the headscarp 2 weeks after the event. The debris flowed down gully C1, which is tributary to the adjacent gully to the southwest, which also had a debris flow November 23.  Calculation of the amount of sediment incorporated into the torrent uses survey data from December 9, 1990. The failure zone is separated from the transport zone in order to determine the relative importance of the two sources. For the failure zone, four cross-sections widths were measured, and depths from the original surface to the failure surface were estimated at three points along the cross-section. The volume of the failure was calculated as the product of the area of each cross-section and the section length for each cross-section. Similarly, the transport zone had widths and depths of erosion measured at regular intervals; the product of these measurements provides the volume incorporated into the debris flow in the transport zone. Volume estimates of debris flows must account for debris deposited along the track (Rollerson, 1984). The debris flow deposited small amounts of sediment at the edges of the forest,  135  FIGURE 5.22 Cl and C6 debris flows  Legend — — — _ Streams  Roads  Debris flow paths  Cutblock  ^ Contours (200' interval) 1^ 0^  1 metres^ 500  136  at the up-slope sides of trees. Deposited sediment is estimated at 10% of the total volume. Table 5.24 shows the results of the volume calculations.  Table 5.24 Cl debris flow volume, November 23, 1990  Failure Zone Subtotal (m 3 )^390 Transport Zone Subtotal (m 3 )^1340 Deposited Sediment (m 3 )^-170  Total C1 Debris Flow Volume (m 3 )^1560  5.4.3 C6 debris flow  The C6 debris flow of November 23, 1990 started as an open slope failure northeast of C6 gully (Figure 5.22). The failure zone is located approximately 180 m above Branch 230 road, on an open slope to the northeast of gully C6. The top edge of the failure is in thick till, approximately 3 m deep. The depth of till lessens towards the lower edges of the failure, where depths average about 1.5 m. The shape of the failure is a broad ellipse, with the length axis 16 m long, and the width axis about 10 m wide. The slope in the failure zone is 26 ° . As with the Cl failure, piping is evident at the headscarp, along the bedrock-till interface. On the bedrock surface, a 1 cm layer of extremely slippery black organic material was found. This  137  material may have contributed to instability at the site.  The debris appears to have flowed through a narrow channel down the open slope, and then entered the gully from the right sideslope 130 m above Branch 230. Some material spread across the open slope to the right of the gully. The slash deposit in the channel remained in place above the area where the debris flow entered the gully (Figure 5.23). The gully channel and sidewalls failed from Station 130 to beneath Branch 230 road. Deposition occurred around the Branch 200 road crossing.  The debris flow incorporated material from four areas: the failure site, the open-slope transport zone, the gully channel, and the sidewalls. Table 5.25 shows the volume calculation for the C6 debris flow. Gully sidewall scour is estimated to be 0.3 m. Depth of failure in the channel is estimated to be 0.5 m, over an average channel width of 4 m. Significant deposits occurred on the open slope to the north-east of the gully and in a few protected areas within the gully. Deposition along the debris flow track is estimated to account for 20% of the total volume.  Debris flows in both C1 and C6 started as open slope debris slides. Both failure sites had 2-3 metre deep soils at the headscarp, where piping was evident. The debris flow in Cl was about 75% larger than the C6 debris flow, primarily a function of the length of gully which was scoured above the Branch 230  138 Figure 5.23 Channel slash above debris flow zone, gully C6  The portion of the channel scoured by debris flow is in lower third of photo. The channel axis runs from the lower left corner of the photo to the center of the photo, where the channel is filled with slash. Sediment fill in the unscoured zone is about 0.5 m deep; slash fill is about 1 m deep. Stadia rod (1.1 m) is located at the edge of the slash, on the middleright of the photo.  139  Table 5.25 C6 debris flow volume, November 23, 1990.  Failure Zone Subtotal (m 3 )  220  Open Slope Zone Subtotal (m 3 )  140  Gully Zone Subtotal (m 3 )  760  Deposited Sediment (m 3 )  Total C6 Debris Flow Volume (m 3 )  -230  900  road. Average sediment yield in the transport zone of the C1 debris flow is 3.5 m 3 /m of channel; average sediment yield for the transport zone of C6 is 5.9 m 3 /m. Fannin and Rollerson (1993) calculated average channel debris yields between 5-10 m 3 /m. Since the Cl and C6 gullies are small, agreement with the lower debris yield rate from Fannin and Rollerson is appropriate.  5.5 Sediment budgets  The sediment budget summarizes the input, storage, and output of sediment in each gully. Budget terms of input from sidewall sources, channel storage changes, fine sediment output, and output of coarse sediment are shown in Table 5.26. The formulation of the budget equation is: Input -6Storage = Fine Output + Coarse Output If the terms do not balance, then error in one or more  140  measurements has been made. The difference between the two sides of the equation is the budget error. If the budget error is positive, then either input has been overestimated, or else storage or output has been underestimated, or a combination of errors. Negative error indicates underestimation of input, or overestimation of the storage or output terms.  Not all budget terms were measured in Year 1, so the balance cannot be computed in that year. Storage change in the slashfull gullies (C4, C8) is assumed to be equal to input, since zero or almost zero coarse outputs were measured. Fine sediment output for these gullies is also assumed to be zero, a result of the trapping capability of the slash.  In almost all cases, the budget error is less than the largest budget term. However, this yields large budget errors in some cases. If the budget error is comparable to the pooled error of the individual budget terms, then the methods used can be presumed to be accurate within the stated errors. If budget error is greater than the pooled error, then the measurement methods are not within the stated error levels. Table 5.27 shows the individual budget term errors, and the pooled error for each gully. The budget error is shown next to the pooled error for comparison.  141  Table 5.26 Sediment budgets  Gully^Input^Storage Fine Sed. Coarse Sed ^Budget Volume^Changel Output^Output^Error (m s )  (m3)  (Ins)  (m3)  (Ins)  Year 1 C2-U  n/a  n/a  n/a  C3-T  1.14  n/a  0.34  C5-T  5.94  n/a  1.78  C4-SF C6-SF C8-SF  n/a 0.27 n/a  n/a n/a n/a  n/a 0.08 n/a  0.40 0.73 25.5 0.01 0.952 0.00  Year 2 C2-U  0.03  -0.18  0.01  1.50  -1.30  C3-T  1.12  3.20  0.34  1.18  -3.60  C5-T  24.1  6.40  7.23  3.23  7.24  C6-T  12.1  -2.40  3.63  2.27  8.60 0  C4-SF  1.76  1.76  0  0.00  C8-SF  0.39  0.39  0  0.02  -0.02  C10-SC  0.00  -14.2  2.84 3  0.77  10.6  C11-SC  0.00  -11.5  2.303  3.18  6.02  142  Table 5.26 Continued, Year 3 Gully  Input^Storage  Fine Sed.  Coarse Sed  Budget  Volume^Changel  Output  Output  Error  (m 3 )^(m3)  (m3)  (m3)  (m3)  C2-U  0.22^1.20  0.07  0.18  -1.23  C3-T  0.92  0.58  0.28  0.30  -0.24  C5-T  8.80  2.60  2.64  1.66  1.90  C6-T  8.69  0.66  2.61  0.20  5.22  C4-SF  2.51  2.51  0  0.00  0  C8-SF  0.49  0.49  0  0.00  0  C10-SC  0.00  0.55  0.00  1.61  -2.16  C11-SC  0.00  1.30  0.00  0.13  -1.46  1 Positive values indicate deposition within the channel. 2 Gully C6 Output volume to November 1990, before debris torrent. 3 Gullies C10 and C11 Fine Output based on erosion of channel material, estimated at 20% fines.  