@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Earth, Ocean and Atmospheric Sciences, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "English, Russell Richard"@en ; dcterms:issued "2009-05-23T16:48:16Z"@en, "1998"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """This research describes lineaments in the southwestern Coast mountains of British Columbia and their effect on landscape evolution and contemporary geomorphic processes. Lineaments mapped on air photographs are reflective of bedrock structure. Three regional trends are evidenced: a Cretaceous, northwest trend; a Tertiary norhteast trend; and an east-west trend which may represent recent crustal convergence between the North American and Juan de Fuca plates. Field and air photo evidence suggests lineaments should be interpreted primarily as either faults or large-scale regional joints. Lineament control of basin axial position is demonstrated. The spatial correlation between stream segments and lineaments is determined using the IDRISI GIS. On average 66% of the stream network overlies lineaments and 51% of lineaments in a basin are overlain by streams. The relationship of lineament length to basin morphometry is assessed by linear regression and compared to the relation between stream length and basin morphometry, Stream length is a better predictor of basin morphometry but lineament length is a good predictor of many parameters and becomes more important as basin area increases. It is speculated that lineaments become more important than streams in determining mean topography as landscape scale increases. 20 sites of large rock landslides and mountain slope deformation are identified in the study area. Rock avalanches and mountain slope deformation are the most common features, other failure types include rockfalls and landslides in surficial materials adjacent to rivers. Lineaments influence these features in three ways: 1) forming landslide headscarps; 2) providing locations about which slope deformation occurs; 3) forming rockfaces capable of shedding large rockfalls. Debris flow and avalanche initiation points are examined in the Seymour watershed north of Vancouver, British Columbia. These are almost twice as likely to occur where streams and lineaments intersect as where a stream only is present. This study suggests that the morphometry of drainage basins in the southwest Coast Mountains and some of the processes operating within them are a consequence of the underlying, tectonically emplaced bedrock structure. It is argued that processes operating at tectonic scales influence, and are reflected in, the landscape patterns around us."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/8077?expand=metadata"@en ; dcterms:extent "29841069 bytes"@en ; dc:format "application/pdf"@en ; skos:note "LINEAMENT CONTROL ON DRAINAGE BASIN DEVELOPMENT, L A R G E R O C K LANDSLIDES A N D M O U N T A I N SLOPE DEFORMATION IN THE SOUTHWEST COAST MOUNTAINS, BRITISH COLUMBIA, C A N A D A . by RUSSELL RICHARD ENGLISH B.Sc.(hons), Imperial College, University of London, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Earth and Ocean Sciences THE UNIVERSITY OF BRITISH C O L U M B I A April 1998 © Russell Richard English, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of /~C\"~&, Oc<2cn 5c*'Q,*%g J The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT This research describes lineaments in the southwestern Coast mountains of British Columbia and their effect on landscape evolution and contemporary geomorphic processes. Lineaments mapped on air photographs are reflective of bedrock structure. Three regional trends are evidenced: a Cretaceous, northwest trend; a Tertiary norhteast trend; and an east-west trend which may represent recent crustal convergence between the North American and Juan de Fuca plates. Field and air photo evidence suggests lineaments should be interpreted primarily as either faults or large-scale regional joints. Lineament control of basin axial position is demonstrated. The spatial correlation between stream segments and lineaments is determined using the IDRISI GIS. On average 66% of the stream network overlies lineaments and 51% of lineaments in a basin are overlain by streams. The relationship of lineament length to basin morphometry is assessed by linear regression and compared to the relation between stream length and basin morphometry, Stream length is a better predictor of basin morphometry but lineament length is a good predictor of many parameters and becomes more important as basin area increases. It is speculated that lineaments become more important than streams in determining mean topography as landscape scale increases. 20 sites of large rock landslides and mountain slope deformation are identified in the study area. Rock avalanches and mountain slope deformation are the most common features, other failure types include rockfalls and landslides in surficial materials adjacent to rivers. Lineaments influence these features in three ways: 1) forming landslide headscarps; 2) providing locations about which slope deformation occurs; 3) forming rockfaces capable of shedding large rockfalls. ii Debris flow and avalanche initiation points are examined in the Seymour watershed north of Vancouver, British Columbia. These are almost twice as likely to occur where streams and lineaments intersect as where a stream only is present. This study suggests that the morphometry of drainage basins in the southwest Coast Mountains and some of the processes operating within them are a consequence of the underlying, tectonically emplaced bedrock structure. It is argued that processes operating at tectonic scales influence, and are reflected in, the landscape patterns around us. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES viii LIST OF FIGURES x ACKNOWLEDGEMENTS xiv CHAPTER ONE INTRODUCTION 1 1.1 Background 1 1.2 Research synopsis and contributions. 2 1.3 Thesis structure 2 CHAPTER TWO LITERATURE REVIEW 4 2.1 Introduction 4 2.2 The relations between lineaments and bedrock structure 5 2.3 Relations between lineaments and drainage basin morphometry 7 2.4 Relations between lineaments and landslides 12 2.5 Conclusions 18 CHAPTER THREE REGIONAL PHYSIOGRAPHY AND GEOLOGY 19 3.1 Physiography 19 3.2 Climate 22 3.3 Local rock types 23 3.4 Tectonic evolution and regional structure 25 3.5 Glaciation and Quaternary deposits 26 CHAPTER FOUR THE LINEAMENT INVENTORY 28 iv 4.1 Introduction 28 4.2 Air photo interpretation of lineaments 28 4.3 The lineament inventory 29 4.3.1 Processing 29 4.3.2 The scale affect 31 4.3.3 Lineament trends 31 4.4 Field and air photo observations 42 4.5 Conclusions 52 CHAPTER FIVE DRAINAGE BASIN ANALYSIS 54 5.1 Background and methodology 54 5.2 Terminology 55 5.3 Sample set basins 56 5.3.1 Mapping of sample set basins : 57 5.3.2 Description of sample set basins 58 5.4 Basin and fan morphometry 60 5.5 Lineament control on drainage basin location 63 5.5.1 Method 64 5.5.2 Results 66 5.5.3 Discussion 74 5.6 Lineament control on drainage pattern 75 5.6.1 Method 76 5.6.2 Results 79 5.6.3 Discussion 83 5.6.4 Directional correlation of lineament and stream trend datasets 86 v 5.6.5 Stream incidence angles 92 5.6.6 Comparison of visual assessment and automatic evaluation of lineament control on streams 94 5.7 Regression analysis of lineament control on basin morphometry 95 5.7.1 Method 95 5.7.2 Results of regression analysis 97 5.7.3 Discussion 113 5.8 Investigation into sediment yield from drainage basins 115 5.8. J Regression analysis of morphometric parameters related to sediment yield. 117 5.8.2 Results of regression analysis 117 5.8.3 Discussion 129 5.9 Summary and conclusions 130 CHAPTER SIX LINEAMENTS AND LANDSLIDES 134 6.1 Introduction 134 6.1.1 Large rock landslides and mountain slope deformation 135 6.2 Airphoto identification of landslides and mountain slope deformation 135 6.3 The landslide and slope deformation inventory 136 6.3.1 Examples of lineament control on large rock landslides and mountain slope deformation 138 6.3.2 Discussion 152 6.4 Investigation into lineament control on small surficial landslides 156 6.4.1 Method of investigating the correlation between lineaments and small surficial landslides 157 6.4.2 Results 161 vi 6.4.3 Discussion 171 6.5 Small surficial landslides as sediment sources and contributors of materials to fans 175 6.6 Conclusions 179 CHAPTER SEVEN DISCUSSION AND CONCLUSIONS 183 REFERENCES 195 APPENDIX I Frequency tables for lineament trend data 205 APPENDIX II Statistical methods 208 APPENDIX III Location and description of sample set basins 212 APPENDIX IV Tabulated sample set data 216 APPENDIX V Results of spatial correlation between lineaments and streams 220 APPENDIX VI Lineament and stream trend data for the sample set Basins 230 APPENDIX VII Matrices of R 2 values for regression analysis 236 APPENDIX VIII Diskette containing regression results (included in back pocket) 246 APPENDIX IX Additional details of landslide and slope deformation inventory sites not discussed in the main text 248 APPENDIX X Results of overlay of streams on lineaments for the seymour watershed 253 vii LIST OF TABLES Table 4.1. Comparison of lineament parameters mapped on high and low altitude air photos. The basin number corresponds to numbers reported in Chapter 5, and the percentage lineament increase in the final column is the increase seen on the lower altitude air photos 31 Table 4.2. Summary information for lineaments. (*) A more realistic value is 0.677 because approximately 2,500 km2 of this block resides in the developed Fraser Lowland 36 Table 5.1. Parameters investigated in the morphometric analysis of the basin - fan system 62 Table 5.2. Assessment of lineament control on basin axis and headwalls 67 Table 5.3. Showing the spatial correlations of streams and lineaments for basin #2 overlying lineaments 80 Table 5.4. Strength of correlations for all stream orders 81 Table 5.5. Showing the average percentages of streams in the sample set basins overlaying lineaments at the two pixel acceptance level, (s.d. = standard deviation) 83 Table 5.6. Orientation data for lineament and stream segment trends in the sample set basins. (* indicates no preferred trend in the data) 89 Table 5.7. Summary data for the two streams in the cited example 93 Table 5.8. Summarizing the correlation results from GIS analysis as compared to visual assessment.94 Table 5.9. Showing the regression analysis of basin area and lineament length variables 96 Table 5.10 Regression equations for relations between lineaments and stream length with other morphometric variables 101 Table 5.10 (Continued) Regression equations for relations between lineaments and stream length with other morphometric variables 102 Table 5.11 Regression equations resulting from sediment yield investigations 118 Table 5.11 (Continued) Regression equations resulting from sediment yield investigations 119 Table 6.1. The large rock landslide and slope deformation inventory 139 Table 6.1. (Continued) The large rock landslide and slope deformation inventory 140 Table 6.2. Results of overlay of landslide initiation points on 20 m buffer zones around lineament pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2.163 vin Table 6.2 (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around lineament pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2 164 Table 6.3. Results of overlay of landslide initiation points on 20 m buffer zones around stream pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m 2 165 Table 6.3 (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around stream pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m 2 166 Table 6.4. Results of overlay of landslide initiation points on 20 m buffer zones around pixels at the intersection of streams and lineaments. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m 2 172 Table 6.4. (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around pixels at the intersection of streams and lineaments. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2 173 ix LIST OF FIGURES Figure 3.1. The location of the study area in southwest British Columbia, Canada 20 Figure 3.2. Detailed geography of the study area: CaR = Capilano Reservoir, SR = Seymour Reservoir, CoR = Coquitlam Reservoir 21 Figure 4.1. Rose diagram showing all lineaments in the inventory 32 Figure 4.2. The lineament inventory map. The inset shows the breakdown of this area into the blocks discussed in the text. The distinction between lineaments and faults indicates features mapped by the GSC from Roddick (1965) 33 Figure 4.3. Rose diagrams for each of the blocks described in the text 35 Figure 4.4. Lineament distribution in the GVRD watersheds. The approximate location of key field sites is also shown: AC = Appian Creek; OC = Orchid Creek; CC = Camp Creek 37 Figure 4.5. Simplified map of northeastern block geology and lineaments: FLG = Fire Lake Group; TIG = Twin Island Group; M = Migmatite; Qdio = Quartz diorite; Gdio = Granodiorite; Al = Alluvium; I = Ice. (Adapted from Roddick 1965) 39 Figure 4.6. A schematic diagram showing the orientation of the maximum principle stress (ai) and the conjugate fault pattern which may develop as a result of this 41 Figure 4.7. Suspected lineaments in alluvium at the mouth of Widgeon Creek near the confluence with Pitt River. These lineaments can be followed into bedrock. (Photo: 30BCB92018 #122) 43 Figure 4.8a. Stereophoto (30BCB92018:134,135) of the Camp Creek basin showing debris cones building from \"canyons\" in the cliff face, (note also the NE trending lineaments responsible for the cliff face on the lower photograph) 45 Figure 4.8b. The Camp Creek basin viewed from the west side of Coquitlam reservoir: The arrows indicates the lineaments mentioned above 45 Figure 4.9. Lineament controlled creek in the Coquitlam watershed (facing east from Coquitlam mainline) 46 Figure 4.10. A large gully (lineament) bisecting the Orchid Creek basin headwall (facing east: This gully is approximately 8 m across) 47 Figure 4.11. Two views of a large lineament behind the Orchid creek headwall of the Seymour watershed: a) facing west from the ridge above the lineament (arrows indicate the lineament), b) the view along the lineament as photographed from the top left of Fig. 4.1 la 48 Figure 4.12. Joints in the eastern headwall of the Appian Creek basin in the Capilano watershed...49 x Figure 4.13. Weathered rock material in situ on the Appian Creek headwall (lens cap indicates scale) 49 Figure 5.1. Distribution of sample set basins 59 Figure 5.2. Stereophoto (BC87098:175,176) showing lineament control of the basin axis and the northern wall in basin #1. Basin #26 is adjacent to the northeast 65 Figure 5.3. Air photo (BC87098:139) showing a laterally persistent lineament controlling the axis of three basins including basin #15 68 Figure 5.4. Stereophoto (BC87098:206,207) showing lineaments on the back and northern walls of basin #22. Note also the major north trending lineament and associated cone 69 Figure 5.5. Stereophoto (BC87098:165,166) showing lineament control in basin #14. Lineament numbers correspond to those mentioned in the text 70 Figure 5.6. Stereophoto(BC87098:178,179) showing basin #16. This basins axis is not attributed lineament control however the southern basin wall appears to be lineament controled. Note how the stream changes course as it reaches this basin wall 72 Figure 5.7. Stereophoto(BC87098:185,186) showing lineament bounding of basin #2. Basin #10 is located adjacent and to the southwest of this basin 73 Figure 5.8. IDRISI image of sample set basin #2 showing streams overlying the lineaments and their surrounding buffer zone 78 Figure 5.9. Orientation of visually assessed lineament controlled streams for the entire dataset and for igneous and metamorphic streams individually 87 Figure 5.10. Plot of Stream length versus lineament length 103 Figure 5.11. Plot of Drainage density versus lineament length 103 Figure 5.12. Plots of basin area versus a) lineament length, and b) stream length 104 Figure 5.13. Plots of basin width versus a) lineament length, and b) stream length 105 Figure 5.14. Plots of basin length versus a) lineament length, and b) stream length 106 Figure 5.15. Plots of maximum basin elevation versus a) lineament length, and b) stream length. 107 Figure 5.16. Plots of maximum basin relief versus a) lineament length, and b) stream length 108 Figure 5.17. Plots of relief ratio versus a) lineament length, and b) stream length 109 Figure 5.18. Plots of basin relief versus a) lineament length, and b) stream length 110 xi Figure 5.19. Plots of basin gradient versus a) lineament length, and b) stream length 111 Figure 5.20. Plot of fan area versus basin area 120 Figure 5.21. Plot of fan gradient versus basin area 120 Figure 5.22. Plot of basin area versus relief ratio 121 Figure 5.23. Plot of fan area versus relief ratio 121 Figure 5.24. Plot of fan area versus fan gradient 122 Figure 5.25. Plot of fan gradient versus relief ratio 122 Figure 5.26. Plots of fan area versus a) lineament length and b) stream length 123 Figure 5.27. Plots of fan gradient versus a) lineament length and b) stream length 124 Figure 5.28. Plot of relief ratio versus drainage density 125 Figure 6.1. Showing sites identified in the landslide and slope deformation inventory: l=Dickson Lake; 2=West Norrish Creek; 3=Lower Coquitlam River; 4=Lower Seymour Valley; 5=The Lions; 6=Chehalis River; 7=Bivouac Mountain; 8=Anne Lake; 9=Stave River I; 10=Stave River II; 1 l=Mount Bonnycastle; 12=Winslow Lake; 13=Winslow Creek; 14=Goat Ridge; 15=Mamquam River; 16=Shale Creek; 17=Bremner Creek; 18=Cheekye Ridge; 19=Rubble Creek; 20=Mount Mason 137 Figure 6.2. Stereophoto (15BC87098:156,157) showing Dickson lake landslide and the west Norrish Creek site. Notice the northeast trending lineament influencing both sites. Additional interpretation of this stereophoto is shown in Figure 6.3 141 Figure 6.3. Interpretation of Dickson Lake landslide and West Norrish Creek mountain slope deformation. Note: traced from airphoto 15BCB87098-156 143 Figure 6.4. The southern slope of the west Norrish Creek site showing the major lineament described in the text. The location of Figure 6.5 is indicated 144 Figure 6.5. An opening (extension joint) along the course of the major lineament at the West Norrish Creek site 145 Figure 6.6. Stereophoto (30BCB92019:41,42) showing The Lions in the Capilano Watershed. Notice the east trending lineament 147 Figure 6.7. Possible interpretation of mountain slope deformation on the southern face of The Lions in the Capilano watershed. The current state of activity at the site is unknown 148 Figure 6.8. Stereophoto (30BCB92019:196,198) showing Bivouac Mountain in the Seymour watershed. Major features of interest are indicated... 149 xii Figure 6.9. Stereophoto (BC87098:232,233) showing Goat Ridge above Britannia Creek where prominant cracks are indicated 150 Figure 6.10. Stereophoto (15BC87096:20,21) showing a large section of the slope above Mamquam River identified as deforming by mass rock creep (Scalia 1995) 151 Figure 6.1 la. Antislope scarp and graben feature at site #15 above Mamquam River (photograph used by permission) 153 Figure 6.1 lb. High angle jointing at site #15 above Mamquam River (photograph used by permission) 153 Figure 6.12. Stereophoto (BC87089:60,61,62) showing Cheekye ridge: several features described in the text are indicated 154 Figure 6.13. The Seymour watershed showing streams and landslide initiation points (from images supplied by the GVRD) 158 Figure 6.14. The Seymour watershed showing streams, landslide initiation points and lineaments 159 Figure 6. 15. The Jamieson Creek landslide, an example of the type of feature included in the GVRD landslide inventory of the Seymour watershed 162 Figure 6.16. Graph showing the distribution of landslides away from target features 167 Figure 6.17. The northern Seymour watershed showing the 0 - 40 m stream buffer, the \"intersecting lineament and stream\" pixels and landslide initiation points 169 Figure 6.18. The southern Seymour watershed showing the 0 - 40 m stream buffer, the \"intersecting lineament and stream\" pixels and landslide initiation points 170 Figure 6.19. The view down a landslide track in the Camp Creek Basin, Coquitlam watershed. The slide track is about 6-7m wide 177 Figure 6.20. A small cone at the site where this landslide enters Camp Creek. The distance across the foreground of this photograph is about 2m 178 xiii ACKNOWLEDGMENTS For the completion of this thesis, and for the last two and a half years of my studies I am indebted to a number of people. I should like to express to my supervisor Dr. K. Wayne Savigny, my sincere thanks for inspiration, guidance, and patience. Also thanks to Dr. Lisel Currie, and Dr. M. Journeay of the GSC, for long discussions, valuable guidance, and for motivation over the last year. Thanks also to Dr. Currie, Dr. R. Beckie and Dick Chase for sitting on my committee and Ken Rood of Northwest Hydraulics for guidance at various stages during this work. Grateful thanks are extended to Matt Boucher who assisted with fieldwork and other thesis related activities last summer. Financial support was provided by the Watershed Management Division of the Greater Vancouver Regional District under the direction of Dr. T. Griffing (grant #554741) and NSERC Operating Grant to Dr. Savigny (#581923). Thanks to Lome Gilmour (GVRD) for providing data and to all GVRD watershed staff who assisted during field activities in particular Dennis, Armond and the two Johns. Bruce Geotechnical Consultants Inc. provided hardware, software, and high altitude air photos for the work conducted during this thesis. I would like to thank my family and friends still in England for all of their encouragement and support over the years that I have been studying and working. Thanks also are extended to friends and colleagues (past) at the Royal School of Mines, Imperial College and in the Geological Sciences Department at the University of British Columbia (present), especially to Dan Walker for hours of discussion, praise, criticism and suggestion, and to the office staff for making the university a little easier to deal with. Special personal thanks are due to Teresa Delvecchio and her family who have been there for me since my first day in Vancouver, and Mark Mauthner for unerring friendship, without these people this would all have been very much harder. Final thanks go to Lynda for getting me this far in the first place. xiv Chapter 1 Introduction CHAPTER 1 INTRODUCTION 1.1 Background This study relates bedrock structure to contemporary process in drainage basins in the southwest Coast Mountains, British Columbia, Canada. Three large drainage basins in the mountains north of Vancouver supply water to the metropolitan areas in their vicinity. These basins are: the Capilano, Seymour, and Coquitlam rivers. Each basin is managed by the GVRD (Greater Vancouver Regional District). Erosion within the basins influences water quality, a prime concern of the GVRD and Vancouver residents. Above-normal levels of sediment in the water supply typically result from landslide activity during heavy rain storms. Structural control is evident in drainage basins of the region. Bedrock structure (described by lineaments) controls drainage patterns, the distribution of large rock landslides, and possibly influences location of debris flow and torrent activity. Establishment of a drainage network in this area requires bedrock incision which occurs preferentially in zones of bedrock weakness such as faults and joints and is primarily due to debris flow activity in steep, high elevation, mountain watersheds. This study quantifies structural control on the drainage pattern, demonstrates its relation to basin geometry, and shows that landslide activity occurs preferentially in structurally controlled areas. 1 Chapter 1 Introduction 1.2 Research synopsis and contributions Research is presented in three parts: a regional lineament inventory; the effect of lineaments on drainage basins; and the affects of lineaments on landslides. The lineament inventory is compiled from air photos. The occurrence of lineaments is described and an attempt made to explain their origins and significance in interpreting regional structure. That lineaments play a major role in the evolution of individual drainage basins is demonstrated by positional relations between lineaments and important basin features, and by regression analysis of allometric relations between lineaments and basin morphometry. Lineament control on both large rock landslides and mountain slope deformation, and debris flows and avalanches are examined. A regional inventory of the former is complemented by a smaller scale study of the latter in the Seymour watershed using a database compiled by the GVRD and a PC (Personal Computer) based GIS (Geographic Information System), IDRISI. The major contributions of this research are: the presentation of a regional lineament inventory and identification of a previously unrecognized regional structural trend; the quantification of lineament control on the drainage partem; presentation of the idea that bedrock structure influences basin morphometry; the presentation of a regional large rock landslide and mountain slope deformation inventory; and the investigation of lineament control on landslides. 1.3 Thesis structure The first chapter details research aims and thesis structure. The second details relevant literature. Chapter three describes physiography, geology, and climate of the regional study area. Chapter four examines lineaments. In Chapter five, analysis of individual basins is described to 2 Chapter 1 Introduction elucidate relationships between lineaments and drainage basin evolution. The effect of lineaments on landslides is examined in Chapter six. Chapter seven summarizes the research and presents conclusions along with suggestions for further research. Supplemental information, and data whose volume would interrupt the flow of text are presented in appendices. Additionally, a diskette is included which contains the results of regression analysis. The volume of this data is too large for inclusion as a printed appendix. 3 Chapter 2 Literature review CHAPTER 2 LITERATURE REVIEW 2.1 Introduction A review of literature posed three questions that act as a framework about which this thesis is constructed. Firstly: Do lineaments reflect bedrock structure? If this is proven, it can be shown that the regional stresses responsible for bedrock structure can be linked to present day processes influenced by lineaments. Secondly: Is there a direct relation between lineaments and drainage basin morphometry? If this is proven it will show that the fundamental landscape unit is controlled by the rock mass and its inherent properties. Finally: Is there a relation between lineaments and landslides in the study area? Demonstrating such a relation for large rock landslides will show that bedrock structure controls large-scale catastrophic landscape mass wasting processes. Small surficial landslides are the most common, point sourced, mass wasting events in the region today and are responsible for the redistribution of significant quantities of surficial materials. It is interesting to speculate to what extent structural control influences the initiation of these contemporary processes. This chapter is divided into three sections based on the questions presented above and is written to show how each question evolved from previous research. 4 Chapter 2 Literature review 2.2 The relations between lineaments and bedrock structure The literature suggests that lineaments are ubiquitous and typically representative of bedrock structure. This section defines the term lineament, reviews existing literature on lineaments and states conclusions drawn from the literature specifically relating to the first of the questions posed above. The term lineament is well established in the literature. In Wertz (1968) lineaments are used to aid in the search for ores in the Basin and Range province in southeast Arizona. The same author later reported lineaments in the northern Cordillera that are described as more significant planes of weakness in the earth's crust than faults (Wertz 1974). It is important, therefore, to clarify use of the term lineament. The most formal definition found in the literature appears inadequate to properly describe the features examined in this research. Mollard (1988) presented a review of fracture lineament research and applications on the western Canadian plains with case studies illustrating lineament mapping applications in geological, hydrogeological and geotechnical investigations. As seen on air photos, a fracture lineament is defined as a composite linear topographic, drainage, vegetation, moisture, and tonal feature (Mollard 1988). Other investigators have used the terms fracture trace, fractures, lineament, surface lineament, and photolineament interchangeably (Mollard 1988). Cronin et al. (1990, p. 232) define a [drainage] lineament as a \"...laterally persistent trend of sub-parallel drainage segments.\" While Mollard's (1988) definition forms a good basis for defining the features examined in this research, lineaments need not necessarily represent fractures. It is preferable to use the term lineament alone as being non-genetic. Hence a formal definition of the term lineament, as it applies in this research is: a non-genetic term describing a feature recognizable on an air photo, or other remotely sensed image, that is essentially linear, and comprises a contiguous set of topographic, drainage, vegetation, moisture and/or tonal features. Other terms used are: lineament pattern, describing the 5 Chapter 2 Literature review overall distribution of lineaments, and lineament set that describes a group of lineaments with similar orientation. It has become a common practice to analyze lineaments in terms of regional patterns. This has been facilitated by high altitude airborne and spaceborne remote sensing platforms and tools. Burdick and Speirer (1980) reported the detection of faults and other linear features from LANDSAT imagery. Field checking of approximately 31 lineaments in south-central Wyoming revealed that the lineaments mapped could not safely be considered faults without field investigation. However, features that proved not to be faults were anthropogenic such as roads or buried pipelines. LANDSAT was also used to map lineaments and propose a neotectonic model for the Nicoya Peninsula, Costa Rica by Hare and Gardner (1985). These authors considered lineaments to represent structural breaks. As a further example, Harris (1991) used remotely sensed lineaments to assist with interpretation of regional tectonics in gold exploration in Nova Scotia. In studies that bear comparison with pattern analysis of drainage networks it is recognized that fractal geometry is well suited to the analysis of fracture systems, e.g., Hestir et al. (1990) and Barton et al. (1990). Hestir et al. noted that the iterated function system, a standard means of producing fractal images, provided a means of generating \"geologically realistic\" fracture patterns. The evolution of natural fracture patterns is an iterative process dependent upon the presence of pre-existing fractures (Barton et al. 1990). Barton et al. (1990) showed that fracture networks from dissimilar settings of lithology, age, and tectonic setting evolve fractal dimensions in the range of 1.6 -1.8. Quart and Sen (1994) identified one of the most important implications of a fracture pattern as the flow of fluids through the fractured media. In southwestern Saudi Arabia the natural fracture pattern exhibits significant anisotropic permeability and maps of fracture density were used to show potential groundwater recharge locations and provide large-scale rock quality zonations. 6 Chapter 2 Literature review Fracture pattern research has also helped reveal the connection between drainage patterns and lineaments. Cronin et al. (1990) analyzed drainage segments mapped from LANDSAT to identify structural trends extending across multiple watersheds. Subsequent comparison with drainage segments showed that a number of drainage segments had been initiated along high-angle faults, joint sets or lithologic discontinuities. Birdseye and Christians (1988) examined drainage lineaments in late Pleistocene and Holocene sediments in Louisiana. Linear physiographic features included: \"...stream channels, natural levees, stream valleys, rectangular drainage patterns and terrace scarps\" (Birdseye and Christians 1988, p. 1109). They further state that the orientations of mapped joint systems are similar to the orientations of the lineaments suggesting an underlying structural control. They conclude that joints \"...may provide paths of weakness along which surface drainage might develop preferentially. Thus, joints probably exert an important control on the geomorphology of the region\" (Birdseye and Christians 1988, p. 1109). In conclusion, lineaments reflect underlying regional structure, and can be used to aid interpretation of both prehistoric and neotectonic stress regimes. Lineaments have typically been identified as either faults, joints (both reflecting bedrock structure) or as anthropogenic features. In areas remote from human development, however, a structural interpretation appears reasonable. 2.3 Relations between lineaments and drainage basin morphometry It is not classically recognized that bedrock structure may be a controlling factor in drainage basin morphometry. However, examination of the literature suggests that it is possible to infer this. Geological structure can play an important part in the evolution of topography. Even where structure is not visible it may be responsible for landform features. Geomorphology has been used to 7 Chapter 2 Literature review characterize the structure of the lithosphere at three scales: continental; regional; and local. At the continental scale it is now recognized that tectonic and geomorphic processes in landscape evolution are strongly coupled (Howard et al. 1994). Much work in this field has been done in New Zealand on the Southern Alps (e.g., Adams 1985, Kamp 1988) where surface geomorphology is considered representative of differing modes of subduction below the Hikurangi Margin. At a regional scale the Basin and Range province in the United States provides a good example of a structurally controlled landscape. Fault scarps, and folds, provide examples of smaller scale geomorphic features. Currently work is underway at the Geological Survey of Canada (GSC) in Vancouver to investigate the evolution of the landscape pattern of southern British Columbia with respect to large-scale crustal features such as fault blocks (Journey and Currie, pers. comm. 1995, 1996). Recent modeling developments have demonstrated the importance of the mechanical behavior of the lithosphere in evolution of topography in collisional mountain systems (e.g., Koons 1995). Modeling mean topography resulting from collisional stresses and strains has been successful in generating gross approximations of the landscape at the wavelength of mean topography. Although erosional processes modify the landscape at lower wavelengths they are often dependent on mechanical properties of the rock mass, in particular heterogeneity's caused by structural patterns. Structural control of topographic features has been identified in the form of entrainment of rivers by growing fault blocks or by preferential incision along fault planes (Bloom 1991). Koons (1995) described the particular conditions that must be met for the entrainment of rivers by growing structures explaining how a ridge valley system parallel to the dominant fault and orogen strike will result where tectonic, rather than erosional processes, are primarily responsible for topography. The drainage basin is recognized as a fundamental landscape unit, the morphology of which has classically been interpreted as a function of the drainage pattern. If erosional processes concentrate near the drainage pattern which, in turn, develops in areas of bedrock weakness, then the drainage 8 Chapter 2 Literature review pattern may reflect the nature of the structural trends inherent in the bedrock. This could be interpreted as evidence that landscape morphology is influenced by bedrock structure. Horton (1945) formulated laws explaining drainage composition of a basin and initiation of the channel network. Later workers have concentrated on refining Horton's laws applying them to their studies; for example Strahler (1952), Woldenberg (1969), and Nikora (1994). Researchers recognize a need to accurately define drainage paths in a network. One reason for this is that catchment flood properties are predictable from catchment geomorphology. Rinaldo et al. (1994) investigate the relation between basin geometry and hydrologic response attempting to extract salient geomorphic parameters of a basin by measuring the hydraulic response. Results indicate the hydraulic response is \"imprinted\" in the shape of the basin. Many properties of the hydrologic response can be predicted from the width function (W(x)), the relative proportion of basin area at distance V from the basin outlet. Further examples are presented by Helmlinger et al. (1993), e.g., mean channel length is used to predict time of concentration of a basin and mean annual flood is often related to basin area. The arrival of GIS technologies as a watershed management tool should prove invaluable in rapid characterization of a drainage network and estimation of hydrologic parameters. The GVRD is currently formulating a GIS database for detailed watershed modeling. The state-of-the-art in drainage pattern extraction is currently performed using DEMs (Digital Elevation Models). Drainage patterns derived from DEMs are typically more detailed than those mapped by conventional means. Concavity and slope of the topography are examined to postulate the presence of a stream channel, or potential for initiation of such. Helmlinger et al. (1993) review the use of DEM data for the extraction of channel networks. The scale of available topographic maps often prohibits the mapping of lower order streams. Attempts have been made to relate properties of a drainage pattern, and/or basin, to parameters such as sediment yield and measures of river discharge. Church et al. (1989) analyze suspended 9 Chapter 2 Literature review sediment load from basins in British Columbia and reported a relationship between specific sediment yield (\"the quantity of sediment passing a monitored river cross section per unit area drained upstream of that section per unit time\" Owens and Slaymaker 1992, p. 147) and basin area, and basin area and main stream length. This suggests that in British Columbia the specific sediment yield of a basin is related to the length of the major stream. Basins studied ranged upwards of 10 km2. It was found that to a threshold area of approximately 30,000 km2 the specific sediment yield increased with increasing basin area. This is opposite to conventional thinking that specific sediment yield should decrease as sediment is returned to storage along the river course. Owens and Slaymaker (1992) subsequently refined the limits of this study showing that sediment yield increases with area in basins over one square kilometer. Below this, sediment yield increases with decreasing elevation, consistent with the idea of low sediment yields in small alpine basins (Owens and Slaymaker 1992). Recently, fractal geometry has been used to investigate drainage patterns and drainage basin properties (Nikora 1994). Fractal structure of drainage patterns is demonstrated by La Barbera and Rosso (1989), Robert and Roy (1990) and Karlinger and Troutman (1992). La Barbera and Rosso show natural channel networks have fractal dimensions between 1.5 - 2.0 with average values between 1.6 - 1.7. The fractal dimension approaches two when maturity is reached \"...in the absence of geologic, topographic or hydrologic constraints\" (Karlinger and Troutman 1992, p. 1975). A fractal dimension of two suggests that the pattern is space filling. This is consistent with the requirement that every point in a basin be drained (Karlinger and Troutman 1992). It has been shown that fractal dimensions of a basin can be estimated using Horton's laws (e.g., Helmlinger et al. 1993). Considering the multi-scale nature of drainage basin morphology, Nikora (1994) examined drainage basin shape from the perspectives of self-similarity and self-affinity. Results indicated that self-affinity is a more likely solution to drainage basin morphology because the process of basin formation is directionally controlled by gravity (Nikora 1994). 