Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The influence of bedrock type on the magnitude, frequency and spatial distribution of debris torrents… Sterling, Shannon M. 1997

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

Item Metadata


831-ubc_1997-0119.pdf [ 13.51MB ]
JSON: 831-1.0087976.json
JSON-LD: 831-1.0087976-ld.json
RDF/XML (Pretty): 831-1.0087976-rdf.xml
RDF/JSON: 831-1.0087976-rdf.json
Turtle: 831-1.0087976-turtle.txt
N-Triples: 831-1.0087976-rdf-ntriples.txt
Original Record: 831-1.0087976-source.json
Full Text

Full Text

THE INFLUENCE OF BEDROCK TYPE ON THE MAGNITUDE, FREQUENCY AND SPATIAL DISTRIBUTION OF DEBRIS TORRENTS ON NORTHERN VANCOUVER ISLAND by SHANNON M . STERLING B.Sc. (Hons.) McGill University, 1993 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE (Department of Geography) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A February, 1997 ® Shannon M . Sterling In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) 11 A B S T R A C T This thesis examines the influence of lithology on debris torrent occurrence. The analysis covers a thirty-year period in 80 supply-limited basins distributed in the 400 km 2 Tsitika River watershed, on northern Vancouver Island, British Columbia. Two bedrock types occur in the watershed, the Igneous Intrusive and the extrusive Karmutsen formations, covering forty-nine and fifty-one percent respectively. The debris torrent source basins are unlogged. The frequency data were obtained in the field using dendrochronological evidence of debris torrents. Field data were compared with data derived from air photographs, the latter were found to be unrepresentative of debris torrent occurrence and were not used. Al l study basins were digitised from 1 : 20 000 Terrain Resource Inventory Maps (TRIM), and were characterised by selected morphometric parameters. Results show that geology exerts significant control over the temporal and spatial occurrence of debris torrents in the Tsitika watershed; the Karmutsen formation is more prolific. Geology also was found to exert significant control over the runout area and volume of debris torrents. Climate, morphometry and surficial materials do not appear to be confounding parameters. Differences in weathering rates, infiltration patterns and detrital grain-size distribution associated with the two bedrock types are believed to account for the differences in debris torrent behaviour. CONTENTS ABSTRACT LIST OF TABLES LIST OF FIGURES LIST OF GRAPHS LIST OF PHOTOS A C K N O W L E D G E M E N T S DEDICATION 1. INTRODUCTION 1.1 Research question 1.2 Mass movement and debris flow 1.2.1 Debris torrents 1.2.2 The debris torrent basin as a system 1.3 Inputs I: Controls on sediment delivery to the gully 1.3.1 Weathering 1.3.2 Sediment delivery to the gully 1.4 Inputs II: Triggers of debris torrents 1.5 Outputs I: Debris torrent deposits 1.6 Outputs II: Spatial occurrence, magnitude and frequency 1.7 Relevance of debris torrents I V 2. THE TSITIKA WATERSHED 23 2.1 Topography and geology 23 2.2 Glacial history 30 2.3 Climate 31 2.4 Vegetation 33 2.5 Rock weathering and soils 35 2.6 Surficial materials and mass wasting processes 38 2.7 Fires 40 2.8 Fish and forestry 41 3. METHODS 43 3.1 Study Design 43 3.2 Debris torrent basin as a system: topographical description 50 3.3 Inputs I: Sediment sources 55 3.4 Inputs II: Triggers of debris torrents 56 3.5 Ouputs I: Debris torrent deposits 58 3.6 Outputs II: Magnitude and frequency 60 4. RESULTS A N D DISCUSSION 69 5. CONCLUSIONS 102 REFERENCES 107 APPENDIX 1: M O R P H O M E T R Y P A R A M E T E R S (NOT USED) STATISTICS 115 APPENDIX 2: RUNOUT AREAS: 7° A N D 10° SLOPE DEFINITIONS 117 List of Tables 1. Characteristics of the Igneous Intrusive and Karmutsen Formations 29 2. Climate stations on northern Vancouver Island 32 3. Land tenure classification for the Tsitika watershed 41 4. Airphotographs of the Tsitika Watershed 46 5. Sample statistics 48 6. Information sources for intensive and non-intensive studies 50 7. Morphometric and Drainage Network Parameters Studied 51 8. Debris torrent volumes 64 9. Occurrence of debris torrent basins 69 10. Lithological composition of debris torrent deposits 71 11. Debris torrent runout lengths and areas 78 12. Dependance of runout area (y) on debris torrent volume (x) 76 13. Debris torrent occurrence in past 30 years 79 14. WSC gauging stations in the Tsitika Watershed 82 15. Maximum Discharge for Tsitika River, Catherine and Russell Creeks 86 16. Debris torrent occurrence per year 88 V I List of Figures 1. Parts of the gully system 7 2. Location of the Tsitika watershed 24 3. Bedrock geology of the Tsitika watershed 27 4. Biogeoclimatic zones of the Tsitika watershed 34 5. Distribution of clearcut areas in the Tsitika watershed 42 6. Areas not ground checked 44 7. Location of debris torrent study basins 49 8. Digitising resolution for a 2.5 km 2 study basin 54 9. Location Water Survey of Canada (WSC) Gauges 57 Vll List of Graphs 1. Lithological composition of debris torrents 72 2. Runout areas and torrent volume 77 3. 30-year debris torrent frequency: separated by bedrock type 77 4. Volumes of most recent event: comparison of bedrock type 80 5. Volumes of debris torrents: most recent torrent 80 6. Maximum daily flow for WSC gauge Tsitika River (08HF004) 84 7. Number of days with precipitation greater than 25 mm: Alert Bay (AES 1020270) 84 8. Extreme daily rainfall (mm) by year: Alert Bay (AES 1020270) 85 9. Monthly rainfall totals (1992) Russell and Catherine Creeks 85 10. Peak discharge for storm events, comparison of Catherine and Russell Creeks 87 11. Debris torrent occurrence per year: percentage of all torrents in 30 years 87 12. Debris torrent occurrence per year: comparison of lithology 90 13. The effect of basin area on frequency: separated by lithology 92 14. The effect of aspect on frequency: separated by lithology 92 15. The effect of the proportion of basin in snow-zone: separated by lithology 93 16. The effect of basin relief on frequency: separated by lithology 93 17. The effect of the number of first order streams on frequency: separated by lithology 94 18. Area of study basins separated by bedrock type 97 19. Aspect of study basins separated by bedrock type 97 20. Proportion of basin in snow-zone separated by bedrock type 98 21. Relief of study basins separated by bedrock type 98 22. Number of first-order streams in basin separated by bedrock type 99 viii List of Photos 1. Type II, transport limited, gully with 20 m. high ravelling till cliffs 9 2. U-shaped valley in the upper Tsitika watershed 25 3. Tree scar used for dating debris torrent occurrence: scar 3 years old 61 4. Clearcut on a debris torrent fan 65 5. 20 000 m 3 debris torrent: upper channel 66 6. 20 000 m 3 debris torrent (in photo 5) from aerial photograph (1:19 000) 67 7. Debris avalanche entering into gully 74 8. Bottom of debris avalanche from photo 7, gully and debris torrent deposition over road 75 ix A C K N O W L E D G E M E N T S This thesis has been a pleasure to work on and to write: this is indeed a feat, and its credit goes to the people who have worked with me and supported me through the process. A sincere thanks to NSERC and Forestry Canada for entrusting me with scholarship funding. Many thanks to Olav Slaymaker, a great supervisor: letting me run with the bit in my teeth, but always there to wisely provide the wider picture and to keep me on the track. Mike Church provided many wise, profound and helpful comments on the project. Les Lavkulich's enthuasiam, inspiration and encouragement were a huge help in the last leg of this thesis. The field seasons were productive and great fun. I thank my field assistants for sharing their love of the beautiful Tsitika Watershed and their true interest in the project: Claire Tweeddale, and John Fraser, with help from Geoff Green, Ian Rose, and Sharom Tafazoli. Rick Guthrie was a great support and friend in the field. MacMillan Bloedel has been generous with map and air photo access. Thanks to Jon Slater and Michel de Bellefeuille of Eve River Division, and Simon Lanoix of Woodlands Division helped with the airphotographs. Thanks to the M & B crew who helped us in the field when we got two flat tires or got stuck in an immense waterbar! Thanks to Catherine Griffiths, Tony Cheong and Rob Hudson for support in the Geographic Information Systems section. Analysis of tree cores was aided by George Brown and Dr. Renne at the Forestry Centre, Forestry Canada, Victoria Branch. Paul Commandeur helped out with field indicators of vegetation regrowth and with project design. Thanks to Rob Hudson who provided logistical and intellectual support. Friends have been a wonderful support during this time: Angela McCarthy, Rhys Evans, Craig Forcese, Katrina and Tanya Kucey, Brendan Bell, and Yvonne Martin. D E D I C A T I O N x These years of work and their product are dedicated to my mother, Lynne Sterling, who supported and worried throughout the field campaign, and who gave much encouragement during the write-up stage. She fought incredibly hard to live to see me finish this work, but did not make it. 1 1 Introduction 1.1 Research Question What is the nature of the connection between bedrock weathering rates and hillslope sediment transfer processes? In this thesis this question is examined in a narrower scope: are different weathering rates caused by different underyling bedrock type reflected in the magnitude, frequency and occurrence of debris torrents? The spatial distribution of debris torrents is strongly influenced by lithology in the Tsitika River watershed, British Columbia. The influence of lithology on magnitude and frequency of these debris torrent events over the past thirty years is the central theme of this thesis. The theme is approached in the following sequence: 1) to document debris torrent occurrence over space, 2) to show variation according to lithology in debris torrent occurrence over time, 3) to characterise the morphometry of the torrenting basins, 4) to show variation according to lithology in the morphometry of torrenting basins, and 5) to explore the effect of morphometry on sediment supply by lithology. The above takes form in three hypotheses: (1) there will be more basins which produce debris torrents in the extrusive bedrock area than in the intrusive bedrock area; (2) there will be larger and more frequent debris torrents in the extrusive bedrock type than in the intrusive bedrock area; and (3) the basins which produce larger numbers of debris torrents will have values of morphometrical parameters which are associated with larger sediment supply and runoff generation. This study examines the nature of debris torrents in basins where there is no human impact; the study basins are unlogged. Impacts of human activity may be interpreted in light of these results. Lithology is not the sole influence on debris torrent occurrence, of course. In this chapter we 2 introduce of other parameters in the debris torrent system which affect the relation between lithology and debris torrent occurrence. 1.2 Mass Movement and Debris Flow Gravity is the principal propulsive agency in mass movements. Mass movement includes four main categories: flow, slide, fall and creep. Debris flow is a type of mass movement which involves the movement of water and sediment, including organic debris, down a steep slope or a confined channel. The mechanics of debris flow differ greatly from turbulent open channel flow of water for the velocity difference between solid and fluid are not significant [Takahashi, 1981], and there is no separation of the sediment load into suspended load and bedload. Debris flows include a wide range of phenomena. Jordan [1994] identified a number of categories of debris flows which occur in different environments with different initiating mechanisms, amongst which rainstorm-initiated debris flows in steep creek channels in humid, forested environments are of interest for this study. Debris avalanches, another form of mass movement, are common on the west coast of North America. They differ from debris flows in that they have a much lower water content and, unlike debris flows, they are not capable of flowing as slurries under their own weight [Pierson and Costa, 1987]. 1.2.1 Debris Torrents Debris torrents are a form of debris flow which are both channelised and coarse-grained. Debris torrents are sometimes called 'channelised debris flows' [Evans, 1982]; however, fine-grained 3 channelised debris flows are not debris torrents. Debris torrents are composed of poorly-sorted, non-plastic, water-charged rock, soil (which lacks a fine-grained fraction, particularly clay), and a large organic debris content - in some cases up to 50 % [Swanston & Swanston, 1976, Church, 1983; Slaymaker, 1988], Debris torrents are highly fluid in character and move rapidly down steep, confined stream channels which are anywhere between 1st and 5th order [Oden, 1994], Debris torrents in the humid Pacific Northwest can be viewed as a variant of debris flow in which much fine material has been removed by fluvial washing of sediment during the gully charging process (see page 13) [Church, 1996 - pers. comm.]. Debris torrents are geomorphologically important, they are common, powerful, and well-connected to fluvial systems. Their occurrence is common in small, steep catchments in the Coast Mountains and Cascade Range of British Columbia, and in the coastal Pacific Northwest of the United States and southeast Alaska [Slaymaker and McPherson, 1977, Eisbacher and Clague 1984; Evans and Lister, 1984; VanDine, 1985]. Debris torrents have amazing erosive power. The sediment entrainment is irreversible, and solids and water flow together as a single-phase slurry [Costa, 1988]. In this way, debris torrents can transport a tremendous amount of material and thus be very destructive [Church, 1983]. Debris torrents appear to be the dominant mechanism of sediment transport from hillslopes to stream channels in coastal British Columbia [Swanston and Swanson, 1976], Much of the debris accumulates in sediment fans at the base of the source gully; where the fans intersect rivers and streams, debris torrents have important impacts on the stream and river morphology. Additions of debris masses modify the geometry and flood potential of the receiving river on the valley floor: on the upstream side accumulations of debris create lakes, swamps, or low-grade flood plains; at the toe of the fan the river channel is pushed against the far embankment and thus causes erosion, on the downstream side the coarse bedload added by debris torrents can create 4 an unruly braided reach [Eisbacher and Clague, 1984, p. 14]. Debris Torrent Studies in British Columbia Considering their geomorphic importance, we know little about debris torrents. It is only recently that debris torrents have been recognized as important geomorphic processes in humid, temperate regions [Swanston and Swanson, 1976]; this delay is likely due to debris torrents commonly being obscured by dense vegetation. Debris torrents also are thus far unpredictable. One complication is that the hydrological relations normally used to predict flood occurrence and magnitude cannot be applied to debris torrents [Miles and Kellerhals, 1981]; another complication is the large diversity of debris supply mechanisms [Dagg, 1987]. Also, separate phenomena are often not treated separately in the literature (for example, debris flows / debris torrents, and supply / transport limited torrents (see section 1-4.1)); because most investigators do not make the distinction, what is evidence for what is not a straightforward issue. Even sparser is the accumulated knowledge on the influence of bedrock type on debris torrent occurrence. In reviewing the literature, one finds that most information is in the form of comments in research which is pursuing other hypotheses. The influence of lithology on debris torrents has not been unequivocally demonstrated. At certain scales of investigations, or in some regions, the effect of lithology does not appear [see Rood, 1984], but there is little doubt that different rocks are susceptible to erosion at different rates and therefore susceptible to producing debris torrents with different magnitudes and frequencies. Because studies on the influence of bedrock type on debris torrents are not available, we consider the larger scale of the influence of bedrock type on mass movements. A1978 storm in the Queen Charlotte Islands triggered significantly different rates of mass wasting on various bedrock formations [Schwab, 1983]. Volcanic rocks showed the highest incidence of failure. Plutonic rocks ranked next in terms of failure frequency, but the slides were relatively small. In this study, Schwab [1983] find that heavily jointed and strongly fissured volcanic rocks produced the highest rate of large mass movements. Eisbacher and Clague [1984] and Evans [1984] note that large landslides are relatively more frequent in volcanic complexes than in other geologic environments. In view of the history of local studies we note the following characteristics: • emphasis on engineering components. Much work has focused on the debris torrents which present a hazard to highways and residential development in the Howe Sound and Fraser Valley areas [for example, Miles and Kellerhals, 1981; Thurber Consultants Ltd., 1983; Evans and Lister, 1984; Hungr et al., 1984, VanDine, 1985; Hungr et al., 1987]. These studies have tended to focus on the small scale engineering aspects of debris torrents, pursuant to the design of mitigation works along major highways. • poor regional coverage. In our geologically heterogeneous province only five general areas have been studied for debris torrents. In addition to the Howe Sound and Fraser Valley areas, discussed above, other sites of study include the Kicking Horse River valley [Jackson et al., 1989], and the Squamish and Lillooet River valleys [Jordan, 1994; Jakob, 1996]. Also, the Queen Charlotte Islands has been the study area for the joint Federal-Provincial Fish/Forestry Interaction Programme (FFJP), initiated in 1981. Much of our province's information on mass wasting is based on these studies from the Queen Charlottes. VanDine [1985] reviewed the state of knowledge of debris torrents in the southern Canadian Cordillera, in terms of the controlling factors, torrent hazard and hazard mitigation. • emphasis on the 'spectacular' debris torrents, or those which have claimed lives. • poor coverage of the variety of bedrock types which exist in the region. Most work has focused on coarse-textured debris torrents, mostly derived from relatively competent plutonic and metamorphic rocks, since these are the dominant lithologies in most areas where 6 engineering work has been done [Jordan, 1994]. The volcanics have only been recently recognised [e.g., Jordan, 1994] though not distinguished between debris torrents and flows. • poor coverage of the geomorphic and hydrological properties of individual torrent-prone basins [Bovis and Dagg, 1988]. • a lack of studies with a large sample size (more than 10 basins). 