143  Table 5.27 Total measurement error  Gully Input Storage Fine Sed. Coarse Sed Pooled Budget Error Error^Error^Error^Error Error (m 3 )  (m3)  (m3)  (m3)  (m3)  (m3)  Year 2 C2-U  0.02  0.09  0.00  0.25  0.27  -1.30  C3-T  0.23  1.60  0.17  0.68  1.76  -3.60  C5-T  4.84  3.20  3.62  0.19  6.84  7.24  C6-T  2.07  1.20  1.82  0.04  3.01  8.60  C4-SF  0.41  0.00  0.41  0  C8-SF  0.14  0.01  0.14  0.02  C10-SC  7.10  1.42  0.98  7.31  10.6  Cul-SC  5.75  1.15  2.40  6.34  6.02  Year 3 C2-U  0.10  0.60  0.03  0.02  0.61  -1.23  C3-T  0.25  0.29  0.18  0.02  0.42  -0.24  C5-T  2.04  1.30  0.57  0.19  2.49  1.90  C6-T  1.74  0.33  1.09  0.04  2.08  5.22  C4-SF  0.57  0  0.00  0.57  0  C8-SF  0.18  0  0.01  0.18  0  C10-SC  0.00  0.28  0.00  0.01  0.28  -2.16  Cul-SC  0.00  0.65  0.00  0.40  0.76  -1.46  144  In most cases, the budget error is greater than the pooled error (Figure 5.24). In four cases, the budget error is larger than all of the budget terms. The largest errors are associated with the largest budget terms. Torrented gullies' largest error term is input. Slash-clear gullies' largest error term is storage change. Both these terms extrapolate samples of change over a large area, which can result in multiplied errors. Sediment output error is usually small, unless the trap was overwhelmed. The magnitude of input and storage errors indicate that further refinement of gully sediment budgets would emphasize more accurate measurements of these two terms. In addition, fine sediment output needs further quantification.  5.6  Summary  Sediment budgets and complete comparisons between gullies are available for Year 2 and Year 3. Year 1 had at least two major storms (November 9-12, November 23); two debris flows occurred on monitored gullies. One gully, C5 (torrented) had a very large output of sediment during both major storms, yet did not produce a torrent. Total output for C5 in Year 1 is 25 m 3 ; other measured gully outputs did not total more than 4 m 3 in any year.  Sediment budgets in Years 2 and 3 show the amount of input, storage, and output for each gully. Two slash-clear gullies, 2 slash-full, 3 torrented and 1 unlogged gully were monitored.  145  Figure 5.24 Budget error vs. pooled error  Pooled error (m ^ 3)  146 Since only one unlogged gully was effectively monitored, unlogged response for all sediment budget terms should be considered more tentative than the results from other treatment types. Budget error is usually greater than pooled error, suggesting that not all terms were well measured. Alternatively, not all sediment transfers or storage elements were recognized. Input and storage terms appear in most cases to be the most likely source of major error.  Input is not altered by in-channel CWD and slash conditions. Consequently, differences in input are not attributable to the primary treatment effect. Secondary effects, implicit in the treatment classification, affect the amount of sediment input. Torrented gullies have the largest proportion of disturbed (bare) sidewall area. Recently torrented gullies have the greatest input volume (C5 and C6). Input from the oldest torrented gully, C3, is similar to that of the slash-full gullies (C4 and C8), indicating torrented sidewalls will recover (particularly with replanting efforts), perhaps over a period of 15 years. The unlogged gully, C2, has little input, mostly from a slump along the channel. Its input terms are most similar to those of the slash-clear gullies, where input is apparently close to zero.  Input was not monitored directly in the slash-clear gullies, but the high volumes of slash on the sidewalls probably prevented any input. No slumps, bare ground, or rainsplash erosion were observed in these gullies. Therefore, the  147 assumption of zero input appears justified.  Channel storage changes are different in Year 2 compared to Year 3. Year 3 shows consistent small amount of storage in almost all cross-sections, regardless of treatment type. The small number and size of precipitation events is a probable cause of the increase in channel storage in Year 3, since sediment may be delivered to the channel, but not transported down the channel. Year 2 results vary by treatment type. Slashclear gullies show clear evidence of erosion, whereas older torrented gullies (C3 and C5) generally show deposition in the cross-sections. The newly torrented gully C6, and the unlogged gully C2 have more variable behavior, somewhere between older torrented gullies and the slash-clear gullies. No reason is evident for the similarity in behavior of C2 and C6.  Output of coarse sediment is greatest in slash-clear gullies, and least in slash-full gullies. Unlogged and torrented gullies have less sediment output than slash-clear gullies and more sediment output than slash-full gullies. No significant difference exists in sediment outputs from unlogged and torrented gullies.  Fine sediment output is dependent upon input. The torrented gullies have the greatest input, and hence the greatest fine sediment output. Slash-clear gullies have the next largest output of fine sediment. Unlogged and slash-full gullies have little or no fine sediment output.  148  Debris flows occurred November 23, 1990 in gullies Cl and C6. Total volumes were 1560 m 3 and 900 m 3 , respectively. Both debris flows originated in soils 2-3 metres deep, with the bedrock-till interface as the failure plane. Piping was evident at the headscarps of both failures. Both failures originated on open slopes.  149 CHAPTER 6. DISCUSSION  6.1 Budget accuracy  In several of the gully sediment budgets, the budget error exceeds the pooled error. Since the pooled error is the average cumulative error from all sources, a budget error greater than the pooled error indicates either inadequate identification of storage zones or sediment transfers, or else severe errors in estimating or measuring budget terms. When individual budgets are examined, both types of errors can be found. Severe error in measurement terms is considered more common than inadequate identification of storage zones or sediment transfers. Both input and storage terms require integration of point measurements to volumes; the assumptions incorporated into volume estimates may result in severe errors. Gullies within a treatment type tend to have the same type of error, consequently, error will be discussed by treatment type.  6.1.1 Unlogged: gully C2  Errors in gully C2 are primarily measurement errors. Year 2 has the greater budget error, which is almost five times the pooled error. In Year 2, the sediment trap was located in a culvert pool. No input to the channel was recorded, a result of well vegetated sidewalls. The recorded volume of eroded channel  150 sediment does not account for all of the volume of sediment output. Some erosion of channel sediment is recorded, but not enough to account for the sediment output. Channel banks are sometimes steep and relatively high (1 m) in this gully, so unnoticed collapse of a portion of the bank is possible. However, a more likely source of the sediment is the channel within the roadcut, which is steep, and apparently unstable. Since the sediment from the roadcut is not included in the budget components, roadcut sediment delivered to the trap will result in error.  In Year 3, the sediment trap was repositioned above the roadcut. The budget error is smaller in this year, but still greater than the pooled error. The most likely source of error is in the storage measurement, which shows an increase of 0.84 m 3 , despite an input of only 0.22 m 3 . Since the unlogged sidewalls show no signs of erosion in this gully, input is probably accurate. Consequently, the storage term is more likely to be in error. Although error is large in relative terms, the actual volume amount of the error is less than 1.3 m3 .  6.1.2 Torrented gullies: C3, C5, and C6  In gully C3, the budget error is excessive in Year 2, but acceptable in Year 3. In Year 2 the storage term is the most likely source of error, since it is the largest of all terms.  151  In gullies C5 and C6, input is almost certainly over-estimated. In both gullies a high proportion of the sidewalls is bare. Gully C6 has the highest rates of rainsplash pin erosion (median C6 erosion rate, Year 2: 29 mm; median for all pins, Year 2: 18 mm). If either erosion rate or the amount of eroding area were overestimated, the error would be multiplied in the volume total.  As a basin size increases, there is an increase in the potential for temporary sediment storage (Schumm, 1977). Sidewalls in gully C5 are very long; consequently, it is not necessarily correct to assume that sediment eroded at any location on the sidewall profile will reach the channel. Sediment eroded from upper portions of the sidewall may be redeposited on the lower segments. In addition, it appears that gullies C5 and C3 have another storage zone located between the sidewall and channel (Figure 6.1). This area is subject to fluvial inundation, but appears to store sediment between large events. Some selective removal of finer sediment was observed in these areas, but they were not monitored quantitatively in this study. Since storm events were moderate in Year 2 and only one large event occurred in Year 3, significant amounts of sediment could accumulate in these channel-margin zones. The very large output of sediment from gully C5 in Year 1 may be a result of erosion of stored sediment from these areas.  