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 COLUMBIAbyTHOMAS HUGH MILLARDB.Sc.(Honours), The University of British Columbia, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(Department of Geography)We accept this thesis as conformingto the required standard THE UNIVERSITY OF BRITISH COLUMBIAOctober 1993© Thomas Hugh Millard, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of ^GeographyThe University of British ColumbiaVancouver, CanadaDate^October 14,1993DE-6 (2/88)i iABSTRACTThis study examines sediment storage and transfers in gulliesof coastal British Columbia, and how logging affects sedimentstorage and transfers. Both fluvial and debris flow transportof sediment occur in gullies, and the amount of fluvialtransport of sediment which occurs will affect the magnitude ofa subsequent debris flow. Coarse woody debris (CWD) may affectthe storage and transfer of sediment in the gully channel, andlogging can affect the supply and type of CWD.To determine whether logging affects storage and transfer ofsediment in gullies, sediment budgets were constructed forgullies 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 thechannel, and output from the channel estimated. Significantdifferences between treatment types occurred, summarized below.Treatment classes grouped together in brackets did not havesignificant differences.iiiBudgettermGreatest LeastInput Torrented Slash-full (Unlogged and Slash-clear)Storage Torrented Unlogged Slash-clearOutput Slash-clear (Torrented and Unlogged) Slash-fullOne objective of the study was to assess the effectiveness andfeasibility of cleaning slash from the gully channels. To beeffective, cleaning slash must either reduce the magnitude of adebris flow in a treated gully, or else reduce the likelihoodof initiation of a debris flow in the treated gully. Removal ofslash will reduce the volume of a subsequent debris flow byabout 15 percent, simply from the reduction in the amount ofCWD. Reduction in sediment stored in the treated channel mayreduce the volume of a debris flow by a further 4 percent.There is no evidence that removing slash will decrease thelikelihood of initiation of a debris flow.ivTABLE OF CONTENTSAbstract^  iiTable of Contents^  ivList of Tables  viList of Figures^  xiiiAcknowledgements  xChapter 1. Introduction^  11.1 Thesis statement and objectives^ 4Chapter 2. Gully morphology and processes  62.1 Gully morphology^  62.2 Sediment transport processes^ 102.2.1 Headwall and sidewall processes^ 102.2.2 Channel processes Debris flow initiation^ CWD and channel processes^ 182.3 Summary^  20Chapter 3. Study Area  213.1 Location and topography^  213.2 Bedrock and surficial geology^ 243.3 Climate and hydrology  253.4 Vegetation and forestry activities^ 293.5 Description of gullies^  32Chapter 4. Study design and measurement program^ 394.1 Study design^  394.2 Gully sediment budget estimates  424.2.1 Input measurements^  464.2.1.1 Surface erosion  464.2.1.2 Slump erosion  484.2.2 Storage measurement  484.2.3 Output measurement^  49Chapter 5. Results^  535.1 Input  585.1.1 Surface erosion^  585.1.1.1 Comparison of rainsplash pinresponse between gullies^ 615.1.1.2 Rainsplash pin erosion ratesfor individual gullies^ 665.1.1.3 Sidewall area for calculationof rainsplash erosion^ 665.1.1.4 Calculation of rainsplashinput volume^  695.1.2 Input volume from slumping^ 715.1.3 Total sediment input  735.25.3Sediment storage in gully channels^5.2.1 Errors in cross-section measurements ^5.2.2 Variation in storage change betweentreatment groups^5.2.3 Storage volume changes in channels^Output^V78808084875.3.1 Coarse sediment output^ 875.3.1.1 Gully morphology scaling factor 885.3.1.2 Variation of sediment outputbetween treatment groups^ 915.3.1.3 Volume of coarsesediment output^ 985.3.2 Fine sediment output 1005.3.2.1 Method^ 1005.3.2.2 Results 1025.3.3 Sediment transport distances^ 1125.3.4 Precipitation and sediment output^ 1155.3.4.1 Method^ 1155.3.4.2 Results 1185.4 Debris flows^ 1255.4.1 Coquitlam precipitation^ 1265.4.2 Cl debris flow^ 1345.4.3 C6 debris flow 1365.5 Sediment budgets 1395.6 Summary^ 144Chapter 6. Discussion and conclusion^ 1496.1 Budget accuracy^ 1496.1.1 Unlogged gullies 1496.1.2 Torrented gullies 1506.1.3 Slash-full gullies^ 1536.1.4 Slash-clear gullies 1536.1.5 Budget error summary 1546.2 Sediment storage and transfers in gullies^ 1556.2.1 Input^ 1556.2.2 Storage 1586.2.3 Output 1636.2.4 Fluvial and debris flow transportof sediment^ 1646.3 Transferability of results to larger gullies 1666.4 Study results: implications forforest^management 1686.4.1 Slash-clearing effects ondebris flows magnitude^ 1686.4.2 Slash-clearing effects ondebris flow frequency 1706.4.3 Assessment of effectiveness ofslash clearing^ 1716.5 Conclusion^ 172References^ 176Appendix 181viLIST OF TABLESTable 3.1 Gully dimensions^ 33Table 3.2 Sidewall materials 35Table 3.3 Channel materials^ 36Table 5.1 Date of installations 55Table 5.2 Rainsplash pin distribution statistics^ 60Table 5.3 Eroding rainsplash pin response^ 67Table 5.4 Sidewall areas^ 68Table 5.5 Volume of surface erosion^ 70Table 5.6 Error in surface erosion volume^ 72Table 5.7 Slump volumes for Coquitlam gullies 73Table 5.8 Total sediment input^ 75Table 5.9 Sediment input by treatment groups^ 77Table 5.10 Storage volume changes in channels 85Table 5.11 Sidewall and channel factors^ 89Table 5.12 Gully-scale-factor^ 90Table 5.13 Scaled sediment output by time period^ 92Table 5.14 Volume of coarse sediment output 99Table 5.15 Example of trap weight reconstruction^ 102Table 5.16 Sediment samples, gullies C3 and C5 103Table 5.17 Fine sediment retained at trap,gullies C3 and C5^ 108Table 5.18 Volume of fine sediment output^ 111Table 5.19 Comparison of C6 precipitation withCoquitlam Lake and Coquitlam River stations^ 117Table 5.20 Estimated C6 precipitation = coefficient *(River station) + constant^ 118TableTable5.215.22viiSlope failures in GVWD basins, 1990-1991^ 126Maximum one-day precipitation totals and returnperiods, November, 1990^  128Table 5.23 Antecedent temperatures and precipitation,November, 1990^ 132Table 5.24 Cl debris flow volume, November 23, 1990^ 136Table 5.25 C6 debris flow volume, November 23, 1990^ 139Table 5.26 Sediment budgets^ 141Table 5.27 Total measurement error^ 143Table 6.1 Old and revised input terms for sediment budgets. 156Table 6.2 Old and revised treatment group means^ 156Table 6.3 Ranking of sediment budget terms bytreatment group^ 175viiiLIST OF FIGURESFigure 2.1 Valley sideslope^  7Figure 2.2 Gully morphology  9Figure 3.1 Location of study area^  22Figure 3.2 Cedar Creek study area  23Figure 3.3 Annual precipitation, Coquitlam Lake andCoquitlam River stations, 1985-1992^ 26Figure 3.4 Mean monthly precipitation, Coquitlam Lakeand Coquitlam River stations, 1985-1992^ 26Figure 3.5 Maximum daily precipitation, Coquitlam Lakestation, 1924-1992^  28Figure 3.6 Daily maximum precipitation and return periodfor Coquitlam Lake station, 1924-1992^ 28Figure 3.7 Daily precipitation events greater than80 mm, for Coquitlam Lake and CoquitlamRiver stations,1985-1992  30Figure 3.8 Partial duration series, Coquitlam Lake andCoquitlam River stations, 1985-1992^ 31Figure 3.9^Channel slash in Gully C6^  38Figure 4.1 Conceptual model of sediment storageand transfers^  41Figure 4.2^Sediment input, storage, and output fora gully channel  44Figure 4.3^Sediment traps^  51Figure 5.1 Filled sediment trap^  56Figure 5.2 Storm precipitation greater than 80 mm,Years 1, 2, and 3  57Figure 5.3 Histograms of rainsplash pins,Years 1, 2, and 3^  59Figure 5.4^Histograms of rainsplash pins, by gully,Years 2 and 3 combined  62Figure 5.5^Rainsplash pins, slope angle, and soil type ^ 65ixFigure 5.6 Channel margin slump in gully C2^ 74Figure 5.7 Active and inactive cross-section zones^ 79Figure 5.8 Active zone width and area change^ 82Figure 5.9^Cross-section area change, Years 2 and 3^ 83Figure 5.10 Cumulative sediment output, unlogged,torrented, and slash-clear gullies^ 93Figure 5.11 Cumulative sediment output, torrented andunlogged gullies^  96Figure 5.12 Cumulative sediment output, torrented andslash-clear gullies  97Figure 5.13 Sidewall grain-size distributions^ 105Figure 5.14 Sidewall and trap grain-size distributions^ 107Figure 5.15 Channel sediment transport distances^ 113Figure 5.16 Maximum storm precipitation andsediment output^  119Figure 5.17 Threshold precipitation requiredfor sediment output  122Figure 5.18 Fluvial and debris flow precipitationthresholds^  124Figure 5.19 Location of Coquitlam Basin slope failures^ 127Figure 5.20 Maximum precipitation, Coquitlam Lakeand River stations, November, 1990^ 130Figure 5.21 Debris flow causing precipitation  133Figure 5.22 Cl and C6 debris flows^  135Figure 5.23 Channel slash above debris flow zone,gully C6^  138Figure 5.24 Budget error vs pooled error^  145Figure 6.1 Channel-margin storage zones in gully C5^ 152Figure 6.2^Sediment storage in gully C6^  160Figure 6.3 Old trees on bedrock in gully C2  162xACKNOWLEDGEMENTSThis 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 wasalways available when I needed his help. His efforts ensuredthe success of the project. My second reader, and actingsupervisor in my first year, was Dr. Michael Church. Asalways, he provided appropriate and useful guidance.Funding was provided by the Fish/Forestry Interaction Program.I would like to thank Steve Chatwin for his organizationalefforts, and Dan Hogan for all his help.Field work for this project was usually strenuous, andfrequently in unpleasant, and sometimes treacherous, weatherconditions. Many thanks are due to Scott Babakaiff, BrendaGriffiths, Sue Young, Scott Davidson, John Matechuk, MarianOden, Lars Uunila, and Craig Nistor. Not all the time wasspent scrambling up and down hillsides; many interesting anduseful discussions added to my knowledge and ideas for theproject.The Greater Vancouver Water District assisted in the project.I would particularly like to thank Derek Bonin for his helpwith project organization, as well as financial assistance forcleaning of the slash-clear gullies. Thanks also to the GVWDcrew led by Roland Totsauer, and the crew of Hans Lee Timber.xiBrian Fast of B.C. Hydro provided data from the Coquitlam Lakeand Coquitlam River stations.Terry Cooper provided his artistic and linguistic talents whenmine did not suffice; hence, I would like to thank him aswell.Finally, I would like to thank Brenda Griffiths, for all herhelp and support, as I thought out loud, to understand what Iwas doing, and why.1CHAPTER 1 INTRODUCTION AND OBJECTIVESGullies are an important component in the sediment transfersystem in the mountainous coastal region of British Columbia,but their dynamics are not well understood. Hillslope failureswhich either enter into a gully or occur within a gully maydevelop into a debris flow (also known as debris torrent),which in turn may result in extensive damage to salmonidstreams downstream of the gully. Sediment delivered to thegully between debris flow events may be stored or transportedthrough the gully, which may affect subsequent debris flows. Animportant control of sediment storage and transport is theamount of coarse woody debris (CWD, also called large organicdebris, LOD). Logging affects the amount and type of CWD in agully, therefore, greater understanding of how CWD affectssediment storage and transport in gullies is required.How trees and logging affect sediment storage and transfers ingullies are not well understood. Gullies combine features ofboth hillslopes and channels, hence their behavior can becomplex. Logging in and around gullies removes the source oflarge trees, and at the same time increases the amount ofsmaller woody debris. Often cable yarding systems are locatedin line with gullies, to take advantage of the greaterclearance between ground level and the cable. This can resultin large amounts of slash (woody debris remaining after logsare removed) deposited in the gully. Increased volumes of slash2may be responsible for initiation of debris flows (Swanston andSwanson, 1976, Sauder and Wellburn, 1987). Sidewalls of gulliesmay be important sources of sediment, and are affected by bothrooting strength reduction after logging, and yardingdisturbance. Hence the amount of sediment delivered to a gullychannel can be affected by logging, as can the movement ofsediment as it interacts with the logging slash introduced intothe channel.The frequency of debris flows and other mass movements usuallyincreases notably in steepland areas following logging.O'Loughlin (1972) reported a 2.3 fold increase in slopefailures in clearcut areas (roads excluded), compared withsimilar unlogged terrain. Young (1992) measured slope failurerates for two areas in the Queen Charlotte Islands (QCI). Slopefailures in clearcut areas increased 1.8 and 4.6 times theunlogged areas (roads excluded). Young found the increase ingully related failures to be greater than the increase in openslope failures. Rood (1984), in a much broader study, reportedan overall increase in slope failures of 30 times in clearcutterrain compared to unlogged terrain in the QCI (again, roadsexcluded). The volume of material entering streams from debrisflows increased by a factor of 69 when clearcut areas werecompared with unlogged areas. The average size of debris flowsentering streams more than doubled in clearcut areas, and thefrequency of debris flows increased by a factor of 27 times.Debris flows accounted for 67% of the volume of sedimententering streams from clearcut areas.3Sediment delivered from gullies into higher order streams canhave important effects on salmonid habitat. Salmonids requirestable and clean gravel for spawning; once fry are hatched,they require a variety of habitats to survive both summer andwinter stream conditions (Tripp and Poulin, 1986a, 1986b). Aprimary component of stream habitat is CWD, which providesdiverse stream morphologies and cover for juvenile fish. Debrisflows which originate on hillslopes, and can scour streamchannels for long distances, can have severe impacts on bothgravel and CWD in the stream. In some cases spawning gravel isalmost completely removed with the debris flow; in others theproportion of fine sediment becomes deleterious to egg survival(Tripp and Poulin, 1986b). Debris flows reduce the amount ofCWD in the stream channel (Tripp and Poulin, 1986a). Hogan(1986) found reduced pool-riffle spacings, altered pool andriffle heights, and smaller pieces of CWD in debris torrentedstreams. Streams subject to debris flows had poor salmonid eggand juvenile overwinter survival due to gravel scour and winterhabitat loss (Tripp and Poulin, 1992). Thus debris flows are animportant factor in determining channel morphology and fishhabitat.1.1 Thesis statement and objectivesSteep gullies of coastal British Columbia are important linksin the sediment transfer system. Storage and transport of4sediment may be affected by the amount of CWD in the gully,which in turn is affected by logging. This study proposes tomonitor the rates of sediment movement in gullies, and toexamine how CWD and logging affect sediment storage andtransfers.The specific objectives of this study are:1. To monitor sediment movement and storage in gullies,including logged, logged and debris torrented, andunlogged gullies.2. To assess how logging, and logging slash, affectsediment movement and storage in gullies.3. To assess the effects and feasibility of removinglogging slash from gullies.Gullies are selected for treatment groups based on logginghistory, debris flow history, and the amount of CWD or slash inthe gully. To determine the effect of slash within loggedgullies, one treatment group will be composed of logged gullieswith all slash removed after falling and yarding of timber iscompleted. How slash-clearing affects sediment storage andtransport has not been investigated.A sediment budget, composed of input, storage, and output, isconstructed for each gully. Sediment budgets are useful sinceeach term within the budget defines the sediment transfer intoor out of a specific morphological component of the gully.Thus, specific effects of slash or slash-clearing can be5determined. In addition, the sediment budget provides aseparate method of checking measurement errors, sincemeasurement error will result in an unbalanced budget.To determine differences between treatment groups, individualsediment budget terms will be compared. Specific hypotheses tobe tested will be:1. Sediment input is equal in all gully treatment groups.2. Sediment storage is equal in all gully treatmentgroups.3. Sediment output is equal in all gully treatment groups.Results of this study should increase our understanding of howsediment is stored and transferred in gullies, and how loggingaffects gully behaviour.6CHAPTER 2. GULLY MORPHOLOGY AND PROCESSESGullies are composed of distinct morphological units whichstore sediment; within each unit, there exists a set ofprocesses which mobilize and transport sediment. Since thisstudy examines the storage and transport of sediment withingullies, it is necessary to define the morphological units of agully, and to describe the important processes which occurwithin each unit.2.1 Gully morphologyGullies in British Columbia often occur on oversteepenedsideslopes 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 fingertiptributaries' as first-order channels; later researchersextended the network to include zero order drainages or hollows(Dietrich and Dunne, 1978, Dietrich et al, 1987). Zero-orderbasins do not have channels, by definition, but are concave inshape, concentrate subsurface water and are subject to massmovements. In most cases gullies are first-order (or sometimeshigher), but in others more closely resemble zero-order basins.This variation in form may be due in part to the amount ofsediment discharged to the gully, and the relative amounts ofsurface and subsurface flow. Gullies are usually a few hundred7FIGURE 2.1 Gullies on a valley sideslopeGullies visible on photo left. The cutblock boundary is locatedalong a very deep gully incised into bedrock. A second gully,not as deep as the other, is located just to the right of thegully on the cutblock boundary. Note that each gully has few,if any, tributary basins.8metres in length, and from less than 1 hectare to severalhectares in area.Gullies can be divided into distinct morphological units(Figure 2.2): headwall, sidewalls, and channel; these units aregenerally easy to define in the field. Gullies originate at theheadwall, generally a steep triangular failure plane scallopedinto the hillside. Headwalls are generally steeper than 30degrees. Beneath the headwall, the gully is composed of achannel and adjacent sidewalls. The sidewalls and channelgenerally form a U, V, or trapezoidal cross-section. Sidewallsrange from shallow, with slopes less than 15 ° , to steep, withslopes greater than 50 ° . Channels range in slope from about 10 °to over 35 °The channel is the longitudinal axis of the gully, and is thelowest elevation on a cross-section perpendicular to thelongitudinal axis. The channel slope is defined as the slope ofthe longitudinal axis of the gully. Although described as achannel, there may or may not be banks, fluvial sediment perse, or frequent surface flows of water. The channel both storesand transports sediment and CWD. Sediment from the sidewalls orheadwall is delivered to the channel, then the sediment may betransported variable distances down the length of the channel,depending on channel slope and morphology, CWD, and waterdischarge.FIGURE 2.2 Gully morphology91 0The sidewalls are bounded at their upper edge by the openhillslope and at their lower edge by the channel (Figure 2.2).The sidewall slope is defined as the angle from the channeledge to the top of the sidewall, orthogonal to the gully axis.Sidewalls do not concentrate as much water as the channel. As aresult, sediment movement processes which occur on sidewallsare generally different from those in the channel. Someprocesses, such as debris flows, are primarily a channelphenomenon, but affect the sidewalls as well.2.2 Sediment transport processesThe type of processes which occur in a gully will depend on thebedrock, type and thickness of glacial drift or colluvium,gully plan form, water discharge, and biological effects suchas tree root strength and CWD. The morphology of the gullystrongly controls the location of processes. The headwall andsidewalls are primarily the location for debris slides ordebris avalanches, together with minor mass movements such asravelling or small slumps. The channel is dominated by fluvialtransport of sediment and debris flows.2.2.1 Headwall and sidewall processesDebris slides are the most significant process to which theheadwall and sidewalls are subjected. Debris slides are shallow11planar movements of unconsolidated and usually unsaturatedsediment (Varnes, 1978). Headwall debris slide volume averages900 m 3 , and sidewall debris slides average 380 m 3 , in the QueenCharlotte Islands (Rood, 1984). Analysis of debris slideinitiation usually adopts an infinite slope model for stabilityanalysis. This method assumes a uniform thickness of soil witha well defined shear plane, and then analyses the slope forresisting forces (shear strength) and driving forces (shearstress). The ratio of shear strength to shear stress is calledthe Factor of Safety (F.S); failure occurs or is imminent whenthe F.S. equals 1. The equation can be expressed as:F.S. = C a  + (af  z cos 2B - A )tan 0^(shear strength)af z cosB sinB^ (shear stress)"where Ca = apparent cohesion (kPa), a f = unit weight of soilat field moisture (kN/m3 ), z = vertical thickness of soilmantle (m), B = slope angle (degrees), 0 = internal angle offriction (degrees), [and] A = pore-water pressure at thefailure surface (kPa)" (Sidle and Swanston, 1982).Using the infinite slope model, slope stability decreases witha decrease in apparent cohesion, an increase in slope angle, anincrease in pore pressure, and a decrease in the internal angleof friction. Sidewall and headwall slopes are sensitive tofailure since they are often very steep, and a gullyconcentrates water, resulting in high pore pressures.12Logging can affect slope stability in several ways. Removal oftrees immediately reduces soil weight and wind stress (Brownand Sheu, 1975). In forest soils, the apparent cohesion termincludes rooting strength, which can be a significant factor inshear strength. O'Loughlin (1972) reports a root strength valueof 71% of shear strength for saturated till soil on a slope of35 degrees. Once trees are cut, roots rot and lose strength. Asnew vegetation grows, root strength increases. Reported periodsin which soils are most susceptible to failure are variable.Sidle et al (1985) report a range of from 3 to 10 years aftercutting. Rollerson (1992) considers the period from 6 to 15years most susceptible. Falling and yarding of trees maydisturb both surface and subsurface soil conditions (Sauder etal, 1987). After swales (zero order basins) in a northernCalifornia watershed were logged, macro-pore (pipe) dischargeincreased 3.7 times over expected discharge (Ziemer, 1992). Inaddition, logging