1.2.2 The Debris Torrent Basin as a System The torrent system consists of three parts, one which generates, another which translates, and a third which receives debris. The torrent system is intimately linked... to the generation of debris flow. The understanding of this natural unit is fundamental for mountain hazard appraisal. [Eisbacher and Clague, 1984, p. 15] The channels through which debris torrents travel are called gullies. A gully is a system composed of head walls, transport zones and a fan (figure 1), although the fan may be transient. We separate tributary gullies from the main gully. The frequency and magnitude of debris torrent events may be quite different from those in the main gully. As defined for this thesis, the main gully is always involved in the debris torrents; tributary gullies may or may not be involved in a debris torrent. We also separate recent debris torrent deposits, on which the debris torrent impact on vegetation is still recognisable, from older deposits composing the fan that has accumulated since deglaciation (Holocene fan), on which 'old-growth' forest would exist, assuming no other process has caused vegetative disturbances. There are two fundamental controls on debris torrent occurrence: the rate of sediment supply to the gully up to the threshold amount capable of forming a debris torrent, and the occurrence of a trigger capable of initiating a debris torrent. The first control is intrinsic to the system, the second is extrinsic to the system. Based on these two controls, a useful categorisation divides gullies into 8 type I (supply limited) and type II (transport limited) [Dietrich and Dunne, 1978; Benda and Dunne, 1987]. The system should take the classification of the trunk or main gully, not the tributary gullies (figure 1); a supply-limited system may contain transport-limited tributary gullies. Debris torrent production in type I systems is primarily limited by the sediment supply rate; in these systems when a threshold peak flow event occurs a debris torrent may or may not occur depending on the amount of material available in the gully. Type I systems are mostly bedrock controlled; a debris torrent cleans the channels of their accumulated debris, and in the period immediately following an event the probability of further debris torrents is very low due to lack of readily available debris in the channel [Kellerhals and Church, 1990]. In type I basins, the source of debris torrents is almost always recharge; thus weathering rates are important controls. One should note that these systems are not independent from the second, extrinsic control, but the interval of accumulation is significantly longer than the interval of recurrence of the threshold peak flow; after the threshold amount of sediment has accrued, there may be a period of 'waiting' for the threshold peak flow to occur. Debris torrent production in type II gullies primarily is limited by the generation of the threshold peak flow of torrent initiation in the gully; in these systems peak flow occurrences above the initiation threshold should always produce a debris torrent. Type II basins have channels incised into sources of deep, unconsolidated material (photo 1) or shattered bedrock. Sediment sources in type II systems can be destabilized by the occurrence of a debris torrent and can remain active for decades because there is a relatively unlimited supply of debris in the channel bed and banks [Kellerhals and Church, 1990]. Type II gullies may eventually convert into type I gullies after the source of sediment runs out, or after a long period of low flows allowing the slope to regain stability through vegetation succession [see Bovis and Dagg, 1987 for discussion of the supply mechanisms • 9 Photo 1: Type II Transport Limited Gully with 20 m. High Ravelling Till Cliffs 10 for the two types of gullies]. It is possible that a gully may bridge the two classifications, alternating between the two dominant controls. Such systems would have gullies which are cleaned to bedrock after every debris torrent, with the rate of gully recharge closely approximating that of the interval of climatic events capable of triggering a debris torrent. 1.3 Inputs I: Controls on Sediment Delivery to the Gully The rate of sediment delivery to the gully is a function of rates of weathering and rates of sediment transport of surficial materials. 1.3.1 Weathering Weathering can be a critical rate limiting factor for the movement of sediment in mountain environments [Dietrich and Dunne, 1978]. This may be particularly true in steep lands with shallow soils over competent bedrock [Slaymaker, 1985]. The study of weathering and its effect on availability of plant nutrients, soil development, and soil stratigraphy is advanced; however, the knowledge of weathering as a factor of sediment routing is not [Slaymaker, 1985]. Bedrock type and structure and microclimate are the most significant determinants of weathering effectiveness [Beaty, 1990]. For example, Sharpe [1974] illustrated the importance of lithology in determining weathering rates of debris input to mudflow source areas. Still, the importance of weathering depends upon the system in question. Church and Miles [1987] note that bedrock geology, and thus recharge rates due to rock weathering, provides a significant constraint only if rock weathering is the dominant affecting parameter. It thus is likely that weathering and thus bedrock type is more important in type I than type II gully systems [also, see Hungr et al, 1984]. The properties of rock which control its weathering rate are: 1) structure and porosity 11 [Douglas, 1976], 2) texture [Douglas, 1976], and, 3) mineral composition [Oilier, 1975]. Structure and Porosity The size and distribution of spaces in a rock have a profound effect on its weathering. The water-holding spaces in a rock will generally form an interconnecting network of channels through which water can move. The more that water can access a mineral, the more it will weather it. Thus, if a rock is porous and water can attack all grains, weathering will be faster than if the rock is dense and compact and water can only penetrate from the rock surface. With respect to granite and basalt, joints are common and very important in their weathering [Oilier, 1975]. Joints and, less commonly, faults, constitute lines of weakness and are followed by cliffs, gullies and major depressions [Ryder, 1978], Rock Texture Rock texture controls the amount of surface area on which weathering processes can occur. Differences in degree of crystallization and in size of crystals determine the texture of an igneous rock. In igneous rocks, the size of grains and their packing influences disintegration because the minerals will break down into different sizes. The smaller the grain size, the more surface area available for chemical reaction, and the faster the weathering. Eruptive rocks, cooling rapidly, (such as those in the Karmutsen formation) have small crystals as a rule. Plutonic rocks (such as the Igneous Intrusives), crystallising very slowly, grow large crystals that are plainly visible [Oilier, 1975]. Mineral Composition Minerals differ in their resistance to chemical attack. The mineralogy of igneous rocks can be broken down into two general groups: felsic and mafic. Mafic rocks have high proportions of pyroxene and olivine, and are dark coloured. They are susceptible to weathering, for pyroxene and 12 olivine are the most unstable minerals in the earth's surface conditions [Goldich, 1938]. Mafic rocks are most commonly extrusive, with smaller crystals [Allaby and Allaby, 1990]. Felsic rocks have high proportions of feldspars and silica, and are light coloured. They are composed of the more resistant minerals, such as quartz. Felsic rocks are primarily intrusive in origin, and have cooled relatively slowly. Lithological differences encountered in the study area would thus be due to 1) the Karmutsen Formation having smaller crystals than the Intrusives, and 2) the Intrusive formation containing minerals which are more resistant to weathering. The weathering susceptibility is thought to be an important control on sediment supply rates. 1.3.2 Sediment Delivery to the Gully Debris availability in a gully depends on the 1) character of creek bed and banks, 2) character of adjacent valley walls, particularly the type and distribution of bedrock and unconsolidated material, 3) vegetation type, and 4) land use. The volume of material in the creek bed may be a function of the bedrock type [VanDine, 1985]. Dagg [1987] identified debris torrent recharge sources in Howe Sound for several basins, finding a large diversity of sources and in their relative importance among basins. The main debris sources in his study area are: 1) active talus, rockfall, and rockslide debris, 2) inactive or stable talus, 3) material from debris slides, soil wedge failures, and, 4) material delivered by snow avalanche. Other sources include material delivered by a debris flow in a sub-gully, or material delivered by wind thrown trees. Anthropogenic sources can include logging debris, road fill, or road failure material [VanDine, 1985]. 13 Given a sediment supply, the rate of its delivery to the gully is controlled by 1) slope, 2) type and distribution of bedrock and overburden, 3) vegetation, and 4) land use, both adjacent to the creek and in the drainage basin [VanDine, 1985]. The amount of material available for a debris torrent is the amount of sediment delivered to the gully minus the amount removed by fluvial transport in the channel. Fluvial sediment transport can operate at a high frequency in the gully [Millard, 1993]. To determine the relative importance of fluvial versus debris torrent sediment transport, Millard [1993] compared the fluvial and debris torrent total sediment transport in two Coquitlam Valley gullies, finding that the average transport by individual debris torrent events equalled 600 years of fluvial transport. During the debris torrent recharge period, fine sediment is preferentially removed from gully storage compartments, leaving coarser deposits to be removed by debris torrents [Oden, 1994], It is helpful to distinguish debris torrent material sources which come off the hillside and immediately cause a debris torrent from those which have fallen into the gully in previous weeks, months, or years and then are subsequently entrained into the debris torrent. The former would likely contain more fine sediment, and more organic material, because of less fluvial sorting. The relative proportion of these two types of sources, along with bedrock type, may determine the amount of fines in the debris torrent. The grain size distribution of source material is related to its parent rock. The variation in grain sizes of sediment sources coming from different lithologies can be large. The grain size of the source material will influence thresholds of activation (see page 15), flow, and deposition dynamics. From theory we postulate that the Intrusives have a larger sand fraction, and the Karmutsen would have a larger clay sized fraction; this postulation is supported in the literature, for instance: • In one study of debris flow fans, Beaty [1990] observed marked differences in the size of debris on alluvial fans built by streams coming from areas of different bedrock. Fans below 14 granitic sectors of the range consist in the main of sand, cobbles and numerous larger boulders. Their surfaces are rough, with occasional blocks of extreme size far from the canyon mouths. Fans derived from metamorphic and sedimentary segments of the mountains tend to contain a large proportion of silt, sand, pebbles, and cobbles [p. 71]. • Jordan [1994] found that the Garibaldi Volcanics had a smaller grain size distribution and showed different rheologic properties than the nearby granitic basins. • Bovis and Dagg [1988] have noted that it is typical of diorites to have a very coarse lag material: "in the Coast Plutonic Complex in particular, the resultant lag deposits are very coarse-grained indeed, which promotes the survival of a relatively free-draining layer." Because grain-size affects the flow dynamics of the debris torrent, it may also affect the torrent's interaction with vegetation along the pathway, and therefore the visibility of the debris torrent from air photographs. Due to factors relating to grain size and amount of sediment available, bedrock type can influence the type of mass movement processes. This in turn can cause the dominant sediment recharge processes to vary with the bedrock type in a region [Millard, 1993; VanDine, 1985]. Soils developed on volcaniclastic materials are particularly conducive to creep and earthflow activity. Alternately, fresh granitoids are subject to rockfalls, rockslides, and block glides. Smooth surfaces of sheeted fresh granite encourage debris avalanches or debris slides in the overlying material [Laird, 1996]. In the literature, it was noted that there was a lack of fines in debris torrent deposits. Most the debris torrent studies have been in granitic bedrock areas. The lack of fines found in the deposits may be more a function of granitic origin than of selection of material larger than fines during the 15 debris torrent process. This should be investigated further. 1.4 Inputs II: Triggers of Debris Torrents Debris torrent triggering mechanisms can be hydrologic, structural or 'other'. Hydrologic triggers include: (1) rainstorm [Miles and Kellerhals, 1981, Caine, 1980; Takahashi, 1981; Innes, 1983; VanDine, 1985; Church and Miles, 1987], (2) rain-on-snow [Miles and Kellerhals, 1981; Caine, 1980; Takahashi, 1981; Innes, 1983; VanDine, 1985; Church and Miles, 1987], (3) water release from subglacial or lake storage (jokuhlaups), and (4) high water table conditions on the valley bottom [Okuda et al., 1981]. Structural triggers include (1) logjam bursts, (2) rock fall, (3) instantaneous release of water dammed behind a deposit, and (4) piping of water through the deposit [Takahashi, 1991]. 'Other' triggers include (1) seismic shaking, (2) wind [Wilford and Schwab, 1981], (3) debris avalanches, and (4) wet snow avalanche impact. If a period between debris torrents increases, the magnitude of triggering event needed to start the debris torrent may increase for, as Bovis and Dagg [1988] state, "as the thickness of the rubble layer increases, so the discharge event required for failure must increase." Hydrologic Triggers The climatic scenarios which produce debris torrents are varied. Local (southwestern B.C.) climatic scenarios producing debris torrents have been summarised by Church and Miles [1987]: 1) locally concentrated rainfall, high antecedent moisture, no snowmelt. 2) widespread moderate rainfall and snowmelt, 3) heavy rain onto thawing ground with little snowmelt, 4) apparently unremarkable rain, rain on snow, or snowmelt. Magnitude-frequency analysis has shown that meteorological events associated with debris torrents 16 are usually not rare [Schaefer, 1983; Innes, 1985]. Church and Miles [1987] found 2-3 year frequency precipitation events associated with 50-100 year debris torrent events. Valley bottom precipitation data, however, may not be representative of rainfall inputs into debris torrent basins. Localised cells of intense precipitation in zones of strong orographic forcing may be embedded within otherwise non-extreme storm events. Strunk [1992] supports the view that the occurrence of debris flows is strongly influenced by localized and extreme intensity showers. These showers, unfortunately, cannot be accurately measured unless there is a dense network of rain gages in the study watershed. Nyberg [1985] and VanDine [1985] came to similar conclusions. Instantaneous peak discharge is the key variable in producing debris torrents [Bovis and Dagg, 1988]. High instantaneous peak discharge requires antecedent precipitation to saturate the basin, followed by heavy rainfall and possibly snowmelt [Schaefer, 1983; Slaymaker, 1988]. Basin morphometry parameters also are important regulators of instantaneous discharge; basin area, shape and topography affect the volume and the speed of transfer of the water to the point of interest. The infiltration rates of the basin likely also are important; less permeability provides more volume of water at a higher rate of transfer to the point of interest. 1.5 Outputs I: Debris Torrent Deposits Slope at deposition and distance of travel of debris torrents are important variables when assessing hazards on debris torrent fans. The critical slope, below which transport will cease, is difficult to define because of material and basin variability [Thurber Consultants, 1983]. Takahashi [1981] suggested that torrents generally maintain motion in mountain canyons steeper than 15°. Thurber Consultants [1983] found that most debris fans along Howe Sound, British Columbia, have gradients between 10° and 16°. Hungr et al. [1984] observed events which began deposition near 10-17 12 degrees, including torrents derived from both crystalline and volcanic rocks. Incorporation of abundant large woody debris (LWD) in the leading edge of a debris flow may result in deposition on even steeper slopes [Montgomery and Buffmgton, 1993, p. 43]. There also is a question whether channel confinement is a more or less decisive factor than bed slope angle. Deposition also occurs along the stream channel in the form of levees, which are caused by friction at the boundary of the flow. Debris torrent deposition is affected by debris torrent volume [Takahashi, 1991; Jordan, 1994], bedrock type [Church, 1983; Hungr et al., 1984; Jordan, 1994], basin topography and water content [Thurber Consultants, 1983; Johnson, 1970; Enos, 1977; Iverson, 1985]. Bedrock type determines the grain size distribution. Jordan [1994] found that the permeability of the debris (related to grain size), and hence its rate of consolidation, is an important factor controlling mobility. Basin topographical parameters which are critical to debris torrent deposition include slope [Takahashi, 1981] and the change from confined to unconfined channel. The basin bedrock type can affect all of the above parameters: through weathering rates, the basin bedrock type can affect volume; through rates of erosion, basin bedrock type can affect basin topography; and through the grain size distribution of the material comprising the debris torrent, basin bedrock type can affect the maximum water content of the debris torrent. Jordan [1994] investigated the impact of grain size distribution on debris flow behaviour. He found that fine-textured debris flows behave according to the Bingham flow model, while coarse-textured debris flows can be better described by a granular, or dilatant, flow model. He used a clay content of about 4% in the matrix (sub-4mm. material) to distinguish the two populations. 18 1.6 Outputs II: Spatial Occurrence, Magnitude and Frequency The spatial occurrence of debris torrent basins reflects an integration of all variables discussed above. The variables of debris torrent occurrence, magnitude and frequency provide a posteriori information of the controls of the debris torrents and of the nature of debris torrent systems. Magnitude is a function of (1) the size of the source area, particularly the length of the channel upstream of the deposition area to the points of origin, (2) the vulnerability of the source materials to be mobilised or rapidly eroded under flood flow conditions, particularly in the channel, (3) the volume of the mass triggering event (if there is one), (4) the volume of debris in the gully bed, and (5) the proportion of sediment delivered to the channel that is transported out of the gully by fluvial transport [Hungr et al., 1984; Millard, 1993]. Attempts to quantify the magnitude with basin morphometric parameters have not been successful. For example, Hungr et al. [1984] attempted to correlate debris torrent magnitude with drainage basin size. Field inspection revealed a large variability, even in the relatively small region of coastal B.C. They concluded that the correlation of debris flow magnitude with basin size could only work as a first approximation, and stressed the importance of channel erodibility in creating differences in debris yield rates. The total runoff volume of a debris flow was correlated to the peak discharge of that debris flow in several gullies surrounding the Sakurajima volcano [Takahashi, 1991]. A good correlation between the two variables is evident, but the wide scatter in these data also suggests that the relation may change depending on many factors such as the shape of hydrograph, the channel conditions, and the various flow characteristics. Supply limited and transport limited basins were not separated in this analysis. Dagg [1987] makes an interesting point that creeks which are able to transport channel debris 19 short distances downstream from the supply points are able to store much more material in marginally stable configurations than would otherwise be possible; these creeks may experience less frequent but larger debris torrents than streams not activitely redistributing their bedloads. Kellerhals and Church [1990] describe debris torrent occurrence (temporal) as a stochastic process, referring to a transport limited example where the channel is fed debris by an individual, very large failure in bedrock or unconsolidated material which may remain active for centuries or even millenia, then finally stabilize. This would produce very long-term episodic behaviour which is superimposed upon other shorter term patterns. Debris torrent occurrence is also dependent upon the time since the previous torrent [Slaymaker, 1988], The establishment of frequency-magnitude relations is a most difficult task. It has not been tackled by many researchers; only Innes [1985] and Johnson et al. [1991] have explicitly addressed this relation. Innes [1985] found that the frequency of debris flows declines exponentially with magnitude. I hypothesize that in supply limited basins, the magnitude may remain relatively constant. A certain amount of material must accumulate in the gully channel for there to be a sufficient hydraulic head to trigger a debris torrent; once this hydraulic head is reached, a relatively common small flood can trigger a debris torrent. This scenario would be consistent with the observations that two-year floods have triggered debris torrents [see section 1-6]. In transport limited basins, the debris torrent occurrence in one basin is likely more dependent upon climatological events, and the size of the debris torrent may change with the size of the climatological event, or the random size of the slope failure off the large supply of sediment which is triggered by the climatological event. To summarise, frequency and magnitude of debris torrents are dependent upon: 1. bedrock type through rate of sediment supply, infiltration rates and possibly basin morphology. 