10 Chapter 2 Literature review Workers have looked for the optimal arrangements of drainage patterns (e.g., Sun et al. 1994) and basins (e.g., Woldenberg 1969). Drainage area is organized into a \"...spatial structure that has a power law distribution of drainage basin areas, self-similar fractal basin boundaries, and a self-similar network structure\" (Sun et al. 1994b, p. 2599). Sun et al. (1994) used principles derived from optimal basin modeling to relate mean annual discharge of a basin to its length-width ratio. The structure of river networks obtained by optimization modeling are similar to those obtained from DEMs. Two key conclusions are drawn from the literature to this point. First, while there is a long history of relating certain topography to structural control it is only recently that geomorphology has been used to infer the tectonic history of an area. This fact has significant implications for landscape studies in active tectonic regions and for studies in areas in which tectonic activity has ceased within one landscape cycle. Second, the characters of a single drainage basin and a network of drainage basins are typically similar suggesting that there may be some underlying, and regionally uniform, controlling factor. This may be bedrock structure. Basin parameters such as length, width, and parameters of the stream network are interrelated and additionally relate to values of mean annual discharge and specific sediment yield. Drainage basin morphologic parameters are easily measured from topographic or digital format maps. With this information it should be possible to make predictions about the hydrological properties of a basin. One potential drawback is that much existing research has been conducted on homogeneous basins or models that take little account of variability in lithology or structure. Studies of real data are likely to provide more reliable results. In summation it seems that topography resulting from fluvial erosion can be controlled by structure in the underlying bedrock. Lineaments are seen as a reflection of this structure. Much work in drainage basin research has demonstrated the close relation between the drainage network and basin 11 Chapter 2 Literature review morphometry. If the drainage network is proven to mirror bedrock structure (lineaments) then it should be possible to demonstrate a more direct relationship between lineaments and basin morphometry. 2.4 Relations between lineaments and landslides Besides providing pathways for drainage development, a plane or zone of weakness in the bedrock may provide a structural discontinuity sufficient to focus catastrophic failure or deformation of the rock mass. This may be particularly true if the slope is oversteepened or has some other internal or external condition which might make the slope susceptible to failure. Evans and Savigny (1994) review landslide occurrence in the Vancouver-Fraser Valley-Whistler region. Characteristic landslide activity in British Columbia includes rock avalanches from high mountain slopes, and volcanic debris avalanches from Quaternary volcanics, (Evans 1992). A brief description of these common landslide events is presented here. Rock avalanches, the most frequent of known historic landslides, are relatively common in certain geomorphic, and geologic environments (Evans 1992). They result from rockslides or falls and involve the rapid downslope movement of bedrock fragments shattered during transport (Evans and Savigny 1994). High magnitude events of this type show high mobility, whereas lower magnitude events behave as rockfalls. Detachment of rock avalanches is favored on steep rock slopes where structural elements combine to form a detachment surface (Evans 1991, for example the 1965 Hope Slide). Volcanic debris avalanches may be initiated on the flanks of volcanoes in the Garibaldi volcanic belt, or on the high margins of lava flows a distance from the source vent. The Rubble Creek landslide is an example (Moore and Mathews 1978). 12 Chapter 2 Literature review Rock slopes adjacent to glaciers seem particularly susceptible to failure (Cruden et al. 1989, Evans and Clague 1988). A recent example is the Kshwan Glacier rock avalanche (Mauthner 1996). A significant amount of non-catastrophic mountain slope deformation is also reported in the Cordillera. Evans and Savigny (1994) reported non-catastrophic mountain slope deformation as having an important impact on Civil Engineering Structures because of the uncertainty of their future behavior. It is thought that slopes may become increasingly unstable with time and ultimately fail catastrophically. The southwest Coast Mountains show much evidence of such deformation for example the Affliction Creek site, about 100 km north of Squamish, is currently evolving and is reported in (Bovis 1982, 1990, and Bovis and Evans 1995). Most landslide events in British Columbia have occurred in remote areas but several have impacted the infrastructure of the province. The Fraser transportation corridor has on several occasions been affected (Evans and Savigny 1994, Savigny 1996). Large landslides can impact structures great distances from their source and are prone to producing secondary effects such as landslide dams (see Clague and Evans 1994, and Evans 1986). Lineament and landslide distributions have been correlated in the lower Fraser Valley (Savigny 1996, Leir et al. 1994, and Leir 1995). Typically in the literature the occurrence of a landslide at a given locality may be partly attributed to structure, however lateral persistence of the feature responsible is rarely described unless there remains obvious potential for further failure. Savigny (1996) has shown that often landslide occurrence is related to the presence of a laterally persistent lineament which is mappable on an air photo. Leir (1995) showed that when compared to factors such as the presence of certain rock types, proximity to a plutonic contact, and certain slope classes, proximity of a lineament was a high predictor in landslide occurrence at a locality. Initiation of a drainage pattern in the southwest Coast Mountains requires bedrock incision. The two primary eroding agents capable of this are: surface water flow; and debris flows. Fluvial 13 Chapter 2 Literature review erosion of channels is dependent on the ability of the stream to scour or pluck bedrock materials. This is a function of drainage area and stream slope (Howard et al. 1994). Horton (1945) stated that erosion on a surface occurs when rainfall intensity exceeds infiltration capacity and the erosive force of overland flow is greater than surface resistance to erosion. The erosive force is a shear stress exerted parallel to the slope by water that increases downslope as more water is added until a threshold value is reached where erosive force becomes larger than surface resistance. This value is dependent on the size of the slope material (Ritter et al. 1995) and is effected by type and density of vegetal cover. Where no vegetation exists soil surfaces form a hard crust that provides a high initial resistance that is progressively destroyed during a precipitation event. Erosion by overland flow is therefore a threshold process (Ritter et al. 1995) which Horton (1945) used to explain the initiation of the drainage system. Horton's model is invoked on a sloping surface as a series of sub-parallel rills, parallel to the slope gradient, form where variations in topography lead to a greater depth of flow and therefore an increase in erosive forces. In humid -temperate regions rills develop first near the base of the slope and extend gradually upslope by headward erosion. As rills develop, one \"master rill\" will become deeper capturing others by destruction of the divides between them. The original pattern is obliterated as downcutting continues until a small stream develops. Ultimately the original slope, parallel to the master channel, is replaced by slopes on either side sloping toward the main drainage line. These side slopes then develop rills and the process is repeated (Horton 1945). Seidl and Dietrich (1992) found scour by debris flows to be the dominant erosional agent on steep, first order channels in the Oregon Coast Range. They demonstrate a change in bedrock erosion processes when stream gradient reaches 0.2. Below this gradient, erosion is primarily due to abrasion and dissolution varying linearly with stream power. Above this, debris flow scour takes over and 14 Chapter 2 Literature review contributing area becomes less important. This has important implications for stream channels that, it appears, will preferentially erode in areas susceptible to debris flow activity. Debris flows and avalanches represent active wasting processes on high elevation slopes in British Columbia. Events in the early nineteen eighties in the Howe Sound area prompted studies into debris flows resulting in a report by Thurber Engineering Ltd. (1983) and several studies by others into various aspects of these features. In addition to presenting a hazard, debris flows deliver large volumes of sediment to catchments. Van Dine (1984, p.44) defines a debris torrent as a mass movement involving water charged, dominantly coarse grained, inorganic, and organic material flowing rapidly down a steep pre-existing confined channel. This is distinct from a debris flow, which occurs on a planar, unconfined slope. Despite the introduction of this distinction early in the literature many authors exclusively use the term debris flow, even for events and studies confined to gullies (for example, Takahashi (1993)). Popular opinion in B.C., argues for the use of the term debris flow only and the term torrent is not used subsequently in this work. Local studies include: Van Dine (1984), Buchanan and Savigny (1990), Fanin and Rollerson (1990), and Jordan (1994). Additionally, Kellerhals and Church (1990) review hazard management on fans in southern British Columbia with special reference to Howe Sound. Costa (1984) provides a comprehensive review of literature prior to 1984. Davies et al. (1992) report on three landscape units involved in a debris flow: the actively eroding gully wall, a steep narrow degrading gully, and a less steep, wider, alluvial-bedded channel or valley. These also represent the transition from a zero order basin through the first order channel to a second or higher order channel. Dietrich et al. (1987) report periodic evacuation of zero order basins by small surficial landslides. Buchanan and Savigny (1990) report that studies in Japan show events on convergent slopes are more common than those on planar or divergent slopes. Hence these events are typically associated 15 Chapter 2 Literature review with creeks and gullies. Events in southern British Columbia often occur in gully sidewalls and move into gullies where they will flow in the confined channel. Only one third of events studied by Fannin and Rollerson (1990) initiated and terminated on open slopes. There is a consensus about the factors required for a creek to be a suitable site for debris flow initiation. The creek must have a drainage area within a critical range, be sufficiently steep, and contain debris for mobilization. The critical area identified in the Howe Sound is between 0.4 and 7.0 km2 (Van Dine 1984). Jordan (1994) reports 0.1 to 10 km2 in the Squamish - Lillooet river areas. This latter figure is in line with those reported in Japan by Van Dine (1984). The steep profiles required are reported at between 20° and 57° for Howe sound (Thurber Engineering Ltd. 1983). Most authors cite extreme precipitation as a common triggering mechanism, (e.g., Van Dine (1984), Buchanan and Savigny (1990) and Fannin and Rollerson (1990)), although snowmelt and antecedent precipitation can be major contributing factors. Snowmelt played a major part in the large number of debris flows on the Wasatch Front, Utah during the springs of 1983 and 1984 (Wieczorek et al. 1989). Johnson and Sitar (1990) review in detail the hydrologic conditions leading to debris flow initiation. Other possible triggering mechanisms are mentioned by Fannin and Rollerson (1990) such as rockfalls, planar translational slides or sidewall collapse. However, for proper classification as debris flows or avalanches the materials involved must be saturated. Once mobilized debris flows may flow between 100 and 1000 m, depositing when the gradient becomes insufficient for continued motion, or when the flow becomes unconfined. In studies in the Queen Charlotte Islands, Fannin and Rollerson (1990) showed that slightly more than 50% of events traveling in a gully finally deposited outside that gully. The remaining flows deposited within the gully. The physical properties of debris flows in motion have been extensively studied. Examples include Costa (1984), Bovis and Dagg (1988) Takahashi (1993), and Jordan (1994). The slope of the 16 Chapter 2 Literature review transport and erosion zone is typically larger than 10° and in Howe Sound ranges from 13 - 35° (Van Dine 1984). In southern British Columbia debris flow deposition typically occurs on fans. Thurber Engineering Ltd. (1983) analyzed hazards on fans in Howe Sound. Kellerhals and Church (1990) report on the major findings. Whipple (1993) attempted to interpret debris flow hazard from fan morphology, citing critical factors as the number and spacing of abandoned channels, texture and irregularity of interfluves, form of channel-margin levees, and degree of cross fan convexity. However, Whipple's studies were conducted on unvegetated fans where measurement of such features is easily made from remote sensing imagery. In British Columbia fans are typically well vegetated. Debris flow paths are often easily identifiable for several years and may stay visible for a maximum 50 - 100 years, if recolonizing vegetation is different to surrounding vegetation (Kellerhals and Church 1990). The sedimentology of deposits from these events is variable, and reported by most authors for their own areas. Debris composition depends on the nature of the bedrock and overburden. Deposits typically have a uniform size distribution up to boulders of several meters. These are supported in a matrix of fine grained debris (Costa 1984) that can be winnowed from the deposit leaving it clast supported. Deposits can be extremely variable from creek to creek and Van Dine (1984) reports that three creeks, less than six kilometers apart in Howe Sound produced widely diverse debris. Van Dine (1984) concluded that occurrence of debris flows appears to be unrelated to geology but Davies et al. (1992) report that gullies exhibiting debris flows are typically unstable and may be associated with faulting and crush zones. Several key conclusions are drawn from the above review. Debris flow activity is a primary eroding mechanism in stream channels in mountain watersheds. Although it is traditionally believed that there is little relation between bedrock geology and debris flow activity, it appears that their 17 Chapter 2 Literature review occurrence may be related to structure as suggested by Davies et al. (1992) in symbiotic relation with structurally emplaced stream channels. Most debris flows initiate in the sidewalls of gullies containing stream channels, material is moved to the channel and either flushed to the fan or deposited in the channel where it may later be re-mobilized. In either case there is the potential for fine materials to be winnowed from the deposit. These may proceed farther into the hydrological system and cause sediment problems in the water supply. It is therefore important to determine how closely the occurrence of debris flow activity is related to the location of stream channels known to be structurally controlled. 2.5 Conclusions It is shown by a review of the literature that lineament mapping provides a means of evaluating the structure of an area and can be used to infer a region's tectonic history. It is also established that structure influences topographic evolution of the landscape, including development of the drainage network. In the literature the drainage network is generally attributed responsibility for basin and fan morphometry and conventional views of drainage basin development are typically presented for homogeneous surfaces with little regard for inhomogeneity that may be caused by bedrock structure. Lineaments have been related to landslides in the neighboring Fraser River Valley. Typically a lineament may provide the structural control at a site necessary for catastrophic failure or provide a focus for slope deformation. It is therefore important to try to ascertain whether lineaments control landslide occurrence in the study area. As a contemporary process, debris flows are revealed as the primary force of erosion in steep, low order channels and transport large amounts of sediment to the fan. The location of debris flow and debris avalanche activity may also be related to lineament occurrence. 18 Chapter 3 Regional physiography and geology CHAPTER 3 REGIONAL PHYSIOGRAPHY AND GEOLOGY 3.1 Physiography The study area, located in the southwest Coast Mountains, British Columbia, Canada (Fig. 3.1), comprises most of the 92G, NTS (National Topographic System), 1:250,000 scale map sheet east of Howe Sound and Squamish River and north of the North and West Vancouver urban areas and the north shore of Fraser River. The eastern boundary lies near Harrison Lake and the northern boundary is in Garibaldi Park. Detailed geographical information is shown in Figure 3.2. The following physiographic description is based on the authors' field and airphoto experience. In the south, moderately steep slopes lead to frontal ridges of mountains with summit elevations upward of 1400 m. Several slopes are developed for housing and three of the higher slopes are utilized for ski resorts. Behind these are a series of flat-topped ridges and peaks separated by narrow, northwest trending river valleys. Several valleys open southward to wide flat bottomed, alluvial valleys, e.g., Pitt River, others are directed through narrow bedrock canyons, e.g., Capilano River. Two coastal fjords, Howe Sound and Indian Arm, are respectively west and east of Vancouver. The GVRD watersheds lie in this southern region. Summit elevations and terrain ruggedness increase northward. Northwest trending ridges and valleys persist to central parts of the study area where most rivers are sourced in high, glaciated, 19 Chapter 3 Regional physiography and geology 1/1 Legend S t u d y A r e a V a n c o u v e r F r a s e r Lowland S t r a i t o f Gerogia V a n c o u v e r Is land Ranges FL SG VIR Figure 3.1. The location of the study area in southwest British Columbia, Canada. 20 Chapter 3 Regional physiography and geology 50 N 10 M o r v t r r t 0 49 N_^ _ 123° 25' W _j 49 N 122° V Figure 3.2. Detailed geography of the study area: CaR = Capilano Reservoir, SR = Seymour Reservoir, CoR = Coquitlam Reservoir. 21 Chapter 3 Regional physiography and geology mountain valleys. Numerous icefields, moraine dammed and glacial fed lakes exist north of about 49° 35'N. The northwest is dominated by the Mount Garibaldi volcanic center (elevation approximately 2680 m) from which extend numerous, well-defined lava flows. The northeastern part of the study area contains the northwest trending Harrison Lake-Pemberton Valley. Over the entire area, slopes are moderate to steep and largely till mantled except for the steepest slopes and ridge-tops. Valley glaciers extend short distances down valleys radiating from the icefields. Logging is ubiquitous especially, but not exclusively, in valleys with access from the south and west where the major transport and urban infrastructure exists. Some northern valleys are logged by boat access, for example the northern Pitt River valley. 3.2 Climate Climate information is derived from the GVWD (Greater Vancouver Water District) watershed ecological inventory pilot study, final report (1993) and from Ryder (1981). The climate of the region is controlled by location and topography. A complex relationship exists between amounts of precipitation (both rain and snow), and distance and elevation up the major valleys. Westerly trade winds are responsible for warm and moist Pacific air. Barometric pressure changes frequently are associated with Arctic to sub-tropic air masses into the region resulting in variable weather. Winter frontal systems are associated with cyclonic storms in the Gulf of Alaska. Frontal system precipitation is enhanced orographically. Precipitation increases with elevation from 2500 mm on the coast to 5000 mm at higher elevations (Ryder 1981). The mountains produce a rain 22 Chapter 3 Regional physiography and geology shadow effect with precipitation decreasing to the northeast. The Fraser lowland channels coastal weather systems inland. Rain storms are frequent and data collected by the GVRD show the one-hour, 100 year storm can be expected to yield 50 mm per hour. The one day, 100 year storm yields precipitations of about 12.5 mm per hour and can be expected to provide about 300 mm of precipitation, and the five day, 100 year storm about 640 mm. Both values were exceeded in November 1990. The maximum temperature range for the southern parts of the study area is 32°C. The mean temperature range for Vancouver is 2 - 18°C with a mean January temperature between 1 - 5°C (Roddick 1965). Temperatures decrease inland from the coast. Two seasons are recognized, winter extends from late September through March. Winter storms have a large areal extent and temperature variations in the air mass lead to rainfall at low elevations and snow accumulation in high areas. Transient snowpacks, resulting from fluctuating freezing levels, form and disappear at elevations between 200 and 1000 m. Near Vancouver mean annual snowfall increases at approximately 660 cm per km of elevation. Snow accumulations in Seymour valley have been measured at 2.5 - 4.6 m. Water equivalents at high elevation range from 1-2 m. The regional snowline rises, and snowfall decreases, eastwards as the climate becomes more continental. Semi-arid conditions exist at low elevations on the lee side of the Coast Mountains. 3.3 Local rock types Rocks of the Coast Plutonic Complex occupy 85 - 90% of the study area. Precise lithologies are described by Roddick (1965). The complex comprises mid-Jurrasic to middle Cretaceous granites, granodiorites, quartz diorite, diorite, gabbro and migmatite dated at 167 - 91 Ma (Monger and Journeay 1994). 23 Chapter 3 Regional physiography and geology The Coast Plutonic Complex was intruded into regionally metamorphosed, greenschist facies, Triassic and Jurrasic stratified strata of the Wrangellia and Harrison terranes (see Section 3.4). These metamorphic rocks now occur as fault bounded pendants or septa and are, from oldest to youngest: Twin Island Group; Bowen Island Group; Cultus Formation; Harrison Lake Formation; Fire Lake Group; Helm Formation; occasional Tertiary deposits and extrusives; and Quaternary, Garibaldi Group volcanics. The pre-Tertiary volcanic and sedimentary rocks typically contain more rocks of sedimentary than volcanic origin. Outcrops are often aligned with the dominant northwesterly structural trend. Within these rocks structure typically trends northwesterly with moderate to steep dips. Some smaller pendants of Twin Island Group rocks seem unrelated, structurally, to others and exhibit a northeasterly trend (Roddick 1965). Late Cretaceous to early Tertiary stratified rocks occur in small exposures on the southern slopes of the mountains north of Vancouvers. These southerly dipping rocks represent continental deposits with minor volcanics deposited in Georgia Depression below Fraser lowland. These materials underlie Vancouver but are largely buried below Quaternary deposits (Roddick 1965). The Garibaldi volcanics represent part of a belt trending for 110 km northwest of Mount Garibaldi. Including pre-glacial, intra- and post-glacial deposits these range in age from 3.8 Ma to 1340 yBP (Lawrence et al. 1984). Rocks comprise flows and pyroclastics of varying composition from basalt through andesite and dacite, to rhyodacite (Roddick 1965). 24 Chapter 3 Regional physiography and geology 3.4 Tectonic evolution and regional structure The geological history of the Vancouver area is influenced by the interaction of three crustal blocks, the Coast Mountains, the Vancouver Island Ranges and the Cascade Range mountains. These are separated by two lowland depressions, Strait of Georgia and Fraser Lowland. The recent tectonic evolution of the Canadian Cordillera is best described as a series of accretion events. The local history is separated into three phases (Monger and Journeay 1994). A pre-accretionary stage saw the joined Vancouver Island Ranges and southwestern Coast Mountains separated from North America by basins now forming the southeastern Coast Mountains. By about 100 Ma these blocks had accreted to the continental margin causing crustal thickening and associated uplift and erosion in the Coast Mountains. The syn-, and post-accretion stage lasted until approximately 40 Ma when the continental Cascade Arc (see Fig.4.6) formed as the present subduction regime evolved. Presently, the Cascadia subduction zone lies about 100 km west of Vancouver Island. The subducting plate is at an estimated depth of 70 km below Vancouver (Monger and Journeay 1994, after Rogers and Horner 1991). Strait of Georgia represents the forearc depression and is subsiding at about 1 mm.a'1 (Monger and Journeay 1994). The Garibaldi Volcanics are a part of the Cascade magmatic arc. Regional physiography is believed to have formed mainly in the last ten million years, and is thought to be related to stresses on the plate margin and thermal expansion of the Cascade arc (Monger and Journeay 1994). Further details of the tectonic evolution of the Canadian Cordillera can be found in Gabrielse and Yorath (1991). The dominant regional structure is a northwesterly trending fabric formed when plate motions switched from an orthogonal sense, to dextral relative motions between approximately 85 Ma and the early Tertiary. This occurred contemporaneous with, and following closure of the basins separating 25 Chapter 3 Regional physiography and geology the Coast Mountain and Vancouver Island crustal blocks from North America. At this time a compressional and transpressional regime existed in the southeastern Coast Mountains. In the southwest, deformation was concentrated on northwest-trending, and west- to southwest-vergent dextral shear zones. One, the Harrison Fault Zone, cuts the northeast corner of the study area. This zone separates two distinct tectonostratigraphic terranes, the Wrangellia and Harrison terranes, respectively west and east of the fault zone. The fault zone extends to the base of the crust and at surface forms the Harrison Lake - Pemberton valley (Monger and Journeay 1994). The youngest dated structures in the area are northeast trending, dextral transcurrent faults with associated northwest striking high angle reverse faults (Journeay 1990). These are possibly associated with the eastward subduction of the Juan de Fuca plate and perhaps recording northeast-southwest crustal shortening. These are thought to have been active between 25 and 14 Ma (Journeay 1990). 3.5 Glaciation and Quaternary deposits At the climax of each of three major glaciations in the Quaternary ice covered most of British Columbia. The last glaciation, the Fraser, mantled slopes with till and filled valley floors with glaciofluvial and glaciolacustrine sediments collectively known as Vashon drift. These are overlain in places by Capilano sediments (Hicock and Armstrong 1984). Materials deposited prior to the last glaciation which remained unconsolidated, were extensively scoured and reworked from the valleys and slopes to form the outwash deposits that underlie much of Vancouver and, in places, the Vashon drift. These are the Quadra sands. The southwestern section of the Wisconsinan ice sheet began to decay at about 14 ka and parts of coastal areas were ice-free by 13 ka. Ice had completely withdrawn from the area by 10 ka (Clague 26 Chapter 3 Regional physiography and geology 1989). At this time mass wasting and fluvial processes began to redistribute glacial deposits. A period of valley aggradation was followed by downcutting as sediment supplies diminished. By mid to late Holocene times many streams flowed near their present levels. The main depositional areas have since been lake and sea floors, fans and deltas (Clague 1989). 27 Chapter 4 The lineament inventory CHAPTER 4 THE LINEAMENT INVENTORY 4.1 Introduction The focus of the chapter is an inventory of lineaments compiled from air photos of the study area. A discussion of the air photo identification of lineaments is followed by presentation of the lineament inventory. Details of inventory production and interpretation are combined with presentation of the data. Field observations of selected lineaments are described along with suggested interpretations for the features identified. 4.2 Air photo interpretation of lineaments Lineament identification on air photos relies on the observer's ability to detect a combination of features defining the lineament. This imparts a bias on interpretation which might imply automatic detection by computer would be preferable. Although subjective, visual inspection tends to reveal a larger number of features than automated detection (Leir 1995). For photo detection of lineaments a sun angle that allows utilization of shadows revealing subtle differences in relief and textural patterns is preferred (Lillesand and Kiefer 1979). In mountainous terrain the sun must be high enough to illuminate the valleys. A mid-morning or mid-afternoon flight during early summer when the ground is free of snow cover is optimal because the presence of snow and ice will obscure ground detail. 28 Chapter 4 The lineament inventory Certain landscape features enhance lineament detection; straight stream segments and rock walls can be used as guidelines for lineament mapping because these morphological features may be part of a more persistent lineament. Some anthropogenic features may be confused with naturally occurring lineaments: Abandoned logging roads and pipelines may be difficult to identify through a dense tree cover, but have a visible photo expression. During field studies the author found one overgrown logging road that had been mapped as a lineament. Other landscape features, such as alluvial valley fill, and alluvial and colluvial fans and cones, may mask the surface expression of lineaments unless the feature has been recently active. In summary a lineament was mapped if it showed a variation in photo tone and is: straight or nearly straight; persistent over a length which may be reasonably mapped at the scale of the study; and if partly formed by a morphological feature such as a stream segment, continuous beyond the expression of that feature on the photograph. To facilitate further processing curvilinear features were mapped as a series of successive straight lineaments. 4.3 The lineament inventory The lineament inventory described below was compiled for two reasons: firstly, to gain an appreciation of regional structure and secondly to provide a framework within which to study the effects of lineaments on individual drainage basins and landslide events. 4.3.1 Processing The lineament inventory was prepared by air photo investigation of stereo air photos (approximate scale 1:60,000). Lineaments were drawn directly onto each photograph and checked 29 Chapter 4 The lineament inventory twice over a period of three months to ensure identification of all features before transfer to a 1:250,000 scale topographic map for subsequent digitization. Despite easy identification in areas of bare rock, lineaments less than 300 m (5 mm photo-length) in length, were not included in the inventory because of the difficulty in transfering these to the basemap. Photos were examined in stereo during transfer to a basemap and mapping accuracy depends on the certainty with which the endpoints of the lineament can be fixed with respect to topographic features. Placement of lineaments on a map in open terrain, such as a hillslope, is difficult where no distinguishing features exist. There is a maximum placement error of + 250 m (within two pixels in subsequent analysis (see Section 5.6.1.)) in the location of lineaments on the map. This represents a map distance of between 0 - 250 m (0 - 1 mm) on a 1:250,000 scale map. This inaccuracy represents the largest probable ground error. In practice persistent lineaments, intersecting a number of topographic features, can be easily placed in comparison with shorter features visible on a single hillslope. When the accuracy of placement was examined features in the Seymour watershed were found to have a maximum placement error of 150 m (see Section 6.4.1). Lineaments were digitized and UTM (Universal Transverse Mercator) grid referenced in AUTOCAD 12 for WINDOWS (ACAD). A total of 4215 lineaments were transferred to ACAD and exported to IDRISI by way of a data exchange (dxf) file. In IDRISI, a rastorized image file and a vector file was generated. The vector file was processed in a FORTRAN program \"TREND\" to determine lineament orientations. TREND calculates the angle between the first and last points of each line. The output is distributed between 0 and 179°. To construct a rose diagram of lineament data the output file from trend is saved as a TEXT file in windows then displayed graphically by ROSE 1.02. Details of ACAD and IDRISI may be found in the user manuals, Autodesk Inc. (1993) and Eastman (1993) respectively. TREND was written by Leir (1994) for in-house use at the University of 30 Chapter 4 The lineament inventory British Columbia, Geological Sciences Department. Rose 1.02 is shareware software available from Thompson and Thompson, Indiana, USA. 4.3.2 The scale effect When mapping on lower altitude (larger scale) air photos the number of lineaments detectable increases. Table 4.1 shows the result of remapping lineaments in sample set basins for which 1:60,000 and 1:20,000 scale photographs were available and compares number, and length of lineaments mapped at both scales. There is an average threefold increase in the length of lineaments mapped on the larger scale photos. Lineament length is calculated in IDRISI by multiplying the number of lineament pixels by the length of one side of one pixel. In the field the number of lineaments detectable in all ground conditions but bare rock exposure drops appreciably (see Section 4.4). Basin # 1:60,000 scale air photo mapping 1:20,000 scale air photo mapping % lin. increase # of lins. Lin. length (m) Lin. density (/m2) # of Lins. Lin. length (m) Lin. density (/m2) 1 14 8300 0.00752 36 16593 0.01503 199.9 16 5 2440 0.00326 17 11466 0.01533 469.9 25 6 1680 0.00614 18 5034 0.01839 299.6 26 9 4900 0.00903 20 8371 0.01542 170.8 Average 285.05 Table 4.1. Comparison of lineament parameters mapped on high and low altitude air photos. The basin number corresponds to numbers reported in Chapter 5, and the percentage lineament increase in the final column is the increase seen on the lower altitude air photos. 