appears to contribute to pipe collapse andincreased sediment discharge from these pipes. Interruption ofmacro-pore networks may result in locally increased porepressures and subsequent slope failure (Ziemer, 1992).Other processes common on sidewalls are dry ravel, rainsplasherosion, frost heave and small slumps. Dry ravel is themovement of individual clasts or small groups of clasts,(Sauder et al, 1987). Rainsplash erosion and frost heave arecommon on exposed ground, but surface vegetation preventsrainsplash erosion and appears to minimize frost heave. Smallslumps, of less than a few cubic metres, are similar to debris13slides, but may have more complex failure surfaces, from planarto rotational.2.2.2 Channel processesTwo processes are important within a gully channel: fluvialtransport of sediment and debris flows. These two processes canact in sequence, with low magnitude, frequent fluvial eventspunctuated by high magnitude, infrequent debris flows. Theproportion of sediment delivered to the channel that istransported out of the gully by fluvial transport is animportant control on debris flow magnitude.Both zero-order basins and first-order channels can exhibit apattern of deposition of sediment over a long period of timefollowed by a complete evacuation of sediment by debris flow(Dietrich and Dunne, 1978, Benda and Dunne, 1987). A debrisflow usually scours the channel to bedrock, after which a newcycle of sediment and CWD recharge occurs. As sediment enters anewly scoured channel, the rate of fluvial transport ofsediment is greatest, since all water discharge is surfaceflow. As the depth of sediment increases in the channel, theproportion of subsurface flow increases, and less fluvialtransport of sediment occurs (Dietrich and Dunne, 1978; Bovisand Dagg, 1987).The rate of sediment recharge will depend upon the supply of14sediment to the channel, the proportion of material too coarseto be transported by fluvial events, and the amount of sedimenttrapped by CWD. Sediment sources for gullies can have extremelyheterogeneous grain sizes, including large boulders. Sedimentsupplied to the channel may be subject to selective fluvialtransport of finer material, both on the surface and in voidspaces between large grains (Bovis and Dagg, 1988). The resultis a coarse lag deposit, with much finer sediment subsurface. Debris flow initiationDebris flows can be defined as a gravitational movement ofsolids with interstitial fluid, where the relative velocitiesbetween fluid and solid are not significant (Takahashi, 1981).Debris torrents are a type of debris flow, where "rapidmovement 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 loggingand slope movement literature, at least in the WesternCordillera, but there is no mechanical difference betweendebris flows and debris torrents (Church and Miles, 1987).Accordingly, the more general term "debris flow" will be usedhere.Debris flows are complex events in that their initiation andmovement depend on a large number of geologic and climaticfactors. Takahashi (1981) cites three main causes of debris15flow: 1) a landslide enters a channel, becomes saturated andturns into a debris flow, 2) a natural dam blocks a gullychannel, eventually collapsing to release a debris flow, and 3)channel sediment becomes unstable and decouples from thestreambed as a debris flow when sufficient water dischargeoccurs. Takahashi states that mobilization of channel debris isthe most common mechanism of debris flow initiation, butevidence indicates that the first or second cause is probablymuch more common in coastal British Columbia. Rood (1990)reported the location for debris flow initiation sites in theQueen 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% fromunknown sources - unknown largely because of dense forest cover(Rood, 1992). In clear-cut terrain, 65% originated as headwallfailures, 16% from sidewall failures, and 12% from locationsoutside the gully. A study of landslides on southwest VancouverIsland shows less than 2% of debris flows in gullies initiatedin the channel (Rollerson, 1984, and 1993). A study fromcoastal Oregon reported debris flows originate in zero-orderbasins, not in first-order channels (Benda, 1990).Not all slope failures which enter a gully necessarily producea 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) observeda much higher ratio of debris flows to slides if only upper-slope locations are considered. Failures originating from16hollows located at the channel head, and entering the channelat an angle less than 45 degrees, initiate debris flows;failures from hollows entering the lower portions of thechannel at a 90 degree angle do not initiate debris flows(Benda and Dunne, 1987). These observations suggest that basingeometry is an important control on debris flow initiation.Other factors affecting debris flow initiation are the size ofthe 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 mostcommonly provided by rainfall or snowmelt, or a combination ofboth. Caine (1980) defined a minimum rainfall intensity curvefor shallow landslides and debris flows, using published datafrom many areas of the world. The threshold for slope failureis 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 specificallyto debris flows; his minimum value is about 25 mm of rain in 24hours. Two curves for slope failures on the Queen CharlotteIslands have been developed, for wet antecedent conditions anddry antecedent conditions. The minimum 24-hour precipitationvalue for wet antecedent conditions is 20 mm; for dryantecedent conditions the minimum value is 30 mm in 24 hours(Hogan and Schwab, 1991a). It should be emphasized that allthese values are minima; in some cases rainfall which producedslope failures was much greater. In addition, my choice of17reporting 24 hour precipitations is somewhat arbitrary, sincedifferent intensities and durations of rainfall are known toinitiate debris flows.Church and Miles (1987) emphasize the importance of parametersdifficult to measure: locally intense precipitation andsnowmelt. The steep mountain fronts and narrow valleys ofsouthwest British Columbia provide effective barriers for airmass movements, forcing the air aloft. Convection cells mayresult, with subsequent heavy precipitation, particularly athigher elevations. For example, yearly precipitation atHollyburn Ridge (elevation 951 m) is more than double that atthe nearby Point Atkinson station (elevation 9 m). Hence thesparse precipitation gauge network, usually located in valleys,may not adequately measure precipitation which actually occursat debris flow sites. Snowmelt in response to risingtemperatures and warm rain may provide significant inputs ofwater to soils. The amount of snow present in a basin can varywith elevation, hence estimating water input from snowmelt isvery difficult, even if snow is known to be present. Church andMiles summarize local debris flows being generated by: "(1)locally concentrated rainfall, high antecedent moisture, nosnowmelt...(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 producedebris flows, very remarkable events are more likely to producedebris flows. CWD and channel processesThe channel is the location where CWD accumulates. Naturalsources of CWD are trees or tree fragments, usually derivedfrom windthrow or sidewall failures. Effects of CWD in steepchannels have not been well investigated. Energy dissipationand sediment trapping are two CWD functions observed(Froehlich, 1973). In stream channels, CWD significantlyaffects morphology. Pool and riffle spacing and size, bankfeatures, and sediment storage are all affected by CWD (Hogan,1986). Log jams, typically from debris flows, are able to traplarge volumes of sediment, and control channel morphology bothupstream and downstream for significant distances (Hogan andSchwab, 1991b). CWD in gullies may act in different ways fromstream channels. Since gully walls are steep and narrow, talltrees falling into a gully will probably be suspended over thechannel on at least one end, unless the tree happens to fallparallel to the channel axis. Smaller fragments, such as brokentrunks, stumps, or branches, are more likely to lie across thechannel, an effective position for trapping sediment. Inaddition to their role in supplying CWD, standing trees mayalso act as barriers to sediment or CWD movement.Logging will reduce the long-term supply of large CWD availablefor a gully, but at the same time will dramatically increasethe amount of smaller CWD introduced to the gully. Slashmeasured in three headwater channels in western Oregon (averagesize, 70 hectares) averaged 0.4 m 3 /m of channel (Froehlich,191973). Total CWD in these headwater channels increased 2.5 foldafter falling and yarding; fine CWD (0.3-10 cm diameter)increased 4.5 fold. Clearly, logging increases the total volumeof CWD, and in particular, increases the amount of smaller CWD.Although these results are for only one area, similar resultscan be expected in forests of coastal British Columbia, sincetree species and logging methods are similar.Forestry management literature often recommends minimizing theintroduction of slash to gully channels, or removal of slashafter yarding (Froehlich, 1973; Chatwin, 1991; British ColumbiaMinistry of Forests, 1992). Channel slash has been cited as acause of debris flows, primarily from debris jams which failduring high flows and initiate a debris flow (Swanston andSwanson, 1976; Krag et al, 1986; Swanston and Howes, 1991).Only Swanston and Swanson (1976) observed debris flows in whichCWD contributed to the initiation of the event; however, howslash affected the initiation was not stated.2.3 SummaryGullies act as both a source and conduit for sediment transportand are an important link in the sediment transfer system. Theheadwall and sidewalls are source areas for sediment deliveredto the channel. Sediment introduced to the channel may bestored temporarily, and can be subject to frequent fluvialevents and less frequent debris flows. CWD affects storage of20sediment, and hence may influence the magnitude and frequencyof debris flows. Logging changes the type and volume of CWD inthe channel, thus it may also affect the timing and volume ofdebris flows. Hillslope failures and debris flows usually occurduring times of high rainfall or snowmelt (or combination), butno simple relationship between debris flow occurrence andprecipitation is observed.21CHAPTER 3. STUDY AREA3.1 Location and topographyThe study area is located in Coast Mountains approximately 30km northeast of Vancouver (Figure 3.1), within CoquitlamWatershed. The Watershed is managed by the Greater VancouverWater 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 CedarCreek, on a ridge which ranges in elevation from about 600 m tomore than 1200 m (Figure 3.2). The Branch 200 road is locatedalong the bottom of the hillslope; Branch 230 is a switchbackspur road, and reaches mid-slope positions.Bedrock is close to the surface on the steep middle and upperportions of the hillslope. Numerous gullies are incised intothe hillslope. Lower portions of the hillslope have colluvialfan or cone shaped debris deposits emanating from many of thesegullies. At the apex of the fan, the gully is usually wellexpressed whereas lower portions of the fan may show little, ifany, expression of a gully or channel. In most cases, a clearlydefined channel does not extend from the hillslope to thechannel of Cedar Creek. The colluvial fans are thereforeinfluent flow zones in that water discharge is often not22FIGURE 3.1 Location of study areaGully^ LegendContours (200 foot interval)Stream (surface flow)Stream (subsurface flow)Cl CutblockL 89^Logged, 19_Roads0^metres^1000FIGURE 3,2 Cedar Creek study area24evident in lower sections of a gully, but is visible well aboveits fan.3.2 Bedrock and surficial geologyBedrock in Cedar Creek is mostly gabbro, with some quartzdiorite areas and small sulfide exposures (Roddick, 1965).Bedrock is generally massive and resistant to erosion; however,within recently torrented gullies bands of finely fracturedbedrock trending NE-SW are evident. Faults visible on aerialphotographs show a similar trend.Basal till overlies bedrock in much of the study area. The tillis consolidated and dense; the Branch 230 roadcut shows almostvertical till exposures over 5 m high. In other areas bedrockis 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, arevisible within the roadcut exposures.Colluvial material often overlies the bedrock or till. Thesource of this material is either bedrock or till deliveredfrom upslope, or else in situ weathering of the till. Depthsare typically less than 2 m, except in the fans at the base ofgullies, where depths may be much greater.253.3 Climate and hydrologyThe south Coast Mountains have a cool temperate climate, butare strongly affected by orography. In Vancouver, near sea-level, mean annual precipitation is about 1400 mm. On the NorthShore Mountains, 3500 mm of precipitation is recorded annuallyat elevations of 1200 m. Low elevations within the mountainsmay receive comparable precipitation totals, as a result oftopographic confinement of air. In winter the freezing level isoften between sea-level and the mountains tops. Snow fallsabout 80 days of the year at the crest of North ShoreMountains, and in Vancouver about 15 days per year (Hay andOke, 1973; Wright and Trenholm, 1969). These ranges inprecipitation totals, temperature and snowfall are similar tothose experienced in nearby Coquitlam Basin.A weather station has operated at the Coquitlam Lake Dam since1924. The station operated as an Atmospheric EnvironmentService station until 1982. Since then, B.C. Hydro has operatedthe station, collecting hourly information, although the dataare incomplete until 1985. The mean annual precipitation forthis station from 1924 to 1992 was 3490 mm. B.C. Hydro alsoestablished a second weather station within the valley, onCoquitlam River above Coquitlam Lake (Figure 3.1). It has beenoperating since 1984, with essentially complete data since1985. The total annual precipitation from both stations for theperiod 1985-1992 is shown in Figure 3.3. Monthly meanprecipitation is shown in Figure 3.4. The Coquitlam River40003500 -3000 -Z 2500-20001500 -H0 1000 -tat 500 -01985 1986 1987 1988 1989 1990 1991 1992^AVG500 -400 -300 -200 -100600•14. AI14I ••• •04•26FIGURE 3.3 Annual precipitationCoquitalm Basin stations, 1985-1992YEARgo Lake stationElev. = 161 mRiver stationElev. = 280 mAL A:FIGURE 3.4 Mean monthly precipitationCoquitlam Basin stations, 1985-1992JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DECMONTHLake station V V ■■• River stationElev. 161 m^Elev 260 m27station has the greater total precipitation in all years, aswell as consistently higher monthly means. This may be relatedto elevation, since the River station is at an elevation of 280m, compared to the Lake station at 161 m. Alternatively,constriction of air as the valley narrows in its headwaters mayaccount for the greater precipitation at the River station.Since gully sediment transport is associated mainly with shortperiods of intense precipitation, the frequency of largeprecipitation events is of interest. The maximum dailyprecipitation for Coquitlam Lake station for 1924-1992 is shownin Figure 3.5. Figure 3.6 shows the return period and magnituderelationship for these data. The largest event occurred in1989; the second largest in 1990, the year this study began.The maximum event in 1991 is the same as the mean for theperiod, 135 mm, and the maximum daily precipitation for 1992 is116, lower than the mean annual maximum. Therefore, 1990 was anexceptional year, and 1991 and 1992 were average and belowaverage, respectively.Large precipitation events may occur more than once a year; sopartial-duration series (De Ploey et al, 1991) have beenconstructed for both B.C. Hydro stations for the period of1985-1992. The minimum value chosen for the partial durationseries is based on a 24-hour maximum precipitation of about 90mm, which can be identified as a threshold for sedimenttransport (Section 5.3.4). Since the partial-duration seriesare constructed using daily totals, rather than 24 hour maxima,I^I^I1970^1980^1990 200019601920 1930 1950194028FIGURE 3.5 Maximum daily precipitationCoquitlam Lake station, 1924-1992YEARFIGURE 3.6 Maximum daily precipitation and return periodCoquitlam Lake station, 1924-19921^ 10^ 100RETURN PERIOD (years)29a correction factor of 1.13 is applied to the maximum 24 hourprecipitation (Linsley et al. 1975), resulting in a total dailyvalue of about 80 mm. Results for the 2 stations, grouped byyear, are shown in Figure 3.7. Figure 3.8 shows the partialduration series for both stations, based on the same data setsas Figure 3.7.Results are similar for the two stations, although theCoquitlam River station has more frequent large events. Onaverage, 3.1 significant events (greater than 80 mm) in a yearcan be expected at the Lake station, and 5.7 significant eventsat the River station. The 1989 and 1990 events at the Lakestation exceed all other events at both stations.Large events can occur at any time of the year, but autumn andwinter account for the majority of large events. November hasthe greatest number of significant events (44% of Lake stationevents, 24% of River station events), and October to Februaryhas 81% of the events at both stations combined.3.4 Vegetation and forestry activitiesThe study area lies within the Submontane Wetter MaritimeCoastal Western Hemlock Biogeoclimatic zone (GVWD Forest covermaps, Slaymaker et al, 1992). Dominant tree species within thestudy area are western hemlock (Tsuga heterophylla), westernredcedar (Thuja plicata), grand fir (Abies grandis) and Douglasl0O —^801985 1986^1987^1988^1989 1990^1991^1992El=••••.30Figure 3.7 Daily precipitation events greater than 80 mm, 1985-19923.7a Coquitlam Lake station3.7b Coquitlam River station280260— 240220O• 200 -• 180I-4fal 160140• 120YEAR31FIGURE 3.8 Partial duration series for daily precipitationCoquitlam Basin stations, 1985-1992^280^^260^240-0 220-200-180-160-140-1 20-100-800t^I^i^l^t t^t^1^t^1^till1 10RETURN PERIOD (years)I t^t^1^Itti100X Lake station^+ River station32fir (Pseudotsuga menziesii). Higher elevations have mountainhemlock (Tsuga mertensiana).Logging in the study area began in 1977. Figure 3.2 indicatesdates when individual patches were logged. High-lead yardingwas used in all of the earlier logging. In the most recentlylogged area, where gullies C10 and C11 are located, a Wyssenskyline system was used. All areas are replanted within twoyears of logging. Regeneration is well established in almostall areas, except the most recently logged areas, and in somegullies which have had slope failures or debris flows.3.5 Description of gulliesThe locations of gullies are shown in Figure 3.2. Gullies varyin length, primarily dependent upon how close to the top of theridge they begin. Since gully headwalls are located in forestedterrain, the location of the exact start of each gully is notusually observed on aerial photographs. Some gullies weresurveyed to the top of their headwalls, but this was notpossible in all gullies. As a result, the highest reaches of agully may not be accurately represented on the map.Gullies were surveyed for channel and sidewall slope, sidewalllength, channel material, and sidewall vegetation cover. Table3.1 summarizes the spatial dimensions of each gully. Wherepossible, channel length is the surveyed distance from theTable 3.1 Gully dimensionsGully^Channel Mean^Mean^Mean^Totaland^Length^Channel Sidewall Sidewall^Area2Type 1^Slope^Length^Slope(n) (degrees) (m)^(degrees) (ha)C1-U^330s 28.5 7.4^16.4 0.30C2-U^350e 27.2 9.6^26.4 0.59C3-T^900m 19.9 10.5^23.8 1.63C5-T^500m 24.1 17.0^38.1 1.13C6-SF,T^260e 32.0 7.2^28.1 0.30C4-SF^140s 29.3 12.0^19.5 0.29C8-SF^195s 33.7 6.4^20.8 0.21C10-SC^250m 30.5 6.4^28.9 0.24CU-SC^500m 29.9 7.6^31.6 0.641 U: unlogged; T:clear.torrented; SF: slash-full; and SC: slash-2 Gully area approximate, measured by average width (topsidewall to top sidewall) and channel length.s Surveyed distance.e Surveyed distance, plus estimated distance.m Distance measured from Figure 3.2.3334sediment trap to the headwall of the gully. Some gullies extendalmost to the top of the ridge and their headwalls are visiblein aerial photographs. Channel lengths for these gullies aremeasured from Figure 3.2. If total channel length was notsurveyed, and location of the gully headwall is not visible inaerial photographs, an estimate of total length, based onsurveyed distance, is used. Gully C3 has two gullies jointogether; channel length for C3 is the combined length. Thetorrented gullies C3 and C5 are the largest gullies selected;the slash-full and slash-clear gullies are generally thesmallest. The largest gullies tend to have the lowest channelslopes, but sidewall slope is not associated with gully size.Sediment traps for gullies C3 and C5 are both located on thefan; this is one reason why the larger gullies have lowerchannel slopes.Tables 3.2 and 3.3 summarize the sidewall and channel materialsin each gully. Gully cross-profiles were surveyed at 10-30 mdistances along each longitudinal profile. At each station, thesidewall and channel materials were noted. Materials noted wereclassified as bedrock, bare sediment, vegetation, and slash.Sidewall sediment is generally very heterogeneous, with grainsizes ranging from clay to boulders. Channel sediment istypically cobbles and boulders, with some finer sediment.Proportions of each material at each station were estimated at100%, 85%, 70%, 50%, 30%, 15%, or 0%. The amount of sidewall orchannel covered by each material is the cumulative proportionestimated at each station.Table 3.2 Sidewall materialsGully Bedrock^Sediment Vegetation Slash 1Type (%)^(%) (%) (%)Cl-U 0^0 100 0C2-U 0 2 98 0C3-T 0^19 71 10C5-T 2 0 58 42 0C6-T 44^33 15 8C4-SF 0 6 30 64C6-SF 0^16 0 84C8-SF 0 10 12 78C10-SC 3^0 3 94C11-SC 0 2 33 651 Slash includes naturally introduced CWD.2 Vegetation includes seeded grasses and legumes.3536Table 3.3GullyChannel materials.Bedrock^Sediment Vegetation SlashlType (96) (%) (%) (%)Cl-U 0 89 0 11C2-U 52 42 0 6C3-T 0 97 0 3C5-T 0 100 0 0C6-T 90 5 5 0C4-SF 0 0 0 100C6-SF 33 2 0 65C8-SF 0 8 2 90C10-SC2 78 22 0 0C11-SC? 0 83 0 171 Slash includes naturally introduced CWD.2 Surveyed in May, 1992; slash clearing occurred the previousautumn.37Sidewall materials vary according to gully group. The unloggedgullies have almost completely vegetated sidewalls. Slash-fulland slash-clear sidewalls are mostly covered in slash, withsome vegetation and smaller bare areas. Torrented gullies varyin the nature of their sidewall materials. C3 was logged in1977, and since then, growth of planted conifers has vegetatedmost of the sidewalls. C5 is partly covered in seeded grasses,but also has bare soil. Gully C6 has mostly bedrock and baresoil, but C6 is located higher on the hillslope than either C3or C5; consequently bedrock is closer to the surface. Thesurveyed portions of C3 and C5 are in colluvial fans, wherebedrock is not close to the surface.Channel materials are similar in all gullies except for slash-full gullies. Unlogged, torrented, and