2. basin morphometry and drainage characteristics, which control peak flow, sediment 20 transfer, potential energy of flow and ability to trap small-cell rainstorms. 3. type and distribution of unconsolidated material in the basin 4. type and rate of recharge processes to the gully [Hungr et al., 1984; Bovis and Dagg, 1988]. 5. vegetation type and distribution in the basin 6. land use in the basin 7. magnitude and frequency of trigger events in the basin 8. time passed since the last debris torrent in the basin. Bovis and Dagg [1988] discuss that the occurrence of a debris torrent may in part be due to the time since the last debris torrent, in that the building up of material in the gully would require a larger trigger event for debris torrent initiation; in this way, this relation would be non-linear 1.7 Relevance of Debris Torrents The magnitude and frequency of debris torrents are of interest for assessment of hazards on fans and in steep channels, for investigation of sediment contributions to rivers, for forestry planners, and for design of engineering structures to withstand debris flows. Salmon Habitat Debris torrents may result in extensive damage to salmonid streams downstream of the gully. Salmonids require stable and clean gravel for spawning; once fry are hatched, they require a variety of habitats to survive both summer and winter stream conditions [Tripp and Poulin, 1986a, 1986b]. A primary component of stream habitat is coarse woody debris (GWD) which provides diverse stream morphologies and cover for juvenile fish. Tripp and Poulin [1986a and 1986b] studied the influence of debris torrents on fish habitat in the Queen Charlotte Islands. They found that debris torrents can scour stream channels for long distances, having severe impacts on both gravel and CWD in the stream. In some cases they found that spawning gravel is almost completely removed with the debris flow; in others the proportion of fine sediment becomes deleterious to egg survival. They also found 21 that streams subject to debris flows had poor salmonid egg and juvenile overwinter survival due to gravel scour and winter habitat loss. Forestry Land-use decisions related to both timber harvesting and lowland development should be made only after careful consideration of the potential dangers of recurrent erosion and deposition of debris torrents. Poor land use practices, especially logging road construction, are implicated as broad environmental influences on debris torrent incidence [Thurber Consultants, 1983]. Schwab [1983] made a comparison of mass movement activity and found that there was a 3.5-fold increase in failures in clearcut units over non-clear cut units on slopes of less than 36 degrees. Also, Rood [1984] found a 34 times increase in the frequency of landslides due to clearcut logging and the associated road building in an area of steep and generally unstable terrain in the Queen Charlotte Islands of British Columbia. Many of the landslides entered gullies and triggered debris torrents. Hazard Debris torrents are among the most dangerous natural hazards that affect humans and properties [Takahashi, 1991]. The important engineering subjects of debris flow/torrent research are the identification of hazardous ravines, prediction of occurrence time, estimation of hazardous areas on the fan, design criteria of structures, and warning and evacuation systems [Takahashi, 1991]. Debris torrents are a major natural hazard in many parts of British Columbia [Hungr et al., 1984]. Owing to their relatively common occurrence, and potential for long reach and high velocities, debris torrents present a major hazard to developments and structures built in the vicinity of mountain creeks and on alluvial fans [Miles and Kellerhals, 1981]. The danger may not be obvious, since the occurrence of debris torrents is irregular on any individual creek and long periods of dormancy often permit full reestablishment of forest cover over affected areas [Hungr et al., 1984]. 22 The rapid growth of urban regions and recreational facilities in British Columbia has caused the construction of roads, telecommunication devices and railways in areas susceptible to debris torrents. Kellerhals and Church [1990] make a good summary of the debris torrent hazard situation in British Columbia: As in many other mountainous areas, fans are often the only gently sloping surfaces below ridge top levels. To the untrained eye these fans may look like the most suitable sites fpr a wide variety of developments ranging from farming to housing, industrial facilities, or transportation routes. Needless to say, first impressions often deceive and fan surfaces are increasingly being recognized as potentially very hazardous. The indiscriminate development of fans has already led to considerable economic losses and loss of life. Along the Squamish highway, in British Columbia, between 1981 and 1984, five major debris torrents occurred, resulting in the loss of life of twelve persons and the destruction of or damage to nine bridges and six houses. The town of Port Alice on Northern Vancouver Island was inundated by 22 000 m 3 of debris torrent material in 1973 and again in 1975 [Nasmith and Mercer, 1979]. At Camp Creek near Revelstoke, a 76 000 m 3 debris torrent killed four people in 1968. At Britannia Creek in Howe Sound a debris torrent killed 37 people in 1921 [Geological Survey of Canada, 1993]. Japan is another area notorious for debris torrents. In July 1938, Kobe, Japan was hit by rainstorms that produced debris torrents that killed 461 people and destroyed 100 000 homes. Additional disastrous landslides occurred in 1961 and 1967. In September 1945 the Makurazaki typhoon blasted Kure and produced 'rocky mudflows' (debris torrents?) that resulted in death for 1154 people. In Japan 1257 persons out of 4598 who were killed by natural disasters during twenty years between 1967 and 1987 were killed by debris flows/torrents [Takahashi, 1991]. In Japan the ratio claimed by debris flows/torrents and the other sediment hazards such as cliff failures and landslides to all the casualties is increasing, because more and more people tend to live in hazardous areas at the foot of mountains [Takahashi, 1991]. 23 The Tsitika Watershed The setting for our research is the Tsitika watershed which is located on northern Vancouver The choice of this setting arose from six considerations: 1) there are equal areas of two bedrock types, with similar topography in both areas; 2) the watershed scale forms natural limits across which information transfers are relatively weak; 3) it is the least logged watershed on the east coast of Vancouver Island, and thus most suitable to a study of the 'natural' occurrence of debris torrents; 4) despite being relatively unlogged, there is very good road access which covers most of the valleys in the watershed; 5) other research studies have been conducted in the watershed (e.g., Maynard, 1991; Davidson, 1996; Guthrie, 1996, and Nistor, 1996); and 6) there has been recent public interest in the watershed due to its draining into Robson Bight, an ecological reserve noted for its orca population. 2.1 Topography and Geology The 400 km2 Tsitika watershed is located on the northeast coast of Vancouver Island (figure 2). The Tsitika River flows in a northerly direction and empties into Johnstone Straight. The main valley in the Tsitika watershed is a broad, coastal glacial valley (photo 2). Floodplains of various widths characterize the tributary valleys in the watershed. The Tsitika watershed has a Ferro-Humic Podzol soil landscape (Humic, Cryorthod, Humic Haplorthod) [Lord and Valentine, 1978]. The Tsitika watershed is located within the northwest-southeast oriented Vancouver Island Range of Mountains [Howes, 1981] and surrounding the watershed are high peaks, including Tsitika Mountain (1648 m), Mount Ashwood (1734 m), Mount Elliott (1575 m), Mount Hapush (1691 m), Mount Russell (1666 m), and Mount Derby (1636 m). The rugged topography of the Vancouver Island range is the result of pre-Pleistocene uplift and dissection of a Tertiary surface [Holland, Photo 2: U-Shaped Valley in the Upper Tsitika Watershed 26 1976]. A large batholith of Island Intrusions underlies the upper 49.2% of the watershed. The lower 50.8% of the watershed is underlain by the extrusive Karmutsen Formation (figure 3). The Karmutsen Formation The Karmutsen Formation is part of the volcanic rich Mesozoic section which characterizes much of western British Columbia, Canada. Volcanic activity which produced the Karmutsen Formation occurred during early stages of the Cordilleran eugeosyncline. The Karmutsen Formation is part of a very thick and complete succession that is neither highly metamorphosed nor strongly folded, and can best be described as a classical suite of eugeosynclinal rocks [Surdam, 1971]. The Karmutsen Group consists mainly of basaltic to andesitic amygdaloidal flows, pillow lavas, pillow breccias, aquagene tuffs, and a few interlava sedimentary rocks. In the Tsitika watershed there are very little of the interlava sediments; where they do occur, they consist of limestones and fine-grained limey tuffs [Surdam, 1971]. Stratigraphically above the interlava sediments a sequence of pillow lavas is usually developed. The sequence can range in thickness from 30 metres, as in the two upper pillow lava horizons, to 3000 metres, as in the basal pillow lava unit [Surdam, 1971]. Where pillow lavas became unstable and slumped, pillow breccias occur [Carlisle, 1963]. Pillow breccias are composed of all sizes of pillow fragments set in a tuffaceous matrix, and they exhibit graded bedding [Surdam, 1971]. Toward the top of the basal unit a few ropy massive volcanic flows are intercalated with pillow lavas and breccias; these thin-layered flows are relatively coarse-grained and very amygdaloidal [Surdam, 1971]. Thin-layered flows grade upward into thick-layered flows ranging in thickness from 20 to 50 feet thick. The thick-layered flows are finer grained and contain a much smaller volume of amygdales than the thin-layered flows [Surdam, 1971]. * modified from Geological Survey of Canada Map 1552A 28 The Igneous Intrusives The Island Intrusions, composed of mainly granodiorite, and quartz monzonite, quartz diorite and quartz feldspar-porphyry, were intruded into the Karmutsen complex in Middle to Late Jurassic time [Muller, 1970]. When the intrusive rocks were emplaced into the Karmutsen Group they were confined to narrow vertical zones; they usually have nearly vertical contacts with Triassic volcanic rocks. Faults and Joints Fault and joint occurrence appears similar for the two bedrock types. The main faults run along the valley bottom, and are evenly distributed between the Igneous Intrusives and the Karmutsen Formation (figure 3) [Howes, 1981]. Rollerson [1996, pers comm] noted that the jointing in the Igneous Intrusives was similar to the Karmutsen Formation. Rollerson [1996, pers comm.] noted in one area exposed from a failure in the Karmutsen volcanics that there was 80% of shear in outcrop area, and that it was highly fractured. Shear zones are a result of tectonic faulting. A moderate fracture spacing, at one metre, was found in both bedrock types [Rollerson, 1996, pers comm]. Table 1 summarises the characteristics of the two bedrock types in the Tsitika watershed. 0) 3 +•» X CD CO X a W5 e o « E i. o to e v. ie +•» S S « a 03 CU > 3 1-S O a s OX) (—1 o U I-CS 3 S3 H IA c o n E I-o u. •R -O (fl c c .E » 1= © c o HE « j | s o o o * e 1 8 i a: o * * o 12 S J co •s «> *: — o a) <j> JQ i 8 » $ ^ T3 J5 •T= ^ X< a> - Q c O O - Q i Q . S Q . a) (0 ° § o c " a> co CD o "o a- > c o CO i i £ co «= co co •° E eg o g 5 8 6 -.2, |S E o) »_ (0 o o ^ >• > = = o E B JP. a a o co o CO to CO Jr. co ±± CD (U „ „ , O 0) — • l i f t CD • .-s g s ± i CO J= Dl <o co "> co WO J 3 " E g CO c - Q CO CN (A 2 § £ 5£ co E « 1 C X) CO > £ CO CD o "-5 E 1— y— '— ro >— rat o-jj j) i i Q_ <z a. ed ling +-» x> c "O ]o CD JO O) >> O) +J =3 c c= o aci tro -*—' ' -> O - CO co 1 1 o ^ CO CO • c o CO 3 CD > CO CO co E g CD co := CO CD CO o co f f i l e 1 e 1 -B i f co o m s > 8 | c is o> c CO co O E CO o a 3 co CD CD 0 co T I CO o co-in cn CD 0 f CO cn g 'c c O 00 00 CO <r- CD . CD f i 1 2 Q O CO CD o O CO 29 c c 0 0 c c o o Q_ Q . E E o o o o l _ 1— o o E E 30 2.2 Glacial History Two glacial intervals and one nonglacial interval have been identified on northern Vancouver Island [Howes, 1981]. During the Fraser glaciation (most recent), several valley glaciers initially coalesced in the main Tsitika valley and flowed northward to Johnstone Strait. At glacial maximum, after 20 600 +/- 330 years BP, ice originating in the Coast Mountains overrode Vancouver Island flowing in a southwesterly direction [Howes, 1981], and the ice stood as high as 1500 m, burying all but the highest mountain peaks [Maynard, 1991]. Deglaciation commenced prior to 12 930 +/- 160 years BP and the area shows evidence that deglaciation was a combination of ice retreat and downwasting of stagnant ice within individual valley systems [Muller et al, 1974; Howes, 1981; Maynard, 1991]. Howes [1981] looked at nearby drainages in northern Vancouver Island and found that as the ice surface lowered by downwasting the uplands emerged and separated the ice mass into discrete valley glaciers. Initially, these glaciers had sufficient thickness to maintain a downvalley flow, as reflected by the orientation of striae observed in the valley bottoms. Later the ice thinned and stagnated, and separated into discrete dead ice masses, disappearing before 9500 years BP [Howes, 1981]. Coastal areas were depressed by the weight of ice during, and immediately after glaciation. They later rose by isostatic rebound, leaving behind a complex marine-influenced zone along much of the coast of B.C. Howes [1981] investigated the maximum relative lateglacial sea-levels for parts of northern Vancouver Island. In the adjacent Eve River valley, Howes used deltaic deposits and marine silts to estimate that the maximum observed elevation is 152 metres above present sea level. There are no contemporary glaciers contained within the Tsitika basin. 31 2.3 Climate The Tsitika watershed has a moist, cool climate with abundant rainfall and snowfall, and mild winter temperatures. Most of the precipitation over the study area falls during winter and is associated with the passage of frontal rainstorms embedded in the generally south-westerly circulation; over 70% of the annual precipitation falls from October to March. Climatological data coverage is poor; there are three data collection sites in the Tsitika valley bottom with four years of record. Orographic precipitation is not measured, but it seems reasonable to assume that orographic effects and flow convergence (or divergence) due to funnelling effects of valleys cause regional gradients in precipitation. The highest elevations within the study area are prone to snow avalanches. The Atmospheric Environment Service [1995] reports climate normals for fifteen stations on northern Vancouver Island (Table 2) — the nearest stations are Chatham Point, which is about 40 km east along the coast, and Alert Bay, which is 25 km west along the coast. There are no long-term stations within the watershed or nearby in the Vancouver Island Ranges, but the Resource Analysis Branch operated several short-term stations that have been roughly adjusted to normals (Table 2). A l l long-term climate stations in the area ~ which are at low elevations in valley bottoms — have mean annual temperatures of around 8°C. Maximum precipitation occurs during winter, with most falling as rain between October and February; minimum precipitation occurs in July and August. Snowfall accounts for less than 5% of the precipitation at all stations but is a greater percentage of precipitation at high elevations. Snow water equivalents at nearby low-level snow courses (Elk River 3B04; 270 m) show maximum accumulations on March 1, with the pack depleted by May 1 each year. Snow courses at higher elevations (Newcastle Ridge 3B14; 1170 m, Wolf River (upper) 3B17; 1490 m) have their maximum snow water equivalent on May 1. Melt typically occurs during May but most of the snow pack remains on June 1. Annual normal precipitation declines from around 32 Table 2: Climate Stations on Northern Vancouver Island Station Name Latitude Longitude Elevation Period of Record Annual Precipitation (mm) Snowfall (mm) Maximum Daily Precipitation (mm) Atmospheric Environment Service Stations1 Alert Bay 50.35 126.56 63 1913-90 1610 70 116 Benson Lake 50.22 127.14 145 - 3310 155 182 Bull Harbour 50.55 127.57 14 1921-88 2190 63 167 Cape Scott 50.47 128.26 70 1965-90 2749 65 146 Chatham Point 50.20 125.26 23 1958-90 2185 98 114 Coal Harbour 50.36 127.30 57 1968-90 1978 85 120 Estevan Point 49.23 126.33 7 1908-90 3181 42 218 Gold River Townsite 49.47 126.03 117 ' - 2721 164 149 Holberg 50.39 128.00 579 1956-90 3156 272 179 Holberg Fire Dept. 50.39 127.39 46 1967-90 3957 100 223 Port Alice 50.23 127.27 21 1924-90 3346 51 234 Port Hardy A 50.41 127.22 22 1944-90 1871 63 154 Quatsino 50.32 127.39 8 1895-1990 2509 53 162 Spring Island 50.00 127.25 11 - 3155 40 176 Tahsis 49.55 126.39 5 1952-88 3911 64 207 Resource Analysis Branch (Short-Term) Stations in Adam/Eve Watershed2 Adam 376 1973-75 2006 - -Eve 338 1973-75 1549 - -Mount Cain 1 414 1973-75 2006 - -Mount Cain 2 701 1973-75 2032 - -Mount Cain 3 1170 1973-75 2743 - -1. Climate Normals are for 1961 to 1990, except at Benson Lake, Holberg, Gold River Townsite and Spring Island, where they are from 1951 to 1980. 2. Normals for the Resource Analysis Branch stations are from Howes 1981 33 3500 to 4000 mm. on the west coast of the Island to less than 2000 mm on the east coast (table 2). Normal precipitation in the Tsitika watershed seems to vary from around 2000 mm. near the mouth, increasing with elevation to around 2700 to 3000 mm. in the upper watershed. Average precipitation over the watershed is thought to be less than 2500 mm. Maximum daily rainfall decreases from over 200 mm. on the west coast to a little more than 100 mm on the east coast of Vancouver island. The upper value may be appropriate for high-elevation portions of the watershed [Northwest Hydraulic Consult, 1996]. There are frequent freezing level fluctuations during the winter. Freezing levels can rise or fall rapidly over the 1734 m elevation range of the study area. Seasonal snowpacks generally form at higher levels but in the early part of the winter it is likely that transient snowpacks form and disappear several times. The rapid melting of a widespread snowpack during a severe frontal rainstorm can increase the rate of runoff significantly. 