152 Figure 6.1^Channel-margin storage zones in gully C5  Channel margin storage zones defined by grain size: sidewalls are fine grained, the main channel is extremely coarse grained, and the channel margin zones are intermediate between the sidewall and channel grain sizes. Two channel margin zones are visible in the photo: 1) the middle to upper left portion of the photo, and 2) the lower right portion of the photo. Stadia rod (1.1 m) in middle left of photo.  153 6.1.3 Slash-full gullies: C4 and C8.  Budget error in these gullies is somewhat artificially set at zero. Since there is virtually no measured output, and storage is defined equal to input, the budget error is, at least partly by definition, zero. Input terms for these gullies are moderate, and probably close to actual amounts.  6.1.4 Slash-clear gullies: C10 and C11.  Slash-clear gullies show dramatic changes in Year 2, and moderate changes in Year 3. Year 3 budgets are reasonably accurate. For Year 2, both gullies have erosion of channel sediment far above the measured outputs. At gully C10, the sediment trap was not installed until one month after crosssection monitoring began. Three large storms (24 hour precipitations of 93, 126 and 143 mm) occurred before the trap was installed. As a result, most of the sediment output from channel erosion was not measured. The sediment trap at C11 was installed at the same time as the cross-sections; however, the large volume of sediment eroded from the channel overwhelmed the trap, and some of the output was lost. Using channel storage changes, actual output (fine and coarse sediment combined) from gully C10 is probably about 14 m 3 , compared to the 3.6 m 3 estimated. In gully C11, actual output is probably 11 m 3 , compared to an estimated 5.5 m 3 .  154 6.1.5 Budget error summary  The estimated error for each budget term indicates that input, storage, and fine sediment output have the greatest errors. In all but one case, either input error or storage error is the greatest component of pooled error. In most gullies, estimated input or storage error is equal to or greater than the measured coarse output. Further refinement of these terms is necessary for greater accuracy in the sediment budget. In this study, fine sediment output is estimated indirectly. Increased accuracy would result from direct measurement of fine sediment output. Coarse sediment output is generally adequately measured, except in large storms which overwhelm the traps.  Given the large budget errors, do the comparisons of budget terms between treatment types still hold? For storage and coarse sediment output, the comparisons are still valid. Storage terms are compared using changes in the area of each cross-section; most error in storage terms will result from applying cross-section area changes along a length of channel. For output, comparisons are made for periods when sediment traps were operating effectively. This method avoids the major source of coarse sediment output error, which occurs when sediment traps are overwhelmed.  Comparison of the input terms between treatment types uses the input volume, normalized by sidewall area. Since input volumes are almost certainly overestimated in the torrented gullies,  155 the comparisons may not be valid. To compare input terms, the torrented gullies' input terms are adjusted by the amount of the budget error. Table 6.1 shows the old and revised input terms for each gully. Table 6.2 shows the old and revised input means for each treatment class, with Years 2 and 3 combined.  Although the revised mean input of torrented gullies is about half of the old mean input, there are still significant differences between the treatment groups. The same conclusions are reached for both the revised inputs and the old inputs: torrented gullies have the greatest input, slash-full gullies have the next greatest input, and unlogged and slash-clear gullies have very little or no sediment input.  6.2 Sediment storage and transfers  Two of the objectives of this study were to monitor the storage and transfer of sediment in gullies, and to investigate how slash and CWD affect the storage and transfer of sediment in the gully channel. The study examined the input, storage and output of sediment from gully channels, and a brief discussion of each of these terms is given below.  6.2.1 Input  Input of sediment is clearly related to treatment type. Torrented gullies have the most input, and slash-full gullies  156 Table 6.1  Gully  Old and revised input terms for sediment budgets  Old Input New Input^Old Input^New Input Year 2  Year 2  Year 3  Year 3  C2-U  0.03  0.03  0.22  0.22  C5-T  24.1  16.9  8.74  8.74  C6-T  12.1  3.5  8.69  3.47  C3-T  1.00  2.00  0.92  0.92  C4-SF  1.76  1.76  2.51  2.51  C8-SF  0.39  0.39  0.49  0.49  C10 SC 1  0  0  0  0  C11-SC  0  0  0  0  -  1)^Input for slash-clear gullies not monitored, estimated to be zero.  Table 6.2 Old and revised treatment group inputs Treatment n^Old Mean Old Var. ^New Mean New Var. (mm)  Unlogged  (mm2  )  (mm)  (mm2)  2  0.065  0.004  0.065  0.002  Slash-full 1 6  0.546  0.067  0.615  0.101  Torrented  6.18  13.7  3.45  1.41  0  0  0  0  4  Slash-clear 4  1 C3 included in slash-full treatment group for input.  157  have the next greatest input. Both the unlogged and the slashclear gullies have very minor input of sediment. The torrented gullies have the greatest input of sediment because they have the greatest amount of bare sidewall area. This is not a result of the primary treatment effect (the amount of CWD or slash within the channel), but nonetheless is characteristic of torrented gullies. Revegetation of the sidewalls of C3 with conifers has reduced the amount of bare area to an amount similar to that found in slash-full gullies about 15 years after torrenting; as a result, C3 has sediment inputs similar to the inputs of the slash-full gullies.  Sediment inputs in the slash-full gullies and slash-clear gullies are different for two reasons. The primary reason is the yarding method used. Slash-full gullies were highlead yarded; slash-clear gullies were skyline yarded. Highlead yarding generally results in greater surface disturbance to soils (Sauder and Wellburn, 1987). The second, and less important reason for differences in input, is that placement of channel slash on the sidewalls in slash-clear gullies added protection to the sidewalls. However, both the slash-full and slash-clear gullies have sideslopes of low to moderate steepness; therefore, input of sediment from logged gullies with steeper sideslopes may be greater.  In the unlogged gully, C2, input appears to originate entirely from the channel margins. Sideslopes are completely covered with vegetation or litter; as a result, they are resistant to  158 rainsplash erosion. Trees indicate that active creep occurs in the gully. The major process which delivers sediment to the channel is probably creep to the channel margins, followed by bank collapse.  6.2.2 storage  Storage of sediment is different in most treatment groups. Slash-full gullies stored about the same amount of sediment as torrented gullies. Slash-clear gullies had large amounts of channel erosion in Year 2; in Year 3, very little change occurred, and was mostly deposition. The unlogged gully showed variable response: in Year 2, the average cross-section response was erosion and in Year 3, minor deposition occurred at all cross-sections.  Storage of sediment in the slash-full gullies will probably continue for some time. Gully C4 is very small, and probably has the least water discharge of all the gullies. Slash is very thick (about 1 m at the trap), and there is no surface flow of water. As a result, sediment delivered to the channel will not be transported at all. Gully C8 is larger, slash loads are smaller, and surface flows are sometimes visible. Local transport of sediment has been observed in this gully, hence as input of sediment continues, some output of material may occur.  