STREAM CHANNEL STABILITY IN THE QUEEN CHARLOTTE ISLANDS: AN EXAMINATION OF SCHUMM'S FLUVIAL MODEL by RICHARD GRAHAM ROBERTS B.Sc. (Hons.), The University College of Wales, Aberystwyth, 1981 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Geography) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 1984 © Richard Graham Roberts, 1984 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree th a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree t h a t permission f o r ex t e n s i v e copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the head o f my department or by h i s or her r e p r e s e n t a t i v e s . I t i s understood t h a t copying or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not be allowed without my w r i t t e n p ermission. Department of CJQ <^ C C X p W ^ j The U n i v e r s i t y of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date rVuc^vr^V V ^ L f ABSTRACT Based upon inv e s t i g a t i o n s of ephemeral, sand-bed streams i n semi-a r i d areas, Schumm (1977) proposed a model of the f l u v i a l system i n which long periods of r e l a t i v e s t a b i l i t y are separated by b r i e f episodes of major morphologic change (termed "dynamic metastable equilibrium"). The l a t t e r c onstitute "episodic behaviour" and are i n i t i a t e d when the system crosses a "geomorphic threshold". The resultant sediment transport-limited conditions may induce a se r i e s of progressively damped c u t - a n d - f i l l cycles, which Schumm termed "complex response". In t h i s study, the s t a b i l i t y of perennial, gravel-bed streams i n the per-humid Queen Charlotte Islands was investigated, i n order to test Schumm's f l u v i a l model i n a strongly contrasting environment. Compared with semi-arid stream systems, forest streams i n the Queen Charlotte Islands are l e s s s e n s i t i v e , and exhibit greater r e s i l i e n c e , to major hydrologic events. In general, a condition of sediment supply-l i m i t a t i o n and "dynamic equilibrium" probably has prevailed i n forested watersheds throughout the past 8,000 years. Schumm's model i s overly elab-orate i n such s i t u a t i o n s . However, the behaviour of four stream systems i n the Queen Charlotte Islands, whose s e n s i t i v i t y was increased by streamside logging a c t i v i t i e s , approaches that of semi-arid stream systems i n c e r t a i n respects. The land-use change crossed an " e x t r i n s i c geomorphic threshold", lowered an " i n t r i n s i c geomorphic threshold" (stream bank s t a b i l i t y ) , and generated major episodes of channel widening and the development of aggradational "sediment wedges". This i s consistent with Schumm's d e f i n i t i o n of "episodic behaviour". However, the occurrence of only one cycle of aggradation and degradation at any one cross-section i s not compatible with the concept of i i "complex response". A simple, or Mackin-type, adjustment at "grade" i s proposed as a more sui t a b l e d e s c r i p t i o n of the heavily damped relaxation pattern of "sediment wedge" movements; t h i s response i s promoted by the propensity of gravels to "armour", and alders to revegetate these deposits, and thereby i n h i b i t t h e i r entrainment. "Sediment wedges" represent the most notable example of f l u v i a l d i s e q u i l i b r i u m to have occurred i n the Queen Charlotte Islands during the past 8,000 years which complies with the concept of "dynamic metastable equilibrium". But i t i s anticipated that these creeks w i l l return to th e i r pre-disturbance condition of "dynamic equilibrium" due to the operation of stream channel recovery processes. TABLE OF CONTENTS Page ABSTRACT i i TABLE OF CONTENTS iv LIST OF TABLES v i i LIST OF FIGURES ix LIST OF PLATES x i i ACKNOWLEDGEMENTS x i i i CHAPTER 1.0 INTRODUCTION 1 1.1 Sediment transfers in drainage basins 2 1.2 Geomorphic "equilibrium" concepts 3 1.3 Schumm's concepts of landform evolution 6 1.4 Limitations of Schumm's concepts 8 1.5 Factors affecting the efficacy of Schumm's concepts 11 1.5.1 Geomorphic reaction 13 1.5.2 Geomorphic relaxation 16 1.5.3 Geomorphic recovery 18 1.6 Study objectives 19 CHAPTER 2.0 STUDY WATERSHEDS 36 2.1 Location 36 2.2 Geology 38 2.2.1 Glacial history 38 2.2.2 Bedrock and tectonics 39 2.2.3 Physiography 42 2.2.4 Surficial materials 43 2.3 Climate and vegetation 50 2.3.1 Precipitation 50 2.3.2 Wind 53 2.3.3 Vegetation 53 iv Page 2.4 Timber harvesting activities 55 2.4.1 Armentieres Creek 55 2.4.2 Mosquito Creek tributary 60 2.4.3 Mountain Creek 60 2.4.4 Lagins Creek tributary 61 CHAPTER 3.0 STUDY METHODOLOGY 63 3.1 Review of sediment sources and mobiliz-ation processes: Summary 63 3.2 Field measurements 66 3.2.1 Channel morphology 66 3.2.2 Sediments 68 3.3 Air photo analyses 71 3.3.1 Hillslope erosion 72 3.3.2 Riparian erosion 74 3.3.3 Road network 76 CHAPTER 4.0 FLUVIAL "SEDIMENT WEDGES" IN THE QUEEN CHARLOTTE ISLANDS 77 4.1 Sediment budget 77 4.1.1 Sediment sources 77 4.1.2 Sediment storage 98 4.1.3 Sediment mobility 106 4.2 Causes of "sediment wedge" i n i t i a t i o n and development 124 4.3 Summary 132 CHAPTER 5.0 CONCEPTUAL IMPLICATIONS 134 5.1 First hypothesis 136 5.2 Second hypothesis 139 5.3 Third hypothesis 142 CHAPTER 6.0 CONCLUSIONS 150 BIBLIOGRAPHY 152 APPENDIX A REVIEW OF SEDIMENT SOURCES AND MOBILIZATION PROCESSES 171 A.1 Soil creep 172 A.2 Tree throw 177 A.3 Landslides 178 A.4 Surface erosion 183 v Page A.5 Stream bank erosion and large organic debris 185 A.6 Hydrologic changes resulting from timber harvesting 188 APPENDIX B VALUES OF SEDIMENT PRODUCTION AND DELIVERY BY LANDSLIDES 194 v i LIST OF TABLES Page Table 1.1 Factors for and against the efficacy of Schumm's concepts i n the Queen Charlotte Islands Table 2.1 General information for the study watersheds Table 2.2 Logging histories of the study watersheds Table 2.3 Logging activities in the study watersheds Table 3.1 Representative rates of sediment delivery in the Pacific Northwest Table 3.2 Process rates used in this study Table 3.3 Surface, sub-surface and bank sample sites Table 3.4 Available air photography Table 4.1 Volumes of sediment delivered from various sources Table 4.2 Sediment sources in Armentieres Creek watershed Table 4.3 Measurements of changes i n stream channel widths Table 4.4 Mean stream bank heights Table 4.5 Time periods used to assess landslide inputs Table 4.6 Rates of landslide erosion and sediment trans-fer in logged watersheds Table 4.7 Relation between slope failure types and water-shed conditions Table 4.8 Annual rates of sediment delivery by s o i l creep and tree throw Table 4.9 Annual rates of sediment delivery by slope wash Table 4.10 Duration of different rates of road surface erosion Table 4.11 The volume of sediment stored i n "sediment wedges" Table 4.12 Grain size s t a t i s t i c s for "wedge" surface, "wedge" sub-surface, and stream bank sediments 33 40 56 57 64 65 69 73 80 81 83-84 86 89 90 92 94 96 97 104 112 v i i Page Table 4.13 Proportion of sub-surface and stream bank material coarser than 1 mm and 2 mm 115 Table 4.14 Volumes and rates of sediment output from the "sediment wedges" 116 Table 4.15 Volumes of intra-"wedge" sediment transfers 122 Table A.l Creep rates in undisturbed basins 175 Table A.2 Creep rates in disturbed basins 176 Table A.3 Calculated rates of sediment transport by tree throw in forested watersheds 180 Table A.4 Classes of mass movement 180 Table A.5 Debris slide erosion in forested areas of the Pacific Northwest 182 Table A.6 Rate of debris slide erosion relative to for-ested areas 182 Table A.7 Rates of sediment production by road surface erosion 186 Table A.8 Sediment storage behind in-stream obstructions 189 Table A.9 Hydrologic changes following clearcut and patchcut logging of watersheds in the Pacific Northwest 190-191 Table B.l Armentieres Creek watershed: Landslides 195 Table B.2 Mosquito Creek tributary watershed: Landslides 196 Table B.3 Mountain Creek watershed: Landslides 197 Table B.4 Lagins Creek tributary watershed: Landslides 198 v i i i LIST OF FIGURES Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 1.5 Figure 1.6 Figure 1.7 Figure 1.8 Figure 1.9 Figure 1.10 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Diagrammatic representation of equilibria Erosional evolution of drainage basins Pr e c i p i t a t i o n and streamflow v a r i a b i l i t y i n r e l a t i o n to mean annual preci p i t a t i o n (M.A.P.) The e f f e c t of r e a c t i o n and r e l a x a t i o n times in systems in which the rate of energy input i s time dependent Relationship between the magnitude-frequency distribution of applied stress and the strain rate Simple and complex relaxation paths with identical relaxation times The interaction between sediment availability and the recovery rate relative to the frequency of hydrologic events (E) that t r i g g e r geomorphic change Geomorphic threshold, concepts Diagrammatic representation of a "sediment wedge" Flow diagram of study design Location of the study watersheds in the Queen Charlotte Islands Frequency distribution curve of total annual precipitation in Tasu Sound, 1964-1983 Seasonal distribution of mean annual precip-itation in Tasu Sound, 1964-1983 Recurrence interval of precipitation events with magnitudes greater than 80 mm in 24 hours in Tasu Sound, 1972-1982 Logging activities in the Armentieres Creek watershed Logging activities in the Mosquito Creek t r i b -utary watershed Page 5 9 12 14 15 17 20 21 23 35 37 51 51 52 58 58 i x Page Figure 2.7 Logging activities in the Mountain Creek water-shed 59 Figure 2.8 Logging activities in the Lagins Creek t r i b -utary watershed 59 Figure 4.1 Location of the "sediment wedges" i n Armentieres Creek 78 Figure 4.2 Location of the "sediment wedge" in Mosquito Creek tributary 78 Figure 4.3 Location of the "sediment wedge" in Mountain Creek 79 Figure 4.4 Location of the "sediment wedge" i n Lagins Creek tributary 79 Figure 4.5 Longitudinal profile of the "sediment wedges" in Armentieres Creek 99 Figure 4.6 Longitudinal profile of the "sediment wedge" in Mosquito Creek tributary 100 Figure 4.7 Longitudinal profile of the "sediment wedge" in Mountain Creek 101 Figure 4.8 Longitudinal profile of the "sediment wedge" in Lagins Creek tributary 102 Figure 4.9 Types of sediment storage adjustment 105 Figure 4.10 Particle size distribution of "sediment wedge" and stream bank material in Armentieres Creek 108 Figure 4.11 Particle size distribution of "sediment wedge" and stream bank material i n Mosquito Creek tributary 109 Figure 4.12 Particle size distribution of "sediment wedge" and stream bank material in Mountain Creek 110 Figure 4.13 Particle size distribution of "sediment wedge" surface material in Lagins Creek tributary 111 Figure 4.14 Schematic diagram of "sediment wedge" movement and stream terrace formation by the downstream progression of a sediment "wave" 120 Figure 5.1 Hypothetical sequence of events 135 x Page Figure 5.2 Schematic diagram of the r e l a t i o n between relaxation, recovery, and equilibrium concepts 147 Figure A.l Sediment transfers i n P a c i f i c Northwest watersheds 173 x i LIST OF PLATES Page Plate 1.1 Aerial photograph of Mountain Creek before and after logging and "sediment wedge" inception 24-25 Plate 1.2 A e r i a l photographs of streamside logging a c t i v i t i e s and "sediment wedge" growth i n Mosquito Creek tributary 26-27 Plate 1.3 Oblique air photograph of the "sediment wedge" in Mosquito Creek tributary 28 Plate 1.4 Photographs of each of the "sediment wedges" 29-30 Plate 1.5 An example of severe stream bank erosion by Mosquito Creek tributary 31 Plate 2.1 Examples of organic stream bank deposits along Mountain Creek 45-46 Plate 4.1 An example of a stream terrace i n Mountain Creek 119 x i i ACKNOWLEDGEMENTS Financial support for this study was provided by The University of British Columbia, through a University Graduate Fellowship; the Natural Sciences and Engineering Research Council of Canada, through Operating Grant A7950 to Dr. M. Church; and the Fish/Forestry Interaction Program, sponsored by the British Columbia Ministry of Environment, British Columbia Ministry of Forests, and Canada Department of Fisheries and Oceans. I would like to thank my supervisor, Dr. M. Church for his guid-ance and support throughout t h i s study. The advice and comments of Dr. H. 0. Slaymaker also are greatly appreciated. I wish to acknowledge L. Davies, C. Horner, and C. Robertshaw for their assistance in the f i e l d ; K. M. Rood for sharing his photogrammetric expertise; D. L. Hogan for many useful discussions; and numerous members of the Fish/Forestry Interaction Program research team for helpful suggestions. Finally, I extend my thanks to C. Hubbard for typing the thesis and S. Grimmond for her help in i t s f i n a l production. 1.0 INTRODUCTION The sediment budget provides a framework within which the trans-fers of sediment by h i l l s l o p e and channel processes can be integrated. Unfortunately, q u a n t i f i c a t i o n of the sediment input, storage, and output components i s complicated by the unsteady nature of process rates and land-form changes in time and space. Over the past century, this variability has been treated within various models of landscape development. During the past decade, S. A. Schumm and his students have advocated the concept of "dynamic metastable equilibrium". Schumm proposed that the evolution of the f l u v i a l system i s not progressive but episodic, with long periods of relative stability separated by brief periods of major a c t i v i t y " a f t e r the crossing of i n t r i n s i c and extrinsic geomorphic thresholds (Schumm, 1979, and prior references there-in). After crossing a threshold, the f l u v i a l system seeks a new equilibrium by alternately aggrading and degrading in a series of progressively damped c u t - a n d - f i l l cycles, which Schumm termed "complex response". However, he acknowledged that these concepts may be appropriate only in areas of high rel i e f and high sediment production, where the stream channel i s unable, in the short term, to transport a l l of the sediment supplied (Schumm, 1977). Hence, f i e l d evidence of his concepts, so far, has been drawn only from ephemeral channels in semi-arid regions, where such transport-limited con-ditions certainly prevail. The appropriateness of Schumm's concepts in other climatic and geomorphic environments, and hence their value as a general geomorphological model, depends upon the occurrence of morphologic and sedimentologic reaction, relaxation, and recovery characteristics that encourage transport-l i m i t e d conditions. Consequently, an adequate description of f l u v i a l 1 phenomena in some environments may require the modification or abandonment of his conceptual framework, but neither possibility has yet been examined rigourously. It i s , therefore, the objective of t h i s thesis to examine the v a l i d i t y of Schumm's conceptual framework i n a c l i m a t i c and geomorphic setting that contrasts with that in which i t was conceived. The development of perennial, gravel-bed streams in the per-humid Queen Charlotte Islands provides such an opportunity. 1.1 Sediment transfers in drainage basins In t h i s thesis, "sediment transfer" denotes the movement of sediment without regard for mechanisms, while "sediment transport" refers to movement associated with a particular mechanism. A better knowledge of sediment transfers i n drainage basins i s prerequisite to a more complete understanding of the integration between hillslope and stream channel processes of sediment erosion, transport, and deposition. Progress in this f i e l d of research w i l l enable resource manage-ment strategies and erosion control measures to be implemented more effectively. The sediment budget is a quantitative statement of the sediment transfers i n a geomorphic system, usually a drainage basin, and has the basic form: I + A S = 0 where I = sediment input; A S = change in sediment storage; and 0 = sediment output. 2 Recently, sediment budgets have been applied widely to attempt to understand geomorphic processes and the effects of land management in the forested drainage basins of the Pacific Northwest region of North America (e.g. D i e t r i c h and Dunne, 1978; Swanson et al . , 1982a, 1982c). Unfortunately, the construction of a sediment budget i s com-plicate d by the s p a t i a l and temporal v a r i a t i o n i n the resistance and response of the geomorphic system to changes initiated by forcing events of differing magnitude and frequency; forcing events include weather and seis-mic events, and more generally, c l i m a t i c , base-level (i.e. eustatic, i s o s t a t i c and tectonic) and land-use changes. The complex int e r a c t i o n between geomorphic systems and forcing events i s manifested in the unsteady nature of process rates and landform changes. In particular, the exceedance of process and landform thresholds results in episodic geomorphic behaviour. 1.2 Geomorphic "equilibrium" concepts Much attention has been directed toward conceptualizing the evo-l u t i o n of the f l u v i a l system, which represents the integrated r e s u l t of several sediment transfer processes in a drainage basin. These various concepts e n t a i l the introduction of d i f f e r e n t ideas of geomorphic "equilibrium". Although the term "equilibrium" has been in the geomorphological l i t e r a t u r e for more than a century, a structured framework with formal d e f i n i t i o n s has been developed only during the l a s t quarter; Chorley and Kennedy (1971) defined "equilibrium" as a condition i n which some kind of balance i s maintained. In a f l u v i a l context, Gi l b e r t (1880), Davis (1902), and Mackin (1948) used the word "graded" to describe a stream in "equilibrium" or, i n engineering terminology, a stream "in regime". Mackin (1948, p.471) defined 3 "grade" as follows: "A graded stream i s one in which, over a period of years, slope i s d e l i c a t e l y adjusted to provide, with a v a i l a b l e discharge and with prevailing channel c h a r a c t e r i s t i c s , just the v e l o c i t y required for the transportation of the load supplied from the drainage basin. The graded stream i s a system in equilibrium." Leopold and Bull (1979) subsequently modified Mackin's definition to include hydraulic roughness, and thus emphasize that i n t e r n a l adjustments involve more than just slope and veloc i t y . The concept of "grade" can also be applied to hillslopes (Carson and Kirkby, 1972). Strahler (1952a), Hack (1960), Chorley (1962), Schumm and Lichty (1965) and Howard (1965) broadened Mackin's conceptual framework and intro-duced the terms "steady-state equilibrium" and "dynamic equilibrium" to describe the condition of an "open" geomorphic system; Chorley and Kennedy (1971) defined the latter as a system with boundaries which allow the import and export of both mass and energy. This concept i s applicable to streams, slopes, or entire landscapes. "Steady-state equilibrium" has been defined as the condition of an open system wherein properties are invariant when considered with refer-ence to a given time scale, but within which i t s instantaneous condition may oscillate due to the presence of interacting variables (Chorley and Kennedy, 1971; see Figure 1.1). In contrast, "dynamic equilibrium" i s the circum-stance i n which fluctuations occur about a constantly changing system condition which has a trajectory of unrepeated "average" states through time (Chorley and Kennedy, 1971; see Figure 1.1). Where thresholds allow occasionally large fluctuations to i n i t i a t e new regimes of "dynamic equilib-rium" , a condition of "dynamic metastable equilibrium" exists (Chorley and Kennedy, 1971; see Figure 1.1). A FIGURE 1.1: Diagrammatic representation of e q u i l i b r i a ( a f t e r Chorley and Kennedy, 1971) . The notion that the environment may be i n a state of "dynamic metastable equilibrium" was f i r s t incorporated into a systems framework, and thence introduced to geomorphology, by Chorley and Kennedy (1971), although the existence of thresholds had been recognized previously (e.g. Wolman and Miller, 1960). The abil i t y of this concept to describe adjustments of the fl u v i a l system has been investigated most notably by Schumm and his students during the past decade (e.g. Schumm, 1973, 1976, 1977, 1979, 1981; Schumm and Parker, 1973; Patton and Schumm, 1975, 1981; Womack and Schumm, 1977; Bergstrom and Schumm, 1981; Begin and Schumm, 1984). 1.3 Schumm's concepts of landform evolution Schumm has argued that landscape evolution i s not slow and con-tinuous, but episodic, with long periods of relative stability separated by bri e f periods of major a c t i v i t y , due to the crossing of geomorphic thres-holds. A geomorphic threshold represents a l i m i t of landform stability that can be exceeded by the progressive change of either an e x t r i n s i c variable (e.g. climate, base-level, land-use) or some property of the landform i t s e l f (e.g. slope angle, degree of weathering, vegetation cover); these correspond to an "extrinsic geomorphic threshold" and an "intrinsic geomorphic thres-hold", respectively (Schumm, 1979). In both cases, the s i n g u l a r i t y of landforms prevents the i d e n t i f i c a t i o n of a sharp threshold, so that the notion of a threshold "zone" i s more r e a l i s t i c (Begin and Schumm, 1984). Once a geomorphic threshold has been exceeded and abrupt channel change initiated, the f l u v i a l system may seek a new equilibrium by alter-nately aggrading and degrading in a series of damped cut-and-fill cycles, which Schumm termed "complex response". "Complex response" was observed i n i t i a l l y by Schumm and Parker (1973) after they lowered s l i g h t l y the base-level of a drainage system 6 developed in a sand-filled container. Incision occurred f i r s t at the basin mouth, leaving terraces on either side, and then progressed upstream along the drainage network, with the main channel conveying upstream sediment in increasing quantities. The i n a b i l i t y of the newly cut main channel to transport a l l the sediment promoted stream aggradation, braiding, and lateral erosion, thereby partially destroying the existing terraces. As the tributaries became adjusted to the new base-level, sediment loads decreased and renewed incision occurred downstream. Thus, within hours of a single base-level lowering, multiple cut-and-fill cycles and a flig h t of terraces were produced. The existence of geomorphic thresholds, episodic behaviour, and complex response in natural f l u v i a l systems has been identified by Patton and Schumm (1975, 1981), Womack and Schumm (1977), and Bergstrom and Schumm (1981) in the semi-arid western United States; however, only Bergstrom and Schumm (1981) and Patton and Schumm (1981) observed, rather than inferred, a cu t - f i l l - c u t sequence of events over time at-a-station. External agencies (namely secular c l i m a t i c changes and/or human land-use changes) had been i d e n t i f i e d previously as catalysts of arroyo formation i n the American southwest (as summarized i n Cooke and Reeves, 1976), but Schumm was the f i r s t geomorphologist to propose that, in semi-arid areas, abrupt channel change could also occur without a change in an external variable. In his view, episodes of aggradation and degradation should be expected to occur simultaneously along a channel, and possibly be out-of-phase between channels (Patton and Schumm, 1981). Such episodic behaviour has been ob-served at a range of timescales, from days (Bergstrom and Schumm, 1981), to years (Patton and Schumm, 1981) and decades (Womack and Schumm, 1977). Channel adjustments at these timescales are related to the exceedance of 7 intrinsic thresholds, while the less frequent crossing of extrinsic thres-holds produces changes of greater magnitude. As a r e s u l t , a complete stratigraphic record should exhibit a hierarchy of sedimentation cycles, whose dimensions and spatial extent reflect rejuvenation events of different magnitude; hence, a major event (e.g. climatic change) should produce a large fining-upward cycle of regional extent, within which local cut-and-f i l l activity (e.g. complex response) i s represented by smaller cycles of limited areal extent (Schumm, 1981). A complex history of sedimentation certainly prevailed in the American southwest during the last 5,000 years, as the three most recent a l l u v i a l deposits in the stratigraphic record are spatially diachronous and exhibit significant temporal overlap (Haynes, 1968). A necessary condition for episodic behaviour and complex response in theoretical dynamic process-response models (Hey, 1979), experimental drainage basins (Schumm and Parker, 1973), and natural f l u v i a l systems (Schumm 1977) i s the exceedance of the sediment transporting capacity of the stream channel. This condition i s most common in areas of high r e l i e f and high sediment production; the f l u v i a l system, therefore, is presumed to be in "dynamic metastable equilibrium" during the early stages of landform development (see Figure 1.2). "Dynamic equilibrium" may prevail later on (see Figure 1.2), when, with lower r e l i e f , more mature weathering of detritus, and lower rates of sediment delivery to the stream, input material may be moved through the system relatively directly. 1.4 Limitations of Schumm's concepts Although the sparsely vegetated, semi-arid areas from which Schumm obtained his f i e l d data are characterized by transport-limited con-ditions, other climatic regimes are not. Schumm (1977, p.87-88) commented 8 FIGURE 1.2: Erosional evolution of drainage basins (after Schumm, 1977) . 9 that: " r i v e r s draining humid regions are competent to transport sediment yielded to them from vegetated valley sides ... Under these circumstances only a major c l i m a t e change, d i a s t r o p h i s m , or man's a c t i v i t i e s can cause major changes i n the type of sediment moved from the system ... [Thus,] i t i s presumed that the discharge of sediment from basins in humid regions i s r e l a t i v e l y constant, except when a catastrophic hydrologic event or eustatic or tectonic changes i n t e r f e r e with the normal progress of landscape denudation and movement of sediment through the system". Hence, under these supply-limited conditions: "the applicability of the [episodic evolution] model to perennial streams and the humid regions i s unknown" (Schumm, 1976, p.82). but, certainly: " i n humid regions of low r e l i e f , where there i s an abundant water supply, i t i s less likely that episodic erosion or deposition w i l l occur or that threshold conditions can exist" (Schumm, 1981, p.22). As a consequence, Schumm (1977, p.87-88) cautions that: " i t i s probably only i n sub-humid and semi-arid regions that pulses of sediment move out of tectonically undisturbed drainage basins at present", and that, perhaps, the episodic evolution model: "is applicable only to high sediment-producing areas of the drylands or the tropics". Furthermore, Schumm coll e c t e d his f i e l d data from r e l a t i v e l y small drainage systems (less than 1,000 km^ in area), where rates of change are rapid enough to be recorded morphologically and sedimentologically (Schumm, 1976). In large drainage systems, rates of change may be s u f f i c i e n t l y slow, and the distance of transport so long, that a damping effect would attenuate and obscure any episodic behaviour (Schumm, 1981). The preceding discussion emphasized that "dynamic metastable equilibrium", "geomorphic thresholds", "episodic behaviour", and "complex 10 response" are useful concepts, but that details of their mode of operation, as described by Schumm and co-workers, are representative only of ephemeral streams in semi-arid regions. Hence, in other regions, different geomorphic and/or climatic conditions may necessitate the modification or abandonment of Schumm's conceptual framework. 1.5 Factors affecting the efficacy of Schumm's concepts S p e c i f i c c l i m a t i c and geomorphic factors have been i d e n t i f i e d that encourage applied stresses, and the corresponding responses, to be highly variable, so that transport-limited conditions may occur. F i r s t l y , i t has been shown that as mean annual p r e c i p i t a t i o n decreases, the frequency distributions of the r a i n f a l l and resultant stream-flow become more right-skewed, so that the difference between the modal and less frequent discharges increases (Chorley and Kennedy, 1971; Baker, 1977; see Figure 1.3). Therefore, in more arid environments, a larger percentage of the total sediment load i s likely to be transported by infrequent flows, as a consequence of increased flow variability (Wolman and Miller, 1960). In addition, as the occurrence of intense r a i n f a l l s and peak discharges over large areas i s less l i k e l y i n semi-arid than i n humid regions, the integration of semi-arid drainage systems occurs only during very rare intensive events that cover the entire drainage basin (Wolman and Gerson, 1978); during less intense events, the movement of material i n a semi-arid watershed i s sporadic and out-of-phase in time and space. Streamflow also becomes increasingly variable as drainage area i s reduced (Wolman and Miller, 1960) and, therefore, to some extent, a decrease in area in a humid region i s comparable with a regional change from a humid to a more arid climate, as reflected in the increased importance of episodic processes (Wolman and Gerson, 1978). 11 FIGURE 1.3: Precipitation and streamflow va r i a b i l i t y in relation to mean annual precipitation (M.A.P.) (after Chorley and Kennedy, 1971; Baker, 1977). (Note: the axes in Figure B are reversed from convention so that both diagrams have an index of magnitude on the x-axis and frequency on the y-axis). A- A O c C 3 O C • H CD u X bO I-H U cd CU (J C - H |«g-Time 14 FIGURE 1.5: R e l a t i o n s h i p between the magnitude-frequency d i s t r i b u t i o n o f a p p l i e d s t r e s s and the s t r a i n r a t e ( a f t e r Baker, 1977). (Note: two s t r e s s t h r e s h o l d s (e.g. t h r e s h o l d o f sediment entrainment) are i n d i c a t e d ) . A. I n f r e q u e n t sediment t r a n s p o r t due p r i m a r i l y t o stream bed "armouring" o i — i e a CD u <44 o +J u X o o o ft c C to CD CD c 3 3 a cr cr u CD CD /—% 4-> f-i fH +-> <4H U •M o C IO ft CD to to 6 > CD c • H CD rt v > 4-> ?H CD to 4-> to tO 10 T3 4-> m CD C C o M rt CD 4-> e CD 10 CD • H +J 4-> T3 rt T3 rt CD CD M to •H i—1 C m ft • H o • ft rt CD rt v ' +-> +-> to •H CD o 4-> 4-> m c rt o rt o 3 c +J cr c CD o •H 3 3 • rt cr T3 H CD O +-> U U CD CO CL. v—> rt O A p p l i e d s t r e s s (e.g. magnitude o f streamflow) B. I n f r e q u e n t sediment t r a n s p o r t due p r i m a r i l y to skewed str e a m f l o w d i s t r i b u t i o n . A p p l i e d s t r e s s (e.g. magnitude o f streamflow) 15 associated with the progressive decline in root strength to the threshold of inst a b i l i t y . Finally, because of the f i n i t e velocities of water and sediment, a "reaction time" may also be specified when the input and response differ i n locus (Allen, 1974); for example, there i s a time lag involved between the erosion of sediment at one location and i t s deposition at a point down-stream. 