2.4 Vegetation The Tsitika drainage area comprises the following biogeoclimatic zones: Coastal Western Hemlock, Mountain Hemlock, and Alpine Tundra [Krajina, 1959, 1965] (figure 4). The Coastal Western Hemlock Zone, wetter subzone (CWHw) This is the most extensive vegetation zone in the Tsitika watershed, occurring from sea level to 790 metres (600 metres feet under adverse conditions). The CWHw is represented by communities of western hemlock (Tsuga heterophylla), amabilis (balsam) fir {Abies amabilis), and western red cedar {Thujaplicata). In the northern lower portion of the valley on drier sites up to 30 metres in elevation, communities of western hemlock, western red cedar and (rarely) Douglas-fir {Pseudotsuga memiesi) also occur. The wetter subzone character of the vegetation in the lower part Figure 4: Biogeoclimatic Zones of the Tsitika Watershed 35 of the watershed may not only be a result of direct precipitation but also of reduced evapotranspiration during the dry season due to fog banks entering from Johnstone Strait [Roemer, 1972]. Mountain Hemlock Zone (MH) The Mountain Hemlock Zone receives more total precipitation than the Coastal Western Hemlock Zone, much of it as snow. There are two subzones in the Mountain Hemlock Zone: Mountain Hemlock Zone, upper, or parkland subzone (MHu) and Mountain Hemlock Zone, lower, or forested subzone (MH1). MH1 is located below and continuous with the M H u and reaches down to 790 metres (in poorly drained areas to 600 metres). In the MH1 the forests are characteristically dense and productive. The generally high moisture content and low temperature of the soils result in slow decomposition of litter and a relatively high proportion of organic matter in most soils. Mountain hemlock (Tsuga mertensiana) is the most common tree. Amabilis (balsam) fir {Abies amabilis) and yellow cedar {Chamaecyparis nootkatensis) are other major trees of the zone, with the yellow cedar generally confined to habitats with abundant moisture [Jones and Annas, 1978]. Alpine Tundra Zone (ATc) This zone is fragmentary and restricted to the mountain tops above 1370 metres which form the boundaries around the drainage. Here there is no tree or shrub growth. 2.5 Rock Weathering Because of recent Pleistocene glaciations, rock weathering systems in the Tsitika watershed have a relatively short history and therefore involve primary weathering conditions. Below we present the primary weathering processes for the primary rock constituents of the Karmutsen formation (basalt and andesite), and for the primary rock constituents of the Intrusive formation 36 (granitoids). Basalt Basalt is the most susceptible to decomposition of all rocks [Goldich, 1938], as it is primarily composed of orthosilicate and inosilicates of olivine and pyroxene. Basalt is attacked first along joint planes, leading eventually to spheroidal weathering. The presence of open vesicles or spaces left by gas bubbles in many lavas gives additional opportunities for weathering attack. The weathering products of lavas are generally fine textured. As there is no quartz in the original rock, the ultimate weathering product is often a deep, brown, base rich, heavy soil. The high proportion of clay often creates impermeable soils. 'Floaters' of basalt occasionally occur in the clay. Where the volcaniclastic rocks have been extensively altered, Peath and others [1971] and Youngberg and others [1975] have identified highly unstable soils containing high concentrations of expandable clays. Soils derived from lava flows are generally stonier, coarser textured, better drained, and much more stable than soils derived from volcaniclastic bedrock [Swanston and Swanson, 1977]. Andesite Andesite is also very susceptible to decomposition. This effusive igneous rock of intermediate composition is composed predominantly of plagioclase feldspars of the oligoclase-andesine end of the range, and of some combination of augite, orthopyroxene, and hornblende [Allaby & Allaby, 1990]; there is no quartz. Plagioclase feldspar is the most common phenocryst, but pyroxene, amphibole, or biotite may appear. Andesite has fewer ferromagnesian minerals than basalt. Andesitic lavas are fine-grained, and the rock is laminated (flow banded) and jointed. Its structure is similar to that of basalt. Andesite can give rise to sand upon weathering, and gives rise to shallower and poorer soils than basalt. Otherwise, the weathering processes and products are intermediate between basalt and 37 rhyolite, with it not being as weatherable as basalt. Granitoids Granitoids are relatively resistant to weathering since they are composed largely of durable minerals (quartz, hornblende, feldspar) arranged as a cohesive fabric of interlocking crystals. Typical granite is made up of quartz, orthoclase feldspars and micas (muscovite, biotite and hornblende); in consequence most granite is light coloured [Oilier, 1975]. Granodiorite's predominant minerals are plagioclase feldspar and quartz, with some alkali feldspar, biotite, hornblende; sphene, apatite and magnetite as its accessory minerals. Other coarse-grained igneous rocks like granodiorite weather in a manner similar to granite so long as they have quartz [Oilier, 1975]. Quartz Diorite consists of quartz, plagioclase feldspars (Oligoclase - andesine), and some ferromagnesian minerals (pyroxenes and hornblende). Quartz monzonite consists of plagioclase feldspar, orthoclase feldspar, pyroxene, and biotite. The minerals of granitic rock weather according to the sequence of plagioclase feldspar, biotite, potassium feldspar, muscovite and quartz [Durgin, 1977]. Biotite is a particularly active agent in the weathering process of granite; biotite and feldspar weather first, causing microfractures and pores to form. Groundwater then can leach out the resulting colloids and clays. As granitoid decomposes, it decreases in bulk density and shear strength [Matsuo et al., 1968]. Biotite also expands to form hydrobiotite that helps disintegrate rock into grus [Wahrhaftig, 1965; Isherwood and Street, 1976]. Feldspars break down by hydrolysis and hydration into kaolin clays and colloids, which may migrate from the rock [Durgin, 1977]. Micas weather to other clay minerals. Muscovite and quartz grains weather slowly and usually form the skeleton of saprolite. Because granite and granodiorite have a granular texture, they frequently exhibit granular disintegration, producing sandy soils and tills. It is interesting to note that the physical appearance and properties of granitic rock 38 change as weathering progresses [Ruxton and Berry, 1957; Deere and Patton, 1971; Clayton and Arnold, 1972; Durgin, 1977]. To summarise, the Karmutsen formation has the following properties: 1) small crystal size (due to faster rate of cooling), which leads to a larger surface area, faster weathering rates, and clayey soils 2) less stable minerals, leading to faster weathering rates. 3) poorly drained soils on gentler slopes (low infiltration rates) The Intrusive formation has the following properties: 1) large crystal size (due to slower rate of cooling), which leads to a smaller surface area, slower weathering rates, and sandy soils 2) more stable minerals, leading to slower weathering rates 3) well drained soils (high infiltration rates) It is thus likely that the Karmutsen formation weathers more rapidly than the Intrusive formation. It is assumed that there is not much difference in physical weathering between the two bedrock types in the study; with all evidence supporting similar degrees of jointing between the two bedrock types, this assumption does not appear unreasonable. 2.6 Surflcial Materials and Mass Wasting Processes We have a situation [in areas of coastal B.C.] in which over steepened, tectonically active mountain slopes are developed on largely weak rocks, with surface veneers or blankets of recent noncohesive sediments, in a perhumid environment. Sediment transfer is apt to be dominated by major, episodic events under natural circumstances. [Church, 1983, p. 13-14]. 39 Surficial Materials In the Tsitika River watershed, surficial materials consist primarily of glacial till, colluvium and alluvium among large areas of exposed bedrock. There are extensive mantles of glacial till covering middle and lower slopes in the whole watershed. Consistent with the weathering theory, we find that texture of surficial materials in the Tsitika watershed varies with underlying bedrock type. Upper landscapes of the Karmutsen areas are dominated by shallow sandy/clayey textured colluvium and glacial till. Mid-slope positions of the Karmutsen areas generally contain either shallow rubbly textured colluvium or shallow to moderately deep clayey glacial till; deeper till deposits dominate lower landscape positions [Jungen and Lewis, 1978]. The colluvial matrix is more silty on slopes derived from volcanic bedrock [Roemer, 1972]. Upper landscapes of the Intrusive areas of the watershed are dominated by shallow, sandy-textured colluvium and glacial till. Mid-slope landscapes in the Intrusive areas are usually occupied by shallow sandy till deposits. In the Intrusive areas, residue from prolonged physical and chemical weathering provides a coarse textured, acid parent material [Jungen and Lewis, 1978]. The differences in textures likely reflect in differences in infiltration rates; the more clayey, impermeable Karmutsen surficial materials are likely have slower infiltration rates than the Intrusive, sandier materials [Rollerson, pers. comm., 1996]. Mass Movements In the post-glacial period erosion and deposition on slopes and in valley bottoms by colluvial and fluvial processes continue to modify the landscape. Mass movement processes are common on the glacially oversteepened upper valley walls; post-glacial slope modification by weathering and mass movements results in the formation of talus slopes, debris avalanche and debris flow deposits on the lower valley slopes, and a mantle of bedrock-derived colluvium on the steep, upper valley sideslopes. 40 Maynard [1991] inventoried 1211 sediment sources in the Tsitika watershed through field checking of the valley bottom and remote sensing of the mid and upper slope areas. He found that sediment delivery mechanisms are mainly debris avalanches, debris torrents and ravelling faces. These were found to be common in thick till along creek and gully banks but the largest failures initiate on steep, upper colluvial-bedrock slopes. It was found that upper bowls of all the tributary valleys are infilled with coarse rockslide debris which is basically inactive as a sediment source. Hudson [1996] investigated the suspended sediment response to the same storms for three sub-basins in the Tsitika watershed, one solely in a Karmutsen area, one with a dominantly Intrusive area, and one with a mix of the two. The basin in the Karmutsen area had much larger suspended sediment measurements than the other two for the same storm. This finding supports the hypothesis that the difference in weatherability in the Karmutsen and Intrusive bedrock types is strongly reflected in the amount of sediment being transported on hillslope and valley bottom sediment transport processes. 2.7 Fires Unlike the nearby Nimpkish Valley, which has experienced many severe forest fires within the past one thousand years, Tsitika watershed forests have remained virtually undamaged by fire [Canadian Forest Products, 1973]. The general absence of Douglas-fir from the watershed indicates that this area has never experienced extensive fires. Charcoal particles in soil profiles have been found in only two instances, in lower Tsitika valley and in the Claud Elliott area, close to locations of Douglas-fir [Roemer, 1972]. 41 2.8 Fish and Forestry The Tsitika River and tributaries support one of the most diverse fisheries resources on Vancouver Island. Anadromous salmon species that return to the river each year include approximately 6000 pink salmon, 1200 chum salmon, 2000 coho salmon, and a small number of sockeye and chinook salmon [Ministry of Forests, 1978]. Anadromous trout species include approximately 3500 summer and winter run rainbow trout (steelhead) and an unknown number of cutthroat trout and Dolly Varden char. Resident stream and lake populations include rainbow and cutthroat trout, Dolly Varden char and prickly sculpin. Pacific lamprey is also found in the drainage [Ministry of Forests, 1978]. Timber licences and pulp leases were awarded prior to 1912 in the Tsitika watershed [Ministry of Forests, 1978]. These lands, subject to various fees and taxes since that time, were included in Tree Farm License ( T F L . ) agreements signed in the late 1950's and early 1960's. Three forest companies now have tree farm licences in the Tsitika: MacMillan Bloedel (MB) (TFL 39), Canadian Forest Products (CFP) (TFL 37), and Western Forest Products (TFL 25) (table 3). Figure 5 shows the distribution of clearcuts in the Tsitika Valley as of spring 1995. Table 3: L a n d Tenure Classification for the Tsitika Watershed TFL Licensee Area of Crown-Granted Land (acres) Area of Temporary Tenures (acres) Area of Crown Land (acres) Total 25 WFP | | | | | | | | | | | | | | | :^ 2254 706 2960 37 CFP llllllllllllllllll 7425 16738 24163 39 MB 91 14632 55730 70453 Total 91 * 24311 73174 97576 5 0 5 10 Kilometres Figure 5: Distribution of Clearcut Areas in the Tsitika Watershed 43 3. Methods 3.1 Study Design Population Definition The study population comprises all basins in the Tsitika watershed which show evidence of producing a debris torrent. Basins with snow avalanches impacting the main gully (figure 1) are excluded, because regular avalanches remove vegetation from which we gather dendrochronological information, they conceal boundaries of the deposits, and the process confuses information on debris torrent causes. Debris torrents in small gullies (#8 in Jordan's [1994] classification - see page 2), which are less than one kilometre long are also excluded from the population. Such debris torrents have small volumes (< 5000 m3), with little or no fan development. The small debris torrent systems were not included because: 1) the source basin is small and difficult to delineate; 2) different controls may govern the system; and 3) if the source basin were clearcut, there is no dendrochronological information. In contrast, the gullies in the population have a distinct source area, usually with tributaries which produces large debris torrents (as in figure 1). None of the basin source areas has been clearcut. The study population was inventoried through field checking and historical air photograph interpretation. Three valley areas were not accessible by foot (figure 6). Lower portions of the accessible gullies and fans were hiked, and all gullies were checked by a helicopter survey. Required and supportive evidence was used in the field inventory to recognise debris torrent deposits. Required evidence consists of: 1) the presence of lateral levees of coarse deposits bordering the relatively flat or less confined channel and terminal lobes on the fan surfaces [Costa, 1988; Kellerhals and Church, 1990], Figure 6: Area Not Ground Checked 45 2) the presence of logs or boulders deposited on either side of the channel or across the channel at elevations that cannot be reached by normal floods [Kellerhals and Church, 1990, p. 341], 3) the presence of scoured bedrock above the highest conceivable flood levels, 4) the presence of scars on trees located along the channel above maximum flood levels, and 5) the presence of recovery species (Red Alder) along the main gully channel and fan. Supportive evidence consists of: 1) the presence of large lone boulders on fan surfaces [Costa, 1988], or boulders rolled against trees on either side of the channel [Kellerhals and Church, 1990], 2) the exposure of unstratified, very poorly sorted sediments, sometimes reverse-graded [Costa, 1988], 3) the presence of debris jams of wood and boulders with some logs splintered, shattered, or broken [Kellerhals and Church, 1990], and 4) the presence of pieces of wood buried under boulders. Air photograph information was used for two purposes when constructing the inventory: (1) to inventory areas not field checked, and (2) to check against field data. Because of the massive scouring, it is thought that most debris flow torrent tracks remain easily identifiable on photos for several years. If the tree species that colonize scoured tracks differ from the surrounding tree cover, the track may remain identifiable for 50 to 100 years [Smith et al., 1986]. The inventory information also was cross-checked against Maynard's [1991] sediment source inventory. Required and supportive air photograph evidence of debris torrent was used. Required evidence consists of: 1) the creation of a channel through the trees between successive air photograph sets (debris, torrents can clear out the vegetation), or 2) fresh deposits seen through trees or on clear cut, differing from the previous air photograph. Supportive evidence consists of: 1) the presence of new landslides in the basin source area, or 2) the presence of the alders (recovery species) on fan in a linear or fan shape. 46 For low frequency debris torrent basins, air photograph dating is more accurate than in high frequency basins for it is less likely that more than one debris torrent occurred between the air photograph years of coverage. Table 4 lists the air photographs used. If a debris torrent was visible on any of these air photographs, the basin was included in the population. Table 4: A i r Photographs of the Tsitika Watershed Year Company Scale Type Comments 1961 M. & B. 1 :19 000 Black & White 1962 M.&B. 1:19 000 Black & White 1972 Federal Government 1 :120 000 Black & White 1973 Canfor 1 :19 000 Black & White 1973 B.C. Government 1 :20 000 Black & White 1977 M.&B. 1 :19 000 Black & White 1978 B.C. Government 1 :20 000 Black & White 1979 M. & B. 1 :19 000 Black & White 1987 M.&B. 1 :19 000 Colour 1987 Canfor 1 :19 000 Black & White 1993 Canfor 1 :19 000 Black & White Orthophoto 1994 M.&B. 1 :19 000 Colour * M. & B.: MacMillan and Bloedel Limited Canfor: Canadian Forest Products Limited Geological Mapping An accurate geological map of the Tsitika watershed is required. We mapped the bedrock type at all road-cuts and along all study gullies. From this we created a map showing the boundary between the Karmutsen Formation and the Igneous Intrusive Formation (figure 3), which improves upon the existing Geological Survey of Canada map 15 52A. We find the two bedrock types have equal areas. Our map differs significantly from the Geological Survey of Canada map. Indeed, 47 reliance on erroneous maps may have confounded people's ability to discern the difference in debris torrent activity because of erroneous geological boundaries. It is recommended that detailed geological maps be made of an area in question before conducting such mass movement inventories. Each basin in the population is classified by the underlying lithology. Fifty-four debris torrent basins, however, contain both bedrock types. In these cases, the basin is classified by lithology of the debris torrent material, as determined by a Wolman Count of 50 stones at the bottom and the top of all the gullies. The areas which are inaccessible (figure 6) are not near the geological boundaries and therefore did not require field confirmation of lithology. Our population consists of 138 debris torrent basins in the Tsitika watershed; 94 Karmutsen and 44 Intrusive. Because the Intrusive bedrock areas are more accessible than the areas underlain by Karmutsen formation there likely is an underestimation of the number of debris torrent basins in the Karmutsen area. Fifty-one basins in the Karmutsen area and two in the Intrusive area were removed under the snow-avalanche criterion. The study covers a thirty-year history of the debris torrent basins, from the summer of 1963 to the summer of 1994. The year is chosen to span from August 1 to July 31 for three reasons: 1) most debris torrents occur between October and February, and thus this 'debris torrent year' is one over which the information transfer of debris torrents is relatively weak; 2) the dendrochronological dating season is easiest during July and August, when the dark winter ring of the past winter is highlighted by the most recent spring summer growth of the light ring, and 3) our field work was conducted during July and August. The 30 year time frame was chosen because it is the longest period of time over which the deposits are easily distinguished in the field; beyond thirty years, deposits can become densely vegetated and difficult to find. Eighty basins were chosen from the population to be subjected to a study of the 30 year debris 48 torrent magnitude and frequency. These eighty basins were divided up into two study groups, named Intensive and Non-Intensive (figure 7). During the inventory basins with evidence of having produced more than one debris torrent were noted. The basins showing the most activity were chosen for their intensity, forming 25% of the study population. Al l basins with evidence of more than one debris torrent are included in the intensive group. The intensive study requires field work; ten Karmutsen basins show evidence of multiple debris torrents from the air photograph inventory, but, because they are located in the inaccessible zone could not be included in the intensive study; all of the active Intrusive basins were accessible. A third of the less active basins were chosen for a non-intensive study. Table 5 shows the sample population statistics, and table 6 summarises the information sources for the two sample types. Table 5: Sample