Storage of sediment in the torrented gullies is probably only  159 temporary, until the next large storm provides flows sufficient for fluvial transport. The channels of gullies C3 and C5 appear to possess a stable configuration of large boulders and cobbles. Gully C5 had very large output of sediment in Year 1 (25 m 3 ), but the channel itself did not appear to change (based on casual observation since cross-sections were not in place at the time). Although gully C3 torrented about 10 years earlier than C5, the channels are very similar. Consequently, torrented gully channels appear to be stable features with little change over time.  Gully C6 is the most recently torrented gully, and since it is located higher on the hillside than gullies C5 and C3, bedrock is closer to the surface. Irregularly sloping bedrock now composes most of the channel of gully C6. Storage is limited to those sections of the bedrock channel where the gradient is low, or else in sections where debris flow material was deposited along a protected sidewall (Figure 6.2). Small sediment storage structures are developing in a few areas. Whether these storage elements are stable remains to be seen.  Slash-clear gullies showed significant erosion of their channels during the first year after slash-clearing occurred. In Year 3, little or no erosion occurred. Erosion of the channel may have resulted from excessive disturbance of the gully channel by work crews. Both C10 and C11 did not have an active channel before slash-clearing operations began. Channel  160 Figure 6.2^Sediment storage in gully C6  Debris flow sediment is stored in middle right of photo, visible as a pile of boulders and CWD. The gully was seeded after the debris flow of November 23, 1990. Entrance of debris flow is visible as uppermost seeded area on right side of photo. Channel slash is visible in upper third of the photo.  161 material was generally boulders and cobbles, but covered in vegetation and litter; water flow was probably almost always subsurface. After slash-clearing, surface flow of water resulted in erosion, initially of fine organic debris and sediment, and later of larger material. Erosion did not continue in Year 3. In part, this is due to self-armouring of the channel; in part it is also due to the very low magnitude storms in Year 3.  Storage in unlogged gully C2 shows variable response. The channel of C2 is composed of sections of bedrock interspersed with sections of sediment. CWD does not appear to be the main determinant of where sediment is stored; rather, it is the location of live trees, which act as anchors for CWD, that determine sediment storage. At least two of these trees (minimum age 230 years) grow on bedrock, almost directly in the gully channel (Figure 6.3). Consequently, it appears that this gully has not had a debris flow for at least 230 years. Sediment storage in this gully is limited, since it occurs only in restricted pockets. Channel sediment is very limited in this gully for one of two possible reasons: 1) input of sediment is limited, or 2) CWD, in the necessary configuration for sediment storage, is limited. Limited input of sediment would suggest that this gully is very stable, even though sections of the sidewalls are very steep. If storage is limited due to a lack of effective CWD, then fluvial transport of sediment may remove as output most sediment delivered to the gully channel. It is  162 Figure 6.3^Old trees on bedrock in gully C2  The tree on the left of photo is directly on bedrock. The tree on photo right appears to be on bedrock as well. The channel is between the two trees, less than 5 metres apart. Stadia rod (2.1 m) leans against the left tree.  163 not known which possibility is the more likely, or whether a combination of factors is operating here.  6.2.3 Output  Coarse sediment output is greatest in the slash-clear gullies, least in the slash-full gullies, and intermediate in the torrented and unlogged gullies. Fine sediment output is greatest in the torrented gullies, least in the slash-full and unlogged gullies, and intermediate in the slash-clear gullies.  Both coarse and fine sediment output from slash-clear gullies originate from channel deposits. Whether output of large amounts of sediment will continue is uncertain. In gully C10, the channel has eroded to bedrock in almost all locations; further output of sediment will be restricted to the limited channel deposits, or else erosion of the sidewalls, a process not yet evident. In gully C11, the channel is now armoured with large boulders in most areas. Migration of the channel is possible in this gully, which could result in renewed sediment output. Some erosion of the sidewalls has already begun (observed August, 1993).  The torrented gullies are capable of very large output (C5, Year 1). But in Years 2 and 3, when comparison is possible, they produced less sediment than the slash-clear gullies. Whether the output of slash-clear gullies would have been as  164 great as the output of C5 in Year 1 is not known. The torrented gullies appear to have zones of temporary storage of sediment; these zones may have been flushed out in Year 1, thereby reducing yields in Years 2 and 3. Torrented gullies are probably the ones most likely to have variable outputs, since input from bare sidewalls and slumps can be large, and sediment can be temporarily stored along channel margins.  The unlogged gully, C2, has an output of sediment similar to that of torrented gullies C3 and C6. Most of this output occurred in two measurement periods, when the road culvert trap was used. The roadcut appears to be a likely source of sediment. Thus, C2 may actually produce less sediment than the torrented gullies.  The slash-full gullies clearly produce the least amount of sediment: either no sediment, or almost no sediment. Slash in effect creates subsurface flows; inputs of sediment to the channel do not come in contact with fluvial discharges.  6.2.4 Fluvial transport and debris flow transport of sediment  Both fluvial transport and debris flow transport of sediment occur in gullies. Fluvial transport events are of low magnitude and high frequency, whereas debris flows are of high magnitude, low frequency. Comparison of the two processes will indicate the relative importance of each process.  165  To compare fluvial transport with debris flow transport, the average output of sediment from unlogged and torrented gullies in Years 2 and 3 is compared with the average debris flow volume from gullies Cl and C6. Fluvial outputs from both torrented and unlogged gullies are used, since there were no apparent differences between the sediment outputs in torrented and unlogged gullies. However, sediment output from gully C5 is excluded, since output from this gully is an order of magnitude greater than that measured in most other unlogged and torrented gullies. Average fluvial sediment output from the torrented and unlogged gullies is 2.1 m 3 per year; the average debris flow volume is 1230 m 3 . Therefore, the average debris flow volume is equivalent to 600 years of fluvial output.  The comparison of debris flow volume with fluvial transport is, of course, very approximate. Although no differences in output are apparent between torrented and unlogged gullies, the C2 gully may actually have a lower output than the torrented gullies. Since input volumes are much greater in torrented gullies than the unlogged gully, output is probably greater as well. If this is true, then the average debris flow could be equivalent to much more than 600 years of output from the forested gullies. Average fluvial transport is determined from only three gullies, with just two years of data. Average debris flow volume is calculated from only two debris flows. Both averages use small samples of processes which are extremely variable. For all these reasons, comparisons between fluvial  166  and debris flow output are very tentative.  The sediment budget approach might provide further clarification of the relative importance of debris flow and fluvial transport processes. At any point in time, the total sediment volume available for either fluvial transport or debris flow transport could be defined as all the sediment in the headwall, sidewalls, and channel of the gully. (This approach would exclude sources outside of the gully, which would be very difficult to define). If sediment is defined as all material above bedrock, and if, over the course of a few thousand years, the volume of sediment created from bedrock weathering is negligible, then all sediment delivered from both fluvial and debris flow processes must come from this limited sediment source. (Alternatively, a bedrock weathering rate could be estimated). The volume of sediment transported by any one process, relative to the total source volume, may clarify the importance of debris flow and fluvial transport processes.  