slash-clear channels areeither bedrock, sediment, or a combination of both. The gullyC10 channel was primarily bedrock when surveyed, however afterslash-clearing in the previous autumn, the channel was mostlysediment. Slash-full channels are clearly dominated by slash(Figure 3.9), although some bedrock or sediment may be present.38Figure 3.9^Channel slash in Gully C6Vertical view of the gully. Stadia rod visible in center bottomof photo; numbers on rod are in decimeters. Photo scale changesdrastically from top to bottom.39CHAPTER 4. STUDY DESIGN AND MEASUREMENT PROGRAM4.1 Study designThe most general objective of this study is to understand howsediment and CWD are stored and mobilized in gullies. The morespecific objective of this study is to understand how loggingslash affects the storage and mobilization of sediment. Fourtreatment 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 slashconditions in the channel. Comparison of Group A with Group Dwill show how adding slash to a gully affects sediment storageand transport, relative to the natural behaviour of an unloggedgully. Comparison of Group A with Group B will show how theabsence of slash affects sediment storage and transport, andwill allow an assessment of the role of intentional CWD removalfollowing harvest, as a factor in sediment management in loggedareas. Gullies in both Group B and Group C do not have much40slash or CWD in their channels; comparison of these two groupswill show how similar slash-cleared gullies are to torrentedgullies.Secondary differences between treatment groups concernvariations in logging and debris flow history. Groups A and Bare almost completely equivalent, with the removal of slashfrom the channel being the most important difference. Somesecondary effects on sediment movement may occur from theplacement of slash on sidewalls of Group B gullies, and fromdifferences in yarding methods. Group C has been logged, likeGroups A and B, but in addition has had a debris flow occursince logging. This implies changes to the amount of sedimentstored in the channel, as well as sidewall conditions. Group Dis not logged: the sidewalls as well as the channel are intheir natural states. For these reasons, differences insidewall and channel sediment and vegetation conditions, notrelated to slash or CWD, may affect behaviour of the treatmentclasses.Within this experimental framework, the main focus of study isthe monitoring of sediment mobilization and transport, as wellas an assessment of which changes have resulted from treatmenteffects, particularly the presence or absence of logging slashin the channel of the gully. Sidewalls are treated separatelyfrom channels, since sidewalls in different treatment groupsmay respond differently, but not as a result of the primarytreatment effect.414.2 Gully sediment budget estimatesA sediment budget defines the input, change in storage, andoutput for an individual sediment storage element. If the gullychannel is defined as the storage element, a sediment budgetwill define the amount of sediment entering the channel, thechange in channel storage, and the output of sediment from thechannel. The quantification of these terms for each gullypermits testing of the specific hypothesis, stated in Chapter1, to determine differences in treatment groups."A sediment budget for a drainage basin is a quantitativestatement of the rates of production, transport, and dischargeof detritus." (Dietrich et al, 1982). Storage elements withinthe drainage basin are identified, and the rates of input andoutput resulting from transport processes are measured. Thesediment budget equation can be expressed as (Roberts andChurch, 1986):I -()S = 0where I is the volume of sediment input, CSS is the volumechange in storage (an increase is positive), and 0 is theoutput. All terms are for an identical period of time. Eachterm refers to a specific storage element, with the output fromone 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 conceptualmodel of storage elements and sediment transfers in a gully.TransferprocessesStoragesitesBedrock TillDebris slidesDebris flowsFrostactionChannel andBank deposits42.Figure 4.1 Conceptual model of sediment storage and transfers in gullies43Although all sediment transfers should be identified, apractical sediment budget will focus on the largest sedimenttransfers. A complete sediment budget includes dissolvedmaterial transfers (Dietrich and Dunne, 1978), but dissolutionof sediment is probably not important in the short term. Thisis especially true of intrusive rock, typical of the studysite. Consequently sediment can be considered unchanged, giventhe three year period reported in this study. Similarly, creepwill be considered minor over the period considered.Since the channel is the primary element of interest, thesediment budget will be constructed with reference to thechannel as storage element (Figure 4.2). Input of sediment tothe channel is from sidewalls of the study reach; erosion pinsmonitor sediment input from sidewall sources. Input from thechannel upstream is not monitored, but will be shown to beminor. Change in storage is defined as an increase or decreasein channel sediment within the study reach; storage change ismonitored using cross-sections. Output is defined as the volumeof 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 andaccurately measured, the budget should balance. In practise,each budget term has error associated with it. The error foreach term can be estimated, and the pooled error is calculated(the square root of the sum of all terms squared). If thebudget balances within the pooled error, the budget can beUpstreamInputOpen slopeErosion pins ---Open slopeMeasuredCross-sectionTrapDeposition bar/Channel OutputFIGURE 4.2 Sediment input, storage, and output for a gully channel4445considered to be an adequate representation of sedimentmovement and storage within the gully. If the budget imbalanceis greater than the pooled error, either the model does notsufficiently represent the actual storage elements andtransport processes, or else measurement procedures are notsufficiently accurate.Since construction of a sediment budget requires a balancing ofthe input, storage, and output terms, the length of the gully,above the sediment trap, which contributes to these terms mustbe estimated. Sediment input to the channel is subject totransport in high rainfall or snowmelt events (or both). If allother factors remain equal, the distance that sediment istransported in the channel should be an inverse function of thecalibre of sediment introduced to the channel; thus an accuratemodel of sediment input into a gully would have input zoneswhich vary with sediment size. Without further detailed study,the length of gully which contributes sediment must be somewhatarbitrarily set.In general, the cross-sections monitor storage changes upstreamof the trap for distances of 50 to 100 m. The upper limit ofthe cross-sections often coincides with a change in the gullymorphology or vegetation, and provides a convenient break todefine the study section of the gully, as well as the upperlimit for sediment input from the sidewalls. If the uppercross-section is not a significant location, the greatestlength of gully study section is set at 100 m.464.2.1 Input measurementsInput of sediment is defined as eroded sidewall sediment whichenters the gully channel. Although sediment from the upstreamchannel may enter the channel study reach, this term could notbe measured without disturbing the sediment regime of the studyreach. If input from the channel above the study reach is animportant term, then the storage and output terms will totalmore than the sidewall input. It will be shown that since inputis not under-estimated, input from the channel above the studyreach may be ignored. Two main types of sidewall erosion arerecognized: slumps and surface erosion. These are combined togive a total input of sediment to the channel. Surface erosionSurface erosion is a result of rainsplash erosion, ravelling,or frost heave, and can occur over broad areas of exposedmineral soil on gully sidewalls. The widespread occurrence ofsurface erosion requires a monitoring program which samples thedepth of sidewall retreat at several locations, then calculatestotal surface erosion volume as the product of average retreatand the total area of the eroding sidewall. The amount ofsurface erosion is monitored with erosion pins, and the amountof eroding area is estimated during gully surveys.Erosion pins are metal spikes 250 mm long and 6 mm thick,47hammered into the ground perpendicular to the surface.Approximately 50 mm of pin is left exposed, and each pin isidentified with a numbered plastic tag. Pins are placed ingroups of 8 to 16 individuals, spaced approximately 0.5 mapart, with two rows extending across the sidewall slope. Pinsusually are located on the lower portions of the sidewallslope, near the channel, so as to prevent excessive disturbanceof bare gully sideslopes. In some cases these sites are theonly accessible locations.Pin exposure above the ground is always measured on thedownhill side of the pin. Repeated measurement of the exposedportion of the downhill side of the pins establishes the amountof erosion or deposition over the specified measurement period.If a pin has fallen out of the ground, the previous measurementof the pin length is used to calculate the minimum amount oferosion required to cause the pin to fall.In some sets of pins, several may be found lying on the groundat one measurement time. If these pins exhibit a clustereddistribution, the erosion of these pins is treated as a slump,rather than erosion typical of the entire sidewall. As aresult, the erosion pins are separated into two groups: slumppins and the rainsplash pins. The slump pin volume is estimatedby multiplying the slump area by slump depth, as determined bythe minimum amount of erosion required to cause the pin tofall. Individual pins which have fallen out are treated asanomalies and ignored for calculation of erosion, unless there48is clear evidence of significant erosion at those sites. Somepossible reasons for pins falling out are animal disturbance,or falling wood or rocks dislodging the pin. Slump erosionSlumps in Coquitlam gullies are generally small shallowfailures which frequently occur near the channel margins.Individual slumps are identified during monitoring of erosionpins and cross-sections. Each slump is measured for width,depth, and length, and a volume total is calculated. Mostslumps are less than 1 m3 in volume.4.2.2 Storage measurementStorage of sediment in a gully channel is monitored at measuredcross-sections. Cross-sections were installed during the summerand fall of 1991, with 3 to 5 cross-sections in each gully,usually less than 10 m apart. Cross-sections consist of a metalstake 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 stationsis measured. Repeated survey of each station reveals changes inthe ground elevation. Each change in elevation is multiplied byhalf the distance between the two adjacent measurementstations. Total area change for a cross-section is the sum of49individual station area changes. If the cross-section stakesare located high on the gully sidewalls, some stations aredeleted from the summation, since these stations do notrepresent changes in storage within the channel.To calculate the change in sediment volume in the gully, theareal storage change is multiplied by a representative lengthof gully channel. Each gully is surveyed, and the position ofthe cross-sections noted. If the channel is consistent in slopeand sediment size, the length of gully represented by aspecific cross-section is half of the length between the twoadjacent cross-sections. Cross-sections at the top or bottom ofthe gully have a representative length equal to the distancebetween the cross-section and the adjacent cross-section. Somegullies have discrete storage areas, separated by lengths ofbedrock channel. In these cases, the length of channelrepresented by each cross-section is simply the total length ofthe storage zone sampled by the individual cross-section. Onceeach cross-section has a storage volume change calculated, thevolume change for the entire gully is then the sum of theindividual storage changes.4.2.3 Output measurementCoarse sediment output is monitored using traps positionedacross the channel, at the lower end of the gully study reach.Traps are of two types, but are similar in function. Both50screen and culvert traps create a pool of water in the gullychannel in which coarse sediment is able to settle (Figure4.3). Sediment which remains suspended in water as it passesthrough the pool will be lost. Hence an estimate is required ofthe amount of fine sediment carried past the trap. Section5.3.2 contains details of how fine sediment output isestimated.Screen traps use a geotextile material, Amoco Siltstop, as aporous dam to create a local impoundment of water (Figure4.3a). The Siltstop is 1 m in height, and is supported byreinforcing bars driven into the ground approximately 0.5 mdeep. Approximately 20 cm of the bottom of the screen materialis anchored by large rocks. Siltstop is a fairly permeablematerial, and it was originally expected that all waterdelivered to the trap would pass through the screen, trappingthe sediment behind. In practise the permeability of the screenrapidly decreases as fine sediment and algae clog the pores.Rock spillways are constructed at the screen edges to permitpassage of water. The best location for a screen trap isbetween large boulders, which prevent erosion of the channelbanks as water is discharged around the sides of the screen.Culvert traps are simply the basin excavated on the uphill, orditch side of the logging roads above a culvert (Figure 4.3b).The basins are excavated for the purpose of trapping sedimentto prevent clogging of the culvert. Culvert traps have greater51FIGURE 4.3^Geotextile screen and culvert sediment trapsa) Screen trapb) Culvert trap52capacity to trap sediment than the screen traps, however, therecan be additional input of sediment from ditch sources.Both types of traps use short sections of reinforcing rod tomonitor the amount of sediment deposited. "Deposition" bars arehammered into the trap basin vertically, with between 0.5 m and1 m of bar exposed. Between three and twelve bars are installedin a trap, depending on the size of trap. Repeated measurementsof the height of the exposed section of each rod records thedepth of sediment deposited.The locations of the deposition bars and the boundaries of thetrap basin are mapped using triangulation. The depositionalarea corresponding to each bar is calculated using Theissenpolygons. The volume of sediment deposited in each polygon isthe product of the depth of deposition and the area of thepolygon. Total trap deposition volume is simply the sum of thepolygon volumes.53CHAPTER 5 RESULTSInstallations were monitored beginning in the summer of 1990,and continued to May, 1993. Partial gully monitoring systemswere installed in the first year and by December 1991, thecomplete network was established. Slash-clearing of gulliesoccurred in September and October of 1991. Slash-clearingoccurred a year later than planned, due to a logging fatalityand subsequent logging moratorium. Since the autumn and winterperiods are the time of greatest storms and sediment movement,yearly periods run from spring to spring. Data were collectedfor three years:Year 1: July, 1990 to May 22, 1991Year 2: May 23, 1991 to May 27, 1992Year 3: May 28, 1992 to May 26, 1993The study design called for 3 replicates in each treatmentgroup (unlogged, logged and slash-full, logged and slash-clear,and logged and torrented). Site features, debris flows, andlogistic problems resulted in most groups having less thanthree replicates. Only two appropriate old-growth gulliesexisted at the study site at the start of the experiment.Originally, a third gully (C7) was located in the unloggedarea, but the channel had been subject to recent massmovements, and consequently, was not loaded with CWD. TheNovember 23, 1990 storm resulted in a debris flow in gully Cl,54which left only one unlogged gully. Similarly, three slash-fullgullies were monitored at the beginning of the period, but C6torrented at the same time as Cl, resulting in only two slash-full gullies over most of the study period. Removal of slashfrom gullies was expensive, and was limited to two gullies. Thestudy began with three torrented gullies, C3, C5, and C9. GullyC9 had torrented before logging occurred in 1990-1991; yardingin 1991 resulted in increased disturbance and slash loads toC9. As a result, C9 had to be abandoned, and C6 replaced it asthe third torrented gully.Table 5.1 shows the installation dates of sediment traps,cross-sections, and erosion pins in each gully. Sediment trapsdid not necessarily operate continuously after installation,since large events often either filled or destroyed them(Figure 5.1). However, cross-sections and erosion pins operatedcontinuously after their installation.Sediment movement can be expected to be related to the numberand size of large precipitation events. Figure 5.2 shows allstorms with daily precipitation greater than 80 mm in eachyear. Year 1 clearly had much larger events than Years 2 and 3.Year 3 had very few large events, and Year 2 had several largeevents, particularly at the Coquitlam River station. Therefore,if sediment movement is driven by the number and size ofprecipitation events, Year 1 should have the greatest sedimentmovement, and Year 3 the least. Since sediment traps were55Table 5.1 Date of installation (day/month/year)Gully Sediment Cross- ErosionTrap Sections PinsCl-U 17/7/90 none nonelC2-U 17/7/90 13/6/91 2/8/91C3-T 26/7/90 12/6/91 26/7/90C4-SF 19/7/90 13/6/912 8/8/91C5-T 19/7/90 12/6/91 26/7/90C6-SF,T 23/7/90 7/8/91 26/7/90C8-SF 6/10/90 none2 7/8/91C10-SC 2/12/91 2/12/91 none 3Cll-SC 2/11/91 2/11/91 none31) Cl monitoring discontinued after debris flow 23/11/90.2) C4 and C8: channel slash prevents effective cross-sectionmonitoring.3) C10 and C11: sidewall slash prevents installation of erosionpins.56Figure 5.1 Filled sediment trap, gully C3Sediment trap at gully C3 filled after November 9-11, 1990storm. Volume of sediment retained in trap was 0.58 m 3 . Screenhas been pulled away from center support. Stadia rod (1.1 m) inleft-center of photo.260—E 240 -220 -0 200 --H180 -4--)Q, 160 --H0 140-a)P4 120 -›-, 100 --H0 80 -60260—0 240 -2200 200 --H0 180 -4-)160--H140 -C-14 120 ->1 100 --H80-60=57Figure 5.2 Storm precipitation greater than 80 mm, Years 1, 2, and 35.2a Coquitlam Lake station0^1^2^3^4Year5.2b Coquitlam River station0^1^2^ 4Year58monitored frequently in Years 1 and 2, sediment output responseto individual storms will be examined (Section 5.3.4).5.1 Input5.1.1 Surface erosionRainsplash pins are compared to examine their response in eachmeasurement year. Histograms of all rainsplash pins for eachyear are shown in Figure 5.3. Distribution statistics for therainsplash pins are given in Table 5.2. Pins generallyregistered a small net deposition in Year 1, despite thelargest storms having occurred in this year. Little differenceis observed in pin responses between Years 2 and 3; both Year 2and Year 3 have greater rates of erosion than Year 1, the yearwith the greatest storms. Erosion pins can act as barriers todownward movement of sediment, and hence the first year mayrepresent a period during which the slope equilibrated to thepresence of the pins. Pins are probably in equilibrium with theslope in the second year after installation. Since thedistributions of pin erosion tend to be negatively skewed,median values will be used for determining differences inrainsplash pin response.Figure 5.3 Histograms of rainsplash pins, Years 1, 2, and 3c^,---5.3b Year 230-125-105 -85 -65 -45 -25 -5 15 35 55 7559Net change (mm/yr; erosion is negative)Table 5.2 Rainsplash pin distribution statistics60Yearnmean (mm) 1median (mm) 1standard dev. (mm)skewnesskurtosis1 Year 2 Year 377 121 1090.4 -12.3 -8.31 -5 -425.1 25.3 26.2-0.4 -1.1 -1.20.7 1.7 4.5number eroding2^35^71^69percent eroding^45.5 58.7 63.3mean erosion (mm) 2^-19.2^-26.8^-20.4median erosion (mm) 2 -12 -18 -131 Positive values are deposition, negative values are erosion.2 Number of pins eroding is those pins which have negativevalues. The mean erosion and median erosion are based onnegative values only.615.1.1.1 Comparison of rainsplash pin response between gulliesRainsplash pin response varies widely between gullies. Figure5.4 shows the distribution of net change for each gully. InFigure 5.4, net change is combined for Year 2 and Year 3, whenall erosion pin sets were operational.All sets of pins were installed in bare soil areas, except formost of the C2 pins. C2 pins installed in vegetated slopes showan average change of +2 mm, that is, net deposition. No spatialpattern of deposition results exists on these sidewalls. Sinceno 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 pinson bare ground. The greater value in this study may be a resultof the vegetation surface, which compresses a variable amountwhen the pin is measured, or else a result of checking pins forfrost heave. Frost heave partially ejected some pins from theslope after cold weather periods. These pins could be easilypushed back into the hole which was created when the pin wasejected. All pins were checked for frost heave by pushingslightly on them, which, if the ground is soft can increase theamount of pin inserted into the ground.All pins installed in slash-full and torrented gullies arelocated on bare ground. Rainsplash pins in slash-full andtorrented gullies show similar responses. The modal class forall gullies is 0 to -20 mm. Some gullies have negatively skewed50 -(C2-8 I40 -6030 -20 -10 -kN N - 130 - 110 -90 -70 -50 -30 -10 10 30 50 70In = 48-130-110 -90 -70 -50 -30 -10 10 30 50 70Figure 5.4 Histograms of rainsplash pins, by gully, Years 2 and 3626050403020100NET CHANGE (mm/yr; Erosion is negative)Figure 5.4 continued63I(IIIEiIfic64distributions (C3-T, C4-SF, and C6-T); the other gullies (C5-T,C8-SF) have positively skewed distributions. There is noevident separation between distributions of slash-full andtorrented gullies. Since bare soil areas may have severalorigins (debris torrent, sidewall failure, loggingdisturbance), the presence of some bare soil does notnecessarily correlate with treatment type.Since all pin installations are in bare ground (except for theC2 pins), rainsplash pin response at different sites should besimilar. Differences in response may be caused by either slopeangle 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 bysoil type and slope angle. C horizon pins tend to be located onsteeper slopes, but the range in erosion rates is similar. Thegreatest erosion rate is for the set of B horizon pins on thesteepest slope, but this group is composed of only 4 pins. Ifthese pins are excluded, no trend in the B horizon pins isevident. The C horizon pins appear to have a trend, but slopesless than 45 degrees would have significant deposition if thedata were extrapolated to