4.3.3 Lineament trends Figure 4.1 is a rose diagram representation of the lineament data with the calculated statistics and Figure 4.2 shows the lineament map derived from the study. 31 Chapter 4 The lineament inventory Arithmetic Plot Number of Points 4215 Class Size 5 Maximum Percent 6 Vector Mean 17.43 Vector Magnitude 1344.61 Consistency Ratio 0.3190 Figure 4.1. Rose diagram showing all lineaments in the inventory. 32 Chapter 4 The lineament inventory 122 25 W 122 W 50\" N 49 N 10 Ktocyettrs 0 50 N I / ^ I. NE \\ 122 25 V L e o e n d ^ L i n e a n e n i ; F a u l t ^ 9 Figure 4.2. The lineament inventory map. The inset shows the breakdown of this area into the blocks discussed in the text. The distinction between lineaments and faults indicates features mapped by the GSC from Roddick (1965). 33 Chapter 4 The lineament inventory The rose diagram (Fig. 4.1) shows a preferred trend of 17.43 ± 3.73° (see below) and peaks in the data representing east-west and northwest-southeast trends. The data is grouped in five degree classes and tabulated in Appendix I. Statistics were compiled in ROSE 1.02 using standard techniques for evaluating directional data (see Davis (1986), and summary examples in Appendix II). The vector mean is the direction of the resultant vector if all lineaments are placed end to end in classical \"head to tail\" fashion. Standard error is calculated according to Davis (1986, p.325) at the 95% confidence interval. The magnitude of the resultant vector is its length assuming each vector (lineament) is of unit length. Consistency ratio is a means of standardizing the vector magnitude and measuring the distribution of the data and varies from zero to one. A value near one implies the data is tightly grouped. The data is tested for randomness to determine whether a preferred trend exists. Calculation of the standard error is detailed in Appendix II. From Figure 4.2 several observations are made: There is a higher lineament density in the region between approximately 49° 23' N . and 49° 45' N . This region contains many of the longer lineaments in particular several large east-west trending lineaments. North of 49° 45' N . , the map area can be split into two regions, east and west of 122° 35' W. To the west the density of mapped lineaments is reduced by numerous ice fields and glaciers. Farther east is a proliferation of northwest trending lineaments following Harrison Lake - Pemberton Valley. Lineament density is lower in the Southern area because of the presence of several wide, flat bottomed, alluvial filled valleys, e.g., those of Pitt, and Stave rivers. Figure 4.3 shows the lineament trend data for the four blocks discussed and Table 4.2 summarizes lineament data for the regional study and these areas. The GVRD watersheds represent a subset of the regional study area and the field area used in this study. 34 Chapter 4 The lineament inventory Southern Block Central Block Arithmetic Plot Number of Points 487 Class Size 5 Maximum Percent 7 Vector Mean 23.08 Vector Magnitude 209.20 Consistency Ratio 0.4296 Northwest Block Arithmetic Plot Number of Points 614 Class Size 5 Maximum Percent 7 Vector Mean 10.59 Vector Magnitude 218.92 Consistency Ratio 0.3565 Arithmetic Plot Number of Points 2546 Class Size 5 Maximum Percent 7 Vector Mean 22.66 Vector Magnitude 895.45 Consistency Ratio 0.3517 Northeast Block Arithmetic Plot Number of Points 625 Class Size 5 Maximum Percent 5 Vector Mean 346.01 Vector Magnitude 184.34 Consistency Ratio 0.2949 Figure 4.3. Rose diagrams for each of the blocks described in the text. 35 Chapter 4 The lineament inventory The number of lineaments calculated for the individual blocks totals 4272 (57 (1.35%) of the lineaments are repeated). Lineaments were separated in ACAD by selecting the required block and deleting all others before export to IDRISI. In doing this some overlapping lineaments may be included in two sections. A visual assessment of each block was made and the most apparent overlapping lineaments removed. The data confirms the visual assessment of lineament distribution. The central block has the highest lineament density and the southern block the lowest. GVRD watersheds overlap the southern and central blocks and visual assessment of the lineament distribution within the watersheds reveals a slightly higher lineament concentration in the northern parts of the basins (Fig. 4.4). Site Area (km2) #of lins. Total length of lins. (Km.) Average length (km.) Lineament density (/km2) Mean Trend Orientation Regional area 10,900 4215 8266.3 1.961 1.319 17.43±3.73° Southern block 4000 487 1015.84 2.086 0.001* 23.08±7.95° Central Block 4000 2546 5303.06 2.083 1.326 22.66±4.34° Northeast block 1160 625 1010.38 1.617 0.871 166.01±10.52° Northwest block 1740 614 1080.8 1.760 0.621 10.59±8.5° GVRD watersheds 596 474 794.36 1.676 0.751 30.2±8.46° Table 4.2. Summary information for lineaments. (*) A more realistic value is 0.677 because approximately 2,500 km2 of this block resides in the developed Fraser Lowland. The larger the sample size the more constrained is the mean trend. The standard error for the entire data set is sufficient to capture all but the mean of the northeastern block and GVRD watersheds. 36 Chapter 4 The lineament inventory 123 15 U 122 40 V 49 40 N 49 20 N 123' 15 v Capi lano y ' )/ 7 R i v e r S e y m o u r R ive r 122 40 v S c a l e in k i l o m e t e r s Coqui t l a m R i v e r : 0 10 20 Figure 4.4. Lineament distribution in the GVRD watersheds. The approximate location of key field sites is also shown: AC = Appian Creek; OC = Orchid Creek; CC = Camp Creek. 37 Chapter 4 The lineament inventory The GVRD watershed dataset contains the means of both the southern and central blocks. A more rigorous comparison of data is made by testing for equivalence of the mean directions statistically. This is done using the two-sample test of Watson and Williams (Mardia 1972, Davis 1986). Only results of these tests are detailed here, an example calculation is found in Appendix II. The south and central blocks derive from the same statistical population but the two northern blocks differ significantly both from those in the two southern blocks (which were joined for the purposes of analysis) and from each other. The GVRD sample set cannot be tested in the same fashion because it is not independent from the south-central dataset. The primary difference between the two northerly blocks is the presence of long, northwesterly trending lineaments in Harrison Lake-Pemberton Valley, the distribution of which corresponds with the outcrop of the Jurassic Fire Lake Group and other metamorphic rocks in the area (Fig. 4.5). Statistical similarities and differences between areas and variations in structural style may be due to either variable rock type response to a particular lineament generating stress field or, more likely, to differences in the stress history of particular rocks. As an example, the northwesterly trend of the major rivers dates from the Jurassic (Journey pers. comm. 1995). This trend is not particularly evident anywhere on the lineament map except where Jurassic and pre-Jurassic rocks occur. It is reasonable to conclude that the these rocks have retained an older structural trend while the younger rocks of the Coast Plutonic Complex have only been subjected to younger stress fields. This younger stress field has imprinted itself over the Jurassic and pre-Jurassic trend as is demonstrated by the additional presence of a northeasterly trend in the northeastern block. This trend is likely to be representative of the Tertiary trend identified by Journeay (1990). Details of the origin and nature of these trends were presented in Chapter 3, Section 3.4. 38 Chapter 4 The lineament inventory Figure 4.5. Simplified map of northeastern block geology and lineaments: F L G = Fire Lake Group; TIG = Twin Island Group; M = Migmatite; Qdio = Quartz diorite; Gdio = Granodiorite; A l = Alluvium; I = Ice. (Adapted from Roddick 1965). 39 Chapter 4 The lineament inventory The northeastern block is the only block underlain by significant amounts of metasedimentary rock. The northwestern block contains areas of pre-Jurassic rocks which, when combined with the presence of numerous icefields are sufficient to disrupt the lineament patterns away from trends seen in the dominantly crystalline igneous, remaining area. Other areas of metasedimentary rocks do exist, notably on the eastern flanks of Howe Sound but these are areally small. The east-west trend visible in the central and southern blocks may represent the youngest regional trend. This trend does not correlate with a recognized structural trend. The present convergence vector of the Juan de Fuca and the North American plate is oriented at approximately (060°) northeast-southwest (Fig. 4.6). If it is assumed that the convergence vector represents the axis of maximum principal stress then the angle at which faulting is expected is given by: 0 = 45°-(|>/2 Where <|> = tan'u, and u. is the coefficient of sliding friction. For most rocks at the large-scale <|> is taken as 30°, hence dextral faulting can be expected oriented at approximately 030° (north-northeast) and sinistral faulting oriented approximately 090° (east-west). The presence of north-northeast faults in this region would be obscured by the Tertiary trend already identified above, however, east-west oriented lineaments dominate the area southwest of the Harrison Lake - Pemberton Valley. It is suggested that the east-west trend may be a young, previously unrecognized structural trend in the southwest Coast Mountains. These lineaments are laterally extensive and cross-cut both 40 Chapter 4 The lineament inventory Figure 4.6. A schematic diagram showing the orientation of the maximum principle stress ( 40% correlation. This increases substantially when streams within the one pixel acceptance level are taken. Then, 76% of the basins show a > 40% correlation and 56% of the basins show a > 60% correlation. Basins #2, 20 and 25 have a greater than > 80% correlation. These values again increase in the two pixel acceptance level: 72% of the basins show a > 60% correlation and 56% of the basins show correlation's at the > 80% level. When only first order streams are considered there is an improvement in the strengths of the correlations. Again six (24%) of the sample set basins show a > 40% correlation with the direct overlay, but now three (12%) of the basins are at the > 50% level with one at the > 60% level. Within one pixel, 84% of the sample basins lie above the > 40% correlation level. Basin #2 shows 100% correlation of first order streams and lineaments and six (24%) of the basins show a correlation of > 80%+. Within the two pixel zone, two basins show 100% correlation (Basins #2 and 4). Eight of the basins are now at the > 80% correlation level. Second order streams show an initial decrease in the strength of correlations. Only four (16%) of the basins show a correlation at the > 40% level when direct overlay is required. This number rises to 17 (68%) when the one pixel zone is used. In this class one of the basins (#2) shows a 100% correlation and 12 (48%) of the basins show a correlation at the > 60% level. At the same level 16 (64%) of the basins show a correlation at the two pixel acceptance level and three basins show 100% correlation. Al l of the sample set basins have first and second order streams. Only 10 of the 25 samples have third and two of the samples have fourth order streams (basins #19 and 29). For third order streams only basin #18 shows a > 40% correlation for streams directly overlying lineaments. Within the one pixel acceptance level five (20%) of the basins show correlation at the same (> 40%) level. 82 Chapter 5 Drainage basin analysis This figure is constant for the two pixel acceptance level but the correlation for basin #18 increases from 68% to 76%. Table 5.5 shows the average percentages of streams in the sample set basins overlaying lineaments. An average of 66% of stream pixels per basin are correlated with lineaments. When broken down by lithology 71% of stream pixels correlate with lineaments in igneous basins while 46% correlate in metamorphic basins. On average for the sample set basins 51% of all lineaments correlate with stream pixels. These numbers are respectively 57% and 28% for igneous and metamorphic basins. Al l si Ore ream ers 1st Order 2nd Order 3rd Order mean s.d. mean s.d. mean s.d. mean s.d. Al l basins 65.71 20.99 68.47 22.45 63.25 26.57 37.71 24.51 Igneous basins 70.74 19.17 73.21 21.35 66.75 26.29 51.24 22.13 Metamorphic basins 45.57 18.11 49.52 15.74 49.26 22.81 17.41 18.85 Table 5.5. Showing the average percentages of streams in the sample set basins overlaying lineaments at the two pixel acceptance level, (s.d. = standard deviation). 5.6.3 Discussion The results appear to indicate that there is a high correlation between streams and lineaments in the sample set basins. The correlation is significantly better in the igneous basins. For the whole sample set an average of 51% of the lineament length within a basin is within the two pixel acceptance level for correlation with streams (for reasons described below this is likely to be a minimum value). In other words about half of the lineament length in each basin is occupied by stream channels. This generalization is more reliable in igneous basins and less so in metamorphic basins. 83 Chapter 5 Drainage basin analysis Nearly 73% of first order streams in igneous basins correlate with lineaments and 67% of second order channels do likewise. In metamorphic basins these numbers decrease to 49% for both first and second order streams. In igneous basins 50% of third order streams lie in lineaments. This evidence seems consistent with the hypothesis that lineaments exert a significant control on the drainage pattern. In particular this seems true of lower order streams. While it would seem initially that correlation within two pixels is imprecise, it is acceptable within limits of mapping accuracy of ± 50 m. In the field lineaments can be many meters wide, for example see Figure 4.10. When mapped on the computer differences in the selection of digitized points for streams and lineaments can be sufficient to cause the two to separate over part of their course even if they are obviously related. Neither streams nor lineaments were intentionally digitized to follow the same path in cases where this appeared to be the case. There are two problems with this means of analysis. The main problem with the methodology is that some pixels representing streams which should be included are missed due to sinuosity of the stream of greater amplitude than is included in the two pixel acceptance level. This is not a major problem in this analysis because few of the streams in the study area exhibit much sinuosity except where a stream meanders between two lineaments as is the case of Basin #14. Higher order streams in the larger basins present a greater risk for this kind of error but these streams are generally not well correlated with lineaments. These streams tend to be incised in glacial and alluvial deposits that mask the effects of lineaments in the rock mass. An additional error is the inclusion of pixels which should not be added to the results of the analysis. This results from locations where lineaments intersect streams obliquely. In this case the lineament has no effect on the stream in question in terms of controlling orientation. In general when a stream/lineament crossover occurs a total of five stream pixels will be counted which should not be included in the results. To solve this problem a visual assessment is made of the number of crossover 84 Chapter 5 Drainage basin analysis points in a basin. For each crossover five pixels were subtracted from the results for the appropriate stream order. The reported correlations (Table 5.4) are best considered as a minimum correlation strength because there are a number of streams in basins which parallel lineament trends but which are not themselves seen as being lineament controlled because of the restrictions imposed on lineament mapping (Section 4.2). This appears to be the case for 15 additional stream segments from seven basins. Five basins (#5, 16, 19, 24 and 29) consistently show low correlation's between streams and lineaments. Basins #5, 19 and 29 are the largest basins in the study set and each of these basins is developed on metamorphic rocks (Basins #5 and #19) or has variable lithologies (Basin #29). Two other basins developed on metamorphic rocks (basins #14 and 21) consistently appear in the middle range of correlation values. There is, therefore, some evidence for concluding that lineament control is not as prevalent in metamorphic basins. A more compelling reason for the low correlation is the larger size and greater (inferred) maturity of these three basins. In a larger basin it may be expected that there would be a much greater variability in the possible stream courses since every point in the basin must be drained and if a suitable lineament is not available the stream must make its own course. Larger basins also tend to have flatter bottoms where streams will incise alluvial, colluvial and glacial deposits, much more so than in the high basin walls therefore, a larger proportion of the stream network will be flowing in areas where lineament effects are masked. A further possible explanation is that the metamorphic rocks on which the larger basins are developed are, softer rocks than the crystalline igneous rocks of the other basins. These, in addition to being more easily erodable, may have primary structures which obscure lineament effects. In basin #16 the main stream appears lineament controlled but it plots poorly in the correlation tables because of the location of the lineament. The lineament is mapped at the top of the 85 Chapter 5 Drainage basin analysis divide but the entire southern wall of the basin is likely to be lineament controlled. The stream flows at its base. Unfortunately this places the stream at a considerable distance from the mapping of the lineament. This kind of error could be remedied by subjectively mapping lineaments, however the controlling effect on the stream is interpretive only (although strongly suggested), and is not known for sure. Basin #24 plots here probably because of the low number of lineaments in the basin, one of which does control, in part, the course of the major stream. In basin #24 the major stream does not follow the basin axis. 5.6.4 Directional correlation of lineament and stream trend datasets In order to separate the orientation data for lineament controlled streams it was necessary to visually determine the orientations of lineaments that controlled streams and manually separate them from the rest of the data. This could not be done automatically. 153 lineaments were visually interpreted as controlling stream segments. A rose plot (Fig. 5.9) shows the orientations of lineament controlled streams. Three notable peaks occur, approximately north-south, northeast-southwest, and east-west. The northwest-southeast trend occupied by the major river valleys (Fig. 3.2) appears rarely paralleled in the sample set basins. When this plot is divided between basins developed on igneous and metamorphic rocks the mean trends are broadly similar although, statistically, there is no preferred trend in the metamorphic basins. The statistical relations between the data for stream trends and lineament trends in the sample set basins were analyzed by the same statistical methods described in Chapter 4, and Appendix II. Table 5.4 shows the mean directions for the data sets relevant to this analysis. Summary data for calculations and vector statistics are presented in Appendix VI. 86 Chapter 5 Al l lineament controlled streams Drainage basin analysis igneous basins Arithmetic Plot Number of Points Class Size Maximum Percent Vector Mean Vector Magnitude Consistency Ratio N 153 5 8 21.94 26.97 0.1763 metamorphic basins Arithmetic Plot Arithmetic Plot Number of Points 85 Number of Points 68 Class Size 5 Class Size 5 Maximum Percent 7 Maximum Percent 10 Vector Mean 22.90 Vector Mean 20.68 Vector Magnitude 15.33 Vector Magnitude 11.66 Consistency Ratio 0.1803 Consistency Ratio 0.1715 Figure 5.9. Orientation of visually assessed lineament controlled streams for the entire dataset and for igneous and metamorphic streams individually. 87 Chapter 5 Drainage basin analysis The overall trend of lineaments in these basins can be accommodated within that of the entire regional study site (17.43 ± 3.73°). When tested for equivalence of means the sample set lineaments are shown to come from the same lineament population as the regional data set (described in Section 4.6.3). As well as demonstrating that the sample set is representative of the overall lineament pattern in the region this shows that remapping of the lineaments in these basins at 1:20,000 scale does not alter observer bias in the orientations of lineaments mapped from the photographs. Stream segments in general show no preferred trend and neither do lineaments that control streams in metamorphic basins (Table 5.4). Neither first, nor second order streams show a preferred trend until they are split into datasets indicative of their geology (Table 5.4). Third order streams do show a preferred trend. When all stream segments are considered one might account for the lack of a preferred trend by recognizing that any particular order of stream might bias the data against a preferred trend by being randomly oriented. The distribution of the number of streams of a given order in a basin is controlled by a power law (Ritter et al. 1995). One would expect to see a relation between lineaments and streams more strongly evidenced in the lower order streams because first order streams are typically higher on the slopes (where there is little surface cover) and are more likely to have trends closely reflecting bedrock structure. Intermediate slopes containing second order streams typically show thickening surficial deposits which may mask the effects of some lineaments and allow for a more random distribution of stream orientations. Analysis revealed that second order streams have no preferred orientation, while the first and third order streams do. The trend of third order streams may be explained by lineament control of the position of the basin axis, i.e., although these streams may not flow on a lineament, or within the two 88 Chapter 5 Drainage basin analysis Description of dataset # of lineaments/ segments Mean trend + standard deviation Lineament trends 509 11.31+17.44 Igneous basins 195 28.8±24.13 Metamorphic basins 314 0.15±20.10 Stream segment trends* 882 150.33+52.05 Igneous basins 230 20.79+43.31 Metamorphic basins 652 167.1±39.87 Lineaments visually controlling stream orientations 153 21.94+35.74 Igneous basins 85 22.9±47.42 Metamorphic basins* 68 20.68+55.98 First order stream segments* 559 147.22+56.49 Igneous basins 171 24.42±30.27 (NNE) Metamorphic basins 388 176.37±43.98 (N-S) Second order stream segments* 168 342.88±101.09 Igneous basins* 52 346.6+105.79 (NNW) Metamorphic basins* 116 338.49+181.77 (NNW) Third order stream segments 96 172.35±25.69 Igneous basins 15 200.84±30.32 (NNE) Metamorphic basins 81 163.7+27.04 (NNW) Table 5.6. Orientation data for lineament and stream segment trends in the sample set basins. (* indicates no preferred trend in the data). 89 Chapter 5 Drainage basin analysis pixel acceptance level (Section 5.6.1), the orientation of these streams may be closely constrained by lineaments. An alternate explanation for the apparent random orientation of all first and second order streams is the bias in the number of these streams occurring in the larger metamorphic basins. In each instance there are more than double the number of segments in the metamorphic basins. Since these basins have apparently been subject to at least two phases of lineament inducing stress (see Section 4.5) one would expect the lineaments and hence streams to have a random orientation. This also explains the lack of a preferred trend in the lineaments visually controlling streams in the metamorphic basins. When the lineaments controlling stream segments in both igneous and metamorphic basins are compared there is an equivalence which suggests a preferred lineament orientation of 21.94 ± 35.74° occupied by streams in both subsets of basins. This is from a population comprising almost entirely first order streams with an approximately 55% in igneous basins. When the trends for all first order streams in the sample set basins are compared with the trends for all lineaments in igneous basins it is found that the mean trends agree closely. This is also true of comparing all igneous basin, first order streams with igneous basin lineaments but is not true of the same comparison within metamorphic basins. This might imply that lineament control is preferred in igneous basins. There is no evidence for equivalence of means in any of the second or third stream order datasets. When the numbers are generalized to compass directions (shown selectively in Table 5.6) we see that first and third order streams in igneous basins tend to trend to the NNE whereas second order streams have a mean NNW trend (although this is not a statistically preferred direction). In metamorphic basins first order streams trend approximately north-south, and second and third order 90 Chapter 5 Drainage basin analysis streams trend northwesterly. It is known that a northeast trend is in agreement with the Tertiary structural trend in the region whereas a northwest trend agrees with the older trend in the region. The statistical evidence suggests that there is a preferred direction for the stream segments which are visually assessed to be lineament controlled however this dataset comprises only 18% of all mapped stream segments in the sample set basins. The data suggests that the preferred direction is 21.94 ± 35.74°, and this is close to the mean trend of lineaments in the regional inventory (Chapter 4). There is no preferred trend in the data for all sample set basins. There is a preferred trend for each subset divided on lithology. The mean orientation for all stream segments in the igneous basins (20.79 ± 43.31°) is similar to that in the metamorphic basins (167.1 ± 39.87°) but these datasets cannot be shown to derive from the same statistical population. The conclusion to be drawn from this is that in the igneous basins the orientation of stream segments is strongly controlled by lineaments. In the metamorphic basins where at least two different lineament sets have been imprinted on one another the orientation of stream segments is consequently more varied. While the average orientations seem to agree well there is sufficient statistical variation in the samples that a test for the equivalence of means does not show a good correlation. The calculated resultant magnitude is 0.0876 and the pooled value is 0.0527. This is a difference of 0.0349 whereas other similar calculations show better correlations. Although it cannot be statistically proven it does seem that there is a preferred orientation for lineament controlled streams in the area and that this is a NNE trend which is similar to the Tertiary structural trend experienced by the region. This trend is found dominantly in first order streams and suggests that the orientation of the main streams was fixed by the Tertiary probably along the northwest structural fabric and that the lower order streams, which are responsible for headward erosion of the basins, are influenced by the more recent lineament trends (Tertiary and later). 91 Chapter 5 Drainage basin analysis 5.6.5 Stream incidence angles The spatial correlation, and the similarities in trends between stream channels and lineaments is compelling reason for accepting the hypothesis that many stream channels are lineament controlled. However, the question of natural flow paths on the hillslopes should not be ignored. One means of determining whether a stream is following the expected (i.e. due to gravity alone) path down a hillslope is to examine the junction angles of the streams to determine whether they match those predicted by the topography. To do this Horton (1945, p. 349) established the following geometric relation: Cos z c = tan sc/tan ss Where zc is the angle between two streams; sc is the slope of the parent stream and ss is the ground slope or the slope of the tributary stream. While this formula was found by Horton to reasonably predict stream entrance angles it was found to be too sensitive for practical use with the data collected here. The problem is the sensitivity of the formula to small variations in the measurements of the gradients in question. This is compounded by the fact that small changes in the ratio on the right hand side of the equation cause large changes in the angle produced on the left. This meant that when the errors in measuring the gradients of the two streams making up a single junction are accounted for, the range of angles predicted is very large, easily encompassing both the angle of incidence measured on the map and that predicted by the formula. The errors in measuring the gradient of the stream lie in the correct assignment of elevation and length. The elevation is taken as accurate to within half a contour interval and the length to within one half of a millimeter (it is customary to measure to within one half of the lowest gradation on the measuring device) which represents 25 m. As an example the junction at the western end of 92 Chapter 5 Drainage basin analysis basin #2 is considered, (see Fig. 5.9). Both streams are judged to be lineament controlled. The data for each stream is summarized in Table 5.7. Upper elevation (m) Stream length (m) Lower elevation (m) Maximum slope Minimum slope Northern stream 1320+20 375125 1120+20 34° 25° Southern stream 1400+20 550+25 1120120 31° 17° Table 5.7. Summary data for the two streams in the cited example. The 'worst case scenario' that will result in the greatest range of predicted values occurs when the two most varied and the two most similar gradients are placed into the formula. These result in predicted angles of between 0 and 65°. This range encompasses the measured angle of 28°. While it is recognized that errors will be reduced over longer length streams, where a length variation of 1 25 m is less consequential, analysis of all stream intersections in basin #2 showed little improvement on the use of the formula. Additionally many streams that are lineament controlled along their entire length to the junction with a higher order stream are not long, but rather short first order features. In conclusion, while measurement of incidence angles is valuable in determining for certain whether natural flow laws are being observed or whether some other control is in effect, the sensitivity of the predicting formula is too great for the measurements made in this analysis. Accuracy could perhaps be improved by the utilization of a DEM. This would improve the accuracy of measures and could automatically calculate gradients. However it is felt that the results of the spatial overlay and of the analysis of trend data are sufficient evidence to accept the hypothesis that stream paths are lineament controlled. 93 Chapter 5 Drainage basin analysis 5.6.6 Comparison of visual assessment and automatic evaluation of lineament control on streams In order to compare the accuracy of visual assessment of lineament control on streams, the assessment reported in Table 5.2 is compared to the calculated percentage of streams that correlate with lineaments reported in Appendix V. The range of values assigned to those basins in each of the three classes in Table 5.2 is evaluated. The results are presented in Table 5.8. Visual classification number of basins mean of stream/lin. correlation standard deviation. Poor 2 33.5 3.5 Moderate 10 55.1 20.13 High 13 78.7 10.77 Table 5.8. Summarizing the correlation results from GIS analysis as compared to visual assessment. In order to establish the typical values for each class resulting from visual inspection the overlap of the standard deviations from each class was bisected and class ranges drawn at the midpoints. This resulted in the following classification: Poor visual correlation: 0 - 36 % of streams correlate with lineaments Moderate visual correlation: 37 - 72 % of streams correlate with lineaments High visual correlation: 73 - 100 % of streams correlate with lineaments This provides an indication of the range of values that may be expected from a visual interpretation of structural control conducted on 1:60,000 scale air photos in this environment. Comparison of the data will show that basin #16, is assigned a moderate lineament control whereas the actual amount of correlation is only 19%. This is largely due to the placement error described in Section 5.6.3 (page 85). Note that basin #2, has the highest correlation at 93% (see also Figure 5.7). 94 Chapter 5 Drainage basin analysis 5.7 Regression analysis of lineament control on basin morphometry It has been shown (Section 5.6.2) that an average 66% of the stream network is a reflection of the lineament pattern inherent in the underlying rock mass. It is, therefore, reasonable to speculate that basin morphometric parameters might also be connected to the underlying structural pattern. In order to assess the validity of this hypothesis the relation between the geometry of the basin, and the lineaments is examined by regression analysis. If the drainage network of a basin is controlled by the lineament pattern then the parameters that describe the lineament characteristics of a basin should exhibit similar relations to basin geometry as the drainage pattern. 5.7.1 Method Each morphometric parameter measured is tested against each of the other parameters by means of regression analysis performed in QPRO. The results of the analysis provide the values necessary for a regression line through the data. An example of regression output is presented in Table 5.9 (for details of linear regression see Ferguson 1976). The R 2 value may be thought of as an assessment of the strength of the correlation, or fit, of the predicted line through the data, its value is used as a guideline for further investigation. It is important to note that it is only a guideline since it is possible for R 2 values to be dramatically influenced by outliers in the data. If the R 2 value is high, near one, then the graph is examined for the presence of outliers. It is also often the case that each of the relations checked: linear; log-linear; and log-log have similar R 2 values. In this case a graphical examination of the data is necessary to correctly determine the nature of the relation. In order to investigate the data it was grouped into three datasets: 95 Chapter 5 Drainage basin analysis Independent Dependent Constant Standard. R 2 X Coefficient Standard variable (X) variable (Y) Error of Y Estimate. Error of Coefficient Basin Area. Lin. Length 4709.86 9395.06 0.96 1.69E-03 6.80E-05 Lin. Length Basin Area. -2318010.58 5445861.71 0.96 569.21 569.21 Log(Basin A.) Lin. Length -341303.51 25878.33 0.73 58639.89 7454.04 Lin. Length Log(Basin A.) 5.92 0.38 0.73 1.24E-05 1.58E-06 Basin Area. Log(Lin. Length) 3.64 0.44 0.54 1.65E-08 3.18E-09 Log(Lin. Length) Basin Area. -114485550.8 19563204.19 0.54 32729578.9 6309561.4 Log(Basin A.) Log(Lin. Length) -1.23 0.27 0.83 0.81 0.08 Log(Lin. Length) Log(Basin A.) 2.31 0.30 0.83 1.02 0.10 Table 5.9. Showing the Regression Analysis of Basin Area and Lineament Length Variables. Set one: All data Set two: Basins larger than one square kilometer Set three: Basins smaller than one square kilometer Matrices were compiled (Appendix VII) showing the R 2 values for each of the data sets tested. It was found that three broad groups of data showed high R 2 values. First, sets of variables related to the areal and dimensional character of the basin such as basin area, and basin width, and length in general returned high correlations, as did those of relief. Second, there were high correlations between parameters describing the drainage network and key aspects of the basin morphometry. Finally there were high correlations between parameters describing the lineament patterns and the same key aspects of basin morphology. A certain number of relations may be thought of as primary mathematical relations. That is where one of the attributes directly contributes to the calculation of the other. An example is the lineament density parameter, where both lineament length and basin area are required to calculate the density. In this case the lineament length should be proportional to basin area and lineament density. The former should show up as a positive correlation and the latter a negative correlation in the regression analysis. When the regression correlations for all of the sample set basins are examined 96 Chapter 5 Drainage basin analysis (Appendix VII), it will be seen that lineament density does not correlate well with either basin area or lineament length. This is because the simple linear regression assumes that the two variables are related by a single explanatory variable (Ferguson 1976). Where the value of the independent variable (lineament density) is dependent upon more than one dependent variable then multiple regression is required to show the relationship. No multiple regression was conducted during this research. In the case of the primary mathematical relations the influence of each factor on the other is already established by its mathematical formulation. Examination of Table 5.2. shows which of the variables are directly related mathematically. Secondary mathematical relations also exist. It is possible to rewrite some of the relations to deduce unknown values if all other values are known for example, it is possible to relate basin length and cone apex elevation in the following manner: Basin Length = (Maximum Basin Elevation - Cone Apex Elevation)/Basin Gradient However in this case all other parameters must be known by observation, since basin length is required to calculate the basin gradient. Hence many parameters can be related mathematically. 