6.3 Transferability of results to larger gullies  The gullies chosen in this study were selected partly on the basis of basin size. In fact, the gullies chosen are some of the smallest in the study area. Screen sediment traps have limited storage potential; therefore, to trap most of the sediment output from the gully, output must be limited in volume. This limitation was not as severe in gullies with road  167 culvert traps. Since the relative importance of processes may be different in gullies of larger size, some consideration of how the results of this study would apply to gullies of larger size is necessary.  Larger gullies have larger sidewalls. These sidewalls may have greater disturbance from falling of trees into the gully, and from the yarding of logs out of the gully. Both slash-full and slash-clear gullies have small gully sidewalls. In C10, a section beneath the study reach has much longer sidewalls; disturbance to these sidewalls is more severe than disturbance within the study reach. Consequently, input in larger logged gullies may be proportionally greater.  Larger gullies have greater discharges of water, and the potential sediment transport capability is also greater. The slash-full gullies are small; discharge in these gullies is not large enough to transport any of the slash within the channels. In larger gullies, some transport of slash may be possible, resulting in a more clumped distribution of woody debris. In these larger gullies, sediment in the channel could also be transported, at least over short distances. Channel conditions could be much different from those observed in gullies C4 and C8.  168 6.4 Study results: implications for forest management  To be effective, management of forestry activities must achieve an objective, and be cost effective. This study has included an assessment of the feasibility and effectiveness of cleaning slash debris from gully channels, with the objective of reducing debris flow impacts on receiving stream channels. Two aspects of debris flows may be affected: their initiation, and hence the frequency with which they occur, and the amount of sediment incorporated into the debris flow - the debris flow magnitude. For slash-clearing to be effective, it must either reduce the magnitude of debris flows in gullies which are treated, or else it must reduce the likelihood of a debris flow occurring in a treated gully.  6.4.1 Slash-clearing effects on debris flow magnitude  Logging slash effectively traps sediment in the slash-full gullies studied. Coarse sediment output in these gullies is essentially zero. Thus, the presence of slash will increase the volume of sediment stored in the channel. Clearly, if a debris flow occurs in one of these gullies at a later time, the magnitude of the event will be greater than if the slash were not present. However, is this increase in magnitude significant? Input of sediment to the two slash-full gullies averaged 0.01 m 3 /m of channel length/year. Most failures in logging areas occur within 15 years of logging. Increased  169 volume of sediment after 15 years would be about 0.20 m 3 /m of channel. The Cl and C6 debris flows mobilized 4.70 m 3 /m of channel. Consequently, if a debris flow were to occur 15 years after logging, it would increase the volume of sediment scoured from the gully channel and sidewalls by about 4 percent.  Increased debris flow magnitude from slash appears to be potentially significant. Slash in these gullies averaged about 0.5 m deep and 5 m wide in the channel area. Assuming a void ratio of 0.5, the volume of slash in these channels is about 1.20 m 3 /m of channel. Sediment mobilized in the gully C6 debris flow from sidewalls and channel was 5.90 m 3 /100 m of channel. Slash incorporated into the debris flow probably increased the volume of the C6 debris flow by 20 percent. This result considers only gully sediment and CWD sources; if the open slope sources are included, slash increased the volume of the C6 debris flow by about 15 percent.  Slash-clearing can reduce the volume of channel sediment which was stored prior to logging. Removal of logging slash resulted in erosion of the channel in the slash-clear gullies. Average loss of channel sediment in gullies C10 and C11 was about 0.17 m 3 /m of channel in two years. Since the Cl and C6 debris flows mobilized 4.70 m 3 /m of channel, channel erosion in the slashclear gullies would reduce debris flow sediment volume by less than 4 percent. If a slash-cleared gully is scoured by a debris flow many years after cleaning, then there would probably be a greater reduction in the debris flow volume as a result of  170 channel erosion, but not at the rate measured in the first two years.  6.4.2 Slash-clearing effects on debris flow frequency  For the presence or absence of slash to affect the frequency of debris flows, there must be some effect on the initiation of a debris flow. Debris flows occur when a debris slide enters a gully and channel debris becomes destabilized (Rood, 1990, Benda, 1990). Slash, and the resulting increased sediment load, may affect the stability of the channel debris. Rood (1990) observed that on average, 3.3 (forested) or 3.2 (clear-cut) debris slides entering a gully resulted in one debris flow. If logging, or logging slash, alters the susceptibility of the gully channel to mass movement, then there should be a change in the average number of debris slides required to initiate one debris flow. The nearly equal ratios of debris slides to debris flows in clear-cut and forested areas suggest that logging, and logging slash, do not affect channel susceptibility to initiation of debris flows.  Channel debris does not appear to be more susceptible to debris flows if slash is introduced, but does erosion of a slashcleared channel change susceptibility to debris flows? In gully C10, removal of almost all sediment has resulted in an extremely stable bedrock channel. Conversely, if erosion is severe, undercutting of sidewalls may result, increasing the  171 rate of sidewall failures. This may result in increased incidence of debris flows.  6.4.3 Assessment of effectiveness of slash cleaning  Slash-clearing occurred in two stages in the Coquitlam gullies. The skyline was initially used to haul large slash pieces out of the gully, after which a hand crew carefully removed all slash larger than 1 cm diameter. The skyline crew was composed of 3 chokermen, one operator of the winch, and one man at the road. The hand crew was composed of 5-6 workers. Each phase of the operation took a day. This level of cleaning is more than would be considered in any industrial operation, but provides a guideline as to the cost of slash-cleaning.  As stated before, to be effective, slash-clearing must either reduce the magnitude or frequency of debris flows. Slash would increase the magnitude of sediment in a debris flow by only 4 percent, if a debris flow occurred 15 years after logging. Slash-clearing will reduce the magnitude of sediment in a debris flow by 4 percent, 2 years after slash-clearing. If slash is retained in the gully, the volume of the debris flow will increase by about 15 percent from added CWD.  Slash-clearing cannot be expected to reduce the frequency of debris flows, since there is no evidence that the presence of slash affects the incidence of debris flows. There may be an  172 increase of debris flow frequency in slash-cleared gullies where channel erosion is severe enough to undercut sideslopes.  If the objective of slash-clearing is to reduce the volume of sediment incorporated into debris flows, or to reduce the incidence of debris flows, then the benefits are either too small, or perhaps even negative, to warrant the cost of slashcleaning. On the other hand, if an objective of slash-cleaning is to reduce the volume of CWD in a debris flow, then there is a real benefit associated with slash-cleaning.  Effectiveness of the operation must be assessed over a set of gullies as a unit. Not all gullies may be susceptible to debris flows; yet unless we can identify sensitive gullies, slashclearing would need to be carried out on all gullies to achieve any benefits. In return, we run the risk of increased erosion from the slash-cleared gullies.  