include less steep slopes. Either atrend for C horizon pins is valid only for slopes steeper than45 degrees, or else the trend is spurious.65Figure 5.5 Rainsplash pinsSlope angle and soil type10CC C0B CBBB T CB CCCBB-50 -B-6020I^ I^ I^ I^ I30 40 50 60 70Slope angle (degrees)B= B Horizon, C= C Horizon, T= Till80665.1.1.2 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. Sincesome volume of sediment is almost certainly eroded fromsidewalls through surface erosion, the median rainsplash pinresponse is not an adequate measure of erosion. To calculatethe volume of sediment input from surface erosion, only therainsplash pins which show a negative (eroding) net yearlychange will be used. The total bare area is adjusted by theproportion of rainsplash pins that are eroding. Table 5.3 showsthe median erosion rate (only negative values used), the totalnumber of pins in the gully, the number of pins with negativevalues, and the percentage of pins which register net erosion. Sidewall areas for calculation of rainsplash erosionTo calculate rainsplash input volume, the study reach sidewallarea which is eroding must first be estimated. Total studyreach sidewall area was measured during survey of the gully,and at the same time the amount of bare area was estimated. Thebare area total for each gully is multiplied by the fractionalproportion of rainsplash pins showing net erosion (Table 5.3)to obtain an eroding area for each gully. During surveying, thebare area was visually estimated as a percentage of thesidewall during the survey of the gully. Each survey segment(10 - 30 m) has the amounts of bare area and vegetation types67Table 5.3 Eroding rainsplash pin response.Gully^Median^Eroding^Total^FractionErosion^Pins Pins^ErodingYear 1C3-TC5-T(mm)-16.0-4.071516270.440.56C6-SF -23.5 12 30 0.40Year 2C2-U -1.0 2 23 0.09C3-T -5.0 7 14 0.50C5-T -18.5 20 27 0.74C6-T -29.0 27 30 0.90C4-SF -13.5 10 16 0.62C8-SF -7.0 4 10 0.40Year 3C2-U -5.0 14 24 0.58C3-T -16.5 6 14 0.43C5-T -14.0 6 15 0.40C6-T -21.0 26 30 0.87C4-SF -15.5 12 16 0.75C8-SF -7.0 5 10 0.5068present estimated (0, 15, 30, 50, 70, 85, 100%). Error isestimated at 10%, about half the range between classes. Resultsare shown in Table 5.4.Gully C5 is partly vegetated in grasses and legumes. A factorof 0.5 has been applied to C5 total sidewall area to accountfor the presence of these grasses.Table 5.4 Sidewall areasGully^Total^Bare^PercentSidewall^Sidewall^BareArea (m2 )^Area (m2 )^AreaC2-U 2040 15 1C3-T 1910 90 5C5-T 3120 1560 50C6-T Y2 1130 450 40C4-SF 2570 210 8C6-SF Y1 1010 0 0C8-SF 1760 140 8C10-SC 670 0 0C11-SC 1500 0 0Gully type clearly has an effect on the amount of bare area.Neither slash-clear gully has any bare area. Average bare69area for slash-full gullies is 5%. Gully C3 has only 5% barearea, compared with 40% (C6-T) and 50% (C5-T) for the othertorrented gullies. Vegetation growth on the gully sidewallssince the debris flow event in C3 has resulted in a similaramount of bare area as in the slash-full gullies. In terms ofits sidewall erosion, gully C3 is more appropriately classifiedas a slash-full gully. Therefore, a distinction must be madebetween gullies which have torrented recently and those whichtorrented many years ago. Calculation of rainsplash input volumeThe bare area in each gully (Table 5.4) is multiplied by thepercentage of pins showing erosion (Table 5.3) to obtain theeroding area. The eroding area is then multiplied by the medianrainsplash pin erosion depth (Table 5.3) to calculate an inputvolume of sediment for each gully. Table 5.5 shows the totalbare area, the eroding area in each gully, the medianrainsplash erosion rate, and the estimated volume of sedimenteroded from the sidewalls of each gully. Torrented gullies C5and C6 have the greatest input, and slash-full gullies have thenext greatest input. The volume figures in Table 5.5 will beused 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 thearea of the bare sidewall. Maximum error would occur when bothTable 5.5 Volume of surface erosionGully Bare^Percent^Eroding Median ErosionArea^Pins^Area Erosion Volume(m2)^Eroding^(m2) Rate (mm) (m3)Year 1C5-T 1560^56^874 -4.0 -3.49C3-T 91^44^40 -16.0 -0.64C6-SF 29^48^14 -23.5 -0.33Year 2C2-U 17^9^1 -1.0 -0.00C5-T 1560^74^1160 -18.5 -21.4C6-T 450^90^403 -29.0 -11.7C3-T 90^50^46 -22.0 -1.00C4-SF 210^62^130 -13.5 -1.76C8-SF 141^40^56 -7.0 -0.39Year 3C2-U 17^58^10 -5.0 -0.05C5-T 1560^40^625 -14.0 -8.74C6-T 450^87^390 -21.0 -8.15C3-T 90^43^40 -16.5 -0.64C4-SF 210^75^158 -15.5 -2.44C8-SF 140^50^70 -7.0 -0.497071errors are in the same direction: that is, when both the areaand the erosion are overestimated or underestimated. Error inestimating area is 10%, and error in erosion pin measurement is+/- 2 mm. The error volume is the difference between theerosion volume (Table 5.5) and the volume obtained when botharea and erosion rate have been underestimated. Table 5.6 showsthe estimated errors. Largest errors are associated with thelargest eroding areas.5.1.2 Input volume from slumpingSlumps observed in the Coquitlam gullies were generally lessthan 0.5 m deep and a few square metres in area. Slumps wereinferred when groups of erosion pins had moved from theiroriginal locations, or from the appearance of a fresh scar onthe sidewall. Calculation of slump volume in pin array areasuses the pin array to determine the areal extent of the slump,with depth assumed to be 0.2 m unless the slump scar indicatesgreater 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 sedimentfrom slumping each year. Errors in slump measurements areestimated 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 torrentedgullies, apart from one slump in gully C2, and a minor one in72Table 5.6 Error in surface erosion volumeGully Eroding^Erosion^Minimum^Erosion^ErrorArea Less Rate less Volume^Volume^Volume10 %^(m2 ) 2 mm (m3) (m3) (m3)Year 1C5-T 788 -2.0 -1.58 -3.49 -1.91C3-T 36 -14.0 -0.50 -0.64 -0.14C6-SF 13 -21.5 -0.27 -0.33 -0.06Year 2C2-U 1.7 -0.0 -0.00 -0.00 -0.00C5-T 1040 -16.5 -17.2 -21.4 -4.22C6-T 363 -27.0 -9.80 -11.7 -1.90C3-T 41 -20.0 -0.82 -1.00 -0.18C4-SF 117 -11.5 -1.35 -1.76 -0.41C8-SF 51 -5.0 -0.25 -0.39 -0.14Year 3C2-U 9 -3.0 -0.03 -0.05 -0.02C5-T 562 -12.0 -6.74 -8.74 -2.00C6-T 349 -19.0 -6.64 -8.15 -1.51C3-T 35 -14.5 -0.51 -0.64 -0.13C4-SF 142 -13.5 -1.91 -2.44 -0.53C8-SF 63 -5.0 -0.32 -0.49 -0.1873gully C4. The C2 slump occurred along the channel bank, andalthough the input is to the channel, the location of the slumpis not strictly the sidewall (Figure 5.6). Of the torrentedgullies, C5 clearly has the most active sidewalls in terms ofslumping. It is worth noting that C5 sidewalls are the longestand steepest sidewalls of all the gullies.TableGully5.7 Slump volumesYear 1for Coquitlam gulliesYear 2^Year 3(m3 ) (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)C6-T 0 -0.45(0.17) -0.54(0.23)C4-SF 0 0 -0.07(0.04)C8-SF 0 0 0C10-SC n/a 0 0CIA-SC n/a 0 05.1.3 Total sediment inputThe total input of sediment to each gully is the sum of thesurface erosion input and slump inputs. Table 5.8 shows totalinput to each gully, and to allow comparison of gullies on an74Figure 5.6 Channel margin slump in gully C2Large log in upper left of photo defines the left bank of thechannel. Smaller log in center of photo had broken in 1990,exposing sediment which has slumped into channel. Stadia rod is1.1 m long.75Table 5.8GullyTotal sediment inputRainsplash^Slump^TotalInput^Input^InputSidewallAreaTotalNormalized(m3 ) (m3) (m3) (m2) Input (mm)Year 1C2 -U n/a 1.00 n/a 2040C5-T 3.49 2.45 5.94 3120 1.90C3-T 0.64 0.50 1.14 1910 0.60C4-SF n/a n/a n/a 2570C6-SF 0.27 0 0.27 1010 0.27C8-SF n/a 0 n/a 1760Year 2C2-U 0.0 0.03 0.03 2040 0.02C5-T 21.4 4.22 24.1 3120 7.71C6-T 11.7 0.45 12.1 1130 10.7C3-T 1.00 0.12 1.12 1910 0.59C4-SF 1.76 0.00 1.76 2570 0.69C8-SF 0.39 0.00 0.39 1760 0.22C10-SC 1 0 0 0 670 0C11-SC1 0 0 0 1500 076Table 5.8GullyContinuedRainsplashInputSlumpInputTotalInputSidewallAreaTotalNormalized(m3 ) (m3) (m3) (m2 ) Input (mm)Year 3C2-U 0.05 0.17 0.22 2040 0.11C5-T 8.74 0.07 8.80 3120 2.82C6-T 8.15 0.54 8.69 1130 7.69C3-T 0.64 0.28 0.92 1910 0.48C4-SF 2.44 0.07 2.51 2570 0.98C8-SF 0.49 0.00 0.49 1760 0.28C10-SC 1 0 0 0 670 0C11-SC 0 0 0 1500 01) Gullies C10 and C11 do not have erosion pins; zero input isbased on complete vegetation and logging slash cover of thesidewalls, and no observed slumps.77equal basis, the total sediment volume is normalized bydividing by the total sidewall area in each study reach.Sediment input for each treatment type is summarized in Table5.9. Gully C3 is included with the slash-full gullies since theamount of bare area, and input, are similar to those in gulliesC4 and C8.Table 5.9 Sediment input by treatment groupsTreatment^n^Mean^Variance(mm)^(mm2 )Unlogged 2 0.065 0.004Slash-full 7 0.546 0.067Torrented 5 6.18 13.7Slash-clear 4 0 0Clear differences exist between treatment types. Torrentedgullies have the greatest input of sediment, as would beexpected of gullies which have the greatest areas of baresidewalls. Slash-full gullies produce the next greatest amountof sediment; this is also reflected in the amount of bare area,since slash-full gullies have the greatest amount of bare areanext to the torrented gullies. The unlogged gully and theslash-clear gullies are similar. Since almost all the inputinto the unlogged gully came from a bank slump, the true amountof sidewall erosion approaches the amount of sidewall erosion78exhibited by the slash-clear gullies, that is, measurably zero.5.2 Sediment storage in gully channelsChanges in sediment storage within gullies are the result ofprocesses which affect the movement of sediment within thegully channel. During the study period major fluvial transportand debris flow events occurred. Unfortunately, cross-sectionswere not in place in Year 1 to measure changes when the largestevents occurred. The storms of November, 1990 affected channelsto an unknown degree. Cross-section monitoring during Year 2and Year 3 showed moderate changes in most gullies.Cross-sections are separated into two zones. The active zone isthe middle area of the cross-section, where fluvial transportof sediment occurs. The inactive zones are at the ends of thecross-section, above the channel area. Figure 5.7 shows theactive and inactive zones in a typical cross-section. Onlychanges within the active zone are considered, unless a majorevent caused clear change in the inactive zone. Since both Year2 and Year 3 did not have major storms, no significant changesoccurred in the inactive zone of any cross-section.4 A3.83.65TNACTIVE^ACTIVE4.84.6>C -ACTUAL CHANGEFigure 5.7Active and inactive cross-section zones0^1^2^3^4^5^6^7^8Station (m)79—I— OCT 19/91 X JAN 11/92805.2.1 Errors in cross-section measurementError in cross-section measurements is a significant problem.The error in an individual elevation measurement has beenequated to the D90 in stream channel surveys (Hogan, 1992). Insome of the gully cross-sections, this standard may result inan estimated error of decimetres. Since this error is so largeas to hide most instances of real change, an alternative methodof 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 (eitherapparent deposition or apparent erosion) should be 0.50. Givena station with change in one direction, the probability of twoadjacent stations recording change in the same direction is0.25, assuming no actual change has occurred. This standardwill be adopted to indicate whether or not real change hasoccurred: if three adjacent stations all show change in onedirection, then the change is taken as real. As furtherconfirmation, the cross-section profiles are examined todetermine whether the changes make "geomorphic" sense. Figure5.7 also shows actual change and measurement error.5.2.2 Variation in storage change between treatment groupsAll cross-sections were in place by November, 1991, and weremonitored regularly through to May 1993. Changes in cross-81section area are summarized for Years 2 and 3. Year 2 changesrefer to the period October, 1991 to May, 1992. Most changeoccurs during the autumn and winter seasons, when storm eventsare most frequent. Change in cross-sectional area is used sincedeposition or erosion tends to occur in discrete areas of across-section; as a result, no normalization based on cross-section width is necessary. Figure 5.8 shows there is norelationship between active zone width and amount of areachange. The greatest area changes are recorded for active zonesof 2-3 m width, not for cross-sections of greater width.Figure 5.9 shows the amount of area change at each cross-section in each gully for Years 2 and 3. Year 3 shows a smallbut consistent pattern of deposition in almost all cross-sections. Year 3 results for all gullies are not significantlydifferent (ANOVA test, probability of equal means = 0.87).Year 2 has a much greater range of changes, with greateramounts of both erosion and deposition, and changes appear tobe related to treatment type. An ANOVA test shows significantdifference in Year 2 (probability of equal means = 0.01).Torrented gullies C3 and C5 show small to moderate amounts ofdeposition in almost all cross-sections. Conversely, the slash-clear gullies show a moderate to large amount of erosion in allbut one of the cross-sections (C10, cross-section XS-3). By theend of Year 2, almost all of the channel of C10 had eroded tobedrock, which was close to the surface before treatment (i.e.slash removal). Cross-section 3 in C10 had a boulder roll into0 1^2^3^4^5^60.2 -0.1-(NI0tn -0.1 -g4-0.2 -v0P -0.3 --0.4--0.5--0.6XXX^4‘xX ++ x X^X^ x^*x X X++XFigure 5.8Active zone width and area change0.3Active zone width (m)82+ YEAR 2 X YEAR 3Figure 5.9 Cross-section area change, Years 2 and 35.9a Year 2830.40.3— 0.2N0.1= XS-3=ME0^0m0It4 -0.100 -0.2mIt0P -0.34)w -0.4-0.5-0.6=lirC2 - U C3 - T C5 - T^C6 - T C10 - SC Cll - SCGully5.9b Year 30.40.3(Ni^0.20.1W0)^0A(1:14 -0.10W -0.2mMP -0.30-4-) -0.4-0.5-0.6=C2 - U C3 - T C5 - T C6 - T C10 - SCC11 - SCGully84the cross-section, which caused sediment deposition around it.This effect is local, and bedrock is exposed within 2 m eitherside of the cross-section. Deposition shown at cross-section 3is anomalous for C10 as a whole. The unlogged gully, C2, has awide range of change for Year 2, as does C6, the most recentlytorrented gully. Overall average response for these gulliesfalls between the responses of the slash-clear gullies and thetorrented gullies C3 and C5.5.2.3 Storage volume changes in channelsThe change in volume of channel sediment is used in thecalculation of the sediment budget. To calculate the change involume of sediment stored in a gully, the change of volume ofsediment represented by all cross-sections in the gully issummed. Change in volume represented by a cross-section is theproduct of the areal change for the cross-section and thelength of channel represented by that cross-section. In mostcases, the length of the represented channel is half thedistance between adjacent cross-sections. In some cases therepresented channel length is less, if bedrock exposure orother evidence indicates lack of erosion or deposition. Table5.10 shows the volume change for each cross-section and thetotal volume change for each gully. Error in volume estimatesdepends upon how representative the cross-section change is forthe channel length, in addition to the measurement error forthe cross-section. Since storage change between cross-sections85Table 5.10 Storage volume changes in channelsGully & Channel Area Storage Change Volume ChangeX-section Length Year 2^Year 3 Year 2^Year 3(m) (m2)^(m2) (1113) (m3)C2-2 4.0 -0.03^0.03 -0.11 0.13C2-3 6.1 -0.34^0.02 -2.09 0.12C2-4 8.5 -0.03^0.07 -0.27 0.59C2-5 11.5 0.20^0.03 2.29 0.38C2 Unlogqed Total -0.18 1.20C3-1 10.5 0.15^-0.06 1.55 -0.68C3-2 11.9 0.04^0.07 0.45 0.83C3-3 7.2 0.05^-0.06 0.39 -0.45C3-4 5.8 0.16^0.14 0.94 0.81C3-5 6.6 -0.03^0.01 -0.18 0.07C3 Torrented Total 3.20 0.58C5-2 7.4 0.21^0.00 1.60 0.00C5-3 7.4 0.06^0.09 0.44 0.68C5-4 9.8 0.03^0.11 0.29 1.10C5-5 14.8 0.23^1.6 0.05 3.450.78C5-6 18.1 0.03^0.00 0.63 0.00C5 Torrented Total 6.40 2.6086Table 5.10 ContinuedGully &^ChannelX-section^LengthArea Storage Change^Volume ChangeYear 2^Year 3^Year 2^Year 3(m) (m2) (m2) (m3) (m3)C6-1^7.0 -0.18 0.00 -1.27 0.00C6-2^5.0 -0.23 0.09 -1.14 0.46C6-3^4.0 0.01 0.05 0.05 0.20C6 Torrented^Total -2.40 0.66C4 Slash-full^Total 1 1.76 2.44C8 Slash-full^Total l 0.39 0.49C10-1^17.0 -0.32 0.00 -5.40 0.00C10-2^7.8 -0.50 0.07 -3.90 0.55C10-3^4.0 0.11 0.00 0.46 0.00C10-4^10.4 -0.51 0.00 -5.32 0.00C10 Slash-clear^Total -14.2 0.55C11-1^15.1 -0.13 0.04 -1.99 0.42C11-2^7.4 -0.23 -0.06 -1.72 -0.43C11-3^9.8 -0.24 0.04 -2.37 0.44C11-4^16.7 -0.32 0.05 -5.39 0.88C11 Slash-clear^Total -11.5 1.301) Gullies C4 and C8, storage equal to input.87is unknown, error may be considerable, and is estimated at 50percent.Slash accumulation in C4 and C8 prevented effective cross-section monitoring. Since no sediment output was measured inthese gullies (Section 5.3), storage is assumed to be equal toinput.5.3 OutputThere are two types of sediment output: coarse sedimentmeasured at the sediment traps, and fine sediment output. Finesediment output is carried past the traps in suspension.Despite this shortcoming, the sediment traps provide aneffective measure of gully response to storms, andconsideration 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 determininghow much of the gully contributes to sediment output.5.3.1 Coarse sediment outputSediment traps were installed in July and August, 1990 atgullies Cl, C2, C3, C4, C5, C6, and C8. The C11 trap wasinstalled November 2, 1991, and C10 trap was installed December2, 1991. Some traps were subject to events which filled ordestroyed them, consequently not all traps record all output.88In particular, storm events in November, 1990 filled ordestroyed sediment traps in gullies C1, C2, C3, and C6. Gully morphology scaling factorComparison of sediment output between gullies must take accountof gully size, since output volume is expected to beproportional to gully size. Sediment yield studies usuallyexpress yield as an equivalent depth of sediment throughout thebasin (Schumm, 1977). However, in a set of gullies of equalbasin area, variation in morphological dimensions betweengullies may strongly influence the areas actually producingsediment. To compare sediment output, gully morphologicalparameters which might affect sediment supply andsusceptibility to movement are considered. Two gully facets areused: sidewalls and the channel. Sidewalls which are longer orsteeper should deliver more sediment, if all other factors areequal. Similarly, longer or steeper gully channels are alsomore likely to deliver sediment to the trap. A scale factor forsediment 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 sidewalland channel factors are scaled against the average of all gullysidewall or channel factors. The gully-scale-factor is then theaverage of these two factors. Table 5.12 shows the scaled89Table 5.11 Sidewall and channel factorsGully Channel Average Channel Average Average SidewallLength Channel Factor^Sidewall Sidewall FactorSlope^Length^Slope(rn) (deg.) (m) (deg.)Cl-U 136 28.7 65.3 7.4 16.3 2.08C2-U 104 27.2 47.5 9.6 26.4 4.27C3-T 190 19.9 46.3 17.0 24.0 4.27C4-SF 140 29.3 68.5 12.0 19.5 4.00C5-T 136 24.1 77.5 17.0 38.1 10.5C6-SF,T 126 32.0 66.7 7.2 28.1 3.39C8-SF 118 33.7 65.4 6.4 20.8 2.27C10-SC 53 30.5 26.9 6.4 29.2 3.12C11-SC 86 28.6 41.1 7.2 29.9 3.59Mean 56.2 4.16Table 5.12 Gully-scale-factorsGully^Scaled^Scaled^GullyChannel^Sidewall^ScaleFactor^Factor^Factor RankCl-U 1.16 0.50 0.83 7C2-U 0.85 1.02 0.94 4C3-T 0.82 1.03 0.92 5C4-SF 1.22 0.96 1.09 2C5-T 1.38 2.52 1.95 1C6-SF,T 1.19 0.81 1.00 3C8-SF 1.17 0.55 0.86 6C10-SC 0.48 0.75 0.61 9C11-SC 0.73 0.86 0.80 8Mean 1.00 1.00 1.009091sidewall and channel factors, and the average gully-scale-factor. The average gully-scale-factor serves to normalizesediment output between gullies. Variation in sediment output between treatment groups.Sediment output is calculated for each gully for a set ofperiods. Not all gullies have data from the same set ofperiods, since traps were not always working, and the slash-clear gullies were not instrumented until autumn 1991. Table5.13 lists the start and end dates for each period, and thescaled (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 C8produces measurable sediment only twice, out of twelvemeasurement periods. C8 output during these periods is justslightly greater than that attributable to error.The other three types of gullies (unlogged, torrented, andslash-cleared), all produce significant but variable amounts ofsediment. 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 gulliesC2, C3, C5, C6, C10 and C11. Gaps in the lines indicate periodswhen the traps were not functioning. Steep line slopes indicatehigh rates of sediment output. Since gullies do not report forTable 5.13^Scaled sediment output by time periodDate Period ^(d/m/y 1 2 3 4 5 6 7 8 9 10 11 12Start 8/90 1/11/90 18/11/90 6/91 9/8/91 29/8/91 5/9/91 17/11/91 22/12/91 12/1/92 26/1/92 19/11/92End 31/10/90 17/11/90 2/1/91 8/8/91 28/8/91 4/9/91 16/11/91 21/12/91 11/1/92 25/1/92 26/5/92 25/5/93GullyCl-U 0.03 0.13C2-U 0.04 0.03 0.11 0.02 0.18 1.22 0.47 0.00 0.00 0.16C3-T 0.12 0.67 0.02 0.00 0.48 0.00 0.25 0.00 0.41 0.15 0.26C5-T 1.22 3.89 5.89 0.55 0.16 0.44 0.21 0.18 0.03 0.12 0.41 0.76C6-T* 0.28 0.12 0.05 0.14 0.03 0.18C4-SF 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00C8-SF 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.04 0.00C10-SC 0.03 0.59 0.65 2.46Cu-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.0^1^2^3^4^5^6^7^8^9^10Time period (not equal lengths)11 121412 -+^ C2-U^--X-- C3-T^---0--- C5-T- C6-T^Y^ C10-SC^C11-SCFIGURE 5.10 Cumulative sediment outputU, T, and SC gullies93I94all periods, comparison of the cumulative total is not alwaysappropriate.Immediately evident from Figure 5.10 is the very large outputrecorded from gully C5 in periods 2 and 3, representing theNovember 11 and November 23, 1990 storms, respectively. Ratesof sediment output from gully C5 in other periods are notexcessively large; thus, the scale factor for this gully isappropriate. Results from other traps in these two periods aresketchy, since almost all traps filled in either the first orthe second of the two periods; however, the balance of evidencesuggests that sediment output from gully C5 was anomalousduring the two November 1990 events.Slash-clear gullies C10 and C11 show steep rises in Figure 5.10during periods 9 - 12. These two gullies produce more sedimentthan all others, except gully C5, despite the fact they werenot monitored in Year 1 when the largest storms occurred.To examine the differences between treatment groups, individualgullies are compared over the same set of periods. The unloggedand slash-clear gullies cannot be compared directly, sincethere are only two common measurement periods. Unlogged gullyC2 is compared with torrented gullies C3 and C5 using sedimentoutput from periods 1,2,4,5,6,7,8,9 and 12. Gully C6 isincluded for periods 7,8,9, and 12. Period 3 is not includedsince both gully C2 and gully C3 traps had filled by that time.Cumulative sediment outputs for the comparison periods are95shown in Figure 5.11.No difference is apparent between the unlogged gully, C2, andtorrented gullies. The cumulative sediment output of C2 isbetween the cumulative output of C3 and C5, and gully C6 showsa 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; thiscoincides with relocation of the trap to the road culvert pool,and hence these outputs may be affected by roadcut sediment. Ifthis is the case, gully C2 output would be less than in thetorrented gullies. Since C2 is the only unlogged gully (exceptfor the short Cl record), C2 must be assumed to berepresentative of unlogged gullies. Sediment outputs from gullyCl in the first two periods are similar to gully C2 outputs(Table 5.13), which supports the assumption of C2 beingrepresentative of unlogged gullies.Torrented gullies and slash-clear gullies are compared forperiods 9-12 (Figure 5.12). Torrented and slash-clear gullieshave different responses. Slash-clear gullies have similartotals, although patterns of sediment output vary. Averageslash-clear sediment output is 5.0 times greater than theaverage torrented sediment output. The slash-clear output isactually greater than shown in Figure 5.12, since both C10 andC11 traps lost sediment; therefore, the increase in slash-clearoutput of 5.0 times over the torrented total is regarded as aminimum. If the torrented data and the slash-clear data areGr-XX^87 ----3 -0 2 -1 —FIGURE 5.11 Cumulative sediment outputUnlogged and torrented gullies0^1^2^4^5^6^7^8^9^12Time period (not equal lengths)C2-U X C3-T 0 C5-T ^ C6-T96+^C3-T^---X --- C5-T^---0--- C6-TC10 -SC Y^ C11 -SC4.54 --PO 3.5 -a,41O•^3--Pow 2 . 