5.7.2 Results of regression analysis Before examination of individual relationships between variables, two conclusions were drawn from this analysis. First, there is a significant difference in the strength of the relations, as expressed by their variance (R2), between each of the three datasets considered. With a few exceptions the correlations are better in basins larger than one square kilometer than in either of the other datasets. Second, log-log relations are more useful because outliers bias the data in linear relationships. This is particularly the case with basins #19 and 29 which cause regression lines for linear relationships to be drawn between two groups of data points (a cluster of data at one end of the 97 Chapter 5 Drainage basin analysis graph and two data points at the other). A trend derived from two clustered data points lacks the credibility of the log-log graphs which show a developed trend including all points. This section is split into two sub-sections the first describes general relations found by the analysis and the second compares the significant relations between the lineament and stream length and density, and other morphometric parameters of the basins. Examination of the matrices presented in Appendix VII will show that there are a large number of relations that have been tested (209 sets of variables for each matrix). Only a small number of variables are related. The relations between parameters which might relate directly to sediment yield are described in detail in Section 5.8: These are fan gradient and area, and basin area and relief ratio. Discussion of basin area and relief ratio is also found below. Throughout the following discussions R 2 values are reported for the larger basins and refer to log-log relationships unless otherwise indicated. 5.7.2.1 General relations There is a weak positive relation between basin area and fan area (R2 = 0.57) indicating that as the size of the basin increases, the amount of material in the fan also increases. This is expected since the fan is the deposition site for material moving from the basin. The larger the area of the basin the larger the drainage network, and the more extensive the erosion and deposition on the fan. Basin length and basin width are strongly, positively correlated (R2 = 0.94) suggesting that there is a tendency for basins to get wider as they become longer. Maximum basin relief and basin relief are both positively correlated with basin area (R2 = 0.63 and R 2 = 0.69) because smaller basins tend to occur higher in the mountains. Conversely there is a negative relation between the relief ratio (indicating overall steepness of the basin) and basin areal properties. Relief ratio will begin to decrease as soon as erosion begins, either by headward 98 Chapter 5 Drainage basin analysis erosion (increasing basin length) or by relief lowering (decreasing maximum basin elevation). Hence a mature basin will have a lower relief ratio than a more youthful basin. Because of the sensitivity of relief ratio to changes in basin length or height this parameter is probably the most useful in determining the amount of erosion that has been undergone by a basin. As such we may expect a negative correlation between this parameter and the fan area. Unfortunately in each of the three datasets the relations are too poorly correlated to say this for sure but the regression indicated relation is negative (R2 = 0.42, 0.39, and 0.08, respectively for the entire dataset, large and small basins). This is discussed more fully in Section 5.8. Basin gradient is negatively correlated with basin length (R2 = 0.87) and width (R2 = 0.82) because increasing the area of the basin decreases its slope. Basin gradient is positively correlated with relief ratio (R2 = 0.99) because gradient is dependent on basin length and height. In summary these observations show that increasing the areal parameters of a basin by erosion will decrease the relief ratio. This will also necessitate the removal of material from the basin which appears to increase the area of the fan. Hence these parameters (areal and relief) may be seen as a measure of the amount of material depleted from the basin. Interpreting these relations in the context of accepted landscape models it is demonstrated that as a basin increases in size, by headward erosion, the length and width of the basin increase. Headward erosion continues until the divide is reached at which point basin length and the width become fixed and erosion continues by relief lowering. This process is most clearly expressed by the relief ratio of the basin and is consistent with the Davisian model of landscape evolution. While a decrease in the relief of the basin probably occurs throughout the period of headward erosion it is only when the basin reaches the divides that it becomes the dominant process. Because of the lack of plateaus in the study area it is inferred that at the present time all basins probably have their areal parameters fixed and are only undergoing relief lowering. 99 Chapter 5 Drainage basin analysis 5.7.2.2 Relations between lineament and stream, and other morphometric parameters The hypothesis being tested is that basin morphometry is related to lineaments. These relations are examined in this section by comparison with the relations between the stream length and basin morphometry. Stream length is the more classically accepted control of drainage basin morphometry. Results of regression analysis show high R 2 values when stream length and lineament length are regressed against certain morphometric variables, these are: fan area; basin area; maximum basin relief; relief ratio; basin length; basin width; fan gradient; basin relief; and basin gradient (see figures 5.11 - 5.19). Regression equations for all relations described below are presented in Table 5.10. The log-log relationship between stream length and lineament length (R2 = 0.82) is shown in Figure 5.10. Drainage density is less clearly related to lineament length (Fig, 5.11). The two parameters are negatively correlated suggesting that a larger lineament length in a basin does not increase the length of the stream network. This is initially surprising but might be accounted for by the fact that in large basins significantly more of the basin area is taken up by surficial, valley filling materials which mask the effects of lineaments on the lower slopes and at the valley floors. Therefore the majority of the lineament length visible and available for exploitation by streams is confined to a small area of the basin on the upper slopes, hence, even if every lineament in this area is occupied by a stream the effect on the overall drainage density will be minor. In the larger basins drainage density and lineament density are strongly correlated (R2 = 0.89). The relations between basin area and lineament length (R2 = 0.83) and stream length (R2 = 0.97) is shown in Figure 5.12. When lineament length is compared to basin area two outliers are seen in the smaller basins (basins #24, and 27). These basins are also among those with the lowest lineament density values as are basins #19 and 29 which also deviate from the major trend. When 100 Chapter 5 Drainage basin analysis Regression equations showing the relations between lineament and stream length, and other mornhometric variables Stream length versus lineament length All basins; Log (Stream length) = 0.72.Log (lineament length) + 0.96 (R2 = 0.82) Large basins; Log (Stream length) = 0.90.Log (lineament length) + 0.23 (R2 = 0.78) Small basins; (R2 = 0.0) Drainage densitv ersus lineament length All basins; Log (Drainage density) = 1.02.Log (lineament length) + 2.3 (R2 = 0.66) Large basins; Log (Drainage density) = 1.36.Log (lineament length) + 0.89 (R2 = 0.71) Small basins; Log (Drainage density) = 0.19.Log (lineament length) + 2.24 (R2 = 0.07) Basin area versus lineament length All basins; Log (Basin area) = 1.02.Log (lineament length) + 2.3 (R2 = 0.83) Large basins; Log (Basin area) = 1.36.Log (lineament length) + 0.89 (R2 = 0.97) Small basins; Log (Basin area) = 0.19.Log (lineament length) + 2.24 (R2 = 0.02) Basin area versus stream length All basins; Large Log (Basin area) = 1.39.Log (stream length) + 1.05 (R2 = 0.97) basins: Small Log (Basin area) = 1.47.Log (stream length) + 0.67 (R2 = 0.96) basins: Log (Basin area) =1.31 .Log (stream length) +1.32 (R2 = 0.74) Basin length versus lineament length All basins; Large Log (Basin length) = 0.49.Log (lineament length) + 1.35 (R2 = 0.68) basins Small Log (Basin length) = 0.67.Log (lineament length) + 0.60 (R2 = 0.92) basins:: Log (Basin length) = -0.1 .Log (lineament length) + 3.33 (R2 = 0.03) Basin length versus stream length All basins; Large Log (Basin length) = 0.71.Log (stream length) +0.61 (R2 = 0.88) basins; Small Log (Basin length) = 0.74.Log (stream length) + 0.49 (R2 = 0.91) basins: Log (Basin length) = 0.90.Log (stream length) - 0.01 (R2 = 0.03) Basin width versus lineament length All basins; Large Log (Basin width) = 0.59.Log (lineament length) + 0.82 (R2 = 0.88) basins: Small Log (Basin length) = 0.76.Log (lineament length) + 0.12 (R2 = 0.97) basins: Log (Basin length) = 0.17.Log (lineament length) + 2.21 (R2 = 0.23) Basin width versus stream length All basins; Large Log (Basin width) = 0.77.Log (stream length) + 0.22 (R2 = 0.94) basins: Small Log (Basin width) = 0.82.Log (stream length) - 0.01 (R2 = 0.95) basins: Log (Basin width) = 0.41.Log (stream length) +1.41 (R2 = 0.27) Maximum basin elevation versus lineament length All basins; Large Log (Maximum basin elevation) = 0.08.Log (lineament length) + 2.89 (R2 = 0.22) basins: Small Log (Maximum basin elevation) = 0.17.Log (lineament length) + 2.47 (R2 = 0.75) basins: (R2 = 0.0) Table 5.10 Regression equations for relations between lineaments and stream length with other morphometric variables. 101 Chapter 5 Drainage basin analysis Maximum basin elevation versus stream lensth All basins; Large Log (Maximum basin elevation) = 0.lO.Log (stream length) + 2.81 (R2 = 0.24) basins: Small Log (Maximum basin elevation) = 2.48.Log (stream length) + 0.18 (R2 = 0.67) basins: (R2 = 0.0) Maximum basin relief versus lineament length All basins; Large Log (Maximum basin relief) = 0.12.Log (lineament length) + 2.60 (R2 = 0.40) basins: Small Log (Maximum basin relief) = 0.22.Log (lineament length) + 2.19 (R2 = 0.66) basins: (R2 = 0.0) Maximum basin relief versus stream length All basins; Large Log (Maximum basin relief) = 0.8.Log (stream length) + 2.42 (R2 = 0.53) basins: Small Log (Maximum basin relief) = 0.25.Log (stream length) + 2.09 (R2 = 0.74) basins: Log (Maximum basin relief) = 0.21 .Log (stream length) + 2.32 (R2 = 0.10) Relief ratio versus lineament lensth All basins; Large Log (Relief Ratio) = -0.37.Log (lineament length) + 1.25 (R2 = 0.64) basins: Small Log (Relief Ratio) = -0.46.Log (lineament length) + 1.59 (R2 = 0.80) basins: Log (Relief Ratio) = 0.12.Log (lineament length) - 0.36 (R2 = 0.06) Relief ratio versus tream lensth All basins; Large Log (Relief Ratio) = -0.53.Log (stream length) +1.80 (R2 = 0.82) basins: Small Log (Relief Ratio) = -0.48.Log (stream length) + 1.60 (R2 = 0.74) basins: Log (Relief Ratio) = -0.68.Log (stream length) + 2.33 (R2 = 0.40) Basin relief versus lineament length All basins; Large Log (Basin relief) = 0.17.Log (lineament length) + 2.35 (R2 = 0.46) basins: Small Log (Basin relief) = 0.26.Log (lineament length) + 1.93 (R2 = 0.72) basins: (R2 = 0.0) Basin relief versus stream lensth All basins; Large Log (Basin relief) = 0.25.Log (stream length) +2.07 (R2 = 0.64) basins: Small Log (Basin relief) = 0.30.Log (stream length) + 1.82 (R2 = 0.79) basins: Log (Basin relief) = 0.60.Log (stream length) + 1.05 (R2 = 0.40) Basin gradient versus lineament lensth All basins; Large Log (Basin Gradient) = -0.33.Log (lineament length) + 1.01 (R2 = 0.66) basins: Small Log (Basin Gradient) = -0.41.Log (lineament length) + 1.33 (R2 = 0.76) basins: Log (Basin Gradient) = -0.46.Log (lineament length) +0.12 (R2 = 0.12) Basin gradient versus stream length All basins; Large Log (Basin Gradient) = -0.46.Log (stream length) +1.46 (R2 = 0.81) basins: Small Log (Basin Gradient) = -0.43.Log (stream length) + 1.33 (R2 = 0.69) basins: Log (Basin Gradient) = -0.34.Log (stream length) + 1.06 (R2 = 0.20) Table 5.10 (Continued) Regression equations for relations between lineaments and stream length with other morphometric variables. 102 Chapter 5 Drainage basin analysis Stream length vs. Lineament length (R 2 = 0.82) 1E6 £ 1E5 ^ Ar-j, c 1E4 (L> - 4 — » | 1000 03 •a 100 i—i 10 10 < 1 km2: R 2 = 0.0 I i 11 n | M i l l 1 1 I I I I III 1 1 I I I I I l | 1 1 I I I III 100 1000 1E4 ' 1 E 5 1E6 Stream length (m) Log (S.L.) = 0.72.Log(L.L.)+0.96 Figure 5.10. Plot of Stream length versus lineament length. C P 1 km2; R 2 = 0.71 1—1 1 1 1 H-H 1—1 1 1 1 UN 1— 4 1 1IH| 1 1 1 M I N I 1 h-H-H-H 10 100 1000 1E4 1E5 1E6 Lineament Length (m) Log(D.D.) = -0.3.Log(L.L.) - 1.34 Figure 5.11. Plot of Drainage density versus lineament length. 103 Chapter 5 Drainage basin analysis CO m Basin Area Vs. Lineament Length (All Data: R 2-0.83) 1E9 1E8 1E7 1E6 1E5 1E4 1000 < 1 km2; R2 = 0.19 —f I—1-H-H+t 1—H-H-HHf- I 11 III 1—I I I I I l+j 1-10 100 1000 1E4 1E5 1E6 Lineament Length (m) Log(B.A.) = 1.02Log(L.L.) + 2.3 1E5 Basin Length Vs. Stream Length (R 2 = 0.88) 1E4 c 1000 c 100 10 -t 1—I I I I l+t t 1—I—H-H+f- -4—1 I I I I |— 100 1000 1E4 1E5 Stream Length (m) 1E6 Log(B.L.) = 0.71.Log(S.L.)+0.61 Figure 5.12. Plots of basin area versus a) lineament length, and b) stream length. 104 Chapter 5 Drainage basin analysis Basin Width Vs. Lineament Length (R 2 -0 .90 ) 1E5 1E4 3 1000 I 100 10 > 1 km2: R 2 = 0.97 < 1 kiii2: R 2 = 0.23 - i — I - H + + H | 1—t-M-H-H) F — + - H + + H ) 1—i M I N I ] i—i i i H Hi 1—i 11111+ 1 10 100 1000 1E4 1E5 1E6 Lineament Length (m) LogfB.W.) = 0.59.Log(L.L) + 0.82 Basin Width Vs. Stream Length (R 2 =0.94) 1E5 10 100 -+— t - H - H - H 1 1 | _ ) H - H H ^ (— 1000 1E4 Stream Length (m) -+++i 1—I—i—I M i l l 1E5 1E6 LogfB.W.) = 0.77Log(S.L.)+0.22 Figure 5.13. Plots o f basin width versus a) lineament length, and b) stream length. 105 Chapter 5 Drainage basin analysis a 1E5 S 1E4 to g 1000 c | 100 Basin Length Vs Lineament Length (All Data:R 2 =0.76) 10 Cm < 1 km2: R 2 = 0.03 > 1 kniz: R /=0.93 1—t—1-H-H+l H 1—l-H-H-M- 1 1--+-H-H+I 1 1 1 1 1 1 III 1 1—I-H-H+ 10 100 1000 1E4 1E5 1E6 Lineament Length (m) Log(B.L) - 0.49.Log(L.L.) + 1.35 Basin Length Vs. Stream Length ( R 2 = 0.88) 1E5 S 1E4 g 1000 H-l c 8 100 > 1 km2: R 2 = 0.91 < I km2: R2 = 0.42 10 100 -i—i--i--t-t-i-t-| f — i — i - i - H - H - -f—M-H-H-1000 1E4 1E5 Stream Length (m) 1E6 Log(B.L.) = 0.71.Log(S.L.)+0.61 Figure 5.14. Plots of basin length versus a) lineament length, and b) stream length. 106 Chapter 5 Drainage basin analysis a Max . Basin Elevation Vs . L i n . Length (R 2= 0.22) 10 100 1000 1E4 1E5 1E6 Lineament Length (m) — Log(MBE) = 0.08.Log(L.L.) + 2.89 Max . Basin Elev. Vs. Stream Length (R2 = 0.24) 1E4 > .£ 1000 in 100 < 1 km2; R 2 = 0.0 > 1 km2: R 2 = 0.67 - ) — H - H - H 1—1 1 I I I I I | 1 1 1 I I I I H 1—I 1 I I I I I 100 1000 1E4 1E5 Stream Length (m) 1E6 Log(M.B.E.) = 0.1 .Log(S.L)+2.81 Figure 5.15. Plots o f maximum basin elevation versus a) lineament length, and b) stream length. 107 Chapter 5 Drainage basin analysis 1E4 Max. Basin Relief Vs. L i n . Length (R 2 =0.40) 10 100 1000 1E4 1E5 1E6 Lineament Length (m) Log(M.B.R.) = 0.12.Log(L.L.) + 2.6 1E4 Max . Basin Relief Vs. Stream Length ( R 2 = 0,53) § 1000 } B Vi m 100 X 10 100 < lkni2: R 2 = 0.10 > 1km2: R 2 = 0.74 1000 1E4 1E5 Stream Length (m) 1E6 Log(MBR) = 0.18.Log(S.L.) + 2.42 Figure 5.16. Plots of maximum basin relief versus a) lineament length, and b) stream length. 108 Chapter 5 Drainage basin analysis 10 * -1 IB 0.1 Relief Ratio Vs . Lineament Length ( R 2 = 0.64) 1km2: R 2 = 0. 06 > lkni2: R 2 = 0.80 - i — i i M i n i 1—i i M i n i 1—i—i—t-H-++f 1—i i i 1111| 1—i i i 1111-10 100 1000 1E4 1E5 Lineament Length (m) 1E6 — Log(RR) = -0.37.Log(LL)+l .25 Relief Ratio Vs . Stream Length (R 2 =0.82) 10 Log(R.R.) = -0.53.Log(S.L.) + 1.80 Figure 5.17. Plots of relief ratio versus a) lineament length, and b) stream length. 109 Chapter 5 Drainage basin analysis Basin Relief Vs. L i n . Length (R 2 = 0.46) 1E4T 10 H I M l+H 1 1—t—H-H+) 1 1 I I i I l l | 1 1 I I I l+H 1 1—H 10 100 1000 1E4 1E5 Lineament Length (m) 1E6 Log(B.R.) = 0.17.Log(L.L.) + 2.35 Basin Relief Vs . Stream Length (R2=0.64) 1 E 4 i ) -| 1 1 l-t-H-H-) 1 1 H - M + H 1 1 1 I I I I I | 1 1 l-H-H+l 100 1000 1E4 1E5 1E6 Stream Length (m) — Log(B.R.) = 0.25.Log(S.L.) + 2.07 Figure 5.18. Plots of basin relief versus a) lineament length, and b) stream length. 110 Chapter 5 Drainage basin analysis Basin Gradient Vs . Lineament Length (R2 = 0.66) 10 c3 o 1 .S S3 0.1 10 < 1 kni2: r2 = 0.12 H-H+t 1 1 ! I I 1 1 I I M i l 1 i I I 1 I III 100 1000 1E4 1E5 Lineament Length (m) Log(B.G.) = -0.33.Log(L.L.)+1.01 1E6 10 O 1 c 0.1 Basin Gradient Vs . Stream Length (R 2 = 0.81) 100 < 1 km2; R2=0.19 —i 1—I—l-t-H-j— I I I I I I 1 1— I I H 1000 1E4 1E5 Stream Length (m) Log(B.G.) = -0.46.Log(S.L.) + 1.46 1E6 Figure 5.19. Plots of basin gradient versus a) lineament length, and b) stream length. Chapter 5 Drainage basin analysis Figure 5.11 is examined these outliers are not apparent suggesting that even where no significant lineament network is developed, as is the case in basins #24 and 27, a drainage network develops that is sufficient to fully drain the basin. This implies that a lack of lineaments in a basin does not inhibit the formation a stream network. Increases in basin area are matched by an increases in the stream length (Fig. 5.12) however, an increase in the lineament length relies on the pre-existence of lineaments in the rock mass into which the basin expands. Lineaments may allow preferential basin growth along their course but the evidence suggests that although basin area and lineament length are strongly correlated (R2 = 0.83) there is no required lineament length necessary for the development of a particular basin area. The typical elongate shape of a drainage basin shows that basins develop preferentially in one direction. Basin width is more accurately predicted by both stream length (R2 = 0.94) and lineament length (R2 = 0.90) than is basin length (R2 = 0.88, and 0.76 respectively). This is shown by comparison of figures 5.13, and 5.14. The data shows lineament length (R2 = 0.75) to be a better predictor of maximum basin elevation than stream length (R2 = 0.67). These relations are shown in Figure 5.15. The regression line relating maximum basin elevation and stream length is slightly biased by basins #19 and 29 which are significantly outlying from the other data. Maximum basin elevation in the entire dataset is not well predicted by either lineament length (R2 = 0.0) or stream length (R2 = 0.0) in the smaller basins. Maximum basin relief is better predicted by stream length (R2 = 0.74) than lineament length R 2 = 0.66) as shown in Figure 5.16. The two largest basins (#19 and 29) may again be biasing the equation of the regression line although their removal from the dataset does not significantly improve the strength of the correlation. 112 Chapter 5 Drainage basin analysis Stream length better predicts relief ratio (R2 = 0.82 for all data) than does lineament length (R2 = 0.64 all data) however in the larger basins lineament length has the higher R2 value (0.80 as compared to 0.74, Fig. 5.17). The spread of data about the regression line for small basins in both plots suggests that stream length is again the better predictor although R 2 values in both cases are low (0.06 for lineaments and 0.40 for stream length). Basin relief is negatively correlated with lineament and stream length and is better predicted by lineament length in the larger basins (R2 = 0.72, Fig. 5.18) although the value for stream length is similar (R2 = 0.64). Stream length is the better predictor than lineament length of basin gradient (R2 = 0.81 compared to R 2 = 0.66) in the entire data (Fig. 5.19). In general lineament density does not correlate well with basin. Drainage density shows good correlation with relief ratio (R2 = 0.79), basin length (R2 = 0.82), width (R2 = 0.81), and gradient (R2 = 0.79). There are no relations between either lineament density or drainage density in the small basins. Where relations do exist the associated relation is invariably stronger when the linear properties are considered as opposed to the density parameters. 5.7.3 Discussion Stream length is the best predictor of basin areal parameters (basin area, basin length and basin width) when the entire data set is considered. Lineament and stream length show very similar strength correlations in the larger basins where lineament length is the better predictor of all areal parameters. Maximum basin elevation and maximum basin relief are poorly predicted in the entire data set but are reasonably predicted (by lineament length for the former and by stream length for the latter) in the larger basins. Relief ratio and basin relief are both better predicted by stream length except in larger basins where relief ratio is better predicted by lineament length. Basin gradient is 113 Chapter 5 Drainage basin analysis better predicted by stream length except in the larger basins. It appears that while lineament length is not a better predictor than stream length of basin morphometry it is a reasonable predictor in many cases. The drainage network in a basin is developed in response to external factors, specifically climate. Although a suitable lineament will preferentially locate channels and the consequent drainage pattern is a function of the lineament network in the basin, channel network will always evolve regardless of bedrock structure and lithology. This is the primary reason why the lineament length of a basin may not be a good indicator of the morphometry of the basin. The presence of a developed stream network in the absence of lineaments is likely to bias correlations away from the lineament length and toward the stream length in places where suitable lineaments do not exist. On average only 66% of the stream network is lineament controlled the remaining amount probably has a significant affect on the strength of correlations. The stream network is a self contained, isolated network within the basin whereas the lineament network is not isolated at the drainage basin scale but persistent beyond the drainage divides and may only be isolated at the much larger scale regional scale. For example it is known that different structural patterns exist in the crustal blocks adjacent to that containing the study area (Journeay, pers. Comm. 1996). It is speculated here that small basins are likely to have their morphometry controlled by the stream pattern in the fashion classically recognized by geomorphologists. As the size of the basin increases it is more likely to be influenced by lineaments in the rock mass, hence in basins between one and approximately one hundred square kilometers, lineament length becomes more of a controlling factor and is often at least of equivalent importance as stream length. It is possible that if larger basins than these are considered (up to the scale of the major rivers) lineaments will be found to be even more important in controlling topography and morphometry. 114 Chapter 5 Drainage basin analysis Finally there is no apparent distinction in the data examined between igneous and metamorphic basins. This is true of all of the relations examined in this section suggesting that the areal and relief parameters of drainage basins in the study area are not significantly controlled by rock type at this scale. 5.8 Investigation into sediment yield from drainage basins One research aim was to determine whether morphometric parameters can be used to make inferences about the nature of landslide activity within the basin or if any correlations can be drawn between these morphometric parameters and sediment yield from the basin. The former is best judged from air photos. Recent debris flow and avalanche activity in a basin is apparent on these and has been noted for the sample set in Appendix III. It is recognized that not all debris flow or avalanche deposition occurs on the fan (Fannin and Rollerson 1990), some material may go into storage in the basin. However, over a long time scale much of this material should ultimately reach the fan. If sediment yield from the basin can be linked to basin or fan morphometry then only select parameters are required from a topographic map to estimate annual sediment yield. Schumm (1954) demonstrated that relief ratio could be used to estimate sediment yield. Ryder (1971a) suggested that the gradient of upper one fourth of the fan as an indicator of sediment yield from basins. Church et al. (1989) examined sediment yield from large basins in British Columbia (> 10 km2) and formulated the following expression relating basin area to sediment yield: Specific Sediment Yield (Mg km\"2 day\"1) = Basin Area (km 2 ) 0 6 115 Chapter 5 Drainage basin analysis This formula provides an indirect means of calculating specific sediment yield by manipulation of the value of basin area. For this reason it was decided to investigate the relationship between basin area and other morphometric parameters that might be related to sediment yield. No attempt was made to convert basin areas to sediment yield values by the above formula. The primary obstacle in attempting a study of sediment yield in this area is the lack of calibrating data. Three studies are known on sediment yield from drainage basins in southwest British Columbia These are: Church et al (1989); Owens and Slaymaker (1992); and Millard (1986). Owens and Slaymaker (1992) provide sediment yield estimates for three small alpine basins each of drainage area less than one square kilometer. Millard (1986) provides sediment yield estimate for gullies in the Coquitlam watershed, but his study is too limited to be widely applied over the sample set basins. Although the study performed by Church et al. (1989) was conducted in British Columbia evidence in the literature suggests that relationships between morphometric parameters may vary regionally, especially where climatic changes are important, and that relations may change in basins above and below one square kilometer. The latter point is consistent with the idea that at time scales of 10° to 104 years (i.e. since deglaciation) sediment yield may not accurately measure erosion and primary denudation rates in small basins but that in larger basins sediment yield will tend toward being a measure for denudation of the land surface (see Owens and Slaymaker 1993, p. 153 for a discussion). The basins examined in Church et al. (1989) study appear to be high order drainage basins, hence the basins examined in this thesis may be comparatively considered subdrainages. No basins higher than fourth order are included in the sample set. This suggests that the relationship identified by Church et al. (1989) may need modification for \"sub\"-drainage basins in the southwest Coast Mountains. Unfortunately there is a lack of sediment yield data with which to make these modifications. 116 Chapter 5 Drainage basin analysis Previous literature has identified three parameters that might reflect sediment yield from a basin: fan gradient (Ryder 1971a); basin area (Church et al. 1989); and relief ratio (Schumm 1954). Although Ryder (1971a) has pointed out that fan area is not a good indicator of sediment yield because of subsequent erosion it is considered here because an initial observation of this study was the large size of the fans building into the valleys in the study area. It was initially suggested that fan area might reflect the amount of sediment output from the basin. Additionally Bull (1964) has shown basin area and fan area to be related in California for basins developed over mudstone, shale and sandstone. 5.8.1 Regression analysis of morphometric parameters related to sediment yield To investigate sediment yield in the absence of sediment yield data it is necessary to accept one parameter as definitely related to the amount of sediment leaving the basin. This parameter is chosen to be the basin area since area must depend on the amount of material removed by primary erosion (the direct weathering of the bedrock) and because previous authors have correlated basin area and sediment yield (e.g., Church et al (1989)). In this examination basin area is compared to fan area, gradient, and relief ratio. If it can be demonstrated that these parameters relate well to basin area then it can be inferred that they should also be valid measures of sediment yield. This will provide an alternative to measurement of basin area for estimation of sediment yield. 5.8.2 Results of regression analysis This section is divided into five parts based on the investigated comparisons between the parameters described above. Regression data for each analysis is presented in Appendix VIII and graphical analysis is presented in figures 5.20 to 5.28. Table 5.11 shows the regression equations resulting from these analysis. 117 Chapter 5 Drainage basin analysis Regression equations resulting from sediment vield investigations Fan area versus Basin area All basins; Log (Fan area) = 0.39.Log (Basin area) + 2.96 (R2 = 0.58) Large basins; Log (Fan area) = 0.38.Log (Basin area) + 3.08 (R2 = 0.57) Small basins; Log (Fan area) = -0.18.Log (Basin area) + 6.6 (R2 = 0.03) Fan area versus lineament length All basins; Log (Fan area) = 0.45.Log (lineament length) + 3.65 (R2 = 0.62) Large basins: Log (Fan area) =1.11 Log (lineament length) - 1.94 (R2 = 0.58) . Small basins: Log (Fan area) = 0.38.Log (lineament length) + 1.41 (R2 = 0.07) Fan area versus stream length All basins; Log (Fan area) = 0.56.Log (stream length) + 3.29 (R2 = 0.48) Large basins Log (Fan area) = 1.05.Log (stream length) - 1.79 (R2 = 0.62) Small basins:: Log (Fan area) = -0.13.Log (stream length) + 4.02 (R2 = 0.04) Fan area versus fan gradient All basins; Log (Fan area) = -0.62.Log (Fan gradient) +5.01 (R2 = 0.46) Large basins; Log (Fan area) = -1.02.Log (Fan gradient) + 4.84 (R2 = 0.55) Small basins: Log (Fan area) = 0.19.Log (Fan gradient) - 1.36 (R2 = 0.06) Fan gradient versus basin area All basins; Log (Fan gradient) = -0.44.Log (Basin area) + 2.11 (R2 = 0.60) Large basins: Log (Fan gradient) = -0.46Log (Basin area) + 2.28 (R2 = 0.45) Small basins: Log (Fan gradient) = -0.67.Log (Basin area) + 5.44 (R2 = 0.29) Fan gradient versus lineament lensth All basins; Log (Fan gradient) = -0.44.Log (Lineament length) + 1.07 (R2 = 0.48) Large basins: Log (Fan gradient) = -0.73.Log (Lineament length) + 3.71 (R2 = 0.47) Small basins: Log (Fan gradient) = 0.22.Log (Lineament length) + 3.43 (R2 = 0.14) Fan gradient versus stream lensth All basins; Log (Fan gradient) = -0.62.Log (Stream length) + 1.73 (R2 = 0.61) Large basins: Log (Fan gradient) = -0.67.Log (Stream length) + 3.56 (R2 = 0.48) Small basins: Log (Fan gradient) = -0.49.Log (Stream length) + 2.16 (R2 = 0.35) Fan srdadient versus relief ratio All basins; Log (Fan gradient) = 0.1.04.Log (Relief ratio) - 0.42 (R2 = 0.58) Large basins: Log (Fan gradient) = 0.37.Log (Relief ratio) - 0.09 (R2 = 0.45) Small basins: Log (Fan gradient) = 0.24.Log (Relief ratio) + 0.53 (R2 = 0.35) Basin area versus relief ratio All basins; Log (Basin area) = -0.39.Log (Relief ratio) + 2.24 (R2 = 0.88) Large basins: Log (Basin area) = -0.35.Log (Relief ratio) + 1.95 (R2 = 0.86) Small basins: Log (Basin area) = -0.53.Log (Relief ratio) + 3.05 (R2 = 0.52) Table 5.11 Regression equations resulting from sediment yield investigations. 118 Chapter 5 Drainage basin analysis Fan area versus lineament densitv All basins; Log (Fan area) = -1.8.Log (Lineament density) -0.11 (R2 = 0.02) Large basins: Log (Fan area) = -OALog (Lineament density) - 0.18 (R2 = 0.43) Small basins: Log (Fan area) = 0.04.Log (Lineament denstiy) - 5.19 (R2 = 0.12) Fan area versus drainage densitv All basins; Log (Fan area) = -0.41 .Log (Drainage density) - 0.29 (R2 = 0.41) Large basins: Log (Fan area) = -0.46.Log (Drainage density) - 0.03 (R2 = 0.38) Small basins: Log (Fan area) = 0.04.Log (Drainage density) - 2.57 (R2 = 0.01) Fan area versus relief ratio All basins; Log (Fan area) = 0.13.Log (Relief ratio) + 2.66 (R2 = 0.42) Large basins: Log (Fan area) = 0.18.Log (Relief ratio) + 2.25. (R2 = 0.39) Small basins: Log (Fan area) = - 1.01.Log (Relief ratio) + 0.20 (R2 = 0.08) Fan gradient versus lineament densitv All basins; Log (Fan gradient) = -2.22.Log (Lineament density) + 0.28 (R2 = 0.14) Large basins: Log (Fan gradient) = -0.73.Log (Lineament density) + 3.71 (R2 = 0.47) Small basins: Log (Fan gradient) =..0.88 Log (Lineament denstiy) - 2.01 (R2 = 0.18) Fan gradient versus drainage densitv All basins; Log (Fan gradient) = -2.26.Log (Drainage density) + 0.38 (R2 = 0.43) Large basins: Log (Fan gradient) = -2.37.Log (Drainage density) + 0.31 (R2 = 0.31) Small basins: Log (Fan gradient) = 0.18.Log (Drainage density) - 2.29 (R2 = 0.06) Relief ratio versus lineament densitv All basins; Log (Basin relief) = 0.53.Log (Lineament density) - 2.3 (R2 = 0.29) Large basins: Log (Basin relief) = 0.74.Log (Lineament density) - 2.13 (R2 = 0.81) Small basins: Log (Basin relief) = 1.45.Log (Lineament density) - 2.4 (R2 = 0.40) Relief ratio versus drainage densitv All basins; Log (Basin Gradient) =1.13 .Log (Drainage density) + 2.64 (R2 = 0.79) Large basins: Log (Basin Gradient) = 0.96.Log (Drainage density) - 2.25 (R2 = 0.92) Small basins: Log (Basin Gradient) = 0.4.Log (Drainage density) - 2.37 (R2 = 0.27) Basin area versus lineament densitv All basins; Log (Basin area) = -0.19.Log (lineament density) -1.23 (R2 = 0.21) Large basins: Log (Basin area) = -0.28.Log (Lineament density) - 0.53 (R2 = 0.84) Small basins: Log (Basin area) = -0.81 .Log (Lineament density) - 2.24 (R2 = 0.23) Basin area versus drainage densitv All basins; Log (Basin area) = -0.3.Log (Drainage density) - 0.61 (R2 = 0.84) Large basins: Log (Basin area) = -0.35.Log (Drainage density) - 0.27 (R2 = 0.87) Small basins: Log (Basin area) = -0.12.Log (Drainage density) - 0.43 (R2 = 0.58) Table 5.11 (Continued) Regression equations resulting from sediment yield investigations. 119 Chapter 5 Drainage basin analysis Fan Area Vs. Basin Area (R 2 = 0.58) 1 E 7 , -1E4 --1000 -(—\\— i-H H-H+l- H-t+|-1E4 1E5 1E6 1E7 Basin Area (m 2 ) -H-H-H 1—i-H-mi I 1E8 1E9 Log(C.A.) = 0.39.Log(B.A.)+2.96 Figure 5.20. Plot of fan area versus basin area. Fan Gradient Vs . Basin Area (R 2 = 0.60) 10 c I km2: R 2 =0.77 < 1 km2: R 2 = 0.52 0.1 1 Relief Ratio Log(B.A.)=-2.25.Log(R.R.)+5.82 10 Figure 5.22. Plot o f basin area versus relief ratio. Fan Area Vs. Relief Ratio (R 2 = 0.42) 1E7 1E6 1E5 o3 < C HH 1E4 1000 > I km2: R 2 =0.39 < 1 km2: R 2 =0.08 H i- H 1 1 — I — ( -0.1 1 Relief Ratio 10 Log(F.A.)=-0.79.Log(R.R.) + 5.24 Figure 5.23. Plot o f fan area versus relief ratio. 121 Chapter 5 Drainage basin analysis Fan Area Vs. Fan Gradient (R 2 =0.46) 1E7-j 1E4 1000 \"I ' — -H 1— -H 1 1 i i i i I I | 1 — ( - H - m - H - l 0.001 0.01 0.1 1 10 Fan Gradient — Log(F.A.) = -0.62.Log(F.G.)+5.01 Figure 5.24. Plot of fan area versus fan gradient. 10 Fan Gradient Vs. Relief Ratio (R 2 = 0.58) c •3 O c IX, 0.1 0.01 -0.001 0.1 < I km2: R 2 =0.35 > 1 km2: R 1 =0.45 -f—i—i—i—t— 1 Relief Ratio 10 Log(C.G.)= 1.04.Log(RR)-0.42 Figure 5.25. Plot of fan gradient versus relief ratio. 122 Chapter 5 Drainage basin analysis a Fan Area Vs. Lineament length (R 2=0.62) 1 E 7 i 1E4 f 1 000 T 1 1 I I I ll+l 1 \\—f-H-H+| t l - H H H + l l 1 1 I I I I l-H 1 1 I I I III 10 100 1000 1E4 1E5 1E6 Lineament Length (m) Log(F.A.) = 0.45.Log(L.L.)+3.65 Fan Area Vs . Stream Length (R 2=0.48) 1E7-1E6-2 1E5-< I 1E4-1000 100 > 1 km2: R 2 = 0.62 < 1 km2: R 2 = 0.04 -i 1—I I I l+H 1 1—1~ -I—I I I 111 1—I—I I I 111 1000 1E4 1E5 Stream Length (m) 1E6 Log(F.A.) = 0.56.Log(S.L.)