Statistics Bedrock Type Inten number sive % of pop. Non-lnl number enslve % of pop. Population number Intrusive 11 24 14 33 44 Karmutsen 25 26 30 32 94 Total 36 25 44 32 138 49 Table 6: Information Sources for Intensive and Non-Intensive Studies 50 Type of Information Intensive Study Non-Intensive Study Occurrence (inventory) 1) Ground checking 2) Helicopter survey 3) Event reports 4) Air photographs 5) Other photos 1) Ground checking 2) Helicopter survey 3) Event reports 4) Air photographs 5) Other photos Dating of Torrents (frequency) 1) Tree sampling: scars and minimum age 2) Event reports 3) Air photographs 4) Other photos 1) Event reports 2) Air photographs 3) Other photographs Volume Determination (magnitude) 1) Volume mapping in field 2) Event reports 3) Air photographs 4) Other photos 1) Event reports 2) Air photographs 3) Other photographs 3.2 The Debris Torrent Basin as a System: Topographical Description All eighty study basins were characterised for their morphology and drainage network. The objectives for the morphometric characterisation of the debris torrent basins are (1) to detect any differences in morphology or drainage network characteristics between the lithological study groups, (2) to obtain information on the gully system controls, and (3) to look at impacts of lithology on the morphometry of the basins. We chose two types of drainage basin characteristics, morphometric and drainage network, to generalise the landscape geometry of the debris torrent basins. These data exist in ordinal, interval and ratio form. The source area of the debris torrent systems is hereafter defined as the area draining into the gully above where the channel is bedrock controlled on both sides, above the top of the Holocene fan (as in figure 1). Table 7 shows the morphometric and drainage network parameters studied for each basin. Table 7: Morphometric and Drainage Network Parameters Studied 51 Parameter Description Parameter Description Area (A) drainage area as defined on 1:20 000 TRIM map Maximum Elevation in Basin as defined from TRIM map Aspect the bearing of a line, heading down slope, which divides the drainage area in two and ends at the apex of the fan Relief (H) as defined from TRIM map Proportion of Basin in Rain-on-Snow Zone proportion of basin area between 300 and 800 metres elevation Perimeter as defined from TRIM map Proportion of Basin in Snow-Zone proportion of basin area above 800 metres elevation Maximum Basin Length (LB) the length of the longest line which starts at the fan apex and ends at the drainage divide Hypsometric Integral the percentage area under the dimensionless curve relating h/H and a/A Relief Ratio H/LB Number of First Order Streams number of unbranched tributaries, not including the main debris torrent channel, as represented on 1:20 000 TRIM map Circularity Ratio A/A. (Ac is the area of a circle with the same perimeter as A) Drainage Density (D) length of channel per unit area Stream Order order of main gully, as on 1:20 000 TRIM map Ruggedness Number HD Statistics on the study basins' values for these parameters are presented in Appendix 1. We reduced the number of parameters used to characterise the study basins in the following analysis to five based upon the following criteria: 1) parameters displaying a small range of values over all the study basins were eliminated; with a small range of values, we are unlikely to glean information on the effects that these parameters have on debris torrents. These parameters are: hypsometric integral, perimeter, circularity ratio, stream order, relief ratio, and maximum basin length; and, 52 2) parameters with information over-lap were removed, keeping the parameter with the largest range; the number of first order streams overlaps with drainage density and stream order, and the latter two were removed; maximum elevation and the ruggedness number showed the same results as relief, and the former two were removed; since all basins are located above 300 metres, and thus in either the rain-on-snow or snow zone, the proportion of the basin in the snow zone is used as an indicator of both parameters. We thus are left with area, aspect, proportion of the basin in the snow zone, relief, and number of first order streams in the basin. Area: Drainage area is a surrogate variable which can encompass runoff intensity, available debris and, in some cases, channel steepness [Slaymaker, 1988]: all three of these variables are important in debris torrent occurrence. Mizuyama [1982] and Thurber Consultants [1983] describe the lower end of basin area for basins supporting debris torrents as quite low: 0.1 km 2 for the former and 0.4 km 2 for the latter; they describe basins of 10 km 2 and 7 km 2 for the high end of basin area supporting debris torrents. VanDine [1985] found that torrents in 1-2 km 2 basins have transported the largest quantities of sediment. Aspect: On the local scale, aspect can become a dominant control of precipitation and surface temperature variation [Cheong, 1992], with west aspect being the warmest. Aspect thus affects plant productivity and may affect weathering rates. Snow Zone: The snow zone in south coastal B.C. is defined by British Columbia Ministry of Forests to be the area higher than 800 metres elevation [Hudson, 1995, pers. comm.]. This zone produces the greatest proportion of sediment from bedrock weathering, as it is less vegetated and likely undergoes more physical weathering. The zone likely does not generate much runoff during the winter, as precipitation is trapped as snow. Relief: Relief is an index of the potential energy available in the drainage basin [Chorley et al., 1984], and the speed of trans^ of water and sediment to the point of interest, so long as area does not vary 53 much with relief; in such cases, a greater relief would encourage higher peak flow. Number of First-Order Streams: The number of first-order streams is inversely related to infiltration [Chorley et al., 1984], reflecting the relation between infiltration and surface runoff. The number of first order streams used in this thesis does not include the main debris torrent channel, as shown in figure 1. The number of first-order streams can also be a function of the number of debris avalanches in the basin, for as O'Loughlin [1972] describes, the Coast Mountains landslides have a hydrological significance because they form part of the surface drainage network. Most large landslide scars possess ephemeral stream channels down the length of their longitudinal axes. In all cases investigated, however, air photos show that discernible surface drainage features did not exist on these sites prior to the time of failure. Presumably the pre-landslide downslope drainage was facilitated primarily by subsurface flow and the occurrence of an avalanche represents, therefore, an addition to the surface drainage network. [O'Loughlin, 1972, p. 107] Morphometric and drainage network parameters were obtained from a 1 : 20 000 TRIM map. Each study basin was digitised using the 'Roots' programme, with 10 m contour lines. Figure 8 shows the output of the slope characteristics obtained with this method. Parameter values were obtained through the 'IDRISr geographic information system programme. We looked for differences in morphometric parameters between bedrock type; any difference begs the following questions: 1) are the differences in morphological parameter values due to differences in the erosion rates, through processes which may include debris torrent? 2) or is there simply just a bias in the population due to other factors such as uplift or glaciation patterns? If the differences are due to reasons outlined in (1) above, a study investigating the influence of bedrock type on debris torrent occurrence would not necessarily be biased but would indicate that debris torrents are a function of bedrock type in two different ways: directly, through rates of weathering and grain size characteristics, and indirectly through the control of the bedrock type on 55 the basin morphometry. If the differences were a result of reasons outlined in (2), and if they are due to differences in glaciation or tectonic patterns in the two bedrock areas there would be a bias. For this study we cannot know which factor caused differences in morphological parameters, if any; we can only hypothesize from circumstantial evidence, and ensure that our analysis of the magnitude and frequency section deals with the morphometric question appropriately. Some morphometric parameters are more suitable to show any non-linearities between lithology and the parameter; these are: slope distribution, steepness (a function of erosion rates), drainage density [see Chorley, Schumm and Sugden, 1984, p. 467], relief (given enough time), and number of first order streams (related to drainage density). For the 54 basins which contain both bedrock types there exists a potential confounding element in the morphometric characterisation: while the morphometric parameter represents the whole debris torrent basin, including both bedrock types, the debris torrents in our study originate from only one bedrock type (see section 4.2) . Thus the morphometric characteristics of the bedrock type from which the debris torrent did not originate may confound the results. 3.3 Inputs I: Sediment Sources We inventoried sediment sources in all study basins as visible from the air photographs listed in table 4. The sediment sources connected to each gully, such as debris avalanches, talus deposits, ravelling till deposits, and cut banks, were noted for each of the eighty study basins. We encountered many visibility problems due to tree cover, snow cover and shadows present on the air photographs. Our sediment source inventory was cross-checked against Maynard's [1991] sediment source inventory. Our air photograph sediment source inventory serves two purposes: 1) to examine the 56 distribution of surficial sediment sources among the study basins, and 2) to learn more about the controls on the nature of the systems. 3.4 Inputs II: Triggers of Debris Torrents Climate Data from Atmosphere Environment Service (AES) and Water Survey of Canada (WSC) provide information on hydrologic triggers in the Tsitika watershed. Precipitation and streamflow data were analysed to (1) determine the context of the 30-year period, and (2) detect any difference in climate between the two bedrock areas. The AES Station Alert Bay (ID 1020270) provides total monthly precipitation data which were summed to obtain yearly precipitation totals, adjusted for the 'debris torrent year' (see page 49) from 1963 to 1994. There were five months without record; an estimate was substituted for these five months, using the average for the rest of the years. The five missing months of data are in 1979-80, 1981-82, 1983-84, 1987-88, and 1993-94. Hydrology There are three Water Survey of Canada stream gauges in the Tsitika watershed. Tsitika River below Catherine Creek (08HF004), Catherine Creek (08HF005), and Russell Creek (08HF007). Russell Creek Gauge drains a sub-basin which is primarily Intrusive, the Catherine Creek drains a Karmutsen sub-basin, and the Tsitika River gauge drains most of the watershed (figure 9). The Tsitika River Gauge has the longest record, from 1975 to the present; there are no data for the oldest decade of the study period. The Catherine and Russell Creek gauge records are short, beginning only in 1992. As determined by using the Gumbel distribution, storm period variables at the Tsitika River Figure 9: Location of Water Survey of Canada (WSC) Gauges 58 gauge are as follows: 50-year flood: 895 m3/s, 20-year flood: 660 m3/s, 10-year flood. 510 m3/s, 5-year flood: 395 m3/s, and the 2-year flood: 225 m3/s. The Russell Creek and Catherine Creek gauges provide insight into the spatial variability of the climate events in this system. Russell Creek is located at the far east side of the watershed, where many of the Intrusive bedrock areas are found. Catherine Creek is located in the north west side of the watershed, where many of the Karmutsen bedrock areas are found. 3.5 Outputs I: Debris Torrent Deposits Field Component: Intensive study basins The intensive study gullies were thoroughly field-checked. Each gully in the intensive group was walked from the bottom of the debris torrent deposit up to where the channel became confined by bedrock on both sides. This degree of coverage was useful in that the volume deposited in the form of levees and channel widening downstream of where the gully became unconfined was all included in the volume measurement. Also, many useful dendrochronological samples were found in these upper reaches of the gully. After the channel became confined on both sides, there was often an impassable waterfall. The length of the gully traversed averages one kilometre. Outside the active channel, debris torrent activity in past decades can leave deposits now vegetated by trees [Costa, 1988]. In the field coverage we measured all debris torrent deposits of all ages. During the field coverage the debris torrent deposits were divided into uniform reaches, averaging 30 m long, but ranging from 5 m. to 100 m, the average number of reaches per deposit was 25. We aimed to have a uniform slope, depth and width for every reach. A cross-sectional width was taken at the top and bottom of each reach; the reach width is chosen to be the average of the two cross sectional widths. The depth of the reach is chosen to be the aVerage of all possible sources of 59 depth information in that reach. The length of the reach is chosen to be the length along the centre of the channel between the middle of the two cross sections. As the deposit was traversed, all trees which were a reliable indicator of debris torrent age were sampled. Four hundred and sixty four trees were sampled in total, averaging thirteen per gully. The main limitation to sampling was the availability of sample trees, not the time or energy needed to collect the samples. On debris torrent paths aggrading through the Holocene fan, finding appropriate sample trees was easy, as there were many trees exposed to the torrent path. On debris torrent paths degrading through the Holocene fan, however, there were fewer trees available, as the entrenchment reduced exposure of trees to the torrent path. Runout Areas The area of debris torrent runout was measured in the field for all intensive study basins. Runout areas provide insight into the mechanics and nature of the source material involved in the debris torrent and how they may differ with bedrock type. There is controversy over which slope defines the beginning of the runout zone (see section 1.5). We therefore explored three different slopes (7°, 10°, and 12°) from which to define runout zone, as based on the literature. Hereafter we discuss the 12 degree slope definition of a runout zone; information derived from the other two slope definitions can be found in Appendix 2. Runout areas are normalised by volume. Many debris torrent systems do not have a complete runout zone; in 42% of the basins the debris torrent path intersects with a valley bottom river, and runout zone and volume information is lost. We provide results only for torrents with their complete volume preserved; thus the sample size for runout area is smaller (n=21). Runout areas obtained in the field were compared with air photograph information on the same gully. The deposits were not easily visible on the air photographs. There was poor correlation 60 between field and air photographs information; because of this, non-intensive samples are not included in the analysis. Field examination of the debris torrent deposits showed larger clay fractions in the Karmutsen formation than in the Intrusive formation. With a larger clay fraction, we expect that debris torrents in the Karmutsen formation have different rheological properties, enabling them to flow farther on a shallower slope than the debris torrents composed of Intrusive bedrock. 3.6 Outputs II: Magnitude and Frequency Frequency For the intensive study basins, dendrochronologic evidence in the form of minimum age and scar samples provides information on the ages of all debris torrents. Minimum age samples consist of tree discs or tree cores of hemlock and alder trees. These trees typically begin establishment within one to two years of a debris torrent. The biggest tree growing on the deposit was chosen for sampling; the samples were taken as close to the ground as possible to avoid losing rings. Duplicates were taken whenever possible. Scars are formed when rocks from a debris torrent strike a tree and remove some of the cambium. If the tree is not killed, each year after the scar a new ring grows over the edge of the scar. The scar dates are thought to be very accurate. Scars were sampled by cutting a wedge through the scar edge and counting the rings covering the scar, giving the exact number of years since the scar was created. It is important that the scar was caused by a debris torrent and not by another process. Evidence for the debris torrent origin of the scar was considered to be: 1) rocks in the scar (very reliable) (photo 3) 2) position of the scar relative to the debris torrent deposits: was it likely that a rock could have made a trajectory that would have placed it in a position to scar the tree? 62 3) accounting for other scarring agents nearby, such as: a) animals, b) rockfalls, c) windthrow, d) disease, e) bedload movement, f) road construction. In cases with no rocks in the scar, supporting dates were used to confirm the date in question. For each dendrochronological sample, the spatial relations to levees, other samples, topography, large woody debris, and channel configuration were recorded. This information was used when questions arose of the validity of the sample. The tree samples were sanded and analysed using a platform microscope. Tree rings were counted to give the minimum age of the tree or the age of the scar. The likelihood that the minimum age/scar represents a debris torrent and any laboratory errors were considered when assessing the dates of debris torrent occurrence. Event reports from the local logging companies, Canadian Forest Products and MacMillan Bloedel, from 1988-1994 provided confirmation of some debris torrent dates. The potential for missing a debris torrent increases with the number of debris torrents that have occurred in a basin, as the potential for the destruction of evidence increases. Volume The volume was determined by summing the length, width and depth measurements for each cross section. Depth estimation of landslide deposits is notoriously prone to errors. We estimated depth using the following indicators: 1) width of deposit, and inferred geometry of the old channel, 2) size of the largest clasts (assumed to be directly proportional to the depth; depth at least the b-axis of average stone), 3) depth of deposit at exposed 'steps' along the channel; any exposures, or visibility from step-jumps over logs, 4) presence or absence of buttressed roots; if present, then depth considered to be shallow, 5) nature of deposit: lumpy or smooth; lumpy usually deeper and more variable, requiring more 63 points of estimation; smooth deposits are more suitable to interpolation, and 6) slope: if a shallow slope area followed a relatively steep section, then the deposit was considered to be relatively deep in the shallow slope section; if a shallow slope area is extensive, the depth of the deposit was assumed to decrease with the distance down slope in the shallow slope area. Volumes were calculated for deposits of all ages as identified by rninimum age samples. Table 8 shows the percentage volume of the 2nd, 3rd, 4th most recent deposit as compared to the most recent. The volume measured in the field for older deposits averaged at ten percent of the original volume, the largest was half of the original deposit. Older deposits were often completely overrun by the more recent deposits, the largest non-most recent deposit uncovered was very small. Clearcutting over debris torrent deposits also caused error in reconstructing older debris torrent volume; photo 4 displays an easily visible most recent deposit, with the older deposits totally obscured from clearcutting. Because of the high errors involved in non-most recent debris torrent volume estimations, we have assumed the most recent volume is representative of all debris torrent volumes in the past 30 years. We thus assume little variability in volume in the same basin. The lack of exposure of other debris torrents supports this assumption, for if the