5 -- -1'0^2a)a)73a) 1 . 5 -r-fn:1can^1-0.5 -0 ^ i8 9^10^11Time period (not equal lengths)121FIGURE 5.12 Cumulative sediment outputT and SC gullies9798pooled, a Student's t-test rejects a null hypothesis of equalmeans (2-tail probability of equal means = 0.03). Hence thereis a significant difference between slash-clear and torrentedgully responses.These results show significant differences between mosttreatment groups. Sediment output is greatest in the slash-clear gullies. These gullies have a sediment output at least5.0 times greater than the torrented gullies. Torrented gulliesand the one unlogged gully appear to have similar sedimentoutputs, whereas slash-full gullies have very little or nosediment output. Volume of coarse sediment outputAs noted above, not all sediment output could be measured sincesome events exceeded the trap storage volume, or else destroyedthe trap. Therefore, the volumes recorded represent minima inmany cases. Error is estimated in two ways. If sediment has notbeen lost, then the error is the product of trap area anddeposition bar measurement error (0.005 m). If sediment hasbeen lost, then the amount lost is estimated as the product ofeither 0.2 m or 0.3 m depth (based on depths of sediment intraps which did not fill) and the trap area. Sediment output ispresented in Table 5.14. The sediment output volumes in Table5.14 are used in Section 5.4 to calculate the sediment budget.99Table 5.14Gully andTreatmentVolume of coarse sediment output^ Sediment Output (m 3 ) ^--- Year 1 ---^--- Year 2 --- --- Year 3 --Total Error Total Error Total ErrorCl-U 0.14 1C2-U 0.40 2 0.60 1.50 2 0.25 0.18 0.02C3-T 0.73 2 2.0 1.18 2 0.68 0.30 0.02C5-T 25.5 3.7 3.23 0.19 1.66 0.19C6-T 2.27 3 0.04 0.20 0.04C4-SF 0.01 0.00 0.00 0.00 0.00 0.00C6-SF 0.95 1C8-SF 0.00 0.01 0.02 0.01 0.00 0.01C10-SC 0.774,2 0.98 1.61 0.01Cl1-SC 3.185,2 2.4 0.16 0.401) Torrented November 23,^1990.2) Sediment trap filled or destroyed at some time; volume is aminimum 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, 1991.1005.3.2 Fine sediment outputThe sediment input from various sources cannot be directlycompared with sediment trapped at the lower end of the gully.Sources of sediment, particularly from the sidewalls, cancontain significant amounts of fine material which are carriedin suspension past the trap. In order to balance the sedimentbudget, the output of both coarse and fine sediment must firstbe accounted for. Fine sediment output is estimated bycomparing the size distribution of the source material with thesize distribution of the trapped material. MethodSediment delivered to the channel can do one of three things:it can remain in the channel, it can be carried to the trap andstored there, or it can be carried past the trap. Generally,distance travelled is associated with sediment size: largerparticles are stored in the channel, intermediate sizes arestored at the trap, and fine particles are carried past thetrap. Since fine material is separated from the intermediatesize material at the trap, the relative proportions of fine andintermediate material in the trap and sidewall sediments shouldreflect the amount of lost material.Sidewall and trap sediment are both sampled. The sediment isseparated into coarse ( >16 mm), intermediate (8 - 16 mm) and101fine fractions ( <8 mm). All fine and intermediate material isassumed to reach the trap, and the samples are truncated toinclude only the intermediate and fine fractions. A unit amountof sidewall sediment yields a fraction, X, as trap sediment,and a fraction 1 - X of fine sediment output. Since theintermediate 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 thesidewall and trap samples. The fraction of the truncatedsidewall sample retained at the trap for each size class is:Trapi X = Trapi  * Sidewall 8 16Trap8_16Where:Trapi X = trap sediment of sizei, measured as a percent oftruncated sidewall sample.Trapi = trap sediment of sizei, measured as a percent oftruncated trap sample.Sidewall8 -16 = sidewall sediment of 8-16 mm, measured aspercent of truncated sidewall sample.Trap8_16 = trap sediment of 8-16 mm, measured as percent oftruncated trap sample.Table 5.15 shows an example of the calculation procedure. Inthis example, a unit amount of truncated sidewall materialwould result in 65 % of the sediment being retained at thetrap. The fraction of material carried past the trap insuspension, measured as a fraction of the total sidewall sample(all grain sizes), is then:102Fine Loss % = (1 - Total Trap % 1) * (Sidewallsub-16^weight) 100^(Total Sidewall weight)1 Total Trap as measured as a percent of the truncated SidewallTable 5.15 Example of trap weight reconstructionSize Class^Truncated^Truncated^TruncatedSidewall^Trap^Trap(mm)^Sample, %^Sample, %^% Truncated Sidewall8 - 16 29 44 294 - 8 29 44 292 - 4 14 5 31 - 2 14 4 20.5 - 1 7 3 2< 0.5 7 1 1Total 100 100 655.3.2.1 ResultsSeveral sidewall samples and a single trap sample werecollected for two gullies, C3 and C5 (Table 5.16). The sidewallsamples are from widely spaced locations, and C5 Sidewall 1 isfrom the roadcut. Away from the road, larger samples weredifficult to obtain, so smaller samples were collected.103Table 5.16 Sediment samples, gullies C3 and C5^Gully C3  ^Gully C5^Sample^Weight^Maximum^Weight Maximum(kg)^Representative^(kg)^RepresentativeSize Fraction l^Size Fraction l(mm) (mm)Trap^139^32 - 45 195 32 - 45Sidewall 1^10.4^11 - 16 131 32 - 45Sidewall 2^6.39^11 - 16 6.44 11 - 16Sidewall 3^4.36^8 - 11 9.08 8 - 16Sidewall 4^3.69^8 - 11 11.9 16 - 22Sidewall 5 12.2 16 - 221 Representative sample defined as nclass (Church et al, 1987).=> 100 stones in a size104The sidewall grain-size distributions for each gully are shownin Figures 5.13a and 5.13b. Each gully has a composite averagesidewall sample created. Since most sidewall samples are notlarge enough to characterize the distribution of materiallarger than 16 mm, the larger samples are used to determine thefraction of material greater than 16 mm. For gully C5, sidewallsample 1 is used; for gully C3, the trap sample is used, sincenone of the sidewall samples are sufficiently large. Whentruncated for a maximum size at 16 mm, most samples have allsize classes adequately represented. Three samples are notrepresentative of the 11 - 16 mm size class since they havefewer than 100 stones in the size class, but since they aregenerally close (n = 67, 87, and 94), they will be used. Thecomposited sidewall sample uses all sidewall samples within agully to determine the average for all size classes smallerthan 16 mm. The average sidewall grain-size distribution andthe trap distribution are plotted in Figure 5.14. The grainsize distribution of sediment in trap C6 is also shown.Once the composite sidewall sample has been constructed, it iscompared to the trap sample. Each is truncated at 16 mm, andthe amount of sediment retained and lost at the trap iscalculated. Table 5.17 shows the results for each gully.1009080706050403020,XJ=1100•^FIGURE 5.13a C3 sidewall sedimentCumulative distributions0.01^0.1^1^10^100Grain size (mm)105100908070605040302010FIGURE 5.13b C5 sidewall sedimentCumulative distributions106= =- --+4=^+'00.01^0.1^1^1 0^100Grain size (mm)FIGURE 5.14 Sidewall and trap sedimentCumulative distributions1071 0^1001009080706050403020100,?/,*is,0.01^0.1^1Grain size (mm)^— 1— C5 SIDE^C5 TRAP - * C3 SIDE^C3 TRAP ^ 1^ C6 TRAP108Table 5.17 Fine sediment retained at trap, gullies C3 and C5^ Gully C3 ^Size Class^Composite Composite C3 Trap C3 Trap C3 TrapSidewall Side-Trunc.^Trunc.^Trunc.(mm) % % % % (4191 - 128 7.6 9.564 - 91 10.1 12.645 - 64 9.3 11.632 - 45 9.6 11.922 - 32 9.2 11.516 - 22 11.5 14.411 - 16 4.0 9.3 9.6 33.8 9.38.0 - 11 3.6 8.3 7.3 25.5 8.35.6^- 8.0 3.1 7.2 5.1 17.8 5.34.0 - 5.6 2.4 5.6 2.5 8.7 2.62.8^- 4.0 2.5 5.9 1.8 6.3 1.92.0 - 2.8 2.2 5.2 1.0 3.5 1.01.0 - 2.0 4.4 10.2 0.76 2.7 0.800.5 - 1.0 4.5 10.6 0.22 0.77 0.230.25^- 0.5 4.0 9.5 0.08 0.27 0.080.12^- 0.25 3.8 8.8 0.05 0.19 0.060.06^-^0.12 2.2 5.3 0.05 0.17 0.05Pan 6.0 14.0 0.09 0.33 0.10Total 100 100Total, <16 mm 42.7 100 28.5 100 29.7109Table 5.17 Continued^ Gully C5 ^Size Class^Composite Composite C5 Trap C5 Trap C5 TrapSidewall Side-Trunc^Trunc.^Trunc.(mm) % % % % %191 - 128 7.7 22.264 - 91 10.1 19.445 - 64 4.6 15.332 - 45 7.7 11.522 - 32 4.5 5.216 - 22 5.8 5.211 - 16 5.4 9.0 3.9 18.3 9.08.0 - 11 5.1 8.6 3.9 18.3 8.65.6^- 8.0 4.8 8.1 2.9 13.4 6.54.0 - 5.6 4.6 7.7 2.9 13.4 6.52.8^-^4.0 3.8 6.5 1.7 8.0 3.92.0^-^2.8 3.5 5.9 1.7 8.0 3.91.0 - 2.0 6.6 11.1 2.1 10.0 4.80.5^-^1.0 5.9 9.8 1.1 5.3 2.60.25 -^0.5 4.7 7.9 0.45 2.1 1.00.12^- 0.25 4.7 7.9 0.25 1.2 0.570.06 - 0.12 2.9 4.9 0.23 1.1 0.52Pan 7.5 12.6 0.18 0.85 0.41Total 100 100Total, <16 mm 59.6 100 21.3 100 48.11 Percent of Truncated Sidewall.110The amount of fine sediment loss from each gully is calculatedas: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 andunlogged gullies are assigned a fine sediment loss of 30%. Inslash-full gullies fine sediment loss is zero, since nosediment transport is observed. Fine sediment loss from gulliesC10 and C11 in Year 2 is attributed to erosion of the channelmaterial, which has an estimated 20% fine sediment. Channelsediment in slash-clear gullies had become armoured by Year 3,and in addition, storms in Year 3 were very moderate. Giventhese conditions, fine sediment output in Year 3 for slash-clear gullies is assumed to be zero.The near equivalence of Gullies C3 and C5 in terms of finesediment loss is almost certainly fortuitous. The variabilityof sidewall grain size distributions within one gully, letalone two, suggests that the amount of sediment lost is muchmore 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 allgullies, an error value of 50% is chosen. Table 5.18 shows theinput, fine sediment output, and error for gullies in Years 2and 3. The fine sediment output and error will be used to111Table 5.18 Volume of fine sediment outputGully^Input Fine Sediment Fine Sediment Fine SedimentVolume^Loss^Output Volume^Error VolumeYear 2C2-U(m3 )0.03(%)30(m3)0.01(m3)0.00C5-T 24.1 30 7.23 3.62C6-T 12.1 30 3.63 1.82C3-T 1.12 30 0.34 0.17C4-SF 1.76 0 0 0C8-SF 0.39 0 0 0C10-SC 114. 21 20 2.84 1.42Cll-SC 111. 51 20 2.30 1.15Year 3C2-U 0.22 30 0.07 0.03C5-T 8.80 30 2.64 1.32C6-T 8.69 30 2.61 1.30C3-T 0.92 30 0.28 0.14C4-SF 2.51 0 0 0C8-SF 0.49 0 0 0C10-SC 0 20 0 0Cul-SC 0 20 0 01 Input for Gullies C10 and C11 is erosion of gully channel.112complete a budget for each gully (Section 5.5). Since finesediment output depends upon input, gully C3 is grouped withthe slash-full gullies. Fine sediment output is dependent uponthe input volume; therefore, the torrented gullies have thegreatest loss of fine sediment. Fine sediment outputs in Year 2are generally greater than those in Year Sediment transport distancesSediment transported along the channel may be from two sources:the channel bed material, and sediment introduced from sidewallinput. Channel bed material and introduced sediment of the samecalibre may respond differently to the same flow conditions,since channel sediment fabric may resist transport of sediment.Sediment transport distances were monitored for both channelbed sediment and sediment introduced to the channel. Channelbed sediment transport was monitored in Year 2, and introducedsediment 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 andC6. The stones were painted on the upper surface only, so as tonot disturb them. Stones were painted October 3, 1991;transport distances were measured on November 16, 1991, after astorm which delivered 93 mm of precipitation in 24 hours.Figure 5.15 shows the transport distances and the intermediateaxis of each stone.Figure 5.15Channel sediment transport distances+ C2 ^ C5 — C6113114Almost all the data shown in Figure 5.15 are from C6: 50 stonesregistered movement in C6, C2 had only 2 stones move, and C5had 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 thetrap, over 60 m from their original position. In contrast, inC2, with a channel slope of 27 ° , only two stones moved, just 1metre. Gully C5 showed sediment transport distances of up to 8metres. Since the number of stones which were originallypainted, and the number of stones which remained at the startpoint after the storm were not counted, it is not known whatrecovery rate was obtained.In September of Year 3 painted stones were introduced to thechannel. Gullies C2, C3, C5, C6, C10 and C11 had stones placedin 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 cross-section. Positions of stones were noted subsequently inNovember 1992 and May 1993. The largest storm of Year 3 was 94mm in December 1992.Results are similar for both periods, so May results will bediscussed. Recovery rates of the painted stones are low: anaverage of 45% for the large stones, and 24% for the smallstones. One problem is that the painted stones are often hiddenby larger clasts. Of the stones recovered, 74% had not moved.Transport distances for 22 large stones ranged from 1 to 15metres. Only gully C6 had large stones transported more than 10115metres. Transport distances for 11 small stones ranged from 1to 17 metres; again, only C6 had distances greater than 10metres recorded. With the exception of C6, there does not seemto be any difference between gullies. Given the low recoveryrates, it is not certain whether the transport distancesrecorded represent average results.The distance sediment is transported in the channel wasmonitored for both channel bed material and introducedsediment. Although conducted at different times, both types ofsediment were monitored during periods when the largestmagnitude storm was either 93 or 94 mm of precipitation in 24hours. With the exception of gully C6, no sediment transportdistances greater than 10 metres were measured. Since recoveryrates were low, or not monitored, transport distances greaterthan 10 metres may have occurred, but were not observed. GullyC6 had transport distances greater than 60 metres, aconsequence of its steep bedrock channel. No differences inchannel bed sediment and introduced sediment transportdistances are observed.5.3.4 Precipitation and sediment output5.3.4.1 MethodThe approach in this investigation is to determine whethersediment output from gullies is related to storm intensity.116Stage data from the Coquitlam River station indicate an averagelag of 2 hours from the start of precipitation to the start ofhydrograph rise (10 storms, drainage area 54.7 km2 ). Sincegullies have much smaller drainage basins (about four orders ofmagnitude less), the lag time should be much shorter.Therefore, there should be a strong relationship betweenprecipitation intensity, stormflow, and fluvial sedimenttransport.To examine the relationship between storms and sediment output,the maximum 3, 6, 12, 24, 48, and 72 hour precipitation totalsare determined for individual storms. The storm record betweensediment output measurements is examined, and any sedimentoutput is associated with the largest storm for the period. Inalmost all cases, only one significant storm occurred in anygiven period. The gullies used in this procedure are C2, C3,C5, and C6. Other gullies did not have records of sufficientlength to warrant their use.The short record of precipitation at the study site (C6station: November, 1991 - May, 1992) requires extension throughcorrelation with data from the B.0 Hydro Coquitlam Lake orRiver stations. B.C. Hydro station data from each storm for theperiod that the C6 Station was operating are compared usingmaximum 3, 6, 12, 24, 48, and 72 hour precipitations. Resultsof the correlation analysis are given in Table 5.19.117Table 5.19 Comparison of C6 precipitation with Coquitlam Lakeand River stationsTime^Lake^DF^River^DFPeriod^Station Station(R2 ) (R2)3 Hrs 0.261 13 0.284 156 Hrs 0.600 13 0.617 1512 Hrs 0.328 13 0.581 1524 Hrs 0.552 13 0.594 1548 Hrs 0.265 9 0.620 1372 Hrs 0.238 8 0.538 12The Lake station shows little correlation with C6 station;however, the River station shows a consistently moderatecorrelation, except for the 3 hour summation, which is poorlycorrelated. In all cases, the River station is bettercorrelated with C6 station than the Lake station. Thus, toextend the C6 record, the River station data will be used.Regression equations used to convert River station data to C6data are given in Table 5.20.118Table 5.20 Values for (est. C6) = coeff(River station) +constantTime^River StationPeriod^Coefficient3 hour^0.4636 hour 0.81012 hour^0.77524 hour 0.81748 hour^0.90272 hour ResultsEach storm precipitation periodConstant(mm)11.911.421.226.935.638.9(3,^6,^12,Standard ErrorC6 Estimate (mm)6.117.9114.122.531.447.224,^48,^and 72 hour)is compared with scaled sediment output for each gully. Figure5.16 presents graphs of maximum precipitation intensity andsediment output for the 3 to 72 hour periods. The two outlyingdata points from gully C5 are the November 11 and November 23,1990 storms.There is no strong relationship between precipitation intensityof any time period and the amount of sediment output. Bothsimple regression and multiple regression (of 3 to 24 hourprecipitation with antecedent precipitation for periods up to72 hours) yield poor results (best simple linear regression R22Pn 1.8 -aE 1.6 -O 1.47PZ 1.2 -WZ1-4^1 -AW 0.8 -wA 0.6-W0.4 -0w 0.2 -0041-GI><0^><-0-5^10^15^20^25^30^35MAXIMUM 3 HOUR PRECIPITATION (mm). L,C5 3.90><><0><0-0-0><40-ioal-40^45/IC5 5.910^20^30^40^50^60^70^80MAXIMUM 6 HOUR PRECIPITATION (mm) 5.16 Maximum storm precipitation and sediment output2Pn 1.8PiDP 1.60 1.4P1.2H 1AW 0.8wA 0.6W0.4u 0.2w020^40^60^80^100^120^140^160MAXIMUM 12 HOUR PRECIPITATION (mm)+ C2 x C3 D C5 - C621.8-1.6 -1.471.2 -10.80.6 -0.4 -0.2 -0 0 C5 3.9C5 5.950^100^150^200MAXIMUM 24 HOUR PRECIPITATION250C5 3.9^C5.5.91111.-10.8 -0.6 -0.4 -^ ).<0.2 -a50^100^150^200^250MAXIMUM 48 HOUR PRECIPITATION (mm)00Owl300Figure 5.16 continued1200^50^100^150^200^250MAXIMUM 72 HOUR PRECIPITATION (mm)C2 x C3 D C5^C62a 1.81.60 1.4P1.21W▪ 0.80.6N 0.44o 0.20 300^350121is 0.47 for the 48 hour period; the best multiple regression R 2is 0.55 for 3 hour maximum precipitation with 12 hourantecedent precipitation).The data in Figure 5.16 do not show a strong trend ofincreasing sediment output with increasing storm precipitation;however, examination of the data shows that there is athreshold storm intensity which must be exceeded to transportsediment. The threshold is exhibited in each time period, andin almost all cases, the exceedence of the threshold isassociated with sediment output. The definition of a thresholdfor each time period results in a storm intensity curve for the3 to 72 hour period, which defines the minimum amount ofprecipitation required to transport sediment (Figure 5.17).The threshold value for the 72 hour period is only slightlygreater than the value for the 48 hour period. This suggeststhat the amount of precipitation in a 72 hour period is notimportant for sediment transport in gullies; therefore, stormintensities of less than 2 days duration appear to determinewhether sediment transport will occur.No simple explanation exists for the scatter of data points inFigure 5.16 which exceed the threshold precipitation values forsediment transport. Any relationship which exists can beconfounded by precipitation gauge location relative to thestudy area, rain on snow events, antecedent soil moisture, orsediment supply variation. The storms for the period of C6120100 -6O 804)rts• 60 -wPO• 40-H-d20 -1^i^1^1^1^110^20^30^40^50 60 70Time (hours)00 80122Figure 5.17^Threshold precipitationrequired for fluvial sediment transport123station and B.C. Hydro station comparisons are not large; thus,the correlation for storms used in this data set may not applyto storms of greater magnitude. The storms of November, 1990appear to exhibit strong localized precipitation; very large,localized events are probably not well characterized by theRiver station data. In addition, the study site is about 600 mabove the River station, so snow may fall (or melt) at the sitewhen rain is recorded at the River station. Event sequences mayaffect the amount of sediment available for transport. DuringYear 2 and Year 3, the channel margins appeared to accumulatesediment in some of the gullies (C3, C5); the moderate stormevents of these years may not transport as much sediment as isdelivered to the channel margins.The precipitation threshold for fluvial transport of sedimentin Coquitlam gullies is compared to the threshold for shallowlandslides and debris flows developed by Caine (1980) (Figure5.18). For periods from 6-24 hours, the fluvial transportthreshold is almost parallel to Caine's threshold, and shiftedjust slightly lower. This is a remarkable amount of similaritybetween fluvial processes and mass movement processes. Thealmost parallel lines suggest that hillslope geomorphicresponse to precipitation intensity over a range of durationsis similar, whether the response is of a fluvial or a massmovement nature. However, since the threshold values are sosimilar, it suggests that the Coquitlam gullies would require agreater threshold precipitation for initiation of landslides ordebris flows than the threshold determined by Caine.10I^I^11T111^I^I^I^1111110 100Time (hours)11Figure 5.18 Fluvial and debris flowprecipitation thresholds—1-- Fluvial transport X Mass transportMass movement curve from Caine (1980)1241255.4 Debris flowsDuring autumn and winter of 1990-1991 several large stormsoccurred in the North Shore Mountains. The storms of November9-11, November 23, December 3, 1990, and April 3, 1991 causedmany slope failures in the GVWD basins (Table 5.21; Thurber,1991). The storm of November 23 caused the largest number offailures, most of them in Coquitlam Basin. All Coquitlamfailures occurred in the southern half of the basin, mostly intwo clusters. On the west shore of Coquitlam Lake, opposite themouth of Cedar Creek, were eight failures, and along the lengthof Cedar Creek were nine failures, centered around the studysite (Figure 5.19). Hence the study site was one of the twoareas showing the highest frequencies of failures over theentire North Shore area. Unfortunately, no precipitationrecords were available from the C6 precipitation gauge locatedat the study site during this period. Hourly precipitation fromthe two B.C. Hydro stations are examined instead.The occurrence of debris flows in some of the monitored gulliesprovides an opportunity to compare low intensity, frequentsediment transport events with high intensity, infrequentevents. Location, surficial materials and geomorphic featuresare described for the two monitored gullies which torrented.Volumes of sediment incorporated into debris flows provide acomparison with the two-year sediment budgets based on lowintensity processes.126Table 5.21 Dates and numbers of slope failures in GVWD basinsin 1990-1991.Date   Basin ^Capilano^Seymour^Coquitlam9 - 11/11/90 2 3 323/11/90 1 3 193/12/90 0 1 13/4/91 1 1 05.4.1 Coquitlam precipitationStorm intensities for the two November storms varied widelyover southwest British Columbia. Table 5.22 shows the 1 dayprecipitation total and return period for various stations insouthwest British Columbia for each storm.The Coquitlam stations have the greatest precipitation valuesfor each storm period. The greater precipitation amounts maysimply be a result of a general increase of precipitation inmountainous areas, compared to low-elevation areas such as theFraser Valley. Church and Miles (1987) note the Coquitlam Lakerecord often registers the highest precipitation in the FraserLowland region.Coquitlam RiverClimate Station0-/)IWatershedBoundaryaNC)kce66xx 0xCoquitlam LakeA Climate StationxDam 0 A5 )CDA Nov. 9-11 slidesX Nov. 23 slides0^ 5kmSource: Thurber, 1991127FIGURE 5.19 Location of Coquitlam Basin slope failures128Table 5.22 Maximum one day precipitation and return periods,southwest British Columbia stations lStation^November 9 - 11^November 22 - 23Precip.