+3.29 Figure 5.26. Plots of fan area versus a) lineament length and b) stream length. I23 Chapter 5 Drainage basin analysis a Fan Gradient Vs. Lineament Length (R 2 = 0.48) c cd o [1 10 l 0.1 0.01 0.001 ^ — — • • n 10 H-4+H- H — M - H + H 1—i i i i i n | 100 1000 1E4 1E5 1E6 Lineament Length (m) •Log(F.G.) = -0.44.Log(L.L.)+1.07 Fan Gradient Vs. Stream Length ( R 2 = 0.61) 10 1 0.001 -I h-l-MH-H-H -t—I I I I I M I 1 I I I I H-H 1—< I 100 1000 1E4 1E5 1E6 Stream Length (m) — Log(F.G.) = -0.62.Log(S.L.)+l .73 Figure 5.27. Plots of fan gradient versus a) lineament length and b) stream length. 124 Chapter 5 Drainage basin analysis Relief Ratio Vs. Drainage Density 10 (R 2 =0.79) Figure 5.28. Plot of relief ratio versus drainage density. 125 Chapter 5 Drainage basin analysis 5.8.2.1 Fan area as a measure of sediment yield The hypothesis that fan area indicates sediment yield from a basin is based on an assumption that a constant proportion of the material eroded from the basin is ultimately stored in the fan and similarly a constant proportion of eroded material is transported beyond the fan. A constant proportion of material is also returned to storage within the basin. If this were true these values would be comparable across the region and a 100% regression correlation between basin and fan area would result. The relationship between fan area and basin area is shown in Figure 5.20. It is seen that larger basins tend to build larger fans. R2 = 0.58, however, this relation appears unreliable in the smaller basins where there is a significantly lower correlation and identification of a trend would be difficult. The apparent increased variability in the amount of material deposited on the fans by smaller basins may be due to variation of the slope gradients on which the fans are deposited. Two fans of the same area may have different volumes. For example cones #8 and 20 are highlighted in Figure 5.20. The area of each is comparable (176,000 and 167,000 m2 respectively) but basin area is very different. Basin #20 is twice the size of basin #8. Using the slope of the basin as an indicator of the gradient upon which the cones are built it is seen that cone #8 (slope 27°) rests at 37° and cone #20 (slope 22°) rests at 50°. For this reason in steep mountain watersheds fan area may not be a reliable indicator of sediment production. Larger fans building into flat bottomed valleys may be better for deducing relations since the depositional slope is less variable. Ideally the volume of each fan should be determined. 126 Chapter 5 Drainage basin analysis 5.8.2.2 Fan gradient as a measure of sediment yield There is a general increase in fan gradient with decreasing basin area (Fig. 5.21) Basin area and fan gradient correlate are moderately correlated (R2 = 0.60). This relation is best developed in the larger basins. The two largest basins (#19 and 29) are the greatest outliers and have very different fan gradients. These gradients are 0.0074 (0.42°) and 0.1517 (8.63°) respectively. Fan #29 is a large fan with multiple stream channels which are capable of reworking material and lowering fan gradient. Fan #19 has a single channel building from a narrow valley. Both are fourth order basins with similar basin slopes (8° and 9° respectively). Ryder (1971a) suggested that only the top one fourth of the fan gradient may accurately predict sediment yield. There is insufficient accuracy on the topographic maps chosen for this study to better evaluate fan gradient. 5.8.2.3 Relief ratio as a measure of sediment yield Basin area and relief ratio show excellent overall correlation with relief ratio increasing as basin area decreases. R 2 = 0.87 for the entire data set (Figure 5.22). This suggests that relief ratio is a good predictor of sediment output from both large and small basins and is consistent with the observations of Schumm (1954). There are no significant outliers in the data. 5.8.2.4 Relations between fan area, fan gradient, and relief ratio Al l relations between fan area (R2 = 0.0.42 for the entire dataset), fan gradient (R2 = 0.58 for the entire dataset) and relief ratio are poor reinforcing the idea that fan area is not a good indicator of sediment yield (Figures 5.23 and 5.24). Figure 5.25 shows fan gradient vs. relief ratio (R2 = 0.58 for the entire data set). The graph indicates that as relief ratio decreases, in larger basins, the fan gradient also decreases, consistent with the view that large basins build gentle fans. 127 Chapter 5 Drainage basin analysis 5.8.2.5 Lineament length and stream length as predictors of sediment yield Lineament and stream length were examined in comparison with fan area and gradient. Fan area shows moderate correlation with lineament length (R2 = 0.62, for the entire data set) but the relation is weaker with stream length (R2 = 0.48, for the entire data set) as shown in Figure 5.26. If fan area were a reliable indicator of sediment yield a good relation would be expected since both parameters correlate extremely well with basin area (R2 = 0.83 and 0.97 for the entire data set, for lineament length and stream length respectively). However, fan area appears the least successful indicator of sediment yield except at a gross scale (a large fan generally indicates large sediment output from the basin). It is therefore difficult to explain this high correlation. It may be that the larger the stream length in a basin the more power the major stream has to remove material from the basin to the fan. Additionally larger stream channels on the fan are likely to be less susceptible to avulsion, characteristic of small, and sometimes debris flow dominated fans. Lineament length is highly correlated with the stream length (R2 = 0.82 for all basins and 0.97 in the larger basins, see section 5.7.2.2) largely due to the extent to which the stream network occupies the lineament pattern in a basin. It is likely that a larger lineament network provides more localities for erosion. This is consistent with the hypothesis that lineaments are zones of weakness in the rock mass which provide sites for erosive and wasting processes such as freeze-thaw and rockfall (see Section 4.4) and thus increase the rate of primary erosion in a basin. Hence the larger the lineament network the larger the stream network may be and the more opportunities that may exist for sediment to be introduced to the stream network. Although fan area is thought not to be a good sediment yield indicator it does seem to correlate well with lineament length and reasonably so with stream length. Lineament and stream length are compared to fan gradient in Figure 5.27. Stream length is the stronger correlating parameter because a larger developed stream network in a basin is likely to 128 Chapter 5 Drainage basin analysis develop a larger, alluvially dominated, low gradient fan. In the smaller basins processes other than alluvial may be responsible for deposition on the fan. In particular the fan may be dominated by debris flow activity and such fans have been shown to be generally smaller and steeper than alluvially dominated fans (for example Kochel 1990, Stanistreet and McCarthy 1993). The initially surprising relation is that fan gradient decreases as lineament length increases. This obscures the fact that although there is a greater potential for erosion and possibly more potential for debris flow and avalanche activity in small basins (see Chapter 6) there is also less likelihood of these reaching the fan and being responsible for primary deposition on the fan, which will be alluvially dominated. Fan area; fan gradient; relief ratio; and basin area, were examined in relation to both lineament and drainage density, only the relation between relief ratio and drainage density showed a strong correlation (R2 = 0.92 in the large basins, Fig. 5.27). This is consistent with the findings of Schumm (1954). In general the drainage density is highest in basins with a higher relief ratio. 5.8.3 Discussion This analysis was based on the assumption that sediment yield is related to basin area. In the absence of available sediment yield data for the sample set basins the parameters of fan area, fan gradient, and relief ratio were compared to basin area. Additionally described were the relations between these parameters and stream and lineament length. A good correlation with basin area would suggest that a parameter provides an alternative to measuring basin area for estimation of sediment yield. If the hypothesis that basin area is a predictor of sediment yield is accepted then the following conclusions can be inferred from the analysis described in this section: 1) Fan area is not a good predictor of sediment yield; 2) Fan Volume might prove a better predictor; 129 Chapter 5 Drainage basin analysis 3) Fan gradient is a slightly better predictor of basin area than fan area and may be improved if the assertion of Ryder (1971a) can be investigated; 4) Relief ratio appears to be the best predictor of basin area and by inference sediment yield (R2 = 0.86 for all basins); 5) Sediment yield is perhaps best predicted by relief ratio for basins larger than 1 km2 (R2 = 0.88). Lithology may have implications for the sediment yield from a basin (for example see Schumm 1954) however since almost half of the larger basins are metamorphic there is insufficient data to separate lithologic effects in this analysis. Additionally these analysis do not permit identification of factors which would identify a basin's potential for increased sediment yield due to debris flow or avalanche activity. They also do not permit identification of the amount of each parameter due to contemporary process (as opposed to early post glacial increased process rates). For this reason for accurate prediction of sediment yield relations must be calibrated for the present with field measures of sediment yield as attempted in the study by Church et al. (1989). 5.9 Summary and conclusions Analyses were conducted into the effect of lineaments on drainage basin position, the drainage pattern, basin morphometry and sediment yield. A sample set of 25 basins was used for these investigations. In 22 of the sample set basins the position of the basin axis was found to be lineament controlled. This result is interpreted as indicating that during initial development of the drainage basin a lineament provided the necessary topographic depression for concentration of surface water flow into a master rill and subsequently a main stream channel. These have deeply 130 Chapter 5 Drainage basin analysis incised the lineament in question forming a basin. This has been facilitated by the fact that lineaments are zones of weakness in the rock mass. A spatial correlation was performed in IDRISI to determine the extent to which the drainage pattern mirrored the lineament pattern in a basin. An average 66% of streams are found to overlay lineaments in the sample set basins. These values were respectively 71% in igneous basins and 46% in metamorphic basins. Additionally the directional correlation between lineament and stream segment trends was investigated. Preferred trends were found for stream segments in each subgroup (igneous and metamorphic) of basins and streams occupying lineaments have a preferred trend of 22 ± 36°. This value is close to the preferred lineament trend in the sample set basins 11 ± 17°. Both trends approximate the orientation of the Tertiary structural trend developed in the region. It was found that the preferred trend of third order stream segments (172 ± 26°) trends agree well with the dominant lineament trend consistent with the idea that lineaments control basin axis position. It is speculated that low order streams are concentrated primarily in the younger lineaments since major drainages probably initiated along the oldest lineaments (Cretaceous, northwest trending lineaments) for example, the major rivers in the area. It is concluded from these investigations that lineaments strongly influence the stream network developed in the study area. The reason for this is that lineaments represent a network of weakened zones emplaced in the rock mass by tectonic stresses operative since the Cretaceous. Water has provided the forces necessary to exploit this network. Regression analysis of allometric relations between morphometric parameters showed that while stream length is often the better predictor of morphometric parameters, the lineament length is also a good predictor. This is especially true in basins larger than one square kilometer in area. Investigation was conducted separately into the entire data set and then two subsets: large and small basins. This was based on the observation of Owens and Slaymaker (1992) that sediment yield 131 Chapter 5 Drainage basin analysis characteristics of basins larger and smaller than this area are different. The evidence of this study seems to support this in that in many cases relations are poor in the small basins while in larger basins good correlations are observed. The relations between basin morphometry and sediment yield were reported. In the absence of suitable sediment yield data for the study area it was necessary to assume that the parameter of basin area was related to sediment yield. This assumption is supported by the relation established by Church et al. (1989), however it is likely that this relation would need to be refined with sediment yield data taken from within the study area. This is because it has been shown in the literature that relations vary regionally especially where climate is variable. It appears from this investigation that relief ratio is a good predictor of basin area. This is true for both the large and small basins and appears consistent with the idea that in small basins sediment yield increases with lower elevation (Owens and Slaymaker 1992). Fan gradient is also a reasonable predictor of basin area but fan area proved inappropriate. The area of the fan was initially believed representative of the volume of material leaving the basin. Fan area is well predicted by both lineament length and stream length as is basin area. Examination of lineament effects on drainage basins leads to the proposal of the following hypothesis: The lineament network is not isolated at the scale of individual (1st through 4th order basins) but may be at the regional scale. Because small basins may not be large enough to include a significant proportion of the regional lineament network it is likely that streams play a much more important role in drainage basin development than lineaments at this scale. When basins with areas between one and one hundred square kilometers are examined it is found that lineaments and streams reveal approximately equivalent correlation strengths when compared to other morphometric parameters of the drainage basin. Significantly at this scale much more of the regional lineament pattern is included in the basins. 132 Chapter 5 Drainage basin analysis Finally, although not investigated, it appears that at the regional scale the lineament network may be more responsible for topography than the drainage network in the southwest Coast Mountains. For example, the trends of the major rivers seen in Figure 3.2 follows the northwest lineament trend. Convenient boundaries for the regional scale exist locally in the form of a number of joined, but distinct crustal blocks. Additionally, it has been recognized that \"strain concentration along a series of strain-weakening zones in an orogen produces the first harmonic pattern of ridges and valleys in the complete absence of erosion\" (Koons 1995, p. 399). It is inferred that the major valleys in the study area are developed along major lineaments. Although these are not visible on air photo because of the dominance of surflcial materials, it is likely that mapping from higher altitude imagery would reveal these as lineaments on the basis of the nearly straight alignment of the river network. It is concluded that the proposed hypothesis is consistent with the latest ideas on the topographic evolution of collisional belts. The evidence suggests that lineaments provide a focus for the development of the drainage network. As well as being zones of weakness in which water will preferentially erode they appear to have provided topographic low points for initiation of drainage basins. It is possible that lineaments quickly captured the master rills that will have developed on slopes from the divide soon after these slopes formed by tectonic means. The paths of the major drainages in the study area follow the regions oldest developed structural trend. It is also the pattern predicted by Koons (1995) of a ridge valley system parallel to the dominant fault and orogen strike. It seems clear that the drainage pattern and the consequent basin forms have developed as a result of the structural patterns tectonically emplaced on the landscape of the southwest Coast Mountains. 133 Chapter 6 Lineaments and landslides CHAPTER 6. LINEAMENTS AND LANDSLIDES 6.1 Introduction Savigny (1996) and Leir (1995) have demonstrated a relationship between lineaments, large rock landslides, and mountain slope deformation in southwestern British Columbia. Chapter five has shown that relationships exist between lineaments and contemporary geomorphic processes. One such process may be small surficial landslides (specifically debris flows and debris avalanches) which are ubiquitous in the study area. This chapter examines these relationships further by presentation and investigation of two landslide inventories; Large rock landslides were identified on 1:60,000 scale air photos as part of this work. Small surficial landslides were identified in an area of 132 km2 located in the southern portion of the study area by a study commissioned by the GVRD and conducted by Dr. J.M. Ryder of J.M. Ryder and Associates, Vancouver, B.C. This chapter begins with presentation of the large rock landslide and mountain slope deformation inventory and description of key sites then reports the investigation of small surficial landslides. Conclusions are presented about the nature of landslide activity in the study area along with a comparison of findings with those in the neighboring Fraser Valley where similar studies have previously been conducted, e.g., Savigny (1996) and Leir (1995). 134 Chapter 6 Lineaments and landslides 6.1.1 Large rock landslides and mountain slope deformation It is important to study landslides and mountain slope deformation in the region because it is known that lineaments play a major role in the evolution of landslides in the adjacent Fraser River valley (Savigny 1996, Leir 1995). Lineaments are the primary focus of this thesis and it is desirable to know if similar relations operate in this study area. The primary benefit of a landslide inventory for this area is that it allows the assessment of hazard distribution in the area which has potential implications for regional hazard zonation. For example, areas where slope deformation is suspected may yield future failures. In conjunction with lineament mapping, described in Chapter 4, each air photo was investigated for landslide activity and slope deformation. These were recorded in an inventory style complementing the Fraser River valley study by Dr. K.W. Savigny (Savigny 1996), only failed slopes and slopes exhibiting evidence of gravity induced deformation are included. 6.2 Airphoto identification of landslides and mountain slope deformation The air photo identification of landslides is described in Leir (1995). In this study air photos were examined for landslide scars, and rubble filling valley bottoms. Slope deformation is harder to identify. Uphill facing (antislope) scarps are characteristic of this type of deformation. These scarps have been described as \"...typically between 1 and 6 m high...\" and being \"...located downslope of anomalous ridge-top troughs\" (Bovis 1982, p. 804). These troughs are thought to represent either infilled tension cracks, small graben like structures, or small faults. Tension cracks and downslope scarps are also common. Scarps will typically trend either parallel to the ridge crest or transverse to a mountain flank. Additionally some slopes may exhibit bulging near the toe of the slope. This again indicatives slope deformation. 135 Chapter 6 Lineaments and landslides 6.3 The landslide and slope deformation inventory Landslides were inventoried in the area extending from approximately 122° W. to 123° 25' W. and from 49° N . to 50° N . an area of approximately 10,900 km2. Rock avalanches were the features most commonly identified in the inventory along with mountain slope deformation. Further large landslide types include landslides in surficial deposits adjacent to major rivers and large rockfalls. Twenty sites were identified in the inventory, the site density is therefore 0.0018 sites/km2 if all features are considered. Considering only bedrock features the landslide density is 0.0017 landslides/km2. Further, considering only bedrock failures the density is 0.001 landslides/km2. All landslides are assumed to have occurred postglacially, i.e., in the last 10,000 years because glaciation would have removed evidence of pre-glacial landslide activity. Assuming equal temporal distribution of bedrock failure one landslide will have occurred every 910 years however this figure does not take into account repeated failures from a single site or the possibility that a single triggering mechanism such as an earthquake might have simultaneously induce a number of failures. This has been shown to be the case with a number of rock avalanches in neighboring Washington State (Schuster et al. 1992). Failures in Quaternary sediments are more common often with many small failures producing a single large failure zone. The twenty sites highlighted in the study area are shown in Figure 6.1. Six of these have been previously reported and are only briefly mentioned here with references for further investigation. These are: Site #l:Dickson Lake landslide (Evans 1986); Site #3: lower Coquitlam River (Thurber Engineering Ltd 1985); Site #14:Goat Ridge (currently monitored by Thurber Engineering Ltd., Vancouver); Site #18: Cheekye Ridge (Thurber Engineering Ltd. and Golder 136 Chapter 6 Lineaments and landslides 123 25 V 122 V 50 N 50 N 10 Kllemeters 0 10 20 30 40 49 N 0 / 123 25 W 49 N 122 W Figure 6.1. Showing sites identified in the landslide and slope deformation inventory: l=Dickson Lake; 2=West Norrish Creek; 3=Lo\\ver Coquitlam River; 4=Lower Seymour Valley; 5=The Lions; 6=Chehalis River; 7=Bivouac Mountain; 8=Anne Lake; 9=Stave River I; 10=Stave River II; ll=Mount Bonnycastle; 12=Winslow Lake; 13=Winslow Creek; 14=Goat Ridge; 15=Mamquam River; 16=Shale Creek; 17=Bremner Creek; 18=Cheekye Ridge; 19=Rubble Creek; 20=Mount Mason. 137 Chapter 6 Lineaments and landslides Of the remaining fourteen sites, three represent prehistoric, and now well vegetated, landslides. Six appear as fresh rock avalanche or rockfall debris, four sites exhibit evidence of rock mass deformation and one is the site of slides in surficial materials adjacent to a river. The inventory is presented in Table 6.1. Exact locations and air photo numbers are shown in this table along with a brief description of the feature. Of the sites previously reported all but one (Site #20: Mount Mason) is near human development. While logging is ubiquitous throughout the area urban growth and transportation networks are restricted to the western and southern margins of the region. Accordingly known landslides lie mostly in the these areas. One site (the Cheekye, #18) was discovered by development investigations. 6.3.1 Examples of lineament control on large rock landslides and mountain slope deformation Discussion of five of the sites from the inventory is presented below. These sites were chosen to best illustrate the relationship between lineaments and landslides and mountain slope deformation. They also illustrate the variety of ways in which lineament control on large rock landslides and mountain slope deformation is manifested. Appendix IX provides additional discussion of some of the sites. Dickson Lake (site #1, Fig. 6.1) is the site of a landslide dam caused by a rock avalanche reported in Evans (1986) and Clague and Evans (1994). This slide is shown in Figure 6.2. The area of the slide is 971,280 m 2 and the estimated volume is 25.59xl06 m3 (Evans 1986). The slide occurred in granodiorite of the Coast Plutonic Complex and has not been accurately dated. The material forming the dam is large, blocky, and has been only partially revegetated, the main scarp is clearly visible. A discussion of the stability and draining characteristics of landslide dams in different materials are found in Evans (1986) and Clague and Evans (1994). 138 Chapter 6 Lineaments and landslides i -3 5 •a s *5.s „ .3 § S u 3-2 ^, ~ >*.ts oo •a u S oo = oo •= .S \" .s .s b U o g .o >< c 5 8 3 | in « -° rS 3 «, 3 w £ \"5 K S3 2 'C . -P C U ^ B--2 E o \\3 > s -5 •- a Jc S > P « w ^ t« -o o 3 5 s£ at £ o .2 8 :•§ •= S » » S o * 2 I -I s s CQ U d> a 3 -S - 5 2 T 3 CQ at 0 0 w - i 00 3 * 03 5t 0 0 CN is <=> 00 5t ^ a 11 00 CC as MS DO 5t 140 Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides This landslide and the adjacent site at West Norrish Creek (discussed below) show clear evidence of lineament association. The headscarp of the Dickson lake slide appears to be lineament controlled with the landslide being located at the intersection of a north-northwesterly trending and a northeast trending lineament. The northeast trending lineament is persistent to the West Norrish creek site where it appears to be influencing slope deformation. The proximity of these two features indicates that landslides and mountain slope deformations can occur in the same area and that mass movement features are not always isolated events. Additionally this site demonstrates the potential of following lineaments to locate landslides. This technique yielded good results during inventory mapping. Having identified the Dickson Lake landslide the two lineaments contributing to its headscarp were traced and the second site discovered. To clarify the air photo interpretation of the stereophoto in Figure 6.2, an additional map of the features is included in Figure 6.3. The West Norrish Creek site (site #2) occurs 1-2 km west of the Dickson Lake slide. Two prominent antislope scarps are seen on the southwest face of the peak in question. The lower scarp is contiguous with the northeast trending lineament described above. It is possible that the preexistence of this lineament was responsible for focusing of slope deformation at the site. This lineament forms approximately half of the headscarp of the Dickson lake landslide. Further evidence of deformation was found at the site. Figure 6.4 shows the side of the mountain with the major lineament indicated. Close examination revealed evidence of extension including the presence of open joints and apparently down faulted blocks (Figure 6.5). This slope is deforming toward the southeast but appears presently inactive. Site #5 is on the slopes below The Lions in the Capilano watershed. The southern flank of the eastern Lion is cut by three antislope scarps bisecting the ridge and trending parallel to the main 142 Chapter 6 Lineaments and landslides Creek 1 km approx. Figure 6.3. Interpretation of Dickson Lake landslide and West Norrish Creek mountain slope deformation. Note: traced from air photo 15BCB87098-156. 143 Chapter 6 Lineaments and landslides Figure 6.4. The southern slope of the west Norrish Creek site showing the major lineament described in the text. The location of Figure 6.5 is indicated. 144 Chapter 6 Lineaments and landslides Figure 6.5. An opening (extension joint) along the course of the major lineament at the West Norrish Creek site. 145 Chapter 6 Lineaments and landslides ridge (Fig. 6.6). They also form a part of a more persistent lineament seen trending eastwards. It appears that the slope is peeling away in a toppling motion. A possible interpretation is shown in Figure 6.7. The rear scarp is actively shedding sediment backward onto the slope although whether this suggests current deformation or is simply due to increased exposure due to opening is unknown. Each of the other two antislope scarps is shedding some material to the sides of the ridge but not in the quantities of the third. It may be significant that the uppermost scarp appears active since this would be the most recently opened. These lineaments also control headward erosion of first order streams. These accelerate removal of material from the lineament zone and may be affecting the stability of the feature by increasing hydrostatic pressure in the lineament zones. Site #7 (Fig. 6.8) also occurs within the GVRD watersheds. Bivouac mountain appears to be the site of a prehistoric landslide that may once have dammed the Seymour River. The headscarp appears to be still shedding material to the slope below. A portion of the base of the slide is still vegetation free. Seymour River dramatically narrows as it passes through the debris and significantly alters its course upon reaching the debris limits. Of significance are a number of lineaments shown above the headwall and the presence of some antislope scarps. Goat Ridge above Britannia Creek (Site #14) is currently being monitored by Thurber Engineering Ltd. of Vancouver (Hungr pers. comm. 1996). A large number of lineaments cross the ridgetop and some antislope scarps are visible on the slopes above the confluence of Britannia Creek and the tributary from Sky Pilot Mountain (Fig. 6.9). Site #15, on the Mamquam River northeast of Squamish is a large-scale feature approximately 2 km of the hillslope appears to be slumping into the valley (Fig. 6.10). After discovery of this site it became the subject of a directed studies project by Mr. Phil Scalia at the Geological Sciences Department, University of British Columbia, Vancouver, Canada, (Scalia 1995) 146 Chapter 6 Lineaments and landslides 147 Chapter 6 Lineaments and landslides Figure 6.7. Possible interpretation of mountain slope deformation on the southern face of The Lions in the Capilano watershed. The current state of activity at the site is unknown. 148 Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides 151 Chapter 6 Lineaments and landslides under the supervision of Dr. K.W. Savigny. Mr. Scalia concluded that the slope had moved by mass rock creep as defined by Chigara (1992). Figures 6.11a and 6.11b are two of Mr. Scalia's photographs showing an antislope scarp and graben, and high angle jointing in the rock mass, both features typical of this kind of deformation. Site #18 is the Cheekye Ridge. The headwaters of Cheekye River are sourced in a large amphitheater-shaped excavation into the western flank of Mount Garibaldi (Figure 6.12). This large feature represents a massive landslide scar created by multiple failure events (Evans 1991). Cheekye fan at the confluence of the Cheekye and Squamish rivers is known to be the deposition site for the rubble resulting from the collapse of the mountain flank in a series of volcanic debris avalanches (Evans 1991, Evans and Savigny 1994). The potential for further movement appears to exist, and this possibility has been investigated in a joint report by Thurber Engineering Ltd. and Golder Associates Ltd. (1993). Modeling of potential run out from a future failure has been conducted by Hungr (1995). Several features are highlighted in Figure 6.12. Attention is drawn to the northern front of the ridge which shows a feature that may represent slumping. Scarps reveal where sagging may be occurring. There is no obvious accommodation of this feature near the toe of the slope, however, in most cases early movement is accommodated by compression and bulging in the rock mass (Hutchinson 1988). Additionally a cracks visible near the ridge top and are indicated by the arrows. 6.3.2 Discussion Two of the twenty sites examined in the inventory are recognized for their potential hazard by the geotechnical community in Vancouver. These are the Cheekye Ridge and Goat Ridge sites. It is sometimes believed that at sites where failure has previously occurred the risk of further failures is 152 Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides reduced. This is not necessarily the case on slopes where further deformation features such as tension cracks and antislope scarps are visible behind or near the existing headscarp. Such evidence is traditionally recognized as representing the possibility of further retrogressive failure at a site. Sites where evidence of slope deformation is found are particularly difficult to assess on air photos. Bovis (1982) cites local factors such as structure, lithology, and seismicity as being controls on the amount of deformation that a slope will experience. He also reports that \"...many antislope scarps do not appear to be actively evolving at present\" (Bovis 1982, p. 804), however such interpretation is difficult from air photo and it is likely that many sites in the southwest Coast Mountains have not received ground investigation. It is suggested that the removal of glacial ice reduces support for the valley walls allowing relaxation to occur post deglaciation. Apparent inactivity may imply that the slopes have reatained an equilibrium condition however while the initial disequilibriating effects of deglaciation may have been naturally remediated it is likely that cyclic reactivation of these features occurs by seismic shaking (Savigny, pers. comm. 1996) There is obvious relation between lineaments and landslides at several sites in the area, most notably the headwall of the Dickson Lake slide. One of the two lineaments that converge at this locality is traceable to the area of deformation at the West Norrish Creek site. The antislope scarps below the Lions are also a part of a major lineament traceable for some distance beyond the affected slope. A lineament is also found in the headscarp of the small rock avalanche in the Seymour watershed. It seems that lineaments can both cause and result from slope deformation and landsliding. In the case of Dickson lake it is suggested that the lineaments have controlled the headscarp position. This is similar to findings in the Fraser Valley where lineaments have also been shown to control 155 Chapter 6 Lineaments and landslides failure position (see Savigny 1992, and Leir 1995). At the west Norrish creek site the lineament seems to be the focus about which deformation is occurring. If the lineament represents a fault, or a large joint plane at the site of the Dickson Lake slide it represents a plane of weakness subsequently occupied by the headscarp. At the West Norrish Creek site gravity induced extension is apparently focused about the lineament. Gravitational displacement alone can account for some lineaments such as those in the wall above Winslow Creek where antislope scarps are mapped as small lineaments. Also at the Cheekye Ridge and Britannia Creek sites ridgetop cracks and antislope scarps have produced mappable lineaments. At Rubble Creek the columnar jointing provided a headwall for the volcanic debris avalanche. These features are joints not apparent on 1:60;000 scale air photos but which do represent small lineaments which would be detectable in the field. 6.4 Investigation into lineament control on small surficial landslides It has been recognized (Seidl and Deitrich 1992) that debris flows and debris avalanches (defined earlier as small surficial landlsides) are a primary erosive agent in steep mountain stream channels. It has also been pointed out (Davies et al. 1992) that debris flows are typically associated with unstable gully sidewalls, and that these may represent fault crush zones. It has been demonstrated in the previous section that large rock landslides and slope deformation are commonly associated with lineaments. Results of the following investigation indicate that there may be a relation between lineaments and small surficial landslides. Many debris flows and avalanches in the southwest Coast Mountains occur at the impermeable interface between the lodgment and basal layers in the vashon till. This surface also presents a limit to root penetration. Some events also occur at the till/bedrock interface. 156 Chapter 6 Lineaments and landslides While lineaments clearly influence bedrock failure, their relation to surficial landslides is less obvious. Small surficial landslides may be related to lineaments for the following reasons: 1) Lineaments form gullies in which surface and subsurface water can concentrate; 2) Bedrock weakness in lineament zones often leads to oversteepened gully walls; 3) Small bedrock failures caused by bedrock weakness often occur on oversteepened gully walls and can trigger debris flows. Once initiated, small surficial landslides move preferentially into the nearest gully. Here, they either continue to flow as a torrent or exhaust their momentum and fill the gully with debris which remains in situ until flushed by a subsequent event or re-mobilized by water. Often such an event will dam a creek and subsequently become re-mobilized under a larger hydraulic load. Because debris flows are a primary erosive force they are themselves responsible for exposing deeper portions of the lineament and steepening gully sidewalls, thus the process of gully erosion by means of landsliding is likely to periodically repeat itself. 