variability was large, we would expect some basins to show evidence of much larger debris torrents than the most recent; such older deposits were not found. Non-Intensive Study Reconstruction of the magnitude and frequency was successful only for the intensive study basins; non-intensive air photograph information was found to be unreliable. The comparison of field results with air photograph information shows that often where a debris torrent was clearly visible from ground investigation, there was no visibility at all from the air photographs. One example is a 20 000 m 3 debris torrent (photo 5) which is not visible from the air photograph (photo 6). In the field we found that the zone of tree destruction in the gullies is often very narrow, with only a few trees 64 Ratio of 6th to 1st 0.00 0.47 6th Most Recent Torrent: Volume (m3) LO o i Ratio of :5th to 1st 0.00 0.01 0.00 1.41 6th Most Recent Torrent Volume (m3) O LO lO CM o> CM ! Ratio of 4th to 3rd 0.01 0.00 0.00 0.00 0.01 0.07 0.12 0.61 0.02 4th Most Recent Torrent Volume (m3) O CM O O O O CD IT) <D LO CD CO CM CO CM CM LO CO CO <D CM 3rd Most Ratio of Recent Torrent 3rd to 1st Volume (m3) 0.11 0.00 0.00 0.01 0.01 0.15 0.00 0.01 0.42 0.11 0.00 0.17 0.01 3rd Most Ratio of Recent Torrent 3rd to 1st Volume (m3) 606 25 debris torrents t deposit 30 85 20 60 459 20 20 1978 874 15 356 20 Ratio of 2nd to 1st 0.00 0.02 u. io 0.00 ience of older by most reoei 0.11 0.09 0.45 0.02 0.00 0.00 0.03 0.01 0.01 0.02 0.14 0.07 0.00 0.00 0.08 0,09 2nd Most Recent Torrent Volume (rr>3) 0 100 50 1319 16 clearcut erased all evi older deposits coverec 921 past 30 years 1893 older deposits coverec 1871 no older events older deposits coverec older deposits coverec no older events no older events 126 older deposits coverec 65 20 20 no older events 98 40 older deposits coverec 30 112 815 571 25 0 older deposits coverec 96 212 older deposits coverec Most Recent Torrent Volume (m3) 5485 18423 2732 7190 24227 18019 74087 8377 no debris torrent found ii 23313 6322 4121 3828 3510 949 21591 4000 7277 2666 19675 10000 5455 16669 3015 5030 3854 2225 4684 5882 8644 13215 2065 2317 1276 2280 4791 Basin Number Photo 5: 20 000 m3 Debris Torrent: upper channel 68 being knocked down by debris torrents, and therefore not visible. As mass movements, debris torrents are unique in this respect, they are the only process which can instantaneously move 10 000 m 3 of material without being visible on air photographs. Shadows pose another limitation to air photograph use in British Columbia. In addition to the visibility problems, there is insufficient frequency of air photography to discern in what year the debris torrents occurred. For these reasons only the intensive study basins are discussed in the following chapter. We recall from chapter 1, that some of the studies of debris torrent occurrence which have relied solely on air photographs have noted no difference in occurrence between bedrock types, for instance, Gimbarzevsky [1990], Sutherland Brown [1968], Rood [1984]. We propose that air photograph error may be a confounding factor and also that correlation between bedrock type and debris torrent visibility on air photographs may exist. Due to less fluidity from the grain size distribution and the related water content, it is possible that the Intrusive torrents have a wider zone of destruction, and are thus more visible. Granitic bedrock types may also be more visible than the basaltic bedrock types under forest cover because of their lighter colour. 4. Results and Discussion 69 The first hypothesis states there will be more basins which produce debris torrents in the extrusive bedrock area than in the intrusive bedrock area. The second hypothesis states that there will be more frequent and larger debris torrents in the extrusive bedrock area than the intrusive area. The third hypothesis states that the basins which produce larger numbers of debris torrents will have values of morphometric parameters which are associated with larger sediment supply and runoff generation. Below we examine these hypotheses in sequence, and we also look to sediment sources, depositional patterns, hydrologic triggers, and basin morphometry for any indications as to whether other parameters may be interfering with the connection between sediment supply and bedrock type. Our results show that debris torrent producing basins are twice as prevalent per unit area in the terrain underlain by Karmutsen bedrock as in that underlain by Intrusive bedrock (table 9). Debris torrent producing basins are defined as the study population on page 43. The larger incidence of debris torrent basins in the Karmutsen terrain supports the broader theme that lithology is an important parameter in determining debris torrent occurrence. Biases caused by morphological differences, however, cannot be removed from the variable of debris torrent occurrence due to the binary nature of this variable; further discussion of this issue is found on page 89. Table 9: Occurrence of Debris Torrent Basins Bedrock Type Number of Basins Total Area covered by Bedrock type (km2) Density (#/km2) Karmutsen 94 203.2 0.46 Intrusives 44 196.8 0.22 TOTAL 138 400 0.35 70 Sediment Sources First, we must verify that the debris torrent material is composed of the same lithology as the predominant lithology in the basin, to ensure that our geological boundaries are representative of the debris torrents. The results show that the lithology of the debris torrent deposits is remarkably homogeneous, even in basins containing both bedrock types (table 10, graph 1), and that the lithology of the debris torrent deposits correspond with the lithology of the dominant underlying bedrock in the basin, as derived from our map (figure 3). Thirteen basins produced debris torrent deposits with only one bedrock type in our sample of fifty stones. The average percent of the dominant bedrock type is 93% for the top sample, and 95% for the bottom sample. The homogeneity of lithology in the debris torrent deposits can indicate two things: firstly, in basins which contain both bedrock types, one bedrock type dominates the sediment supply; secondly, any till sources for debris torrents are of homogeneous lithology, and are representative of the underlying bedrock type. This result also lends support to the geological boundaries as they exist on figure 3. The connection between sediment supply and bedrock type has been made from weathering theory (chapters 1 & 2), and it can be stated that the Karmutsen Formation has a higher weathering potential than the Igneous Intrusive Formation, and has a higher potential to produce clay-sized material. Surficial material, however, must be accounted for, particularly with regard to any indications that it may play a confounding role. There are no indications of differences in sediment source distribution between the Karmutsen and Intrusive areas, as based up on air photograph analysis and field checking. However, because the basins are generally located within the vegetated tree zone the error in air photo interpretation is high. While till can be considered to be a dominant surficial material in the study area, it should not be a confounding factor in this study because (1) the air photograph analysis did not detect any difference in till distribution in the two lithologies, (2) our 71 -4-t CO c u m is ja : IB: £ 3 : 3 : 3 E : E : £ u : J2: . 2i 11: 42: re * C O o CO o Q . cu T3 c CO CO . Q _c o CO cu o E o V o XI a. o c cu ro co CO co CD c o 0) o in J5 o. E co CO < re o r -iS E (0 : fM: : « 13; : CO O T- m : in ID ID . 10 . . M 3fc 3t 4fc : O : ~ : Lithological Composition of Debris Torrents Top Sample Graph 1B 73 lithological composition results indicate that there is little mixing of lithologies in the till, and (3) the geotechnical characteristics of till are strongly tied to its parent material [Jungen and Lewis, 1978]. With no indications to the contrary, it is relatively safe to assume that surficial materials are not a confounding factor in this study, while keeping in mind the errors inherent in the air photograph analysis. From air photographs and helicopter surveys, we found three main sources of debris torrent material in the Tsitika watershed, large point sources of colluvium (input by debris avalanches) (photos 7 and 8), unconsolidated material (till) (photo 1), and small scale colluvium inputs, through gully recharge. Only five debris torrents (4%) were visibly triggered by a debris avalanche. The small number of debris torrents triggered by debris avalanches differs from the study in the Cascade range [Buchanan and Savigny, 1990] which found that 83 % of the debris torrents were directly related to debris avalanches. The rest of the debris torrents were fed by ravelling till, gully incision, or sources invisible on the air photographs. The bedrock type may affect the sediment source type (as discussed on pages 12-13). Al l debris avalanches serving as debris torrent triggers were in Intrusive basins, except one; this is consistent with Laird's [1996] observation that granitic, glacially-smoothed bedrock is prone to planar slides. Also, the three Intrusive basins which produced a debris avalanche leading to a torrent produced only one torrent in the 30 year study period, illustrating the particularly episodic nature of supply-limited basins controlled by debris avalanches; conversely, the Karmutsen formation basin which had a debris avalanche also had 4 other debris torrents in the past 30 years. From air photograph evidence it appears that most of the debris torrent basins in the Tsitika watershed are supply-limited at the main gully level. There is no evidence of large sources of material at one point, with the one exception of a transport-limited basin which has a large ravelling till source Photo 8: Bottom of Debris Avalanche (from photo 7), Gully, and Debris Torrent Deposition over Road 76 (photo #1); this basin is treated as an outlier in the analysis. Characteristics of Deposition The depositional characteristics of debris torrents can shed light on the nature of their sediment sources. Karmutsen debris torrents were found to have a consistently larger runout area when compared with their volume than the Intrusive debris torrents (graph 2, table 11). As discussed on page 16, the result can be due to differences in grain-size distribution, water content, or channel steepness. The former appears to be the critical factor and, as Jordan [1994] illustrates, it may be a cause for differences in the second parameter. The third parameter introduces a non-linear component; channel steepness is a function of the erosion taking place within the gully. Also, there a strong relation between runout area and the volume of the torrent; regression results are displayed in table 12. Two outliers were removed, one because it is a transport-limited basin as discussed above in section 4-1, and a second because it is the only debris avalanche caused debris torrent in this data subset. Table 12: Dependence of Runout Area (y) on Debris Torrent Volume (x) Bedrock Type Equation i i i i i i i i i i i i i i i Standard Error of Y Est Standard Error of Coeff. Constant Intrusive y = 0.045 x 0.496 256.1 0.013 0 Karmutsen y = 0.304 x 0.644 1069 0.0321 0 Hypothesis 2 Our second hypothesis states that there will be more frequent and larger debris torrents in the extrusive bedrock region than in the intrusive region. The results clearly support the first part of this hypothesis to be true; the basins producing the most debris torrents in the 30 year period are situated in the Karmutsen formation (graph 3). In combining the debris torrent incidence data per unit time 77 Graph 2 ' - 30 Year Debris Torrent'Frequehcy Graph 3 Table 11: Debris Torrent Runout Areas - only complete deposits are included B a s i n L i thology A r e a (m2) V o l u m e Number k=Karmutsen i=lntrusives Less than or equal to 12 degrees (m3) 4 k goes into Tsitika R. 5 k 5070 18423 6 k 3595 2732 17 k goes into Tsitika R. 19 k goes into Tsitika R. 20 k goes into Tsitika R. 22 k 3373 74087 23 k 3725 8377 37 i no debris torrents in 30 years 39 k goes into Stephanie Cr. 40 i 0 6322 43 i 60 4121 44 i 170 3828 45 i goes into Tsitika R. 48 i 499 949 49 i 2151 21591 51 k goes into Claud Elliot Ck. 52 k 2957 7277 58 i goes into K. Ck. 60 k 5521 19675 61 k 9731 12070 65 k 1803 5455 82 i 950 16669 86 k goes into Boulder Ck. 87 k goes into Boulder Ck. 90 k 1650 3854 91 k goes into Boulder Ck. 97 k 530 4684 99 k goes into Stephanie Ck. 100 k goes into Stephanie Ck. 101 k 3887 13215 125 k 1004 2065 136 k goes into Claire Ck. 137 i 0 1276 149 k 297 2280 150 i 0 4791 79 and area over the past 30 years, we find the occurrence of debris torrents per unit time and area in the Karmutsen region is over 6 times that of the occurrence in the Intrusive region (table 13). Table 13: Debris Torrent Occurrence in past 30 years Bedrock Type Number of Torrents Area (km2) Occurrence (#/km2/yr) Karmutsen 114 203.2 0.019 Intrusives 15 196.8 0.003 TOTAL 129 400 0.011 The latter part of the second hypothesis is that the debris torrents in the Karmutsen region will be larger than those in the Intrusive. Graph 4 displays the debris torrent volumes obtained for this study. As discussed in section 3.6, the volumes obtained are for debris torrents which have not lost volumes into valley bottom streams and rivers, and the volumes represent the most recent debris torrent in each basin. These debris torrent volumes are larger in the Karmutsen basins than those in the Intrusive basins (graph 4, table 8). Given the weatherability, grain size, hydrological and transport threshold differences, the Karmutsen volcanics having larger volumes than the Intrusives is what is expected; however, the difference is not as stark as it was for debris torrent frequency. This may be due to the lower degree of reliability of the volume data, or that in these supply-limited basins an increase in sediment supply results in more debris torrents, and not bigger debris torrents. Thus, with increased sediment transport to the gullies the threshold would cause the basins to flush more frequently, rather than to build up more volume, as discussed on page 19. While the volume required to trigger a debris torrent may be similar among basins, differences in volume can result from differential erosion along the debris torrent path and other similar sediment inputs. An interesting study would be to investigate the impact of bedrock type on runout area, for our results have shown that runout area is a function of bedrock type (grain size) and volume, and the latter of which also Graph 4 Volumes of Debris Torrents Most Recent Torrent 80000 Graph 5 81 appears to be a function of bedrock type. Volume results also shed light on the nature of the debris torrent basins, whether they are supply-limited or transport-limited. Three-quarters of the basins produced debris torrents with a volume less than 10 000 m3; one-quarter of the basins have produced a debris torrent between 10 000 and 25 000 m3, with the exception of the transport limited outlier (graph 5). The distinction in results of the transport-limited system from the other systems lends support to the viability of the type I/type II classification system. Our volumes are likely underestimated because of the loss of deposit into main channels which is up to perhaps 60% for some torrents. Thus we find both parts of hypothesis 2 are supported by the results, with the frequency results more strongly supported than the volume results. There are more errors involved with the volume reconstruction than with the frequency reconstruction. Triggers of Debris Torrents With only five debris torrents triggered by debris avalanches, and no evidence for any triggered by large-scale wind events (as seen in windthrow), it appears that the hydrologic trigger is the most common in the Tsitika watershed. In looking at hydrologic triggers, we have two objectives: the first is to investigate a longer term context of the storms which occurred during the study period, and the second is to determine whether there are any climatic gradients within the watershed which may serve as an bias external to the system in the comparison of bedrock types. The annual flood in the Tsitika watershed results from rainfall, or rain on snow, and typically occurs in fall or winter, between October and February. Unit mean annual maximum instantaneous flows (m3/s per km2) on the east coast of Vancouver Island range from about 0.6 to 1 (table 14). Mean Winter 7-day Low (m3 /s) 4.71 0.24 0.38 Mean Summer 7-day Low (m3 /s) 2.06 0.20 0.24 Recorded Inst. Maximum (m3 /s) 1010 CO CO £ Unit Mean Annual (m3 / km2 ) CO o CN Mean Annual Inst. Maximum (m3 /s) 1^  CO CM CO in Mean Annual Runoff (mm) 1892 1984 1854 Mean Annual Flow (m3 /s) 21.6 . o CN CM Drainage (km2 ) o co CO T— CO Period of Record1 1975RC; 76MS; 77-95RC 1992-95RC 1992-95RC Number 08HF004 08HF007 08HF008 Station Name Tsitika River below Catherine Creek Russell Creek near the mouth Catherine Creek near the mouth CO T J L _ o o CD To c o CO CD CD CO CO co" -a o o a> co M> to O o o & CD O SB D) ZS CD c CD E CO CD D ) CD O) O O CD f2 & CD i or 83 A suitable value for the Tsitika River is thought to be about 1 m3/s per km 2 which provides a mean annual flood of about 400 m3/s at the stream mouth. Based on the flood record from the Tsitika River gauge, the last major flood occurred in 1990, and the previous large flood occurred in 1975, which was the flood of record. In the twenty years of the flow records (1974-1993), there has not been a flood that is as large as the estimated 20 year flow (graph 6). This may indicate that the flows during the study period are lower than normal. Graphs 7 and 8 display precipitation patterns over the 30 year study period, as taken from nearby Alert Bay. Located in a low-lying area, Alert Bay does not share the same topography as the Tsitika Watershed, but serves as an indicator of the arrival of large frontal storms in the region. Graphs 7 and 8 show that the precipitation patterns during the first decade of the study period may be slightly more intense than for those of the last two decades. The comparison of flow and precipitation data from sites on opposite sides of the watershed provide insight into the existence of any strong climatic gradients in the watershed. For this we compare the maximum discharge, monthly rainfall totals, storm rainfall inputs and hydrological response to the storms of the Catherine Creek gauge, which is located in the Karmutsen area with the Russell Creek gauge which is located in the opposite side of the watershed, in the area dominated by Intrusives (figure 9). In comparing the maximum discharges per unit area, we find that the highest maximum discharge alternates between Russell and Catherine Creeks, with little difference between the two (table 15). In all cases, the Tsitika River has a larger discharge per unit area than the other two basins. There is therefore no striking difference between Catherine and Russell Creeks. Graph 6 Graph 7 Extreme Dany Rainfall (mm; by year Alert Bay (1020270) Graph 8 Monthly Rainfall Tutai 1992 R sseil and Cather i ie Ci Graph 9 86 Table 15: Maximum Discharge for Tsitika River, Catherine and Russell Creeks WSC Gauge ^llllillli Maximum Instantaneous Discharge / Area (m3/s/km2) | |§ ! ! f ) ! ! ! l ! ! ! ! Maximum Daily Discharge / Area (m3/s/km2) Date Tsitika River 1992 1.95 Oct 23 1.37 Oct 23-Catherine Creek 1992 1.10 Jan 29 0.77 Oct23 Russell Creek 1992 1.16 Oct 23 0.83 Oct 23 Tsitika River 1993 2.01 Dec 3 1.24 Dec 3 Catherine Creek 1993 1.24 Dec 3 0.77 Dec 3 Russell Creek 1993 0.90 Dec 3 0.75 Dec 3 Graph 9 displays the monthly rainfall totals for the Russell and Catherine Creek valley bottom gauges; they are essentially the same. In 1992 there were 20 storm events in the Tsitika watershed, and the sites at Catherine and Russell Creeks are compared for these 20 storms. The peak rainfall for these storms is similar for the two sites, with the linear regression (y = Catherine, x = Russell, with an r 2 of 0.9: y = 0.962 x fitting the data. We find that the peak discharge per unit area is larger for Russell Creek than for Catherine Creek (graph 10). This difference may be due in part to the extensive network of roads and clearcutting in the Russell Creek basin, and the general absence of roads and clearcuts in the Catherine Creek basin in 1992. The comparison of the number of debris torrents occurring each year (graph 11 and table 16) with the number of storms with greater than 25 mm. precipitation (graph 7) and the maximum 24 hour precipitation (graph 