1.5.2 Geomorphic relaxation S p e c i f i c system mechanisms may also exhibit relaxation, which w i l l delay the completion of the morphological response, following a change of energy input, independently of any accompanying effect due to reaction. The time period separating the beginning of system displacement and the achievement of i t s new equilibrium state i s c a l l e d the "relaxation time" (Chorley and Kennedy, 1971; see Figure 1.4). "Relaxation time" may be interpreted physically as a function of, for example, sediment size i n that the f i n e s t , and hence most mobile, p a r t i c l e s should respond most rapidly to a change of input and should, therefore, have the shortest "relaxation time". Thus, cobble-gravel stream-beds have a longer "relaxation time" associated with th e i r complete morphologic (e.g. slope, bed-form) adjustment to a new equilibrium condition than finer-grained channels. Moreover, the pattern of relaxation may d i f f e r according to particle size. A single, slow adjustment may prevail i f the sediment supply is sufficiently coarse-grained to form an "armour" layer on the streambed surface, and thereby heavily damp any c u t - a n d - f i l l tendencies (see Figure 1.6); such a simple response can be viewed in Mackin's (1948) terms as an adjustment at "grade". In contrast, fine-grained channels may exhibit a FIGURE 1.6: Simple and complex relaxation paths with identical relaxation times. A [c — Relaxation time >| Time 17 relaxation path that repeatedly "overshoots" the new equilibrium condition (see Figure 1.6); t h i s i s compatible with Schumm's concept of "complex response", and may be manifest only in active, or "sensitive", systems. In a greater spatial and temporal context, a "relaxation time" of from 1,000 to 5,000 years may be associated with the decline in post-glacial rates of sediment mobilization (Church and Ryder, 1972) and isostatic u p l i f t (Bloom, 1978), from a maximum immediately after deglaciation. 1.5.3 Geomorphic recovery Finally, the ab i l i t y of a hydrologic event to affect the shape or form of the landscape depends upon not only the magnitude and frequency of the event, but also the magnitude of the processes which operate during the intervals between formative events and which tend to restore the surface of the landscape to the condition that existed previously (Wolman and Gerson, 1978); the time taken for a system to return to i t s o r i g i n a l state on the cessation of a short-term input (e.g. a hydrologic event) i s ca l l e d the "recovery time" (Chorley and Kennedy, 1971). The "recovery time" of the f l u v i a l system is controlled primarily by the rate of revegetation of floodplains and channel bars (Wolman and Gerson, 1978). The a v a i l a b i l i t y of moisture i n humid regions enables relatively rapid revegetation, and hence reconstruction of pre-disturbance channel characteristics (Wolman and Gerson, 1978). In contrast, the recup-erative capacity of channels i n ar i d regions i s l i m i t e d by low rates of revegetation, so that destructive floods produce nearly irreparable, and hence progressive, changes (Wolman and Gerson, 1978). Furthermore, in a given region, the production and persistence of landforms i s determined not only by the absolute magnitudes, frequencies, and durations of i n d i v i d u a l hydrologic events, but also by the i r d i s t r i b -18 ution and timing relative to one another (Beven, 1981). The length of time between events controls the possibility for recovery, while their ordering determines the extent of recovery, and thus, the availability of sediment for successive events (see Figure 1.7). Beven (1981) accounted for the effect of the ordering and timing of hydrologic events on geomorphic recovery by suggesting that system thres-holds change over time (see Figure 1.8). As arid areas are sparsely vegetated and undergo l i t t l e recovery, the ef f e c t of event ordering and timing i s more pronounced, and hence dynamic thresholds are most lik e l y to occur, in semi-arid areas (see Schumm and Lichty, 1963; Burkham, 1972) and humid temperate regions (see Newson, 1980; Beven, 1981); Schumm (1977) treated thresholds as s t a t i c over time, which may be adequate for ar i d systems. 1.6 Study objectives In summary, the concepts of "dynamic metastable equilibrium", "geomorphic thresholds", "episodic behaviour", and "complex response", as presented by Schumm and his students, were derived from ephemeral stream channels in semi-arid areas only. It i s , therefore, the objective of this thesis to examine the validity of Schumm's conceptual framework in a con-trasting geomorphic and climatic setting to that in which i t was conceived. Such an opportunity i s provided in the Queen Charlotte Islands (see Figure 2.1), which are characterized by abundant p r e c i p i t a t i o n , a per-humid s o i l moisture regime, perennial streamflow, varied rock l i t h o l o g i e s (many of which are rea d i l y erodible), and a high l e v e l of seismic and geomorphic a c t i v i t y . Despite the l a t t e r condition, the t y p i c a l l y steep and cobble-gravel bedded stream channels that drain forested mountain watersheds in the Queen Charlotte Islands usually are sediment supply-limited, so that exten-FIGURE 1.7: The i n t e r a c t i o n between sediment a v a i l a b i l i t y and the r e c o v e r y r a t e r e l a t i v e t o the f r e q u e n c y o f h y d r o l o g i c e vents (E) t h a t t r i g g e r geomorphic change ( a f t e r K e l s e y , 1982) . A. Recovery r a t e always f a s t e r than event f r e q u e n c y . Maximum E E E 9> T i lme B. Recovery r a t e always s l o w e r than event f r e q u e n c y . Maximum E E E E E E E E E E 5> T lme C. Combination o f c a s e s A and B. Maximum E E E E E E E E Time 20 FIGURE 1.8: Geomorphic threshold concepts (after Beven, 1981). Constant threshold Time 21 sive channel aggradation i s sporadic, in time and space, and transient. However, timber harvesting and road construction along the moun-tainous west coast of the Queen Charlotte Islands accelerated s o i l erosion on h i l l s l o p e s and sedimentation in streams (Wilford and Schwab, 1982; Schwab, 1983). As a r e s u l t , the Fish/Forestry Interaction Program was i n i t i a t e d to resolve management uncertainties in areas of intensive mass wasting (Poulin, 1983). During the i n i t i a l synoptic survey of watersheds in the Queen Charlotte Islands by the project f i s h e r i e s b i o l o g i s t (E. A. Harding, pers. comm., 1981), and from a subsequent air photo reconnaissance by the writer, i t was noted that four logged watersheds contained partic-ularly large, recent accumulations of f l u v i a l sediment (henceforth c a l l e d "sediment wedges") i n the mid- and down-stream reaches of their trunk streams (see Figure 1.9 and Plates 1.1 to 1.4). "Sediment wedges" are of the order of one kilometre in length, which i s an order-of-magnitude greater than the extent of channel aggradation typically produced by slope failures or by a pool-riffle sequence scaled according to channel width. A "sediment wedge" usually i s defined at i t s upstream end by a scoured reach of channel and at i t s downstream terminus by a steep snout which, unlike many debris flow deposits, i s not marked by a major log jam. The zone of deposition t y p i c a l l y i s an order-of-magnitude wider than the stream channel, which c h a r a c t e r i s t i c a l l y has a braided flow pattern (see Plates 1.2 to 1.4), and i s bordered by severely eroded stream banks (see Plates 1.4C and 1.5); the latter were logged by cross-stream f e l l i n g and yarding, and in-stream skid-ding, techniques. "Normal" f l u v i a l processes and landforms appear to pre-v a i l along the length of stream channel extending above and below a "sediment wedge". The appropriateness of Schumm's concepts to the description of f l u v i a l landforms in the Queen Charlotte Islands depends upon the relative 22 FIGURE 1.9: Diagrammatic r e p r e s e n t a t i o n o f a "sediment wedge" | | "Sediment wedge" Holocene v a l l e y f i l l Bedrock f^p^~ Channel may be i n c i s e d i n t o bedrock Stream t e r r a c e s B r a i d e d f l o w and l a t e r a l bank e r o s i o n "Wedge" r e a c h shown i n F i g u r e 4.14A "Wedge" a g g r a d a t i o n may o v e r - r u n channel banks Steep "wedge" snout PLATE l . l : Aerial photograph of Mountain Creek, before and after logging and "sediment wedge" i n c e p t i o n A. Before logging (1954-1955) B. A f t e r logging (1974) Scale: 1 cm - 324 m Scal e : 1 cm = 334 m P l a t e 1.2: A e r i a l photographs of streamside logging a c t i v i t i e s and "sediment wedge" growth i n Mosquito Creek t r i b u t a r y A. During logging (1964) Scale: 1 cm = 166 m B. A f t e r logging (1976) Scale: 1 cm = 206 m PLATE 1.3: Oblique a i r photograph of the "sediment wedge" i n Mosquito Creek t r i b u t a r y View lo o k i n g upstream from former bridge c r o s s i n g (at 0.27 Km i n Figure 4 .6 ) . Most stream flow has been d i v e r t e d from along the pr e - l o g g i n g channel (at bottom r i g h t of photo) to along the Branch 10 haul road (at bottom l e f t of photo) . Photo taken i n November 1981. 28 PLATE 1.4: Photographs of each of the "sediment wedges" A. Upper "sediment wedge" i n Armentieres Creek: view lo o k i n g upstream (from 1.6 Km i n Figure 4.5) at de-watered channel. B. "Sediment wedge" i n Mosquito Creek t r i b u t a r y : view l o o k i n g downstream (from 0.35 Km i n Figure 4.6) at the j u n c t i o n of the pre-logging channel and the Branch 10 haul road (at r i g h t of photo). Photo taken i n August 1982. PLATE 1.4 (contd.) C. "Sediment wedge" in Mountain Creek: view looking downstream (from 0.75 Km in Figure 4.7) at braided flow and lateral bank erosion (at left of photo). D. "Sediment wedge" in Lagins Creek tributary: view looking upstream (from 0.95 Km in Figure 4.8) at de-watered channel (former haul road). 30 PLATE 1.5: An example of severe stream bank erosion by- Mosquito Creek tributary View looking downstream from 1.25 Km in Figure 4.6. 31 influence of those geomorphic and hydrologic factors that encourage, and those that discourage, episodic and complex f l u v i a l behaviour (see Table 1.1). On the basis of these factors and the occurrence of unusual f l u v i a l phenomena (i.e. "sediment wedges") in four, streamside-logged water-sheds, i t is hypothesized that: 1. in forested watersheds, e x t r i n s i c geomorphic thresholds are rarely crossed, so that sediment transport-limited conditions, and hence major episodes of f l u v i a l disequilibrium, occur infrequently, 2. land use effects, specifically streamside logging activities, may cross an extrinsic geomorphic threshold, lower an int r i n s i c geomorphic thres-hold (stream bank s t a b i l i t y ) , and thereby i n i t i a t e major episodes of steam bank erosion and trunk stream aggradation ("sediment wedges"), 3. sediment transport i n severely aggraded reaches with "armoured" streambeds ("sediment wedges") i s episodic and complex, due to geo-morphic threshold, reaction, relaxation, and recovery effects. To test the validity of these hypotheses, three aspects of "sediment wedges" must be examined: 1. What are the principal sources of the sediment deposited in the "sediment wedges"? 2. How mobile are the "sediment wedges"? 3. What caused the appearance of the "sediment wedges" i n the stream channels? 32 TABLE 1.1: Factors for and against the efficacy of Schumm's concepts in the Queen Charlotte Islands FOR AGAINST 1. Large vertical r e l i e f 1. Perennial streamflow 2. Small drainage basin area 3. High seismic activity 2. Relatively uniform spatial and temporal distribution of rain-f a l l (on average) 4. Coarse-grained streambed sediment 5. Episodic sediment delivery by mass wastage events 3. Well-vegetated drainage basin 4. Sediment transporting capacity of stream is rarely exceeded 33 The steps involved i n addressing each of these questions are summarized, i n r e l a t i o n to the o v e r a l l study design, i n Figure 1.10. In brief, the major sediment transfer processes are examined directly using the methods outlined in Chapter 3.0, while contributions from minor processes are estimated by using sediment delivery rates measured elsewhere i n the Pacific Northwest; the latter are abstracted from the literature, which i s .reviewed in Appendix A and summarized in Tables 3.1 and 3.2. The resultant volume of sediment transferred to the stream channels i s compared with the volume of sediment presently stored in the "sediment wedges" by means of a sediment budget i n Section 4.1. The sediment output inferred from the sediment budget i s discussed in conjunction with f i e l d and air photo evi-dence of "sediment wedge" movement in the f i n a l part of Section 4.1, and the causes of "sediment wedge" appearance are addressed in Section 4.2. Chapter 5.0 synthesizes the results and examines their degree of correspondence with Schumm's f l u v i a l model via a critique of the three conceptual hypotheses. 34 FIGURE 1.10: Flow diagram of study design Morphology and sedirnentology of contemporary "sediment wedges" and stream bank dep-osits assessed by f i e l d studies Historical review of stream channel and stream bank stability and "sediment wedge" development evaluated by air photo analyses and personal communications t Hypothesize the relationship between "sediment wedges" and sediment sources Typical rates of sediment delivery by transfer processes in the Pacific Northwest evaluated from literature Select sediment sources and sediment transfer processes for quantitative examination Measure sediment contributions from major sources by air photo analysis of morphologic evidence and estimate contributions from minor sources by using appropriate Pacific Northwest rates Sediment volume contributed by sources compared to volume stored in "sediment wedges" via a sedi-ment budget Sediment mobility inferred Causes of "sediment wedges" inferred & Conceptual implications 35 2.0 STUDY WATERSHEDS As indicated in the preceding chapter, several logged watersheds in the Queen Charlotte Islands have experienced slope erosion and stream sedimentation problems. The term "study watersheds" refers to the four watersheds (namely Armentieres Creek, Mosquito Creek tributary, Mountain Creek, and Lagins Creek tributary) that contain abnormally large f l u v i a l sediment accumulations (i.e. "sediment wedges") in their trunk streams. 2.1 Location The Queen Charlotte Islands are divided into three natural physiographic units which from east to west, and i n order of increasing vertical r e l i e f , are the Queen Charlotte Lowlands, the Skidegate Plateau and the Queen Charlotte Ranges (Sutherland Brown, 1968; see Figure 2.1). The four study watersheds l i e within the Queen Charlotte Ranges, and their locations are shown in Figure 2.1. Mountain Creek and Lagins Creek tributary share the same head-water divide on Graham Island, but the former discharges directly into Rennell Sound to the north, while the latter flows into the main stem of Lagins Creek and thence into Long Inlet to the southeast. Moresby Island contains the Mosquito Creek tributary and Armentieres Creek watersheds. The former flows northward from Mount Moresby to i t s confluence with the main stem of Mosquito Creek, while the latter empties into Chaatl Narrows to the north. 36 FIGURE 2.1: L o c a t i o n o f the study watersheds i n the Queen C h a r l o t t e I s l a n d s 2.2 Geology 2.2.1 Glacial history In the Pleistocene Epoch, the Queen Charlotte Islands were over-ridden by ice. During the Fraser Glaciation, ice developed as an independ-ent accumulation centre i n the Queen Charlotte Ranges (Sutherland Brown, 1968; Alley and Thomson, 1978). This most recent glaciation culminated in an ice-cap more than 900 metres thick, with some mountain peaks projecting as nunataks, and some coastal areas remaining ice-free (Sutherland Brown, 1968; Clague, 1981; Clague et al., 1982a, 1982b); during this glacial phase, the sea level was lower, relative to the land, than at present, as indicated by former shorelines and associated deposits (Clague, 1981; Clague et al., 1982a, 1982b), archaeological data (Hobler, 1978; Fladmark, 1979), and biological evidence of glacial refugia (Sutherland Brown, 1968; Mathewes and Clague, 1982). The melting of the ice cover, approximately 11,000 years ago, exposed the glacial d r i f t deposits to a variety of geomorphic processes, and unvegetated, glacially oversteepened hillslopes would have been especially susceptible to erosive forces (Church and Ryder, 1972; Alley and Thomson, 1978). Therefore, Pleistocene deposits are poorly preserved on steep valley sides and in upland areas in the Queen Charlotte Islands (Alley and Thomson, 1978), and the most complete g l a c i a l history i s recorded i n the g l a c i o -f l u v i a l sediments on the eastern lowlands (Sutherland Brown, 1968; Alley and Thomson, 1978; Clague et a l . , 1982a). Although there i s l i t t l e d i r e c t evidence of early Holocene geomorphic a c t i v i t y i n the Queen Charlotte Ranges, the general pattern probably i s similar to that suggested for other mountainous parts of western British Columbia (Ryder, 1971a, 1971b; Church and Ryder, 1972; Clague, 1981). Thus, i t i s probable that immediately 38 following deglaciation of the Queen Charlotte Ranges, large volumes of glacial d r i f t were eroded rapidly from glacially oversteepened hillslopes by mass wastage processes. As the supply of e a s i l y erodible d r i f t became exhausted, the geomorphic tempo declined. It i s speculated that the "relaxation time" of these "paraglacial" processes was less than 3,000 years in the Queen Charlotte Ranges. As the Islands had only a thin ice cover during the Fraser Glaciation, and there are no glaciers at present, renewed glacier growth i s unlikely to have occurred during any Neoglacial episode. Thus, for the past 8,000 years at least, the study watersheds have experienced rates of sedi-ment production and transfer conditioned by subaerial processes. Sea level has fallen continuously, relative to the land, during this period, and tid a l records indicate that i t i s currently f a l l i n g at almost 2 mm/yr (Clague et a l . , 1982b). 2.2.2 Bedrock and tectonics The geology of the Queen Charlotte Islands i s detailed by Sutherland Brown (1968). The predominant rock types i n the four study basins are shown in Table 2.1 and discussed briefly below; unless otherwise indicated, the following information i s quoted from Sutherland Brown (1968). Both catchments on Moresby Island are underlain by the Karmutsen Formation, a thick pillowed basalt of Late Triassic age, which i s the funda-mental s t r u c t u r a l unit of the Queen Charlotte Islands. S p e c i f i c a l l y , Mosquito Creek tributary i s underlain by pillow lavas, and Armentieres Creek by massive lavas and derived amphibolites. In addition, both basins have outcroppings of plutonic rocks. The metamorphosed b a s a l t s i n the Armentieres watershed contain lens-shaped syntectonic plutons of Jurassic age and are composed of medium- to coarse-grained foliated hornblende and TABLE 2.1; General information for the study watersheds Physical Characteristic Armentieres Watershed Mosquito trib. Lagins trib, Mountain Drainage Area 3 (km2) Drainage Length^ (Km) Drainage Re l i e f 0 (m) Average Gradient^ j>- Drainage Density (Km/Km ) Highest Stream Order Major Rock Type 3.93 2.75 490 0.18 4.2 4 5.38 5.05 860 0.17 6.1 4 Basalt flows, pillow lavas, pillow breccia and tuff, minor limestone, volcanic sand-stone and shale 5.92 5.60 770 0.14 5.9 4 Dark grey calcareous siltstone, greywacke conglomerate and minor volcanic rocks 12.64 7.30 910 0.12 6.7 Quartz monzonite granite, granod-io r i t e , and quartz diorite Drainage Area: Drainage Length: Drainage Relief: Average Gradient: Highest Stream Order: area on a map enclosed by the watershed boundary length of the main channel from i t s outlet to the basin divide height difference between the main channel at i t s outlet and the highest point at the basin divide. ratio of drainage relief to drainage length, after Strahler (1952b). quartz diorite with migraatitic border zones. Although outcrops are mapped by Sutherland Brown (1968) in the headwaters, f i e l d investigations by the writer revealed their occurrence at lower elevations also. The small, post-tectonic pluton, of Cretaceous or Tertiary age, emplaced in Mount Moresby i s chiefly fine- to medium-grained quartz monzonite. Mountain Creek and Lagins Creek tributary catchments are under-lain mainly by plutonic and sedimentary rocks, respectively, with the latter geologic unit extending into the headwaters of Mountain Creek. The post-tectonic pluton, of Cretaceous or Terti a r y age, present under most of Mountain Creek i s composed of fine-grained quartz monzonite and granite and, due to contact metamorphism, i s associated with a hornfelsic aureole. The sedimentary unit underlying Lagins Creek tributary i s the Longarm Formation of Cretaceous age, which i s composed primarily of dark-grey calcareous siltstones and l i t h i c fine greywackes. The rock formations discussed above usually are well bedded, jointed, or fissured, thereby promoting the penetration of water, and assoc-iated weathering, to great depths (Sutherland Brown, 1968; Alley and Thomson, 1978). As a result of these weathering processes, the bedrock i s highly susceptible to erosion and contributes to the high rate of geomorphic activity prevalent in the Queen Charlotte Islands (Alley and Thomson, 1978). The Queen Charlotte Islands have a long history of f a u l t i n g , as r e f l e c t e d i n the geology and d i s t r i b u t i o n of recent seismic shocks (Sutherland Brown, 1968). In particular, the Rennell Sound-Louscoone Inlet Fault Zone, in the past, has controlled the d i s t r i b u t i o n of the Karmutsen and Longarm Formations, and the location and genesis of syntectonic and post-tectonic plutons (Sutherland Brown, 1968). The Mountain Creek and Lagins Creek tributary basins l i e along this fault zone, and consequently, contain faults and linears which trend northwest-southeast. 41 The Armentieres Creek catchment has a f a u l t along i t s major tributary valley, but i t is not associated with any of the three major sub-parallel fault systems present in the Queen Charlotte Islands. Mosquito Creek tributary has no known faults in i t s watershed. Movements on the Queen Charlotte Fault, which i s part of the circum-Pacific continental margin fault system, are responsible for the highest incidence of seismic activity in Canada (Sutherland Brown, 1968). Between 1899 and 1974 there were 1,268 earthquakes recorded by seismic means for Graham Island, of which 41 could be f e l t by humans (i.e. those equal to and greater than Richter magnitude 4.0) (Alley and Thomson, 1978). The largest earthquake known to have occurred in Canada was a magnitude 8.0 event in 1949, with the epicentre located just off Graham Island (Milne et a l . , 1978). The magnitude-frequency relationship of earthquakes, equal to and greater than magnitude 4.0, generated along the Queen Charlotte-Fairweather (i.e. southeast Alaska) fault system between 1899 and 1975 i s discussed by Milne et a l . (1978). Alley and Thomson (1978) believe that on tenuously stable h i l l -slopes, the shaking stress of earthquakes triggers widespread mass wasting; in particular, the wet conditions preceding the 1949 earthquake probably aggravated the occurrence of slope failures (E. J. Karanka, pers. comm., 1984). 2.2.3 Physiography The influence of bedrock geology and glacial history i s expressed topographically by the steep valley side slopes in the four study water-sheds. Typically, the slopes are concave or straight with gradients ranging from 0.5 (25°) on the footslopes to more than 1.7 (60°) in the headwaters, 42 and averaging between 0.7 (35°) and 0.9 (40°) on the raid-slopes. However, the width of the valley bottoms varies considerably: around half of the Lagins Creek tributary catchment i s low gradient t e r r a i n , so that slope transfer of sediment reaches the creek only in the headwaters; in contrast, Armentieres Creek, Mountain Creek and Mosquito Creek tributary are bordered by a narrow floodplain zone and restricted footslope area, thus allowing the transfer of material from the steep valley walls d i r e c t l y into the creek. In general, the zone of r e l a t i v e l y f l a t t e r r a i n that separates the slopes from the streams increases in width downstream. Cha r a c t e r i s t i c s of the drainage basins are summarized in Table 2.1 and t h e i r means of derivation are indicated beneath. The information was abstracted from 1: 50,000 scale topographic maps and, in order to deter-mine the drainage densities, additional data were obtained from 1: 16,000 scale air photographs; due to their greater resolution, photos were used to evaluate densities of lower order channels i n clearcut areas, i n a manner similar to that of Reid (1981). 2.2.4 Surficial materials The scarcity of a l l u v i a l fan deposits and a l l u v i a l paleo-terraces i n the study watersheds at the present time, suggests that l i t t l e of the sediment mobilized during the "paraglacial" period remains today in the valley bottoms. This may be the combined r e s u l t of the small size and steepness of the study watersheds (see Table 2.1). Most of the sediment probably was evacuated directly to the sea during and immediately following deglaciation when the magnitude, frequency and duration of melt runoff would have been greatest. F i e l d investigations of s u r f i c i a l deposits by the writer were restricted to examinations of stream bank materials and hillslope deposits adjacent to landslide scars. In the study watersheds, the banks consist primarily of f l u v i a l deposits, occasionally underlain by t i l l or bedrock and overlain by colluvium; the f l u v i a l material usually i s poorly sorted and s t r a t i f i e d , with moderately-well imbricated sub-angular to sub-rounded gravels of varied lithologies, in a loose or weakly cemented fine to coarse sand matrix. T i l l i s more compact and finer textured than f l u v i a l deposits, even after weathering, and limited exposures are found along Mountain Creek and Mosquito Creek tributary. Colluvium can be derived from bedrock or t i l l , though more frequently the former, and i s coarser textured than f l u v i a l deposits, with angular to sub-angular gravels and cobbles in a sandy matrix. At the mouths of Armentieres Creek and Lagins Creek, f l u v i a l materials are underlain by limited exposures of sandy gravel deltaic dep-o s i t s , and s i l t y sand marine deposits, respectively. The l i t h o l o g y and texture of a l l s u r f i c i a l materials i s influenced strongly by the parent material from which i t was derived, and consequently, i s highly variable. In addition, peat deposits and possible buried paleosols are present in several stream bank exposures along Armentieres Creek, Mountain Creek and Lagins Creek tributary (see Plate 2 .1 ) . Organic s i l t horizons that are laterally continuous for several metres along the stream bank are presumed to be buried paleosols, rather than weathered sediments, on the basis of th e i r very high organic matter content and the discolouration of the underlying parent material. In the study watersheds, apparent paleosols are overlain either by re-worked colluvial and morainal deposits (implying h i s t o r i c a l mass movements)(e.g. Armentieres Creek, Mountain Creek) or by sandy f l u v i a l sediments (e.g. Mountain Creek, Lagins Creek tributary). In the latter case, paleosols often occur as cumulic horizons (see Plate 2.1B) or are associated with thick peat deposits. Peat develops i n low-energy, 44 PLATE 2.1: Examples of organic stream bank deposits along Mountain Creek A. Thick basal peat layer overlain by f l u v i a l sands and recent accumulation of organic matter. Section located 3.5 Km from basin outlet. B. Fluvial sands and gravels overlain by thick sand deposit containing seven black organic s i l t horiz-ons (possible buried pal-eosols) . Section located 1.0 Km from basin outlet. water-saturated environments, where the decomposition of organic matter is slow and relatively undisturbed. In contrast, the alternation of thick sand deposits and thin organic s i l t horizons (supposedly buried paleosols) im-plies recurrent overbank flood deposition and s o i l formation. Neither the presumed paleosols nor the peat deposits have been dated absolutely, but from rates of contemporary s o i l development, the time required for the paleosols to have developed rudimentary profile characteristics can be estimated (see below). More accurate dating by s o i l stratigraphic tech-niques i s not yet possible in terrain sculpted by frequent, shallow debris slides (Harden et a l . , 1982). The materials mantling the slopes in the study watersheds are largely unknown, except for Mosquito Creek tributary which was mapped in 1979 by T. Lewis (MacMillan Bloedel Ltd.); here, steep bedrock slopes are covered by a colluvial veneer, except near the confluence with Mosquito Creek where the footslopes are masked by a morainal blanket. In the other study watersheds, landslide scars on steep, upper slopes are juxtaposed primarily by a thin colluvial veneer overlying bedrock, while the footslopes are mantled usually by a colluvial blanket; the terms "veneer" and "blanket" denote deposits less than, and more than, 1 metre thick, respectively (Environment and Land Use Committee' Secretariat, 1976). The abundance of colluvium in the study watersheds, and throughout the Rennell Sound region (Alley and Thomson, 1978; Wilford and Schwab, 1982; Schwab, 1983), reflects the intensity of weathering and mass wasting processes in this inherently unstable terrain. In the Queen Charlotte Ranges, the soils have a high moisture content for most of the year, resulting in Regosols, Brunisols or Ferro-Humic Podzols, depending on their state of development (Valentine et a l . , 1978; Schwab, 1983); s o i l development i s hindered by weathering and erosion processes and by the windthrow of large trees (Valentine et a l . , 1978). However, the nature of the s u r f i c i a l material has an important influence on the character of the s o i l : soils developed on colluvium are usually deeper and better drained than those on morainal and gravelly f l u v i a l materials, and lack any cemented horizons (Valentine et a l . , 1978). Due to the per-humid s o i l moisture regime i n these coniferous forests, there i s a net accumulation of organic matter, and Folisols develop in association with the Ferro-Humic Podzols (Valentine et al., 1978). The time required by a s o i l to approach a steady-state condition w i l l vary with the s o i l property being studied, the parent material, and the particular type of s o i l profile that develops in a particular environment (Birkeland, 1974, p.175). In general, Podzols and A-horizon properties form rapidly, and attain steady-state within 100 to 1,000 years (Birkeland, 1974, Figure 8.17), while soils of the Brunisolic order develop rudimentary Podzol characteristics, such as an LFH l i t t e r layer and an underyling Bm-horizon that i s browner or redder than the parent material (Valentine et al., 1978), i n even less time. For example, Ugolini (1968) found that in Glacier Bay, Alaska, freshly deposited t i l l developed an i n c i p i e n t B-horizon a f t e r 55 years and had attained true Podzol characteristics within 150 years. On the west coast of Vancouver Island, B r i t i s h Columbia, sandy beach deposits developed s o i l profiles with Orthic Dystric Brunisol and Orthic Humo-Ferric Podzol c h a r a c t e r i s t i c s after 130 years and 370 years, respectively (Singleton, 1978). Both these regions have a similar climate to that in the Queen Charlotte Ranges, so that soils in the study watersheds should attain recognizable Podzol profiles within 150 to 400 years (L. M. Lavkulich, pers. comm., 1984). The evidence for buried paleosols i s based largely upon the 48 existence of black s i l t y organic horizons, which probably represent earlier LFH l i t t e r layers as, in general, the organic matter in the A-horizon does not p e r s i s t after b u r i a l (Birkeland, 1974, p.24). Because organic matter probably reaches steady-state more rapidly than any other s o i l property (Birkeland, 1974, p.163), the uppermost horizons provide only a minimum estimate of the age of a paleosol. Generally, the buried B-horizon i s the most important for recognizing buried soils (Birkeland, 1974, p.24), and may indicate the degree to which steady-state i s approached by the paleosol profile as a whole. In lower Mountain Creek, for example, discolouration of the parent material between the organic horizons in the stream bank deposits (see Plate 2.IB) implies that Brunisol or Podzol p r o f i l e s have developed, and thus, that each paleosol i s at least 150 to 400 years old; although seven organic horizons are indicated in Plate 2.IB, additional horizons may have been removed by erosion or may be indistinct, due to s o i l immaturity or disturbance, so that the entire sand body i s at least 1,000 years old. Thus, during the past few millenia, pedogenic processes in the downstream reaches of Mountain Creek have been interrupted sporadically by s o i l burial. Along Armentieres Creek and Lagins Creek tributary, buried paleosols occur only singly or i n pairs, and are less well developed than those shown in Plate 2.IB. The lack of cumulic horizons implies that there were fewer episodes of s o i l formation than i n Mountain Creek, or that a number of buried s o i l horizons have been removed by erosion or are too poorly dev-eloped to enable positive field.identification. However, as s o i l burial by individual landslide and overbank flood events i s limited in areal extent, the number of buried paleosols typically should vary along a creek as well as between creeks. 