6.4.1 Method of investigating the correlation between lineaments and small surficial landslides An inventory of surficial landslides was compiled for the GVRD by J.M. Ryder and Associates, Vancouver, BC, as a part of the watershed ecological inventory currently underway in the three GVRD watersheds. A total of 1203 landslides were mapped on air photos from 1957 onward. A DBASE (database software from Borland) catalogue of slide attributes contains information on the terrain polygons (mapped according to the terrain classification system for British Columbia, MOE Manual 10 1988) in which the landslides have initiated, and on the aspect of each polygon. Figure 6.13 shows the Seymour watershed with streams and landslides, geographic localities mentioned in the text are indicated. Figure 6.14 shows lineaments added to this map. The stream network was supplied by the GVRD. 157 Chapter 6 Lineaments and landslides Loch Lomond Dam Figure 6.13. The Seymour watershed showing streams and landslide initiation points (from images supplied by the GVRD). 158 Chapter 6 Lineaments and landslides Loch Figure 6.14. The Seymour watershed showing streams, landslide initiation points and lineaments. 159 Chapter 6 Lineaments and landslides The spatial correlation between lineaments and landslide initiation points, and streams and landslide initiation points, was investigated in the same manner as described in Chapter 5. An investigation of lineament control on streams in the Seymour watershed was also conducted and the results presented in Appendix X. Stream order was not considered in this investigation. Lineaments used were those mapped in the initial 1;60,000 scale air photo inventory and hence include only the major features evident in the watershed. Lineaments were not re-mapped at 1:20,000 scale because the threefold increase in the expected number of lineaments (Section 4.3.2) would dramatically increase the area of buffer zones around lineaments and might obscure any reasonable assessment of relations by including too much of the land area in these zones. Use of the 1:60,000 scale lineaments, however, did present a problem in that the mapping accuracy of the two datasets is significantly different. Lineaments were mapped to within ± 259 m (see section 4.3.1.) which theoretically results in a positional uncertainty of over half a kilometer. A comparison was made between lineament position in the IDRISI image used for analysis, and the original air photo mapping. The maximum positional error observed was about 150 m: most features were accurately placed. Particularly accurate were features located near major streams. Mapping accuracy of the GVRD inventory is assumed to be within ± 10 m. Stream data is also generated at 1 ;20,000 scale. Airphoto mapping was completed at 1:20,000 scale and assuming errors similar to those in this study positional error of approximately ± 10 m is likely. A 20 m pixel resolution was chosen to investigate the Seymour watershed. The 140 m resolution of the original lineament data is too coarse for the landslide features examined and a 10 m resolution is too fine for display in IDRISI for this size basin. Additionally this scale is comparable to that used to investigate sample set basins. All data supplied by GVRD was converted to IDRISI image and vector files and IDRISI's spatial analysis capabilities were used for investigation. When incorporated in the IDRISI database 160 Chapter 6 Lineaments and landslides 1199 pixels were identified as landslide initiation points indicating that 4 small surficial landslides occurred at, or very close to preexisting initiation points (within 20m) and hence could not be distinguished independently. 21 landslide initiation points occurred outside of the boundary of the Seymour watershed as digitized by the author for this study. Al l were very close to the boundary and their omission is due to a small error in digitizing the basin outline. The landslide omissions represent 2 % of the data and are likely to be insignificant in the results. Figure 6.15 shows the Jamieson Creek landslide in the Seymour Watershed: an example of the type of small surficial landslide included in the GVRD inventory. The portion of the Seymour watershed examined extends northward from approximately Seymour Falls dam. The IDRISI calculated area is 131.53 km2, the lineament and stream density are respectively 1.417 (km/km2) and 3.583 (km/km2). The sole purpose of the investigation was to determine if there was any significant structural control on landslide initiation points. No other landscape attributes were included in the analysis except for the stream network which is believed to be in part related to bedrock structure. 6.4.2 Results The results of analysis do not clearly indicate a relation between either lineaments and landslide initiation points or between streams and landslide initiation points. However, the probability of a small surficial landslide occurring is increased when the two features are coincident in an area. Table 6.2 shows the results of spatial correlation between lineaments and landslide initiation points. Table 6.3 shows the results for streams and landslide initiation points. Figure 6.16 shows a graphical representation of the results. The largest number of landslides were initiated between 20 and 40 m from each of the investigated features (lineaments and streams). 161 Chapter 6 Lineaments and landslides Figure 6. 15. The Jamieson Creek landslide, an example of the type of feature included in the GVRD landslide inventory of the Seymour watershed. 162 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 0-20 9322 2.83% 2.83% 36 3.06% 36 3.06% 20-40 26619 8.10% 10.93% 108 9.17% 144 12.23% 40-60 25850 7.86% 18.79% 106 9.00% 250 21.23% 60-80 17514 5.33% 24.11% 71 6.03% 321 27.25% 80-100 16959 5.16% 29.27% 57 4.84% 378 32.09% 100-120 23481 7.14% 36.41% 87 7.39% 465 39.48% 120-140 15361 4.67% 41.08% 56 4.75% 521 44.23% 140-160 17189 5.23% 46.31% 70 5.94% 591 50.17% 160-180 16590 5.05% 51.36% 68 5.77% 659 55.95% 180-200 13723 4.17% 55.53% 54 4.58% 713 60.53% 200-220 12202 3.71% 59.24% 54 4.58% 767 65.11% 220-240 10589 3.22% 62.46% 40 3.40% 807 68.51% 240-260 11689 3.55% 66.01% 38 3.23% 845 71.74% 260-280 11258 3.42% 69.44% 41 3.48% 886 75.22% 280-300 8977 2.73% 72.17% 34 2.89% 920 78.10% 300-320 7890 2.40% 74.57% 23 1.95% 943 80.05% 320-340 7573 2.30% 76.87% 32 2.72% 975 82.77% 340-360 7316 2.22% 79.10% 20 1.70% 995 84.47% 360-380 6678 2.03% 81.13% 21 1.78% 1016 86.25% 380-400 5985 1.82% 82.95% 20 1.70% 1036 87.95% 400-420 5205 1.58% 84.53% 18 1.53% 1054 89.48% 420-440 5630 1.71% 86.24% 14 1.19% 1068 90.67% 440-460 4332 1.32% 87.56% 12 1.02% 1080 91.68% 460-480 3959 1.20% 88.76% 13 1.10% 1093 92.79% 480-500 3779 1.15% 89.91% 13 1.10% 1106 93.89% 500-520 3661 1.11% 91.03% 12 1.02% 1118 94.91% 520-540 3106 0.94% 91.97% 9 0.76% 1127 95.67% 540-560 2464 0.75% 92.72% 7 0.59% 1134 96.27% 560-580 2263 0.69% 93.41% 4 0.34% 1138 96.61% 580-600 2361 0.72% 94.13% 6 0.51% 1144 97.12% 600-620 1885 0.57% 94.70% 5 0.42% 1149 97.54% 620-640 1874 0.57% 95.27% 3 0.25% 1152 97.80% 640-660 1476 0.45% 95.72% 3 0.25% 1155 98.05% 660-680 1455 0.44% 96.16% 5 0.42% 1160 98.48% 680-700 1289 0.39% 96.55% 1 0.08% 1161 98.56% 700-720 1089 0.33% 96.88% 5 0.42% 1166 98.99% 720-740 1143 0.35% 97.23% 1 0.08% 1167 99.07% 740-760 969 0.29% 97.53% 1 0.08% 1168 99.16% Table 6.2. Results of overlay of landslide initiation points on 20 m buffer zones around lineament pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2. 163 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative Zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 760-780 1072 0.33% 97.85% 4 0.34% 1172 99.49% 780-800 796 0.24% 98.09% 3 0.25% 1175 99.75% 800-820 617 0.19% 98.28% 0 0.00% 1175 99.75% 820-840 607 0.18% 98.47% 1 0.08% 1176 99.83% . 840-860 509 0.15% , 98.62% 0 0.00% 1176 99.83% 860-880 492 0.15% 98.77% 0 0.00% 1176 99.83% 880-900 413 0.13% 98.90% 0 0.00% 1176 99.83% 900-920 390 0.12% 99.01% 0 0.00% 1176 99.83% 920-940 384 0.12% 99.13% 1 0.08% 1177 99.92% 940-960 330 0.10% 99.23% 0 0.00% 1177 99.92% 960-980 369 0.11% 99.34% 0 0.00% 1177 99.92% 980-1000 301 0.09% 99.44% 0 0.00% 1177 99.92% 1000-1020 316 0.10% 99.53% 1 0.08% 1178 100.00% 1020-1040 300 0.09% 99.62% - - -1040-1060 244 0.07% 99.70% - -1060-1080 284 0.09% 99.78% - - -1080-1100 206 0.06% 99.85% - - - -1100-1020 195 0.06% 99.91% - - -1120-1140 79 0.02% 99.93% - - - -1140-1160 57 0.02% 99.95% - -1160-1180 46 0.01% 99.96% - - -1180-1200 33 0.01% 99.97% - - - -1200-1220 25 0.01% 99.98% - - -1220-1240 20 0.01% 99.98% - - - -1240-1260 17 0.01% 99.99% - - - -1260-1280 10 0.00% 99.99% - - - -1280-1300 6 0.00% 99.99% - - - -1300-1320 2 0.00% 100.00% - - -328825 100.00% 100.00% 1178 100.00% 1178 100.00% Table 6.2 (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around lineament pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2. 164 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 0-20 23566 7.17% 7.17% 93 7.89% 93 7.89% .20-40 53392 16.24% 23.41% 214 18.17% 307 26.06% 40-60 46866 14.25% 37.66% 193 16.38% 500 42.44% 60-80 30263 9.20% 46.86% 148 12.56% 648 55.01% 80-100 26270 7.99% 54.85% 125 10.61% 773 65.62% 100-120 30713 9.34% 64.19% 122 10.36% 895 75.98% 120-140 18985 5.77% 69.97% 77 6.54% 972 82.51% 140-160 19025 5.79% 75.75% 56 4.75% 1028 87.27% 160-180 16407 4.99% 80.74% 46 3.90% 1074 91.17% 180-200 12276 3.73% 84.47% 40 3.40% 1114 94.57% 200-220 10424 3.17% 87.64% 19 1.61% 1133 96.18% 220-240 7563 2.30% 89.94% 11 0.93% 1144 97.11% 240-260 7106 2.16% 92.11% 14 1.19% 1158 98.30% 260-280 5653 1.72% 93.82% 3 0.25% 1161 98.56% 280-300 3971 1.21% 95.03% 6 0.51% 1167 99.07% 300-320 3129 0.95% 95.98% 2 0.17% 1169 99.24% 320-340 2333 0.71% 96.69% 2 0.17% 1171 99.41% 340-360 2057 0.63% 97.32% 2 0.17% 1173 99.58% 360-380 1577 0.48% 97.80% 3 0.25% 1176 99.83% 380-400 1202 0.37% 98.16% 0 0.00% 1176 99.83% 400-420 900 0.27% 98.44% 2 0.17% 1178 100.00% 420-440 759 0.23% 98.67% - - - -440-460 573 0.17% 98.84% - - - -460-480 492 0.15% 98.99% - - - -480-500 447 0.14% 99.13% - - - -500-520 414 0.13% 99.25% - - - -520-540 339 0.10% 99.36% - - - -540-560 260 0.08% 99.44% - - - -560-580 225 0.07% 99.51% - - - -580-600 213 0.06% 99.57% - - - -600-620 194 0.06% 99.63% - - - -620-640 166 0.05% 99.68% - - - -640-660 137 0.04% 99.72% - - - -660-680 137 0.04% 99.76% - - - -680-700 103 0.03% 99.79% - - - -700-720 79 0.02% 99.82% - - - -Totals 328825 100.00% 99.82% 1178 100.00% 1178 100.00% Table 6.3. Results of overlay of landslide initiation points on 20 m buffer zones around stream pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2. 165 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 720-740 67 0.02% 99.84% - - -740-760 67 0.02% 99.86% - - -760-780 63 0.02% 99.88% - - -780-800 61 0.02% 99.90% - - _ _ 800-820 56 0.02% 99.91% - - - -820-840 51 0.02% 99.93% - - -840-860 51 0.02% 99.94% - - - -860-880 46 0.01% 99.96% - -880-900 40 0.01% 99.97% - - -900-920 35 0.01% 99.98% - - -920-940 32 0.01% 99.99% - - - -940-960 18 0.01% 100.00% - - -960-980 12 0.00% 100.00% - - -980-1000 7 0.00% 100.00% - - -1000-1020 3 0.00% 100.00% - - -Totals 328825 100.00% 100.00% 1178 100.00% 1178 100.00% Table 6.3 (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around stream pixels. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2. 166 Chapter 6 Lineaments and landslides 250 200 T 3 •° 150 0 Number of landslides versus distance from target feature Landslides at distance \"X\" from a stream Landslides at distance \"X\" from a lineamenent 0 200 400 600 800 1000 1200 Distance from target feature Figure 6.16. Graph showing the distribution of landslides away from target features. 167 Chapter 6 Lineaments and landslides No landslides are initiated more than 420 m from a stream and 75 % of features are initiated within 100 m of a stream. 144 (12.23 %) landslides are initiated within 40 m of a lineament, 32% are initiated within 100 m of a lineament. A consideration in assessing the results is the area of the watershed which lies within the buffer zones created. This is shown in each table. For example when lineaments are considered 8% of the Seymour watershed lies within the 0 - 40 m buffer zone and, 12 % of landslides lie within this 8%> of the watershed (Table 6.2). 18 %> of landslides lie in the 23 % of the watershed delineated as being within 40 m of a stream (Table 6.3). These results do not clearly reveal a relationship between either lineaments and landslide initiation points or streams and landslide initiation points. For example, Table 6.3 indicates that all landslides are initiated within a distance of about 400 m from a stream. The 0 - 400 m buffer around streams includes 99 % of the watershed. This indicates that 100 % of landslides occur within 99 % of the watershed. This result is meaningless except in indicating a maximum length a landslide can travel before entering a stream if it moves by the shortest possible route to a stream. (Landslide events may be longer because this analysis does not include the effects of topography). An additional investigation was conducted when the buffer zones containing the greatest number of landslides (i.e., the modal classes) for each of the above investigations were found. In each case this distance was determined to be between 0 - 40 m from either a lineament or stream pixel. The effect of a lineament within the 0 - 40 m stream buffer on landslide distribution was examined by identifying pixels that are both within 40 m of a stream and 40 m of a lineament as target cells and constructing distance buffer zones about them. This allowed the investigation of localities where streams are intersected by lineaments as well as locations where lineaments may control stream sections (Figs. 6.17 and 6.18). 168 Chapter 6 Lineaments and landslides Figure 6.17. The northern Seymour watershed showing the 0 - 40 m stream buffer, the \"intersecting lineament and stream\" pixels and landslide initiation points. 169 Chapter 6 Lineaments and landslides Gibbens Creek Stream buffer zone 1 (0 - 20 m) Stream/lineament target zone 1 (0 - 20 m) Landslide initiation point Stream buffer zone 2 (20 - 40 m) Stream/lineament target zone 2 (20 - 40 m) 2 km Figure 6.18. The southern Seymour watershed showing the 0 - 40 m stream buffer, the \"intersecting lineament and stream\" pixels and landslide initiation points. 170 Chapter 6 Lineaments and landslides Table 6.4 shows the results of this analysis. The occurrence of lineaments and streams together is a better predictor of landslide initiation location than either of them independently. It is seen that in the 0 - 40 m from a stream zone (23 % of the watershed) 307 (26 %) slides occur. 102 (9 % of all landslides) occur within 40 m of a lineament intersecting this zone. This indicates that about one third of the slides in the \"0 - 40 m from a stream\" buffer zone occur in 19% of this zone (7 % of the watershed). The landslide density of the entire zone is 10.0 (landslides/km2), the landslide density of the lineament/stream intersection zone is 17.5 (landslides/km2), a significant increase. Extending the investigation to the 60 m buffer zone revealed approximately constant densities. 159 (32 %) of the 500 landslides in the 0 - 60 m from a stream buffer zone occur within the 17 % of this zone which also contains a lineament. This suggests that a slide is 1.75 times as likely to occur in an area where a lineament intersects or controls a stream as in an area without a lineament. The overall landslide density for the watershed area examined and considering 1178 slides is 8.9561 slides/km2: 0.0036 slides/pixel. Although it is difficult to determine the effects of lineaments on landslide distribution in isolation it is seen that where their occurrence corresponds with the occurrence of a stream there is an increased likelihood of a landslide initiation. It is recognized that there are many other factors contributing the occurrence of small surficial landslides. These have been described in Chapter 2. 6.4.3 Discussion It is apparent that careful interpretation of the data generated by spatial correlation analysis is necessary to obtain geologically valid results from these investigations. Analysis of lineament and landslide distribution alone proved insufficient to determine if a relation existed between the two. However when the presence of a lineament and stream are combined in analysis the results suggest that the presence of a lineament in association with a stream at a location provides a site favorable 171 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 0-20 14375 4.37% 4.37% 58 4.92% 58 4.92% 20-40 8473 2.58% 6.95% 44 3.74% 102 8.66% 40-60 12624 3.84% 10.79% 57 4.84% 159 13.49% 60-80 11388 3.46% 14.25% 58 4.92% 217 18.42% 80-100 13526 4.11% 18.36% 59 5.01% 276 23.43% 100-120 14226 4.33% 22.69% 48 4.07% 324 27.50% 120-140 12360 3.76% 26.45% 55 4.67% 379 32.17% 140-160 14659 4.46% 30.91% 70 5.94% 449 38.11% 160-180 14714 4.47% 35.38% 69 5.86% 518 43.97% 180-200 14028 4.27% 39.65% 50 4.24% 568 48.21% 200-220 12829 3.90% 43.55% 56 4.75% 624 52.97% 220-240 11558 3.51% 47.06% 47 3.99% 671 56.96% 240-260 13599 4.14% 51.20% 49 4.16% 720 61.12% 260-280 12312 3.74% 54.94% 42 3.57% 762 64.68% 280-300 11649 3.54% 58.49% 37 3.14% 799 67.82% 300-320 10329 3.14% 61.63% 33 2.80% 832 70.62% 320-340 10292 3.13% 64.76% 37 3.14% 869 73.77% 340-360 9941 3.02% 67.78% 35 2.97% 904 76.74% 360-380 9414 2.86% 70.64% 25 2.12% 929 78.86% 380-400 9019 2.74% 73.39% 29 2.46% 958 81.32% 400-420 7605 2.31% 75.70% 20 1.70% 978 83.02% 420-440 7905 2.40% 78.10% 26 2.21% 1004 85.23% 440-460 6468 1.97% 80.07% 20 1.70% 1024 86.92% 460-480 5776 1.76% 81.83% 17 1.44% 1041 88.37% 480-500 6055 1.84% 83.67% 11 0.93% 1052 89.30% 500-520 5342 1.62% 85.29% 22 1.87% 1074 91.17% 520-540 4883 1.48% 86.78% 13 1.10% 1087 92.27% 540-560 3950 1.20% 87.98% 8 0.68% 1095 92.95% 560-580 3594 1.09% 89.07% 8 0.68% 1103 93.63% 580-600 3696 1.12% 90.19% 9 0.76% 1112 94.39% 600-620 3065 0.93% 91.13% 5 0.42% 1117 94.82% 620-640 3142 0.96% 92.08% 13 1.10% 1130 95.92% 640-660 2558 0.78% 92.86% 5 0.42% 1135 96.35% 660-680 2555 0.78% 93.64% 5 0.42% 1140 96.77% 680-700 2231 0.68% 94.32% 5 0.42% 1145 97.20% 700-720 1872 0.57% 94.89% 1 0.08% 1146 97.28% 720-740 1868 0.57% 95.45% 5 0.42% 1151 97.70% Table 6.4. Results o f overlay o f landslide initiation points on 20 m buffer zones around pixels at the intersection o f streams and lineaments. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m 2 . 172 Chapter 6 Lineaments and landslides Buffer number Percentage Cumulative Number Percentage Cumulative Cumulative zone of pixels of percentage of slide of total number of percentage Distance in buffer watershed initiation slides slides of slides in m pixels 740-760 1467 0.45% 95.90% 4 0.34% 1155 98.04% 760-780 1440 0.44% 96.34% 2 0.17% 1157 98.21% 780-800 1249 0.38% 96.72% 3 0.25% 1160 98.47% 800-820 1045 0.32% 97.04% 2 0.17% 1162 98.64% 820-840 1031 0.31% 97.35% 4 0.34% 1166 98.98% 840-860 911 0.28% 97.63% 4 0.34% 1170 99.32% 860-880 844 0.26% 97.88% 2 0.17% 1172 99.49% 880-900 688 0.21% 98.09% 1 0.08% 1173 99.57% 900-920 623 0.19% 98.28% 1 0.08% 1174 99.66% 920-940 608 0.18% 98.47% 1 0.08% 1175 99.74% 940-960 526 0.16% 98.63% 2 0.17% 1177 99.91% 960-980 546 0.17% 98.79% 1 0.08% 1178 100.00% 980-1000 466 0.14% 98.93% - -1000-1020 467 0.14% 99.08% - - - _ 1020-1040 463 0.14% 99.22% - - - _ 1040-1060 383 0.12% 99.33% - - -1060-1080 389 0.12% 99.45% - - - -1080-1100 298 0.09% 99.54% - - - -1100-1020 257 0.08% 99.62% - -1120-1140 202 0.06% 99.68% - - -1140-1160 180 0.05% 99.74% - - -1160-1180 156 0.05% 99.78% - - -1180-1200 133 0.04% 99.82% - - _ 1200-1220 113 0.03% 99.86% - - - -1220-1240 98 0.03% 99.89% - - - _ 1240-1260 91 0.03% 99.92% - - -1260-1280 69 0.02% 99.94% - - -1280-1300 56 0.02% 99.95% - - -1300-1320 42 0.01% 99.97% - - - _ 1320-1340 36 0.01% 99.98% - -1340-1360 27 0.01% 99.99% - - - -1360-1380 20 0.01% 99.99% - - - -1380-1400 12 0.00% 100.00% - -1400-1420 7 0.00% 100.00% - - -1420-1440 2 0.00% 100.00% - - -Totals 328825 100.00% 100.00% 1178 100.00% 1178 100.00% Table 6.4 (Continued). Results of overlay of landslide initiation points on 20 m buffer zones around pixels at the intersection of streams and lineaments. 21 landslide pixels are omitted from the results (see text). Each pixel represents 400 m2. 173 Chapter 6 Lineaments and landslides for landslide initiation. The addition of streams indicates that the pixels identified in this analysis are at the topographic low points and field experience in this area suggests that they are likely to be in the vicinity of steep gully walls. The presence of a lineament intersecting or paralleling these walls may be sufficient to initiate a failure. This will likely be either a failure in structurally weakened bedrock (which may have been undercut by rapid bedrock incision in these zones) or the result of concentration of ground and surface waters sufficient to cause a precipitation initiated event. Additionally in the event of seismic activity shaking in these weakened zones may cause rockfalls capable of triggering debris flows. The proximity of these failures to the streams suggests that they will move directly into the stream, at which point they typically flow down the confined stream channel and ultimately deposit on the fan. This a the primary mechanism for sediment transfer from the basin to the fan. Only major lineaments identified on 1;60,000 scale air photos were considered in this analysis therefore only more significant structural features were used to correlate with landslides. Examination of Figure 6.14 shows that several of these major lineaments control streams, e.g., Gibbens Creek and Burwell Creek. Fannin Creek is fault controlled and has been identified as a fault zone by the GSC (Roddick 1965). This fault was not visible on 1:60,000 scale air photos. The fault extends for great length up Seymour Valley (see Fig. 4.2). A number of smaller creeks on both the eastern and western flanks of the valley are also lineament controlled. There is no apparent increase in landslide activity in these creeks as compared to others. Landslide activity seems concentrated near lower order streams. There are some concentrations of landslides about lineaments and stream intersection points easily visible in Figure 6.14, as an example the northern side of Clipper Creek shows a large number of landslides in close proximity to the lineament which successively intersects several subsidiary creeks. As well as this, many areas exist where landslides appear to be initiated preferentially at a particular distance from a 174 Chapter 6 Lineaments and landslides creek or at a uniform elevation above it, e.g., the northern side of Orchid Creek. Therefore more analysis of other attributes is required to determine whether lineaments are at all responsible for landslide activity in these areas. It is likely that increasing the number of lineaments by increasing lineament mapping scale would increase the strength of correlation between lineaments and landslides. Care should be taken that this is not due to the increase in the area of the buffer zones around lineaments. There is likely to be a threshold point at which such analysis would prove counterproductive for this reason. 6.5 Small surficial landslides as sediment sources and contributors of materials to fans The GVRD have found that stream bank erosion and landsliding in gully walls are primary sources of sediment in the Seymour watershed (GVWD watershed ecological inventory pilot study, Final report, March 1993). The GVRD study estimated the amount of material yielded by each sediment source in a basin in the Seymour watershed. Values calculated in the report were based on average annual stream bank retreat rates and the amount of (field) observed material resulting from landslide activity. Identification of landslide initiation sites is of primary importance to the GVRD. Lineaments do not appear directly responsible for the introduction of fine sediment to the water supply (Section 4.4). However it appears that there is a relation between lineaments and the location of landslide initiation points. This suggests an indirect relation between the lineaments and fine sediment. The GVRD assume a fines content of about 20% in landslides (GVRD ecological inventory final report, March 1993). The fine sediment content of a landslide deposit is typically winnowed from the deposit and flushed into the primary river. Fans in the GVRD watersheds typically comprise large quantities of gravel to small boulder size materials with fine materials concentrated in 175 Chapter 6 Lineaments and landslides naturally occurring sediment traps such as behind boulders and large organic debris. Larger materials are common higher in the creeks where boulders of several meters are often encountered. The movement of sediment from the basin to the water supply is a complex process. Upon mobilization in a landslide sediment will move from the initiation point to the nearest stream or gully, which is typically very close (within a few hundred meters), here the materials may be deposited or continue to flow as a torrent until deposition when the gradient of the slope is insufficient for further movement. Sediment will reside at the deposition site until winnowed from the deposit, the last remaining fines are typically those caught in sediment traps which may be obliterated by subsequent debris flow events. Figure 6.19 shows the view down a landslide track in the Camp Creek basin, Coquitlam watershed which has moved directly to the nearest gully. The slide has deposited material in Camp Creek (Fig. 6.20) where it will reside until re-mobilized. The landslide was initiated about 60 m from the creek and has built a small cone into the creek, just visible in the foreground of Figure 6.20 is a small debris levee. It has been pointed out in south-central British Columbia that valley degradation probably ceased around 3-4,000 years after deglaciation (Ryder 1971b). Since this time major erosion has been concentrated in small point sourced landslide events. It is likely that the fans seen in the southwest Coast Mountains have continued to build throughout the last 10,000 years by the process described above with considerable amounts of material being deposited on them by debris flows. 176 Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides Chapter 6 Lineaments and landslides 6.6 Conclusions An inventory of large rock landslides and mountain slope deformation was conducted in the regional study area from 1;60,000 scale air photos. Besides allowing an assessment of regional hazard distribution this allowed the relation between lineaments and landslides to be investigated. Landslides were identified on air photos by the presence of landslide scars and rubble filling the valley bottoms. Slope deformations were identified by the presence of antislope scarps and slope bulging. Landslides and slope deformations were inventoried over an area of approximately 10,900 km2 and twenty sites identified. Six of these have been previously reported. Rock avalanches and slope deformations are the commonest features locally. Other failure types include large rockfalls and surficial slides in Quaternary sediments adjacent to major rivers. The density of bedrock failures in the area is 0.001 failures/km2 and the density of mountain slope deformations is 0.0006 deformations/km2. Because all features have occurred postglacially the temporal distribution of failures is assessed at one every 910 years although this does not preclude the possibilities of repeated failures at a single site or a single triggering mechanism initiating multiple events, for example an earthquake. Lineaments are clearly related to many of the inventory features. The best example presented is the Dickson Lake landslide and associated West Norrish Creek site. Lineaments appear related to inventory features in three ways: 1) Lineaments may form the headscarp of major landslides, for example Dickson Lake; 179 Chapter 6 Lineaments and landslides 2) Lineaments may either provide locations about which slope deformation may occur, for example the West Norrish Creek site, or be caused by slope deformation as is probably the case at Goat Ridge; 3) Lineaments may form rock faces capable of shedding large rockfalls, for example Stave River I (Table 6.1). Mountain slope deformation accounts for seven inventory sites. These features may represent a re-equilibriating response to deglaciation and many may be currently inactive however the state of activity is difficult to assess on air photos. The antislope scarps on the Lions in the Capilano watershed are shedding materials suggesting at least active weathering processes. These features represent possible hazards because their future behavior is uncertain. Mountain slope deformation may ultimately result in catastrophic failure. When the inventory features are compared to landslides and mountain slope deformation in the adjacent Fraser River Valley, it appears that these features are less common in this study area. The inventory compiled by Savigny (1996) identified 58 sites of instability in a 40 km wide corridor centred on Fraser River. Leir (1995) investigated factors that might influence landslide distribution including; rock type, slope, proximity to a (known) fault, proximity to a (mapped) lineament. Using a weights of evidence modelliing technique it was determined that the top five predictors of large rock landslide occurance in his study area were: 1) presence of Custer Gneiss; 2) within 1300 m of a fault trace; 3) within 1200 m of a lineament; 4) presence of Cultus Formation; 5) presence of Chiliwack Formation (Leir 1995, p. 74). It is clear that rock type strongly influenced landslide distribution however the importance of faults and lineaments is also clear. The primary difference between the two study sites is the presence of several major, pervasive fault zones: the Hope, Yale and Vedder. It is likely that these faults are strongly influencing the landslide characteristics of the 180 Chapter 6 Lineaments and landslides Fraser River Valley and that comparison of the two study areas shows that landslides are less frequent farther from major fault zones (Savigny, pers. comm. 1996). Although one major fault zone is present in the study area, the Harrison Fault Zone, examination of Figure 6.1 (showing distribution of inventory sites) does not appear to show a decrease in sites away from this fault zone. It should be noticed however that only a small portion of this fault zone intersects the study area. Other faults mapped in this area (shown on figure 4.2) are smaller features than those seen in the Fraser River Valley study. Although small surficial landsliding is ubiquitous in the regional study it was not reported in the regional inventory. Instead the occurrence of these features was investigated in the Seymour watershed north of Vancouver. The GVRD have found that stream bank erosion and landslides in gully walls to be the primary sources of sediment input to the water supply. As such landslide initiation points represent points of high erosion and sediment supply. An inventory supplied by the GVRD was used to determine the relation between landslide initiation points and lineaments. A total of 1178 small surficial landslide initiation points were used in spatial correlations between lineaments and landslides and streams and landslides. In each case the greatest numbers of landslides were initiated between 20 and 40 m from the feature in question. Analysis of landslide distribution compared to sites where both a lineament and a stream existed showed that a landslide is almost twice as likely to occur in these zones than in a zone occupied.by a stream alone. An average of 10 slides per kilometer were found within 40 m of a stream and 17.5 slides per kilometer were found within 40 m of a lineament/stream intersection. It is concluded that while many other factors may influence landslide distribution in the Seymour watershed there is a correlation between landslide distribution and locations where lineaments intersect streams. These zones are likely to represent areas where lineaments focus surface and ground water into the creeks or where rapid incision of creeks into structurally weakened 181 Chapter 6 Lineaments and landslides bedrock has oversteepened slopes sufficiently to cause bedrock failure capable of initiating debris flow activity. In conclusion lineaments influence landslides of all scales. Following lineaments on an air photo will often lead to identification of landslide or slope deformation features and as such lineament mapping is an important tool in hazard investigation. Additionally it has been previously reported (GVWD watershed ecological inventory pilot study, Final report 1993) that small surficial landslides are a major source of sediment in the Seymour watershed. As such these represent a contemporary process capable of moving materials from a drainage basin to its fan. Fine sediment is often winnowed from deposits on the fan and flushed into the primary river system. 182 Chapter 7 Discussion and conclusions CHAPTER 7 DISCUSSION AND CONCLUSIONS This study was undertaken to evaluate the importance of structure as a control on drainage basin development in the southwest Coast Mountains, British Columbia, Canada. Lineaments mapped from air photos were used to assess the structure of the regional study area and then to investigate the effects of this structure at the smaller, drainage basin, scale. The influence of lineaments on drainage basin position, stream network pattern and basin morphometry was investigated either by assessment of spatial correlations or by regression analysis of variables. Also investigated was the possibility of determining sediment yield from basin morphometric parameters. Lineament influence on large scale geomorphic process was assessed by examining lineament control on large rock landslides and slope deformations. Finally an investigation into lineament control on debris flow and avalanche activity in the Seymour watershed was conducted using an inventory compiled by the GVRD. The term lineament as applied in this research is defined as a non-genetic term describing a feature recognizable on an air photo or other remotely sensed image, that is essentially linear, and comprises a contiguous set of topographic, drainage, vegetation, moisture and tonal features. An inventory of 4215 lineaments was compiled from air photos of the regional study area in order to obtain an understanding of regional structure and to provide the necessary framework for drainage basin studies. Lineaments used in this research were mapped on 1 ;60,000 scale air photos before transfer to topographic maps. Lower altitude photos were used to assist with field work. It was determined that 183 Chapter 7 Discussion and conclusions mapping lineaments on 1;20,000 scale air photos produced an approximately threefold increase in the number of features detectable on 1;60,000 scale air photos. Orientation data for mapped lineaments was provided by a FORTRAN77 program TREND. A preferred orientation of 17.43 ± 3.73° was found for the inventory data. Peaks were also found representing northwest and east-west trends. Two of these trends correlate with known structural trends in the region. The northwest trend is related to the oldest developed (Cretaceous) structural trend. This trend is particularly developed in Jurassic rocks in the northeast corner of the study along the Harrison Lake Fault zone. The northeast trend is probably related to a Tertiary trend developed in the region between 25 and 14 Ma. The east-west trend has not been previously reported and it is suspected that this is the youngest regionally developed trend for three reasons: 1) These lineaments are numerous and persist for tens of kilometers throughout the central and southern study area cutting the youngest rocks in these areas. 2) They appear to cut Quaternary valley fill in parts of the southern map area and some are visible in recent alluvium. (This assertion is currently the subject of further investigation by Thurber Engineering Ltd. under contract from the GVRD). 3) They are correctly oriented to have resulted from stresses induced by the current convergence vector of the Juan de Fuca and North American plates. Investigation of the regional lineament trends demonstrated that it is possible to infer regional structure from lineament mapping. Additionally variations in structural styles can be found by looking for statistical differences between different areas. For example the lineament pattern visible in the northeastern corner of the study is significantly different from that in the rest of the area. 184 Chapter 7 Discussion and conclusions Lineament mapping and analysis in the manner described has suggested that rocks in the study area have been subjected to three phases of lineament inducing stress. These are summarized as follows: 1) Phase one: associated with dextral shear in the Cretaceous produced northwest trending lineaments and is responsible for the trends of the major river valleys; 2) Phase two: associated with northeast-southwest crustal shortening in the Tertiary which emplaced a northeast trending lineament se; 3) Phase three: a more recent east-west trend with lineaments of unknown age which may represent present convergence between the North American and Juan de Fuca plates. Field investigations determined that it is difficult to detect lineaments on the ground in this area because of the nature of the topography and vegetal cover. Where detected they were found to be represented by gullies of various forms including narrow bedrock \"canyons\" in cliff faces, deep cuts in the ridgetops, steep-sided, deeply incised creeks, and small topographic lows traversing hillslopes. Although lacking in situ evidence of movement, it is suspected that many of these features are faults or large scale regional joints. Field investigation also concluded that it is unlikely that these features are directly responsible for fine sediment yield in drainage basins. Instead they contribute to headward erosion of drainage basins by exposing bedrock directly to weathering processes such as freeze-thaw. In order to investigate lineament effects on drainage basins a sample set of 25 were randomly selected from the study area. The basins selected were determined to be representative of a number of variables in the area such as aspect, basin order, elevational ranges and lithology. Drainage divides for sample set basins were mapped from 1:50,000 scale topographic maps. It was possible to transfer additional lineaments from 1:60,000 scale air photos to this map scale, specifically smaller features identifiable on air photos that were not transferable to the original 1:250,000 scale topographic map. It was also determined that hydrology mapped on the 1:50,000 scale topographic maps represented only 185 Chapter 7 Discussion and conclusions higher order streams. In order to represent first order streams in the analysis it was necessary to map these on air photo and transfer them to the map sheets. 