6.5 Conclusion  The objectives of this study were to monitor sediment storage and movement in gullies, and to assess how logging and slash affect sediment storage and movement. Four classes of gullies are studied: 1) Unlogged, with the channel naturally loaded with CWD, 2) Logged and torrented, 3) Logged, channel loaded with slash, and 4) Logged, with slash cleared from the gully channel. One unlogged gully, three torrented gullies, two  173 slash-full, and two slash-clear gullies were monitored. Sediment budgets composed of channel input, storage, and output were constructed for each gully. A parallel study is occurring in the Queen Charlotte Islands, and the results may differ from those observed in the Coquitlam basin study.  A relationship between storm intensity and sediment output could not be established; many problems exist in representing study site gully water discharge using precipitation gauges located several kilometres away. A minimum precipitation threshold for sediment movement was established; this is very similar to the threshold for shallow landslides and debris flows determined by Caine (1980). A 24 hour precipitation of at least 92 mm is required before sediment output will occur from most gullies in the study area. In Year 1, 10 events exceeded this threshold; in Year 2, 12 events, and in Year 3, 3 events exceeded the threshold.  Two very large storms occurred in November of Year 1. The storm of November 9-12, 1990 yielded 226 mm of precipitation in one day. This is the second greatest daily precipitation recorded at the Coquitlam Lake station, which has operated since 1924. The second storm occurred on November 23, which delivered 156 mm of precipitation at the Coquitlam Lake station in one day, and 189 mm at the Coquitlam River station. Snowmelt probably occurred during this second storm. Nineteen slope failures occurred during this storm, most within Cedar Creek basin and on the west shore of Coquitlam Lake opposite Cedar Creek. Two  174 of the study gullies, one unlogged, and one slash-full, were scoured by debris flows during this storm. Sediment output from the debris flows was equivalent to about 600 years of fluvial sediment output.  Sediment budgets developed for each gully show some clear differences between the treatment types. Table 6.3 summarizes the differences; treatment types in brackets are not significantly different.  Input is greatest in torrented gullies because the amount of bare ground is greatest. If revegetation of the sidewalls occurs, input will decrease. Gully C3 has been revegetated, and thus input is similar to that of the slash-full gullies. The unlogged and slash-clear gullies have very little or no sediment input since the sidewalls are well vegetated. Storage changes are not large, except in the slash-clear gullies. Torrented gullies showed moderate but consistent deposition. Slash-full gullies stored all sediment inputs. Channel erosion in the slash-clear gullies produced the greatest output. Torrented gullies, particularly C5, had the next greatest output of sediment. Output from the unlogged gully may be less than the output from the torrented gullies. Since slash-full gullies trap all sediment in the channel, they have essentially zero output.  175 Table 6.3 Ranking of sediment budget terms by treatment types  Budget^Greatest^  Least  term Input^Torrented 1 Slash-full^(Unlogged and Slash-clear) Storage 2 Torrented 3^Unlogged^Slash-clear Output^Slash-clear (Torrented and Unlogged) Slash-full  1 Torrented Gullies C5 and C6; gully C3 included in slash-full. 2 Greatest storage is deposition, Least storage is erosion. Slashfull gullies not compared. 3 Torrented gullies C3 and C5; C6 included with unlogged  The management decision of whether to clear slash from a gully must weigh the possible benefits against the potential costs. Change in the amount of sediment incorporated into a debris flow is probably not significantly affected by either retaining slash in logged gullies, or else removing it. Debris flow size may increase by about 15% if slash is retained in the channel, a result of increased amounts of CWD. If a gully is slashcleared, and no debris flow occurs, then there exists increased risk of sediment input to the channel from the undermining of sidewalls, as well as some possibility of a debris flow as a result of slash clearing. Since slash-clearing may have some negative consequences when used in a gully which would not otherwise produce a debris flow, effective use of gully cleaning should be restricted to gullies which have a high probability of mass instability.  176 REFERENCES Benda, L. 1990. The influence of debris flows on channels and valley floors in the Oregon Coast Range, U.S.A. Earth Surface Processes and Landforms, vol 15, pp 457-466 Benda, L. and T. Dunne. 1987. Sediment routing by debris flow. In Erosion and Sedimentation of the Pacific Rim (Proceedings of the Corvallis Symposium, August, 1987). IAHS Publication No. 165, pp 213-223. Bovis, M.J. 1989. Gully erosion and debris torrent initiation in Coastal British Columbia: An evaluation of the role of woody debris. Unpublished Working Plan prepared for B.C. Ministry of Forests. 5p. Bovis, M.J. and B.R. Dagg. 1987. Mechanisms of debris supply to steep channels along Howe Sound, southwest British Columbia. In Erosion and Sedimentation of the Pacific Rim (Proceedings of the Corvallis Symposium, August, 1987). IAHS Publication No. 165, pp 191-200. 1988. A model for debris accumulationand mobilization in steep mountain streams. Hydrological Sciences Journal, vol 33, pp 589-604. 1992. Debris flow triggering by impulsive loading: mechanical modelling and case studies. Canadian Geotechnical, vol 29, pp 345-352. B.C. Ministry of Forests, B.C. Ministry of Environment, Lands and Parks, Federal Department of Fisheries and Oceans, and Council of Forest Industries. 1992. British Columbia Coastal Fisheries/Forestry Guidelines. B.C. Ministry of Forests, Victoria. 102 p. Brown, C.B. and M.S. Sheu. 1975. Effects of deforestation on slopes. Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, vol 101 GT2, pp 147-165. Caine, N. 1980. Rainfall intensity-duration control of shallow landslides and debris flows. Geografiska Annaler, vol. 62(A), pp 23-27. Chatwin, S.C. 1991. Measures for control and management of unstable terrain. In A guide for management of landslide-prone terrain in the Pacific Northwest. B.C. Ministry of Forests, Land Management Report No. 18, pp 85-166.  177 Church, M.A., D.G. McLean, and J.F. Wolcott. 1987. River bed gravels: sampling and analysis. In Sediment Transport in Gravel-bed Rivers. C.R. Thorne, J.C. Bathurst, and R.D. Hey (Eds.) John Wiley and Sons, Inc. pp. 43-87. Church, M.A. and M.J. Miles, 1987. Meteorological antecedents to debris flow in southwestern British Columbia; Some case studies. In Reviews in Engineering Geology, vol 7, pp 63-79. De Ploey, J., M.J. Kirkby, and F. Ahnert. 1991. Hillslope erosion by rainstorms - A magnitude-frequency analysis. Earth Surface Processes and Landforms, vol 16, pp 399-409. Dietrich, W.E., S.L. Reneau, and C.J. Wilson. 1987. Overview: "zero-order basins" and problems of drainage density, sediment transport and hillslope morphology. In Erosion and Sedimentation of the Pacific Rim (Proceedings of the Corvallis Symposium, August, 1987). IAHS Publication No. 165, pp 27-37. Dietrich, W.E. and T. Dunne. 1978. Sediment budget for a small catchment in mountainous terrain. Zeitschrift fur Geomorphologie, Supp. 29, pp 191-206. Dietrich, W.E., T. Dunne, N.F. Humphrey and L.M. Reid. 1982. Construction of sediment budgets for drainage basins. In Sediment budgets and routing in forested drainage basins. F.J. Swanson, R.J. Janda, T. Dunne, and D.N. Swanston (Eds.) USDA Forest Service, Pacific Northwest and Range Experiment Station, General Technical Report PNW-141. pp 5-23. Environment Canada, 1991a. November, 1990 heavy rains in southern British Columbia. Prepared by E. Coatta, Atmospheric Environment Service, Vancouver, B.C. 1991b. A second occurrence of heavy rains in southern British Columbia, November, 1990. Prepared by E. Coatta, Atmospheric Environment Service, Vancouver, B.C. Fannin, R.J. and T.P. Rollerson. 1993. Debris flows: some physical characteristics and behaviour. Canadian Geotechnical Journal, vol 30, pp 71-78. Froelich, H.A. 1973. Natural and man-caused slash in headwater streams. In Loggers Handbook, vol 33, 9p. Hay, J.E. and T.R. Oke. 1973. The climate of Vancouver. Tantalus, Vancouver, 49 pp.  178 Hogan, D.L. 1986. Channel morphology of unlogged, logged, and debris torrented streams in the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 49, 94 p. 1992. Personal communication. Hogan, D.L. and J.W. Schwab. 