(mm)Return Per.(years)Precip.(mm)Return Per.(years)Abbotsford A 80 9 40 1Chilliwack 99 10 42 1Hope A 173 > 100 116 15North Vancouver 51 1 40 1Squamish 164 n/a 124 n/aVancouver A 35 1 21 1Whistler 72 15 52 2Coquitlam Lake 226 25 2 156 5 2Coquitlam River 144 2-3 3 189 9 31 Data from Environment Canada, 1991a and 1991b, exceptCoquitlam Stations.2 From Maximum Daily Precipitation Series, Figure 4.63 From Partial Duration Series, Figure 4.8129Although the greatest number of slope failures occurred onNovember 23, the greatest amount of precipitation was recordedat the Coquitlam Lake station on November 9. Figure 5.20 showsthe maximum precipitation for 1 hour to 72 hour periods at boththe Lake and River stations. In each period, the Lake stationduring the November 9-11 storm recorded the greatestprecipitation. If local slope failures were simply andpositively related to precipitation recorded at either of thetwo stations, then most slope failures should have occurredduring the November 9-11 storm.Several confounding factors may prevent a simple correlationbetween the amount of precipitation recorded at one of thegauge locations and the number of slope failures whichoccurred. The gauge may not accurately reflect the amount ortype of precipitation at the study site, or snowmelt may occur.Alternatively, the severe storm of November 9-11 may have fullysaturated slopes, thereby increasing the chance of slopefailures in subsequent events.Severe storm precipitation is known to occur within cells ofconvective uplift producing increased precipitation (Church andMiles, 1987). The linear alignment of slope failure locationsalong Cedar Creek and on the west side of Coquitlam Lakesuggests topographic confinement of the storm front withresulting convective cells. As a result, precipitation at thestudy site may have been much greater than that recorded ateither of the two climate stations.400350-0• 300 -GZO• 250-0 200 --H_94 150-0c)s-4• 100 --H• 50-(t)I^I^I^I^I^I^110^20^30^40^50 60 70Time (hours)00 80FIGURE 5.20 Maximum precipitationCoquitlam stations; November, 1990130131Snowpack prior to a storm may cause increased delivery of waterto the soil as warm rain and wind melts the snow, with possibleeffects on slope failure occurrences. Table 5.23 showsantecedent precipitation and temperatures leading up to the daywith the greatest precipitation in each storm. Snowpacks mayhave been present before both storms. In the case of theNovember 9-11 storm, the Lake station antecedent temperaturessuggest rain; however the River station temperatures indicatesome precipitation may have been in the form of snow. The Lakestation is at an elevation of 161 m, the River station at 280m. Higher elevations may have held shallow snowpacks prior tothe warm rain of November 8 and 9. The November 23 storm almostcertainly had an antecedent snowpack. Temperatures at bothstations were close to zero, and temperatures showed a diurnalpattern, suggesting solar radiation heating rather than warmerair mass invasion. Significant precipitation fell on November19 and 21, totalling 82 mm and 85 mm at the Lake and Riverstations, respectively. A rapid rise in temperature wasobserved from 1400 hrs (0.7 ° C) to 2100 hrs (7.0 ° C) November22, coincident with the onset of precipitation. Snowmeltprobably delivered about 80 mm of water during the November 23storm. If all the snow melted in one day, and the same amountof precipitation fell at the study site as fell at the Riverstation (probably a minimum), total water input would be about270 mm. The total water input, whether snowmelt is included ornot, is greater than the threshold intensity defined by Caine(1980) (Figure 5.21).132Table 5.23 Antecedent precipitation and temperature forNovember, 1990 StormsDay Coquitlam Lake StationMin Temp^Max Temp^Prec.(0C )^(0C)^(mm)Coquitlam River StationMin Temp^Max Temp Prec.(0C)^(Cc)^(mm)Nov. 5 1.1 5.7 5 -0.5 4.6 6Nov. 6 1.9 6.1 38 0.4 3.7 27Nov. 7 1.4 6.5 33 0.4 5.5 29Nov. 8 2.9 9.2 93 0.4 6.1 65Nov. 9 6.9 11.1 226 5.5 12.1 144Nov. 19 0.6 2.9 26 0.1 2.2 43Nov. 20 -0.3 4.0 0 -0.2 1.0 0Nov. 21 -0.2 1.1 56 0.1 0.4 42Nov. 22 2.9 8.4 33 0.4 7.3 47Nov. 23 8.5 10.2 156 8.2 10.6 189I^!III10Time (hours)I^I^i^i^I^II1001Figure 5.21 Coquitlam precipitationcompared with Caine's curve1331345.4.2 Cl debris flowThe C1 debris flow of November 23, 1990 originated as a debrisslide about 380 m upslope of Branch 230 (Figure 5.22). Althoughon an open slope, the failure was in direct line with gully C1.The failure was roughly rectangular in plan-form, and soildepths ranged from 0.3 to 2 metres. The debris slide failuresurface was the bedrock-till interface, and sloped between 29 °and 32 ° . Piping was evident at the headscarp 2 weeks after theevent. The debris flowed down gully C1, which is tributary tothe adjacent gully to the southwest, which also had a debrisflow November 23.Calculation of the amount of sediment incorporated into thetorrent uses survey data from December 9, 1990. The failurezone is separated from the transport zone in order to determinethe relative importance of the two sources. For the failurezone, four cross-sections widths were measured, and depths fromthe original surface to the failure surface were estimated atthree points along the cross-section. The volume of the failurewas calculated as the product of the area of each cross-sectionand the section length for each cross-section. Similarly, thetransport zone had widths and depths of erosion measured atregular intervals; the product of these measurements providesthe volume incorporated into the debris flow in the transportzone. Volume estimates of debris flows must account for debrisdeposited along the track (Rollerson, 1984). The debris flowdeposited small amounts of sediment at the edges of the forest,FIGURE 5.22 Cl and C6 debris flowsLegend — — — _ StreamsDebris flow paths^ Contours (200' interval)RoadsCutblock1^ 10 metres^ 500135136at the up-slope sides of trees. Deposited sediment is estimatedat 10% of the total volume. Table 5.24 shows the results of thevolume calculations.Table 5.24 Cl debris flow volume, November 23, 1990Failure Zone Subtotal (m 3 )^390Transport Zone Subtotal (m 3 ) 1340Deposited Sediment (m 3 )^-170Total C1 Debris Flow Volume (m 3 )^15605.4.3 C6 debris flowThe C6 debris flow of November 23, 1990 started as an openslope failure northeast of C6 gully (Figure 5.22). The failurezone is located approximately 180 m above Branch 230 road, onan open slope to the northeast of gully C6. The top edge of thefailure is in thick till, approximately 3 m deep. The depth oftill lessens towards the lower edges of the failure, wheredepths average about 1.5 m. The shape of the failure is a broadellipse, with the length axis 16 m long, and the width axisabout 10 m wide. The slope in the failure zone is 26 ° . As withthe Cl failure, piping is evident at the headscarp, along thebedrock-till interface. On the bedrock surface, a 1 cm layer ofextremely slippery black organic material was found. This137material may have contributed to instability at the site.The debris appears to have flowed through a narrow channel downthe open slope, and then entered the gully from the rightsideslope 130 m above Branch 230. Some material spread acrossthe open slope to the right of the gully. The slash deposit inthe channel remained in place above the area where the debrisflow entered the gully (Figure 5.23). The gully channel andsidewalls 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: thefailure site, the open-slope transport zone, the gully channel,and the sidewalls. Table 5.25 shows the volume calculation forthe C6 debris flow. Gully sidewall scour is estimated to be 0.3m. Depth of failure in the channel is estimated to be 0.5 m,over an average channel width of 4 m. Significant depositsoccurred on the open slope to the north-east of the gully andin a few protected areas within the gully. Deposition along thedebris flow track is estimated to account for 20% of the totalvolume.Debris flows in both C1 and C6 started as open slope debrisslides. Both failure sites had 2-3 metre deep soils at theheadscarp, where piping was evident. The debris flow in Cl wasabout 75% larger than the C6 debris flow, primarily a functionof the length of gully which was scoured above the Branch 230138Figure 5.23 Channel slash above debris flow zone, gully C6The portion of the channel scoured by debris flow is in lowerthird of photo. The channel axis runs from the lower leftcorner of the photo to the center of the photo, where thechannel is filled with slash. Sediment fill in the unscouredzone is about 0.5 m deep; slash fill is about 1 m deep. Stadiarod (1.1 m) is located at the edge of the slash, on the middle-right of the photo.139Table 5.25 C6 debris flow volume, November 23, 1990.Failure Zone Subtotal (m 3 ) 220Open Slope Zone Subtotal (m 3 ) 140Gully Zone Subtotal (m 3 ) 760Deposited Sediment (m 3 ) -230Total C6 Debris Flow Volume (m 3 ) 900road. Average sediment yield in the transport zone of the C1debris flow is 3.5 m 3 /m of channel; average sediment yield forthe transport zone of C6 is 5.9 m 3 /m. Fannin and Rollerson(1993) calculated average channel debris yields between 5-10m3 /m. Since the Cl and C6 gullies are small, agreement with thelower debris yield rate from Fannin and Rollerson isappropriate.5.5 Sediment budgetsThe sediment budget summarizes the input, storage, and outputof sediment in each gully. Budget terms of input from sidewallsources, channel storage changes, fine sediment output, andoutput of coarse sediment are shown in Table 5.26. Theformulation of the budget equation is:Input -6Storage = Fine Output + Coarse OutputIf the terms do not balance, then error in one or more140measurements has been made. The difference between the twosides of the equation is the budget error. If the budget erroris positive, then either input has been overestimated, or elsestorage or output has been underestimated, or a combination oferrors. Negative error indicates underestimation of input, oroverestimation of the storage or output terms.Not all budget terms were measured in Year 1, so the balancecannot be computed in that year. Storage change in the slash-full gullies (C4, C8) is assumed to be equal to input, sincezero or almost zero coarse outputs were measured. Fine sedimentoutput for these gullies is also assumed to be zero, a resultof the trapping capability of the slash.In almost all cases, the budget error is less than the largestbudget term. However, this yields large budget errors in somecases. If the budget error is comparable to the pooled error ofthe individual budget terms, then the methods used can bepresumed to be accurate within the stated errors. If budgeterror is greater than the pooled error, then the measurementmethods are not within the stated error levels. Table 5.27shows the individual budget term errors, and the pooled errorfor each gully. The budget error is shown next to the poolederror for comparison.141Table 5.26 Sediment budgetsGully^Input^Storage Fine Sed. Coarse Sed^BudgetVolume^Changel Output^Output^ErrorYear 1(ms ) (m3) (Ins) (m3) (Ins)C2-U n/a n/a n/a 0.40C3-T 1.14 n/a 0.34 0.73C5-T 5.94 n/a 1.78 25.5C4-SF n/a n/a n/a 0.01C6-SF 0.27 n/a 0.08 0.952C8-SF n/a n/a n/a 0.00Year 2C2-U 0.03 -0.18 0.01 1.50 -1.30C3-T 1.12 3.20 0.34 1.18 -3.60C5-T 24.1 6.40 7.23 3.23 7.24C6-T 12.1 -2.40 3.63 2.27 8.60C4-SF 1.76 1.76 0 0.00 0C8-SF 0.39 0.39 0 0.02 -0.02C10-SC 0.00 -14.2 2.84 3 0.77 10.6C11-SC 0.00 -11.5 2.303 3.18 6.02142Table 5.26GullyC2-UContinued, Year 3Input^StorageVolume^Changel(m3 )^(m3)0.22^1.20Fine Sed.Output(m3)0.07Coarse SedOutput(m3)0.18BudgetError(m3)-1.23C3-T 0.92 0.58 0.28 0.30 -0.24C5-T 8.80 2.60 2.64 1.66 1.90C6-T 8.69 0.66 2.61 0.20 5.22C4-SF 2.51 2.51 0 0.00 0C8-SF 0.49 0.49 0 0.00 0C10-SC 0.00 0.55 0.00 1.61 -2.16C11-SC 0.00 1.30 0.00 0.13 -1.461 Positive values indicate deposition within the channel.2 Gully C6 Output volume to November 1990, before debristorrent.3 Gullies C10 and C11 Fine Output based on erosion of channelmaterial, estimated at 20% fines.143Table 5.27 Total measurement errorGully Input Storage Fine Sed. Coarse Sed Pooled BudgetError Error^Error^Error^Error ErrorYear 2C2-U(m3 )0.02(m3)0.09(m3)0.00(m3)0.25(m3)0.27(m3)-1.30C3-T 0.23 1.60 0.17 0.68 1.76 -3.60C5-T 4.84 3.20 3.62 0.19 6.84 7.24C6-T 2.07 1.20 1.82 0.04 3.01 8.60C4-SF 0.41 0.00 0.41 0C8-SF 0.14 0.01 0.14 0.02C10-SC 7.10 1.42 0.98 7.31 10.6Cul-SC 5.75 1.15 2.40 6.34 6.02Year 3C2-U 0.10 0.60 0.03 0.02 0.61 -1.23C3-T 0.25 0.29 0.18 0.02 0.42 -0.24C5-T 2.04 1.30 0.57 0.19 2.49 1.90C6-T 1.74 0.33 1.09 0.04 2.08 5.22C4-SF 0.57 0 0.00 0.57 0C8-SF 0.18 0 0.01 0.18 0C10-SC 0.00 0.28 0.00 0.01 0.28 -2.16Cul-SC 0.00 0.65 0.00 0.40 0.76 -1.46144In most cases, the budget error is greater than the poolederror (Figure 5.24). In four cases, the budget error is largerthan all of the budget terms. The largest errors are associatedwith the largest budget terms. Torrented gullies' largest errorterm is input. Slash-clear gullies' largest error term isstorage change. Both these terms extrapolate samples of changeover a large area, which can result in multiplied errors.Sediment output error is usually small, unless the trap wasoverwhelmed. The magnitude of input and storage errors indicatethat further refinement of gully sediment budgets wouldemphasize more accurate measurements of these two terms. Inaddition, fine sediment output needs further quantification.5.6 SummarySediment budgets and complete comparisons between gullies areavailable for Year 2 and Year 3. Year 1 had at least two majorstorms (November 9-12, November 23); two debris flows occurredon monitored gullies. One gully, C5 (torrented) had a verylarge output of sediment during both major storms, yet did notproduce 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 inany year.Sediment budgets in Years 2 and 3 show the amount of input,storage, and output for each gully. Two slash-clear gullies, 2slash-full, 3 torrented and 1 unlogged gully were monitored.Figure 5.24Budget error vs. pooled errorPooled error (m ^ 3)145146Since only one unlogged gully was effectively monitored,unlogged response for all sediment budget terms should beconsidered more tentative than the results from other treatmenttypes. Budget error is usually greater than pooled error,suggesting that not all terms were well measured.Alternatively, not all sediment transfers or storage elementswere recognized. Input and storage terms appear in most casesto 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 theprimary treatment effect. Secondary effects, implicit in thetreatment classification, affect the amount of sediment input.Torrented gullies have the largest proportion of disturbed(bare) sidewall area. Recently torrented gullies have thegreatest input volume (C5 and C6). Input from the oldesttorrented gully, C3, is similar to that of the slash-fullgullies (C4 and C8), indicating torrented sidewalls willrecover (particularly with replanting efforts), perhaps over aperiod of 15 years. The unlogged gully, C2, has little input,mostly from a slump along the channel. Its input terms are mostsimilar to those of the slash-clear gullies, where input isapparently close to zero.Input was not monitored directly in the slash-clear gullies,but the high volumes of slash on the sidewalls probablyprevented any input. No slumps, bare ground, or rainsplasherosion were observed in these gullies. Therefore, the147assumption of zero input appears justified.Channel storage changes are different in Year 2 compared toYear 3. Year 3 shows consistent small amount of storage inalmost all cross-sections, regardless of treatment type. Thesmall number and size of precipitation events is a probablecause of the increase in channel storage in Year 3, sincesediment may be delivered to the channel, but not transporteddown the channel. Year 2 results vary by treatment type. Slash-clear gullies show clear evidence of erosion, whereas oldertorrented gullies (C3 and C5) generally show deposition in thecross-sections. The newly torrented gully C6, and the unloggedgully C2 have more variable behavior, somewhere between oldertorrented gullies and the slash-clear gullies. No reason isevident 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 gullieshave less sediment output than slash-clear gullies and moresediment output than slash-full gullies. No significantdifference exists in sediment outputs from unlogged andtorrented gullies.Fine sediment output is dependent upon input. The torrentedgullies have the greatest input, and hence the greatest finesediment output. Slash-clear gullies have the next largestoutput of fine sediment. Unlogged and slash-full gullies havelittle or no fine sediment output.148Debris flows occurred November 23, 1990 in gullies Cl and C6.Total volumes were 1560 m3 and 900 m3 , respectively. Bothdebris flows originated in soils 2-3 metres deep, with thebedrock-till interface as the failure plane. Piping was evidentat the headscarps of both failures. Both failures originated onopen slopes.149CHAPTER 6. DISCUSSION6.1 Budget accuracyIn several of the gully sediment budgets, the budget errorexceeds the pooled error. Since the pooled error is the averagecumulative error from all sources, a budget error greater thanthe pooled error indicates either inadequate identification ofstorage zones or sediment transfers, or else severe errors inestimating or measuring budget terms. When individual budgetsare examined, both types of errors can be found. Severe errorin measurement terms is considered more common than inadequateidentification of storage zones or sediment transfers. Bothinput and storage terms require integration of pointmeasurements to volumes; the assumptions incorporated intovolume estimates may result in severe errors. Gullies within atreatment type tend to have the same type of error,consequently, error will be discussed by treatment type.6.1.1 Unlogged: gully C2Errors in gully C2 are primarily measurement errors. Year 2 hasthe greater budget error, which is almost five times the poolederror. In Year 2, the sediment trap was located in a culvertpool. No input to the channel was recorded, a result of wellvegetated sidewalls. The recorded volume of eroded channel150sediment does not account for all of the volume of sedimentoutput. Some erosion of channel sediment is recorded, but notenough to account for the sediment output. Channel banks aresometimes steep and relatively high (1 m) in this gully, sounnoticed collapse of a portion of the bank is possible.However, a more likely source of the sediment is the channelwithin the roadcut, which is steep, and apparently unstable.Since the sediment from the roadcut is not included in thebudget components, roadcut sediment delivered to the trap willresult in error.In Year 3, the sediment trap was repositioned above theroadcut. The budget error is smaller in this year, but stillgreater than the pooled error. The most likely source of erroris in the storage measurement, which shows an increase of 0.84m3 , despite an input of only 0.22 m 3 . Since the unloggedsidewalls show no signs of erosion in this gully, input isprobably accurate. Consequently, the storage term is morelikely to be in error. Although error is large in relativeterms, the actual volume amount of the error is less than 1.3m3 .6.1.2 Torrented gullies: C3, C5, and C6In gully C3, the budget error is excessive in Year 2, butacceptable in Year 3. In Year 2 the storage term is the mostlikely source of error, since it is the largest of all terms.151In 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 erodingarea were overestimated, the error would be multiplied in thevolume total.As a basin size increases, there is an increase in thepotential for temporary sediment storage (Schumm, 1977).Sidewalls in gully C5 are very long; consequently, it is notnecessarily correct to assume that sediment eroded at anylocation on the sidewall profile will reach the channel.Sediment eroded from upper portions of the sidewall may be re-deposited on the lower segments. In addition, it appears thatgullies C5 and C3 have another storage zone located between thesidewall and channel (Figure 6.1). This area is subject tofluvial inundation, but appears to store sediment between largeevents. Some selective removal of finer sediment was observedin these areas, but they were not monitored quantitatively inthis study. Since storm events were moderate in Year 2 and onlyone large event occurred in Year 3, significant amounts ofsediment could accumulate in these channel-margin zones. Thevery large output of sediment from gully C5 in Year 1 may be aresult of erosion of stored sediment from these areas.152Figure 6.1^Channel-margin storage zones in gully C5Channel margin storage zones defined by grain size: sidewallsare fine grained, the main channel