19 morphometric parameters were measured for each basin and 2 physical attributes, aspect and geology, were recorded. Lineament control on drainage basin position was investigated by determining how many of the sample set basins were located about an axial lineament. It was determined that 22 (88%) of the basins studied could be shown to have a major lineament controlling either the axial line of the basin, the path of the main stream, or (typically) both. The presence of a lineament at the center of a basin suggests that the lineament was either responsible for the positioning of the master rill early in the development of a surface flow regime or captured the master rill very shortly after its conception. A lineament cutting a sloping surface may have provided the topographic low point necessary for concentration of surface water flow. A lineament also provides the zone of weakness in the rock mass necessary for rapid incision of the bedrock. A spatial correlation analysis was conducted to determine to what extent the stream pattern in a basin mirrored the lineament pattern. This was done by creating buffer zones around lineaments and overlaying the stream network to determine the extent of correlation. When all sample set basins and all stream orders were considered an average of 66% of the stream network was found to overlay lineaments. An average of 51% of the lineament network is overlain by streams. In igneous basins the average correlation was 71% and in metamorphic basins 46%. When separated to individual stream orders first order streams showed the highest correlations. Poor correlations were observed for third order and higher (fourth order) streams because these tended to flow in Quaternary sediments which mask the effects of lineaments. Despite this, the orientations of third order streams tended to agree well with lineament trends presumably because these often flow on a basin floor which is controlled by a major lineament. 186 Chapter 7 Discussion and conclusions A directional analysis of lineaments and lineament controlled streams was conducted for the sample set basins. It was found that preferred lineament trends existed and were different in both igneous and metamorphic basins. The average lineament orientation in igneous basins was found to be 28.8 ± 24.13° and in metamorphic basins 0.15 ± 20.10°. Stream segment trends for these sets were determined to be 20.79 ± 43.31° and 167.1 ± 39.87° respectively. Statistically, the data for igneous and metamorphic basins cannot be shown to derive from the same population. Although not statistically proven by these investigations it appears that the overall preferred lineament trend for lower order streams in the sample set basins is approximately north-northeast, similar to the Tertiary emplaced structural trend. Both lineament and stream segment trends are more varied in metamorphic basins, consistent with the idea that these rocks have a longer history and have been subjected to more varied lineament inducing stress conditions. When the results of spatial and directional data on lineaments and streams for the sample set basins are compared it can be concluded that significant evidence exists to state that the drainage pattern in these basins is largely controlled by the distribution of lineaments in the underlying bedrock. The implication of this is that lineaments not only capture the master rill early in the formation of a drainage network on a sloping surface but also will capture rills and channels on the sideslopes as well. This supports the idea that at surface lineaments are expressed as topographic low points, and that they are likely to represent zones of weakness in the rock mass which may be easily eroded. It is also interesting to speculate on die extent to which these lineaments influence groundwater flow. The creeks formed in these zones are very deeply incised and it is likely that the zone has permeability at depth. Having established a connection between bedrock structure and the stream network it was decided to investigate the correlation between lineaments and basin morphometry. Classically the morphometry of a basin has been attributed to the stream pattern and measured in terms of stream 187 Chapter 7 Discussion and conclusions length and drainage density, as well as other parameters of the drainage network. If the stream pattern mirrors structure in the bedrock then there should be a connection between lineaments and basin morphometry. Lineament length and density were chosen to be the most representative measures of bedrock structure because they are most easily measured and the linear nature of lineaments suggests comparison with stream length and drainage density. The sample set basins were examined as a group and were also separated into basins larger than one square kilometer and smaller than this threshold value. Morphometric parameters were linearly regressed against one another and all possible combinations of analysis performed. Three groups of data showed high strength relations: 1) Sets of variables related to the areal and dimensional character of the basin such as basin area and basin width and length returned high correlations with each other as did relief parameters. 2) High correlations (measured in terms of Revalues) existed between stream length and drainage density and certain basin morphometric parameters (typically areal and relief parameters). 3) High correlations existed between lineament length and the same parameters as for stream length. In the entire sample set stream length was found to be a better predictor of basin morphometry than lineament length in many cases. However correlations between lineament length and basin morphometric parameters compared well with those between stream length and the same parameters typically being within a few percent of those found for stream length. Strong correlations (typically R 2 > 75%) existed between lineament length and basin area, length, width, maximum basin elevation, maximum basin relief, relief ratio, basin relief, and basin gradient. In most cases the best correlations were returned in the larger basins. 188 Chapter 7 Discussion and conclusions The strength of most correlations in the smaller basins is poor (R2 < 0.40). Although it was not satisfactorily determined why this should be it may be related to the processes operating in these basins. Sediment yield characteristics are known to be different in small alpine basins than in the larger basins, increased elevation plays a more important role because many of these basins are more rugged and less forested than the larger basins. Previous research indicated that it basins should be examined according to whether they were larger or smaller than one square kilometer (e.g. Owens and Slaymaker 1992). Typically relations are very poor (R2 < 0.4) in the smaller basins but stream length is generally a better predictor of any given parameter than lineament length. In larger basins the strength of nearly all relations is similar when stream length and lineament length are jointly compared to other parameters. It is speculated that in the larger basins lineament influences are more evident. It is possible that one reason why the stream network is a controlling factor in drainage basin development is that the stream network within a basin is a complete and almost isolated network (the only connection to a larger network is by way of the basin mouth) therefore the stream network is responsible for the movement of all materials out of the basin. The lineament network in a basin is not an isolated network. Links exist in this network to multiple basins by lineaments traversing the divides. Lineaments are places where erosion concentrates and sediment is stored, they therefore exert some control on basin processes. The larger a basin becomes the more likely it is to include a connected lineament network. This can be extended to the regional scale where it is possible that an isolated (i.e., distinct) lineament network exists. It is speculated that at the regional scale the lineament network may become more significant than the stream network in morphometric control of the landscape. Some evidence for this hypothesis is found in Koons (1995) where it is demonstrated that a ridge valley system develops parallel to the main orogen strike in a structurally controlled landscape. 189 Chapter 7 Discussion and conclusions A possible explanation for the fact that stronger correlations exist between stream length and basin morphometry than between lineament length and basin morphometry is that a stream network must develop in a basin regardless of the presence of lineaments, in order to satisfy the natural law that every point in a basin be drained. Hence if no lineament exists on a slope a stream must make its own course on that slope. The effect of this is that in a drainage basin without lineaments morphometry will be due to the stream network. In a basin with a very high lineament density morphometry is likely to be more heavily related to the lineament network because of the erodability of the lineaments. In other words, if two identical basins (developed on the same rock under the same climatic conditions) evolve, one with no lineaments and one with a high lineament density, in the basin without lineaments morphometry will be related to the stream network. In the basin with the high lineament density the erodability of the lineaments will more strongly influence morphometry and there will be relations between lineament length and basin morphometry. The relations between stream length and the morphometry of this second basin will no longer be the same. Sediment yield from the sample set drainage basins was investigated for two reasons; first because it provides a means of investigating contemporary process in these basins and secondly, because sediment input to the water supply is a primary concern of the GVRD and the local population. Several morphometric parameters have been previously identified as related to sediment yield from drainage basins for example, basin area (Church et al 1989), fan area (Bull 1962), fan gradient (Ryder 1971) and relief ratio (Schumm 1954). Because sediment yield has not been measured directly for basins in the study area basin area was chosen as reflecting sediment yield from a basin. Church et al. (1989) have identified a relationship between basin area and sediment yield for basins in southwest British Columbia. It was determined that fan area is not a good predictor of basin area (hence sediment yield). If it can be measured fan volume might prove more useful. Relief ratio is the best predictor of sediment 190 Chapter 7 Discussion and conclusions yield because it correlates very well with basin area (R2 = 0.88). Both lineament and stream length correlate well with basin area (R2 = 0.83 and 0.97 respectively). Landslide events of all types represent contemporary landscape processes. An inventory of post-glacial, large rock landslides and mountain slope deformation was completed from 1:60,000 scale air photos over a total area of approximately 10,900 km2. Twenty sites were identified providing a site density of 0.0018 sites/km2. Rock avalanches were found to be the most frequent event locally (8 total) with slope deformations being almost equally distributed (7 total). When only sites representing bedrock failures are considered the event density is 0.001 landslides/km2. Other types of landslide occurring locally are rockfalls and surficial slides adjacent to major rivers. Large rock landslides and mountain slope deformation were inventoried to assess regional hazard and the effects of lineaments on these features. Landslides were identified on air photos by the presence of landslide scars and rubble filling valley bottoms. Slope deformations were identified by the presence of antislope scarps and sometimes slope bulging. It was determined that lineaments influence landslides in three ways: 1) Lineaments may form the headscarps of major landslides; 2) Lineaments may either provide locations about which slope deformation can occur or may be the result of slope deformation; 3) Lineaments may form rockfaces capable of shedding large rockfalls. Mountain slope deformations are important to evaluate because of the uncertainty of their future behavior. It is believed that many of these sites in the southwest Coast Mountains may represent a re-equilibrating response to deglaciation and as such many may be currently inactive. However, seismicity may periodically reactivate sites in the study area (Savigny pers. comm. 1996). Small surficial landslides represent a contemporary process of sediment evacuation from drainage basins. An inventory of landslide initiation points in the Seymour watershed was obtained 191 Chapter 7 Discussion and conclusions from the GVRD. 1178 landslide initiation points were considered in the analysis and these were investigated to determine the correlation between their distribution and the distribution of streams and major lineaments in the watershed. Using the lineaments mapped in the original lineament inventory and streams extracted from the GVRD's database it was determined that 11% of landslides were initiated within 40 m of a lineament and 23% of landslides were initiated within 40 m of a stream. However the area of the watershed included in these buffer zones was extensive and introduced a lack of confidence in the results of the analysis. The intersection points of streams and lineaments were investigated in order to determine whether the presence of a lineament and a steam at the same locality influenced landslide distribution. It was found that a landslides was almost twice as likely to occur within 40 m of a lineament/stream intersection as within 40 m of either a lineament or stream alone. The landslide density within the \"40 m from a stream\" and the \"40 m from a lineament\" buffer zones was 10 landslides/km2 in both cases whereas the landslide density within the \"40 m from a lineament/stream intersection\" buffer zone was found to be 17.5 landslides/km2. The result is considered reasonable because where a lineament intersects a stream the location represents either a site where the lineament controls stream position (in which case rapid incision into the lineament zone oversteepens gully walls to the point of instability), or the lineament is likely to a focus surface flow and groundwater flow to the creek (if the lineament is oblique to the stream). The concentration of water in what is likely to represent a topographic low point may be sufficient to cause a small surficial landslide along the lineament. Once mobilized these small landslides will move preferentially to the nearest creek where they will either continue to flow or will deposit until later remobilized by subsequent landslide events or saturation by water. Ultimately the material contained in the landslide will be deposited on the fan where fine material may be winnowed from the deposit and flushed into the primary stream. Therefore 192 Chapter 7 Discussion and conclusions debris flows and avalanches represent a contemporary means of moving materials from the drainage basin and there is some evidence that their occurrence may be related to lineament and stream interaction. It is recognized that a large number of other factors may contribute to landslide initiation at a given point. Further investigation is recommended into the landslide distribution in the Seymour watershed using the GVRD inventory. If lineament mapping were conducted at 1:20,000 scale it is possible that the correlation between small surficial landslides and lineaments would be improved. In the analysis described only the larger lineaments were included. Additionally it would be beneficial to include other landscape attributes in the analysis such as slope, aspect, and surface material type. If the groundwater flow in these lineaments could be properly investigated the flow patterns around a lineament might be useful in predicting landslide initiation points. Additional groundwater investigations could include the effect of lineaments on regional groundwater flow patterns. For example, is there a regional anisotropy developed in the flow pattern due to the presence of lineaments? Investigation into the depths at which water flows in lineament controlled creeks might also be beneficial. It is likely that creeks developed on lineaments transmit water along their course at some depth below the surface. This could account for the fact that many of the creeks appear dry except during times of heavy rain. Subsurface flow in these zones could have implications for tunneling operations. Investigation into the east-west lineament trend identified in the lineament inventory would hopefully ascertain the age and nature of movement (if any) on these features and would assist with regional structural interpretation. Field investigations and examination of the seismic record could determine the current state of activity on these features. It would also be interesting to see if lineament/stream correlations could be improved by field examination. More work could be conducted on the preferred orientations of stream segments to see if 193 Chapter 7 Discussion and conclusions there if any temporal information can be obtained from structurally controlled stream courses. For example, was there a change in the flow paths of the streams in response to the initiation of phase two deformation. Further investigations into the relation between lineaments and landslides would include rigorous field examination and description of the inventory sites, investigations into other possible causes for the features examined and proper identification of the current state of activity at the sites of mountain slope deformation. It would also be interesting to try to separate the effects of lineaments of different orientations (and thus inferred age) on the inventory features. For example, are all sites of mountain slope deformation associated with one particular lineament trend and what is its age. In conclusion it has been demonstrated that lineament mapping can be used to investigate the structural trends of a region. It has also been shown that lineaments strongly influence the location of drainage basins, and the stream network pattern developed in them. The evidence suggests that the stream pattern is largely a reflection of the underlying bedrock structure. The biggest problem in analyzing the relation between streams and lineaments is that, by necessity, a stream network must form in a basin even in the absence of lineaments. This fact probably obscures, to some extent, the degree to which bedrock structure influences basin morphometry. By demonstrating lineament control on large rock landslides and mountain slope deformations and the suspected influence on debris avalanche and flow locations it is shown that lineaments influence contemporary basin processes. This study suggests that the morphometry of the drainage basins in the southwest Coast Mountains and the processes operating within them are a consequence of the underlying bedrock structure that has been emplaced by tectonic stresses. It is therefore argued that processes operating at tectonic scales influence and are reflected in the landscape patterns around us. 194 REFERENCES 195 Adams, J. 1985. Large-scale tectonic geomorphology of the Southern Alps, New Zealand. In Tectonic Geomorphology, M. Morisawa, and J.T. Hack (eds.), Boston: Allen & Unwin, pp. 105-128. Armstrong, J.E. 1984. Environmental engineering applications of the surficial geology of the Fraser Lowland, British Columbia. Geological Survey of Canada paper 83-23. 54 pps. Autodesk Inc. 1993. AUTOCAD release 12: user's guide. 562 pps. Bagnold, R.A., 1973. The nature of saltation and \"bed-load\" transport in water. Proceedings of the Royal Society of London, Series A. 332, pp. 473-504. Barton, C.C., Channer, D.DeR., and Cande, S. 1990. Fractal scaling of fracture networks in rock. American Geophysical Union 1990 fall meeting, Eos Transactions, AGU 71 (43), p.1595. Birdseye, R.U., and Christians, G.L. 1988. Structural control of drainage in late Pleistocene and Holocene sediments, Ascension Parish, Louisiana. Gulf Coast Association of Geological Societies and Gulf Coast Section SEPM meeting; abstracts, AAPG-Bulletin, 72. (9). p. 1109. Bloom, A.L. 1991. Geomorphology, a systematic analysis of Late Cenozoic landforms. Prentice Hall, 2nd edition. 532 pps. Bovis, M.J. 1982. Uphill-facing (antislope) scarps in the Coast Mountains, southwest British Columbia. Geological Society of America Bulletin, Vol. 93, pp. 804-812. Bovis, M.J. 1990. Rock-slope deformation at Affliction Creek, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences, Vol. 27. pp. 243-254. Bovis, M.J., and Dagg, B.R. 1988. A model for debris accumulation and mobilization in steep mountain streams. Hydrological Sciences Journal, Vol. 33, pp. 589-604. Bovis, M.J and Evans, S.G. 1995. Large-scale mountain slope deformation, upper Lillooet Valley, British Columbia. 48th Canadian Geotechnical Conference, September 25-27, 1995, Vancouver BC. pp. 901-908. Buchanan, P. and Savigny, K.W. 1990. Factors controlling debris avalanche initiation. Canadian Geotechnical Journal, Vol. 27, pp. 659-675. 196 Bull, W.B. 1964. Erosion and sedimentation in a semiarid environment: geomorphology of segmented alluvial fans in western Fresno County, California. United States Geological Survey, Professional Paper 352 - E. pp. 89-129. Burdick, R.G., and Speirer, R.A. 1980. Development of a method to detect geologic faults and other linear features from LANDSAT images. United States Bureau of Mines, Report of Investigations 8413, 74 pps. Chigira, M . 1992. Long-term gravitational deformation of rocks by mass rock creep. Engineering Geology, 32, pp. 157-184. Church, M. , Kellerhals, R., and Day, T.J. 1989. Regional clastic sediment yield in British Columbia. Canadian Journal of Earth Sciences, Vol. 26, pp. 31-45. Clague, J.J. (compiler) 1989. Quaternary geology of the Canadian Cordillera, Chapter 1 in Quaternary Geology of Canada and Greenland, R.J. Fulton (ed.), Geological Survey of Canada, Geology of Canada, no. 1, pp. 15-96. Clague, J.J. and Evans, S.G. 1994. Formation and failure of natural dams in the Canadian Cordillera. Geological Survey of Canada, Bulletin 464, 35 pps. Costa, J.E. 1984. Physical geomorphology of debris flows. Chapter 9 in Developments and applications of geomorphology, J.E. Costa and P.J. Fleisher (eds.), Springer-Verlag, Berlin, Heidelberg, pp. 267-317. Cronin, V.S., Sverdrup, K.A., and Schurter, G.J. 1990. LANDSAT drainage lineaments, seismicity and uplift of the Nanga Parbat-Haramosh Massif, northwest Himalaya. Geological Society of America, 1990 annual meeting, Abstracts with Programs, 22 (7). p. 232. Cruden, D.M., Thompson, S., Bornhold, B.D., Chagnon, J.-Y., Locat, J., Evans, S.G., Heginbottom, J.A., Moran, K., Piper, D.J.W., Powell, R., Prior, D., and Quigley, R.M. 1989. Landslides: extent and economic significance in Canada; in Brabb and Harrod (eds.) Landslides: extent and economic significance. 1989 Balkema, Rotterdam, pp. 1-23. Davies, T.R., Phillips, C.J., Pearce, A.J., and Zhang, X.B. 1992. Debris flow behaviour - an integrated overview, in Erosion, Debris Flows and Environment in Mountain Regions (Proceedings of the Chengdu Symposium, July 1992), IAHS Publication no. 209, 1992, pp. 217-225. Davis, J.C. 1986. Statistics and data analysis in geology. 2nd edition, John Wiley and Sons, 646 pps. 197 Dietrich, W.E., Reneau, S.L., and Wilson C.J. 1987. Overview: \"Zero-order basins\" and problems of drainage density, sediment transport and hillslope morphology. Erosion and sedimentation in the Pacific rim (Proceedings of the Corvallis Symposium, August, 1987). IAHS publication no. 165, pp. 27-37. Eastman, J.R. 1993. IDRISI: A raster based Geographic Information System, Users Guides, Version 4.1, rev. 0, Graduate School of Geography, Clark University, Worcester, Massachusetts, USA. 177 pps. Evans, S.G. 1984. The 1880 landslide dam on Thompson River, near Ashcroft, British Columbia; in Current Research, Part A; Geological Survey of Canada, Paper 84-1 A, pp. 655-658. Evans, S.G. 1986, Landslide damming in the Cordillera of western Canada. Landslide dams: processes, risk and mitigation, (ed.) R.L. Schuster; American Society of Civil Engineers, Geotechnical Special publications, No. 3, pp. 111-130. Evans, S.G. 1992. High magnitude - low frequency catastrophic landslides in British Columbia, in Geological hazards in British Columbia, Proceedings of the Geologic Hazards Workshop, 1991, British Columbia Geological Survey Branch 1992 - 15, pp. 71-98. Evans, S.G., and Clague, J.J. 1988. Catastrophic rock avalanches in glacial environments. Proceedings, Vth International Symposium on Landslides, Vol. 2, pp. 1153-1158. Evans, S.G., and Savigny, K.W. 1994. Landslides in the Vahcouver-Fraser Valley-Whistler region, in Geology and Geological Hazards in the Vancouver Region, Southwestern British Columbia, (ed.) J.W.H. Monger, Geological Survey of Canada, Bulletin 481, pp. 251-286. Fanin, R.J., and Rollerson, T.P. 1993. Debris Flows: some physical characteristics and behaviour. Canadian Geotechnical Journal, Vol. 30, pp. 71-81. Ferguson, R. 1976. Linear regression in geography. Concepts and techniques in modern geography No. 15, 44 pps. Foley, M. 1980. Bed-rock incision by streams. Geological Society of America Bulletin, Part II, 91, pp. 2189-2213. Gaile, G.L., and Burt, J.E. 1980. Directional Statistics. Concepts and techniques in modern geography No. 25, 39 pps. 198 Gabrielse, H. and Yorath, C.J. 1991. Tectonic synthesis, Chapter 18 in Geology of the Cordilleran Orogen in Canada, H. Gabrielse and C.J. Yorath (ed.), Geological Survey of Canada, Geology of Canada, no. 4, pp. 677-705. GVWD Watershed Ecological Inventory Pilot Study (Jamieson-Orchid-Elbow Drainage) Final Report, 18 March 1993, Prepared by Acres International Limited, B.A. Blackwell and Associates, Northwest Hydraulic Consultants, Oikos Ecological Consulting, Phero Tech Inc. Remtec Inc. Hare, P.W., and Gardner, T.W. 1985. Geomorphic indicators of vertical neotectonism along converging plate margins, Nicoya Peninsula, Costa Rica. Tectonic Geomorphology, M. Morisawa and J.T. Hack (eds.) pp. 75-104. Harris, J.R. 1991. Mapping of regional structure of eastern Nova Scotia using remotely sensed imagery: Implications for regional tectonics and gold exploration. Canadian Journal of Remote Sensing, Vol. 17, No. 2, pp. 122-135. Helmlinger, K.R., Kumar, P., and Foufoula-Georgiou, E. 1993. On the use of digital elevation model data for Hortonian and fractal analysis of channel networks. Water Resources Research, 29. (8). p. 2599-2613. Hestir, K., Martel, S.J., and Long, J.C.S. 1990. Generation of fracture patterns using self-similar iterated function systems. American Geophysical Union 1990 fall meeting, Eos Transactions, AGU 71 (43), p. 1595. Hicock, S.R., and Armstrong, J.E. 1984. Vashon Drift: definition of the formation in the Georgia Depression, southwest British Columbia. Canadian Journal of Earth Sciences, Vol. 22, pp. 748-757. Horton, R.E. 1932. Drainage basin characteristics. Transactions of the American Geophysical Union, Vol. 13, pp. 350-361. Horton, R.E. 1945. Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. Bulletin of the Geological Society of America, Vol. 56, pp. 275-370, March 1945. Howard, A.D. 1967. Drainage analysis in geologic interpretation: a summation. American Association of Petroleum Geologists Bulletin, Vol. 51, pp. 2246-2259. Howard, A.D., Dietrich, W.E., and Seidl, M.A. 1994. Modeling fluvial erosion on regional to continental scales. Journal of Geophysical Research, Vol. 99, No. 13, pp. 971-986. 199 Howes, D.E. and Kenk, E. (Contributing eds.) 1988. MOE Manual 10, Terrain classification system for British Columbia (revised edition). Recreational Fisheries Branch, Ministry of Environment and Surveys and Resource Mapping Branch, Ministry of Crown Lands, Province of British Columbia, 90 pps. Hutchinson, J.N. 1988. General report: Morphological and geotechnical parameters of landslides in relation to geology and hydrology. Proceedings, Vth International Symposium on Landslides, Vol. 1, pp. 3-35. Johnson, K.A., and Sitar, N . 1990. Hydrologic conditions leading to debris-flow initiation. Canadian Geotechnical Journal, Vol. 27, No. 6, pp. 789-801. Jordan, R.P. 1994. Debris flows in the southern Coast Mountains, British Columbia: dynamic behaviour and physical properties. Unpublished Ph.D. thesis, Department of Geography, the University of British Columbia, Vancouver, Canada, 258 pps. Journeay, J.M. 1990. A progress report on the structural and tectonic framework of the southern Coast Belt, British Columbia. Geological Survey of Canada, Paper 90-1E, pp. 183-195. Kamp, P.J.J. 1988. Tectonic geomorphology of the Hikurangi Margin: surface manifestations of different modes of subduction. Z. Geomorph, N.F., Suppl. -Bd. 69, pp. 55-67. Karlinger, M.R. and Troutman, B.M. 1992. Fat fractal scaling of drainage networks from a random spatial network model. Water Resources Research, Vol. 28, No. 7, pp. 1975-1981, July 1992. Kellerhals, R., and Church, M . 1990. Hazard management on fans, with examples from British Columbia. Chapter 17 in Alluvial Fans: a field approach, A.H. Rachocki, and M. Church (eds.), John Wiley and Sons Ltd, pp. 335-354. Kochel, R.C. 1990. Humid fans of the Appalachian Mountains. Chapter 6 in Alluvial fans: a field approach. A.H. Rachocki and M . Church (eds.), John Wiley & Sons Ltd. pp. 109-129 Koons, P.O. 1995. Modeling the topographic evolution of collisional belts. Annual Review of Earth and Planetary Sciences, 1995, 23, pp. 375-408. La Barbera, P., and Rosso, R. 1989. On the fractal dimensions of stream networks. Water Resources Research, Vol. 25, No. 4, pp. 735-741. Lawrence, R.B., Armstrong, R.L., and Berman, R.G. 1984. Garibaldi Group volcanic rocks of the Salal Creek area, southwestern British Columbia; Alkaline lavas on the fringe of a 200 predominantly calc-alkaline Garibaldi (Cascade) Volcanic Arc. Journal of Volcanology and Geothermal Research. Vol. 21, 3-4, pp. 255-276. Leir, M.C. 1994. TREND, a FORTRAN77 program for calculating lineament trends from coordinate endpoint ASCII data. Unpublished computer code, Department of geological Sciences, The University of British Columbia, Vancouver, British Columbia, Canada. Leir, M.C. 1995. Airborne Synthetic Aperture Radar, Digital Terrain Models, and Geographic Information Systems: Tools for mapping and managing large landslide hazards in southwestern British Columbia. Unpublished M.A.Sc. thesis, Geological Engineering Program, Department of Geological Sciences, The University of British Columbia, Vancouver, Canada, 102 pps. Leir, M . C , English, R.R., and Savigny, K.W. 1994. Statistics and GIS: Tools for landslide prediction in the lower Fraser Valley, southwestern British Columbia. Canadian Geotechnical Conference Lillesand, T.H., and Kiefer, R.W. 1994. Remote Sensing and Image Interpretation. John Wiley and Sons Inc. Toronto, Canada, 612 pps. Lubowe, J.K. 1964. Stream junction angles in the dendritic drainage pattern. American Journal of Science, Vol. 262, March 1964, pp. 325-339. Mardia, K.W. 1972. Statistics of directional data. Academic Press, 357 pps. Mauthner, T.E. 1996. Kshwan Glacier rock avalanche, southeast Stewart, British Columbia, in Current Research 1996-A; Geological Survey of Canada, pp. 37-44. Millard, T.H. 1986. Sediment in forested and logged gullies, Coastal British Columbia. Unpublished M.Sc. thesis, Department of Geography, University of British Columbia. 217 pps. Mollard, J.D. 1988. First R.M. Hardy Memorial Lecture: Fracture lineament research and applications on the western Canadian plains. Canadian Geotechnical Journal, Vol. 25, pp. 749-767. Monger, J.W.H. and Journeay, J.M. 1994. Basement geology and tectonic evolution of the Vancouver region, in Geology and Geological Hazards in the Vancouver Region, Southwestern British Columbia, (ed.) J.W.H. Monger, Geological Survey of Canada, Bulletin 481, pp. 3-25. 201 Moore, D.P., and Mathews, W.H. 1978. The Rubble Creek landslide, southwestern British Columbia. Canadian Journal of Earth Sciences, Vol. 15, No. 7, pp. 1039-1052. Nikora, V.I. 1994. On self-similarity and self affinity of drainage basins. Water Resources Research, Vol. 30, No. 1, pp. 133-137, January 1994. Owens, P., and Slaymaker, O. 1992. Late Holocene sediment yields in small alpine and subalpine drainage basins, British Columbia, in Erosion, Debris Flows and Environment in Mountain Regions (Proceedings of the Chengdu Symposium, July 1992), IAHS Publication no. 209, 1992. pp. 147-154. Price, N.J., and Cosgrove, J.W. 1990. Analysis of geological structures. Cambridge University Press, 502 pps. Quari, M.Y.H.T., and Sen, Z. 1994. Remotely sensed fracture patterns in southwestern Saudi Arabia and qualitative analysis. Bulletin of the International Association of Engineering Geology, No. 49, pp. 63-72, April 1994. Rinaldo, A., Vogel, G.K., Rigon, R., and Rodriguez-Iturbe, I. 1995. Can one gauge the shape of a basin? Water Resources Research, Vol. 31, No. 4, pp. 1119-1127, April 1995. Ritter, D.F., Kochel, R.C., and Miller, J.R. 1995. Process Geomorphology, 3rd edition. Wm. C. Brown Publishers, 546 pps. Robert, A., and Roy, A.G. 1990. On the fractal interpretation of the mainstream length-drainage area relationship. Water Resources Research, Vol. 26, No. 5, pp. 839-842. Roddick, J.A. 1965. Vancouver North, Coquitlam, and Pitt Lake map-areas, British Columbia with special emphasis on the evolution of the plutonic rocks. Geological Survey of Canada, Memoir 335, 276 pps. plus maps. Ryder, J.M. 1971a. The stratigraphy and morphology of para-glacial alluvial fans in south-central British Columbia. Canadian Journal of Earth Sciences, Vol. 8, pp. 279-298. Ryder, J.M. 1971b. Some aspects of the morphometry of paraglacial alluvial fans in south-central British Columbia. Canadian Journal of Earth Sciences, Vol. 8, pp. 1252-1264. Ryder, J.M. 1981. Geomorphology of the southern part of the Coast Mountains of British Columbia. Z. Geomorph. N.F., Suppl. -Bd. 37, pp. 120-147. 202 Savigny, K.W. 1996. Engineering geology of large landslides in the lower Fraser Valley transportation corridor southwestern Canadian Cordillera. Canadian Geotechnical Journal (in press). Scalia, P.R. 1995. Terrain analysis: Squamish area. Unpublished Directed Studies Report supervised by Dr. K.W. Savigny, Department of Geological Sciences, University of British Columbia, Vancouver, Canada. 17 pps. Schuster, R.L., Logan, R.L., and Pringle, P.T. 1992. Prehistoric rock avalanches in the Olympic Mountains, Washington. Science, Vol. 258. pp. 1620-1621. Schumm, S. 1954. The relation of drainage basin relief to sediment loss. International Association of scientific Hydrology(?) 36 (1), International Union of Geodesy and Geophysics, 10th General Assembly, Rome, Transactions 1, pp. 216-219. Seidl, M.A., and Dietrich, W.E. 1992. The problem of channel erosion in bedrock, in Catena Supplement 23 (ed.), pp. 101-124. Stanistreet, I.G., and McCarthy, T.S. 1993. The Okavango fan and the classification of subaerial fan systems. Sedimentary Geology, 85, pp. 115-133. Strahler, A.N. 1952. Dynamic basis of geomorphology. Geological Society of America Bulletin, 63, pp. 923-938. Sun, T., Meakin, P., and Jossang, T. 1994b. The topography of optimal drainage basins. Water Resources Research, Vol. 30, No. 9, pp. 2599-2610, September 1994. Takahashi, T. 1993. Debris flow initiation and termination in a gully. Hydraulic Engineering 1993 Hsieh wen Shen and others (eds.). Vol. 2, pp. 1756-1761. Thurber Engineering Ltd. 1983. Debris torrent and flooding hazards, Highway 99, Howe Sound. Report to British Columbia Ministry of Transportation and Highways. 25 pps. + appendices. Thurber Engineering Ltd. and Golder Associates Ltd. 1993. Cheekye River terrain hazard and land use study; Final Report, Volume 1. 