49 2.3 Climate and vegetation The following discussion i s necessarily general due to the lack of specific information on the study watersheds. 2.3.1 Precipitation The climate of the Queen Charlotte Islands i s per-humid, with cool summers and mild winters. Annual precipitation varies from 1,500 mm on the eastern plains to more than 5,000 mm on some west coast slopes (Alley and Thomson, 1978). Snow i s ephemeral at sea level and i s confined mainly to elevations above 600 metres (Alley and Thomson, 1978). The v a r i a t i o n i n t o t a l annual p r e c i p i t a t i o n from 1964 to 1983, and the seasonal distribution of mean annual precipitation, in Tasu Sound (see Figure 2.1) i s summarized in Figures 2.2 and 2.3, respectively. It i s apparent i n Figure 2.2 that the frequency d i s t r i b u t i o n of t o t a l annual precipitation i s approximately symmetrical about the mean annual precipit-ation amount (4,220 mm). However, Figure 2.3 illustrates that, like much of the P a c i f i c Northwest (Harr, 1983), there i s an uneven d i s t r i b u t i o n of annual precipitation, with roughly 70% occurring from October through March. Tasu Sound i s located on the west coast of the Queen Charlotte Ranges and probably receives more precipitation than the study watersheds. The difference, however, i s unl i k e l y to induce changes i n the t o t a l or seasonal p r e c i p i t a t i o n d i s t r i b u t i o n s , other than reducing the mean annual precipitation amount. Between January 1972 and December 1982, 58 storm events in Tasu Sound had a magnitude greater than 80 mm i n 24 hours. Analysis of t h i s p a r t i a l - d u r a t i o n series indicates that r a i n f a l l amounts between 100 and 120 mm i n 24 hours recur, on average, once or twice per year (see Figure 50 FIGURE 2.2: Frequency distribution curve of total annual precipitation in Tasu Sound, 1964-1983 (data provided by Environment Canada, Atmospheric Environment Service). x u c 3 cr co u 25 _ 20 _ 15 _ 10 Normal distribution curve 3200 3500 3800 4100 4400 4700 5000 Total annual precipitation (mm) 5300 FIGURE 2.3; Seasonal distribution of mean annual precipitation in Tasu Sound, 1964-1983 (data provided by Environment Canada, Atmospheric Environment Service). o rt 4-> •H o< •H o 200 -i 180 160 140 -120 100 80 line fitted by eye -i 1 r~ 0.1 0.5 1.0 5.0 Recurrence interval (years) 10 20 52 2.4). Caine (1980) found that shallow slope failures in mountainous areas of the world, on average, were triggered by 70 to 100 mm of r a i n f a l l within a 12 to 24 hour period. In Rennell Sound, storms known to have initiated mass wasting generally exceed 120 to 150 mm within a 12 to 24 hour period (Schwab, 1983) and occur annually (Wilford and Schwab, 1982); for example, 264 mass movements resulted from the October 30, 1978 storm, which had two distinct 12 hour periods of intense r a i n f a l l (120 mm and 110 mm) (Schwab, 1983). 2.3.2 Wind The frequency of occurrence of destructive v/inds i s greater in the Queen Charlotte Islands than elsewhere in Canada, with recorded gusts approaching 200 km/hr (Alley and Thomson, 1978). On the west coast, Wilford and Schwab (1982) found that although the winds generally originate in the southwest, trees are blown down by both up- and down-valley winds. 2.3.3 Vegetation The Mosquito Creek tributary and Lagins Creek tributary water-sheds l i e within the Coastal Western Hemlock biogeoclimatic zone, which is the most productive forest zone in British Columbia (Valentine et a l . , 1978; Krajina et a l . , 1982; Banner et a l . , 1983). Sitka spruce (Picea sitchensis), western hemlock (Tsuga heterophylla) and western red cedar (Thuja plicata) are the main forest types, with mountain hemlock (Tsuga mertensiana) and Alaska yellow cedar (Chamaecyparis nootkatensis) present to a lesser extent (Krajina et a l . , 1982). The Mountain Creek and Armentieres Creek watersheds l i e within the wetter Coastal Cedar-Pine-Hemlock zone, which supports low productivity mixtures of red and yellow cedar, western and mountain hemlock, and shore 53 pine (Pinus contorta var. contorta) (Banner et a l . , 1983). The dominant deciduous species i s red alder (Alnus rubra) which rapidly colonizes disturbed sites, such as stream banks, mass movement tracks and abandoned logging roads (Wilford and Schwab, 1982); although this results in the short-term loss of productive coniferous forest sites, the rapid regeneration of a vegetative root system aids in stabilizing residual soils in areas of intensive mass wasting. This is especially c r i t i c a l as the cover of shrubs and herbs i s less than 20% (Smith et a l . , 1983), and the majority of tree roots are confined to surface organic s o i l horizons, seldom extending below 30 cm on well drained sites or below 15 cm on more poorly drained sites (Schwab, 1983). For the Oregon Cascades, Swanson et a l . (1982b) hypothesized that in the f i r s t 5 to 10 years following watershed disturbance, deciduous r i p -arian species, such as red alder, developed more rapidly than conifers on the h i l l s l o p e s ; a f t e r 30 to 60 years, the growth and expansion of the upslope conifers would suppress the deciduous component, due to their low shade tolerance, and establish the successional trend of coniferous domin-ance along streams (Swanson et a l . , 1982b). A similar succession exists in the Queen Charlotte Islands, where 85% vegetation cover i s reached in 40 years in logged areas and in 100 years on landslide surfaces, accompanied by a switch from a predominantly deciduous to coniferous plant community (Smith et a l . , 1983). At present, the logged stream banks and valley fl a t s in the study watersheds support a dense cover of red alder interspersed with young conifers and an understory of shrubs and herbs, while the logged slopes are revegetated primarily by shrubs, herbs and young conifers. Forest cover maps compiled by British Columbia Forest Service personnel indicate that currently forested slopes in the study watersheds support conifers more than 250 years old, and Smith et a l . (1983) found 300 year-old stands elsewhere 54 in the Queen Charlotte Islands; prior to logging, almost the entire area of the study watersheds had a tree cover more than 250 years old. 2.4 Timber harvesting activities The size, location, and date of clearcut areas and roads in the study watersheds, obtained from air photo analyses, are shown in Table 2.2 and Figures 2.5 to 2.8, and logging act i v i t i e s are summarized in Table 2.3. The following sub-sections describe the logging history of each basin. 2.4.1 Armentieres Creek The Armentieres Creek basin has been logged by four operators, namely Goodwin Johnson Ltd. from 1962-1963, Bernie Ralph from 1964-1965, Paul Egger from 1967-1968 and fi n a l l y by Goodwin Johnson Ltd. in 1969 (G. Johnson, pers. comm., 1983; W. Funk, pers. comm., 1983). The only inform-ation available about the early years of logging, near the mouth of the basin, i s that skidding was conducted along t r a i l s , rather than in the creek (W. Funk, pers. comm., 1983; confirmed by air photo analyses) and that cross-stream f e l l i n g was avoided (G. Johnson, pers. comm., 1983). More is known about the logging of the upper watershed between 1967 and 1969: the area was tractor skidded using skid t r a i l s (G. Johnson, pers. comm., 1983; W. Funk, pers. comm., 1983; confirmed from f i e l d investigations and air photo analyses), except where the stream channel and banks were ill-defined as the result of i t s naturally multi-branched condition (G. Johnson, pers. comm., 1983; confirmed by air photo analyses) and/or i t s redirection, during the spring freshets, along skid t r a i l s (G. Johnson, pers. comm., 1983). 55 TABLE 2.2: Logging histories of the study watersheds Basin Date Area Logged Road Length Road Logging Watershed Area Logged Width Method (Km2) Km2 % basin Km Km/Km2 (m) Armentieres 3.93 u, Mosquito °^ tributary 5.38 1962-65 1967-69 Total 1960's 0.29 0.46 0.75 1.08 7.4 11.7 19.1 20.1 0.77 0.43 1.20 5.01 0.20 0.11 0.31 0.93 4.6 5.6 5.0e 9.4 Hand Skidded High-lead Mountain 12.64 1958 1965-67 Total 0.06 1.18 1.24 0.5 9.3 9.8 6.73 0.53 7.5 Skidded High-lead Lagins tributary 5.92 1970-74 1975-77 Total 0.63 10.6 5.50 0.93 11.2 Skidded High-lead aMean: weighted in proportion to the corresponding length of road. TABLE 2.3: Logging activities in the study watersheds Logging Cross- Cross- Avulsion Bridge Watershed to Stream Stream In-Stream Along Collapse Streambank Felling Yarding Skidding Road Armentieres Mosquito -j tributary Mountain Lagins tributary 2 2 0 1 2 0 KEY 2 : extensive 1 : limited 0 : none FIGURE 2.5: Logging a c t i v i t i e s i n the Armentieres Creek watershed FIGURE 2.6: Logging a c t i v i t i e s i n the Mosquito Creek t r i b u t a r y watershed 58 FIGURE 2.7: Logging activities in the Mountain Creek watershed ide FIGURE 2.8: Logging activities in the Lagins Creek tributary watershed 59 2.4.2 Mosquito Creek tributary This watershed was clearcut logged by Rayonier Canada Ltd. (now Western Forest Products Ltd.) beginning in 1963, using high-lead yarding and haul roads (H. Hanson, pers. comm., 1983). As seen from Figure 2.6, the upper two-thirds of the basin contains a single haul road along one side of the creek, therefore necessitating the cross-stream yarding of timber from the opposite slopes. Cross-stream f e l l i n g was avoided whenever possible (H. Hanson, pers. comm., 1983), so that the large amount of in-stream timber v i s i b l e i n the 1964 a i r photos (see Plate 1.2A) may be att r i b u t a b l e to yarding, rather than f e l l i n g , practices. Also of si g n i f i c a n c e are the bridge construction a c t i v i t i e s i n the basin: the most downstream bridge was built in 1962 from approximately 1500 m^ of gravel, and collapsed following logging operations, due partly to i t s sub-standard construction (H. Hanson, pers. comm., 1983). As a result of the large input of gravel, the creek bed aggraded downstream (see Chapter 4.0) and the flow was diverted northward along Branch 10 haul road (see Figure 2.6 and Plate 1.2). Additional sediment probably was supplied to the creek at the former bridge crossing by the surface erosion of convergent haul roads (H. Hanson, pers. comm., 1983). If the three other bridge crossings were constructed similarly, their subsequent collapse would have introduced into the creek a considerable volume of gravel, which may have contributed to the growth of the "sediment wedge". 2.4.3 Mountain Creek The Mountain Creek watershed tree farm licence was owned by British Columbia Forest Products Ltd. (now part of CIPA Industries Ltd.) and i t was clearcut logged in two stages: beginning in 1958, timber was skidded 60 along the multi-branched lower reaches of the creek, u n t i l the Fish and W i l d l i f e Branch enforced the use of high-lead yarding and haul roads to harvest the remaining timber, between 1965 and 1967 (V. Wellburn, pers. comm., 1983; confirmed by a i r photo analyses). From the 1966 a i r photos, cross-stream f e l l i n g i s evident, and as the only haul road l i e s on the east side of the creek, timber from the opposite slopes was yarded across the stream. In the lower reaches, a branch road crosses the creek, but neither f i e l d investigations nor air photo analyses indicated the construction of a bridge. In the upper reaches, the haul road was constructed close to the stream, so that, at present, the creek flows along sections of the road. 2.4.4 Lagins Creek tributary This basin was logged by Robert Leblanc, on contract to Goodwin Johnson Ltd., between 1970 and 1977 (G. Johnson, pers. comm., 1983). From the start of operations until the end of 1974, the steep slopes were high-lead yarded, while the flat t e r ground was clearcut by high-lead and skidder methods (D. McDiarmid, pers. comm., 1983; G. Johnson, pers. comm., 1983); timber was hauled to the beach by a cat and arch skid on the east side of the creek (D. McDiarmid, pers. comm., 1983) and along the creek i n i t s multi-branched upper reaches (G. Johnson, pers. comm., 1983). In 1975, the Fish and W i l d l i f e Branch prohibited cross-stream f e l l i n g and in-stream skidding, so skid t r a i l s were used i n the upper reaches (G. Johnson, pers. comm., 1983) and haul roads were b u i l t elsewhere (G. Johnson, pers. comm., 1983; D. McDiarmid, pers. comm., 1983). The latter were never culverted or elevated properly and, in some areas, the road was lower than the surround-ing ground (D. McDiarmid, pers. comm., 1983). As a res u l t , when the main bridge collapsed during a 1976 winter storm (G. Johnson, pers. comm., 1983; D. McDiarmid, pers. comm., 1983), Lagins Creek tributary, and many of the smaller ones, began flowing down the haul road (D. McDiarmid, pers. comm., 1983), which, as f i e l d investigations show, has become the contemporary stream channel. Truck hauling was completed along an upgraded haul road on the west side of the creek, without further major stream disturbance (D. McDiarmid, pers. comm., 1983). In conclusion, ground yarding and poor road construction influenced s i g n i f i c a n t l y the drainage network within t h i s watershed (D. McDiarmid, pers. comm., 1983). 62 3.0 STUDY METHODOLOGY This chapter outlines the methods used to estimate the volume of sediment delivered to the study creeks and the morphology of the "sediment wedges"; the relation to the overall study design i s summarized in Section 1.6 and Figure 1.10. 3.1 Review of sediment sources and mobilization processes: Summary In t h i s thesis, "sediment production" refers to the amount of sediment moved any distance by a process, while "sediment delivery" denotes only the amount that reaches the stream channel. Rates of sediment delivery by different transfer processes, ab-stracted from published r e s u l t s of other P a c i f i c Northwest studies, are reviewed i n Appendix A and are summarized i n Table 3.1; as s i t e - s p e c i f i c factors cause delivery rates to vary greatly within the Pacific Northwest, a range of rates, rather than an average rate, i s given for each process. In general, the most important processes are landslides, stream bank erosion, s o i l creep in logged terrain, and road surface erosion. From preliminary air photo and f i e l d investigations, landslides and riparian erosion also appear to be major sediment delivery processes in the study watersheds; hence, the delivery of sediment to the study creeks by these processes i s measured directly from air photos. The delivery rates of s o i l creep and road surface erosion shown i n Table 3.1 were measured in basins that are clearcut to a greater extent, and on steeper slopes, and have greater road densities than the study watersheds; thus, the sediment input per unit length of channel by these processes probably would be lower in the study basins, especially along stream reaches that are bordered by a floodplain. Therefore, the ranges of sediment production rates shown i n 63 T A B L E 3 . 1 : R e p r e s e n t a t i v e r a t e s o f s e d i m e n t d e l i v e r y i n t h e P a c i f i c N o r t h w e s t P r o c e s s W a t e r s h e d C o n d i t i o n R a t e o f S e d i m e n t D e l i v e r y t o S t r e a m ( m ^ / c h a n n e l K m / y r ) S o u r c e s S o i l c r e e p F o r e s t e d 0 . 9 - 4 D i e t r i c h & D u n n e ( 1 9 7 8 ) , L e h r e ( 1 9 8 1 , 1 9 8 2 ) , S w a n s o n e t a l . ( 1 9 8 2 a ) . L o g g e d 6 - 4 0 B a r r & S w a n s t o n ( 1 9 7 0 ) , S w a n s t o n ( 1 9 8 1 ) . T r e e t h r o w F o r e s t e d 0 . 1 - 1 R e i d ( 1 9 8 1 ) , S w a n s o n e t a l . ( 1 9 8 2 a ) . ' L a n d s l i d e s 3 F o r e s t e d 3 - • 1 0 R e i d ( 1 9 8 1 ) , M a d e j ( 1 9 8 2 ) , L e h r e ( 1 9 8 1 , 1 9 8 2 ) . L o g g e d a n d R o a d e d 9 - 2 0 b R e i d ( 1 9 8 1 ) , M a d e j ( 1 9 8 2 ) . S u r f a c e E r o s i o n F o r e s t e d 0 . 4 - 0 . 9 L e h r e ( 1 9 8 1 , 1 9 8 2 ) , S w a n s o n e t a l . ( 1 9 8 2 a ) . S l i d e s c a r s 0 . 9 - 5 R e i d ( 1 9 8 1 ) , L e h r e ( 1 9 8 1 , 1 9 8 2 ) . R o a d s 5 - 3 0 R e i d ( 1 9 8 1 ) , L e h r e ( 1 9 8 1 , 1 9 8 2 ) . S t r e a m B a n k s F o r e s t e d 2, - 1 6 0 c R e i d ( 1 9 8 1 ) , L e h r e ( 1 9 8 1 , 1 9 8 2 ) , T o e w s & M o o r e ( 1 9 8 2 b ) . L o g g e d 1 6 0 - 4 2 0 c T o e w s & M o o r e ( 1 9 8 2 b ) . a T h i s t e r m i n c l u d e s d e b r i s s l i d e s , a v a l a n c h e s , f l o w s , a n d t o r r e n t s . H i g h e r r a t e s o f s e d i m e n t p r o d u c t i o n r e p o r t e d b y o t h e r s t u d i e s ( s e e T a b l e A . 6 ) c a n n o t b e e x p r e s s e d i n m / c h a n n e l K m / y r b e c a u s e d r a i n a g e d e n s i t i e s a r e n o t c i t e d . H i g h d e l i v e r y r a t e s r e p o r t e d b y T o e w s a n d M o o r e ( 1 9 8 2 b ) r e f l e c t a v e r a g i n g o v e r s h o r t s t u d y r e a c h e s r a t h e r t h a n a l o n g t h e e n t i r e d r a i n a g e n e t w o r k . 6 4 TABLE 3.2: Process rates used in this study Process Watershed Condition Rate of Sediment Production Source Soil creep Forested slopes 1 - 3 mm/yr Table A.l Tree throw Surface erosion Logged slopes Forested slopes Logged slopes Forested slopes Logged slopes Slide scars Roads: During use First year of dis-use Abandoned 2 - 5 mm/yr 1 - 2 mm/yr n i l Table A.2 Table A.3 Swanson et a l . (1982a) 4-10 m3/Km2 basin/yr Swanson et al. (1982a) Megahan & Kidd (1972b) Lehre (1982). 16 m3/Km2 basin/yr Megahan & Kidd (1972b) 1,000-4,000 m-^ /Km^ scar/yr 10,000-15,000 m^ /Km2 road/yr 1,000-2,000 m^ /Km2 road/yr 100-500 m^ /Km2 road/yr Lehre (1982), Reid (1981). Table A.7 Table A.7 Table A.7 65 Table 3.2 are used to estimate the volume of sediment mobilized in the study watersheds by the less important transfer processes (i.e. s o i l creep, tree throw, surface erosion); the proportion delivered to the study creeks i s then estimated from the length of stream channel flanked b.y steep, s o i l -covered hillslopes and intersected by landslide scars and roads. 3.2 Field measurements The following f i e l d investigations were carried out during the summer low flow period from June through August 1982, and during May 1983. 3.2.1 Channel morphology The morphologic c h a r a c t e r i s t i c s of each "sediment wedge" were described l o n g i t u d i n a l l y and i n plan form by the following survey techniques. Longitudinal profile In order to determine the length of a "sediment wedge", the long profile began at a point downstream, and extended upstream to a point, where "normal" f l u v i a l processes and channel morphology appeared to prevail. The tapered upstream end of a "sediment wedge" was easily identified because the adjoining upper reaches usually were scoured to bedrock, while the downstream terminus was distinguished by the sudden change in channel width and gradient (see Figure 1.9). Due to i t s short length, Armentieres Creek was surveyed to sea l e v e l , while the long p r o f i l e s of the other three creeks were begun at an arbitrarily assigned datum. Standard surveying techniques were employed except for one minor modification: an automatic level and stadia rod were used to obtain height 66 information, but a hip chain (metered cotton thread), rather than stadia, was preferred for distance measurement. The rationale for this choice i s as follows. In order to characterize the long profile of a "sediment wedge", the desired line-of-sight l i e s mid-way between the stream banks rather than along the stream thalweg. Correct height and distance information i s ob-tained i f the l e v e l i s set up along t h i s l i n e , but i n small creeks with heavily vegetated banks and in-stream obstructions, i t i s often necessary to shoot diagonally across the channel. Although height information i s un-affected, the use of stadia hairs overestimates the true distance between two stadia rod positions. To alleviate this problem, the rodman carried a hip chain to measure the straight l i n e slope distance between successive positions, from which the true horizontal distance was determined; as a check, the upper and lower stadia hair readings were also booked. On av-erage, successive stadia rod positions were 30 metres apart. Plan form The principal morphological features of a "sediment wedge" were mapped in plan view by standard surveying techniques, using a plane table, telescopic alidade, stadia rod, metric tape, and hip chain. The telescopic alidade and stadia rod provide height and distance information with the same degree of precision as a theodolite and stadia rod, but a major advantage of a plane table i n small creeks such as these, i s that the rudimentary map can be supplemented rapidly by sketching and annotating additional features of interest while in the f i e l d . For this study, a section of each "sediment wedge", 500 metres or more i n length, was mapped at a scale of 1: 500, using bench marks, established during the long profile traverses, as height data. Observations made during e a r l i e r surveys enabled the chosen sections to include morphologic evidence of accelerated stream sedimentation, such as de-watered channels and discontinuous terraces. 3.2.2 Sediments Analysis of "sediment wedge" and stream bank deposits addresses the sources, modes of deposition, and residence times of material i n the f l u v i a l system. Surface sediment Surface sediments were sampled at both ends of a "wedge", and at two or three approximately equally spaced sites in between (see Table 3.3) in order to describe any downstream trends; channel thalweg material was not sampled. At each s i t e , a 30 metre survey tape was stretched out p a r a l l e l to the direction of flow, and the clast exactly beneath each half metre mark was picked up (after Wolman, 1954). The a, b and c axes of the 60 c l a s t s were then measured for those p a r t i c l e s with a b-axis greater than 8 mm. This i s the minimum size of material that can be removed practicably from the surface by hand (Wolman, 1954; Bray, 1972; Gomez, 1983). As such particles were encountered infrequently, the sample size always exceeded 50, which i s adequate for practical and s t a t i s t i c a l purposes (Bray, 1972). For very large, immovable stones, the a, b and c axes were estimated iii situ. The b-axis measurements were then arranged into half phi classes on the basis of frequency by number, which i s equivalent to frequency analysis by weight of bulk sieve samples (Kellerhals and Bray, 1971). In addition, at each s i t e a q u a l i t a t i v e assessment was made of the angularity, l i t h o l o g y and imbrication of large c l a s t s . As with the 68 TABLE 3.3: Surface, sub-surface and bank sample sites Watershed Distance upstream from start of long profile (metres) Sediment Sample Sites Surface Sub-Surface Bank Lagins Creek tributary 3 1,750 1,300 1,000 600 200 SI S2 S3 S4 S5 Mountain Creek 1,630 1,300 1,000 670 SI S2 S3 S4 SSI SS2 SS3 SS4 B2 B3 B4 Mosquito Creek tributary 1,990 1,580 1,220 940 720 540 SI S2 S3 S4 S5 SS2 SS3 SS4 SS5 SS6 B3 B5 Armentieres Creek 1,700 1,510 1,270 930 600 SI S2 S3 S4 S5 SSI SS3 SS4 SS5 Bl B4 B5 aRain prevented sampling of sub-surface sediments. 69 measurements, these assessments always were made by the writer, i n order to minimize operator variance. Sub-surface sediment Sub-surface sediment was sampled at four or f i v e s i t e s along each "wedge", coi n c i d i n g l a r g e l y with the surface transect s i t e s (see Table 3.3). At each s i t e , the surface grains were removed and a p i t (usually 0.5 metres wide at the top and 0.3 metres deep) was dug, although the precise l o c a t i o n was a r b i t r a r y . The t o t a l weight of material removed was between 80 and 100 Kg, which was the p r a c t i c a l l i m i t for the f i e l d sieving procedure. The amount required to represent accurately the weight d i s t r i b u t i o n of the sediment i s l a r g e l y a function of the heaviest c l a s t included, which at most s i t e s did not exceed 5% of the t o t a l sample weight. This material was then sieved into half phi classes between 64 mm (-6 phi) and 8 mm (-3 phi), and weighed i n the f i e l d using a procedure s i m i l a r to that described by Bray (1972). The c l a s t s which were held on the 64 mm sieve were i n d i v i d u a l l y measured and weighed, while the material f i n e r than 8 mm was weighed, quartered, and sealed i n a sample bag for subsequent laboratory a n a l y s i s . In general, the sample returned weighed 2 to 3 Kg, and losses during f i e l d s i e v i n g were les s than 1% of the i n i t i a l sample weight. In the l a b o r a t o r y , the m a t e r i a l was analyzed by standard techniques, which are described i n d e t a i l by G r i f f i t h s (1967) and Bray (1972). To simplify subsequent c a l c u l a t i o n s , i t was assumed that a l l the s o i l moisture was d i s t r i b u t e d among the sub-8 mm f r a c t i o n , because of the higher s p e c i f i c surface area and the small pore spaces i n aggregate c l u s t e r s of f i n e s ; Bray (1972) made the same assumption. Sedimentation event stratigraphy was examined i n the walls of 70 sub-surface sample pits and in exposed "sediment wedge" faces. A clean, vertical exposure of a "wedge" face was obtained by removing the s u r f i c i a l deposits with a spade, although further sloughing of non-cohesive material during excavation could not be prevented. Stream bank sediment Rapid surveys were conducted along the main stem of each creek, from their outlets up to their unlogged headwaters. Stream bank heights were measured at 50 metre intervals, and exposed sections were evaluated qualitatively in terms of the size (texture), sorting, compaction, and stratigraphy of the material, and the shape, lithology and imbrication of large c l a s t s ; from th i s information, the genetic o r i g i n and potential e r o d i b i l i t y were inferred. Additional observations were made of the position of organic layers (peat deposits) and organic horizons (presumed buried paleosols) relative to the contemporary s o i l , and the genetic origin of the intervening material. At two or three sub-surface sample sites, the adjacent stream bank material was also sampled (see Table 3.3), using the same f i e l d and laboratory procedure for bulk sieve analysis. However, two additional problems were encountered. F i r s t l y , stream banks may contain deposits of varying genetic origin, although this situation occurred infrequently; where fea s i b l e , these materials were sampled i n proportion to th e i r exposed surface area. Secondly, during pit excavation into near-vertical bank faces, much fine material f e l l to the base of the bank; consequently, a tarpaulin was placed beneath the pit which trapped most of the fines. 3.3 Air photo analyses The Federal, Provincial, and privately-flown photography used in this study provides a temporal coverage of 50 years, at scales ranging between 1: 10,000 and 1: 63,360 (see Table 3.4). 