is extremely coarse grained,and the channel margin zones are intermediate between thesidewall and channel grain sizes. Two channel margin zones arevisible in the photo: 1) the middle to upper left portion ofthe photo, and 2) the lower right portion of the photo. Stadiarod (1.1 m) in middle left of photo.1536.1.3 Slash-full gullies: C4 and C8.Budget error in these gullies is somewhat artificially set atzero. Since there is virtually no measured output, and storageis defined equal to input, the budget error is, at least partlyby definition, zero. Input terms for these gullies aremoderate, and probably close to actual amounts.6.1.4 Slash-clear gullies: C10 and C11.Slash-clear gullies show dramatic changes in Year 2, andmoderate changes in Year 3. Year 3 budgets are reasonablyaccurate. For Year 2, both gullies have erosion of channelsediment far above the measured outputs. At gully C10, thesediment trap was not installed until one month after cross-section monitoring began. Three large storms (24 hourprecipitations of 93, 126 and 143 mm) occurred before the trapwas installed. As a result, most of the sediment output fromchannel erosion was not measured. The sediment trap at C11 wasinstalled at the same time as the cross-sections; however, thelarge volume of sediment eroded from the channel overwhelmedthe trap, and some of the output was lost. Using channelstorage changes, actual output (fine and coarse sedimentcombined) from gully C10 is probably about 14 m 3 , compared tothe 3.6 m 3 estimated. In gully C11, actual output is probably11 m 3 , compared to an estimated 5.5 m 3 .1546.1.5 Budget error summaryThe estimated error for each budget term indicates that input,storage, and fine sediment output have the greatest errors. Inall but one case, either input error or storage error is thegreatest component of pooled error. In most gullies, estimatedinput or storage error is equal to or greater than the measuredcoarse output. Further refinement of these terms is necessaryfor greater accuracy in the sediment budget. In this study,fine sediment output is estimated indirectly. Increasedaccuracy would result from direct measurement of fine sedimentoutput. Coarse sediment output is generally adequatelymeasured, except in large storms which overwhelm the traps.Given the large budget errors, do the comparisons of budgetterms between treatment types still hold? For storage andcoarse sediment output, the comparisons are still valid.Storage terms are compared using changes in the area of eachcross-section; most error in storage terms will result fromapplying cross-section area changes along a length of channel.For output, comparisons are made for periods when sedimenttraps were operating effectively. This method avoids the majorsource of coarse sediment output error, which occurs whensediment traps are overwhelmed.Comparison of the input terms between treatment types uses theinput volume, normalized by sidewall area. Since input volumesare almost certainly overestimated in the torrented gullies,155the comparisons may not be valid. To compare input terms, thetorrented gullies' input terms are adjusted by the amount ofthe budget error. Table 6.1 shows the old and revised inputterms for each gully. Table 6.2 shows the old and revised inputmeans for each treatment class, with Years 2 and 3 combined.Although the revised mean input of torrented gullies is abouthalf of the old mean input, there are still significantdifferences between the treatment groups. The same conclusionsare reached for both the revised inputs and the old inputs:torrented gullies have the greatest input, slash-full gullieshave the next greatest input, and unlogged and slash-cleargullies have very little or no sediment input.6.2 Sediment storage and transfersTwo of the objectives of this study were to monitor the storageand transfer of sediment in gullies, and to investigate howslash and CWD affect the storage and transfer of sediment inthe gully channel. The study examined the input, storage andoutput of sediment from gully channels, and a brief discussionof each of these terms is given below.6.2.1 InputInput of sediment is clearly related to treatment type.Torrented gullies have the most input, and slash-full gullies156Table 6.1GullyOld and revised input terms for sediment budgetsOld Input New Input^Old Input^New InputYear 2 Year 2 Year 3 Year 3C2-U 0.03 0.03 0.22 0.22C5-T 24.1 16.9 8.74 8.74C6-T 12.1 3.5 8.69 3.47C3-T 1.00 2.00 0.92 0.92C4-SF 1.76 1.76 2.51 2.51C8-SF 0.39 0.39 0.49 0.49C10 -SC 1 0 0 0 0C11-SC 0 0 0 01)^Inputzero.for slash-clear gullies not monitored, estimated to beTable 6.2 Old and revised treatment group inputsTreatment n^Old Mean Old Var.^New Mean New Var.(mm) (mm2 ) (mm) (mm2)Unlogged 2 0.065 0.004 0.065 0.002Slash-full 1 6 0.546 0.067 0.615 0.101Torrented 4 6.18 13.7 3.45 1.41Slash-clear 4 0 0 0 01 C3 included in slash-full treatment group for input.157have the next greatest input. Both the unlogged and the slash-clear gullies have very minor input of sediment. The torrentedgullies have the greatest input of sediment because they havethe greatest amount of bare sidewall area. This is not a resultof the primary treatment effect (the amount of CWD or slashwithin the channel), but nonetheless is characteristic oftorrented gullies. Revegetation of the sidewalls of C3 withconifers has reduced the amount of bare area to an amountsimilar to that found in slash-full gullies about 15 yearsafter torrenting; as a result, C3 has sediment inputs similarto the inputs of the slash-full gullies.Sediment inputs in the slash-full gullies and slash-cleargullies are different for two reasons. The primary reason isthe yarding method used. Slash-full gullies were highleadyarded; slash-clear gullies were skyline yarded. Highleadyarding generally results in greater surface disturbance tosoils (Sauder and Wellburn, 1987). The second, and lessimportant reason for differences in input, is that placement ofchannel slash on the sidewalls in slash-clear gullies addedprotection to the sidewalls. However, both the slash-full andslash-clear gullies have sideslopes of low to moderatesteepness; therefore, input of sediment from logged gullieswith steeper sideslopes may be greater.In the unlogged gully, C2, input appears to originate entirelyfrom the channel margins. Sideslopes are completely coveredwith vegetation or litter; as a result, they are resistant to158rainsplash erosion. Trees indicate that active creep occurs inthe gully. The major process which delivers sediment to thechannel is probably creep to the channel margins, followed bybank collapse.6.2.2 storageStorage of sediment is different in most treatment groups.Slash-full gullies stored about the same amount of sediment astorrented gullies. Slash-clear gullies had large amounts ofchannel erosion in Year 2; in Year 3, very little changeoccurred, and was mostly deposition. The unlogged gully showedvariable response: in Year 2, the average cross-sectionresponse was erosion and in Year 3, minor deposition occurredat all cross-sections.Storage of sediment in the slash-full gullies will probablycontinue for some time. Gully C4 is very small, and probablyhas the least water discharge of all the gullies. Slash is verythick (about 1 m at the trap), and there is no surface flow ofwater. As a result, sediment delivered to the channel will notbe transported at all. Gully C8 is larger, slash loads aresmaller, and surface flows are sometimes visible. Localtransport of sediment has been observed in this gully, hence asinput of sediment continues, some output of material may occur.Storage of sediment in the torrented gullies is probably only159temporary, until the next large storm provides flows sufficientfor fluvial transport. The channels of gullies C3 and C5 appearto possess a stable configuration of large boulders andcobbles. Gully C5 had very large output of sediment in Year 1(25 m3 ), but the channel itself did not appear to change (basedon casual observation since cross-sections were not in place atthe time). Although gully C3 torrented about 10 years earlierthan C5, the channels are very similar. Consequently, torrentedgully channels appear to be stable features with little changeover time.Gully C6 is the most recently torrented gully, and since it islocated higher on the hillside than gullies C5 and C3, bedrockis closer to the surface. Irregularly sloping bedrock nowcomposes most of the channel of gully C6. Storage is limited tothose sections of the bedrock channel where the gradient islow, or else in sections where debris flow material wasdeposited along a protected sidewall (Figure 6.2). Smallsediment 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 theirchannels during the first year after slash-clearing occurred.In Year 3, little or no erosion occurred. Erosion of thechannel may have resulted from excessive disturbance of thegully channel by work crews. Both C10 and C11 did not have anactive channel before slash-clearing operations began. Channel160Figure 6.2^Sediment storage in gully C6Debris flow sediment is stored in middle right of photo,visible as a pile of boulders and CWD. The gully was seededafter the debris flow of November 23, 1990. Entrance of debrisflow is visible as uppermost seeded area on right side ofphoto. Channel slash is visible in upper third of the photo.161material was generally boulders and cobbles, but covered invegetation and litter; water flow was probably almost alwayssubsurface. After slash-clearing, surface flow of waterresulted in erosion, initially of fine organic debris andsediment, and later of larger material. Erosion did notcontinue in Year 3. In part, this is due to self-armouring ofthe channel; in part it is also due to the very low magnitudestorms in Year 3.Storage in unlogged gully C2 shows variable response. Thechannel of C2 is composed of sections of bedrock interspersedwith sections of sediment. CWD does not appear to be the maindeterminant of where sediment is stored; rather, it is thelocation of live trees, which act as anchors for CWD, thatdetermine sediment storage. At least two of these trees(minimum age 230 years) grow on bedrock, almost directly in thegully channel (Figure 6.3). Consequently, it appears that thisgully has not had a debris flow for at least 230 years.Sediment storage in this gully is limited, since it occurs onlyin restricted pockets. Channel sediment is very limited in thisgully for one of two possible reasons: 1) input of sediment islimited, or 2) CWD, in the necessary configuration for sedimentstorage, is limited. Limited input of sediment would suggestthat this gully is very stable, even though sections of thesidewalls are very steep. If storage is limited due to a lackof effective CWD, then fluvial transport of sediment may removeas output most sediment delivered to the gully channel. It is162Figure 6.3^Old trees on bedrock in gully C2The tree on the left of photo is directly on bedrock. The treeon photo right appears to be on bedrock as well. The channel isbetween the two trees, less than 5 metres apart. Stadia rod(2.1 m) leans against the left tree.163not known which possibility is the more likely, or whether acombination of factors is operating here.6.2.3 OutputCoarse sediment output is greatest in the slash-clear gullies,least in the slash-full gullies, and intermediate in thetorrented and unlogged gullies. Fine sediment output isgreatest in the torrented gullies, least in the slash-full andunlogged gullies, and intermediate in the slash-clear gullies.Both coarse and fine sediment output from slash-clear gulliesoriginate from channel deposits. Whether output of largeamounts 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 limitedchannel deposits, or else erosion of the sidewalls, a processnot yet evident. In gully C11, the channel is now armoured withlarge boulders in most areas. Migration of the channel ispossible in this gully, which could result in renewed sedimentoutput. 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 as164great as the output of C5 in Year 1 is not known. The torrentedgullies appear to have zones of temporary storage of sediment;these zones may have been flushed out in Year 1, therebyreducing yields in Years 2 and 3. Torrented gullies areprobably the ones most likely to have variable outputs, sinceinput from bare sidewalls and slumps can be large, and sedimentcan be temporarily stored along channel margins.The unlogged gully, C2, has an output of sediment similar tothat of torrented gullies C3 and C6. Most of this outputoccurred in two measurement periods, when the road culvert trapwas used. The roadcut appears to be a likely source ofsediment. Thus, C2 may actually produce less sediment than thetorrented gullies.The slash-full gullies clearly produce the least amount ofsediment: either no sediment, or almost no sediment. Slash ineffect creates subsurface flows; inputs of sediment to thechannel do not come in contact with fluvial discharges.6.2.4 Fluvial transport and debris flow transport of sedimentBoth fluvial transport and debris flow transport of sedimentoccur in gullies. Fluvial transport events are of low magnitudeand high frequency, whereas debris flows are of high magnitude,low frequency. Comparison of the two processes will indicatethe relative importance of each process.165To compare fluvial transport with debris flow transport, theaverage output of sediment from unlogged and torrented gulliesin Years 2 and 3 is compared with the average debris flowvolume from gullies Cl and C6. Fluvial outputs from bothtorrented and unlogged gullies are used, since there were noapparent differences between the sediment outputs in torrentedand unlogged gullies. However, sediment output from gully C5 isexcluded, since output from this gully is an order of magnitudegreater than that measured in most other unlogged and torrentedgullies. Average fluvial sediment output from the torrented andunlogged gullies is 2.1 m 3 per year; the average debris flowvolume is 1230 m 3 . Therefore, the average debris flow volume isequivalent to 600 years of fluvial output.The comparison of debris flow volume with fluvial transport is,of course, very approximate. Although no differences in outputare apparent between torrented and unlogged gullies, the C2gully may actually have a lower output than the torrentedgullies. Since input volumes are much greater in torrentedgullies than the unlogged gully, output is probably greater aswell. If this is true, then the average debris flow could beequivalent to much more than 600 years of output from theforested gullies. Average fluvial transport is determined fromonly three gullies, with just two years of data. Average debrisflow volume is calculated from only two debris flows. Bothaverages use small samples of processes which are extremelyvariable. For all these reasons, comparisons between fluvial166and debris flow output are very tentative.The sediment budget approach might provide furtherclarification of the relative importance of debris flow andfluvial transport processes. At any point in time, the totalsediment volume available for either fluvial transport ordebris flow transport could be defined as all the sediment inthe headwall, sidewalls, and channel of the gully. (Thisapproach would exclude sources outside of the gully, whichwould be very difficult to define). If sediment is defined asall material above bedrock, and if, over the course of a fewthousand years, the volume of sediment created from bedrockweathering is negligible, then all sediment delivered from bothfluvial and debris flow processes must come from this limitedsediment source. (Alternatively, a bedrock weathering ratecould be estimated). The volume of sediment transported by anyone process, relative to the total source volume, may clarifythe importance of debris flow and fluvial transport processes.6.3 Transferability of results to larger gulliesThe gullies chosen in this study were selected partly on thebasis of basin size. In fact, the gullies chosen are some ofthe smallest in the study area. Screen sediment traps havelimited storage potential; therefore, to trap most of thesediment output from the gully, output must be limited involume. This limitation was not as severe in gullies with road167culvert traps. Since the relative importance of processes maybe different in gullies of larger size, some consideration ofhow the results of this study would apply to gullies of largersize is necessary.Larger gullies have larger sidewalls. These sidewalls may havegreater disturbance from falling of trees into the gully, andfrom the yarding of logs out of the gully. Both slash-full andslash-clear gullies have small gully sidewalls. In C10, asection beneath the study reach has much longer sidewalls;disturbance to these sidewalls is more severe than disturbancewithin the study reach. Consequently, input in larger loggedgullies may be proportionally greater.Larger gullies have greater discharges of water, and thepotential sediment transport capability is also greater. Theslash-full gullies are small; discharge in these gullies is notlarge 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. Inthese larger gullies, sediment in the channel could also betransported, at least over short distances. Channel conditionscould be much different from those observed in gullies C4 andC8.1686.4 Study results: implications for forest managementTo be effective, management of forestry activities must achievean objective, and be cost effective. This study has included anassessment of the feasibility and effectiveness of cleaningslash debris from gully channels, with the objective ofreducing debris flow impacts on receiving stream channels. Twoaspects of debris flows may be affected: their initiation, andhence the frequency with which they occur, and the amount ofsediment incorporated into the debris flow - the debris flowmagnitude. For slash-clearing to be effective, it must eitherreduce the magnitude of debris flows in gullies which aretreated, or else it must reduce the likelihood of a debris flowoccurring in a treated gully.6.4.1 Slash-clearing effects on debris flow magnitudeLogging slash effectively traps sediment in the slash-fullgullies studied. Coarse sediment output in these gullies isessentially zero. Thus, the presence of slash will increase thevolume of sediment stored in the channel. Clearly, if a debrisflow occurs in one of these gullies at a later time, themagnitude of the event will be greater than if the slash werenot present. However, is this increase in magnitudesignificant? Input of sediment to the two slash-full gulliesaveraged 0.01 m 3 /m of channel length/year. Most failures inlogging areas occur within 15 years of logging. Increased169volume of sediment after 15 years would be about 0.20 m3 /m ofchannel. The Cl and C6 debris flows mobilized 4.70 m 3 /m ofchannel. Consequently, if a debris flow were to occur 15 yearsafter logging, it would increase the volume of sediment scouredfrom the gully channel and sidewalls by about 4 percent.Increased debris flow magnitude from slash appears to bepotentially significant. Slash in these gullies averaged about0.5 m deep and 5 m wide in the channel area. Assuming a voidratio of 0.5, the volume of slash in these channels is about1.20 m 3 /m of channel. Sediment mobilized in the gully C6 debrisflow from sidewalls and channel was 5.90 m 3 /100 m of channel.Slash incorporated into the debris flow probably increased thevolume of the C6 debris flow by 20 percent. This resultconsiders only gully sediment and CWD sources; if the openslope sources are included, slash increased the volume of theC6 debris flow by about 15 percent.Slash-clearing can reduce the volume of channel sediment whichwas stored prior to logging. Removal of logging slash resultedin erosion of the channel in the slash-clear gullies. Averageloss of channel sediment in gullies C10 and C11 was about 0.17m3 /m of channel in two years. Since the Cl and C6 debris flowsmobilized 4.70 m3 /m of channel, channel erosion in the slash-clear gullies would reduce debris flow sediment volume by lessthan 4 percent. If a slash-cleared gully is scoured by a debrisflow many years after cleaning, then there would probably be agreater reduction in the debris flow volume as a result of170channel erosion, but not at the rate measured in the first twoyears.6.4.2 Slash-clearing effects on debris flow frequencyFor the presence or absence of slash to affect the frequency ofdebris flows, there must be some effect on the initiation of adebris flow. Debris flows occur when a debris slide enters agully 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. Iflogging, or logging slash, alters the susceptibility of thegully channel to mass movement, then there should be a changein the average number of debris slides required to initiate onedebris flow. The nearly equal ratios of debris slides to debrisflows in clear-cut and forested areas suggest that logging, andlogging slash, do not affect channel susceptibility toinitiation of debris flows.Channel debris does not appear to be more susceptible to debrisflows if slash is introduced, but does erosion of a slash-cleared channel change susceptibility to debris flows? In gullyC10, removal of almost all sediment has resulted in anextremely stable bedrock channel. Conversely, if erosion issevere, undercutting of sidewalls may result, increasing the171rate of sidewall failures. This may result in increasedincidence of debris flows.6.4.3 Assessment of effectiveness of slash cleaningSlash-clearing occurred in two stages in the Coquitlam gullies.The skyline was initially used to haul large slash pieces outof the gully, after which a hand crew carefully removed allslash larger than 1 cm diameter. The skyline crew was composedof 3 chokermen, one operator of the winch, and one man at theroad. The hand crew was composed of 5-6 workers. Each phase ofthe operation took a day. This level of cleaning is more thanwould be considered in any industrial operation, but provides aguideline as to the cost of slash-cleaning.As stated before, to be effective, slash-clearing must eitherreduce the magnitude or frequency of debris flows. Slash wouldincrease the magnitude of sediment in a debris flow by only 4percent, if a debris flow occurred 15 years after logging.Slash-clearing will reduce the magnitude of sediment in adebris flow by 4 percent, 2 years after slash-clearing. Ifslash is retained in the gully, the volume of the debris flowwill increase by about 15 percent from added CWD.Slash-clearing cannot be expected to reduce the frequency ofdebris flows, since there is no evidence that the presence ofslash affects the incidence of debris flows. There may be an172increase of debris flow frequency in slash-cleared gullieswhere channel erosion is severe enough to undercut sideslopes.If the objective of slash-clearing is to reduce the volume ofsediment incorporated into debris flows, or to reduce theincidence of debris flows, then the benefits are either toosmall, or perhaps even negative, to warrant the cost of slash-cleaning. On the other hand, if an objective of slash-cleaningis to reduce the volume of CWD in a debris flow, then