1991a. Meteorological conditions associated with hillslope failures on the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 73, 36 p. 1991b. Stream channel response to landslides in the Queen Charlotte Islands, B.C.: changes affecting Pink and Chum Salmon Habitat. In Proceedings of the 15th Northeast Pacific Pink and Chum Workshop. Pacific Salmon Commission, Canada Department of Fisheries and Oceans. pp 222-236. Horton, R.E. 1945. Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. Geological Society of America Bulletin vol 56, pp 275-370. Innes, J.L. 1983. Debris flows. Progress in Physical Geography, vol 7, pp 469-501. Krag, R.K., E.A. Sauder, and G.V. Wellburn. 1986. A forest engineering analysis of landslides in logged areas on the QCI. B.C. Ministry of Forests, Land Management Report No. 43, 138 p. Linsley, R.K., M.A. Kohler, and J.L.H. Paulus. 1975. Hydrology for Engineers. McGraw-Hill. New York. pp 356-357. O'Loughlin, C.L. 1972. An investigation of the stability of the steepland forest soils in the Coast Mountains, Southwest British Columbia. Ph.D. thesis. Univ. B.C. Vancouver. Roberts, R.G. and Church, M. 1986. The sediment budget in severely disturbed watersheds, Queen Charlotte Ranges, British Columbia. Can. Journal of Forest Resources. 16: 1092-1106. Roddick, J.A., 1965. Vancouver North, Coquitlam, and Pitt Lake Map Areas, British Columbia. Geological Survey of Canada, Memoir 335. 276 p.  179 Rollerson, T.P. 1984. Terrain Stability Study - TFL 44. Land Use Planning Advisory Team. Woodlands Services, MacMillan Bloedel Ltd. Nanaimo, B.C. 1992. Relationships between landscape attributes and landslide frequencies after logging: Skidegate Plateau, Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 76, 11 p. 1993. Personal communication. Vancouver, B.C. Rood, K.M. 1984. An frequency Charlotte Forests.,  aerial photograph inventory of the and yield of mass wasting on the Queen Islands, British Columbia. B.C. Ministry Land Management Report No. 34. 55p.  1990. Site characteristics and landsliding in forested and clearcut terrain, Queen Charlotte Islands, B.C.. B.C. Ministry of Forests, Land Management Report No. 64, 46 p. 1992. Personal communication. Vancouver, B.C. Sauder, E.A, R.K. Krag, and G.V. Wellburn. 1987. Logging and mass wasting in the Pacific Northwest with application to the Queen Charlotte Islands, B.C.: A literature review. B.C. Ministry of Forests, Land Management Report No. 53, 26 p. Sauder, E.A., and G.V. Wellburn. 1987. Studies of yarding operations on sensitive terrain, Queen Charlotte Islands, B.C.. B.C. Ministry of Forests, Land Management Report No. 52, 45 p. Schumm, S.A. 1977. The fluvial system. John Wiley and Sons, New York. 338 p. Sidle, R.C., A.J. Pearce, and C.L. O'Loughlin. 1985. Hillslope stability and land use. American Geophysical Union. Water Resources Monograph 11, 140 p. Sidle, R.C., and D.N. Swanston. 1982. Analysis of a small debris slide in coastal Alaska. Canadian Geotechnical Journal, vol 19, pp 167-174. Slaymaker, 0., M. Bovis, M. North, T.R. Oke, and J.R. Ryder. 1992. The primordial environment. In Vancouver and Its Region. G. Wynn and T. Oke (Eds.) UBC Press, Vancouver. pp 17-37. Swanston, D.N. and D.E. Howes. 1991. Slope movement processes and characteristics. In A guide for management of landslide-prone terrain in the Pacific Northwest. B.C. Ministry of Forests, Land Management Report No. 18, pp 1-17.  180 Swanston, D.N. and F.J. Swanson. 1976. Timber harvesting, mass erosion, and steepland forest geomorphology in the Pacific Northwest. In Geomorphology and engineering. D.R. Coates (Ed.) Dowden, Hutchinson, and Ross. Stroudsburg, Pa. pp 199-221. Takahashi, T. 1981. Debris flow. Annual Review of Fluid Mechanics, vol 13, pp 57-77. Thurber Engineering. 1991. Geotechnical assessment of 1990-1991 landslide events in Greater Vancouver Water District Watersheds. Unpublished report to Greater Vancouver Water District. 26 p. Tripp, D.B. and V.A. Poulin. 1986a. The effects of mass wasting on juvenile fish habitat in streams in the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 45, 48 p. 1986b. The effects of logging andmass wasting on salmonid spawning habitat in streams in the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 50, 29 p. 1992. The effects of logging and mass wasting on juvenile salmonid populations in streams on the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 80, 38 p. Toy, T.J. 1983. A comparison of the LEMI and erosion pin techniques. Zeitschrift fur Geomorphologie, Supp. 46. pp 25-34. Varnes, D.J. 1978. Slope movement types and processes. In Landslides, analysis, and control. R.L. Schuster and R.J. Krizek. (Eds.) National Academy of Sciences, Transportation Research Board, Special Report 176, pp 11-33. Wright, J.B. and C.H. Trenholm. 1969. Greater Vancouver precipitation. Atmospheric Environment Service, Technical Memo No. 722. 36 pp. Young, S.E. 1992. Slope stability prediction techniques for forest management purposes - A case study from the Queen Charlotte Islands, British Columbia. Master's thesis, University of British Columbia, Department of Geography, Vancouver, B.C. 150 pp. Ziemer, R.R. 1992. Effect of logging on subsurface pipeflow and erosion: coastal California, USA. In Erosion, Debris Flows and Environment in Mountain Regions (Proceedings of the Chengdu Symposium, July, 1992). IAHS Publication no. 209. pp 187-197.  181  APPENDIX  This appendix includes data from the erosion pins, crosssections, and sediment output traps. Erosion pin data are the amount of change, in millimetres, recorded at each pin in each measurement period. Cross-section profiles are shown at three times: autumn 1991, May 1992, and May 1993. Sediment output volumes are shown for each measurement period.  Gully C2^Unlogged: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin Bank  ^ End date of measurement period (d/m/yr) ^ 3/10/91 16/11/91 25/5/92 4/9/92 19/11/92 25/5/93 8/8/91  37 38 39 40 41 42 43 44  L L L L L L L L  2 -2 -1 -2 1 -1 0 -1  2 0 4 -1 0 3 5 0  0 3 0 3 0 -7 -4 0  45 46 47 48 49 50 51 52 53 54  R R R R R R R R R R  -3 0 -1 -1 -1 -3 2 -1 -16 -10  11 5 -28 2 4 6 -3 -3 20 14  -8 -7 -6 10 6 -12 -3  55 56 57 58 59 60  L L L L L L  -7 3 -1 2 1 17  2 -1 3 1 3 -4  -1 -4 0 5 -1 -1  -2 3 1 2 0 9 5 0  4 -2 9 -4 -2 -2 2 1  -8 -2 19 8 8 1 0 15 6  -3 -4 -5 -2 -5 0 -2 12 -2 2  6 8 -1 3 4 3  -4 -4 -2 -4 0 -2  1 0 -12 -2 0 -2 -2 0  0 3 3 4 1 0 2 2 -34  -3 -1 -11 -9 9  -60  -3  6 2 14 13  0 -3 10 -11 5 -1  3 3 -13 18 -5 -4  25/5/92 Pin 47 slump 0.5m * 0.5m * 0.117m = 0.029 m3. 25/5/93: Slumped pins 46 + 47:^0.5m * lm * 0.160m =^.08 m3. 25/5/93: Slumped pins 49 + 50:^.5m * lm * 0.172m = 0.086 m3.  1-, 00  N  Gully C3 Torrented: Erosion pins Erosion amounts in mm, positive numbers indicate deposition ^  End date of measurement period (d/m/yr) ^  Pin Bank^31/10/90 22/05/91 8/8/91^19/10/91 11/1/92^1/2/92^26/5/92^4/9/92^18/11/92 25/5/93 33 R^-8^-36^2^64^-72 34 R^13^2^10^-3^-3 35 R^10^ 3^10^-10 36 R^7^-11^0^5^-17 37 R^-17^ -3^24^-34 38 R^25^33^-47^2^15 39 R^3^-73^3^35^-100 40 R^12^26^-26^0^-1  1^0 3^8 1^3 10^11 1^-2 3^3 4^-16 4^1  -1 7 1 -3 0 -3 65 10  -1 -16 2 1 0 -2 2 -10  41 42 43 44 45 46 47 48  0^-3 -3^5  0 -3  -6 13  -22 -31  1^-2 12^7 3^8 1^3  3 -28 -5 0  -1 34 0 -3  42 -11 13 2  R^10^5^-3^1^12 R^1^15^-29^12^14 R^6^28^28^0 R^-3 R^12^7^5^5^-46 R^-1^ -5^6^-8 R^50^-27^-4^6^-2 R^-19^3^7^3^-6  22/05/91:^pins 35,^37,^44,^and 46 treated as slumps: Volume of 35 and 37 slump is 0.5m*lm*.15m = 0.075m3; Volume of 44 and 46 slump is .5m*lm*0.2m = 0.10m3. 11/1/92 Pin 43:^slump lm * 0.5m * 0.247m = 0.124 m3. 25/5/93: Pins 33 and 35 treated as slumps: lm * lm * .18m = 0.18m3. 25/5/93:^Pin 39 slump lm * 0.5m *^.193m = 0.096m3.  10 -19 -15 -31 3  Gully C4 Slash-full: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin Bank  --- End date of measurement period (d/m/yr)^-19/10/91 8/12/91 25/5/92 4/9/92 25/5/93  71 R 72 R 73 R 74 R 75 R 76 R  16 8 -2 1 7 8  -7 -16 -4 -5 -13 -20  1 -3 -5 -9 -12 -18  -8 -10 -9 -7 10 -4  -5 -23 4 5 -15 8  R R R R R R R R R R  13 8 -18 1 4 1 3 5 5 1  -2 -36 27 6 2 0 -4 -5 -7 -22  -2 -39 -9 -10 1 3 16 -14 -11 -57  -4 -3 -8 -3 2 -17 1 -13 -4 7  4 -9 -10 -16 -13 -22 -35 32 -60  77 78 79 80 81 82 83 84 85 86  25/5/93: Pin 86 slump lm * .5m * .134m = 0.067m3.  