3.3.1 Hillslope erosion Photogrammetric measurements and mapping of mass movement events in the study watersheds were made with a Wild A6 stereo-plotter, at The University of British Columbia. Operating procedures for this machine are given in Rood (1983), and Moffitt and Mikhail (1980) provide a comprehensive discussion of basic photogrammetric theory and practice. It i s the objective of this part of the study to measure the volume of material mobilized by mass movement events on hillslopes and thereby quantify the volume entering the stream channel network. Three types of mass movement are distinguished: debris slides, debris flows and debris torrents, as defined in Section A.3. For purposes of analysis, each mass movement i s sub-divided into three zones: the in i t i a t i o n zone where most s o i l erosion occurs, the run-out (or depositional) zone where most debris i s deposited, and the intervening transport zone, through which upslope material passes and additional s o i l erosion may occur. These zones are d i f f e r e n t i a t e d by changes in slope, and evidence of erosional (e.g. bedrock exposure) and depositional (e.g. large organic debris) features. However, each zone i s not necessarily present in every mass movement; for example, the transport zone may be either absent or indistinguishable from the i n i t i a t i o n zone. The photogrammetric measurement and mapping of mass movements in the Mountain Creek, Mosquito Creek tributary and Armentieres Creek basins was made from the 1982, 1: 10,000 scale air photos; for the Lagins Creek tributary watershed, the 1976, 1: 15,840 scale air photos were ut i l i z e d . By 72 TABLE 3.4: Available air photography Photography Watersheds Covered Date Approx. Source Armentieres Mosquito Mountain Lagins Scale trib. t r i b . 1933 1: 31,680 Federal X X 1937-38 1: 31,680 Provincial X X 1954-55 1: 63,360 Federal X X X X 1964 1: 15,840 Provincial X X 1966 1: 15,840 Provincial X X 1974 1: 63,360 Provincial X X X X 1976 1: 15,840 Provincial X X X X 1979 1: 63,360 Federal X X X X 1982 1: 10,000 Private X X X 73 relocating them on the most recent set, events of different age were mapped at a single 1: 5,000 scale. The resolution of air photos determines the accuracy with which mass movements can be i d e n t i f i e d and measured. In p a r t i c u l a r , small landslides are more easily identified in clearcuts than in forested areas, and their frequency of occurrence and their dimensions are subject to proportionally greater measurement errors due to differences in operator performance. From these considerations, only those landslide scars with a surface area greater than 200 m2 were measured and mapped; the number of smaller slides was noted. The length and width of slide scars were measured from 1: 5,000 scale maps, with a graduated ruler, to iX).5 metres on the ground. A similar precision was obtained from the vertical scale on the Wild A6, but because the depth of most s l i d e s i s less than 1 metre (Smith et a l . , 1983), height information from the Wild A6 was used primarily for slope gradient c a l c -ulations. Consequently, the depth to bedrock i n the i n i t i a t i o n zone was taken as 50 cm, with a standard deviation of 6 cm (R.B. Smith, pers. comm., 1984), and checked from the v e r t i c a l scale, and the depth of s o i l eroded from the transport zone was assumed to be 25 cm (after Rood, 1984). The product of these dimensions gives an estimate of the volume of eroded sedi-ment, and t h i s was compared to the independently calculated volume of deposited sediment in order to corroborate these assumptions. Limited ground checking of mass movement scars substantiated the assumptions and results of air photo analyses. 3.3.2 Riparian erosion The erosion of stream banks in the study watersheds was measured d i r e c t l y from a i r photos that were uncorrected for t i p or t i l t displace-74 ments; the extra time involved in mapping with the Wild A6 was not justified because the absolute displacements are usually small, and the r e l a t i v e displacement of a stream channel i s much less than that of a hillslope. The procedure used in each watershed was as follows: i ) Along the main stem of each creek, the edges of the stream banks were traced from the e a r l i e s t available set of photos (1933 or 1937-38), the set flown immediately before or during logging (1964 or 1966), the set flown af t e r the completion of logging (1976), and the most recent set of photos (1982). i i ) As these tracings were at different scales, they were redrawn to a common scale (usually the largest) by means of a Saltzman projector. i i i ) The width of the stream channel was then measured perpendicular to the stream banks, at 50 metre i n t e r v a l s , with a graduated ruler to within ±0.8 metres on the ground. Where a stream chan-nel was divided about an island, the width of each branch was measured, and then summed, to produce a single value of channel width. The f i r s t cross-section was located at the same place in each set of a i r photos,and always downstream of the "sediment wedge"; channel widths were measured as far upstream as air photo resolution permitted (usually up to second order channels). iv) Temporal changes in channel width prior to logging (from 1930's and 1960's air photos), before and after logging (from 1960's and 1976 air photos), and since the completion of logging (from 1976 and 1982 a i r photos) enabled the area of r i p a r i a n land eroded 75 during each period to be calculated. Locating the edge of the stream banks is a major problem in areas with a dense forest canopy and/or i n the older, poorer resolution a i r photography. In such situations, the crowns of trees assumed to be on or near the bank edge were used to delineate the margins of the stream channel. Qualitative observations were made of changes in stream channel location, due to braiding or river diversion along a logging road, and these were ground checked during subsequent f i e l d investigations. 3.3.3 Road network The length and width of logging roads i n the study watersheds were measured d i r e c t l y from a i r photos, uncorrected for t i p or t i l t displacements, with a graduated ruler to within ±1.0 metre on the ground. The average road width was measured from a i r photos during or shortly after logging, when the road right-of-way was most visible, and the total road length was measured from air photos flown after logging, when the road density was greatest. The results are summarized in Table 2.2. The impact of bridge collapse, and river piracy by logging roads, on stream channel morphology was assessed q u a l i t a t i v e l y by a i r photo analyses and subsequent f i e l d investigations. 76 4.0 FLUVIAL "SEDIMENT WEDGES" IN THE QUEEN CHARLOTTE ISLANDS The sources and mobility of the "sediment wedges" are evaluated in terms of a sediment budget (Section 4.1), while the causes are addressed i n Section 4.2. 4.1 Sediment budget 4.1.1 Sediment sources In the following discussion, f i v e sediment sources, each rep-resenting a group of erosion processes, are considered: stream banks, rapid mass movements, slow mass movements, hillslope surfaces, and road surfaces. The justification of those processes examined directly (rapid mass movements (landslides) and stream bank erosion) and those assessed from representative process rates ( s o i l creep, tree throw, surface erosion), has been made in Section 3.1. The sources pertinent to t h i s study are those that deposit their sediment into, or near to, the drainage network that i s adjacent to, and upstream of, a "sediment wedge"; the location of the "sediment wedge" within each study watershed i s shown in Figures 4.1 to 4.4. The delivery of sediment from various sources, within each water-shed, i s summarized in Table 4.1, and contributions to the three "sediment wedges" i n Armentieres Creek are shown i n Table 4.2. A range of delivery volumes i s presented for each sediment source: landslide and stream bank ranges incorporate the error associated with measurement imprecision, while the other source estimates also reflect the range of representative process rates (indicated in Table 3.2). Furthermore, the total volume of sediment delivered to the contributing drainage network of each "sediment wedge" is presented as an extreme range (the sum of a l l positive and negative 77 FIGURE 4.1: L o c a t i o n o f t h e "sediment wedges" i n A r m e n t i e r e s Creek 78 FIGURE 4.3: Location of the "sediment wedge" in Mountain Creek TABLE 4.1: Volumes of sediment delivered from various sources 3 Armentieres Creek Mosquito Creek tr i b . Mountain Creek Lagins Creek t r i b . Sediment Source xlO nr input % 3 3 x l 0 J m input % 3 3 xlO nr input % 3 3 xlO nr input % STREAM BANKS 3.7 - 7.7 11-13 15.1-31.8 45-48 23.6-49.2 49-48 28.8-60.2 69 LANDSLIDES 22.4 - 35.8 71-61 7.9-12.6 24-19 7.0-11.2 15-11 3.5- 5.5 8-6 Slide scars 1.3 - 5.2 4-9 0.5- 1.9 1-3 0.9- 3.6 2-3 0.3- 1.1 1 Soil creep & tree throw 2.9 - 7.3 9-12 3.9- 9.7 12-15 12.0-30.0 25-29 4.4-11.0 10-13 Slope wash 1.2 - 2.1 4 1.7- 2.9 5-4 3.1- 6.3 6 1.5- 3.0 4 Road surfaces 0.2 - 0.5 1 3.1- 5.8 9 1.2- 2.6 3 3.3- 6.1 8-7 ROAD BRIDGE Not applicable 1.3- 1.6 4-2 Not applicable Not applicable Most extreme range 31.7-58.6 100 33.4-66.3 100 47.8-102.9 100 41.7-86.9 100 Most probable range 37.5-52.7 100 40.5-59.1 100 59.4-91.3 100 48.2-80.5 100 a A l l values rounded to the nearest 100 m Sources whose values are derived from measurements are in capital letters; Sources whose values are derived from representative rates are in small letters. TABLE 4.2: Sediment sources in Armentieres Creek watershed3 Sediment Source Upper "Wedge •a Q xlCr nr input II % Middle "Wedge xlO m input n % Lower "Wedge xl 0 J nr input II % Total x l 0 J nr input % STREAM BANKS 3.7- 7.7 16-18 n i l 0 n i l 0 3.7- 7.7 11-13 LANDSLIDES 16.6-26.6 71-62 4.0- 6.4 76-66 1.8- 2.8 58-50 22.4-35.8 71-61 Slide scars 1.0- 3.9 4-9 0.2- 0.9 4-9 0.1- 0.4 3-7 1.3- 5.2 4-9 Soil creep & tree throw 1.6- 4.1 7-9 0.8- 1.9 15-20 0.5- 1.3 17-22 2.9- 7.3 9-12 Slope wash 0.4- 0.7 2 0.2- 0.4 4 0.6- 1.0 19-17 1.2- 2.1 4 Road surfaces 0.1- 0.1 0.3 N i l - 0.1 1 0.1- 0.2 3-4 0.2- 0.5 1 Most extreme range 23.4-43.1 100 5.2- 9.7 100 3.1- 5.7 100 31.7-58.6 100 Most probable range 27.6-38.9 100 6.1- 8.8 100 3.7- 5.1 100 37.5-52.7 100 a A l l values rounded to the nearest 100 m Sources whose values are derived from measurements are in capital letters; Sources whose values are derived from representative rates are in small letters extremes), and as the most probable range (as a l l extremes are unlikely to arise together); the latter i s calculated by taking the square root of the sum of squared errors, in which a non-statistical error (quoted as an out-side lim i t , such as a measurement uncertainty) i s assumed to be equivalent to a two standard deviation s t a t i s t i c a l error. Stream banks Stream banks are the major sediment source in Mosquito and Lagins Creek t r i b u t a r i e s , and Mountain Creek (see Table A.l). The rate of stream bank erosion was measured from sequential air photography (see sub-section 3.3.2), and the results are summarized in Table A.3. The analysis of each creek on a reach-by-reach basis enables small, but significant, changes in plan form to be identified. The length of a reach was determined by distinctive changes in channel morphology (e.g. at the ends of a "sediment wedge"), s u r f i c i a l geology and by the location of roads and clearcut boundaries. Armentieres Creek was sub-divided into three reaches, each containing a "sediment wedge". Mosquito Creek tributary had the "wedge" in i t s most downstream reach, a haul road constructed along the edge of a stream bank in reach A, and a change of s u r f i c i a l materials (from f l u v i a l to f l u v i a l / c o l l u v i a l ) i n the most upstream reach (B). Mountain Creek i s broken into four reaches: the reach that contains the "sediment wedge" separates the most downstream reach (A) from reach B; i n the most upstream reach (C), the channel pattern i s controlled by bedrock. Similar-ly , i n Lagins Creek tributary the "sediment wedge" defines the upstream l i m i t of reach A and the downstream l i m i t of reach B; the most upstream reach (C) begins at the clearcut boundary, and indicates the rate of stream bank erosion in the forested part of the watershed. 82 TABLE 4.3: Measurements of changes in stream channel widths Watershed Date(s) Mean Channel Width (metres) and % Change Armentieres Cr. Lower Middle Upper "Wedge" "Wedge" "Wedge" mean channel width (m) 1933 32.0 23.2 34.1 1964 39.7 33.4 30.0 1976 26.4 29.0 44.3 1982 30.7 24.0 34.8 % change 1933-64 +24.1 +44.0 -12.0 1964-76 -33.5 -13.2 +47.7a 1976-82 +16.3 -17.2 -21.4 Length of Reach (Km) 0.50 0.70 0.35 Mosquito Cr. trib. "Wedge" Reach A Reach B mean channel width (m) 1964 20.3 25.2 23.6 1976 33.2 24.0 21.0 1982 27.3 25.7 20.8 % change 1964-76 +63.5a _ 4. 8 _ 1 1 # 0 1976-82 -17.8 +7.1 -1.0 Length of reach (Km) 2.05 0.80 0.55 83 TABLE 4.3 (contd.) Watershed Date(s) Mean Channel Width (metres) and % Change Mountain Cr. Reach A "Wedge" Reach B Reach C mean channel width (m) 1933 22.5 20.3 24.7 24.3 1976 48.2 40.9 25.0 14.7 1982 33.2 32.9 17.2 13.1 % change 1933-76 +114.2 +101.53 +1.2 -39.5 1976-82 -31.1 19.6 -31.2 -10.9 Length of reach (Km) 0.90 1.35 1.90 1.10 Lagins Cr. trib. Reach A "Wedge" Reach B Reach C mean channel width (m) 1966 34.0 57.4 56.3 30.4 1976 27.4 58.4 49.7 27.6 % change 1966-76 -19.4 +1.7 -11.7 -9.2 Length of reach (Km) 0.30 1.80 0.35 2.20 aValue used in sediment input calculations 84 It i s apparent from Table 4.3 that for the study creeks i n general, most of the channel width increase (i.e. stream bank recession) occurred in the "sediment wedge" reach, sometime before 1976, and that the channels narrowed between 1976 and 1982, due to revegetation of the "wedge" surface. Other large percentage changes i n channel width usually are associated with the inaccurate delineation of forested stream banks, which are p a r t i c u l a r l y d i f f i c u l t to i d e n t i f y on the low resolution, 1933 a i r photos. In Armentieres Creek, channel width decreased 33% along the lower "wedge" reach between 1964 and 1976, and increased 44% along the middle "wedge" reach between 1933 and 1964 (see Table 4.3). The former i s the r e s u l t of rapid recolonization by streamside vegetation (e.g. red alder), while the l a t t e r i s due c h i e f l y to the inaccurate delineation of forested stream banks on the 1933 a i r photos. Thus, only the stream bank material eroded from the upper "wedge" reach, between 1964 and 1976, was used i n sediment source computations for Tables 4.1 and 4.2. In Mosquito Creek tributary, bank retreat along the "sediment wedge" reach, between 1964 and 1976, was the only source of stream bank material used i n sediment input calculations (see Table 4.1). The 18% decrease in "sediment wedge" width, between 1976 and 1982 i s due largely to the recolonization of red alder along the lowermost 270 metres. In Mountain Creek, the only riparian sediments which were con-sidered to contribute to the "wedge" were those derived from the stream banks adjacent to the "wedge", between 1933 and 1976; the largest increase in width (114%) occurred, during the same period, along the reach downstream of the "wedge", where in-stream skidding was practiced. The decreased widths of a l l reaches, between 1976 and 1982, i s the r e s u l t of rapid re-colonization by streamside vegetation; i n contrast, the 40% decrease i n 85 width, between 1933 and 1976, along reach C reflects the d i f f i c u l t y of using tree crowns to represent the stream bank margins of narrottf channels. In Table 4.3, the "wedge" reach i n Lagins Creek tributary i s shown as having widened only 2% between 1966 and 1976. This value was obtained from the reach of channel sandwiched between reaches A and B (identified on both sets of air photos). However, the reach containing the "wedge" i n 1976 i s a section of a logging road, along which the creek was diverted, and therefore, i s not comparable to the undisturbed, 1966 stream channel. Therefore, a better indication of the amount of "wedge" sediment contributed by "stream bank" erosion i s the difference betv^een the i n i t i a l road width (11.2 metres), and i t s width following r i v e r avulsion (32.6 metres); the latter was obtained from the "wedge" reach, and the former from roads unaffected by flow diversion, on the 1976 a i r photos. The rate of bank retreat adjacent to the road was thus calculated, and i s shown in Table 4.1 as the sediment input from stream banks. In order to calculate the volume of sediment introduced through stream bank erosion, the areal rates of bank retreat (computed from a i r photo analyses) were mult i p l i e d by the mean height of the stream banks (measured in the f i e l d ; see Table 4.4). TABLE 4.4: Mean stream bank heights Watershed Mean Height (m) Armentieres Cr. 1.0 Mosquito Cr. t r i b . 0.9 Mountain Cr. 1.4 Lagins Cr. t r i b . 1.3 86 Rapid mass movements I n i t i a l h i l l s l o p e f a i l u r e s and the subsequent erosion of t h e i r scar surfaces, introduces large volumes of sediment to Armentieres Creek, Mosquito Creek tributary, and Mountain Creek (see Table 4.1); in Armentieres Creek, rapid mass movements account for 60 to 70% of the sediment delivered to the channel (see Table 4.2). The locale of deposition and the timescale employed require c l a r i -fication. For Armentieres Creek and Mountain Creek, only those slides that entered the creek d i r e c t l y were included as sediment sources, while for Mosquito and Lagins Creek tributaries, material deposited originally on the footslopes was also included. The rationale for this inconsistent method-ology i s as follows: i n the headwaters of the two l a t t e r watersheds, several landslides run-out on the footslopes immediately adjacent to the stream channel. As these footslopes are short and r e l a t i v e l y steep, the slide deposits are remobilized by surface erosion processes, and subsequent-ly reach the stream channel. In contrast, the headwater channels of Mountain and Armentieres Creeks are juxtaposed with steep slopes, so that mobilized material invariably reaches the stream channel directly; footslope deposits are present further downstream in these two watersheds, where the valley f l o o r i s wider, but the lack of g u l l i e s and small fans immediately downslope of the deposits suggests that t h e i r subsequent transfer to the stream channel i s prohibited by the low slope gradient. Secondly, landslides that occurred decades before the development of the "sediment wedge" are considered as sediment sources for the following reasons. The i d e n t i f i c a t i o n of older landslides from a i r photos i s re-s t r i c t e d usually to the largest events, where a large area can be distinguished from the undisturbed surrounds by i t s vegetation succession, 87 and headwater events, where recurrent failure or snow avalanching prevents vegetation recolonization. Although most of the fine sediment from such events i s transported rapidly a considerable distance downstream, the coarse fraction i s moved at a much slower rate. Hence, the morphologic impact of large and/or headwater landslides may persist for several years after the actual event, as the supply of coarse material i s depleted slowly. The oldest landslide thought to contribute material to a "sediment wedge" differs between watersheds (see Table 4.5). Mountain Creek and Lagins Creek tributary watersheds have slide scars pre-dating the 1933 a i r photography that probably are more than 100 years old and therefore, are not included as sources of sediment. Further-more, these old scars are presently snow avalanche tracks, and may have originated as such, in which case their volumetric inputs may be much less than indicated i n Table 4.5; Gardner (1983) noted that the erosive e f f e c t s of wet snow avalanches were highly s i t e - s p e c i f i c , but that even under favourable conditions, erosion may be limited. In order to maximize the timespan for landslide analysis in the Armentieres Creek basin, a l l s l i d e s , regardless of age, were included; however, only 1500 m3 of sediment reached the creek from pre-1963 land-s l i d e s . The fact that no landslides prior to 1955 were found i n the Mosquito Creek tributary basin is due, perhaps, to the very poor quality of the 1937-38 air photos. The rates of landslide erosion in the forested and logged parts of the study watersheds are shown in detail in Appendix B, and are summar-ized in Table 4.6; for the Armentieres Creek basin, annual rates of erosion were calculated for the 1964-1982 period. The frequency of occurrence of the d i f f e r e n t f a i l u r e types, under various watershed conditions, i s 88 TABLE 4.5: Time periods used to assess landslide inputs Watershed Time Period of Landslide Contribution Antecedent Landslide Events Total Number Volume (m ) Number Reaching Creek Reaching Creek Armentieres Creek Pre-1963-1982 10 3 1,500 Mosquito Creek tributary 1955-1982 0 0 0 Mountain Creek 1934-1982 8 1 2,300 Lagins Creek tributary 1934-1982 4 3 8,300 Probably more than 100 years old. Although no photos were flown over Lagins Creek tributary in 1982, i t s headwaters appeared in the last two frames of the 1982 Mountain Creek sequence, from which i t was observed that no new landslides had occurred since 1976; as previous landslides were concentrated in the headwaters, and no new slides were found elsewhere in the basin during f i e l d investigations, the timescale of landslide analysis was deemed to extend up to 1982. 89 TABLE 4.6: Rates of landslide erosion and sediment transfer in logged watersheds Period Total Sedj Lment Production Sediment I{ iput to Creek a Sediment D$ ^livery Ratio*3 Watershed of (nr VKm2/yr) (nr 3/Km2/yr) (nr VKm2/yr) Record (Years) Unlogged Logged Ratio Unlogged Logged Ratio TSP SIC Ratio Armentieres Creek 19 607 393 0.6 452 21 0.1 566 370 0.65 Mosquito Cr. t r i b . 28 95 112 1.2 71 63 0.9 98 69 0.70 Mountain Creek 49 78 326 4.2 24 3 0.1 102 22 0.22 Lagins Cr. tr i b . 49 21 Nil Nil 16 Nil N i l 19 14 0.74 Armentieres Creek and Mountain Creek include material that reached the creek directly; Mosquito Creek tributary and Lagins Creek tributary also include material deposited on the footslopes. TSP: Total sediment production from entire (logged and unlogged) watershed; SIC: Sediment input to creek from entire (logged and unlogged) watershed. indicated i n Table 4.7; nearly a l l the f a i l u r e s are c l a s s i f i e d as debris slides, of which more than one-third are gully-related. There are no road-related failures in the study watersheds. In the Mountain Creek and Mosquito Creek tributary basins, clear-cuts produced more sediment than forested areas; this concurs with results reported for the P a c i f i c Northwest i n general (see Table A.6). In Armentieres Creek, the forested area produced more sediment than the clear-cut area, because the average s l i d e volume i s four times larger i n the former, while Lagins Creek tributary experienced no f a i l u r e s i n the low-gradient clearcuts. When the entire basin area i s considered, 65 to 75% of the sediment mobilized by landsliding reached the drainage networks of three study creeks (see Table 4.6). Following the i n i t i a l f a i l u r e , landslide scars are modified by surface erosion processes (rainsplash, sheetwash, dry ravel, r i l l i n g ) . Values shown i n Table 3.2 are applied, in t h i s study, to the scars of landslides that reached the stream channels of Armentieres and Mountain Creeks, and also the footslopes of Lagins and Mosquito Creek t r i b u t a r i e s . It i s assumed that these secondary erosion processes are active for 20 years, following Smith et a l . (1983), and hence the volume of sediment eroded from each scar i s 20,000 to 80,000 m3/Km2. In general, slide scars produce an order-of-magnitude less sedi-ment than the i n i t i a l f a i l u r e s , and less than 10% of the t o t a l volume of sediment delivered to the study creeks (see Table 4.1). As scars produce sediment f i n e r than 2 mm (Reid et al . , 1981), which can be transported rapidly downstream, their impact on channel morphology should be ephemeral and insignificant, compared to that of the bed load calibre material intro-duced by the i n i t i a l failures. 91 TABLE 4.7: Relation between slope failure types and watershed conditions Number of failures under various watershed conditions Watershed Type of Slope Total Failure Number Gully- Gully- Open- Open-Logged Unlogged Logged Unlogged Armentieres Creek Debris slide 49 Debris flow 1 Debris torrent 1 17 18 1 Mosquito Cr. t r i b . Debris slide 10 Debris flow 1 2 1 Mountain Creek Debris slide 57 Debris torrent 1 18 1 16 19 Lagins Cr. tr i b . Debris slide 14 13 92 Slow mass movements The slow processes of s o i l creep and tree throw appear to account for 9-15% of the total sediment delivery to Armentieres Creek, and Mosquito and Lagins Creek tributaries (see Table A.l); in Mountain Creek they deliver almost 30%, as a result of the extensive length of contributing stream bank associated with the atypically high drainage density of low-order channels. For purposes of estimating the input by s o i l creep and tree throw, i t i s assumed that the combined processes of s o i l creep and tree throw result in the movement of the s o i l column by 2 to 5 mm/yr (see Table 3.2). S o i l creep was assumed to extend down to bedrock or a strongly indurated layer, which occurs at a mean depth of approximately one metre (Smith et a l . , 1983). As t h i s depth of movement i s quite shallow, and the slopes are steep, wet, and unstable, the use of a surface rate of s o i l creep rather than a depth-integrated rate, should not strongly bias the subsequent calculations. At least, this w i l l provide a maximum estimate of sediment input by these processes, and thus emphasize their relative insignificance as sediment sources. The combined rate was then applied to the length of stream channel juxtaposed by steep, soil-covered slopes (see Table A.8). In practice, this excluded only those streams that drain from bedrock or snow-covered surfaces above the t r e e l i n e , and higher-order channels that are bordered by low-gradient terrain. From the annual rates of sediment delivery, shown in Table A.8, the total volume of sediment delivered between 193A and 1982 was estimated (see Table A.l). Sediment entered upstream of a "wedge", and prior to "wedge" inception, i s included because stream banks deliver predominantly 93 TABLE 4.8: Annual rates of sediment delivery by s o i l creep and tree throw Length of Watershed Contributing Creep & Throw Stream Bank Input (m-Vyr) (Km)a Armentieres Cr. Upper "Wedge" 16.7 33- 83 Middle "Wedge" 7.8 16- 39 Lower "Wedge" 5.2 10- 26 Total 29.7 60-149 Mosquito Cr. t r i b . 39.7 79-198 Mountain Cr. 122.4 245-612 Lagins Cr. t r i b . 44.8 90-224 Twice the length of contributing stream channel. 94 coarse material (see sub-section 4.1.3), which i s transported slowly. Hillslope surfaces The surface erosion of forested and logged h i l l s l o p e s d e l i v e r s approximately 4 to 6% of the t o t a l volume of sediment reaching the study creeks (see Table 4.1). Annual delivery rates were generated by applying the ranges of representative process rates (see Table 3.2) to the forested and logged basin areas that are adjacent to, and upstream of, the "sediment wedges" (see Table 4.9). The t o t a l volume of sediment delivered between 1934 and 1982 was estimated subsequently (see Table 4.1). As slope wash mobilizes only fine sediment, the impacts on channel morphology should be insignificant compared to those accompanying coarse sediment inputs. Road surfaces To calculate road surface erosion, the road area was measured from a i r photos (the method i s described i n sub-section 3.3.3, and the results are given in Table 2.2) and appropriate rates of erosion were chosen from the literature (see Table 3.2). Only those sections of road that inter-sected the creek upstream of, or along, a "sediment wedge" were included in the analysis. The time periods over which the d i f f e r e n t rates of erosion were employed are given in Table 4.10. Overall, road surface erosion accounts for less than 10% of the t o t a l volume of sediment delivered to the drainage networks i n the study watersheds (see Table 4.1). Road surface erosion may have l i t t l e effect on the morphology of the study creeks, because the fine sediment produced probably i s transported rapidly through the f l u v i a l system. The morphologic TABLE 4.9: Annual rates of sediment delivery by slope wash Contributing Area (Km^) Slope Wash Input (m-Vyr) Watershed Forested Logged Total Forested Logged Total Armentieres Cr. Upper "Wedge" 1.1 0.3 1.4 4- 11 4 8- 15 Middle "Wedge" 0.6 0.1 0.7 2- 6 2 4- 8 Lower "Wedge" 1.4 0.3 1.7 6- 14 6 12- 20 Total 3.1 0.7 3.8 12- 31 12 24- 43 Mosquito Cr. t r i b . 4.3 1.1 5.4 17- 43 17 34- 60 Mountain Cr. 10.9 1.2 12.1 44-109 19 63-128 Lagins Cr. t r i b . 5.2 0.6 5.8 21- 52 10 31- 62 96 TABLE 4.10: Duration of different rates of road surface erosion Period of Time at Erosion Rate (m3/Km2 road/yr) a Watershed 10,000 - 15,000 1,000 - 2,000 100 - 500 Armentieres Cr. Upper "Wedge" 1967-1969 1970 1971-1982 Middle "Wedge" 1967-1969 1970 1971-1982 Lower "Wedge" 1962-1965 1966 1967-1969 1970 1971-1982 Mosquito Cr. trib. 1963-1969 1970 1971-1982 Mountain Cr. 1965-1967 1968 1969-1982 Lagins Cr. trib. 1970-1977 1978 1979-1982 Erosion rates given in Table 3.2: 10,000-15,000 m3/Km2 road/yr: during use 1,000-2,000 m3/Km2 road/yr: f i r s t year of dis-use 100-500 m^ /Km2 road/yr: abandoned 97 impacts also may be less in the study creeks than those experienced else-where in the Pacific Northwest, as road densities in other areas often are more than twice those in the study watersheds. Finally, the collapse of a bridge crossing in Mosquito Creek tributary, delivered approximately 1,450 (-i-150) m3 of predominantly coarse material directly into the stream channel (see Table 4.1), and additional fine sediment probably was contributed by the surface erosion of the haul roads (see sub-section 2.4.2). 4.1.2 Sediment storage The volume of material stored in the "sediment wedges" forms the second component of the sediment budget. The surface of a "sediment wedge" i s described by the longi-tudinal profile. Between bedrock control points, i t s base (representing the idealized pre-"wedge" profile) i s assumed to be a smooth, concave curve, based on the widespread observation that stream profiles usually are concave over considerable distances (Mackin, 1948; Hack, 1957; Miller, 1958; Brush, 1961). The mathematical function that best f i t s the surveyed points up-and down-stream of the depositional zone, and the bedrock outcrops along the "sediment wedge", was found by attempting logarithmic and polynomial curve f i t s . The long profiles, and best-fit curves, of the "sediment wedges" are shown in Figures 4.5 to 4.8. The pre-"wedge" profile in Armentieres Creek has been delineated by two logarithmic curves, which intersect at a plutonic bedrock outcrop at 1.1 Km i n Figure 4.5, and the base of the "sediment wedge" in Mountain Creek, Mosquito Creek tributary and Lagins Creek tributary has been represented by a second order polynomial. 98 FIGURE 4.5: Longitudinal profile of the "sediment wedges" in Armentieres Creek. FIGURE 4 . 6 : Longitudinal profile of the "sediment wedge" in Mosquito Creek tributary. FIGURE 4 . 