there isa real benefit associated with slash-cleaning.Effectiveness of the operation must be assessed over a set ofgullies as a unit. Not all gullies may be susceptible to debrisflows; yet unless we can identify sensitive gullies, slash-clearing would need to be carried out on all gullies to achieveany benefits. In return, we run the risk of increased erosionfrom the slash-cleared gullies.6.5 ConclusionThe objectives of this study were to monitor sediment storageand movement in gullies, and to assess how logging and slashaffect sediment storage and movement. Four classes of gulliesare studied: 1) Unlogged, with the channel naturally loadedwith CWD, 2) Logged and torrented, 3) Logged, channel loadedwith slash, and 4) Logged, with slash cleared from the gullychannel. One unlogged gully, three torrented gullies, two173slash-full, and two slash-clear gullies were monitored.Sediment budgets composed of channel input, storage, and outputwere constructed for each gully. A parallel study is occurringin the Queen Charlotte Islands, and the results may differ fromthose observed in the Coquitlam basin study.A relationship between storm intensity and sediment outputcould not be established; many problems exist in representingstudy site gully water discharge using precipitation gaugeslocated several kilometres away. A minimum precipitationthreshold for sediment movement was established; this is verysimilar to the threshold for shallow landslides and debrisflows determined by Caine (1980). A 24 hour precipitation of atleast 92 mm is required before sediment output will occur frommost gullies in the study area. In Year 1, 10 events exceededthis threshold; in Year 2, 12 events, and in Year 3, 3 eventsexceeded the threshold.Two very large storms occurred in November of Year 1. The stormof November 9-12, 1990 yielded 226 mm of precipitation in oneday. This is the second greatest daily precipitation recordedat the Coquitlam Lake station, which has operated since 1924.The second storm occurred on November 23, which delivered 156mm of precipitation at the Coquitlam Lake station in one day,and 189 mm at the Coquitlam River station. Snowmelt probablyoccurred during this second storm. Nineteen slope failuresoccurred during this storm, most within Cedar Creek basin andon the west shore of Coquitlam Lake opposite Cedar Creek. Two174of the study gullies, one unlogged, and one slash-full, werescoured by debris flows during this storm. Sediment output fromthe debris flows was equivalent to about 600 years of fluvialsediment output.Sediment budgets developed for each gully show some cleardifferences between the treatment types. Table 6.3 summarizesthe differences; treatment types in brackets are notsignificantly different.Input is greatest in torrented gullies because the amount ofbare ground is greatest. If revegetation of the sidewallsoccurs, input will decrease. Gully C3 has been revegetated, andthus input is similar to that of the slash-full gullies. Theunlogged and slash-clear gullies have very little or nosediment input since the sidewalls are well vegetated. Storagechanges are not large, except in the slash-clear gullies.Torrented gullies showed moderate but consistent deposition.Slash-full gullies stored all sediment inputs. Channel erosionin the slash-clear gullies produced the greatest output.Torrented gullies, particularly C5, had the next greatestoutput of sediment. Output from the unlogged gully may be lessthan the output from the torrented gullies. Since slash-fullgullies trap all sediment in the channel, they have essentiallyzero output.175Table 6.3 Ranking of sediment budget terms by treatment typesBudget^Greatest^ LeasttermInput^Torrented1 Slash-full^(Unlogged and Slash-clear)Storage2 Torrented 3^Unlogged Slash-clearOutput^Slash-clear (Torrented and Unlogged) Slash-full1 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 unloggedThe management decision of whether to clear slash from a gullymust weigh the possible benefits against the potential costs.Change in the amount of sediment incorporated into a debrisflow is probably not significantly affected by either retainingslash in logged gullies, or else removing it. Debris flow sizemay increase by about 15% if slash is retained in the channel,a result of increased amounts of CWD. If a gully is slash-cleared, and no debris flow occurs, then there exists increasedrisk of sediment input to the channel from the undermining ofsidewalls, as well as some possibility of a debris flow as aresult of slash clearing. Since slash-clearing may have somenegative consequences when used in a gully which would nototherwise produce a debris flow, effective use of gullycleaning should be restricted to gullies which have a highprobability of mass instability.176REFERENCESBenda, L. 1990. The influence of debris flows on channels andvalley floors in the Oregon Coast Range, U.S.A. EarthSurface Processes and Landforms, vol 15, pp 457-466Benda, 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 initiationin Coastal British Columbia: An evaluation of therole of woody debris. Unpublished Working Planprepared for B.C. Ministry of Forests. 5p.Bovis, M.J. and B.R. Dagg. 1987. Mechanisms of debris supply tosteep channels along Howe Sound, southwest BritishColumbia. In Erosion and Sedimentation of the PacificRim (Proceedings of the Corvallis Symposium, August,1987). IAHS Publication No. 165, pp 191-200.1988. A model for debris accumulationand mobilizationin steep mountain streams. Hydrological SciencesJournal, vol 33, pp 589-604.1992. Debris flow triggering by impulsive loading:mechanical modelling and case studies. CanadianGeotechnical, vol 29, pp 345-352.B.C. Ministry of Forests, B.C. Ministry of Environment, Landsand Parks, Federal Department of Fisheries andOceans, and Council of Forest Industries. 1992.British Columbia Coastal Fisheries/ForestryGuidelines. B.C. Ministry of Forests, Victoria.102 p.Brown, C.B. and M.S. Sheu. 1975. Effects of deforestation onslopes. Journal of the Geotechnical EngineeringDivision, American Society of Civil Engineers, vol101 GT2, pp 147-165.Caine, N. 1980. Rainfall intensity-duration control of shallowlandslides and debris flows. Geografiska Annaler,vol. 62(A), pp 23-27.Chatwin, S.C. 1991. Measures for control and management ofunstable terrain. In A guide for management oflandslide-prone terrain in the Pacific Northwest.B.C. Ministry of Forests, Land Management Report No.18, pp 85-166.177Church, M.A., D.G. McLean, and J.F. Wolcott. 1987. River bedgravels: sampling and analysis. In Sediment Transportin Gravel-bed Rivers. C.R. Thorne, J.C. Bathurst, andR.D. Hey (Eds.) John Wiley and Sons, Inc. pp. 43-87.Church, M.A. and M.J. Miles, 1987. Meteorological antecedentsto debris flow in southwestern British Columbia; Somecase studies. In Reviews in Engineering Geology, vol7, pp 63-79.De Ploey, J., M.J. Kirkby, and F. Ahnert. 1991. Hillslopeerosion by rainstorms - A magnitude-frequencyanalysis. Earth Surface Processes and Landforms, vol16, 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. InErosion 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 smallcatchment in mountainous terrain. Zeitschrift furGeomorphologie, 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 drainagebasins. F.J. Swanson, R.J. Janda, T. Dunne, and D.N.Swanston (Eds.) USDA Forest Service, PacificNorthwest and Range Experiment Station, GeneralTechnical Report PNW-141. pp 5-23.Environment Canada, 1991a. November, 1990 heavy rains insouthern British Columbia. Prepared by E. Coatta,Atmospheric Environment Service, Vancouver, B.C.1991b. A second occurrence of heavy rains in southernBritish Columbia, November, 1990. Prepared by E.Coatta, Atmospheric Environment Service, Vancouver,B.C.Fannin, R.J. and T.P. Rollerson. 1993. Debris flows: somephysical characteristics and behaviour. CanadianGeotechnical Journal, vol 30, pp 71-78.Froelich, H.A. 1973. Natural and man-caused slash in headwaterstreams. In Loggers Handbook, vol 33, 9p.Hay, J.E. and T.R. Oke. 1973. The climate of Vancouver.Tantalus, Vancouver, 49 pp.178Hogan, D.L. 1986. Channel morphology of unlogged, logged, anddebris torrented streams in the Queen CharlotteIslands. B.C. Ministry of Forests, Land ManagementReport No. 49, 94 p.1992. Personal communication.Hogan, D.L. and J.W. Schwab. 1991a. Meteorological conditionsassociated with hillslope failures on the QueenCharlotte Islands. B.C. Ministry of Forests, LandManagement Report No. 73, 36 p.1991b. Stream channel response to landslides in theQueen Charlotte Islands, B.C.: changes affecting Pinkand Chum Salmon Habitat. In Proceedings of the 15thNortheast Pacific Pink and Chum Workshop. PacificSalmon Commission, Canada Department of Fisheries andOceans. pp 222-236.Horton, R.E. 1945. Erosional development of streams and theirdrainage basins; hydrophysical approach toquantitative morphology. Geological Society ofAmerica 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 forestengineering analysis of landslides in logged areas onthe QCI. B.C. Ministry of Forests, Land ManagementReport No. 43, 138 p.Linsley, R.K., M.A. Kohler, and J.L.H. Paulus. 1975. Hydrologyfor Engineers. McGraw-Hill. New York. pp 356-357.O'Loughlin, C.L. 1972. An investigation of the stability ofthe 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 inseverely disturbed watersheds, Queen CharlotteRanges, British Columbia. Can. Journal of ForestResources. 16: 1092-1106.Roddick, J.A., 1965. Vancouver North, Coquitlam, and Pitt LakeMap Areas, British Columbia. Geological Survey ofCanada, Memoir 335. 276 p.179Rollerson, T.P. 1984. Terrain Stability Study - TFL 44. LandUse Planning Advisory Team. Woodlands Services,MacMillan Bloedel Ltd. Nanaimo, B.C.1992. Relationships between landscape attributes andlandslide frequencies after logging: SkidegatePlateau, Queen Charlotte Islands. B.C. Ministry ofForests, Land Management Report No. 76, 11 p.1993. Personal communication. Vancouver, B.C.Rood, K.M. 1984. An aerial photograph inventory of thefrequency and yield of mass wasting on the QueenCharlotte Islands, British Columbia. B.C. MinistryForests., Land Management Report No. 34. 55p.1990. Site characteristics and landsliding inforested and clearcut terrain, Queen CharlotteIslands, B.C.. B.C. Ministry of Forests, LandManagement Report No. 64, 46 p.1992. Personal communication. Vancouver, B.C.Sauder, E.A, R.K. Krag, and G.V. Wellburn. 1987. Logging andmass wasting in the Pacific Northwest withapplication to the Queen Charlotte Islands, B.C.: Aliterature review. B.C. Ministry of Forests, LandManagement Report No. 53, 26 p.Sauder, E.A., and G.V. Wellburn. 1987. Studies of yardingoperations on sensitive terrain, Queen CharlotteIslands, B.C.. B.C. Ministry of Forests, LandManagement Report No. 52, 45 p.Schumm, S.A. 1977. The fluvial system. John Wiley and Sons, NewYork. 338 p.Sidle, R.C., A.J. Pearce, and C.L. O'Loughlin. 1985. Hillslopestability and land use. American Geophysical Union.Water Resources Monograph 11, 140 p.Sidle, R.C., and D.N. Swanston. 1982. Analysis of a smalldebris slide in coastal Alaska. Canadian GeotechnicalJournal, vol 19, pp 167-174.Slaymaker, 0., M. Bovis, M. North, T.R. Oke, and J.R. Ryder.1992. The primordial environment. In Vancouver andIts Region. G. Wynn and T. Oke (Eds.) UBC Press,Vancouver. pp 17-37.Swanston, D.N. and D.E. Howes. 1991. Slope movement processesand characteristics. In A guide for management oflandslide-prone terrain in the Pacific Northwest.B.C. Ministry of Forests, Land Management Report No.18, pp 1-17.180Swanston, D.N. and F.J. Swanson. 1976. Timber harvesting, masserosion, and steepland forest geomorphology in thePacific 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 FluidMechanics, vol 13, pp 57-77.Thurber Engineering. 1991. Geotechnical assessment of 1990-1991landslide events in Greater Vancouver Water DistrictWatersheds. Unpublished report to Greater VancouverWater District. 26 p.Tripp, D.B. and V.A. Poulin. 1986a. The effects of mass wastingon juvenile fish habitat in streams in the QueenCharlotte Islands. B.C. Ministry of Forests, LandManagement Report No. 45, 48 p.1986b. The effects of logging andmass wasting onsalmonid spawning habitat in streams in the QueenCharlotte Islands. B.C. Ministry of Forests, LandManagement Report No. 50, 29 p.1992. The effects of logging and mass wasting onjuvenile salmonid populations in streams on the QueenCharlotte Islands. B.C. Ministry of Forests, LandManagement Report No. 80, 38 p.Toy, T.J. 1983. A comparison of the LEMI and erosion pintechniques. Zeitschrift fur Geomorphologie, Supp. 46.pp 25-34.Varnes, D.J. 1978. Slope movement types and processes. InLandslides, analysis, and control. R.L. Schuster andR.J. Krizek. (Eds.) National Academy of Sciences,Transportation Research Board, Special Report 176, pp11-33.Wright, J.B. and C.H. Trenholm. 1969. Greater Vancouverprecipitation. Atmospheric Environment Service,Technical Memo No. 722. 36 pp.Young, S.E. 1992. Slope stability prediction techniques forforest management purposes - A case study from theQueen Charlotte Islands, British Columbia. Master'sthesis, University of British Columbia, Department ofGeography, Vancouver, B.C. 150 pp.Ziemer, R.R. 1992. Effect of logging on subsurface pipeflow anderosion: coastal California, USA. In Erosion, DebrisFlows and Environment in Mountain Regions(Proceedings of the Chengdu Symposium, July, 1992).IAHS Publication no. 209. pp 187-197.APPENDIXThis appendix includes data from the erosion pins, cross-sections, and sediment output traps. Erosion pin data are theamount of change, in millimetres, recorded at each pin in eachmeasurement period. Cross-section profiles are shown at threetimes: autumn 1991, May 1992, and May 1993. Sediment outputvolumes are shown for each measurement period.181Gully C2^Unlogged: Erosion pinsErosion amounts in mm, positive numbers indicate deposition^ End date of measurement period (d/m/yr) ^Pin Bank 8/8/91 3/10/91 16/11/91 25/5/92 4/9/92 19/11/92 25/5/9337 L 2 2 0 -2 4 1 038 L -2 0 3 3 -2 0 339 L -1 4 0 1 9 -12 340 L -2 -1 3 2 -4 -2 441 L 1 0 0 0 -2 0 142 L -1 3 -7 9 -2 -2 043 L 0 5 -4 5 2 -2 244 L -1 0 0 0 1 0 245 R -3 11 -8 -3 -3446 R 0 5 -2 -447 R -1 -28 -548 R -1 2 -8 19 -2 -3 -6049 R -1 4 -7 8 -5 -150 R -3 6 -6 8 0 -1151 R 2 -3 10 1 -2 -9 652 R -1 -3 6 0 12 9 253 R -16 20 -12 15 -2 1454 R -10 14 -3 6 2 -3 1355 L -7 2 -1 6 -4 0 356 L 3 -1 -4 8 -4 -3 357 L -1 3 0 -1 -2 10 -1358 L 2 1 5 3 -4 -11 1859 L 1 3 -1 4 0 5 -560 L 17 -4 -1 3 -2 -1 -425/5/92 Pin 47 slump 0.5m * 0.5m * 0.117m = 0.029 m3. 1-,25/5/93: Slumped pins 46 + 47:^0.5m * lm * 0.160m =^.08 m3. 00N25/5/93: Slumped pins 49 + 50:^.5m * lm * 0.172m = 0.086 m3.Gully C3 Torrented: Erosion pinsPinErosion amounts in mm, positive numbers indicate deposition^  End date of measurement period (d/m/yr) ^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 25/5/9333 R -8^-36^2^64^-72 1^0 -1 -134 R^13 2 10 -3 -3 3 8 7 -16 1035 R 10 3^10^-10 1^3 1 236 R 7^-11^0 5^-17 10 11 -3 1 -1937 R^-17 -3^24^-34 1^-2 0 0 -1538 R 25^33^-47 2 15 3 3 -3 -2 -3139 R 3^-73 3^35^-100 4^-16 65 240 R^12 26^-26 0 -1 4 1 10 -10 341 R 10^5^-3^1^12 0^-3 0 -6 -2242 R^1 15^-29 12 14 -3 5 -3 13 -3143 R 6^28 28^044 R -345 R^12^7^5^5^-46 1^-2 3 -1 4246 R -1 -5 6 -8 12 7 -28 34 -1147 R 50^-27^-4^6^-2 3^8 -5 0 1348 R^-19 3 7 3 -6 1 3 0 -3 222/05/91:^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.Gully C4 Slash-full: Erosion pinsErosion amounts in mm, positive numbers indicate deposition--- End date of measurement period (d/m/yr)^--Pin Bank 19/10/91 8/12/91 25/5/92 4/9/92 25/5/9371 R 16 -7 1 -8 -572 R 8 -16 -3 -10 -2373 R -2 -4 -5 -9 474 R 1 -5 -9 -7 575 R 7 -13 -12 10 -1576 R 8 -20 -18 -4 877 R 13 -2 -2 -4 478 R 8 -36 -39 -3 -979 R -18 27 -9 -8 -1080 R 1 6 -10 -3 -1681 R 4 2 1 2 -1382 R 1 0 3 -17 -2283 R 3 -4 16 1 -3584 R 5 -5 -14 -13 3285 R 5 -7 -11 -4 -6086 R 1 -22 -57 725/5/93: Pin 86 slump lm * .5m * .134m = 0.067m3.Gully C5 Torrented: Erosion pinsErosion amounts in mm, positive numbers indicate depositionEnd date of measurement period (d/m/yr) ^Pin Bank 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 25/5/931 L 15 -52 5 4 -7 -19 -3 -4 -3 112 L 6 -8 9 14 -3 -8 2 4 18 53 L 26 -12 4 3 -1 -23 1 -2 0 54 L 12 -14 6 5 -10 -5 0 4 -7 -35 L 34 -15 2 0 -5 11 -66 13 -2 306 L -4 2 3 9 -11 -31 -19 3 -18 -1357 L -9 12 11 5 -20 -16 -9 9 -9 428 L -3 14 6 -5 18 2 -18 8 -4 69 L 45 10 2 -2 2 -23 -10 -12 -37 4710 L 0 -35 -4 -5 11 -25 8 27 -5 111 L 6 -15 7 8 -5 -60 4 -12 4 5612 L -30 37 0 -7 -3 -16 12 -19 38 -4413 L 28 -6 -16 12 2 -10 -13 11 -6 -1314 L 10 -23 -5 11 1 -2 0 -2 -1 215 L -3 -15 -5 9 -18 -32 7 -11 3 -2916 L 8 -210Gully C5 Torrented: Erosion pinsErosion amounts in mm, positive numbers indicate deposition^  End date of measurement period (d/m/yr) ^Pin Bank 31/10/90 22/05/91 8/8/91^3/10/91^16/11/91 1/2/92^26/5/92^4/9/92^19/11/92 25/5/9317 R 34^-5^BURIED?18 R 6^-7^20^10^-4^-5219 R 30^-34^10^-13 19 9^-17 -320 R 10 -14 9 4^7^-1 -3^-1621 R 14^-40^77^-62 17 1^26 -4222 R 16 4 -19 -1^-1^2 3^-523 R 2^13^-21^0 -3 -7^3 -524 R 6 21 -1 18^-5^-5 -11^-625 R -2^NOTE26 R -2 7^2^7^-5^5^127 R 228 R 4^ 1^-7^3^-21^-4^-529 R 1 2 -5 14 5 -20 -430 R 331 R -1^ 7^11^-26^-10^1^1832 R -14 8 2 -6 2 -7^-36NOTE: PINS 25 - 32: Treated as slump rather than surface erosion. Slump volume = 4m * 2m * .3m =PIN 18, 31/10/90 treated as slump of .1m3.2.4m3.25/5/93 Pin 6: slump lm * .5m * 0.135m = 0.068 m©3.Gully C6 Slash-full, Torrented: Erosion pinsErosion amounts in mm, positive numbers indicate depositionEnd date of measurement period (d/m/yr)Pin Bank 31/10/90 03/06/91 27/8/91 3/10/91 25/5/92 4/9/92 18/11/92 26/5/9349 R 6 -41 4 -2 -74 -5 -2 -6250 R 40 -16 -18 24 1 0 -251 R -1 -2 -5 1 -7 -46 3152 R 39 11 -50 33 8 0 -2353 R 10 -72 -3 -9 -3 -3 -3954 R 4 25 -10 -5 -44 -12 -255 R 25 -14 -7 -5 -26 1 0 -3556 R 13 -67 -45 5 4 -12 1057 R 14 -13 0 6 -18 9 -1358 R 5 -17 -1 5 -12 -9 759 R 16 13 17 -23 14 -1860 R 7 -4 11 -8 -3 -5 461 R -1 -3 3 -3 -15 2 -1062 R 7 -1 9 -29 -9 -9 -28Pin65666768697071727374757677787980BankLLLLLLLLLLLLLLLLGully C6 Slash-full, Torrented: Erosion pinsErosion amounts in mm, positive numbers indicate deposition^ End date of measurement period (d/m/yr)^^31/10/90^03/06/91 27/8/91^3/10/91^25/5/92^4/9/92^18/11/9223 6^3 -47 -8 18^-8 -3 -13^-14^123^-14^-10^-48 -6 113 -9 7 -38^-6^-72^-2^-15 -33 -1 720^-18 4^ -6^28^-103 -7^-19 -1^-29 15^28^-28 -1^-24^121^-19 -4^-24 -3 43 -44^-2 -27^1^-611^-58 15 -10 -2 -1-5 -41^-15^-89^-10^41^-3 -40 -33 -1 18-3 -22^-11 -7^-21^5-4^22 -53^-31 -8 -9-2 5^-14 -30^0^726/5/9326-148-13-820-40-1-212-96-55-26-107-38NOTE: 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 pinsErosion amounts in mm, positive numbers indicate deposition-- End date of measurement period (d/m/yr) -Pin Bank 29/8/91 25/1/92 17/9/92 18/11/92 26/5/9361 L 3 2 -23 1 -562 L 8 1 4 -1 -1863 L 0 -7 0 -5 664 L 7 -1 -5 -3 -265 L -1 3 4 -2 -566 L 0 -9 4 13 1167 L -8 13 7 -5 568 L -1 -6 1 0 1269 L 7 -8 18 -370 L 1 0 -18 6 -25Gully C2, XS-24.81904.6 -4.4 -i540 1 2^3Station (m)3.6 -3.45 ^4.8 -4.6 -i 4.4 -e 4.2 -- '-I4)ni^4-a)w,--1 3.8-3.6 -3.4 -3.2  0Gully C2, XS-32^2.5^3Station (m)0.5^1^1.5 3.5^4^4.5^5-1--- 3/10/91 ---*--- 25/5/92 ^x^ 25/5/93-1--- 3/10/91 ---*--- 25/5/92 ^x^ 25/5/933.50.5^1^1.5^2^2.5Station (m)3 40 4 . 5191Gully C2, XS-44.8 -4.6 -4.4 -4.2 -4-3.8 -3.6 -3.45Gully C2, XS-54.80 1 2 3^4Station (m)♦ 3/10/91^25/5/92     25/5/935 6 74.6-4.4 -004.2 -0:10 4-W3.8 -3.6I— 3/10/91^25/5/92 ^x^ 25/5/935 60^1^2^3^4Station (m)19/10/91 --*-- 26/5/92^)4^  25/5/936 70^1^2^3^4^5Station (m)192Gully C3, XS-154.8 -4.6• 4.4 -4.2 -co 4-W 3 . 8 -3.6 -3.4 -3.2Gully C3, XS-2♦ 19/10/91^26/5/92     25/5/931^2^3^4Station (m)---11--- 19/10/91^>PE^26/5/925x^ 25/5/93Gully C3, XS-4Gully C3, XS-3193i0^2^4^6^8^10^12Station (m)---41— 19/10/91 -->,E -- 26/5/92^x^ 25/5/937^8x^ 25/5/939x^ 25/5/932^4^6^8^10Station (m)4--- 16/11/91 --*-- 26/5/924.84.6 -4.4 -i 4.2 -4 -3.8 -3.6 -3.4 -3.2 -Gully C3, XS-53^0 1^2^3^4^5^6Station (m)4,--- 19/10/91 --*-- 26/5/92Gully C5, XS-21940 121954.8 ^4.64.4 -4.2 -g4-q0▪ 3.8 ->•3.6 -• 3.4 -3.2 -0Gully C5, XS-31^2^3^4^5^6^7Station (m)-I--- 16/11/91 --*-- 26/5/92^x^ 25/5/938Gully C5, XS-41^2^3^4^5^6^7Station (m)-4--- 16/11/91 * 26/5/92   25/5/93Gully C5, XS-5Gully C5, XS-64.84.6-_ 4.4 -Eg 4.2 -0-14)›co^4-wHILI3 . 8 -3.6 -3.41960^1^2^3^4^5^6^7^8Station (m)4--- 16/11/91 ------ 26/5/92^x^ 25/5/930^1^2^3^4^5^6Station (m)4--- 16/11/91 --,*-- 26/5/92^x^ 25/5/934.8 ^4.6 -4.4 -i 4.2 -4-3.8 -3.6 -3.4 -3.2197 C6, XS-130^1^2^3^4Station (m)4--- 3/10/91 --w -- 25/5/925^6x^ 26/5/93Gully C6, XS-23.2 1^1^ 11^,0.0^0.5^1.0^1.5^2.0^2.5^3.0^3.5Station (m)—I— 3/10/91 --w- 25/5/92 ^x 26/5/934.01985 Gully C6, XS-34.8 -4.6 -i— 4.4 -g0 4.2 -o0^4 -HW3.8 -3.6 -3.4^I^,^I0.0^0.5^1.0^1.5^2.0^2.5^3.0Station (m)---4--- 3/10/91 --* - 25/5/924.03.5 4.5 5.0x^ 26/5/934.00.5^1.0^1.5^2.0^2.5Station (m)4---- 9/10/91 --w - 7/5/92 x^ 26/5/933.0^3.5199Gully C10, XS-1Gully C10, XS-20.5^1.0^1.5^2.0^3.0^3.5Station (m)---4,--- 9/10/91 ---w -- 7/5/92^x^ 26/5/93Gully C10, XS-32000.5^1.0^1.5^2.0^2.5^3.0^4.0Station (m)-'- ^-- w - 7/5/92   26/5/934.6 ^4.4 -i 4.2-g0 4 -4 )>1-1 33.6 -3.40.0Gully C10, XS-40.5^1.0^1.5^2.0^2.5^3.0^3.5^4.0^4.5Station (m)---+.-- 9/10/91 --w - 7/5/92^x^ 26/5/935.^1.5^2.0^2.5Station (m)9/10/91 --w-- 7/5/92^x^ 26/5/93Gully C11, XS- -0 4.5-04.)4.4-4.3 -4.2-4.1^0.0^0.5 4.0201^4.8^4.7 -4.6 -4.5 -4.4-4.3 -4.2-4.1 -4-3.9 -^3.8^0.0^0.5^1.0 1.5^2.0^2.5^3.0^3.5^4.0Station (m)9/10/91 - w - 7/5/92^x^ 26/5/930-d4)rd>04.5Gully Cii, XS-24.0 6.03.0Station (m)----1--- 9/10/91 --w-- 7/5/92^x^ 26/5/93202Gully C11, XS-3Gully C11, XS-44.84.6 -3.8 -3.6^■^0.0^1.0 2.0 3.0Station (m)----f-- 9/10/91 --w 7/5/92^x^ 26/5/934.0 5.0 6.0Summary of sediment deposition volumes (cubic metres) at the trapsSediment accumulated at traps from previous measurement date^ Date (dd/mm/yr)^Gully 27/10/90 31/10/90 17/11/90 2/12/90 29/3/91 22/05/91 8/07/91 8/8/91 28/8/91 4/9/91 16/11/91Cl 0.024 0.004 0.112C2 0.040 0.000 0.026 0.105 0.234 0.020 0.176 1.147 0.465C3 0.114 0.620 0.019 -0.008 0.452C4 0.005 0.000 0.000 -0.001 -0.001C5 2.369 0.000 7.583 11.503 3.960 0.060 0.778 0.240 0.322 0.848 0.410C6 0.051 1.209 0.284C8 0.010 0.000 0.000 0.000C10CliNotes1): Cl and C6 torrented Nov. 23/90.2): C3 full 17/11/90.3): 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.Summary of sediment deposition volumes (cubic metres) at the traps continued.Sediment accumulated at traps from previous measurement date^ Date (dd/mm/yr)^Gully 21/11/91 8/12/91^21/12/91 11/1/92^25/1/92 1/2/92 8/2/92 26/5/92 4/9/92 18/11/92 25/5/93ClC2 0.000^-0.021^0.004 0.033 0.150C3 0.216^0.003^-0.004 0.378 0.136 0.050 0.005 0.241C4 0.002 0.002 0.000 0.000 0.000C5 0.327^0.000^0.009^0.062^0.232 0.000 0.801 0.024 1.479C6 0.135^0.022^0.005^0.051^0.144 0.000 0.026 0.004 0.183C8 0.020^0.000^0.000 0.000 0.034 0.000 0.000 0.000C10 0.019^0.361 0.044 0.350 0.108 1.498C11 2.710^0.230 0.017 0.315 0.000 0.000 0.160Notes12): C5 21/12/91 includes 16/12/91 measurement of .0013.13): C2 deposition area destroyed 11/1/92.14): C8 measurement 26/5/92 actually 7/5/92.15): C11 installed 2/11/91.16): C10 installed 2/12/91.17): C11 sediment measured Jan 11,^1992 = 2.7 m3, half est. to be mulch. Using accumulation of 1.318): 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