8) provides further insight the supply-limited or transport-limited classification of the study basins. During the 1970s and the 1980s the debris torrent occurrence in the Karmutsen basins has peaked every six to seven years. Comparing the debris torrent occurrence to number of days with precipitation greater than 25 mm., we find little correlation between debris torrent incidence and the number of day swith precipitation greater than 25 mm. We find a better, 87 Graph 10 Graph 11 Table 16: Debris Torrent Occurrence per Year Spring Year (incl. prev. fall) Number of Debris Torrents Percentage of 30 Year total (128) 1994 10 7.8 1993 1 0.8 1992 0 0.0 1991 24 18.8 1990 1 0.8 1989 3 2.3 1988 4 3.1 1987 3 2.3 1986 1 0.8 1985 11 8.6 1984 3 2.3 1983 1 0.8 1982 4 3.1 1981 4 3.1 1980 1 0.8 1979 3 2.3 1978 12 9.4 1977 4 3.1 1976 3 2.3 1985 5 3.9 1974 4 3.1 1973 4 3.1 1972 1 0.8 1971 6 4.7 1970 3 2.3 1969 0 0.0 1968 2 1.6 1967 3 2.3 1966 1 0.8 1965 3 2.3 1964 3 2.3 89 yet still weak, correlation between the debris torrent incidence per year and the extreme daily rainfall each year. The yearly occurrence of debris torrents appears to be only weakly controlled by rainfall patterns; this result is consistent with a supply-limited characterization of the debris torrent basins. Graph 11 also displays the nature of the errors involved in our frequency reconstruction. The decay curve in the graph is not likely due to a reduction of debris torrent occurrence in the 1960's and 1970's, for there is no theoretical explanation for the latter, and it is not supported by available climate information. A better explanation is that the degree of error associated with this method of frequency increases with time past, increasing with the frequency of torrents occurring in the basin in question. Errors are likely to arise when newer torrents destroy dendrochronological information (i.e. trees), until all evidence of a past debris torrent is obscured. If we assume that the 1990-91 debris torrent occurrence is the normal 7-year occurrence, then we can apply a decay curve to the reliability of the data set. An equation that would fit such a curve is y = 25 - 0.024 x2. This is likely an overestimation of the decay of the reliability, because it is likely that the 1990-91 torrent occurrence is probably rarer than a 7-year occurrence. If we consider 1990-91 an outlier we could interpret that the reliability is good back to 1979, or 15 years, and after we begin to lose the reliability of the frequency reconstruction. The separation of graph 11 into bedrock types illustrates the errors for each bedrock type (graph 12). The Karmutsen has far more debris torrents than the Intrusives and also appears to have more error in frequency reconstruction, as shown by its decay curve. This type of error only strengthens the results, as there is a higher chance that Karmutsen basins are producing more debris torrents than we have recorded. Debris Torrent Occurrence per Year Separated by Lithology Graph 12 91 Basin Morphometry The investigation of basin morphometry completes the second hypothesis, and addresses the third hypothesis, that the basins which produce larger numbers of debris torrents will have values of morphometric parameters which are associated with larger sediment supply and runoff generation. The first part of the second hypothesis, that the frequency of debris torrent occurrence will be higher in the Karmutsen basins than in the Intrusive basins, must be evaluated in the context of the morphometry of the individual basins, for it is possible that these parameters may cause a bias in the results. Our results show that the difference between the two bedrock types is actually strengthened upon comparing the above frequencies with the basin morphometric parameter values (graphs 13, 14, 15,16 and 17). For every value of every morphometric parameter encountered, all of the Karmutsen basins have a greater or equal frequency than do the Intrusive basins at that level of the morphometric parameter. One should note that two or more basins with the same x and y values are shown as only one symbol. Thus the frequency of debris torrents is higher in the Karmutsen basins, independent of the morphometric parameters. The graphs also display results on the impact of morphometrical parameters on debris torrent frequency, which addresses the third hypothesis. Results show that: Area: The maximum frequency is inversely related to basin area after 1 km 2 for both the Intrusive and Karmutsen basins. A possible interpretation of this is that the storage effects may drop off with increasing basin area. The results also are consistent with VanDine's [1985] finding that 1-2 km 2 basins produce the most sediment; our apparent sediment production peak is slightly less than 1 km 2. The results are consistent with the greater vulnerability of Karmutsen lithologies to weathering; hence the ability to produce more sediment in smaller gully systems. The study basin areas fall within the ranges described by Mizuyama [1982] and Thurber Consultants [1983], as discussed on page 52. 92 fllll IIB pH * PB-s -Itcojllli LL The Effect ofJBasin Area on Frequency Separated by Lithology x X X X X< X X X X X X X XX • ^; x • • x • x x x x x D A e a i k n ° i K i TI i • im v Graph 13 The Effect of Aspect on Frequency, Separated by Lithology •ralpll co 7 X 2 I MlMl X X X X X * X X X X X • X X x x x IcD^n MPS • X LL m • • • • 120 §|ao 300 K ul I t vt Graph 14 c l l l l l iolll ls . : : ft D 5 '1 ^';v^3^'2-' : ' s 400 The Effect of Basin Relief on Frequency, Separated by Lithology X X X 600 X X X X X X X X X X X X • X • X X X X X X X X H i 800 » 0 0 0 - f ^ - * i - ^ • V ^ 1 2 0 0 # ^ ^ ^ J 1 4 0 0 - " - - ^ -Kcirmutsen • Intrusive^ Graph 16 The Effect of the # of 1st Order Streams on Frequency by Lithology f a l l c X X X X X X X X X X •>, • X X • • • LL • • • • • • X N i nit si i iJIstffams W & i P ^ • ^^^^^^^^^^^^^^^^^^^^^^^ 111 | Graph 17 95 Aspect: The basins with the largest debris torrent frequencies face southwest and east. As discussed in section 4.1, southwest corresponds with the dominant airmass direction. West also corresponds with the warmest temperatures, and thus where chemical weathering rates would be the highest. Thus, a possible interpretation of this result is that dominant airmass direction and warmest temperatures encourage greater frequencies of debris torrents. The Intrusive formation has debris torrents only in basins facing in an easterly direction; the basins facing southwest did not have any debris torrents. Proportion of basin in Snow Zone: Debris torrent frequency increases slightly with proportion of area in the snow zone. The snow zone may represent an area of larger sediment supply, as discussed on page 52, and with the basins being supply limited, this result appears logical. This relation, however, is not sensitive; storage of precipitation in the form of snow and related lowering of peak flows may inhibit debris torrent initiation in this zone. The Karmutsen basins have more debris torrents at the lower proportions of snow-zone, implying that they may be less supply limited than the Intrusive basins. The basins with the highest frequencies in both bedrock types have 100% of their basin in the snow zone. Relief: Relief is not strongly correlated with debris torrent frequency. If anything, there appears to be a slight decrease in debris torrent frequency with increasing relief; this is true for both bedrock types. A reason for the insensitivity of debris torrent frequency with relief may be that the basins with the highest relief also have the largest areas. Number of First Order Streams: There is an inverse relation between the number of first order streams in a basin and the frequency of debris torrents. Relating the number of first order streams with debris avalanche occurrence, as discussed on page 55, this result supports the observation that debris avalanches are not the dominant sediment transfer mechanism for debris torrents in the Tsitika 96 watershed. The insensitivity of debris torrent occurrence to number of first order streams is also likely a function of area, in that the basins with the largest number of first order streams have the largest area. The morphometric statistics and graphs of all the debris torrent producing basins are displayed in Appendix 1. Here we discuss and compare morphometric parameter values for Karmutsen and Intrusive debris torrent basins as shown in graphs 18, 19, 20, 21 and 22. Basin Area: Basins underlain by Karmutsen bedrock produce debris torrents more commonly than those basins underlain by Intrusives; in addition, a larger proportion of both the smallest and of the largest basins are underlain by Karmutsen. Possible interpretations are that the Karmutsen basins produce sediment more readily; hence a smaller basin will sustain debris torrents; and the weathered Karmutsen material is finer grained than that of the Intrusives, hence the sediment is more mobile and debris torrents also can be sustained in larger basins. The smallest basin in both bedrock types is similar: in Karmutsen bedrock it is 0.19 km 2, and in the Intrusive bedrock it is 0.13 km 2. Aspect: For both bedrock types, the predominant aspect of the debris torrent basins is southwest. Again, a possible interpretation is that the occurrence of debris torrent producing basins is affected by the direction of the dominant air masses, as they come from the southwest. Proportion of Basin in Snow-Zone: Debris torrent basins predominantly have 60 to 100 % of their basin above 800 metres in elevation, defined as the snow-zone. In these higher elevations, larger temperature ranges and thus weathering cycles occur, and these areas are therefore likely to produce more sediment than the lower elevation areas. The high proportion of basins in the snow zone which produce debris torrents may be related to the increased sediment supply at these higher elevations. The results also show that more Karmutsen basins have 80 % of their area over 800 metres than do the Intrusive basins. The four steepest basins in the Tsitika watershed are widely distributed: one Area of Study Ba~ins Separated by Bedrock type Graph 18 Aspect of Debris Torrent Basins Separated by Bedrock Type Graph 19 Proportion of Basin in Snow Zone Separated by Bedrock Type Graph 20 Relief of Study Basins Graph 21 Graph 22 100 is in the north west quadrant in a low elevation tributary watershed (Karmutsen), the second is in a tributary valley in higher elevations in the south west quadrant (Karmutsen), the third lies in the south east quadrant adjacent to the upper Tsitika River (Intrusive), and the fourth lies in the Russell tributary headwaters in the north east quadrant (Intrusive). The wide distribution of steep basins indicates no predominance of available steep debris torrent basins in a bedrock type. Relief: While there is no striking relation displaying the impact of relief on debris torrent basin occurrence, we find that the average relief is larger for basins underlain by Karmutsen bedrock than by Intrusive bedrock. Greater potential energy would be expected to encourage the occurrence of debris torrents. Number of First Order Streams: While most debris torrent producing basins have none to two first order streams (as defined in table 7), we find that two basins underlain by Karmutsen bedrock contain a large number of first-order streams (16 and 18). This result may indicate the greater vulnerability of the Karmutsen formation to stream dissection. Any comparison of debris torrent basin occurrence between two areas never can be directly related to one parameter, as two areas are never exactly alike. While the Intrusive and Karmutsen formations are quite well-suited for a comparison, one should note that the Karmutsen bedrock may have more areas of a higher elevation and elevation range than the Intrusive bedrock. * Our results are more inconclusive for the third hypothesis than for the first and second, the values of morphometric parameters which are associated with larger sediment supply and runoff generation do not always produce more debris torrents. Application of the Findings The results from this study should be applied only to basins with similar geologic history, and 101 for basins in which sediment supply is the limiting factor. There are two main reasons why the bedrock can influence the rate of debris torrent production: sediment supply and geotechnical characteristics. The relative influence of these two variables has not been studied, and this is an important caveat in the study. It should also be noted that the geotechnical characteristics of bedrock can change with time [see Durgin, 1977]. The results should also only be applied to the type of gullies studied - not to shallow slides in small gullies on open hillslopes which are often caused by forestry activities (as described by Jordan [1994]. An important finding is that any further matters in areas with similar vegetative characteristics which invoke number or volume of debris torrents probably should be restricted to the most recent 15 years. 102 5. Conclusions Geology is a significant control on the frequency, and runout characteristics of debris torrents in the Tsitika watershed. Geology also appears to be a significant control on the spatial and volume characteristics. The Karmutsen formation produced twice as many debris torrents per unit area as the Intrusive formation. The frequency of debris torrents was always greater in Karmutsen formation basins given the same value of every morphometric parameter studied. Combining the magnitude and frequency results we can infer that geology is an important control on the sediment supply rates of debris torrents to fluvial systems. Differences in debris torrent behaviour between the Karmutsen and Intrusive formations can be due to differences in (1) infiltration rates and thus peak flow behaviour, (2) sediment supply rates due to differences in weathering rates, and (3) the transportability of the sediment as dictated by grain-size distribution. These differences can act at either of the two stages of debris torrent production: (1) the supply of sediment to the gully and (2) the transport of material in the gully. The morphometric parameter results are consistent with the hypothesis that the Karmutsen formation is more susceptible to debris torrents than the Intrusive formation. Some relations between the morphometrical parameters and debris torrent frequency were found; this confounds the direct linear relation between lithology and debris torrent occurrence. It is not the objective of this study to establish a predictive relation between morphometric parameters and debris torrent occurrence. In order to establish such a relation one needs the morphometric parameter values of basins which have not recently produced a debris torrent. Basins which have not produced a debris torrent may indeed have the same morphometric parameter values as the basins which have produced debris torrents. The lack of information on basins which give no evidence of having debris torrents recently forms a significant stumbling block to the effective 103 prediction of debris torrents. A l l available information supports our assumption that the study basins are supply-limited (type I) (with the exception of one outlier), and therefore are suitable for a study of the influence of bedrock on debris torrent temporal and spatial occurrence. Field evidence of uniform volume indicates that the debris torrent systems produce more debris torrents rather than bigger debris torrents with an increase in sediment supply. Closer examination of first-order fingertip tributaries of each basin would demonstrate the existence of a large number of transport-limited type II basins. The presence of till does not appear to affect the impact of lithology on debris torrent occurrence. Till lithological composition was found to be spatially representative of its underlying bedrock type and till is thought to retain the same geotechnical characteristics as the underlying bedrock type. Bedrock type may influence the dominant type of sediment source. We found more debris avalanche initiated debris torrents in the Intrusive areas than in the Karmutsen areas. Evidence supports the dominance of hydrologic triggers in the production of debris torrents. Climatic variations within the basin are undetected, but are assumed not to be a source of external bias for this study. The thirty year debris torrent period was a practical time frame for easily distinguishing the debris torrents in the field. The plant growth made debris torrents up to this age easy to identify and date; beyond this time period, debris torrents tend to destroy most of the evidence or previous debris torrents, and the regrowth of vegetation obscures the evidence. Evidence showed, however, that the reliability of the data decreased significantly after fifteen years of reconstruction due to the destruction of vegetation by successive debris torrents. The large spatial scale of the study on debris torrent occurrence was helpful in providing debris torrent regional information. However, the 104 establishment of causal relationships is more difficult at such scales. Flows may be lower than normal for the last twenty-years of the study period; the time over which the frequency reconstruction information is most reliable. Although we do not know the relative importance of all the parameters which control debris torrents, regional information from the spatial and temporal occurrence of debris torrents provides valuable indicators of future direction of study. We found that the air photograph information, even with a relatively good number of years of coverage, was inadequate to use as a database for this type of investigation on the grounds that vegetation cover is so extensive. Further Research Needed 1) We have established some relations showing the importance of morphometric parameters, but we have not shown the relative importance of morphometric parameters. Further research should focus on how more direct relations are confounded by less direct interactions within the system. 2) Variations in infiltration rates may be the most important regulator of debris torrent activity. Further study should investigate the importance of infiltration rates. 3) It is very difficult to reconstruct the volumes of successive debris torrents in a single basin because of the most-recent debris torrent obscuring the other deposits. We have had to assume little variability in volumes produced by a single basin through time. A long-term 105 study on the variability of volumes produced by one basin would be very useful to determine the validity of this assumption. Such a study should be separated by source type and gully type (supply or transport limited). 4) Investigate non-linear relations between lithology, morphology and debris torrent occurrence. 5) Investigate the morphometric parameters for basins which do not produce debris torrents, so we can get more information on which parameters are predictive of debris torrent occurrence. For example, all the basins of a given order (for example first to third order) can be digitised in the study area, and the morphometric parameters for the basins which produce debris torrents within a given time frame (e.g. 30 years) can be compared with those not producing debris torrents. 6) It is strongly recommended that other similar studies be conducted in other areas to distinguish spatial variations of the influence of bedrock type on debris torrent occurrence. 