8 : Longitudinal profile of the "sediment wedge" in Lagins Creek tributary. The volume of a "sediment wedge" was estimated by measuring i t s length and depth from the long profile plots (Figures 4.5 to 4.8), and obtaining widths from air photos (see sub-section 3.3.2). Rather than using gross mean values of width and depth, each "wedge" was divided into sub-units with a 50 metre base length, and the mean values of width and depth for each sub-unit were used to compute their individual volumes; the sum of these increments equals the total volume of the "sediment wedge": u = n S t = Z W u D u L u u = 1 where S t = total volume of the "sediment wedge" (m ) u = number of sub-unit n_ = maximum number of sub-units = mean width of sub-unit (m) Du = mean depth of sub-unit (m) L u = length of sub-unit (50 metres) This procedure calculates the volume of a rectangle, although most stream channels are more semi-circular in cross-section; thus, rect-angular "sediment wedge" volumes (see Table 4.11, column A) are maximum estimates, to which channel shape adjustments can be made (see Figure 4.9). Two shape adjustments are given in Table 4.11: column B indicates the semi-circular "sediment wedge" volumes, and column C gives the appro-priate volumetric reduction for stream bank materials at the angle of repose (35°), as unconsolidated sands and gravels can not maintain a vertical face. Moreover, i f the creek i s incised into the "sediment wedge", the volume occupied by the stream channel should be deducted. The approximate length and mean width of a creek can be found from the long profile and plane table surveys, respectively, but as the depth of incision was not measured i n the f i e l d , the mean depths of each "sediment wedge" were 103 TABLE 4.11: The volume of sediment stored in the "sediment wedges"3 A B C D E F G "Sediment Wedj >e" Semi-Rectangular Circular Triangular Stream L.O.D. "Sediment Wedge" Watershed Mean Mean "Wedge" "Wedge" Subtraction Channel Storage Volume Length Width Depth Volume Volume Volume Volume Volume B-D-E A -C-D-E (m) (m) (m) (xl0 3m 3) (xl0 3m 3) (xl0 3m 3) (xl0 3m 3) (xl0 3m 3) (xl0 3m 3) (xl0 3m 3 Armentieres Cr. Upper "Wedge" 160 35 1.6 9.0 6.2 0.6 0.6 Nil 5.6 7.8 Middle "Wedge" 300 27 1.2 9.4 6.9 0.6 0.5 Nil 6.4 8.3 Lower "Wedge" 770 22 1.5 25.3 16.8 2.5 1.4 0.1 15.3 21.3 Total 43.7 29.8 3.7 2.5 0.1 27.2 37.4 Mosquito Cr. tr i b . 2050 28 1.7 99.9 57.0 8.5 8.6 0.2 48.2 82.6 Mountain Cr. 1300 36 0.8 39.5 29.0 1.2 3.9 0.1 25.0 34.3 Lagins Cr. tr i b . 1650 16 1.5 40.0 28.1 5.3 7.5 0.2 20.4 27.1 3A11 values rounded to the nearest 100 m' GURE 4.9: Types of sediment storage adjustment. c ~f~ / Semi-circular "wedge" volume Triangular subtraction volume Stream channel volume Large Organic Debris (L.O.D.) storage volume 105 obtained from the long profile plots (see Table 4.11). From f i e l d observ-ations, i t i s known that the stream channels were incised infrequently to these depths, and thus, a stream channel depth half that of the "sediment wedge" was used for a f i r s t approximation. The volumes calculated for stream channels with a rectangular cross-section are presented in Table 4.11 (column D), but i n r e a l i t y they w i l l be somewhat less, depending upon the actual cross-sectional shape. Further reduction of "sediment wedge" volumes i s possible i f bedrock intrudes into the smooth, concave pre-"wedge" p r o f i l e s . Un-fortunately, apart from the adjustments made already for bedrock outcrops; there i s no basis for an objective evaluation of bedrock micro-topography. Finally, the volume of material stored behind log jams, or other in-stream obstructions, prior to the development of the "sediment wedges", should be subtracted from the contemporary "wedge" volumes. Other studies have found two orders-of-magnitude v a r i a t i o n i n the volume of sediment trapped behind large organic debris (L.O.D.), but with the exclusion of P i t l i c k (1982), the mean (and modal) value i s 3 m3/30 m channel (see Table A.8). The volumes of antecedent storage were estimated by applying t h i s value to the length of each "sediment wedge" (see Table 4.11, column E). Therefore, a range of possible "sediment wedge" volumes exists, depending on channel shape, stream incision, and antecedent storage factors. In Table 4.11, columns F and G indicate two possible volumetric estimates, and the actual "wedge" volumes probably l i e within this range. 4.1.3 Sediment mobility This sub-section discusses the mobility of the material i n the "sediment wedges", and represents the output term i n the sediment budget 106 equation. As the rate of sediment discharge was not measured d i r e c t l y i n the f i e l d , the output i s inferred from the estimated input and storage volumes, and measured sedimentologic and morphologic parameters. Sedimentology The plots of "wedge" surface, "wedge" sub-suirface, and stream bank sediment samples are shown in Figures 4.10 to 4.13, and t h e i r associated grain size s t a t i s t i c s are summarized in Table 4.12. In general, the mean size of the surface material i s greater than that of the underlying sediments; the l a t t e r are always c l a s s i f i e d as gravels, while the former are gravels and/or cobbles (e.g. Mountain Creek). Bank sediments are of gravel c a l i b r e , but the size range i s greater than that present in the sub-surface sediments. Furthermore, surface material in Lagins Creek tributary has a downstream fining trend, which i s also shown, to a lesser extent, in the other creeks; size trends are not apparent in the sub-surface or stream bank sediments. The standard deviations indicate that sub-surface sediments commonly are sorted as well as, or better than, the bank material, but that both are less well sorted than the surface sediments. A^similar pattern i s present in the skewness values, in that sub-surface sediments are positively skewed (indicating a t a i l of fines) usually to the same extent as, or lesser than, the bank material, while the surface sediments are only slightly skewed. Using these s t a t i s t i c a l parameters, surface sediments can be discriminated from the underlying material, but the latter can not be dis-tinguished readily from stream bank sediments. As the surface sediments represent a layer only one grain in thickness, almost the entire volume of a 107 FIGURE 4.10: Particle size distribution of "sediment wedge" and stream bank material in Armentieres Creek. Particle diameter (phi) FIGURE 4.11: Particle size distribution of "sediment wedge" and stream bank material in Mosquito Creek tributary. Particle diameter (phi) FIGURE 4.12: Particle size distribution of "sediment wedge" and stream bank material in Mountain Creek. Particle diameter (phi) FIGURE 4.13: Particle size distribution of "sediment wedge" surface material in Lagins Creek tributary. TABLE 4.12: Grain size s t a t i s t i c s for "wedge" surface, "wedge" sub-surface and stream bank sediments 3 Watershed S i t e Distance Geometric Mean (mm) Mean (phi) Standard Deviation (phi) Skewness (phi) Upstream (m) Surf Sub-Surf Bank Surf Sub-Surf Bank Surf Sub-Surf Bank Surf Sub-Surf Bank 1 1700 66.3 25.8 43.4 -6.05 -4.69 -5.44 1.50 2.10 2.76 0.04 0.27 0.19 2 1510 73.5 -6.20 1.14 0.10 3 1270 60.5 31.3 -5.92 -4.97 1.02 2.42 0.30 0.35 4 930 45.9 20.7 29.2 -5.52 -4.37 -4.87 1.26 2.53 2.54 -0.05 0.31 0.45 5 600 37.3 23.3 6.8 -5.22 -4.54 -2.76 1.01 1.99 2.19 0.09 0.26 0.20 1 1990 113.3 -6.83 1.24 0.41 2 1580 56.5 20.7 -5.82 -4.37 1.08 1.95 0.05 0.32 3 1220 46.5 27.7 19.3 -5.54 -4.79 -4.27 1.14 2.40 2.76 0.04 0.41 0.33 4 940 35.3 17.1 -5.14 -4.10 1.13 2.16 -0.06 0.34 5 720 78.2 25.3 43.1 -6.29 -4.66 -5.43 1.17 2.03 2.30 0.11 0.33 0.40 6 540 32.2 -5.01 .2.34 0.49 1 1630 112.2 19.8 -6.81 -4.31 1.25 2.37 0.30 0.27 2 1300 69.6 27.3 32.4 -6.12 -4.77 -5.02 1.27 2.61 2.66 0.04 0.46 0.68 3 1000 76.6 20.3 29.0 -6.26 -4.34 -4.86 1.72 2.47 2.41 0.15 0.20 0.44 4 670 86.8 22.9 45.6 -6.44 -4.52 -5.51 . 1.41 2.40 2.26 0.21 0.3 0.60 1 1750 76.6 -6.26 1.09 0.11 2 1300 60.5 -5.92 1.29 0.23 3 1000 50.2 -5.65 1.01 ' 0.07 4 600 39.1 -5.29 1.26 0.07 5 200 44.9 -5.49 0.90 0.15 Armentieres Cr. Mosquito Cr. t r i b . Mountain Cr. Lagins Cr. t r i b . a F o l k and Ward (1957) graphic measures were used: Mean size: (^16 + ^50 + ^ 8 4 ^ 3 Standard deviation: ( $ g 4 - ithg)/^ + ($95 - 84 - 2 * 5 Q j/2 ( 0 8 4 - A , . ) ] [«|>5 + (|)95 - 2 /2 0g5 _ ty5fi where ( b n = the grain size ( i n phi units) at which n per cent of the d i s t r i b u t i o n i s coarser. ( J ) = - logomm "sediment wedge" i s composed of sub-surface material. Consequently, the underlying material i s a better indicator of the source, and the extent of reworking, of "wedge" deposits. The s i m i l a r grain size c h a r a c t e r i s t i c s of sub-surface and bank sediments suggests that the former may be derived largely from the latter, and with l i t t l e subsequent f l u v i a l reworking, but due to the limited number of sediment samples analyzed, this association i s tenuous. On theoretical grounds, the patterns of sorting and skewness are in general accordance with those expected to r e s u l t from the erosion, transport, and deposition of material derived from a single sediment source (see McLaren, 1981). Thus, the sedimentary evidence i s consistent with the conclusions drawn in sub-section 4.1.1, that stream banks are the single most important sediment source in Mountain Creek, Mosquito Creek tributary, and Lagins Creek t r i b -utary. The mobility of "wedge" deposits i s determined largely by the stability of the surface sediments, because only after their mobilization can the underlying material be f l u v i a l l y resorted. Commonly, these are weakly imbricated, resulting in an "armoured" surface layer, which i s chara-cterized by larger material, and a lower fines content, than found i n the sub-surface deposits. These differences are due to f l u v i a l sorting pro-cesses and/or sampling limitations, in that clasts as large as those on the surface may also be encountered beneath i f larger volumes of material were sampled. The absence of fine sediment in the "armour" layer i s the result of t h e i r hydraulic winnowing from around stable surface c l a s t s and the intrusion of fines further into the streambed (Einstein, 1968; Beschta and Jackson, 1979; Bray and Church, 1980; Carling and Reader, 1982; Milhous, 1982; Parker and Klingeman, 1982; Gomez, 1984). Surface sediments exhibit a weak downstream-fining trend, 113 implying sorting by f l u v i a l processes, but due to the l i m i t e d sampling program, this relationship i s tenuous; the sorting trend remains weak even when a l l 19 surface samples are pooled in a single data set. Surface clasts were predominantly sub-angular in shape, varying from angular to sub-rounded, and did not show any systematic downstream-rounding trends. These observations suggest that the clasts have not moved far from their source and/or that they have a short residence time in the f l u v i a l system and/or that the reaches are too short to accomplish sub-stantial particle abrasion during transport. Apart from the imbrication and coarsening of surface c l a s t s , there was no c l e a r l y expressed stratigraphy i n the "sediment wedge" material; this may be a consequence of i t s rapid deposition and/or coarse texture. The mode of sediment transport i s dependent largely on particle size, with material finer than 1 to 2 mm representing suspended load, and the coarser f r a c t i o n becoming bed load. In terms of the "wedge" sub-surface sediments, approximately 90% i s coarser than 2 mm, and 94% i s coarser than 1 mm (see Table 4.13). As more than 90% of the "wedge" material i s of bed load c a l i b r e , the mobility of a "sediment wedge" i s determined primarily by bed load transport processes. Inputs/Storages Sediment output can be quantified indirectly as the difference between the estimated input and storage volumes. Hence, Table 4.14 gives the volumetric differences between the ranges of sediment inputs (from Table 4.1) and storages (from Table 4.11), and the subsequently inferred mean annual rate of sediment transport between 1934 and 1982. 114 TABLE 4.13; Proportion of sub-surface and stream bank material coarser than 1 mm and 2 mm Mean % Coarser than 1 mm Mean % Coarser than 2 mm Watershed Sub-Surface Banks Mean Sub-Surface Banks Mean Armentieres Cr. 94.5 90.7 92.9 90.9 85.6 88.6 Mosquito Cr. t r i b . 94.8 92.5 94.1 90.6 89.1 90.2 Mountain Cr. 93.0 93.7 93.3 88.2 89.9 88.9 Mean 94.1 92.3 93.4 89.9 88.1 89.2 115 TABLE 4.14: Volumes and rates of sediment output from the "sediment wedges"a Sediment Input "Sediment Wedge" Sediment Output (1934-1982) Watershed Volume Volume (xlO 3 m3) (xlO 3 m3) (xlO 3 m3) (m3/yr) Armentieres Cr. Upper Wedge" Middle "Wedge" Lower "Wedge" Total 27.6 - 38.9 6.1 - 8.8 3.7 - 5.1 37.5 - 52.7 5.6 - 7.J 15.3 - 21.3 27.2 - 37.4 19.8 - 33.3 400 - 700 6.4 - 8.3 19.8 - 35.7b 400 - 700b 10.2 - 17.6C 600 - l,000c 0.1 - 25.5 Nil - 500 Mosquito Cr. t r i b . Mountain Cr. Lagins Cr. t r i b . 40.5 - 59.1 59.4 - 91.3 48.2 - 80.5 48.2 - 82.6 25.0 - 34.3 Ni l - 10.9 Nil - 200 25.1 - 66.3 500 - 1,400 20.4 - 27.1 21.1 - 60.1 400 - 1,200 a A l l values rounded to the nearest 100 m ^Includes the volume of sediment exported from the upper "wedge" reach. cVolume of sediment gained between 1965 and 1982 116 The results from Mountain Creek and Lagins Creek tributary in d i -cate a sediment transport rate of approximately 400 to 1,400 m /yr; Madej (1978) found a similar rate in a creek subjected to an increase in sediment load following land-use disturbance. In Mosquito Creek tributary, the volumetric range of the "sedi-ment wedge" overlaps that of the sediment inputs, implying a very low rate of sediment transport ( n i l to 200 m /yr); this i s corroborated by the ab-sence of terraces along the creek, and the dense vegetation cover downstream of the "wedge" nose and along the Branch 10 haul road. In Armentieres Creek, the volume of material transferred between the three "sediment wedges" was used to estimate the rate of bed load trans-port (see Table 4.14). Between 1934 and 1982, the upper "wedge" reach received 19,800-33,300 m more sediment than i t stored, implying a mean annual output of 400-700 m . The volume of sediment stored in the middle "wedge" i s comparable to the volume of sediment delivered from the adjacent land. Therefore, the middle "wedge" reach acted as a transportation conduit for the upper "wedge" output, and delivered 19,800-33,300 m of sediment to the lower "wedge" reach. Of this amount, 10,200-17,600 m contributed to the development of the lower "wedge", and 100 - 25,500 m was transported further downstream. The latter range implies a mean annual sediment export rate, between 1934 and 1982, of n i l - 500 m , and as the lov/er "wedge" has developed since 1965, the former range implies a mean input rate of 600-1,000 m3/yr. Estuarine sedimentation was also examined, from hydrographic charts and air photos, at the mouths of Mountain, Lagins, and Armentieres Creeks. As no evidence of aggradation was found (e.g. see Plate 1.1), i t i s concluded tentatively that an insignificant volume of bed load reaches their •a estuaries; for,example, in Armentieres Creek, the 100-25,500 m of sediment 117 exported from the lower "wedge", whose terminus i s 180 metres from the mouth of the creek, would only result in an average estuarine aggradation of up to 8 mm/yr. Morphology Along the "sediment wedges" in Mountain Creek and Lagins Creek tributary are discontinuous and unpaired terraces (see Plate 4.1). The heights of these terrace levels above the surface of the "sediment wedges" were measured during f i e l d surveys, and linear interpolation between points facilitates the delineation of the terrace "profiles" on Figures 4.7 and 4.8. The volumetric difference between present "wedge" profiles and terrace "profiles" i s estimated with the same procedure as described in sub-section 4.1.2: incremental volumes are computed and then summed, using measurements of length and depth from long profile plots, and width from air photos. Observations from a i r photos suggest that the terraces were produced episodically by the slow, downstream movement of a sediment "wave", and are not the remnants of a single episode of aggradation; that i s , by travelling downstream as a zone of local aggradation, a single sediment "wave" can leave residual deposits as a sequence of terraces along the stream, identical to those that would be produced by a single, stream-length aggradation/incision episode (see Figure 4.14). Consequently, much of the coarse sediment present in the downstream portion of a "wedge" originates from stream incision, and thereby terrace formation, in the upstream "wedge" deposits. This i s demonstrated in Mountain Creek by comparing the most downstream "wedge" reach (0.35 - 0.80 Km in Figure 4.7) with the adjacent 118 PLATE 4.1: An example of a stream terrace in Mountain Creek View looking upstream from 0.5 Km in Figure 4.6. 119 FIGURE 4.14: Schematic diagram of "sediment wedge" movement and stream terrace formation by the downstream progression of a sediment "wave" 120 upstream reach (0.80 - 1.20 Km in Figure 4.7). The lower reach contained 3 20,600 m of sediment in 1976, which, from air photo evidence, appears to be the year of maximum aggradation. Of this amount, 11,200 m can be accounted for by the erosion of adjacent stream banks, and the remaining volume (9,400 m ) i s assumed to be bed load delivered from the upstream reach (see Table 4.15). However, the volumetric difference between the "wedge" surface and the terrace "profile" in the upstream reach implies that 19,000 m of sediment has been transported downstream (see Table 4.15). This apparent 3 discrepancy between the assumed input to the lower reach (9,400 m ) and the 3 estimated output from the upper reach (19,000 m ) can be explained in terms of the episodic and "wave-like" nature of bed load transport. Air photo analyses indicate that in 1976, part of the upstream reach (0.80 - 0.90 Km in Figure 4.7) was severely aggraded, but that by 1982 the zone of maximum aggradation was farther downstream (0.65 - 0.75 Km in Figure 4.7). From these observations, i t appears that the terraces found along the upper reach were produced by the downstream progression of a sediment "wave"; thus, the total volume of material exported from the upper reach i s less than that suggested by the terrace "profile" (19,000 m ). 3 By 1982, approximately 11,100 m of "wedge" material remained in 3 the lower reach, which implies the export of 9,500 m of sediment since o 1976; this corresponds to a mean bed load transport rate of 1,600 m /yr. Between June 1982 and May 1983, no major morphologic changes were noted along Mountain Creek. However, a large rain storm on January 26-27, 1984 destabilized the stream bed and banks i n i t s downstream reaches, resulting in severe bank erosion, displacement of in-stream large organic debris, and channel aggradation by fine gravels (L. Beaven, pers. comm., 1984). The latter observation suggests disturbance and transport of the downstream "wedge" deposits, while the presence of recently-uprooted alders 121 TABLE A.15: Volumes of intra-"wedge" sediment transfers 3 Watershed Reach (Km) "Wedge" (xlO 3 Present Volume m3) Removed Sediment Input (xlO 3 m3) Banks Bed Load Mountain Cr. 0.35 - 0.80 0.80 - 1.20 11.1 9.5 19.0 11.2 9.Ab Lagins Cr. tr i b . 0.15 - 0.70 0. 70 - 1.A0 1. A0 - 1.70 1.70 - 1.85 15.8 A.6 7.0 1.9b 1.9 10.8 5.7 6.5 10.l b a A l l values rounded to the nearest 100 m3 Walues derived by the arithmetic difference between the estimated values (unmarked) 122 on the stream estuary (L. Beaven, pers. comm., 1984), implies that some sediment may have reached the sea. Thus, i t appears that the Mountain Creek "wedge" i s again advancing downstream in the form of a sediment "wave". In Lagins Creek tributary, the same procedure i s used to quanti-fy the sediment transfers along the length of the entire "wedge". Four reaches can be distinguished morphologically: the most downstream reach (0.15-0.70 Km i n Figure 4.8) developed a "wedge" after a 1976 winter storm had diverted the creek along a haul road (see sub-section 2.4.4); the adja-cent and the most upstream reaches (0.70 -1.40 Km, and 1.70 - 1.85 Km, in Figure 4.8) are sections of haul road along which the creek was diverted in 1976, and terraces were found; the intervening reach (1.40 - 1.70 Km i n Figure 4.8) i s a length of natural stream channel and lacks terraces. Air photo observations suggest that sediment has been transported from the three upper reaches into the lower reach since 1976, and quanti-tative support i s given by the estimated volumetric transfers of sediment (see Table 4.15). In 1982, the lower reach had an approximate "wedge" volume of 15,800 m3, of which 5,700 m3 i s derived from the erosion of con-tiguous stream banks; the remaining 10,100 m3 presumably represents the bed load influx from upstream since 1976, and i s equivalent to a mean transport rate of 1,700 m3/yr. This input volume corresponds closely to the estimated volume of sediment exported from the three upstream reaches between 1976 and 1982 (10,800 m3 or 1,800 m3/yr; see Table 4.15). Therefore, two-thirds of the coarse sediment present in the lower reach of Lagins Creek tributary was transported rapidly i n the form of a bed load "wave", following creek diversion and incision along the upstream haul road. The short-term sediment transport rates derived from morphologic evidence in Mountain Creek (1,600 m3/yr) and Lagins Creek tributary (1,700-123 1,800 m /yr) agree reasonably with the long-term rates estimated from input/ storage volumes (400-1,400 m /yr); differences are explicable in terms of the widely acknowledged temporal variability of sediment transport. Terraces are absent in Mosquito Creek tributary, but, as men-tioned i n sub-section 2.4.2, the main bridge crossing collapsed after logging was completed, introducing 1,300-1,600 m of bed load c a l i b r e material (see Table 4.1), and additional fine sediments, into the downstream channel reach. This material has been transported only 270 metres (see Figure 4.6 and Plate 1.2B), due to i t s coarse texture and the diversion of most of the stream-flow along a haul road. The total volume of sediment presently deposited downstream of the former bridge crossing i s estimated to be 2,600-4,200 m3. This requires the addition of 1,300-2,600 m3 of material, of which 90% represents bed load, since the time of bridge collapse. From air photo analyses, i t i s known that the bridge collapsed sometime between 1964 and 1974. If i t i s assumed that the bridge failed in 1969, then the mean bed load transport rate i s 100-200 m /yr. This i s an order-of-magnitude lower than the transport rate found in Mountain Creek and Lagins Creek tributary, and gives quantitive support to the previous quali-tative proposal (from input/storage estimates) of negligible sediment trans-port in Mosquito Creek tributary. 4.2 Causes of "sediment wedge" in i t i a t i o n and development As demonstrated in Section 4.1, in three of the four study water-sheds, the single most important sediment source i s stream bank erosion, which accounts for more than half of the total sediment input. Without stream bank retreat, there would be significantly less sediment available to in i t i a t e a "wedge", and moreover, insufficient lateral room for "wedge" 124 development. Therefore, the factors leading to "sediment wedge" inception and growth are closely related to those causing stream bank destabilization. Stream banks may become destabilized by a natural or logging-induced increase in streamflow and/or stream sedimentation, and/or as the result of timber harvesting activities. An analysis of pr e c i p i t a t i o n and stream runoff trends i n the Queen Charlotte Islands, during this century, i s reported by Karanka (1984). Unfortunately, the u t i l i t y of his results i s seriously limited by the lack of a long-term, homogeneous c l i m a t i c and hydroraetric data base. Despite this l i m i t a t i o n , Karanka (1984) noted that by combining the cumulative percent deviation curves of individual precipitation data series, there was an overall regional pattern of short-term fluctuations. Since 1946, there have been two periods of generally above normal (1958-1966 and 1974-1980), and two periods of generally below normal (1946-1957 and 1967-1973), precip-i t a t i o n (Karanka, 1984). Compared with the long-term (20-40 year) deviations that characterize the South Coastal p r e c i p i t a t i o n record, the Queen Charlotte Islands have experienced much more frequent fluctuations. Karanka attempted to relate periods of above normal precipitation with known episodes of slope f a i l u r e and increases i n channel width (and hence, "sediment wedge" formation). In part i c u l a r , he suggested that the 1958-1966 period of above normal precipitation and runoff coincided with the onset of logging and channel enlargement in three of the four study water-sheds (Armentieres Creek, Mountain Creek, and Mosquito Creek tributary). Except for a very small area in Mountain Creek, logging and channel enlarge-ment in the study watersheds were initiated during the early to mid 1960's. The years between 1961 and 1966 were marked by a year-to-year " f l i p " between above and below normal pr e c i p i t a t i o n amounts, and hence, channel changes were not coincident with a period of sustained above normal precipitation. 125 Karanka (1984) noted that a major episode of slope f a i l u r e s occurred in 1949, during a period of generally below normal precipitation, due to the combination of uncommonly wet s o i l moisture conditions and the exceptional, magnitude 8.0 earthquake. Many slope failures also occurred in 1978, during a period of below normal runoff because of the intensity of the October 30 rain storm (Karanka, 1984). Thus, in forested basins, channel enlargement may occur only in response to intense storm events, as the magnitude of secular climatic changes may be insufficient to trigger severe bank erosion. Hydrologic changes resulting from timber harvesting were not investigated in this study, but as indicated in Section A.6, the changes accompanying logging elsewhere in the Pacific Northwest have been varied and contradictory. Moreover, as the present study watersheds were largely high-lead yarded and have low road densities, the area of s o i l compaction i s limited, and therefore, peak flows are unlikely to be augmented by rapid overland flow. Similarly, logging eliminates the evapotranspiration and interception of precipitation by the forest canopy, but as these processes are at a minimum during winter, when the s o i l moisture status i s at a maximum, peak flows are unlikely to be increased significantly. Further-more, in cases where stream bank erosion has occurred, high flows per se have not been blamed (Rice, 1981; Toews and Moore, 1982b; Lyons and Beschta, 1983). Stream channel aggradation (natural or logging-induced) may de-tabilize stream banks by creating channel changes that deflect streamflow against the banks, which are then undercut and retreat by cantilever c o l l -apse; the added sediment may enhance aggradation and so the cycle may be repeated. In general, only point source erosion processes (e.g. landslides, 126 stream bank erosion) deliver sufficiently large volumes of sediment to short reaches of channel rapidly enough to i n i t i a t e serious, local aggradation; non-point source erosion processes (e.g. s o i l creep, slope wash) have much lower instantaneous rates of sediment delivery. As the rate of sediment delivery to the study creeks by land-slid e s i s lower i n clearcut areas than i n forested areas (see Table 4.6), stream sedimentation can not be attributed to logging activities on h i l l -slopes. The largest landslide identified in the Mountain Creek basin del-ivered approximately 8,300 m3 of sediment into the trunk stream, circa 1950, when the basin was unlogged. However, bank retreat and "wedge" development did not occur until the onset of logging, 20 years later. Therefore, stream bank destabilization in the study watersheds appears unlikely to be the result of natural or logging-induced increases in streamflow or sediment delivery by hillslope failures. In contrast, the s p e c i f i c logging practices used i n the four study basins (see Section 2.4) have encouraged stream bank destabilization. Three lines of evidence support this conclusion: the absence of severe bank erosion and "sediment wedges" i n forested basins i n the Queen Charlotte Islands, during the past 300 years;the coincidence of severe bank retreat, "wedge" development, and logging in the study basins; and the occurrence of significant riparian erosion in other basins in the Pacific Northwest that have been subjected to similar timber harvesting practices. The condition of currently forested watersheds i n the Queen Charlotte Islands has been investigated in the f i e l d by Hogan (1984) and by the writer. High rates of sediment transport and substantial bed-form (i.e. pool-riffle) adjustments may occur during high flow events in these streams, but episodes of severe channel aggradation are infrequent. For example, channel aggradation i n Hangover Creek (see Figure 2.1), following a major 127 rain storm on January 26-27, 1984, was limited to the local i n f i l l i n g of pools (Hogan, 1984; L. Beaven, pers. comm., 1984). Similarly, Government Creek and Security Creek (see Figure 2.1), at present, do not have notice-able channel aggradation or stream bank erosion problems, and air photo surveys by Rood (1984) and by the writer, of other forested basins indicate likewise; in general, stream bank retreat in forested watersheds i s confined to local "pockets" of erosion between old-growth trees. Stream channel sta b i l i t y in forested watersheds in the Queen Charlotte Islands also i s manifest over longer timescales. In the Hangover, Government and Security Creek watersheds, there are 9, 13, and 27 slope failures, respectively, that probably exceed 100 years of age (K.M. Rood, pers. comm., 1984). The proportion of these that delivered sediment to the creek i s not known. However, as the streamside vegetation in these water-sheds i s more than 250 years old (according to forest cover maps compiled by British Columbia Forest Service personnel), episodes of major riparian erosion apparently have not occurred during the past three centuries. In addition, the lack of stream terraces in these creeks implies that the streams usually are competent to transport the sediment load supplied by localized riparian erosion and episodic slope failures. In forested watersheds, changes in stream channel morphology that persist for decades are restricted chiefly to reaches affected by debris flows and torrents. In the Queen Charlotte Islands, such events comprise only about 10% by number of a l l slope failures (Rood, 1984), occur mainly in short, headwater reaches, and typically have a depositional zone an order-of-magnitude shorter than a "sediment wedge". Consequently, most of the drainage nework in forested basins has not undergone any net morphologic change during the past three centuries. 