Gully C5 Torrented: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16  Bank L L L L L L L L L L L L L L L L  31/10/90 22/05/91 8/8/91 15 6 26 12 34 -4 -9 -3 45 0 6 -30 28 10 -3 8  -52 -8 -12 -14 -15 2 12 14 10 -35 -15 37 -6 -23 -15 -210  End date of measurement period (d/m/yr) ^ 26/5/92 4/9/92 19/11/92 25/5/93 3/10/91 16/11/91 1/2/92  5 9 4 6 2 3 11 6 2 -4 7 0 -16 -5 -5  4 14 3 5 0 9 5 -5 -2 -5 8 -7 12 11 9  -7 -3 -1 -10 -5 -11 -20 18 2 11 -5 -3 2 1 -18  -19 -8 -23 -5 11 -31 -16 2 -23 -25 -60 -16 -10 -2 -32  -3 2 1 0 -66 -19 -9 -18 -10 8 4 12 -13 0 7  -4 4 -2 4 13 3 9 8 -12 27 -12 -19 11 -2 -11  -3 18 0 -7 -2 -18 -9 -4 -37 -5 4 38 -6 -1 3  11 5 5 -3 30 -135 42 6 47 1 56 -44 -13 2 -29  Gully C5 Torrented: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32  Bank R R R R R R R R R R R R R R R R  ^ End date of measurement period (d/m/yr) ^ 31/10/90 22/05/91 8/8/91^3/10/91^16/11/91 1/2/92^26/5/92^4/9/92^19/11/92 25/5/93 34^-5^BURIED? 6^-7^20^10^-4^-52 30^-34^10^-13^19^9^-17^-3 10^-14^9^4^7^-1^-3^-16 14^-40^77^-62^17^1^26^-42 16^4^-19^-1^-1^2^3^-5 2^13^-21^0^-3^-7^3^-5 6^21^-1^18^-5^-5^-11^-6 -2^NOTE -2^ 7^2^7^-5^5^1 2 4^ 1^-7^3^-21^-4^-5 1^ 2^-5^14^5^-20^-4 3 -1^ 7^11^-26^-10^1^18 -14^ 8^2^-6^2^-7^-36  NOTE: PINS 25 - 32: Treated as slump rather than surface erosion. Slump volume = 4m * 2m * PIN 18, 31/10/90 treated as slump of .1m3. 25/5/93 Pin 6: slump lm * .5m * 0.135m = 0.068 m©3.  .3m = 2.4m3.  Gully C6 Slash-full, Torrented: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin 49 50 51 52 53 54 55 56 57 58 59 60 61 62  Bank R R R R R R R R R R R R R R  End date of measurement period (d/m/yr) 18/11/92 26/5/93 31/10/90 03/06/91 27/8/91 3/10/91 25/5/92 4/9/92 -62 6 -41 4 -2 -74 -5 -2 -2 1 0 24 40 -16 -18 -7 -46 -2 -5 1 31 -1 -23 8 0 -50 39 11 33 -39 10 -72 -3 -9 -3 -3 -5 -44 -2 -10 -12 4 25 -7 -5 -26 -35 1 0 -14 25 10 5 4 -12 -67 -45 13 14 5 16 7 -1 7  -13 -17 13 -4 -3 -1  0 -1 17 11 3 9  6 5 -23 -8 -3 -29  -18 -12 -3 -15 -9  9 -9 14 -5 2 -9  -13 7 -18 4 -10 -28  ^  Gully C6 Slash-full, Torrented: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80  Bank L L L L L L L L L L L L L L L L  ^ End date of measurement period (d/m/yr) ^ 31/10/90^03/06/91 27/8/91^3/10/91^25/5/92^4/9/92^18/11/92 26/5/93 26 23^6^3^-47^-8^1 -14 8^-8^-3^-13^-14^12 8 3^-14^-10^-48^-6^1 -7 -13 13^-9^7^-38^-6^ -8 2^-2^-15^-33^-1^7 20 20^-18^4^ -6^28^-10 3^-7^-19^ -1^-29^1 -40 5^28^-28^ -1^-24^1 -1 21^-19^-4^-24^-3^4 -21 3^-44^-2^-27^1^-6 2 11^-58^15^-10^-2^-1 -96 -5^-41^-15^-89^-10^4 -55 1^-3^-40^-33^-1^18 -26 -3^-22^-11^ -7^-21^5 -107 -4^22^-53^-31^-8^-9 -38 -2^5^-14^-30^0^7  NOTE: 25/5/92 Pins 50, 52, and 53 reresent slump of lm * 2m * .2m = 0.4m3. 26/5/93 Pins 51, 54, 56: slump 3m * lm * 0.158m = 0.475 m©3. 26/5/93 Pin 71: slump lm * .5m * 0.126m = 0.063m©3.  Gully C8 Slash-full: Erosion pins Erosion amounts in mm, positive numbers indicate deposition  Pin Bank 61 62 63 64 65 66 67 68 69 70  L L L L L L L L L L  -- End date of measurement period (d/m/yr) 29/8/91 25/1/92 17/9/92 18/11/92 26/5/93 3 8 0 7 -1 0 -8 -1 7 1  2 1 -7 -1 3 -9 13 -6 -8 0  -23 4 0 -5 4 4 7 1 -18  1 -1 -5 -3 -2 13 -5 0 18 6  -5 -18 6 -2 -5 11 5 12 -3 -25  190  Gully C2, XS-2  4.8  i  4.6  -  4.4  -  3.6  -  3.4  0  1  2^3 Station (m)  4  5  -1--- 3/10/91 ---*--- 25/5/92 ^x^ 25/5/93  Gully C2, XS-3  5^ 4.8 4.6 i 4.4  -  e 4.2 - -I 4) ni^4a) ,--1 3.8w '  3.6 3.4 3.2  0  0.5  ^ ^ 1 1.5  -1---  2^2.5^3 Station (m)  3.5  ^ ^ ^ 4 4.5 5  3/10/91 ---*--- 25/5/92 ^x^ 25/5/93  191 Gully C2, XS-4  5 4.8 4.6  -  4.4 4.2  -  43.8 3.6 3.4  0  0.5^1^1.5^2^2.5 Station (m)  3.5  3  4  4.5  I— 3/10/91^25/5/92 ^x^ 25/5/93  Gully C2, XS-5  4.8 4.64.4 0 0  4.2 -  0:1 0  W  43.8 3.6  0  1  2  ♦  3^4 Station (m)  5  3/10/91^25/5/92  6 25/5/93  7  192  Gully C3, XS-1  5 4.8  -  4.6 • 4.4  -  4.2 co 4-  W  3.8-  3.6 3.4  -  -  3.2 ^ 0  1^2^3^4 6 5 Station (m) 26/5/92^)4^ 25/5/93 19/10/91 * --  --  Gully C3, XS-2  0  ^ ^  1  2^3^4^5 6 Station (m) ♦ 19/10/91^26/5/92  7 25/5/93  193  Gully C3, XS-3  i  0  ^  ^ ^ ^ 2^4^6 10 8 12 Station (m) ---41— 19/10/91 -- >,E -- 26/5/92^x^ 25/5/93  Gully C3, XS-4  1^2^3^4 5 Station (m) ---11--- 19/10/91^>PE^26/5/92  x^ 25/5/93  194  Gully C3, XS-5  4.8 4.6  -  4.4 i 4.2 4 3.8 3.6 3.4 3.2  -  -  -  -  -  -  3^ 0^1^2^3^4^5^6 Station (m) 4,--- 19/10/91 --*-- 26/5/92  7^8  9  x^ 25/5/93  Gully C5, XS-2  0  ^ 2^4^6 8^10 Station (m) x^ 25/5/93 4--- 16/11/91 --*-- 26/5/92  12  ▪• 195  4.8 ^  Gully C5, XS-3  4.6 4.4 g q  0  4.2 43.8  > 3.6 •  -  3.4 3.2 -  0  ^ ^ 6 7 1^2^3^4^5 Station (m) -I--- 16/11/91 --*-- 26/5/92^x^ 25/5/93  8  Gully C5, XS-4  ^ ^ ^ 1^2^3^4 5 6 7 Station (m) -4--- 16/11/91 * 26/5/92 25/5/93  196  Gully C5, XS-5  0  ^  ^ 1^2^3^4^5^6^7 8 Station (m) 4--- 16/11/91 ------ 26/5/92 ^x^ 25/5/93  Gully C5, XS-6  4.8 4.6_ 4.4 E g 4.2 -01 4) co ›^4w H ILI  3.8-  3.6  -  3.4 ^ 0  ^ ^ 1^2^3^4 5 6 Station (m) ^x^ 4--- 16/11/91 --,*-- 26/5/92 25/5/93  197 4.8 ^ 4.6  -  4.4  -  C6, XS-1  i 4.2 -43.8  -  3.6  -  3.4 3.2 3  ^ 4 0^1^2^3 Station (m) 4--- 3/10/91 -- w -- 25/5/92  5  ^  6  x^ 26/5/93  Gully C6, XS-2  3.2  1^1^ 11^, 0.0^0.5^1.0^1.5^2.0^2.5^3.0^3.5 Station (m) —I— 3/10/91 -- w - 25/5/92 ^x 26/5/93  4.0  198  Gully C6, XS-3  5 4.8  -  4.6  -  i — 4.4 g 0 4.2 o 0^4 H W 3.8  -  3.6  -  -  -  -  3.4^I^,^I 0.0^0.5^1.0^1.5^2.0^2.5^3.0 3.5 4.0 4.5 Station (m) x^ 26/5/93 ---4--- 3/10/91 --* - 25/5/92  5.0  199  Gully C10, XS-1  0.5^1.0^1.5^2.0^2.5 Station (m) 4---- 9/10/91 -- w - 7/5/92  3.0^3.5  4.0  x^ 26/5/93  Gully C10, XS-2  0.5^1.0^1.5^2.0^3.0^3.5 Station (m) ^x^ ---4,--- 9/10/91 ---w -- 7/5/92 26/5/93  200  Gully C10, XS-3  ^ ^ 3.0 4.0 0.5^1.0^1.5^2.0^2.5 Station (m) ^ w 7/5/92 26/5/93 -'--  4.6 ^ 4.4  -  Gully C10, XS-4  -  i 4.2g 0  4-  4)  > 1-1  3  3.6  -  3.4 0.0  ^ ^ 0.5^1.0^1.5^2.0^2.5^3.0^3.5 4.0 4.5 Station (m) ---+.-- 9/10/91 w 7/5/92^x^ 26/5/93 --  -  5.0  ^  201  Gully C11, XS-1  4.8 4.7 4.6  -  0 4.50  4.)  4.44.3  -  4.24.1 ^ 0.0^0.5  1.0^1.5^2.0^2.5 3.0 3.5 Station (m) 9/10/91 w 7/5/92^x^ 26/5/93 --  ^4.8 ^ 4.7  4.0  --  Gully Cii, XS-2  -  4.6 4.5  -  4.40  -d  4.3 -  > 0  4.2-  4) rd  4.1  -  43.9  -  3.8 ^ 0.0^0.5^1.0  ^ 1.5^2.0^2.5^3.0^3.5 4.0 Station (m) 9/10/91 - w 7/5/92^x^ 26/5/93 -  4.5  202  Gully C11, XS-3  ----1---  4.0 3.0 Station (m) 9/10/91 -- w -- 7/5/92^x^ 26/5/93  Gully C11, XS-4  4.8 4.6  6.0  -  3.8 3.6^■^ 0.0^1.0^2.0 f  ---- --  4.0 3.0 5.0 Station (m) 9/10/91 --w 7/5/92^x^ 26/5/93  6.0  Summary of sediment deposition volumes (cubic metres) at the traps  Sediment accumulated at traps from previous measurement date Date (dd/mm/yr) ^ ^ Gully Cl C2 C3 C4 C5 C6 C8 C10 Cli  27/10/90 31/10/90 17/11/90 0.024 0.040  2.369 0.051 0.010  0.004 0.000 0.114 0.005 0.000  2/12/90 29/3/91  0.112 0.026 0.620  0.105  0.234  7.583  11.503  3.960  0.000  22/05/91 8/07/91  0.000 0.060  8/8/91  0.778  0.000  Notes 1): Cl and C6 torrented Nov. 23/90. C3 full 17/11/90. 2): 3): C2 29/3/91 screen full, probably losing sediment by 2/12/90. 4): Installation of all new bars at site C5 27/5/91. 5): C3 excavated 03/06/91. 6): New installation at C2 2/8/91. 7): New deposition bars in C6 7/8/91. 8): C2 8/8/91 measurement actually 12/8/91. 9): C3 and C4 28/8/91 measurement actually 29/8/91. 10): C4 4/9/91 measurement actually 6/9/91. 11): C8 4/9/91 measurement actually 23/9/91.  0.020 0.019 0.000 0.240  28/8/91  0.176 -0.008 -0.001 0.322  4/9/91  1.147 0.452 -0.001 0.848 1.209 0.000  16/11/91  0.465  0.410 0.284  Summary of sediment deposition volumes (cubic metres) at the traps continued.  Gully Cl C2 C3 C4 C5 C6 C8 C10 C11  Sediment accumulated at traps from previous measurement date ^ Date (dd/mm/yr) ^ 21/11/91 8/12/91^21/12/91 11/1/92^25/1/92 1/2/92 26/5/92 4/9/92 8/2/92  0.000^-0.021^0.004 0.216^0.003^-0.004 0.002 0.327^0.000^0.009^0.062^0.232 0.135^0.022^0.005^0.051^0.144 0.020^0.000^0.000 0.019^0.361 2.710^0.230  0.378 0.000 0.000  0.026 0.000 0.044 0.017  0.136 0.002 0.801 0.004 0.034 0.350 0.315  0.050 0.000 0.024 0.000 0.000  18/11/92 25/5/93  0.033 0.005 0.000  0.000 0.108 0.000  0.150 0.241 0.000 1.479 0.183 0.000 1.498 0.160  Notes 12): 13): 14): 15): 16): 17): 18):  C5 21/12/91 includes 16/12/91 measurement of .0013. C2 deposition area destroyed 11/1/92. C8 measurement 26/5/92 actually 7/5/92. C11 installed 2/11/91. C10 installed 2/12/91. C11 sediment measured Jan 11, ^1992 = 2.7 m3, half est. to be mulch. Using accumulation of 1.3 4/9/92 includes 17/9/92 measurements; Gully C5 until 19/6 only.  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items