106 REFERENCES Allaby, A. , and M . Allaby, 1990. The Concise Oxford Dictionary of the Earth Sciences. Oxford, University Press. Atmospheric Environment Service, 1995. Climate normal data for stations Alert Bay and Chatham Point. Government of Canada. Beaty, C.B., 1990. Anatomy of a White Mountains debris-flow - the making of an alluvial fan. In Alluvial Fans: a field approach. A H . Rachocki and M . Church, (eds). John Wiley & Sons, Chichester United-Kingdom, pp. 69-89. Benda, L. , and T. Dunne, 1987. Sediment routing by debris flow. Erosion and sedimentation in the Pacific Rim (Proceedings of the Corvallis Symposium, August, 1987), IAHS Publ. 165, pp. 213-223. Bovis, M.J . , and B R . Dagg, 1987. Mechanisms of debris supply to steep channels along Howe Sound, southwest British Columbia. Erosion and sedimentation in the Pacific Rim (Proceedings of the Corvallis Symposium, August, 1987), IAHS Publ. 165, pp. 191-200. Bovis, M.J. , and B.R. Dagg, 1988. A model for debris accumulation and mobilization in steep mountain streams. Hydrological Sciences Journal, 33. pp. 589-604. Canadian Forest Products, 1973. "Tree Farm Licence 37, Management Plan for the Integrated Resources of the Upper Claude Elliott and Tsitika Watersheds" in Tsitika-Schoen Resources Study, MOF, B.C. Caine, N , 1980. The rainfall intensity duration control of shallow landslides and debris flows. Geografiska Annaler, 62A, pp. 23-27. Carlisle, D., 1963. Pillow breccias and their aquagene tuffs, Quadra Island, British Columbia. Journal of Geology, 71 (71), pp. 48-71. Cheong, A . L . , 1992. Quantifying drainage basin comparisons within a knowledge-based system framework. Master's thesis, Department of Geography, University of British Columbia. Chorley, R. J., S. A. Schumm, and D.E. Sugden, 1984. Geomorphology. Methuen, London and New York. 589 p. Church, M . , 1983. Concepts of sediment transfer and transport on the Queen Charlotte Islands. Fish/Forestry Interaction Program, Working Paper, 2/83. Church, M . , and M.J. Miles, 1987. Meterological antecedents to debris flow in southwestern British 107 Columbia. In J.E. Costa and G.F. Wieczorek, eds. Debris flows and avalanches. GSA, Reviews in Engineering Geology, 7, pp. 63-79. Clayton, J.L., and J.F. Arnold, 1972. Practical grain size, fracturing density, and weathering classification of intrusive rocks of the Idaho batholith. U.S. Dept. Agriculture Forest Service Gen. Tech. Rept. INT-2, 17 p. Costa, J.E., 1988. Rheologic, geomorphic, and sedimentologic differentiation of water floods, hyperconcentrated flows, and debris flows. In V.R. Baker, R.C. Kochel, and P C . Patton (eds.) Flood Geomorphology. Wiley-Interscience, New York, pp. 113-122. Dagg, B.R., 1987. Debris supply to torrent-prone channels on the east side of Howe Sound, British Columbia. Master's thesis, Department of Geography, University of British Columiba, 243 P Davidson, S., 1997. Master of Science thesis (in progress). Department of Geography, University of British Columbia. Dawson, G.M. , 1887. Report on the geological examinatoin of the northern part of Vancouver Island and adjacent coasts. Geological Survey of Canada Annual Report, New Serv. 2, pp. 1B-107B. Deere, D .U. , and F.D. Patton, 1971. Slope stabiliity in residual soils. Panamerican Conf. Soil Mechanics and Foundation Engineering, 4th, Caracas 1971, Proc, pp. 87-170. Dietrich, W.E., and T. Dunne, 1978. Sediment budget for a small catchment in mountainous terrain. Z. Geomorphol., Supplementband. 29. pp. 191-206. Douglas, I , 1976. Lithology, landforms and climate, in E. Derbyshire (ed), Geomorphology and Climate. John Wiley and Sons, London, pp. 345-366. Durgin, P.B., 1977. Landslides and the weathering of granitic rocks, in Geological Society of America Reviews in Engineering Geology, 3. pp. 127-131. Eisbacher, G.H., and J.J. Clague, 1981. Urban landslides in the vicinity of Vancouver, British Columbia, with special reference to the December 1979 rainstorm. Canadian Geotechanical Journal, 18, pp. 205-216. Eisbacher, G.H., and J.J. Clague, 1984. Destructive mass movements in high mountains: hazard and management. Geological Survey of Canada, Paper, 84-16, 230 p. Enos, P., 1977. Flow regimes in debris flow. Sedimentology, 24, pp. 133-142. Evans, S.G., 1982. Landslides and surficial deposits in urban areas of British Columbia: a review. 108 Canadian Geotechnical Journal 19, pp. 269-288. Evans, S.G., 1984. The landslide response of tectonic assemblages in the southern Canadian Cordillera. Proceedings, IVInternational Symposium on Landslides, Toronto, 1984. Vol. 1, pp. 495-502. Evans, S.G. and D R . Lister, 1984. The geomorphic effects of the July, 1983 rainstorms in the southern Cordillera and their impact on transportation facilities. Geological Survey of Canada, Paper No. 84-10 pp. 223-235. Fyfe, W.S., F.J. Turner and J. Verhoogen, 1958. Metamorphic reactions and metamorphic facies. Geological Society of America Memoir 73, 260 p. Geological Survey of Canada, 1993. Landslides in British Columbia. Information Circular, 1993-7. Q.P. 96409. Gimbarzevsky, P., 1988. Mass wasting on the Queen Charlotte Islands: A regional inventory. B.C. Min. For. Res. Note No. 29. ISSN 0702-9861, Victoria, B.C. , 96 p. Goldich, S.S., 1938. A study in rock weathering. Journal of Geology, 46, pp. 17-58. Groulx, B. , 1993. An examination of morphology and sediments of three adjacent steepland creeks in the Tsitika Valley. Unpublished Honours thesis, University of British Columbia, Vancouver, B.C. Gunning, H.C., 1932. Preliminary report on the Nimpkish Lake Quadrangle, Vancouver Island, British Columbia. Geological Survey of Canada Summary Report, 1931, pt. A , pp. 22-35. Guthrie, R , 1996. Masters of Science thesis (in progress) Deparment of Earth and Ocean Sciences, University of Victoria, British Columbia, Canada. Holland, S. S., 1976. Landforms of British Columbia. British Columbia Department of Mines and Petroleum Resources, Bulletin 48. Howes, D.E., 1981. Terrain Inventory and Geological Hazards: Northern Vancouver Island. APD Bulletin 5. Province of British Columbia, Ministry of Environment, Terrestrial Studies Branch, Victoria, B.C. Hudson, R.O., 1996. Use of Automated Turbidity Monitoring and Suspended Sediment Sampling in Sediment Budget Research in the Tsitika River Watershed. Paper presented at Automatic Water Quality Monitoring Workshop, Richmond, B.C. , February 12-13, 1996. Hungr, O., 1988. Notes on Dynamic Anaysis of Flowslides. In Landslides. Proceedings of 5th International Symposium on Landslides. 109 Hungr, O., G.C. Morgan, and R. Kellerhals, 1984. Quantitative analysis of debris torrent hazards for design of remedial measures. Canadian Geotechnical Journal, 21. pp. 663-677. Hungr, O., G.C. Morgan, D.F. VanDine, and D R . Lister, 1987. Debris flow defenses in British Columbia. In J.E. Costa, and G.F. Wieczorek (eds.) Debris Flows/Avalanches: Process, Recognition, and Mitigation. Geological Society of America, Reviews in Engineering Geology, V E , pp. 201-222. Innes, J.E., 1983. Debris flows. Progress in Physical Geography,!. pp. 469-501. Innes, J.L., 1985. Magnitude frequency relations of debris flows in northwest Europe. Geografiska Annaler, 67 A (1-2), pp. 23-32. Isherwood, D., and A. Street, 1976. Biotite-induced grussification of the Boulder Creek Granodiorite, Boulder County, Colorado. Geol. Soc. America Bull., 87, pp. 366-370. Iverson, R . M . , 1985. A constitutive equation for mass movement behaviour. Journal of Geology, 93, pp. 143-160. Jackson, L.E., M . Church, J.J. Clague, G.H. Eisbacher, 1985. Slopes hazards in the southern coast mountains of B.C. Field Trip Guide. Canadian Association of Geographers Jackson, L.E. , O. Hungr, J.S. Gardner, C. Mackay, 1989. Cathedral Mountain debris flows, Canada. Bulletin of the International Association of Engineering Geology, 40, pp. 35-54. Jakob, M . , 1996. Ph.D. Thesis, Deparment of Geography, University of British Columbia. Johnson, A M . , 1970. Physical Processes in Geology. Freeman, New York. Johnson, P. A. , R.H. McCuen, T V . Hromadge, 1991. Magnitude and frequency of debris flows. Journal of Hydrology, 123, pp. 68-82. Jones, R.K., and R. Annas, 1978. Vegetation. In K.W.G. Valentine, P.N. Sprout, T.E. Baker and L . M . Lavkulich (eds.) The Soil Landscapes of British Columbia. Ministry of Environment, Victoria, British Columbia, pp. 35-45. Jordan, P., 1994. Debris flows in southern Coast Mountains. PhD Thesis, Department of Geography, University of British Columbia, 260 p. Jungen, J.B., and T. Lewis, 1978. The Coast Mountains and Islands. In The Soil Landscapes of British Columbia. Valentine, K . W . G , P.N. Sprout, T.E. Baker, and L . M . Lavkulich (editors). B.C. Ministry of Environment, Victoria, British Columbia, pp. 101-120. 110 Kellerhals, R., and M . Church, 1990. Hazard management on fans, with examples from British Columbia, in Alluvial Fans: A Field Approach (A. Rachocki and M . Church, eds.). Krajina, 1965. Vegetation and Environment of the Coastal Western Hemlock Zone on Northern Vancouver Island, British Columbia, Canada. British Columbia Provincial Museum, Victoria, B.C. Laird, R , 1996. Forest Road Deactivation Manual. Ministry of Forests, British Columbia, Canada. Lord, T.M., and K.W.G. Valentine, 1978. The Soil Map of British Columbia. In K.W.G. Valentine, P.N. Sprout, T.E. Baker and L . M . Lavkulich (eds.) The Soil Landscapes of British Columbia. Ministry of Environment, Victoria, British Columbia, p. 100 Matsuo, S., Nishida, K. , and Yamashita, S., 1968. Weathering of the grahite soils and its influence on the stability of slope. Kyoto University Faculty of Engineering Memo 30 (2), pp. 85-93. Maynard, D., 1991. Tsitika River Watershed: Sediment Source Inventory. Unpublished Report for Research Branch Ministry of Forests. Vancouver-Victoria, B.C. Maynard, D., 1993. Gully Stability Classification System for Coastal British Columbia. Ministry of Forests, British Columbia. Miles, M.J. and R. Kellerhals, 1981. Some engineering aspects of debris torrents. In Proceedings of the 5th Canadian Hydrotechnical Conference, May 28-31, 1981, Fredericton, N.B. Canadian Society for civial Engineering, Fredericton, N .B . , pp. 395-420. Millard, T.H., 1993. Sediment in forested and logged gullies, Coastal British Columbia. Unpublished Masters thesis, Department of Geography, Univerisity of British Columbia. Ministry of Forests, British Columbia, 1978. Tsitika Watershed Integrated Resource Plan. Summary report, Volume II. Tsitika Planning Committee. Mizuyama, T., 1982. Analysis of sediment yeidl and transport data for erosion control works. Recent Debelopments in the Explanation and Prediction of Erosion and Sediment Yield. Proceedings Exeter Symposium, July. IAHS Publication no. 137, pp., 177-182. Montgomery, D., and J. Buffington, 1993. Channel Classification, Prediction of Channel Response and Assessment of Channel Condition. Timber, Fish and Wildlife, Report RFW-WH10-93-002. prepared for the SHAMW committee of the Washington State Timber/Fish/Wildlife Agreement. Muller, J.E., 1970. Northern Vancouver Island, British Columbia, in Report of Activities, April to October, 1969; Geological Survey of Canada, Paper 70-1, Part A, pp. 44-49. I l l Muller, J.E., K . E . Northcote and D. Carlisle, 1974. Geology and Mineral Deposits of Alert Bay -Cape Scott map Area, Vancouver Island, British Columbia. Canadian Geological Survey Paper 74-8, 77pp. Nakona, T., 1974. Natural hazards: report from Japan, in White, G.F., ed., Natural hazards - local, national, global. New York, Oxford Univ. Press, pp. 231-343. Nasmith, H.E., and A.G. Mercer, 1979. Design of dykes to protect against debris flows at Port Alice, British Columbia. Canadian Geotechnical Journal, 16, pp. 748-757. Northwest Hydraulic Consultants, 1996. Stream and Fish Habitat Restoration Plan, Eve River Watershed Restoration Program. Steelhead Watershed Restoration Corp., January, 1996. Unpublished Draft Report. Nistor, C , 1996. Masters thesis, Department of Geography, University of British Columbia, Canada. Nyberg, R., 1985. Debris flows and slush avalanches in northwestern Swedish Lappland, distribution and geomorphological significance. Meddelanden fran Lunds Universitets Geografiska Institution Avhandlinger 97, 222 p. Oden, M.E . , 1994. Debris recharge rates in torrented gullies. Masters thesis, University of British Columbia, Department of Geography Okuda, S , H. Suwa, K. Okunishi, K. Yokoyama, and K. Ogawa, 1981. Synthetic observation on debris flow. Part 7, Annuals of Disaster Prevention Research Institute, Kyoto University, No. 24B-l,pp. 157-204. (In Japanese). Oilier, C D . , 1965. Some features of granite weathering in Australia. Zeitschrift fur Geomorphologie, 11, pp. 103-108. Oilier, C D , 1975. Weathering. Geomorphology Text 2. Longman, London. O'Loughlin, C.L., 1971. An investigation of the stability of the steep land forest soils in teh Coast Mountains, southwest British Columbia [Ph.D. dissert], Vancouver, University of British Columbia, 147 p. O'Loughlin, C.L., 1972. A preliminary study of landslides in the Coast Mountains of Southwestern British Columbia. In O. Slaymaker and H.J. McPherson, eds. Mountain Geomorphology, Tantalus Press, Vancouver, pp. 101-111. Peath, R.C., M E . Harward, E.G. Know, and C.T. Dyrness, 1971. Factors affecting mass movement of four soils in the western Cascades of Oregon. Soil Sci. Soc. America Proc, 35, pp. 943-947. 112 Pierson, T.C. and J.E. Costa, 1987. A rheologic classification of subaerial sediment-water flows. In Debris Flows/Avalanches: Process, Recognition and Mitigation. J.E. Costa and G.F. Wieczorek (eds). Geological Society of America, Reviews in Engineering Geology 7, pp. 1-12. Roemer, D., 1972. Botanical Description of the Tsitika River Watershed. In Tsitika-Schoen Resources Study. Ministry of Forests, British Columbia. Rollerson, T., 1996. Personal communication, June, 1996. Rood, K , 1984. An aerial photograph inventory of the frequency and yield of mass wasting on the Queen Charlotte Islands, British Columbia. British Columbia Ministry of Forests, Land Mangement Report, 34. Ruxtoa, B P . , and L. Berry, 1957. Weathering of granite and associated erosional features in Hong Kong. GeolSoc. America Bull. 68, pp. 1263-1291. Ryder, J.M., 1978. Geology, Landforms and Surficial Materials. In K.W.G. Valentine, P.N. Sprout, T.E. Baker and L . M . Lavkulich (eds.) The Soil Landscapes of British Columbia. Ministry of Environment, Victoria, British Columbia, pp. 11-34. Schaefer, D.G., 1983. Meteorological conditions associated with slides and floods in the Lower Mainland. Atmospheric Environment Service. Pacific Regional Office, Vancouver. 5 p. Schwab, J., 1983. Mass wasting: October-November 1978 storm, Rennell Sound, Queen Charlotte Islands, British Columbia. British Columbia Ministry of Forests Research Note, No. 91. Sharpe, D R . , 1974. Mudflows in the San Juan Mountains, Colorado: flow constraints, frequency and erosional effectiveness. Unpublished Master's thesis. University of Colorado. Slaymaker, O., and H.J. McPherson, 1977. An overview of geomorphic processes in the Canadian Cordillera. Z. Geomorphol. 21(2), pp. 169-186. Slaymaker, O., 1985. Slope erosion and mass movement in relation to weathering in geochemical cycles. In Lerman, A. , and M . Meybeck, (eds) Physical and Chemical Weathering in Geochemcial Cycles. N A T O ASI Series, Vol 251., pp. 83-111 Slaymaker, O., 1988. The distinctive attributes of debris torrents. Hydrological Sciences Journal 33, pp. 567-73. Smith, R.B., P R . Commandeur, and M.W. Ryan, 1986. Soil, vegetation and forest growth on landslides and surrounding logged and old-growth areas on the Queen Charlotte Islands. British Columbia Ministry of Forests, Land Management Report, 41, 95 p. 113 Strunk, H . , 1992. Reconstructing debris flow frequency in southern Alps back to A D 1500 using dendrogeomorphological analysis in Debris flows and environment in mountain regions IASH Pub. 209, pp. 299-306. Surdam, R.C., 1971. The stratigraphy and volcanic history of the Karmutsen Group, Vancouver Island, B.C. In Progress in Geology at the University of Wyoming, Laramie, pp. 15-26. Sutherland Brown, A. , 1968. Geology of the Queen Charlotte Islands, British Columbia. British Columbia Department of Mines and Petroleum Resources, Bulletin no 54, 226 p. Swanson, F J . , and D.N. Swanston, 1977. Complex mass-movement terrains in the western Cascade Range, Oregon. In Geological Soceity of America Reviews in Engineering Geology, 3. pp. 113-124. Swanston, D.N. and F.J. Swanson, 1976. Timber harvesting, mass erosion, and steepland forest geomorphology in the Pacific Northwest, in Coates, D R . [ed.] Geomorphology and engineering. Stroudsburg, Pa., Dowden, Hutchinson, and Ross, pp. 199-221. Takahashi, T., 1981. Debris flow. Ann. Rev. FluidMech, 13, pp. 57-77. Takahashi, T., 1991. Debris flow. International Association for Hydraulic Research, Monograph Series. A. A. Balkema, Rotterdam, 165 p. Thurber Consultants Ltd. 1983. Debris Torrent and Flooding Hazards, Highway 99, Howe Sound. Report to Ministry of Transport and Highways, Victoria, British Columbia. Tripp, D. , and V. Poulin, 1986a. The effects of logging and mass wasting on salmonid spawning habitat in streams on the Queen Charlotte Islands. B.C. Min. For. Lands, Land Manage. Rep. No. 50. Tripp, D., and V. Poulin, 1986b. The effects of mass wasting on juvenile fish habitats in streams in the Queen Charlotte Islands, Land Management Report 45, B.C. Ministry of Forests and Lands, 48 p. Tripp, D., and V. Poulin, 1992. The effects of logging and mass wasting on juvenile salmonid populations in streams on the Queen Charlotte Islands. B.C. Ministry of Forests, Land Management Report No. 80, 38 p. VanDine, D.F., 1985. Debris flows and debris torrents in the southern Canadian Cordillera. Canadian Geotechnical Journal, 22, pp. 44-68. Wahrhaftig, C , 1965. Stepped topography of the southern Sierra Nevada, California. Geol. Soc. AmericBull. 76, pp. 1165-1190. 114 Wilford, D.J., and J.W. Schwab, 1981. Soil mass movements in the Rennel Sound area:, Queen Charlotte Islands, British Columbia. Unpublished report, British Columbia Ministry of Forests, Victoria. Youngberg, C.T., M E . Harward, G.H. Simonsen, and D. Rai, 1975. Nature and causes of stream turbidity in a mountain watershed, in B. Bernier, and C.H. Winget, eds., Forest soils and forest land management. Quebec, Laval Univ., pp. 267-283. APPENDIX 1: Morphometric Parameter Statistics Karmutsen Formation Mean Stan dard Error Median Mode Standard Deviation Variance Range Mini mum Maxi mum Area (km2) 0.92 0.15 0.52 NA 0.96 0.92 0.41 0.12 4.2 Aspect (degrees) NA NA NA 225 NA NA 351 3 354 Proportion of Basin in Rain-on-Snow Zone 0.23 0.29 0.17 0 0.19 0.040 0.74 0 0.74 Proportion of Basin in Snow Zone 0.77 0.028 0.83 1 0.18 0.034 0.70 0.30 1 Hypsometr ic Integral 0.53 0.009 0.525 0.52 0.058 0.003 0.29 0.36 0.65 Number of First Order Streams NA NA 2 1 NA NA 6 1 7 Drainage Density 0.99 0.084 0.87 NA 0.49 0.24 2.30 0.17 2.5 Ruggedne ss Number 871 69 696 NA 449 201448 1764 226 1990 Maximum Elevation in Basin (m) 1410 28.9 1380 1300 189.3 35848.6 700 1000 1700 Relief (m) 897.6 36.5 900 780 239.4 57327.8 880 460 1340 Perimeter (m) 5057 355 4520 2600 2328 5418983 9720 1920 11640 Maximum Basin Length (m) 1652 113 1520 800 733 537112 3440 560 4000 Relief Ratio 0.59 0.023 0.57 NA 0.15 0.022 0.80 0.27 1.06 Circularity Ratio 0.36 0.010 0.36 NA 0.070 0.0049 0.32 0.18 0.50 Stream Order NA NA 2 1 NA NA 3 1 3 Intrusive Formation Mean Stan dard Error Median Mode Standard Deviation Variance Range Mini mum Maxi mum Area (km2) 0.66 0.11 0.52 NA 0.48 0.23 1.98 0.16 2.1 Aspect (degrees) NA NA 106.6 NA NA NA 341 17 358 Proportion of Basin in Rain-on-Snow Zone 0.33 0.06 0.31 0 0.25 0.064 1.01 0 1.01 Proportion of Basin in Snow Zone 0.77 0.12 0.69 NA 0.51 0.26 2.47 0.10 2.57 Hypsometr ic Integral 0.53 0.01 4 0.53 0.50 0.057 0.0033 0.241 0.384 0.625 Number of First Order Streams NA NA 2 1 NA NA 6 1 7 Drainage Density 0.88 0.11 0.87 NA 0.49 0.24 2.30 0.17 2.47 Ruggedne ss Number 748 99 752 NA 432 186256 1880.5 96.5 1977 Maximum Elevation in Basin (m) 1281 40.8 1270 1000 182.3 33237 650 1000 1650 Relief (m) 786 38.0 780 700 170 28795 600 500 1100 Perimeter (m) 4593 309 4230 NA 1381 1.9E+06 5160 2820 7980 Maximum Basin Length (m) 1551 110 1430 1200 493.6 245641 1780 860 2640 Relief Ratio 0.53 0.02 8 0.543 NA 0.126 0.016 0.455 0.328 0.783 Circularity Ratio 0.36 0.02 1 0.35 NA 0.093 0.0087 0.32 0.18 0.50 Stream Order NA NA 2 1 NA NA 2 1 3 117 Appendix 2: Debris Torrent Runout Areas - only complete deposits are included Basin Number Lithology ::k=Karmutsen:: i=lntrusives Area (m2) Less than or equal to (in degrees) 12 10 7 Volume (m3) 4 k goes into Tsitika R. 5 k 5070 4162 2508 18423 6 k 3594.8 2951.1 1778.2 2732 ' 17 k goes into Tsitika R. 19 k goes into Tsitika R. 20 k goes into Tsitika R. 22 k 3372.8 3372.8 1265.6 74087 23 k 3725.3 3058 1842.6 8377 37 i no debris torrents in 30 years 39 k goes into Stephanie Cr. 40 i 0 0 0 6322 43 i 60.1 60.1 0 4121 44 i 170 118.4 23 3828 45 i goes into Tsitika R. 48 i 498.6 165.3 126.2 949 49 i 2151.3 2151.3 2151.3 21591 51 k goes into Claud Elliot Ck. 52 k 2957.4 2957.4 638.3 7277 58 i goes into K. Ck. 60 k 5521.2 5521.2 3907.4 19675 61 k 9730.8 9175.1 0 12070 65 k 1802.6 990.8 990.8 5455 82 i 949.7 949.7 0 16669 86 k goes into Boulder Ck. 87 k goes into Boulder Ck. 90 k 1649.9 476.3 238.1 3854 91 k goes into Boulder Ck. 97 530 492.2 82.1 4684 99 k goes into Stephanie Ck. 100 k goes into Stephanie Ck. 101 k 3886.5 2265 784.4 13215 125 k 1003.6 742 677.6 2065 136 k goes into Claire Ck. 137 i 0 0 0 1276 149 k 296.7 8.5 0 2280 150 i 0 0 0 4791 


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