128 Logging a c t i v i t i e s are implicated as the triggering agent i n stream bank destabilization because of their temporal correspondence with riparian retreat and "sediment wedge" development in the study watersheds. Forest cover maps show that, prior to logging, almost the entire area of the study watersheds had a tree cover more than 250 years old. Linear tracts of younger vegetation, identified from air photos, occur sporadically in the study watersheds and presumably represent revegetated landslide scars or snow avalanche tracks. The number of landslides, estimated to be more than 100 years old, that reach the study creeks does not exceed 3 per basin (see Table 4.5). As only one revegetated failure delivered more than 4,000 m3 of sediment, episodes of major stream channel aggradation and stream bank erosion i n the study watersheds, during the 300 year period prior to logging, probably were limited in magnitude and duration. The absence of old a l l u v i a l terraces, and the lack of deciduous and young coniferous trees beside the study creeks i n the 1930's, supports t h i s contention. The old-growth forest root network may have reinforced the stream banks, so that even large landslide inputs could not easily induce riparian erosion; for example, the morphologic impact of the delivery of 8,300 m3 of sediment into Mountain Creek by a single debris slide, circa 1950, i s not discernible on the 1954-1955 air photos, although the slide scar i s s t i l l bare (see Plate 1.1A). Thus, despite the probable occurrence of exceptional seismic and hydrologic events during the past three centuries, the most extensive and persistent episode of channel widening and aggradation in the study xvater-sheds (i.e. "sediment wedge" development) occurred only after the onset of logging. The absence of "wedges" in other, logged basins implies that on-site circumstances are of paramount importance. The logging practices used i n the study watersheds are summarized in Table 2.3: the common feature i s the considerable extent of logging a c t i v i t y i n the r i p a r i a n zone. From investigations conducted throughout the Pacific Northwest (see Section A.5), i t i s well established that: "harvesting operations that potentially cause damage to banks include f e l l i n g across streams, yarding through or across streams, machine operation near streams, and the removal of vegetation which has roots that strengthen s o i l structure" (Chamberlin, 1982, p.21). A l l of these act i v i t i e s apply to the study watersheds, but Chamberlin (1982, p.13) notes in particular that: "the breakdown and destruction of stream banks by f e l l i n g and yarding are among the most persistent of direct impacts of harvesting, and they are the most d i f f i c u l t to avoid when streamside f e l l i n g or skidding and cross-stream yarding occur." Thus, destabilization of the stream banks in the study watersheds was initiated by logging act i v i t i e s (especially f e l l i n g and yarding) in the ripa r i a n zone. Subsequent erosion of the stream banks i s f a c i l i t a t e d by streamflow, but identification of "channel forming" discharges i s a major problem. An extreme climatic event need not necessarily produce a s i g n i f i -cant geomorphic response, while a less exceptional climatic event may yield a greater geomorphic response i f the f l u v i a l system i s more sensitive to disturbance, due to logging act i v i t i e s in the riparian zone. The i n i t i a l erosion of destabilized stream banks, shortly after logging, may be accomplished by moderate streamflows. For example, Bourgeois (1981) attributed the formation of large gravel bars in the Mamin River, Queen Charlotte Islands, to the erosion of sand and gravel banks, by moderate to high flows, following timber removal from islands within the 130 river floodplain. Larger flows, however, were required to i n i t i a t e the development of "sediment wedges" i n Lagins Creek tributary and M i l l e r Creek, which i s located on the east coast of Graham Island and was logged to the streamside in i t s lower reaches during the early 1970's. Miller Creek developed a very small "wedge" i n the logged section after a rain storm i n August 1983 had i n i t i a t e d severe stream bank erosion (L. Beaven, pers. comm., 1983), and Lagins Creek tributary was diverted along a haul road by a 1976 winter storm (see sub-section 2.4.4). In Armentieres Creek, Mountain Creek, and Mosquito Creek t r i b -utary, bank erosion was initiated between circa 1970 and 1974. Although the streamflows responsible are not known, the lack of severe rain storms docu-mented during this period implies that exceptional hydrologic events are not a prerequisite for "wedge" inception. However, f i e l d evidence suggests that high flows are a necessary, but not always sufficient, condition for continued bank retreat and "sedi-ment wedge" growth. That i s , after the passage of a sediment "wave" through a reach, the stream channel adopts a single-thread thalweg position, and alders revegetate the surface of the "wedge", thereby s t a b i l i z i n g the re-maining deposits and hindering further bank erosion. For example, the mid-and up-stream portions of the "sediment wedges" i n Armentieres Creek, Mountain Creek, and Mosquito Creek tributary, decreased their active widths between 1976 and 1982 (see Table 4.3), despite a severe rain storm on October 30, 1978. In contrast, depositional reaches (i.e. the downstream portion of the "sediment wedges") are characterized by stream channel branching and lateral streamflow deflection (see Plate 1.4C), and hence, are susceptible to l o c a l stream bank erosion and sediment transport i n response to large 131 hydrologic events. For example, extensive bank erosion and sediment trans-port occurred along lower Mountain Creek during a severe rain storm on January 26-27, 1984 (L. Beaven, pers. comm., 1984). 4.3 Summary The following points summarize the sources, causes, and mobility of the "sediment wedges" present in the trunk streams of the four study watersheds. 1. The most important sediment delivery processes are stream bank erosion and landslides, which together account for two-thirds to four-fifths of the total sediment input: stream bank erosion delivers 15, 50, 50, and 70%, and landslides deliver 65, 20, 15, and 10% of the total volume of sediment supplied to Armentieres Creek, Mosquito Creek tributary, Mountain Creek, and Lagins Creek tributary, respectively, between 1934 and 1982 (see Table 4.1). Road and hillslope surface erosion, s o i l creep, and tree throw account for the remaining volume. 2. The total volume of sediment delivered to the drainage networks in the study watersheds equals or exceeds the volume stored in the "sediment wedges" (see Table 4.14). The volumetric difference indicates a long-o term (i.e. 1934-1982) sediment output rate of 400-1,400 m /yr from o Mountain Creek and Lagins Creek tributary, nil-500 m /yr from o Armentieres Creek, and nil-200 m /yr from Mosquito Creek tributary. 3. Air photo and f i e l d observations suggest that bed load i s transported episodically downstream as a sediment "wave", so that much of the down-stream "wedge" material i s derived from the erosion and transport of upstream "wedge" deposits. Morphologic evidence implies short-term (i.e.. since "wedge" inception), intra-"wedge" sediment transport rates of 1,600-1,800 m^/yr in Mountain Creek and Lagins Creek tributary, 500-900 m3/yr i n Armentieres Creek, and 100-200 m3/yr i n Mosquito Creek tributary; 90% of this material i s transported as bed load. 4. I n i t i a l stream bank de s t a b i l i z a t i o n appears to have been caused by logging activities in the riparian zone, and especially by f e l l i n g and yarding timber across the stream channel. The streamflow discharges necessary for subsequent bank erosion and "sediment wedge" mobilization vary temporally and spatially with respect to the degree of sta b i l i t y of the "sediment wedge". Contemporary bank erosion and "wedge" move-ments occur only during major hydrologic events along downstream reaches that are aggraded and unstable, as, for example, was observed along the lower reaches of Mountain Creek in response to a large rain storm in January 1984. 133 5.0 CONCEPTUAL IMPLICATIONS The preceding chapter has indicated that sediment transport, and associated morphologic adjustments, in four small, gravel-bed mountain streams in the Queen Charlotte Islands, have been in disequilibrium since the onset of logging in the 1960's. In contrast, currently forested water-sheds experienced comparatively minor and transient episodes of channel widening and aggradation during the same period. The general evidence suggests, therefore, that Schumm's f l u v i a l model may not be pertinent to stream channels draining forested terrain, but may be relevant in circum-stances where channel stab i l i t y i s disrupted by a land-use change (e.g. streamside logging). The more severe the disruption, the greater the potential for compliance with Schumm's model. In the Queen Charlotte Islands, "sediment wedges" are the foremost example of major f l u v i a l dis-turbance, and thus, provide the most favourable opportunity for Schumm's concepts to be relevant in a contrasting environment to that in which they were conceived. Hence, the degree of analogy, and aspects of dissimilarity between Schumm's concepts and the observed behaviour of "sediment wedges" are the focus of this chapter. As the sequence of events observed in the study watersheds since the 1960's may represent high frequency fluctuations about some kind of longer-term trend (see Figure 5.1), contemporary f l u v i a l activity may not reflect the overall, long-term development of the f l u v i a l system. For conceptual purposes, therefore, the temporal perspective i s expanded to encompass the past 8,000 years, when subaerial processes probably have dictated the mode and rate of f l u v i a l activity in the Queen Charlotte Ranges (see sub-section 2.2.1). An implication of the above i s that sediment budgets representative of the past 25 years and the past 8,000 years differ 134 FIGURE 5.1: Hypothetical sequence of events (after Church, 1980). A Long term mean Linear trend Cyclic trend High frequency fluctuations => Time 135 in terms of the relative unimportance of logging-related sediment transfers over the longer timescale. F l u v i a l adjustments during the past three years, and the past five decades, have been observed directly in the f i e l d , and from air photos, respectively, while the frequency of major landslides and f l u v i a l disturb-ances during the past three centuries, and the past eight m i l l e n i a , were inferred from vegetation changes and supposed buried paleosols, respective-l y . Naturally, observations made over years and decades provide a more accurate and detailed description of geomorphic activity than inferences drawn over centuries and mille n i a . Thus, most of this chapter examines Schumm's concepts within the time frame of the l a s t half century, as does Schumm himself. The following discussion i s organized with respect to the three hypotheses proposed in Chapter 1.0. 5.1 First hypothesis The f i r s t hypothesis i s that i n forested watersheds, e x t r i n s i c geomorphic thresholds are rarely crossed, so that sediment transport-limited conditions, and hence major episodes of f l u v i a l disequilibrium, occur i n -frequently. At the shortest pair of timescales (years and decades), episodes of stream bank erosion and channel aggradation in currently forested water-sheds (e.g. Hangover, Government, and Security Creeks) usually are highly localized and short-lived (see Section 4.2). Similarly, episodes of major channel widening and aggradation in the study watersheds were not discern-ible during the 30-year period of pre-logging air photo coverage; even the delivery of 8,300 m3 of sediment to Mountain Creek by a single debris slide, c i r c a 1950, had no observable morphologic impact (see Plate 1.1A). In 136 general, the largest and most persistent morphologic changes i n forest streams i n the Queen Charlotte Islands are produced by infrequent debris flows and torrents, whose impacts are confined usually to short, headwater reaches (see Section 4.2). Hence, i t i s proposed that during the past half century, forest streams have experienced infrequent and spatially-limited episodes of major f l u v i a l disturbance in response to slope failures and flood events generated by the exceedance of extrinsic geomorphic thresholds. The streams apparent-ly were competent to transport most of the sediment supplied, as stream terraces are rare i n Hangover, Government and Security Creeks, and result invariably from the retention of sediment behind large organic debris, rather than from stream incision per se. As resultant channel changes were short-lived, there has been no net morphologic change along most of the drainage network in forested basins during the past 50 years. Because the concept of "dynamic equilibrium" implies both st a b i l i t y and the a b i l i t y to adjust, i t i s an appropriate framework for such situations, where the stream channel i s capable of rapidly regaining the characteristics that existed prior to a disturbance. The age and d i s t r i b u t i o n pattern of vegetation can be used to infe r the l e v e l of geomorphic a c t i v i t y i n unlogged watersheds during the past three centuries (see Section 4.2). An air photo survey of revegetated landslide scars in the study watersheds suggests that the average number of slides older than 100 years i s 5.5 per basin (see Table 4.5). In 16 other basins that have drainage areas less than 13 Km2 (the drainage area of the largest study watershed), the average number of revegetated failures that probably exceed 100 years of age i s 5.1 per basin (K.M. Rood, pers. comm., 1984). Thus, despite the probable occurrence of major seismic and 137 hydrologic events during the l a s t 300 years, the existence of only 5 or 6 slope f a i l u r e s per basin t e s t i f i e s to the infrequent exceedance of an ex-tr i n s i c geomorphic threshold. On average, one-third of the revegetated slope failures reached the study creeks (see Table 4.5), and only one f a i l u r e delivered more than 4,000 m3 of sediment. The limited impact of these inputs i s implied by the absence of old a l l u v i a l terraces, and the lack of young vegetation beside the study creeks i n the 1930's. The predominantly old-growth streamside vegetation cover, and the lack of stream terraces, in the Hangover, Govern-ment, and Security Creek watersheds (see Section 4.2) also suggests long-term stream channel s t a b i l i t y . Therefore, episodes of major channel aggradation and widening during the past three centuries probably were limited in number, magnitude, and duration. Hence, the intr i n s i c geomorphic threshold of stream bank s t a b i l i t y has been crossed.rarely, and forest streams i n the Queen Charlotte Islands have maintained a condition of "dynamic equilibrium", during the past 300 years. The occurrence of episodes of hillslope and stream channel dis-turbance in the study watersheds, during the past eight millenia, i s implied by the existence of paleosols beneath c o l l u v i a l and f l u v i a l deposits (see sub-section 2.2.4). In the lower reaches of Mountain Creek, the alternation of thick sand deposits and seven buried s o i l horizons (see Plate 2.IB) implies that at least seven episodes of overbank flood deposition have occurred at t h i s s i t e during the past few m i l l e n i a . In other reaches of Mountain Creek, and along Armentieres Creek and Lagins Creek tributary, buried paleosols occur either singly or in pairs, which implies fewer epi-sodes of s o i l burial. The number of buried paleosols typically should vary along and between creeks, as individual landslide and overbank flood events impact a limited area within a watershed, and may be non-synchronous between 138 watersheds. As landsliding and overbank flooding occur usually in response to low frequency-high magnitude climatic and hydrologic events (see sub-section 2.3.1), episodes of activity are initiated by the crossing of an extrinsic geomorphic threshold. It i s d i f f i c u l t to establish an accurate chronology, and hence the frequency of occurrence, of these threshold crossings during the past eight millenia, because of the incomplete and spatially highly variable nature of the stratigraphic record. Moreover, the absence of f l u v i a l paleo-terraces in the study watersheds suggests either that these episodes of sediment delivery did not cause severe channel aggradation, or that the evidence has been obliterated subsequently by hillslope and channel erosion processes. Hence, during the last 8,000 years, forest streams in the Queen Charlotte Islands ordinarily were competent to transport the sediment load injected by episodic landslides. Major periods of f l u v i a l disequilibrium probably were infrequent, highly localized, and transient. It i s concluded, therefore, that Schumm's f l u v i a l model does not accurately describe the behaviour of forested drainage systems, which have maintained a condition of "dynamic equilibrium" throughout the past eight millenia. In contrast, i t i s apparent that since logging, the study creeks have been disturbed from their former equilibrium condition and have dev-eloped unusual f l u v i a l phenomena, namely "sediment wedges". The conceptual implications of the i n i t i a l appearance, and subsequent movement, of the "sediment wedges" are addressed by the second and third hypotheses, respect-ively. 5.2 Second hypothesis The second hypothesis i s that land-use effects, specifically 139 streamside logging act i v i t i e s , may cross an extrinsic geomorphic threshold, lower an intrinsic geomorphic threshold (stream bank stability), and thereby i n i t i a t e major episodes of stream bank erosion and trunk stream aggradation (i.e. "sediment wedge" formation). In Chapter 4.0, i t was indicated that shortly after streamside timber harvesting in the study watersheds, the stream banks retreated rapid-ly and supplied much of the sediment incorporated into the "sediment wedges". The s p e c i f i c logging practices were i d e n t i f i e d as the c r i t i c a l factor i n the d e s t a b i l i z a t i o n of the stream bank deposits: t h i s anthro-pogenic land-use change corresponds conceptually to the exceedance of an extrinsic geomorphic threshold. Although nearby watersheds also have been logged, the stream banks were disturbed much less severely, so that, con-ceptually, vegetation removal lowered, but did not cross, the e x t r i n s i c geomorphic threshold. Thus, the severity of the logging practices used in the study watersheds (i.e. cross-stream f e l l i n g , cross-stream yarding, and in-stream skidding) increased the sensitivity of the stream banks to f l u v i a l erosion processes, to the extent that subsequent episodes of major riparian erosion occurred during flows that previously had been capable only of minor damage. In addition to the physical disturbance of the stream banks during vegetation removal in the study watersheds, the subsequent decay of the root network results in a loss in root strength, and hence a lowering of the resistance of the stream banks to erosion. The l a t t e r corresponds to the progressive decline of the threshold of stream bank st a b i l i t y (in s o i l mechanics terminology) and of an intrinsic geomorphic threshold (in Schumm's conceptual terminology). Consequently, there i s a "reaction time" assoc-iated with the threshold-controlled temporal delay between the devegetation, and the subsequent erosion, of stream banks; logged h i l l s l o p e s i n Rennell 140 Sound have a "reaction time" of 4 to 6 years (Schwab, 1983), and a s i m i l a r lag time may apply to post-logging riparian retreat in the study watersheds. As the extent of ri p a r i a n retreat i n other logged basins i n the Queen Charlotte Ranges i s comparatively minor, i t i s apparent that lowering the intrinsic geomorphic threshold (via root decay) i s largely ineffectual unless the logging-related extrinsic geomorphic threshold has been crossed previously. In order to create the sediment transport-limited conditions that have persisted for the past quarter century in the study creeks, the i n i t i a l volume of delivered sediment had to be large enough to exceed the sediment-transporting capacity of the streams. The combination of a high flow event and an abundant supply of easily-erodible stream bank and/or road-related sediment probably would satisfy this requirement. For example, the erosion of the haul road along which Lagins Creek tributary was diverted by a 1976 rain storm (see sub-section 2.4.4), and the input of sediment from the collapse of the Branch 10 haul road bridge in Mosquito Creek tributary (see sub-section 2.4.2), illustrates the major role that single, large hydrologic events and haul roads have played in "sediment wedge" ini t i a t i o n . In summary, streamside logging activities in the study watersheds crossed an e x t r i n s i c geomorphic threshold and lowered an i n t r i n s i c geo-morphic threshold (i.e. the threshold of stream bank s t a b i l i t y ) . This facilitated the most extensive and persistent changes in channel morphology (i.e. "sediment wedge" formation) to have occurred during the past 25 years in the Queen Charlotte Islands. The extent to which subsequent "sediment wedge" evolution i s compatible with Schumm's model i s discussed with respect to the third hypothesis. 141 5.3 Third hypothesis The t h i r d hypothesis i s that sediment transport in severely aggraded reaches with "armoured" streambeds (i.e. "sediment wedges") i s episodic and complex, due to geomorphic threshold, reaction, relaxation, and recovery effects. In the study creeks, the surfaces of the "sediment wedges" are "armoured" by the concentration, and often the imbrication, of large clasts (see sub-section 4.1.3). The sub-surface "sediment wedge" material i s thus shielded from entrainment u n t i l the "armour" layer i s disturbed. This occurs during infrequent, high flow events so that contemporary f l u v i a l adjustments in the study creeks are dominated by the episodic transport of bed load. For example, a major rain storm i n January 1984 was the f i r s t event i n at least 20 months to have produced any apparent change i n the morphology of the "sediment wedge" in the lower reaches of Mountain Creek (see sub-section 4.1.3). The concept of an i n t r i n s i c geomorphic threshold of sediment entrainment i s redundant i n such situations, as the "armour" layer i s disturbed only during high flow events and, thus, corresponds conceptually to the crossing of an extrinsic geomorphic threshold. In semi-arid areas, ephemeral stream channels also exhibit "episodic behaviour", but primarily because of the highly skewed frequency distribution of streamflow, rather than because of high sediment entrainment thresholds (see Figure 1.5). "Complex response" i s the term used by Schumm to describe s i t -uations i n which a single event triggers a series of damped c u t - a n d - f i l l cycles. This concept implies that the pattern of sediment transport i s complex in space and time: at a fixed point in time, the magnitude of the response i s damped i n a downstream direction within a watershed, and may 142 vary between watersheds; at a single location, more than one phase of aggradation and/or degradation occurs over time. Air photo and f i e l d observations of Mountain Creek and Lagins Creek tributary indicate that "sediment wedge" material i s transported as a slow-moving bed load "wave", whose magnitude i s attenuated in a downstream direction (see sub-section 4.1.3). However, the pattern of "sediment wedge" movement since logging involves the transport not only of the sediment delivered originally, but also of the subsequent sediment inputs from h i l l -slope and r i p a r i a n erosion. Large volumes of additional sediment were delivered to Mountain Creek and Lagins Creek tributary only at sites where the arrival of the sediment "wave" induced aggradation, braiding, and later-a l bank erosion; hence, the effect of the input was always to magnify the residual logging-related response. In contrast, the attenuation of the sediment "wave" derived from the upper "wedge" i n Armentieres Creek was masked downstream by the irregular superimposition of highly localized and episodic landslide inputs. Moreover, in 1975, a major debris flow augmented the volume of sediment stored i n the Armentieres Creek upper "wedge", by one-quarter to one-third, but a log jam prevented the downstream transport of material; i t s release upon disintegration of the log jam may complicate further the pattern of movement of the middle and lower "wedges". Subsequent sediment additions do not vio l a t e the d e f i n i t i o n of "complex response" i f they are a by-product of "sediment wedge" movements or are a delayed response to the o r i g i n a l event (i.e. logging). For example, riparian erosion beside the sediment "wave" i n Mountain Creek and Lagins Creek tributary i s comparable with the temporary feedback loop between channel aggradation and widening that Patton and Schumm (1981) observed in ephemeral stream channels. Landslide inputs into Armentieres Creek, how-ever, were uncoupled from f l u v i a l a c t i v i t i e s and were initiated in unlogged 143 terrain, and thus, are not a "complex response" in the Schumm sense. Despite the identification of a sediment "wave" in Mountain Creek and Lagins Creek tributary, there i s no evidence that more than one episode of aggradation or degradation occurred at any cross-section; the input of stream bank sediment, as a result of sediment "wave" arrival, represents a single, uninterrupted episode of channel aggradation. There are two poss-i b l e reasons for t h i s lack of evidence. F i r s t l y , the products of e a r l i e r c u t - a n d - f i l l cycles (e.g. multiple terraces) may have been destroyed by later events, and air photo coverage i s too infrequent to demonstrate their occurrence. Secondly, the propensity of gravels to be deposited during the early stages of flood recession, and to "armour" streambeds, may result in the rapid attenuation, and subsequent stabilization, of the bed load "wave". If the l a t t e r reason i s correct, as the writer contends, then "sediment wedge" movements in Mountain Creek and Lagins Creek tributary have been slow and without the at-a-station cut-and-fill tendencies that define "complex response" in the Schumm sense. Some coarse material may have moved indepen-dently of the main "wave"^ but the absence of multiple terraces implies that the accompanying phases of channel aggradation were comparatively minor and were not separated by periods of stream incision. However, the potential for "complex response" exists in Mountain Creek, because a large volume of "sediment wedge" material i s stored behind two log jams i n the upper portion of the "wedge". In these upper reaches, the sediment "wave" has passed through and the stream channel, at present, i s incised into the residual "wedge" deposits. Upon disintegration of the log jams, the release of sediment may cause severe channel aggradation immediately downstream. This would represent a second " f i l l " phase at one place in response to the same event (i.e. logging) and, thus, i s compatible 144 with Schumm's definition of "complex response". The release of the upper "wedge" material in Armentieres Creek, upon disintegration of the log jam, would not constitute "complex response" as the log jam, and much of the sediment trapped behind i t , resulted from a debris flow initiated in un-logged terrain. In summary, Schumm's definition of "complex response" i s incom-patible with contemporary "sediment wedge" movements, which are heavily damped in comparison with the f l u v i a l adjustments that Schumm observed in more "sensitive", semi-arid systems. The slow, attenuated "sediment wedge" response i s better interpreted in terms of Mackin's (1948) concept of the "graded" stream. Mackin (1948, p.478) stated that "degrading" i s down-cutting approximately at "grade" and that "aggrading" i s upbuilding approx-imately at "grade". Hence, the relatively simple relaxation pattern of "sediment wedges" may be considered as an adjustment at "grade" toward the new equilibrium condition (see Figure 1.6); the latter may be attained at a cross-section only after the passage of a sediment "wave", and within a watershed only after the complete attenuation of the "wave". As the Queen Charlotte Islands climate i s not limiting to veg-etation growth, the surfaces of residual, "sediment wedge" deposits, in parts of a l l four study creeks, are currently supporting young alders. As vegetation growth increases the resistance to erosion of the deposits, a self-enhancing feedback loop probably w i l l develop, so that with time, the "wedge" deposits w i l l become increasingly stable and well-vegetated; this process i s the exact reverse of that which accompanied logging, and con-sequently, corresponds conceptually to the gradual raising of the intrinsic geomorphic threshold. It might be expected, therefore, that the study creeks w i l l eventually return to a condition of "dynamic equilibrium"; the 145 time taken by the f l u v i a l system to return to i t s pre-disturbance condition i s called the "recovery time" (see sub-section 1.5.3). The extent of recovery of a "sediment wedge" i s a function chief-ly of the time available for vegetation to re-establish. It was shown in sub-section 4.1.3 that much of the downstream "wedge" material i s derived from the erosion, and subsequent transport as a bed load "wave", of upstream "wedge" deposits. Thus, upstream reaches are impacted by sedimentation, and begin recovery, before downstream reaches. This aspect of spatial complex-ity i s manifest in Mountain Creek, which has, at the present time, relative-l y stable upstream reaches, with a revegetated "wedge" surface and an entrenched stream channel, and unstable lower reaches, where aggradation, braiding, and lateral bank erosion are occurring (see Plate 1.4C). The occurrence of intervening major events may delay recovery of the study creeks (see Figure 5.2) and thereby complicate this simple pattern of increasing revegetation i n an upstream dire c t i o n . For example, i n Armentieres Creek, the upper "sediment wedge" has only recovered slightly from the major 1975 debris flow (see Plate 1.4A), while the surfaces of the middle and lower "wedges", which were not affected directly by the debris flow, are almost entirely revegetated; the latter reaches may yet be dis-turbed by the remobilization of the upper "wedge" deposits, following the disintegration of the log jam. Similarly, disintegration of the log jams in the upper reaches of Mountain Creek may renew aggradation, and delay recovery, in the downstream reaches. The length of the temporal delay depends upon the s p e c i f i c characteristics (i.e. magnitude, frequency, duration, ordering and timing), and resultant impacts, of the hydrologic events. Event timing i s particularly important in steep, geomorphically active terrain, where the "recovery time" of the system may be much longer than the recurrence 146 FIGURE 5.2: Schematic diagram of the relation between relaxation, recovery, and equilibrium concepts. A. The impact of logging bO C •H C w •p c A ; c cd CL) e •H i d CD g i/> cd CD O 4 - 1 c •H o > Start of logging Relaxation time N Recovery time Time B. The impact of logging and major hydrologic events Start of logging Major hydrologic events Time 147 interval of climatic events that trigger geomorphic response (Kelsey, 1982; see Figure 1.7) For example, landslides in Rennell Sound are triggered by rain storms that, on average, occur annually (see sub-section 2.3.1), while the scars take several decades to recover f u l l y (see sub-section 2.3.3) The location and areal extent of the delay i s determined by the particular sediment delivery process. Slope failures occur sporadically in time and space, so that the input locations are not easily predicted and their effects usually are highly localized; for example, the 1975 debris flow in Armentieres Creek impacted only the upper "wedge" reach. Bank retreat rates along a "wedge" depend upon the degree of spatial and temporal coincidence of logged stream banks that approach the threshold of in s t a b i l -i t y and hydrologic events that are sufficiently large to cross this thres-hold and induce erosion. In p a r t i c u l a r , as l a t e r a l bank erosion i s encouraged along aggraded reaches, stream banks flanking a sediment "wave" are especially susceptible to erosion; for example, severe riparian erosion occurred along the aggraded, lower reaches of Mountain Creek in response to a major rain storm in January 1984 (see sub-section 4.1.3). In general, "wedge" deposits should be eroded less frequently and less severely during the later stages of recovery because of the rising intrinsic geomorphic threshold associated with the stabilizing vegetation cover. Finally, stream bank erosion and "sediment wedge" movements are sporadic in time and space not only within a watershed, but also between watersheds. For example, as logging commenced at different times in the different watersheds (see Table 2.2), "sediment wedge" formation was non-synchronous. In particular, Lagins Creek tributary developed a "wedge" about a decade later than the other study creeks. The degree of channel recovery also differs between, as well as within, watersheds; for example, Mosquito Creek tributary i s more f u l l y recovered than the other study 148 ' creeks. In summary, the location and magnitude of "sediment wedge" move-ments vary temporally within each study creek, and episodes of aggradation and degradation may occur simultaneously along a creek, but be out-of-phase in different creeks. Moreover, the relevance of Schumm's terms "intrinsic geomorphic threshold" and "complex response" i s limited by streamflow com-petency to entrain stream bed and bank sediments. Therefore, f l u v i a l adjustments in these gravel-bed channels may be best described as a simple, or Mackin-type, response to a change in an extrinsic variable. 