146 pages + Tables, Plates and Figures. Thurber Engineering Ltd. 1985. Westwood Plateau area escarpment and gravel study - phase 2: report to Ministry of Lands, Parks and Housing Land Development Branch. Thorn, C.E. 1988. Introduction to theoretical geomorphology. Allen & Unwin, Inc., 247 pps. 203 VanDine, D.F. 1984. Debris flows and debris torrents in the Southern Canadian Cordillera. Canadian Geotechnical Journal, Vol. 22, pp. 44-68. Wertz, J.B. 1968. Structural elements of ore search in the Basin and Range province, southeast Arizona: domes and fracture intersections. American Society of Mining Engineers, Transactions, Vol. 241. pp. 276-291, September 1968. Wertz, J.B. 1974. Detection and significance of lineaments and lineament intersections in parts of the northern Cordillera. Proceedings of the First International Conference on the New Basement Tectonics, A. Robert, S. Hodgson, Parker Gay Jr., and J.Y. Benjamins (eds.), Salt Lake City, Utah, June 3-7, 1974, pp.42-53. Whipple, K.X. 1993. Interpreting debris-flow hazard from study of fan morphology. Hydraulic Engineering '93, Hsieh Wen Shen and others (eds.). Vol. 2, pp. 1302-1307. Wieczorek, G.F., Elliott, W.L., and Ellen, S.D. 1989. Debris Flows and hyperconcentrated floods along the Wasatch Front, Utah, 1983 and 1984. Bulletin of the Association of Engineering Geologists, Vol. XXVI, No. 2, 1989, pp. 191-208. Woldenberg, M.J. 1969. Spatial order in fluvial systems: Horton's laws derived from mixed hexagonal hierarchies of drainage basin areas. Geological Society of America Bulletin, Vol. 80, pp. 97-112, January 1969. 204 APPENDIX I Frequency Tables for Lineament Trend Data 205 Distribution Frequency % 0-4 156 3.70 5-9 244 5.79 10-14 272 6.45 15-19 277 6.57 20-24 251 5.95 25-29 242 5.74 30-34 212 5.03 35-39 151 3.58 40-44 164 3.89 45-49 122 2.89 50-54 99 2.35 55-59 92 2.18 60-64 69 1.64 65-69 55 1.30 70-74 56 1.33 75-79 62 1.47 80-84 72 1.71 85-89 84 1.99 90-94 107 2.54 95-99 82 1.95 100-104 65 1.54 105-109 72 1.71 110-114 44 1.04 115-119 55 1.30 120-124 55 1.30 125-129 74 1.76 130-134 85 2.02 135-139 102 2.42 140-144 89 2.11 145-149 88 2.09 150-154 90 2.14 155-159 87 2.06 160-164 91 2.16 165-169 106 2.51 170-174 125 2.97 175-179 118 2.80 Total 4215 100.00 Frequency - distribution for all lineaments in the study area. 206 Frequency Distribution Southern Central Block Northeast Northwest Block Block Block 0-4 19 90 18 33 5-9 28 166 27 32 10-14 36 171 29 40 15-19 34 184 25 43 20-24 27 164 26 38 25-29 37 162 18 27 30-34 30 136 26 24 35-39 25 100 10 17 40-44 35 100 12 17 45-49 11 74 16 22 50-54 12 69 11 10 55-59 7 62 11 12 60-64 10 42 7 9 65-69 6 39 6 6 70-74 5 32 7 11 75-79 3 36 11 12 80-84 7 49 7 9 85-89 8 60 4 12 90-94 14 68 10 15 95-99 8 62 4 8 100-104 8 45 8 4 105-109 11 41 15 7 110-114 5 27 6 6 115-119 7 32 7 8 120-124 8 23 21 3 125-129 2 37 26 9 130-134 4 39 30 14 135-139 5 49 33 15 140-144 5 46 28 11 145-149 5 36 34 13 150-154 5 44 21 22 155-159 7 28 29 25 160-164 7 52 22 13 165-169 13 53 23 21 170-174 20 69 16 22 175-179 13 59 21 24 Total 487 2546 625 614 Frequency - distribution for lineaments in the sample blocks. 207 APPENDIX II Statistical Methods 208 Test for randomness We must assume a Von Mises distribution. The null hypothesis and its alternative are therefore: Ho:k=0 H,:k>0 The value of R is calculated using either ROSE 1.02 or methods described in Davis (1986). This is compared to a critical value at the desired level of significance obtained from table 5.7 in Davis (1986, pp. 324). If R is greater than the critical value the null hypothesis is rejected and the observations are assumed to come from a population with a preferred orientation (Davis 1986). In the case of large values of 'n' the critical value is distributed as chi-square with 2 degrees of freedom and is approximated by 2nR2 (Mardia 1972). Example calculation: For the southern block of lineaments described on page 36. The sample size V is 487. R=0.4296. 2nR2 = 2.487 x 0 x 4296 =179.76 The critical value from tables of Chi-square is found to be 5.99. Because 179.76 is greater than 5.99, the null hypothesis is rejected and a preferred trend is implied. 209 Calculation of standard error A confidence angle about the mean direction of the sample data can be calculated based on the standard error of the estimate of mean direction. In this fashion both the size of the sample and the dispersion are taken into account. First R is calculated and K estimated from tables (e.g. Davis 1986, p.323). Then the standard error is calculated according to the following formula; Se = l/(nRK) , / 2 This gives a result in radians which must be converted to degrees. The calculated value is a measure of the chance variation expected in the sample in estimates of the mean direction. Assuming the estimation errors are normally distributed; 6>±Za.Se should capture the true population mean a % of the time. Za is determined from standard normal distribution tables. The value is 1.96 for the 95% confidence level. Example calculation: For the entire lineament dataset n = 4215, R= 0.319 and K is estimated to be 0.67587. Hence: Se= 1/(4215 x0.319x0.67587)m = 0.0331722 rads. = 1.9° Thus Za.Se = 3.725°. 210 Test for equality of two sets of directional data The equality of two mean directions can be tested by comparing the vector resultants of the two groups to that produced when the two groups are combined. If the two groups are drawn from the same population the resultant of the pooled sample should be approximately equal to the sum of their two resultants (Davis 1986). If the mean directions are significantly different then the pooled resultant will be shorter than the sum of their resultants. R is calculated using the following formula: R = (R, + R2)/n The hypothesis being tested is: H o ^ A u U r v , ) = (XTHIVJ) .HI : (AI /* ,V , ) * Q.^i2v2) Example calculation: To test the equality of the southern and central blocks described on page 35. R t = 209.2 (for the southern data) and R 2 = 895.45 (for the central data), n = 3033; the sample size of the pooled dataset. R =.(209.2 + 895.45)/3033 = 0.3642 This value is compared to that calculated by ROSE 1.02 for the pooled dataset. This value is 0.3642. Since the two values are equal it is concluded that they are from the same population. More rigorous testing is provided by the Watson - Williams test described by Mardia (1972). 211 APPENDIX III Location and Description of Sample Set Basins 212 a. W u J3 4s « 5 . _ O • -C G O - i f S o o I s? 0 0 > to lack of usceptible cj fc *S al, littl latedc one SU| > •-3 V Large colluvial c vegetation would avalanche. Cone mostly coll may be avalanch been logged. E 6 > c £ u •3 S \" 8 3 E >• £ c —; O 4) 2 w > E T3 « •s > -e a S 3 2 «f \"c ^ n - \" i on «J > en wi — 2 & o > > , 3 .2! « 1 -a Q O « ^ E _ tf ° a g x. § 0 . 3 CQ S o o 00 E g •o c w o 1 .E •d .£ :epl icti b e C J es Undi; — o. c rt 3 2 X) u « XI 2 3 3 in « c S ed I Jj 3 8 s > £ it c xi : l ' ~ \"O 1 x: E C5 o o Cu, c o T3-T3 e C o CO o -1 o H D CQ 5t 0O v 2 aa » oa % oo n aa * aa * oa % 0 0 V I CQ » O 2 CQ i t CO 3t u. O O O ON a OS O 3 'C X co C O i- oi s e 1? s co o J - o 1 < J < Oi i i 3 3 o 3 8 ON O ? is 0 C S ^ I 213 c c Z Q c 1 G -2 •s 2 a 5 ° 2 .s § g § I S *o « — 31 60 ^ > a °^ s 3 S t 8.: 3 3 e 8- S 5 >» « .H y> rt = o \"5 c ' 5 « •O -rn • § 1 2 c£ u 5T « 2 E w O \" S'F j -S .> .0 a wT to u U S ^3 C Si *« o 6 H J5 rt C L , D O O 0 0 °1 : 1 ; w on 5 3 I O rt I'l l 0 0 o 1 s 3 S t 5 •C vr i - ? U S--o E E S bo .•2 >. i~ a. 3 o 3 Hi X> C 214 IS £. I 1 & £ C ± I 3 2 2 i 2 ^ 3 J3d E .2 O O O N E -5 ti5 E \"8 « o p c OO CQ at o g 3 5 3 -6 00 ™ — 1 \" S > 5 -d * -s GO O > 4 i 1> - C K g sz 5 S3 u « t i CQ at -o .S •S 8 U « = K - c O O 3 CQ at O i_ rj 3 > J 3 L O 1-5 E S o. •= 11 •S -2 C XL P o 0 ^ w 1 • • o f §8*; i l . f i C C U ^ CQ at •= 2 8\" « « 3 Q oo 00 ON CQ at ON O •a a •« ^ oo O C3 c oo ra D a — O CO *-• o -^ *-» a ^ S « \" N ! O CTJ . — w < / > — « = \" r « i u J ; ^ I- « 6 0 \" O \"oo C 73 > .S3 « __ ™ o .2 6 o > -a in ? o 3 c l « p u ej H o > a oo ^ o S ° o o « r-. « oo i2 M o) .2 -5 Jj a s o c •— 5 \"> 5 2 c c X) o « c - 3 > C oo CJ E o •o x: H 2 -a o c o o cx, (/) £ o TO on - C z i2 5 £ 2 2 1 5 APPENDIX I V Tabulated Sample Set Data 216 CI o o i S o i« o 3 2 I- g 2 Q U •a S« u OS D 15 I t Q u < u o t/3 217 c o •S 8. 8 9S b £ J! US \"s „ |« « ~ I -a £ < < 4! o o o O H «2 c. n. E o a. I— 00 o o. o f— O 'CL O H | — O H 10 2 \\< i E en o o O O H O H 8 5 E o s I-i T3 D . 3 \"a \"£j 03 3 O c o U a .S II S3 o t> — • \" 9- on I 2 M Q •SS2 •a n 3 5 05 « o 1= ? o O H D < O o 3 < 3 < -a c o oo _o u 3 o on 218 Cone U Geology Aspect 1 453 Granodiorite. east-southeast 2 174 Quartz diorite. northeast 3 205 Quartz diorite. northwest 4 248 Granodiorite. northwest 5 281 Migmatite west 6 639 Quartz diorite. south-southwest 7 221 Granodiorite. southeast 8 470 Granodiorite. southeast 9 314 Quartz diorite. north-northeast 10 162 Quartz diorite. west-northwest 14 41 Twin Island/Gambier Groups southeast 15 18 Quartz diorite. east-southeast 16 45 Quartz diorite. southeast 17 445 Quartz diorite. southwest 18 53 Quartz diorite. east-southeast 19 285 Fire Lake Group southeast 20 305 Quartz diorite. north 21 186 Gambier Group northeast 22 484 Quartz diorite. west 24 303 Quartz diorite. southwest 25 89 Quartz diorite. northwest 26 99 Granodiorite. northeast 27 153 Quartz diorite. southwest 28 71 Quartz diorite. west 29 229 Migmatite, quartz diorite, Harrison Lake Formation, and Twin Island Group south-southwest SOURCE GEOLOGIC MAP TOPOGRAPHIC MAP Thematic Information for sample set basins. 219 APPENDIX V Results of spatial correlation analysis between lineaments and streams 220 DATA FOR A L L STREAM ORDERS Overlying lineaments Correlation Basins Frequency % Cumulative % % Frequency 40-49 2,3,4,6,9,20 6 24% 6 24% 30-39 7,18,21,25,28 5 20% 11 44% 20-29 1,10,14,22,26 5 20% 16 64% 10-19 5,15,17,24,27,29 6 24% 22 88% 0-9 8,16,19 3 12% 25 100% Within 40m Correlation Basins Frequency % Cumulative % % Frequency 90-99 2 1 4% 1 4% 80-89 20,25 2 8% 3 12% 70-79 6,4,26,28 4 16% 7 28% 60-69 1,3,9,10,15,17,27 7 28% 14 56% 50-59 7,14,18,21,22 5 20% 19 76% 40-49 0 0% 19 76% 30-39 8 1 4% 20 80% 20-29 5,24,29 3 12% 23 92% 10-19 16,19 2 8% 25 100%. Within 60m Correlation Basins Frequency % Cumulative % % Frequency 90-99 2,25 2 8% 2 8% 80-89 4,6,10,17,20,26 6 24% 8 32% 70-79 1,3,9,15,27,28 6 24% 14 56% 60-69 7,14,18,21 4 16% 18 72% 50-59 8,22 2 8% 20 80% 40-49 0 0% 20 80% 30-39 5,24,29 3 12% 23 92% 20-29 19 1 4% 24 96% 10-19 16 1 4% 25 100% Results of overlay of streams on lineaments: All stream orders. 221 DATA FOR 1ST. ORDER STREAMS Correlation Basins Frequency % Cumulative % % Frequency 60-69 6 1 4% 1 4% 50-59 2,4 2 8% 3 12% 40-49 3,20,25 3 12% 6 24% 30-39 9,21,28 3 12% 9 36% 20-29 1,7,10,14,18,22,26 7 28% 16 64% 10-19 5,8,15,17,24,27,29 7 28% 23 92% 0-9 16,19 2 8% 25 100% Within 40m Correlation Basins Frequency % Cumulative % % Frequency 100 2 1 4% 1 4% 90-99 4,25 2 8% 3 12% 80-89 6,20,26 3 12% 6 24% 70-79 3,17,28 3 12% 9 36% 60-69 1,8,10,15,22 5 20% 14 56% 50-59 7,21,27 3 12% 17 68% 40-49 9,14,18 3 12% 20 80% 30-39 5 1 4% 21 84% 20-29 19,24,29 3 12% 24 96% 10-19 16 1 4% 25 100% Within 60m Correlation Basins Frequency % Cumulative % % Frequency 100 2,4 2 8% 2 8% 90-99 25 1 4% 3 12% 80-89 6,10,17,20,26 5 20% 8 32% 70-79 1,3,22,27,28 5 20% 13 52% 60-69 7,8,14,15,21 5 20% 18 72% 50-59 9,18 2 8% 20 80% 40-49 5 1 4% 21 84% 30-39 19 1 4% 22 88% 20-29 24,29 2 8% 24 96% 10-19 16 1 4% 25 100% Results of overlay of streams on lineaments: First order streams. 222 DATA FOR 2ND. ORDER STREAMS Overlying lineaments Correlation Basins Frequency % Cumulative % % Frequency 60-69 2 1 4% 1 4% 50-59 7,9,20 3 12% 4 16% 40-49 0 0% 4 16% 30-39 18,21,26 3 12% 7 28% 20-29 3,6,10,14,15,17,25,27,28 9 36% 16 64% 10-19 1,24,29 3 12% 19 76% 0-9 4,5,8,16,19,22 6 24% 25 100% Within 40m Correlation Basins Frequency % Cumulative % % Frequency 100 2 1 4% 1 4% 90-99 1,9 2 8% 3 12% 80-89 20 1 4% 4 16% 70-79 25,28 2 8% 6 24% 60-69 6,7,15,17,18,27 6 24% 12 48% 50-59 10,14,26 3 12% 15 60% 40-49 21,29 2 8% 17 68% 30-39 3,24 2 8% 19 76% 20-29 4 1 4% 20 80% 10-19 19,5,22 3 12% 23 92% 0-9 8,16 .2 8% 25 100% Within 60m Correlation Basins Frequency % Cumulative % % Frequency 100 1,2,9 3 12% 3 12% 90-99 15,25 2 8% 5 20% 80-89 17,20 2 8% 7 28% 70-79 6,7,14,18,26,27,28 7 28% 14 56% 60-69 10,21 2 8% 16 64% 50-59 29 1 4% 17 68% 40^J9 3 1 4% 18 72% 30-39 4,5,24 3 12% 21 84% 20-29 8,22 2 8% 23 92% 10-19 16,19 2 8% 25 100% Results of overlay of streams on lineaments: Second order streams. 223 DATA FOR 3RD. ORDER STREAMS Overlying lineaments Correlation Basins Frequency % Cumulative % % Frequency 40-49 18 1 10% 1 10% 30-39 0 0% 1 10% 20-29 14 1 10% 2 20% 10-19 1,10 2 20% 4 40% 0-9 2,5,16,19,22,29 6 60% 10 100% Within 40m Correlation Basins Frequency % Cumulative % % Frequency 60-69 18 1 10% 1 10% 50-59 10 1 10% 2 20% 40-49 ; 1,14,22 3 30% 5 50% 30-39 0 0% 5 50% 20-29 2 1 10% 6 60% 10-19 16 1 10% 7 70% 0-9 5,19,29 3 30% 10 100% Within 60m Correlation Basins Frequency % Cumulative % % Frequency 70-79 18 1 10% 1 10% 60-69 10 1 10% 2 20% 50-59 14,22 2 20% 4 40% 40-49 1 1 10% 5 50% 30-39 16 1 10% 6 60% 20-29 2 1 10% 7 70% 10-19 0 0% 7 70% 0-9 5,19,29 3 30% 10 100% Results of overlay of streams on lineaments: Third order streams. 224 B a s i n « 1 Stream Order Length % 1«t_ 2nd 3rd l o t 920 16.97% 24.73% 2nd 0 0 00% 0.00% 3rd 4*0 8 8 6 % 51.06% 4 l h 0 0.00% Isolated 1400 25.83% 1st / l in 940 17.34% 25 27% 2nd/ l ln 100 1.85% 13.16% 12.77% 3rd/1in 120 2.21% 4th / l in 0 0.00% Over iy inq L i n t . 1160 21.40% 1«1*40 1360 25.09% 36.56% 2nd«40 620 11.44%| 61.83% 81.58% 3rdi-40 320 5.90% 94 74% 34 04% 41h*40 0 0.00% 4 6 8 1 % l With in 40 m 2300 42.44% 1st««0 600 9.23% 13.44% 2nd*«0 40 0.74%| 75.27% 5.26% 3rd*60 20 0.37% 100.00% 2.13% 4lh>«0 0 0.00% 48.94%| Wi th in 60 m 560 10.33% Stream Length 5420 S o n Una. H w i t h l n 40m 2 1 * . 64% 3720 760 940 S w r t h i n 60m 74% % o c c u p i e d l ins . 48% B a s i n tt 2 Stream Order Length % IKL 2nd 3rd 1st 0 0.00% 0.00% 2nd 0 0.00% 0.00% 3rd 420 7.42% 72.41% 4th 0 0.00% Isolated 420 7.42% 1at / l in 1920 33 92% 51.61% 2nd/1in 860 15.19% 63.24% 3rd/ l in 0 0 00% 0.00% 4tt i / l in 0 0.00% Over iy inq L i n s . 2780 49.12% 1*1*40 1800 31.80% 48.39% 2nd*40 500 8.83% | 100.00% 36.76% 3rd+40 120 2.12% 100.00% 20 69% 4th+40 0 0.00% 20.69% I Wi th in 40 m 2420 42.76% 1st*«0 0 0.00% 0.00% 2nd*60 0 0.00% | 100.00% 0.00% 3rd«60 40 0.71% 100.00% 6.90% 4th«60 0 0.00% 27.59%l W i th in 60 m 40 0.71% Stream Length 5660 %on l ine. S w H h l n 40m 49% 92% 3720 1360 580 S w r t h i n 60m 9 3 % % o c c u p i e d l ine. 56% B a s i n # 3 Stream Order Length % 1st. 2nd 1st 1080 19.15% 23.48% 2nd 540 9.57% 51.92% 3rd 0 0.00% 4th 0. 0.00% Isolated 1620 28.72% 1st/I in 2060 36.52% 44.78% 2ndfl in 240 4.26% 23.08% 3rdflin •o 0.00% 4th / l in 0 0.00% Overiyinq L ins . 2300 40.78% 1st«40 1240 21.99% 26,96% 2nd*40 160 2.84%| 71.74% 15.38% 3rd . 4 0 0 0.00% 38.46%| 4th«40 0 0.00% With in 40 m 1400 24.82% 1.1.60 220 3.90% 4.78% 2nd«60 100 1.77%| 76.52% 9.62% 3rd . 6 0 0 0.00% 48.08% I 4th*«0 0 0.00% With in 60 m 320 5.67% Stream Length %on l ine. %wrthin 40m 5640 4 1 % 66% 4600 1040 %wi th in 60m 7 1 % %occup ied l ins . 47% B a s i n #4 Stream Order Length % 1 s t 2nd 1st 0 0.00% 0.00% 2nd 280 18.67% 66.67% 3rd 0 0.00% 4th 0 0.00% Isolated 280 18.67% 1st / l in 620 41.33% 57.41% 2nd / l in 20 1.33% 4.76% 3rd(lin 0 0.00% 4tWl in 0 0.00% Over iy inq L i n s . 640 42.67% 1«t*40 400 26.67% 37.04% 2nd«40 80 5.33% | 94.44% 19.05% 3rd»40 0 0.00% 23 81%) 41h*40 0 0.00% Wi th in 40 m 480 32.00% 1st*60 60 4.00% 5.56% 2nd*60 40 2.67% | 100.00% 9.52% 3rd . 6 0 0 0.00% 33.33% I 4th*60 0 0.00% Wi th in 60 m 100 6.67% Streem Length %on l ins . 1500 43% S w K h i n 40m 75% 1080 420 %wi th in 60m 8 1 % % o c c u p i e d l ine. 36% B a s i n # 5 Stream Order Length % 1 s t 2nd 3rd 1st 5460 34.51% 54.60% 2nd 2200 13.91% 66.27% 3rd 2340 14.79% 93.60% 4th 0 0.00% Isolated 10000 63.21% 1st / l in 1900 12.01% 19.00% 2ndfl in 120 0.76% 3 .61% 3rd/ l in 0 0.00% 0.00% 4th/ l in 0 0.00% Overiy inq L i n s . 2020 12.77% 1st«40 1760 11.13% 17.60% 2nd*40 500 3.16%| 36.60% 15.06% 3rd . 40 60 0.38% 18.67% 2.40% 4th«40 0 0.00% 2.40% With in 40 m 2320 14.66% 1st»«0 880 5.56% 8 80% 2nd<«0 500 3.16%| 45.40% 15.06% 3rd»«0 100 0 6 3 % 3 3 7 3 % 4.00% 4th*60 0 0 0 0 % 6.40% With in 60 m 1480 9 36% Stream Length 15820 V o n l ins . %w«hln 40m 13% 27% 10000 3320 2500 %wfthln 60m 37% %occupi«d l ins . 15% B a s i n # 6 Stream Order Length % 1 s t 2nd 1st 120 6.25% 12.77% 2nd 260 13.54% 26.53% 3rd 0 0.00% 4th 0 0.00% Isolated 380 19.79% 1st/1in 580 30.21% 61.70% 2ndf l in 220 11.46% 22.45% 3rd / l in 0 0.00% 4thf l in 0 0.00% Over iy inq L i n s . 800 41.67% 1 s H 4 0 200 10.42% 21.28% 2nd*40 400 20.83% | 82.98% 40.82% 3rd*40 0 0.00% 63.27% 4th«40 0 0.00% Wi th in 40 m 600 31.25% 1st . 6 0 40 2.08% 4.26% 2 n d . 6 0 100 5 .21%| 87.23% 10.20% 3rd . 6 0 0 0.00% 73.47% 4th«60 0 0.00% Wi th in 60 m 140 7.29% Stream Length %on l ins . 1920 42% % w H h i n 4 0 m 73% 940 980 % w i t h i n 60m 80% % o c c u p i e d l ins . 84% Results of spatial correlations between lineaments and streams for individual basins. (lins. = lineaments) 225 B a s i n « 7 6 . . i n « i 2nd Stream Order l e n g t h % 1«t 2nd Stream Order Length % 1 s t 1«t 620 2 0 8 1 % 38.75% l e t 200 15 63% 30.30% 70.97% 2nd 380 1275% 27.54% 2nd 440 34 38% 3rd 0 0 0 0 % 3rd 0 000% 4th 0 0 0 0 % 4th 0 0 00% Isolated 1000 33 56% Isolated 640 50 00% 1st / l in 360 12 08% 22.50% 1*1/lln 60 6.25% 12.12% 2nd/1in 700 2 3 4 9 % 50.72% 2nd/1in 0 0.00% 0 0 0 % 3rd/ l in 0 0 00% 3rd/1in 0 0.00% 4th / l in 0 000% 4th/1in 0 0.00% Over ly inq L i n * . 1060 35.57% Over ly inq U n a . 80 6.25% 1e1*40 460 1544% 28 75% 1st . 4 0 340 26.56% 51 52% 2nd«40 240 8.05% | 51.25% 17.39% 2nd *40 60 4.69% 63.64% 9.68% 3rd«40 0 0 00% 68.12%l Jrf»40 0 0.00% 9.68% I 41h*40 0 0.00% 4th<40 0 000% With in 40 m 700 23.49% Wi th in 40 m 400 31.25% 1et»80 160 5.37% 1 0 0 0 % 1st««0 40 3.13% 6 0 6 % 2 n d . 6 0 60 2.01%| 61.25% 4 35% 2 n d - 6 0 120 9.38% 69.70% 19.35% 3rd*«0 0 0 0 0 % 72.46%| 3rd««0 0 0.00% 29 0 3 % | 4th*60 0 0 0 0 % 4thf«0 0 0.00% Wi th in 60 m 220 7.38% Wi th in 60 m 160 12.50% Stream Length 2980 Stream L e n g t h 1280 200.00% V o n l ins . 3614 %on Una. 6% S w i t h i n 4 0 m 5914 1600 1380 % wi th in 40m 38% 660 620 S w i t h i n 60m 6614 S w i t h i n 60m 50% %occup(ed l ins . 3714 % o c c u p i e d l ins . 42% B a s i n #9 B a s i n tt 10 Stream Order Length % 1 s t 2nd Stream Order Length % 1 s t 2nd 3rd 1st 780 29.77% 46.99% 1st 1540 9.99% 13.87% 2nd 0 0 00% 0.00% 2nd 680 4.41% 31.78% 3rd 0 0.00% 3rd 740 4.80% 33.94% 4th 0 0.00% 4th 0 0.00% isolated 780 29.77% Isolated 2960 19.20% 1st / l in 600 19.08% 30.12% 1et / l in 2780 18.03% 25.05% 2nd / l in 540 20.61% 56.25% 2nd / l in 620 4.02% 28.97% 3rdnin 0 0.00% 3rd / l in 240 1.56% 11.01% 4th / l in 0 0.00% 4th/1in 0 0 0 0 % Over ly inq L i n s . 1040 39.69% Over ly inq L i n s . 3640 2 3 6 1 % 1st*40 300 11.45% 18,07% 1st*40 4620 29.96% 41.62% 2nd«40 400 15.27% 48.19% 41.67% 2nd»40 640 4.15% 66.67% 29.91% 3rd*40 0 0.00% 97.92% 3rd*40 880 5.71% 58.88% 40.37% 4th*40 0 0.00% 4th*40 0 0.00% 51.38% Wi th in 40 m 700 26.72% Wi th in 40 m 6140 39.82% 1st*«0 so 3.05% 4.82% 1st*60 2160 14.01% 19.46% 2nd*«0 20 0.76% 53.01% 2.08% 2nd*«0 200 1.30% 86.13% 9.35% 3rd*60 0 0.00% 100.00%| 3rd 460 320 2.08% 68.22% 14.68% 4th»60 0 0.00% 41h460 0 0.00% 66.06% With in 60 m 100 3.82% Wi th in 60 m 2680 17.38% Stream Length 2620 200.00% Stream Length 15420 200.00% %on l ins . 40% %on l i ns . 24% % w i t h i n 40m 66% 1660 960 S w i t h i n 40m 63% 11100 2140 2180 % w i t h i n 60m 70% S w i t h i n 60m 8 1 % ^ o c c u p i e d l ins . 100% % o c c u p i e d l i ns . 56% B a s i n # 14 B a s i n #15 Stream Order Length % 1 s t 2nd 3rd Stream Order Length % 1 s t 2nd 1st 1720 18.42% 33.59% 1st 1040 18.18% 30.41% 2nd 520 5.57% 21.49% 2nd 220 3.85% 9.57% 3rd 900 9.64% 50.00% 3rd 0 0.00% 4th 0 0.00% 4th 0 0.00% Isolated 3140 33.62% Isolated 1260 22.03% 1st / l in 1100 11.78% 21.48% 1 st / l in 640 11.19% 18.71% 2ndfl in 480 5.14% 19.83% 2nd / l in 460 8.04% 20.00% 3rd/ l in 360 3.85% 20.00% 3rd / l in 0 0.00% 4th/)in 0 0.00% 4th / l i n 0 0.00% Over ly inq L ins . 1940 20.77% Over ly inq L ine . 1100 19.23% 1st*40 1380 14.78% 26.95% 1ai*40 1460 25.52% 42.69% 2nd*40 960 10.28%| 48.44% 39.67% 2nd 440 940 16.43%| 61.40% 40,87% 3rd*40 480 5.14% 59,50% 2 6 6 7 % 3rd 440 0 0.00% 60.87% | 4th«40 0 0.00% 4 6 6 7 % | 4th»40 0 0.00% With in 40 m 2820 30.19% With in 40 m 2400 41.96% 1st«60 920 9.85% 17.97% 1et*60 280 4.90% 8,19% 2nd*60 460 4.93% | 66.41% 19.01% 2nd««0 680 11.89%| 69.59% 29.57% 3rd»«0 60 0 64% 78,51% 3.33% 3rd .60 0 0.00% 90.43% | 4th4«0 0 0 00% 50 00%| 4th460 0 0 0 0 % Wi th in 60 m 1440 15,42% Wi th in 60 m 960 16.78% Stream Length 9340 200 00% Stream Length 5720 200 00% V o n l ins . %wfthin 40m 21% 51% 5120 2420 1800 %on l i ns . % w r t h i n 4 0 m 19% 6 1 % 3420 2300 %wtthin 60m % o c c u p i e d line. 66% 48% X w i t h l n «0m % o c c u p i e d l ine . 78% 58% Results of spatial correlations between lineaments and streams for individual basins (Continued). (lins. = lineaments) 226 B a s i n U 16 Stream Order Length % 1 s t 2nd 3rd 1at 1260 40.91% 82.89% 2nd 800 25.97% 86.96% 3rd 440 14.29% 68.75% 4th 0 0.00% Isolated 2500 81.17% 1et/l in 20 0.65% 1.32% 2nd/ l in 0 0.00% 0.00% 3rd/ l in 20 0.65% 3.13% 4 t M i n 0 0.00% Over iy inq Line. 40 1.30% 1st+40 100 3.25% 6 58% 2nd+40 20 0.65%| 7.89% 2 1 7 % 3rd+40 100 3.25% 2.17% 15.63% 4th+40 0 0.00% 18.75%l With in 40 m 220 7.14% 1st+80 140 4.55% 9.21% 2nd+60 100 3.25% | 17.11% 10.87% 3rd+60 80 2.60% 13.04% 1 2 5 0 % 4th+60 0 0.00% 31.25%| Wi th in 60 m 320 10.39% Stream Length 3080 200.00% S o n l ine . S w i t h i n 40m 1% 8 % 1520 920 640 S w i t h i n 60m 19% S o c c u p i e d l ins . 24% B a s i n # 17 Stream Order Length % l e t 2nd 1st 200 6.54% 10.64% 2nd 140 4.58% 11.86% 3rd 0 0.00% 4th 0 0.00% reolated 340 11.11% 1st/t in 320 10.46% 17.02% 2nd/ l in 260 8.50% 22 0 3 % Srdflin 0 0.00% 4th/1ln 0 0.00% Over iy inq U n s . 580 18.95% 1st +40 1020 33.33% 54.26% 2nd+*0 520 16.99% | 71.28% 44.07% 3rd*40 0 0.00% 66.10% 4th+40 0 0.00% With in 40 m 1540 50.33% 1st+60 340 11.11% 18.09% 2nd+60 260 8.50% | 89.36% 22 0 3 % 3rd+60 0 0.00% 88.14% 4th+60 0 0.00% With in 60 m 600 19.61% Stream Length < 3060 200.00% %on l ine . 19% S w i t h i n 40m 6 9 % 1880 1180 S w i t h i n 60m 8 9 % S o c c u p i e d l ins . 6 2 % B a s i n # 18 B a s i n * 19 Stream Order Length S 1 s t 2nd 3rd Stream Order Length S 1 s t 1st 2700 26.01% 45.00% 1st 34560 39.89% 63.20% 2nd 700 6.74% 28.46% 2nd 9080 10.48% 3rd 460 4.43% 23.96% 3rd 10940 12.63% 4th 0 0.00% 4th 9360 10.80% Isolated 3860 37.19% Isolated 63940 73.80% Istyiin 1480 14.26% 24.67% 1st / l in 3460 3 9 9 % 6.33% 2nd/ l in 900 8.67% 36.59% 2nd/1in 680 0.78% 3rd/ l in . 780 7.51% 40.63% 3rdflin 40 0.05% 4th / l in 0 0.00% 4th / i in 40 0.05% Overiy inq L i n s . 3160 30.44% Over iy inq L ins . 4220 4.87% 1st+40 1360 13.10% 22.67% 1et+40 11280 13.02% 20.63% 2nd+40 700 6.74%| 47.33% 28.46% 2nd+40 520 0.60% | 26.96% 3rd+40 520 5.01% 65.04% 27.08% 3rd+40 280 0.32% 4th+40 0 0.00% 67.71%| 4th+40 240 0.28% Within 40 m 2580 24.86% With in 40 m 12320 14.22% 1st+80 460 4.43% 7.67% 1st+60 5380 6.21% 9.84% 2nd+60 160 1.54%| 55.00% 6.50% 2nd+60 340 0.39% | 36.80% 3rd+60 160 1.54% 71.54% 8.33% 3rd+60 260 0.30% 4th+60 0 0.00% 76.04%| 4th+60 180 0.21% Within 60 m 780 7.51% Within 60 m 6160 7.11% Stream Length 10380 200.00% Stream Length 86640 200.00% S o n t ins. S w i t h i n 40m S w i t h i n 60m S o c c u p i e d l ins . 30% 55% 6 3 % 5 5 % 6000 2460 1920 S o n l ins . S w i t h i n 40m S w i t h i n 60m S o c c u p i e d l ins . 5 % 19% 26% 15% 54680 10620 11520 B a s i n ft 20 Stream Order Length S 1 s t 2nd 1st 140 7.69% 17.50% 2nd 180 9.89% 17.65% 3rd 0 0.00% 4th 0 0.00% isolated 320 17.58% 1st / l in 320 17.58% 40.00% 2nd/1in 520 28.57% 50.98% 3rdflin 0 0.00% 4th / l in 0 0.00% Overiy inq L ins . 840 46.15% 1st440 320 17 58% 40.00% 2nd+40 300 16.48%| 80.00% 29.41% 3rd . 4 0 0 0.00% 80.39%| 41h+40 0 0.00% With in 40 m 620 34.07% 1st+60 20 1.10% 2.50% 2nd+60 20 1.10%| 82.50% 1.96% 3rd»60 0 0.00% 82.35% | 41h»60 0 0.00% With in 60 m 40 2.20% Stream Length 1820 200.00% S o n t ins. 46% S w i t h i n 40m 80% 800 1020 S w i t h i n 60m 8 2 % S o c c u p i e d l ins . 7 7 % B a s i n #21 Stream Order Length S 1st. 2nd 1st 1560 22.87% 30.83% 2nd 600 8.80% 34.09% 3rd 0 0.00% 4th 0 0.00% Isolated 2160 31.67% 1st/ l in 1580 23.17% 31.23% 2nd/ l in 520 7.62% 29.55% 3rd/ l in 0 0.00% 4th / l in 0 0.00% Overiy inq Line. 2100 30.79% 1st+40 1360 19.94% 26.88% 2nd+40 240 3.52% | 58.10% 13.64% 3rd+40 0 0.00% 43.18% 4th+40 0 0.00% With in 40 m 1600 23.46% 1st+60 560 8.21% 11.07% 2nd+60 400 5.87% | 69.17% 22.73% 3rd .60 0 0.00% 65.91% 4th+60 0 0.00% With in 60 m 960 14.08% Stream Length 6820 200.00% S o n l ins . 3 1 % S w i t h i n 40m 54% 5060 1760 S w i t h i n 60m 6 8 % S o c c u p i e d l ins . 49% Results of spatial correlations between lineaments and streams for individual basins (Continued) (lins. = lineaments) 227 B e . I n « 22 3rd Stream Order Length V 1 s t 2nd 1.1 2400 1869% 27.71% 7 6 7 0 % 2nd 2700 21.03% 42.42% 3rd 280 2 18% 4th 0 0 0 0 % Isolated 5380 41.90% le t / t in 2540 1978% 29.33% 2nd/1in 60 0.47% 1.70% 9 0 9 V 3rd/ l in 60 0.47% 4th / l in 0 0.00% Ov . f ty inq Line. 2660 20.72% 1.1-»0 3060 23.83% 35.33% 2nd»40 420 3.27%| 64.67% 11.93% 3rd . 40 220 1.71% 13 64% 33 33% 4th<40 0 0 0 0 % 42.42VI W H h i n 4 0 m 3700 28.82% 1et««0 660 5.14% 7.62% 2nd««0 340 2.65%| 72.29% 9 6 6 % 3td»60 100 0.78% 23.30% 15.15% 4 th .60 0 0.00% 57.58% | W i th in 60 m 1100 8.57% Stream Length 12840 V o n l ine. Vwr th in 40m 2 1 V 5 0 V 8660 3520 660 V w i t h i n 60m 58% V o c c u p i e d l ine. 32% B a a l n * 24 2nd Stream Order Length V 1 s t 1st 880 36.67% 70.97% 65.52% 2nd 760 31.67% 3rd 0 0.00% 4th 0 0 .00V Isolated 1640 68.33% 1st / l in 120 5.00% 9.68% 2ndf l in 180 7.50% 15.52% 3rrJ/1in 0 0.00% 4thAin 0 0.00% Over ly inq L i n s . 300 12.50% 1st*40 180 7.50% 14.52% 2nd*40 200 8 . 3 3 V | 24.19% 17.24% 3rd«40 0 0.00% 32.76% 4th*40 0 0.00% Wi th in 40 m 380 15.83% 1st«60 60 2.50% 4.84% 2nd»60 20 0.63% | 29.03% 1.72V 3rd*60 0 0.00% 34.48% 4th*eo 0 0.00% Wi th in 60 m 80 3 .33V Stream Length 2400 153.23% 167.24V V o n l i ns . V w r t h i n 40m 13V 2 8 % 1240 1160 V w r t h i n 60m 3 2 V V o c c u p i e d l i ns . 70% B a s i n # 25 Stream Order Length % 1st, 2nd 1st 0 0.00% 0.00% 19.35% 2nd 120 8.70% 3rd 0 0.00% 4th 0 0.00% Isolated 120 8.70% 1st/ l in 340 24.64% 44.74% 2nd/ l in 140 10.14% 22.58% 3rdflin 0 0.00% 4th/ l in 0 0.00% Overlyinq L ins . 480 34.78% 1.1*40 420 30.43% 55.26% 2nd*40 260 18.84%| 100.00% 41.94% 3rd*40 0 0.00% 64.52% | 4th*40 0 0.00% Within 40 m 680 49.28% 1st»60 0 0.00% 0.00% 2nd*60 100 7.25%l 100.00% 16.13% 3rd*60 0 0.00% SO.65%1 41hi60 0 0.00% Within 60 m 100 7.25% Stream Length 1380 V o n l ins . 35% V w i t h i n 40m 84% 760 620 V w i t h i n 60m 9 1 V V o c c u p i e d l ins . 75% B a s i n « 26 26 2nd Stream Order Length V 1st. 1st 220 8.53% 12.36% 27.50% 2nd 220 8.53% 3rd 0 0.00% 4th 0 0.00% Isolated 440 17.05% IstTlin 520 20.16% 29.21% 2nd / l in 240 9.30% 30.00% 3rd/ l in 0 0.00% 4th / l in 0 0.00% Over ly inq L i n s . 760 29.46% 1st«40 940 36.43% 52.81% 2nd*40 180 6.98%| 82.02% 22.50% 3rd»40 0 0.00% 52.50% | 4th*40 0 0.00% Wi th in 40 m 1120 43.41% 1st*60 100 3.88% 5.62% 2nd»60 160 6.20% | 67,64% 20.00% 3rd*60 0 0.00% 72.50% | 4th*60 0 0.00% Wi th in 60 m 260 10.08% Stream Length 2580 V o n l ins . 29% V w i t h i n 40m 73% 1780 800 V w i t h i n 60m 83% V o c c u p i e d l ins . 44% B a s i n 0 27 Stream Order Length V 1st. 2nd 1st 620 17.51% 25.20% 2nd 240 6.78% 22.22% 3rd 0 0.00% 4th 0 0.00% Isolated 860 24.29% IstTlin 340 9.60% 13.82% 2nd/1in 240 6.78% 22.22% 3rd/l in 0 0.00% 4th/ l in 0 0.00% Overlyinq L ins . 580 16.38% 1st MO 1040 29.38% 42.28% 2nd .40 500 14.12%| 56.10% 46.30% 3rd»40 0 0.00% 68.52%| 4th»40 0 0.00% Within 40 m 1540 43.50% 1st«60 460 12.99% 1 8 7 0 % 2nd»60 100 2.82%| 74.80% 9.26% 3rd*60 0 0.00% 77.78% | 41h«60 0 0.00% Within 60 m 560 1582% Stream Length 3540 V o n line. 16% Vwrth in 40m 60% 2460 1080 V w i t h i n 60m 7 6 V V o c c u p i e d l ins. 100% B a s i n # 28 Stream Order Length V 1 s t 2nd 1st 280 13.33% 22.22% 2nd 200 9.52% 23.81% 3rd 0 0.00% 4th 0 0.00% Isolated 480 22.86% 1st / l in 400 19.05% 31.75% 2nd/1in 240 11.43V 28.57% 3rd / l in 0 0.00% 4th/I in 0 0.00% Over ly inq L i n s . 640 30.48% 1st t40 520 24.76% 41.27% 2nd«40 360 ,17.14%| 73.02% 42.86% 3rd»40 0 0.00% 71.43% 4th*40 0 0.00% Wi th in 40 m 880 41.90% 1st««0 60 2.86% 4.76% 2nd*60 40 1.90%| 77.78% 4 76% 3rd*60 0 0.00% 76.19% 4th»<0 0 0.00% Wi th in 60 m 100 4.76% Stream Length 2100 V o n l ine. 30% V w i t h i n 40m 7 2 V 1260 840 V w r t h i n 60m 77% V o c c u p i e d l ins . 29% Results of spatial correlations between lineaments and streams for individual basins (Continued). (lins. = lineaments) 228 B a s i n # 29 1st. 2nd 3rd 4th Stream Order Length % 1st 50780 45.17% 70.20% 2nd 8760 7.79% 46.35% 3rd 7620 6.78% 91.81% 88 0 4 % 4th 11340 10.09% Isolated 78500 69.83% 1st / l in 8980 7.99% 12.41% 2nd/ l in 3400 3.02% 17.99% 1 4 5 % 3rd/ l in 120 0.11% 0.00% 4th/1in 0 0.00% Overly inq t i n s . 12500 11.12% 1st+40 8960 7 5 7 % 12.39% 2nd«40 5260 4.68%| 24.80% 27.83% 3rd*40 280 0.25% 45.82% 3.37% 41h*40 900 0.80% 4,82% 6 9 9 % With in 40 m 15400 13.70% 6.99%l 1st+60 3620 3.22% 5.00% 2nd+60 1480 1.32%| 29.80% 7.83% 3rd»60 280 0.25% 53.65% 3.37% 4 th t60 640 0.57% 8.19% 4.97% Wrthin 60 m 6020 5.35% 11.96%l Stream Length 112420 %on l ins . %wrth in40m 11% 25% 72340 18900 8300 12880 %withtn 60m 30% V o c c u p i e d l ins . 17% Results o f spatial correlations between lineaments and streams for individual basins (Continu (lins. = lineaments) 229 i APPENDIX V I Lineament and stream trend data for the sample set basins 230 Freauency Distribution Entire Sample Set Igneous Basins Metamorphic Basins 0-4 17 6 11 5-9 30 8 22 10-14 37 11 26 15-19 25 10 15 20-24 16 7 9 25-29 16 6 10 30-34 19 < 8 11 35-39 18 9 9 40-44 22 11 11 45-49 16 7 9 50-54 13 6 7 55-59 9 5 4 60-64 5 4 1 65-69 5 0 5 70-74 7 3 4 75-79 9 3 6 80-84 6 5 1 85-89 15 9 6 90-94 19 10 9 95-99 14 8 6 100-104 9 2 7 105-109 10 3 7 110-114 10 4 6 115-119 6 3 3 120-124 11 3 8 11 125-129 12 1 130-134 19 4 15 135-139 14 0 14 140-144 15 3 12 145-149 8 4 4 150-154 8 3 5 155-159 12 3 9 160-164 8 3 5 165-169 15 7 8 170-174 17 8 9 175-179 17 8 9 Total 509 195 314 Frequency - distribution of lineaments in the sample set basins 231 Distribution Entire Sample Set Frequency Igneous Basins Metamorphic Basins 0-4 25 5 20 5-9 32 10 22 10-14 27 7 20 15-19 17 1 16 20-24 22 6 16 25-29 34 12 22 30-34 20 5 15 35-39 27 9 18 40-44 25 7 18 45-49 39 12 27 50-54 23 6 17 55-59 25 9 16 60-64 22 8 14 65-69 28 6 22 70-74 15 4 11 75-79 26 3 23 80-84 23 5 18 85-89 30 6 24 90-94 35 5 30 95-99 21 3 18 100-104 32 5 27 105-109 26 8 18 110-114 24 3 21 115-119 32 11 21 120-124 19 2 17 125-129 20 4 16 130-134 27 9 18 135-139 16 1 15 140-144 21 5 16 145-149 18 3 15 150-154 22 10 12 155-159 17 4 13 160-164 15 6 9 165-169 21 7 14 170-174 29 11 18 175-179 27 12 15 Total 882 230 652 Frequency - distribution of stream segments in the sample set basins 232 Frequency Distribution Al l lineaments Igneous Basins Metamorphic controlling streams Basins 0-4 5 2 3 5-9 13 6 7 10-14 7 4 3 15-19 5 3 2 20-24 2 1 1 25-29 0 0 0 30-34 6 , 3 3 35-39 7 3 4 40-44 10 5 5 45-49 5 4 1 50-54 6 4 2 55-59 3 3 0 60-64 1 1 0 65-69 1 0 1 70-74 4 2 2 75-79 4 2 2 80-84 2 1 1 85-89 6 4 2 90-94 9 3 6 95-99 2 1 1 100-104 2 0 2 105-109 3 2 1 110-114 3 3 0 115-119 3 3 0 120-124 5 2 3 125-129 3 1 2 130-134 4 4 0 135-139 2 0 2 140-144 0 0 0 145-149 1 1 0 150-154 2 1 1 155-159 3 1 2 160-164 6 2 4 165-169 4 3 1 170-174 7 4 3 175-179 7 6 1 Total 153 85 68 Frequency - distribution of lineament controlled stream segments in the sample set basins 233 Frequency Distribution 1st. order Ig- Meta. 2nd. order Ig Mcta. 3rd. order Ig Meta. 0-4 16 4 12 5 1 4 2 0 2 5-9 20 9 11 6 1 5 2 0 2 10-14 21 6 15 4 2 2 0 0 0 15-19 8 1 7 4 0 4 3 0 3 20-24 16 5 11 2 1 1 2 0 2 25-29 24 11 13 5 1 4 4 0 4 30-34 11 4 7 5 1 4 0 0 0 35-39 17 7 10 2 2 0 4 0 4 40-44 14 6 8 6 1 5 3 0 3 45-49 26 10 16 4 2 2 5 0 5 50-54 14 3 11 6 3 3 2 0 2 55-59 15 7 8 4 2 2 4 1 3 60-64 14 5 9 6 2 4 2 1 1 65-69 19 5 14 4 1 3 4 0 4 70-74 10 4 6 1 0 1 3 0 3 75-79 16 2 14 5 1 4 3 0 3 80-84 15 5 10 6 0 6 2 0 2 85-89 17 4 13 4 1 3 6 1 5 90-94 25 5 20 5 1 4 5 0 5 95-99 11 2 9 3 0 3 7 2 5 100-104 22 2 20 3 2 1 7 1 6 105-109 16 4 12 6 3 3 2 1 1 110-114 18 2 16 4 1 3 1 0 1 115-119 20 6 14 5 3 2 5 3 2 120-124 9 1 8 5 0 5 2 1 1 125-129 12 3 9 4 1 3 3 0 3 130-134 15 5 10 8 3 5 2 1 135-139 11 1 10 3 0 3 1 0 1 140-144 13 5 8 4 0 4 3 0 3 145-149 14 2 12 1 0 1 2 1 1 150-154 12 7 5 6 3 3 0 0 0 155-159 11 1 10 6 3 3 0 0 0 160-164 9 4 5 5 3 2 1 0 1 165-169 12 4 8 6 3 3 1 0 1 170-174 17 9 8 9 3 6 1 0 1 175-179 19 10 9 6 1 5 1 1 0 Total 559 171 388 168 52 116 96 15 81 Frequency - distribution of stream segments in the sample set basins (Ig. = Igneous basins; Mcta. = Metamorphic basins) 234 Dataset Sample size Maximum percent Vector Mean Vector Magnitude Consistency Ratio Lineament trends 509 7 11.31 101.54 0.1995 Igneous basins 195 5 28.80 45.81 0.2349 Metamorphic basins 314 8 .015 69.19 0.2203 Stream segment trends 882 4 150.33 46.47 0.0527 Igneous basins 230 5 20.79 27.82 0.1209 Metamorphic basins 652 4 167.10 49.43 .00758 Lins visually controlling stream orientations 153 8 21.94 26.97 0.1763 Igneous basins 85 22.90 15.33 0.1803 Matamorphic basins 68 10 20.68 11.66 0.1715 First order stream segments 559 4 147.22 35.89 0.0588 Igneous basins 171 6 24.42 33.71 0.1972 Metamorphic basins 388 5 176.37 36.08 0.0930 Second order stream segments 168 5 342.88 10.27 0.0611 Igneous basins 52 5 346.60 5.61 0.1078 Metamorphic basins 116 5 338.49 4.76 0.0411 Third order stream segments 96 7 172.35 29.30 0.3052 Igneous basins 15 20 200.84 9.05 0.6034 Metamorphic basins 81 7 163.70 25.52 0.3151 Summary vector statistics for lineament and stream trend data from the sample set basins 235 APPENDIX VII Matrices of R 2 values for regression analysis ( 236 237 238 £ c « ^ ^ J\" _c *c oo — oc ^ « •*» \"* ^ ca 5? ^ ^ oo r- v-i — r- c— TJ 1 -Nl -vf V ) CO \"6 C~ £ • <. .° >° •? >~ '.' .° ^ ';V ... ... ... ... „~ ... \"^ S ^? (*1 f O CIS \"O —1 - c n E g \" r-~ ci y- ^\\ o- « n M c <--* « ^ n <-i OS o> o . r-u c M «\"v >o o\\ -^ K >o r~ - - - -^ N —> T <»% T I~1 ICQ irt ~ \" ~ *\"* ~ ~~ N« S° «-l >© o< V- - o xo xo >^ .» -.0 2? ^ ^ ^ ^ ^ 5 - £ S 5 3 S> 2 h 2 2 ~ -\"ir^^roc-t — v~> — *r 0 3 » •© c- 00 <7> —. f-J \"-I — «J f J O. M S i ? J * 7 1^ r i w o . o c o > > o c o o o o c ~ t t 5 ^ — r- t~ >^ 00 £ c r~ -™ «j «i v-, t>, o ifi i n 1^ 1 o r-*(~-t~-e>. O i i ^ - O C - s O > — • t (\\| — * \"O <0 I—r«\"i >© t~- f l fM *0 a. ' < O. 'O -u\"i r< r « o « o « c -= u.[.j)uj|-j|ca|.j|.j — r . — h> v% ^ -«T »-> -20m/s (72 km/h). The headwall of the slide comprised steep columnar joints within the cooled lava. The base of the slide corresponded closely to an unconformity surface (Moore and Mathews 1978) and it is likely that groundwater flow at this surface contributed to failure. There remains the potential for further failure at the site. The Mount Mason Slide (site #20) is briefly described in Evans (1986) and in Clague and Evans (1994). The slide has an estimated area of 450,000 m2 and an estimated volume of 10.5 xl0 6m 3 (Evans 1986). The slide has not been accurately dated and occurs on or near the contact between coast plutonic migmatite and quartz diorite. 250 251 252 APPENDIX X Results of overlay of streams on lineaments for the Seymour watershed 253 Buffer zone distance in m Number of stream pixels Percentage of total Cumulative percentage of stream pixels stream pixels 0-20 676 2.87% 2.87% 20-40 1871 7.94% 10.81% 40-60 1786 7.58% 18.39% 60-80 1277 5.42% 23.81% 80-100 1210 5.13% 28.94% 100-120 1599 6.79% 35.73% 120-140 1056 4.48% 40.21% 140-160 1195 5.07% 45.28% 160-180 1208 5.13% 50.40% 180-200 973 4.13% 54.53% 200-220 907 3.85% 58.38% 220-240 758 3.22% 61.60% 240-260 783 3.32% 64.92% 260-280 712 3.02% 67.94% 280-300 530 2.25% 70.19% 300-320 555 2.36% 72.55% 320-340 557 2.36% 74.91% 340-360 491 2.08% 76.99% 360-380 468 1.99% 78.98% 380-400 406 1.72% 80.70% 400-420 391 1.66% 82.36% 420-440 445 1.89% 84.25% 440-460 327 1.39% 85.64% 460-480 287 1.22% 86.86% 480-500 292 1.24% 88.09% 500-520 306 1.30% 89.39% 520-540 260 1.10% 90.50% 540-560 188 0.80% 91.29% 560-580 185 0.79% 92.08% 580-600 192 0.81% 92.89% 600-620 166 0.70% 93.60% 620-640 145 0.62% 94.21% 640-660 117 0.50% 94.71% 660-680 153 0.65% 95.36% 680-700 150 0.64% 96.00% 700-720 107 0.45% 96.45% 720-740 83 0.35% 96.80% 740-760 82 0.35% 97.15% 760-780 101 0.43% 97.58% 780-800 65 0.28% 97.85% 800-820 59 0.25% 98.10% 820-840 65 0.28% 98.38% 840-860 49 0.21% 98.59% 860-880 51 0.22% 98.80% 880-900 53 0.22% 99.03% 900-920 39 0.17% 99.20% Results of overlay of streams on lineaments for the Seymour Watershed. 254 920-940 31 0.13% 99.33% 940-960 26 0.11% 99.44% 960-980 31 0.13% 99.57% 980-1000 15 0.06% 99.63% 1000-1020 17 0.07% 99.70% 1020-1040 9 0.04% 99.74% 1040-1060 9 0.04% 99.78% 1060-1080 11 0.05% 99.83% 1080-1100 18 0.08% 99.90% 1100-1020 7 0.03% 99.93% 1120-1140 6 0.03% 99.96% 1140-1160 6 0.03% 99.98% 1160-1180 4 0.02% 100.00% 23566 100.00% Results of overlay of streams on lineaments for the Seymour Watershed (Continued) 255 "@en, "Acc. material: one floppy disk (Koerner Library)."@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1998-11"@en ; edm:isShownAt "10.14288/1.0088464"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Geological Sciences"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Lineament control on drainage basin development, large rock landslides and mountain slope deformation in the southwest coast mountains, British Columbia, Canada"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/8077"@en .