149 6.0 CONCLUSIONS Based upon investigations of ephemeral stream channels in semi-arid regions, Schumm (1977) proposed that the evolution of the f l u v i a l system i s dominated by long periods of relative stability and brief periods of major morphologic change, during which sediment transport-limited con-ditions prevail. This model i s appropriate in a semi-arid environment because of the skewed magnitude-frequency distribution, and localized occur-rence, of precipitation, the sparse vegetation cover, and the predominance of sand-bed streams. As a result of these factors, semi-arid stream channels are "sensitive" to disturbance and recover their pre-disturbance configuration slowly, and a large proportion of the sediment load i s trans-ported by infrequent flows. These adjustments are embodied within Schumm's concepts of "dynamic metastable equilibrium", "episodic behaviour", "geo-morphic thresholds", and "complex response". The objective of this thesis was to examine the relevance of Schumm's f l u v i a l model in a contrasting environment to that in which i t was developed. The evolution of perennial, gravel-bed streams in the per-humid Queen Charlotte Islands provided such an opportunity. Compared with semi-arid stream systems, forest streams in the Queen Charlotte Islands normally are "less sensitive", and exhibit greater resilience, to major hydrologic events, and have essentially maintained a condition of sediment supply-limitation and "dynamic equilibrium" during the past 8,000 years. Schumm's f l u v i a l model i s overly elaborate i n such circumstances. However, the behaviour of perennial, gravel-bed streams approach-es that of semi-arid, sand-bed channels in some respects when they are modified in certain ways. For example, streamside logging activities in the 150 four study watersheds increased the "sensitivity" of the stream banks to erosion by moderate streamflows. This resulted in major episodes of channel widening and the development of aggradational "sediment wedges". As "geo-morphic thresholds" were crossed during "sediment wedge" inception, the study watersheds may be said to have exhibited "episodic behaviour" in the Schumm sense, and created sediment transport-limited conditions. The latter caused a spatially and temporally complex pattern of sediment transport in the study creeks. However, only one cycle of aggradation and degradation occurred at any one cross-section, so that "sediment wedge" movements are more compatible with a simple, or Mackin-type, adjustment at "grade", than with Schumm's concept of "complex response"; the heavily damped response of the "sediment wedges" i s promoted by their rapid revegetation and "armouring". In terms of geomorphic "equilibrium" concepts, contemporary f l u v i a l adjustments in the study watersheds represent the single most signi-ficant disturbance of the f l u v i a l system to have occurred in the Queen Charlotte Islands during the past eight millenia. This i s consistent with the concept of "dynamic metastable equilibrium". However, due to the op-eration of stream channel recovery processes, i t i s anticipated that the study creeks w i l l eventually return to their pre-disturbance condition of "dynamic equilibrium". 151 BIBLIOGRAPHY ALLEN, J.R.L. (1974) Reaction, Relaxation and Lag i n Natural Sedimentary Systems: General Principles, Examples, and Lessons Earth-Science Reviews, V.10, p.263-342. ALLEY, N.F. and B. Thomson (1978) Aspects of Environmental Geology, Parts of Graham Island, Queen Charlotte Islands Province of British Columbia, Ministry of the Environment, Resource Analysis Branch, B u l l e t i n No. 2, 64 pp. ANDERSON, E.W. and N. J. Cox (1978) A Comparison of Different Instruments for Measuring Soil Creep Catena, V.5, p.81-93. ANDERSON, H.W. (1971) Relative Contributions of Sediment from Source Areas, and Transport Processes in Forest Land Uses and Stream Environment, J. T. Krygier and J. D. Hall (Eds.), Symposium Proceedings, Oregon State University, Corvallis, p.55-63. ANDERSON, H.W., M. D. Hoover and K. G. Reinhart (1976) Forests and Water: Effects of Forest Management on Floods, Sedimentation, and Water Supply U.S.D.A. Forest Service, General Technical Report PSW-18, 115 pp. BAKER, V.R. (1977) Stream-Channel Response to Floods, with Examples from Central Texas Geological Society of America Bulletin, V.88, p.1057-1071. BANNER, A., J. Pojar and R. Trowbridge (1983) Ecosystem C l a s s i f i c a t i o n of the Coastal Western Hemlock Zone, Queen Charlotte Island Subzone (CWHg), Prince Rupert Forest Region, B r i t i s h Columbia Unpublished manuscript, Province of British Columbia, Ministry of Forests, Smithers, British Columbia, 255 pp. BARR, D.J. and D. N. Swanston (1970) Measurement of Creep in a Shallow, Slide-Prone T i l l Soil American Journal of Science, V.269, p.467-480 BEGIN, Z.B. and S. A. Schumm (1984) Gradational Thresholds and Landform Singularity: Significance for Quaternary Studies Quaternary Research, V.21, p.267-274. BERGSTROM, F.W. and S. A. Schumm (1981) Episodic Behaviour in Badlands in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.478-492. 152 BESCHTA, R.L. (1978) Long-Term Patterns of Sediment Production Following Road Construction and Logging in the Oregon Coast Range Water Resources Research, V.14, p.1011-1016. BESCHTA, R.L. and W. L. Jackson (1979) The Intrusion of Fine Sediments into a Stable Gravel Bed Journal of the Fisheries Research Board of Canada, V.36, p.204-210. BEVEN, K. (1981) The Effect of Ordering on the Geomorphic Effectiveness of Hydrologic Events in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.510-526. BIRKELAND, P.W. (1974) Pedology, Weathering, and Geomorphological Research Oxford University Press, New York, 285 pp. BLOOM, A.L. (1978) Geomorphology: A Systematic Analysis of Late Cenozoic Landforms Prentice-Hall, Englewood C l i f f s , New Jersey, 510 pp. B0CKHEIM, J.G., T. M. Ballard and R.P. Willington (1975) Soil Disturbance Associated with Timber Harvesting in Southwestern British Columbia Canadian Journal of Forest Research, V.5, p.285-290. BOSCH, J.M. AND J. D. Hewlett (1982) A Review of Catchment Experiments to Determine the Effect of Vegetation Changes on Water Yield and Evapotranspiration Journal of Hydrology, V.55, p.3-23. BOURGEOIS, W.W. (1981) River Channel and Streambank Changes with Logging - Mamin River Unpublished manuscript, MacMillan Bloedel Limited, Woodlands Services, Nanaimo, British Columbia, 10 pp. BRAY, D.I. (1972) Generalized Regime-Type Analysis of Alberta Rivers Unpublished Ph.D. thesis, Dept. of C i v i l Engineering, University of Alberta, Edmonton, 232 pp. BRAY, D.I. and M. Church (1980) Armored Versus Paved Gravel Beds Journal of the Hydraulics Division, American Society of C i v i l Engineers, V.106, p.1937-1940. BROWN, G.W. and J. T. Krygier (1971) Clear-Cut Logging and Sediment Production in the Oregon Coast Range Water Resources Research, V.7, p.1189-1198. 153 BRUSH, L.M., Jr. (1961) Drainage Basins, Channels, and Flow Characteristics of Selected Streams in Central Pennsylvania United States Geological Survey, Professional Paper 282-F, p.145-181. BRYANT, M.D. (1980) Evolution of Large, Organic Debris After Timber Harvest: Maybeso Creek, 1949 to 1978 U.S.D.A. Forest Service, General Technical Report PNW-101, 30 pp. BURKHAM, D.E. (1972) Channel Changes of the Gila River in Safford Valley, Arizona, 1846-1970 United States Geological Survey, Professional Paper 655-G, 24 pp. CAINE, N. (1980) The R a i n f a l l Intensity-Duration Control of Shallow Landslides and Debris Flows Geografiska Annaler, V.62A, p.23-27. CARLING, P.A. and N.A. Reader (1982) Structure, Composition and Bulk Properties of Upland Stream Gravels Earth Surface Processes and Landforms, V.7, p.349-365. CARSON, M.A. and M. J. Kirkby (1972) Hillslope Form and Process Cambridge University Press, London, 475 pp. CHAMBERLIN, T.W. (1982) Influence of Forest and Rangeland Management on Anadromous Fish Habitat in Western North America: Timber Harvest U.S.D.A. Forest Service, General Technical Report PNW-136, 30 pp. CHENG, J.D., T.A. Black, J. de Vries, R.P. Willington and B.C. Goodell (1975) The Evaluation of I n i t i a l Changes in Peak Streamflow Following Logging of a Watershed on the West Coast of Canada in IAHS Publication No. 117, Symposium Proceedings, Tokyo, Japan, p.475-486. CHORLEY, R.J. (1962) Geomorphology and General Systems Theory United States Geological Survey, Professional Paper 500-B, 10 pp. CHORLEY, R.J. and B.A. Kennedy (1971) Physical Geography: A Systems Approach Prentice-Hall, London, 370 pp. CHURCH, M. (1980) Records of Recent Geomorphological Events i n Timescales i n Geomorphology, R.A. Cullingford, D.A. Davidson, and J. Lewin (Eds.), J. Wiley and Sons, Chichester, p.13-29. CHURCH, M. (1983) Concepts of Sediment Transfer and Transport in the Queen Charlotte Islands Working Paper 2/83, Fish-Forestry Interaction Program, Vancouver, British Columbia, 23 pp. 154 CHURCH, M. and J. M. Ryder (1972) Paraglacial Sedimentation: A Consideration of Fluvial Processes Conditioned by Glaciation Geological Society of America Bulletin, V.83, p.3059-3071. CLAGUE, J.J. (1981) Late Quaternary Geology and Geochronology of B r i t i s h Columbia, Part 2: Summary and Discussion of Radiocarbon-Dated Quaternary History Geological Survey of Canada, Paper 80-35, 41 pp. CLAGUE, J.J., R.W. Mathewes and B.G. Warner (1982a) Late Quaternary Geology of Eastern Graham Island, Queen Charlotte Islands, British Columbia Canadian Journal of Earth Sciences, V.19, p.1786-1795. CLAGUE, J., J. R. Harper, R. J. Hebda and D. E. Howes (1982b) Late Quaternary Sea Levels and Crustal Movements, Coastal British Columbia Canadian Journal of Earth Sciences, V.19, p.597-618. COOKE, R.U. and R. W. Reeves (1976) Arroyos and Environmental Change in the American South-West Oxford University Press, Oxford, 213 pp. DAVIS, W.M. (1902) Base-Level, Grade, and Peneplain Journal of Geology, V.10, p.77-111. DENNY, CS. and J.C. Goodlett (1956) Microrelief Resulting from Fallen Trees in Surficial Geology and Geomorphology of Potter County, Pennsylvania, CS. Denny (Ed.), United States Geological Survey, Professional Paper 288, p.59-68. DE VRIES, J. and T. L. Chow (1978) Hydrologic Behaviour of a Forested Mountain Soil in Coastal British Columbia Water Resources Research, V.14, p.935-942. DIETRICH, W.E. and T. Dunne (1978) Sediment Budget for a Small Catchment in Mountainous Terrain Z e i t s c h r i f t fur Geomorphologie N.F., Suppl.. Bd. 29, p.191-206. DIETRICH, W.E., T. Dunne, N.F. Humphrey and L.M. Reid (1982) Construction of Sediment Budgets for Drainage Basins in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.5-23. EINSTEIN, H.A. (1968) Deposition of Suspended Particles in a Gravel Bed Journal of the Hydraulics Division, American Society of C i v i l Engineers, V.94, p.1197-1205. 155 ENVIRONMENT AND LAND USE COMMITTEE SECRETARIAT (1976) Terrain Classification System Province of British Columbia, Ministry of the Environment, Resource Analysis Branch, Victoria, 54 pp. EVEREST, F.H. and R. D. Harr (1982) Influence of Forest and Rangeland Management on Anadromous Fish Habitat in Western North America: Silvicultural Treatments U.S.D.A. Forest Service, General Technical Report PNW-134, 19 pp. EVERETT, K.R. (1963) Slope Movement, Neotoraa Valley, Southern Ohio Ohio State University, Institute of Polar Studies, Report No. 6, 16 pp. EYLES, R.J. and R. Ho (1970) Soil Creep on a Humid Tropical Slope Journal of Tropical Geography, V.31, p.40-42. FIKSDAL, A.J. (1974) A Landslide Survey of the Stequaleho Creek Watershed Fisheries Research I n s t i t u t e , College of F i s h e r i e s , U n i v e r s i t y of Washington, Seattle, Supplement to Final Report, FRI-UW-7404, 8 pp. FLADMARK, K.R.(1979) The Early Prehistory of the Queen Charlotte Islands Archaeology, V.32, p.38-45. FOLK, R.L. and W. C. Ward (1957) Brazos River Bar: A Study in the Significance of Grain Size Parameters Journal of Sedimentary Petrology, V.27, p.3-26. FREDRIKSEN, R.L. (1970) Erosion and Sedimentation Following Road Construction and Timber Harvest on Unstable Soils in Three Small Western Oregon Watersheds U.S.D.A. Forest Service, Research Paper PNW-104, 15 pp. GARDNER, J.S. (1983) Observations on Erosion by Wet Snow Avalanches, Mount Rae Area, Alberta, Canada Arctic and Alpine Research, V.15, p.271-274. GILBERT, G.K. (1880) Report on the Geology of the Henry Mountains United States Geographical and Geological Survey of the Rocky Mountain Region, Dept. of the Interior, Washington, Second Edition, 160 pp. GOMEZ, B. (1983) Representative Sampling of Sandy Fluvial Gravels Sedimentary Geology, V.34, p.301-306. GOMEZ, B. (1984) Typology of Segregated (Armoured/Paved) Surfaces: Some Comments Earth Surface Processes and Landforms, V.9, p.19-24. 156 GRAY, D.H. (1970) Effects of Forest Clear-Cutting on the Stability of Natural Slopes Association of Engineering Geologists Bulletin, V.7, p.45-66. GRAY, D.H. and W.F. Megahan (1981) Forest Vegetation Removal and Slope Stability in the Idaho Batholith U.S.D.A. Forest Service, Research Paper INT-271, 23 pp. GRESSWELL, S., D. Heller and D.N. Swanston (1979) Mass Movement Response to Forest Management i n the Central Oregon Coast Ranges U.S.D.A. Forest Service, Resource Bulletin PNW-84, 26 pp. GRIFFITHS, J.C. (1967) Scientific Method in Analysis of Sediments McGraw-Hill, New York, 508 pp. HACK, J.T. (1957) Studies of Longitudinal Stream Profiles in Virginia and Maryland United States Geological Survey, Professional Paper 294-B, p.45-97. HACK, J.T. (1960) Interpretation of Erosional Topography in Humid Temperate Regions American Journal of Science, V.258A, p.80-97. HARDEN, D.R., F. Ugolini and R.J. Janda (1982) Weathering and S o i l P r o f i l e Development as Tools i n Sediment Routing and Budget Studies in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.146-149. HARR, R.D. (1976a) Forest Practices and Streamflow in Western Oregon U.S.D.A. Forest Service, General Technical Report PNW-49, 18 pp. HARR, R.D. (1976b) Hydrology of Small Forest Streams in Western Oregon U.S.D.A. Forest Service, General Technical Report PNW-55, 15 pp. HARR, R.D. (1979) Effects of Timber Harvest on Streamflow in the Rain-Dominated Portion of the Pacific Northwest in Proceedings of the Workshop on Scheduling Timber Harvest for Hydrologic Concerns, U.S.D.A. Forest Service, P a c i f i c Northwest Region, Portland, Oregon, 44 pp. HARR, R.D. (1980) Streamflow After Patch Logging i n Small Drainages Within the B u l l Run Municipal Watershed, Oregon U.S.D.A. Forest Service, Research Paper PNW-268, 16 pp. 157 HARR, R.D. (1983) Potential for Augmenting Water Yield Through Forest Practices in Western Washington and Western Oregon Water Resources Bulletin, V.19, p.383-393. HARR, R.D. and F. M. McCorison (1979) I n i t i a l Effects of Clearcut Logging on Size and Timing of Peak Flows i n a Small Watershed in Western Oregon Water Resources Research, V.15, p.90-94. HARR, R.D., R. L. Fredriksen and J. Rothacher (1979) Changes in Streamflow Following Timber Harvest in Southwestern Oregon U.S.D.A. Forest Service, Research Paper PNW-249, 22 pp. HARR, R.D., A. Levno and R. Mersereau (1982) Streamflow Changes After Logging 130-Year-Old Douglas F i r i n Two Small Watersheds Water Resources Research, V.18, p.637-644. HARR, R.D., W.C. Harper, J.T. Krygier and F.S. Hsieh (1975) Changes i n Storm Hydrographs After Road Building and Clear-Cutting i n the Oregon Coast Range Water Resources Research, V . l l , p.436-444. HARRIS, D.D. (1971) Preliminary Evaluation of Effects of Logging on Hydrologic Characteristics of the Three Principal Watersheds in Forest Land Uses and Stream Environment J.T. Krygier and J.D. Hal l (Eds.), Symposium Proceedings, Oregon State University, C o r v a l l i s , p.244-245. HAYNES, C.V., Jr. (1968) Geochronology of Late-Quaternary Alluvium in Means of Correlation of Quaternary Successions, R.B. Morrison and H.E. Wright,Jr. (Eds.), University of Utah Press, Salt Lake City, p.591-631. HEEDE, B.H. (1972) Flow and Channel Characteristics of Two High Mountain Streams U.S.D.A. Forest Service, Research Paper RM-96, 12 pp. HEEDE, B.H. (1981) Dynamics of Selected Mountain Streams i n the Western United States of America Zeitschrift fur Geomorphologie N.F., Suppl. Bd. 25, p.17-32. HEEDE, B.H. (1983) Sediment Source Areas After Timber Harvest in Mixed Conifer in Proceedings of the Third Annual American Geophysical Union Front Range Branch Hydrology Days, Fort Collins, Colorado, p.139-156. HETHERINGTON, E.D. (1982) Effects of Forest Harvesting on the Hydrologic Regime of Carnation Creek Experimental Watershed: A Preliminary Assessment in Hydrological Processes of Forested Areas, Proceedings of the Canadian Hydrology Symposium '82, Fredericton, New Brunswick, p.247-267. 158 HEWLETT, J.D. (1982) Principles of Forest Hydrology University of Georgia Press, Athens, 183 pp. HEWLETT, J.D. and J.D. Helvey (1970) Effects of Forest Clear-Felling on the Storm Hydrograph Water Resources Research, V.6, p.768-782. HEY, R.D. (1979) Dynamic Process-Response Model of River Channel Development Earth Surface Processes, V.4, p.59-72. HOBLER, P.M. (1978) The Relationship of Archaeological Sites to Sea Levels on Moresby Island, Queen Charlotte Islands Canadian Journal of Archaeology, V.2, p.1-13. HOGAN, D.L. (1984) Stream Channel Morphology: Comparison of Logged and Unlogged Watersheds in the Queen Charlotte Islands Working Paper 16/84, Fish-Forestry Interaction Program, Vancouver, British Columbia (in press). HORNBECK, J.W. and K.G. Reinhart (1964) Water Quality and Soil Erosion as Affected by Logging in Steep Terrain Journal of Soil and Water Conservation, V.19, p.23-27. HOWARD, A.D. (1965) Geomorphological Systems - Equilibrium and Dynamics American Journal of Science, V.263, p.302-312. KARANKA, E.J. (1984) Pr e c i p i t a t i o n and Stream Runoff Trend Analysis for the Queen Charlotte Islands Fi n a l Report, Fish-Forestry Interaction Program, Vancouver, B r i t i s h Columbia, 34 pp. (in press). KELLER, E.A. and F.J. Swanson (1979) Effects of Large Organic Material on Channel Form and Fluvial Processes Earth Surface Processes, V.4, p.361-380. KELLER, E.A. and T. Tally (1979) Effects of Large Organic Debris on Channel Form and Fluvial Processes in the Coastal Redwood Environment jLn Adjustments of the F l u v i a l System, D.D. Rhodes and G.P. Williams (Eds.), Kendall-Hunt, Dubuque, Iowa, p.169-197. KELLERHALS, R. and D.I. Bray (1971) Sampling Procedure for Coarse Fluvial Sediments Journal of the Hydraulics Division, American Society of C i v i l Engineers, V.97, p.1165-1180. 159 KELSEY, H.M. (1978) Earthflows in the Franciscan Melange, California Geology, V.6, p.361-365. KELSEY, H.M. (1980) A Sediment Budget and an Analysis of Geomorphic Process i n the Van Duzen River Basin, North Coastal California, 1941-1975 Geological Society of American Bulletin, Part I I , V.91, p.1119-1216. KELSEY, H.M. (1982) Influence of Magnitude, Frequency, and Persistence of Various Types of Disturbance on Geomorphic Form and Process in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-41, p.150-153. KIRKBY, M.J. (1967) Measurement and Theory of Soil Creep Journal of Geology, V.75, P.359-378. KOTARBA, A. (1970) The Morphogenic Role of Foehn Wind in the Mts. Studia Geomorph. Carpatho-Balcanica, V.4, p.171-186. KRAG, R.K. (1980) A Method to Estimate Risk of Soil Erosion to Logging Sites in the Kootenay Area of British Columbia Forest Engineering Research Institute of Canada, Technical Report No. TR-38, 55 pp. KRAJINA, V.J., K. Klinka and J. Worrall (1982) Distribution and Ecological Characteristics of Trees and Shrubs of British Columbia Faculty of Forestry, University of British Columbia, Vancouver, 131 pp. LEHRE, A.K. (1981) Sediment Budget of a Small California Coast Range Drainage Basin Near San Francisco in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.123-139. LEHRE, A.K. (1982) Sediment Budget of a Small Coast Range Drainage Basin i n North-Central California in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.67-77. LEOPOLD, L.B. and W. B. B u l l (1979) Base Level, Aggradation, and Grade Proceedings of the American Philosophical Society, V.123, p.168-202. LEWIS, L.A. (1976) Soil Movement in the Tropics - A General Model Zeitschrift fur Geomorphologie N.F., Suppl. Bd. 25, p.132-144. 160 LISLE, T.E. (1981) Recovery of Aggraded Stream Channels at Gauging Stations i n Northern California and Southern Oregon in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.189-211. LISLE, T.E. (1982) Effects of Aggradation and Degradation on Riffle-Pool Morphology in Natural Gravel Channels, Northwestern California Water Resources Research, V.18, p.1643-1651. LYONS, J.K. and R. L. Beschta (1983) Land Use, Floods, and Channel Changes: Upper Middle Fork Willamette River, Oregon (1936-1980) Water Resources Research, V.19, p.463-471. MACKIN, J.H. (1948) Concept of the Graded River Geological Society of America Bulletin, V.59, p.463-512. MADEJ, M.A. (1978) Response of a Stream Channel to an Increase in Sediment Load Unpublished M.Sc. thesis, Dept. of Geological Sciences, University of -Washington, Seattle, 111 pp. MADEJ, M.A. (1982) Sediment Transport and Channel Changes in an Aggrading Stream in the Puget Lowland, Washington in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.97-108. MARSTON, R.A. (1982) The Geomorphic Significance of Log Steps in Forest Streams Annals of the Association of American Geographers, V.72, p.99-108. MATHEWES, R.W. and J.J. Clague (1982) Stratigraphic Relationships and Paleoecology of a Late-Glacial Peat Bed from the Queen Charlotte Islands, British Columbia Canadian Journal of Earth Sciences, V.19, p.1185-1195. McLAREN, P. (1981) An Interpretation of Trends in Grain Size Measures Journal of Sedimentary Petrology, V.51, p.611-624. MEGAHAN, W.F. (1982) Channel Sediment Storage Behind Obstructions i n Forested Drainage Basins Draining the Granitic Bedrock of the Idaho Batholith in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.114-121. 161 MEGAHAN, W.F. (1983) Hydrologic Effects of Clearcutting and Wildfire on Steep Granitic Slopes in Idaho Water Resources Research, v.19, p.811-819. MEGAHAN, W.F. and W.J. Kidd (1972a) Effect of Logging Roads on Sediment Production Rates in the Idaho Batholith U.S.D.A. Forest Service, Research Paper INT-123, 14 pp. MEGAHAN, W.F. and W.J. Kidd (1972b) Effects of Logging and Logging Roads on Erosion and Sediment Deposition from Steep Terrain Journal of Forestry, V.70, p.136-141. MEGAHAN, W.F. and R. A. Nowlin (1976) Sediment Storage i n Channels Draining Small Forested Watersheds i n the Mountains of Central Idaho in Proceedings of the Third Federal Inter-Agency Sedimentation Conference, Denver, Colorado, p.4.115-4.126. MERSEREAU, R.C. and CT. Dyrness (1972) Accelerated Mass Wasting After Logging and Slash Burning in Western Oregon Journal of Soil and Water Conservation, V.27, p.112-114. MILES, M.J. and R. Kellerhals (1981) Some Engineering Aspects of Debris Torrents i n Proceedings of the F i f t h Canadian Hydrotechnical Conference, Canadian Society for C i v i l Engineering, Fredericton, New Brunswick, p.395-420. MILHOUS, R.T. (1982) Effect of Sediment Transport and Flow Regulation on the Ecology of Gravel-Bed Rivers jLn Gravel-Bed Rivers: ^ F luvial Processes, Engineering and Management, R.D. Hey, J.C. Bathurst and CR. Thorne (Eds.), J. Wiley and Sons, Chichester, p.819-842. MILLER, J.P. (1958) High Mountain Streams: Effects of Geology on Channel Characteristics and Bed Material New Mexico State Bureau of Mines and Mineral Resources, Memoir 4, 53 pp. MILNE, W.G., G.C. Rogers, R.P Riddihough, G.A. McMechan and R.D. Hyndman (1978) Seismicity of Western Canada Canadian Journal of Earth Sciences, V.15, p.1170-1193. MOFFITT, F.H. and E.M. Mikhail (1980) Photogrammetry, Third Edition Harper and Row, New York, 648 pp. MOORE, M.K. (1980) Streamside Management: A Decision-Making Procedure for South Coastal British Columbia Province of British Columbia, Ministry of Forests, Land Management Handbook No. 1, 88 pp. 162 MORRISON, P.H. (1975) Ecological and Geomorphological Consequences of Mass Movements in the Alder Creek Watershed and Implications for Forest Land Management Unpublished B.A. thesis, University of Oregon, Eugene, 102 pp. MOSLEY, M.P. (1981) The Influence of Organic Debris on Channel Morphology and Bedload Transport in a New Zealand Forest Stream Earth Surface Processes and Landforms, V.6, p.571-579. NARVER, D.W. (1972) A Survey of Some Possible Effects of Logging on Two Eastern Vancouver Island Streams Fisheries Research Board of Canada, Technical Report No. 323, P a c i f i c Biological Station, Nanaimo, British Columbia, 55 pp. NEWSON, M. (1980) The Geomorphological Effectiveness of Floods - A Contribution Stimulated by Two Recent Events in Mid-Wales Earth Surface Processes, V.5, p.1-16. O'LOUGHLIN, C.L. (1969) Streambed Investigations in a Small Mountain Catchment New Zealand Journal of Geology and Geophysics, V.12, p.684-706. O'LOUGHLIN, C.L. (1972) An Investigation of the S t a b i l i t y of Steepland Forest S o i l s i n the Coast Mountains of Southwest British Columbia Unpublished Ph.D. thesis, Faculty of Forestry, University of B r i t i s h Columbia, Vancouver, 147 pp. O'LOUGHLIN, C.L. (1974) A Study of Tree Root Strength Deterioration Following Clearfelling Canadian Journal of Forest Research, V.4, p.107-113. PARKER, G. and P.C. Klingeman (1982) On Why Gravel Bed Streams Are Paved Water Resources Research, V.18, p.1409-1423. PATTON, P.C. and S.A. Schumm (1975) Gully Erosion, Northwestern Colorado: A Threshold Phenomenon Geology, V.3, p.88-90. PATTON, P.C. and S.A. Schumm (1981) Ephemeral-Stream Processes: Implications for Studies of Quaternary Valley F i l l s Quaternary Research, V.15, p.24-43. PITLICK, J. (1982) Sediment Routing in Tributaries of the Redwood Creek Basin: Northwestern California Redwood National Park, Research and Development, Technical Report No. 8, Areata, California, 26 pp. 163 POULIN, V.A. (1983) A Research Approach to Solving Fish-Forestry Interactions i n Relation to Mass Wasting on the Queen Charlotte Islands Working Paper 1/83, Fish-Forestry Interaction Program, Vancouver, British Columbia, 20 pp. REID, L.M. (1981) Sediment Production From Gravel-Surfaced Forest Roads, Clearwater Basin, Washington F i s h e r i e s Research I n s t i t u t e , College of Fisheries, University of Washington, Seattle, Final Report FRI-UW-8108, 247 pp. REID, L.M., T. Dunne and C.J. Cederholm (1981) Application of Sediment Budget Studies to the Evaluation of Logging Road Impact Journal of Hydrology (New Zealand), V.20, p.49-62. RICE, R.M. (1981) A Perspective on the Cumulative Effects of Logging on Streamflow and Sedimentation i n Cumulative Effects of Forest Management on California Watersheds: An Assessment of Status and Need for Information, Proceedings of the Edgebrook Conference, R.B. Standiford and S.I. Ramacher (Eds.), Berkeley, California, p.36-46. RICE, R.M. (1982) Sedimentation in the Chaparral: How Do You Handle Unusual Events? in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141 p.39-49. RICE, R.M. and P.A. Datzman (1981) Erosion Associated with Cable and Tractor Logging i n North-Western California in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.362-374. RICE, R.M. and G.T. Foggin I I I (1971) Effect of High Intensity Storms on Soil Slippage on Mountainous Watersheds in Southern California Water Resources Research, V.7, p.1485-1496. RICE, R.M., F.B. T i l l e y and P.A. Datzman (1979) A Watershed's Response to Logging and Roads: South Fork of Caspar Creek, California, 1967-1976 U.S.D.A. Forest Service, Research Paper PSW-146, 12 pp. ROBERTS, R.G. (1984) Major F l u v i a l Disturbances i n Logged Watersheds i n the Queen Charlotte Islands, Bri t i s h Columbia Working Paper 17/84, Fish-Forestry Interaction Program, Vancouver, British Columbia (in press). 164 ROOD, K.M. (1983) Operating Manual for the Wild A6 Unpublished manuscript, Dept. of Geography, University of British Columbia, Vancouver, 46 pp. ROOD, K.M. (1984) Episodic Mass Wasting in the Queen Charlotte Islands, British Columbia Working Paper 4/83, Fish-Forestry Interaction Program, Vancouver, British Columbia, 38 pp. (in press). ROTHACHER, J. (1970) Increases i n Water Yield Following Clear-Cut Logging i n the P a c i f i c Northwest Water Resources Research, V.6, p.653-658. ROTHACHER, J. (1971) Regimes of Streamflow and Their Modification by Logging i n Forest Land Uses and Stream Environment, J.T. Krygier and J.D. Hall (Eds.), Symposium Proceedings, Oregon State University, Corvallis, p.40-54. ROTHACHER, J. (1973) Does Harvest i n West Slope Douglas-Fir Increase Peak Flow i n Small Forest Streams? U.S.D.A. Forest Service, Research Paper, PNW-163, 13 pp. RYDER, J.M. (1971a) The Stratigraphy and Morphology of Para-Glacial A l l u v i a l Fans i n South-Central British Columbia Canadian Journal of Earth Sciences, V.8, p.279-298. RYDER, J.M. (1971b) Some Aspects of the Morphometry of Paraglacial A l l u v i a l Fans in South-Central British Columbia Canadian Journal of Earth Sciences, V.8, p.1252-1264. SAUNDERS, I. and A. Young (1983) Rates of Surface Processes on Slopes, Slope Retreat, and Denudation Earth Surface Processes and Landforms, V.8, p.473-501. SCHUMM, S.A. (1973) Geomorphic Thresholds and Complex Response of Drainage Systems in Fluvial Geomorphology, M. Morisawa (Ed.), Publications in Geomorphology, State University of New York, Binghamton, p.299-310. SCHUMM, S.A. (1976) Episodic Erosion: A Modification of the Geomorphic Cycle i_n Theories of Landform Development, W.N. Melhorn and R.C. Flemal (Eds.), Publications i n Geomorphology, State University of New York, Binghamton, p.69-85. SCHUMM, S.A. (1977) The Fluvial System J. Wiley and Sons, New York, 338 pp. 165 SCHUMM, S.A. (1979) Geomorphic Thresholds: The Concept and i t s Applications Transactions of the Institute of British Geographers, V.4, p.485-515. SCHUMM, S.A. (1981) Evolution and Response of the Fluvial System, Sedimentologic Implications Society of Economic Paleontologists and Mineralogists, Special Publication No. 31, p.19-29. SCHUMM, S.A. and R.W. Lichty (1963) Channel Widening and Flood-Plain Construction Along Cimarron River in South-western Kansas United States Geological Survey, Professional Paper 352-D, p.71-88. SCHUMM, S.A. and R.W. Lichty (1965) Time, Space, and Causality in Geomorphology American Journal of Science, V.263, p.110-119. SCHUMM, S.A. and Parker, R. (1973) Implications of Complex Response of Drainage Systems for Quaternary Alluvial Stratigraphy Nature (London), Physical Science, V.243, p.99-100. SCHWAB, J.W. (1983) Mass Wasting: October-November 1978 Storm, Rennell Sound, Queen Charlotte Islands, British Columbia Province of British Columbia, Ministry of Forests, Research Note No. 91, 23 pp. SINGLETON, G.A. (1978) Weathering in a Soil Chronosequence Unpublished Ph.D. thesis, Dept. of Soil Science, University of British Columbia, Vancouver, 206 pp. SMITH, R.B., P.R. Commandeur and M.W. Ryan (1983) Natural Revegetation, S o i l Development and Forest Growth on the Queen Charlotte Islands - Progress Report. Working Paper 7/83, Fish-Forestry Interaction Program, Vancouver, British Columbia, 46 pp. SPANGLER, M.G. and R.L. Handy (1982) Soil Engineering, Fourth Edition Harper and Row, New York, 819 pp. STATHAM, I. (1977) Earth Surface Sediment Transport Oxford University Press, Oxford, 184 pp. STRAHLER, A.N. (1952a) Dynamic Basis of Geomorphology Geological Society of America Bulletin, V.63, p.923-938. STRAHLER, A.N. (1952b) Hypsometric (Area-Altitude) Analysis of Erosional Topography Geological Society of America Bulletin, V.63, p.1117-1142. 166 SUTHERLAND BROWN, A. (1968) Geology of the Queen Charlotte Islands, British Columbia Province of B r i t i s h Columbia, Dept. of Mines and Petroleum Resources, B u l l e t i n No. 54, 226 pp. SWANSON, F.J. and CT. Dyrness (1975) Impact of Clear-Cutting and Road Construction on Soil Erosion by Landslides in the Western Cascade Range, Oregon Geology, V.3, p.393-396. SWANSON, F.J. and R.L. Fredriksen (1982) Sediment Routing and Budgets: Implications for Judging Impacts of Forestry Practices in Sediment Budgets and Routing in Forested Drainage Basins, F.J. Swanson, R.J. Janda, T. Dunne and D.N. Swanston (Eds.), U.S.D.A. Forest Service, General Technical Report PNW-141, p.129-137. SWANSON, F.J. and G.W. Lienkaemper (1978) Physical Consequences of Large Organic Debris in Pacific Northwest Streams U.S.D.A. Forest Service, General Technical Report PNW-69, 12 pp. SWANSON, F.J., R.L. Fredriksen and F.M. McCorison (1982a) Material Transfer in a Western Oregon Forested Watershed in Analysis of Coniferous Forest Ecosystems in the Western United States, R.L. Edmonds (Ed.), Hutchinson Ross, Stroudsburg, Pennsylvania, p.233-266. SWANSON, F.J., G.W. Lienkaemper and J.R. Sedell (1976) History, Physical Effects, and Management Implications of Large Organic Debris in Western Oregon Streams U.S.D.A. Forest Service, General Technical Report PNW—56, 15 pp. SWANSON, F.J., M.M. Swanson and C. Woods (1981) Analysis of Debris-Avalanche Erosion in Steep Forest Lands: An Example from Mapleton, Oregon, USA in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.67-75. SWANSON, F.J., S.V. Gregory, J.R. Sedell and A.G. Campbell (1982b) Land-Water Interactions: The Riparian Zone in Analysis of Coniferous Forest Ecosystems in the Western United States, R.L. Edmonds (Ed.), Hutchinson Ross, Stroudsburg, Pennsylvania, p.267-291. SWANSON, F.J., R.J. Janda, T. Dunne and D.N. Swanston (Eds.) (1982c) Sediment Budgets and Routing in Forested Drainage Basins U.S.D.A. Forest Service, General Technical Report PNW-141, 165 pp. SWANSTON, D.N. (1967) Debris Avalanching in Thin Soils Derived from Bedrock U.S.D.A. Forest Service, Research Note PNW-64, 7 pp. SWANSTON, D.N. (1969) Mass Wasting in Coastal Alaska U.S.D.A. Forest Service, Research Paper PNW-83, 15 pp. 167 SWANSTON, D.N. (1970) Mechanics of Debris Avalanching in Shallow T i l l Soils of Southeast Alaska U.S.D.A. Forest Service, Research Paper PNW-103, 17 pp. SWANSTON, D.N. (1971) Principal Mass Movement Processes Influenced by Logging, Road Building, and Fire i n Forest Land Uses and Stream Environment, J.T. Krygier and J.D. H a l l (Eds.), Symposium Proceedings, Oregon State University, Corvallis, p.29-39. SWANSTON, D.N. (1974a) The Forest Ecosystem of Southeast Alaska. 5. Soil Mass Movement U.S.D.A. Forest Service, General Technical Report PNW-17, 22 pp. SWANSTON, D.N. (1974b) Slope Stability Problems Associated with Timber Harvesting in Mountainous Regions of the Western United States U.S.D.A. Forest Service, General Technical Report PNW-21, 14 pp. SWANSTON, D.N. (1980) Influence of Forest and Rangeland Management on Anadromous Fish Habitat in Western North America: Impacts of Natural Events U.S.D.A. Forest Service, General Technical Report PNW-104, 27 pp. SWANSTON, D.N. (1981) Creep and Earthflow Erosion From Undisturbed and Management Impacted Slopes in the Coast and Cascade Ranges of the Pacific Northwest, U.S.A. in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.76-94. SWANSTON, D.N. and F.J. Swanson (1976) Timber Harvesting, Mass Erosion and Steepland Forest Geomorphology in the Pacific Northwest in Geomorphology and Engineering, D.R. Coates (Ed.), Dowden, Hutchinson and Ross, Stroudsburg, Pennsylvania, p.199-221. THORNE, C.R. (1982) Processes and Mechanisms of River Bank Erosion in Gravel-Bed Rivers: Fluvial Processes, Engineering and Management, R.D. Hey,J.C. Bathurst and C.R. Thorne (Eds.), J. Wiley and Sons, Chichester, p.227-271. TOEWS, D.A.A. and M.K. Moore (1982a) The Effects of Streamside Logging on Large Organic Debris in Carnation Creek Province of British Columbia, Ministry of Forests, Land Management Report No. 11, 29 pp. TOEWS, D.A.A. and M.K. Moore (1982b) The Effects of Three Streamside Logging Treatments on Organic Debris and Channel Morphology of Carnation Creek in Proceedings of the Carnation Creek Workshop: A Ten-Year Review, G.F. Hartman (Ed.), Nanaimo, British Columbia, p.129-153. 168 UGOLINI, F.C. (1968) S o i l Development and Alder Invasion in a Recently Deglaciated Area of Glacier Bay, Alaska in Biology of Alder, J.M. Trappe, J.F. Franklin, R.F. Tarrant and G.M. Hansen (Eds.), Proceedings of the Northwest Scientific Association Fortieth Annual Meeting, Pullman, Washington, p.115-140. VALENTINE, K.W.G., P.N. Sprout, T.E. Baker and L.M. Lavkulich (1978) The Soil Landscapes of British Columbia Province of British Columbia, Ministry of the Environment, Resource Analysis Branch, Victoria, 197 pp. VARNES, D.J. (1978) Slope Movement Types and Processes i i i Landslides: Analysis and Control, R.L. Schuster and R.J. Krizek (Eds.), Transportation Research Board, Washington, D.C, Special Report No. 176, p.11-33. WILFORD, D.J. and J.W. Schwab (1982) S o i l Mass Movements i n the Rennell Sound Area, Queen Charlotte Islands, British Columbia in Hydrological Processes of Forested Areas, Proceedings of the Canadian Hydrology Symposium '82, Fredericton, New Brunswick, p.521-541. WOLMAN, M.G. (1954) A Method of Sampling Coarse River-Bed Material Transactions of the American Geophysical Union, V.35, p.951-956. WOLMAN, M.G. and R. Gerson (1978) Relative Scales of Time and Effectiveness of Climate i n Watershed Geomorphology Earth Surface Processes, V.3, p.189-208. WOLMAN, M.G. and J.P. M i l l e r (1960) Magnitude and Frequency of Forces in Geomorphic Processes Journal of Geology, V.68, p.54-74. WOMACK, W.R. and S.A. Schumm (1977) Terraces of Douglas Creek, Northwestern Colorado: An Example of Episodic Erosion Geology, V.5, p.72-76. WU, T.H. and D.N. Swanston (1980) Risk of Landslides i n Shallow S o i l s and i t s Relation to Clearcutting i n Southeastern Alaska Forest Science, V.26, p.495-510. WU, T.H., W.P. McKinnell I I I and D.N. Swanston (1979) Strength of Tree Roots and Landslides on Prince of Wales Island, Alaska Canadian Geotechnical Journal, V.16, p.19-33. YOUNG, A. (1960) Soil Movement by Denudational Processes on Slopes Nature (London), V.188, p.120-122. 169 YOUNG, A. (1963) Soil Movement on Slopes Nature (London), V.200, p.129-130. YOUNG, A. (1974) The Rate of Slope Retreat in_ Progress i n Geomorphology, E.H. Brown and R.S. Waters (Eds.), I n s t i t u t e of British Geographers, Special Publication No. 7, p.65-78. ZIEMER, R.R. (1981a) Storm Flow Response to Road Building and Partial Cutting in Small Streams of Northern California Water Resources Research, V. 17, p.907-917. ZIEMER, R.R. (1981b) Roots and the Stability of Forested Slopes in Erosion and Sediment Transport in Pacific Rim Steeplands, T.R.H. Davies and A.J. Pearce (Eds.), IAHS Publication No. 132, p.343-361. ZIEMER, R.R. and D.N. Swanston (1977) Root Strength Changes After Logging in Southeast Alaska U.S.D.A. Forest Service, Research Note PNW-306, 10 pp. 170 APPENDIX A REVIEW OF SEDIMENT SOURCES AND MOBILIZATION PROCESSES 171 This appendix reviews b r i e f l y the major sediment sources and mobilization processes investigated in other studies, primarily those in the Pacific Northwest, to help determine which processes should be investigated i n t h i s study; a more comprehensive l i t e r a t u r e review can be found in Roberts (1984). The processes are discussed i n d i v i d u a l l y , due to the diversity of the study methods and their results, but the interactions among sediment transfer processes in steep, headwater basins are illustrated in Figure A.l, and summary rates of sediment delivery are compared in Table 3.1. A.l Soil creep S o i l creep i s the term used to describe the group of processes that result i n the slow downslope translation of s o i l , without the development of discrete failure planes (Swanson et al., 1982a;. Dietrich et a l . , 1982). These processes include the continuous, gravitational shear stress on the s o i l mass; the displacement of s o i l by expansion and contraction from wetting and drying, freezing and thawing, or weathering; the transfer of wind stress to the s o i l mantle via standing trees; and the spalling and raveling of surface debris (Dietrich et al., 1982). The steady movement in response to gravitational stress alone i s called "continuous" creep, but when combined with episodic s o i l movement due to variations in s o i l moisture, temperature and ice content, larger total displacements occur, termed "seasonal" creep (Carson and Kirkby, 1972, p.275). Most studies of s o i l creep have investigated "seasonal" creep (Carson and Kirkby 1972, p.286; Anderson and Cox, 1978) because the compara-t i v e l y r a pid rate of s o i l movement i s more e a s i l y measured. Therefore,"continuous" creep i s not considered per se in the following discussion. 172 FIGURE A.l: Sediment transfers in Pacific Northwest watersheds (after Reid, 1981; Church, 1983) Bedrock In-situ Organic matter production Weathering Mixing <£• Drift Soil in-situ Transfer Slope TRANSFER PROCESSES 1. rockfalls 2. mass wastage processes (debris slides, avalanches, flows, torrents) 3. earthflows 4. s o i l creep 5. surface erosion processes (slope wash, dry ravel, tree throw) 6. stream bank erosion and collapse Slope - slope transfer Slope - stream transfer «=• Streambank - stream transfer Stream Sediment transport in channel (debris flow, bed load, suspended load) Channel bed erosion 173 In a watershed with an undisturbed vegetation cover, the rate of s o i l creep i s influenced primarily by the s o i l moisture content (Carson and Kirkby, 1972, p.275). On theoretical grounds, the creep rate would also be expected to increase l i n e a r l y with the sine of the slope angle, but no measurements of s o i l creep have p o s i t i v e l y demonstrated t h i s r e l a t i o n (Carson and Kirkby, 1972, p.289; Young, 1974; Statham, 1977, p.112); evidently, other c o n t r o l l i n g factors on individual slopes (Statham, 1977, p.112), and the strong relationship with s o i l moisture (Saunders and Young, 1983), mask this simple geometric dependence. Increased rates of creep due to high s o i l moisture conditions usually are confined to the upper 30 cm (Carson and Kirkby, 1972, p.275; Wu et al., 1979) to 50 cm (Barr and Swanston, 1970), with appreciable movement extending t y p i c a l l y to 20 cm depth (Young, 1974; Statham, 1977, p.112; Saunders and Young, 1983). In general, measured velocity p r o f i l e s on creeping soils show a rapid decay in movement with depth (Carson and Kirkby, 1972, p.295; Statham, 1977, p . I l l ) , but, i n d e t a i l , are complex and do not agree well with theory (Statham, 1977, p.111). In undisturbed catchments with humid temperate climates and slope gradients of 0.2 to 0.6, and where the depth to bedrock or to a strongly indurated layer i s less than one metre, such as i n the Queen Charlotte Islands (Smith et a l . , 1983), surface movement l i e s t y p i c a l l y within the narrow range, 1-3 mm/yr (Young, 1974; Anderson and Cox, 1978; Saunders and Young, 1983; see Table A.1). Creep movement can be stated a l t e r n a t i v e l y i n volumetric terms, as cra^/cm width of slope/year. Con-version between these units depends upon the depth and velocity profile of movement, but where both are reported, they usually are within a numerical range of x 0.5 to x 3 (Saunders and Young, 1983); volumetric rates of 174 TABLE A.l; Creep rates in undisturbed basins Climate Movement of Surface or Upper 5 cm (mm/yr) Volumetric Movement (cm /cm/yr) Location Source Temperate maritime 1-2 0.5 Northern England Young (1960, 1963)a Temperate maritime 1-2 2.1 Scotland Kirkby (1967)a Temperate maritime 1.3 Northern England Anderson & Cox (1978) Temperate continental 1 6.0 Ohio, U.S.A. Everett (1963)a Mediterranean 1.1 4.4 California, U.S.A. Lehre (1982) Rainforest 5 12.4 Malaya Eyles & Ho (1970)b Rainforest 4 7-9 Puerto Rico Lewis (1976)b aStudies used by Madej (1978, 1982) to characterize rates of creep in western Washington bStudies used by Dietrich and Dunne (1978) to characterize rates of creep in the Oregon Coast Range TABLE A.2: Creep rates in disturbed basins Depth of Movement of Surface Location Movement or Upper 5 cm (mm/yr) Source Pre- Post-(m) Disturbance Disturbance Southwest Oregon 3.2 - 3.7 2.8 5.7 Swanston (1981) 3.8 9.3 1.5 6.3 Southeast Alaska 0.15-0.46 - 6.4 Barr & Swanston (1970) 176 sediment movement by creep are of the order of 1-10 cm /cm/yr in temperate latitudes (Young, 1974; Saunders and Young, 1983; see Table A.1). Due to the importance of woody vegetation in maintaining slope s t a b i l i t y , clearcutting increases the creep rate on slopes (Barr and Swanston, 1970; Gray, 1970; S\«mson and Dyrness, 1975; Wu et a l . , 1979; Swanston, 1981), with available data indicating surface movement of the order of 6-7 mm/yr (Barr and Swanston, 1970; Swanston, 1981; see Table A.2). A.2 Tree throw Tree throw i s defined as the downslope movement of organic and inorganic material by li v i n g and recently fallen trees (Swanson et a l . , 1982a). The importance of tree throw i s demonstrated by the presence of exposed rood wads, s o i l mounds and pits in stands of large, old-growth forest on steep slopes in the Pacific Northwest (Reid, 1981; Swanson et a l . , 1982a; Dietrich et a l . , 1982). The uprooting of trees i s an episodic process that occurs during wind storms with return periods of several years or decades (Swanson et a l . , 1982a). In particular, the Queen Charlotte Islands are character-ized by frequent, strong winds, whose potential for destruction i s greater than at most places in Canada (Alley and Thomson, 1978). Due to the sporadic nature of tree throw, estimates of the annual rate of sediment transfer require long-term records (Swanson et a l . , 1982a). Fortunately, dendrochronologic observations and the stage of decay of root wads, s o i l mounds and pits of uprooted trees, and their subsequent revegetation, provide a record of events for the past 80 years (Reid, 1981) to 150 years (Swanson et a l . , 1982a). However, only those trees uprooted adjacent to a stream, or on steep valley slopes above a stream, contribute to the transfer of sediment from the hillsides into the stream channel 177 (Reid, 1981). The volume of sediment delivered to the stream i s estimated from rootwad dimensions: Swanson et a l . (1982a) assumed, and Reid (1981) calculated from 55 treefalls, that each tree throw event that impinged on the stream channel delivered 2 m and 4 m of s o i l , respectively. From o these data, sediment delivery rates of 0.1 m /channel Km/yr in western Oregon and 1 m /channel Km/yr in western Washington were found by Swanson et a l . (1982a) and Reid (1981), respectively. It has recently been proposed that tree throw in forested catchments i s as important as s o i l creep in terms of the rate of sediment transport (Dietrich et a l . , 1982; see Table A.3). However, in logged water-sheds, tree throw i s eliminated as a s i g n i f i c a n t process for several decades, u n t i l trees are again large enough to be subject to blowdown (Swanson et a l . , 1982a). A.3 Landslides Soil mass movements are a dominant geomorphic process in much of the Pacific Northwest and have been investigated extensively in Alaska (Swanston, 1967, 1969, 1970, 1971, 1974a, 1974b; Wu et a l . , 1979; Wu and Swanston, 1980), California (Rice and Foggin, 1971; Rice et a l . , 1979; Kelsey, 1978, 1980; Rice and Datzman, 1981; Lehre, 1981, 1982; Rice, 1982; P i t l i c k , 1982), Idaho (Megahan and Kidd, 1972a, 1972b; Gray and Megahan, 1981), Oregon (Fredriksen, 1970; Gray, 1970; Anderson, 1971; Brown and Krygier, 1971; Mersereau and Dyrness, 1972; Morrison, 1975; Swanson and Dyrness, 1975; Beschta, 1978; Dietrich and Dunne, 1978; Gresswell et a l . , 1979; Swanson et a l . , 1981; Swanston, 1981; Swanson et a l . , 1982a; Swanson and Fredriksen, 1982; Lyons and Beschta, 1983), Washington (Fiksdal, 1974; 178 Madej, 1978, 1982; Reid, 1981; Reid et a l . , 1981), and British Columbia (O'Loughlin, 1972; Alley and Thomson, 1978; Wilford and Schwab, 1982; Schwab, 1983). Mass movements can be grouped into four classes, based on the mechanics of failure, the geometry of the sliding surface, and the movement process: creep, slides, flows and f a l l s (Swanston, 1974a); s o i l creep has been discussed in Section A.1. Slides and flows are classified usually on the basis of material type, moisture content, and rate of movement into five groups (Varnes, 1978; Wilford and Schwab, 1982; see Table A.4); in this thesis, these five groups of processes are collectively terms "landslides". Slump-earthflows develop on deep s o i l s , beginning as rotational failures and slowly moving downslope through a combination of slumping and flowage (Swanston, 1971). They are a dominant progressive failure type on engineered slopes (Spangler and Handy, 1982) and are a major slope erosion process on deep residual soils in the Coast and Cascade Ranges of northern California, Oregon, and Washington (Swanston, 1981), and in the Queen Charlotte Islands (Wilford and Schwab, 1982). The rate and timing of movement are highly variable depending on local s o i l properties, clay and water content, and the degree and depth of parent material weathering (Swanston, 1981). For example, Swanston and Swanson (1976) found movement rates between 25 and 250 mm/yr in the western Oregon Cascades, while Kelsey (1978) reported rates of movement up to 27 metres/yr in northern California. Debris slides and avalanches are collectively called debris slides i n t h i s thesis. They are shallow f a i l u r e s involving the rapid downslope movement, by sliding and rollin g , of near-saturated s o i l , forest debris and weathered bedrock (Swanston, 1971; Wilford and Schwab, 1982). A debris flow involves the rapid mass movement of water-saturated material by true flow processes (Swanston, 1971; Wilford and Schwab, 1982), while 179 TABLE A.3: Calculated rates of sediment transport by tree throw in forested watersheds Location Rate of Movement Source (mm/yr) Central Appalachians, Penn. 1.5 Denny & Goodlett (1956)* Tatra Mountains, Poland 2 Kotarba (1970)* Olympic Mountains, Wash. 1.8 Reid (1981)a Puget Lowland, Wash. 1.0 Madej (1982)b Cascade Mountains, Ore. 0.04 Swanson et a l . (1982a) aResults quoted in Dietrich et a l . (1982) DRate estimated from results quoted in Dietrich et a l . (1982) TABLE A.4: Classes of mass movement Process Soil Moisture Content Rate of Movement Soil Creep Increasing Increasing Slump-Earthflow Debris Slide Debris Avalanche Debris Flow Debris Torrent 180 channel-scouring debris torrents are generated when debris sl i d e s within steep gullies are coupled with concentrated storm runoff and the failure of log jams (Swanston, 1969, 1971, 1974a, 1980; Gresswell et a l . , 1979; Miles and Kellerhals, 1981; Wilford and Schwab, 1982). Reported rates of debris s l i d e erosion i n forested areas of the P a c i f i c Northwest are summarized in Table A.5, and range from 9 to 72 m3/Km2/yr. Studies in California have reported rates of erosion up to 20 times greater than those elsewhere (e.g. Rice and Foggin, 1971; Kelsey, 1980; Lehre, 1982; P i t l i c k , 1982), due largely to the interaction between readily erodible bedrock and high intensity storms. However, most attention has focussed on debris slide erosion during and following timber harvesting, and i t s impact on stream sediment-ation. It has been widely observed that debris slide erosion from roads is greater than that from clearcuts and that both yield greater erosion rates than forested areas (Fredriksen, 1970; Anderson, 1971; Brown and Krygier, 1971; Swanston, 1971, O'Loughlin, 1972; Megahan and Kidd, 1972b; Morrison, 1975; Swanson and Dyrness, 1975; Swanston and Swanson, 1976; Beschta, 1978; Swanson et al., 1981; Lyons and Beschta, 1983; Schwab, 1983; see Table A.6). The high erosion rates i n clearcut areas reported by Lyons and Beschta (1983) were associated also with a major storm event while those i n New Zealand r e f l e c t , i n part, the e r o d i b i l i t y of the highly dissected te r r a i n (Swanson et a l . , 1981). In general, landslide hazard on slopes increases i n direct proportion to the amount of vegetation removed, and i s greatest 3-6 years after logging, corresponding to a significant loss in root strength at this stage (O'Loughlin, 1974; Swanston, 1974a; Ziemer and Swanston, 1977; Gray and Megahan 1981; Ziemer, 1981b; Schwab, 1983). Slides along road rights-of-way are related to the interruption of surface and sub-surface drainage 181 TABLE A.5: Debris slide erosion in forested areas of the Pacific Northwest Location Soil Eros (m3/Km' jion Rate 7yr) Source Southern Coast Mountains, B.C. 11 O'Loughlin (1972) Puget Lowlands, Wash. 9 Madej (1982) Olympic Mountains, Wash. 72 Fiksdal (1974) Olympic Mountains, Wash. 28 Reid (1981) Coast Ranges, Ore. 28 Swanson et a l . (1981) Cascade Ranges, Ore. 36 Swanson & Dyrness (1975) Cascade Ranges, Ore. 45 Morrison (1975) Cascade Ranges, Ore. 59 Swanson et a l . (1982a) TABLE A.6: Rate of debris slide erosion relative to forested areas Rate Location Clearcut Road Source Queen Charlotte Islands, B.C. 14 x 25 x Schwab (1983) Southern Coast Mtns., B.C. 2 25 O'Loughlin (1972) Puget Lowlands, Wash. 22 Madej (1982) Olympic Mtns., Wash. - 4-7 Reid (1981) Coast Ranges, Ore. 4 125 Swanson et a l . (1981) Cascade Ranges, Ore. 4 49 Swanson & Dyrness (1975) Cascade Ranges, Ore. 3 344 Morrison (1975) Cascade Ranges, Ore. 23 27 Lyons & Beschta (1983) Notown, New Zealand 22 - O'Loughlin & Pearce (1976)a aCited in Swanson et;al. (1981) 182 networks, and the changed d i s t r i b u t i o n of slope material by cu t - a n d - f i l l construction (Swanson and Dyrness, 1975; Reid, 1981). After the i n i t i a l f a i l u r e , debris s l i d e scars undergo modification by rainsplash, sheetwash, dry ravel, and r i l l i n g (Reid, 1981; Reid et a l . , 1981; Lehre, 1982). Using erosion pins, Lehre (1982) and Reid (1981) measured net surface lowering of 5 (±1) and 16 (±4) mm/yr, respect-i v e l y , which correspond to erosion rates of 1,000-4,000 m3/Km2 scar/yr. Erosion of slide scars may continue for 10 years or more, until the scar i s revegetated (Reid et al., 1981), but in total, secondary sediment production is an order-of-magnitude less than the i n i t i a l failure (Reid, 1981; Lehre, 1982). Finally, f a l l s are the very rapid downslope movement of s o i l and rock by free f a l l i n g , bounding and r o l l i n g . For individual coarse pa r t i c l e s , t h i s process i s called dry ravel, and occurs i n response to cycles of wetting and drying, and freezing and thawing (Swanston, 1974a). Dry ravel i s a ubiquitous erosional process i n southern C a l i f o r n i a (Swanston, 1971; Rice, 1982) and i s locally dominant in exposed alpine areas where scree material accumulates at the foot of steep slopes. However, in terms of the relative importance of mass movement processes in the Pacific Northwest, dry ravel annually transfers less sediment than s o i l creep, and several orders-of-magnitude less than slide and flow events (Swanson et al., 1982a). A.4 Surface erosion Under forested conditions, overland flow i s rarely generated because s o i l i n f i l t r a t i o n capacities greatly exceed normal prec i p i t a t i o n rates (Harr, 1976b; Chamberlin, 1982); for example, i n coastal B r i t i s h Columbia, coarse-textured forested mountain s o i l s have i n f i l t r a t i o n 183 capacities greater than 42 mm/hr, whereas r a i n f a l l intensities are usually less than 26 mm/hr (de Vries and Chow, 1978). But i f the forest flo o r i s disturbed by yarding, road building, or slash burning, i n f i l t r a t i o n may be impaired (Swanston, 1971; Harr et a l . , 1975; Anderson et a l . , 1976; Chamberlin, 1982; Everest and Harr, 1982) and overland flow may occur (Chamberlin, 1982; Hewlett, 1982, p.62), resulting in higher sediment pro-duction after forest floor disturbance (Anderson et al., 1976; Heede, 1983). The rates of sediment delivery to stream channels by surface erosion on forested hillslopes in northern California (Lehre, 1982), western Oregon (Swanson et al., 1982a), and central Idaho (Megahan and Kidd, 1972b) are 4, 5, and 10 m^ /Km^ /yr, respectively. But following logging in central Idaho, Megahan and Kidd (1972b) noted a 60% increase in sediment delivery, to 16 m 3/Km2/yr. Several studies have addressed the s o i l disturbance and surface erosion associated with different yarding and roading a c t i v i t i e s . In a review paper, Chamberlin (1982) indicated that helicopter or balloon yarding caused the least, and ground skidding the most, s o i l disturbance, with s k y - l i n e and high-lead systems i n between. For example, i n northwestern C a l i f o r n i a , s o i l erosion i s four times greater on tractor-yarded slopes than on cable-yarded slopes (Rice and Datzman, 1981), but on steep terrain in British Columbia, even high-lead yarding can cause severe s o i l disturbance (Bockheim et al., 1975; Chamberlin, 1982). In effect, site characteristics determine which yarding system w i l l minimize the general level of s o i l disturbance and erosion (Krag, 1980; Rice and Datzman, 1981; Chamberlin, 1982), while differences in operator performance provide random perturbations (Rice et al., 1979; Rice and Datzman, 1981). In addition to s o i l disturbance on slopes, sediment i s eroded 184 by overland flow on road surfaces and backcuts, as a result of reduced i n f i l t r a t i o n through their compacted surfaces and interception of sub-surface flow by road cuts (Hornbeck and Reinhart, 1964; Reid, 1981; Megahan, 1983); for example, Reid (1981) measured i n f i l t r a t i o n capacities of 0.5 mm/hr and 1.0 mm/hr for heavily-used gravel road surfaces and abandoned road surfaces, respectively. Rates of sediment production by road surface erosion decline rapidly with road dis-use, and are summarized in Table A.7. Overall, the tabulated values are remarkably s i m i l a r , in view of the con-tra s t i n g environmental conditions and land management practices; the consistently greater rates of surface erosion in the Idaho Batholith are the result of the p a r t i c u l a r l y erodible s o i l s (Megahan and Kidd, 1972a). In general, there i s an approximate order-of-magnitude decrease in erosion, from 10,000 m3/Km2 road/yr during logging, to 1,000 m3/Km2 road/yr the f i r s t year of dis-use, and 100 m3/Km2 road/yr thereafter. A.5 Stream bank erosion and large organic debris Most of the work on stream bank erosion has dealt with meandering, lo\tf-gradient rivers where sediment type and hydraulic forces are the principal controls (Thorne, 1982). In contrast, mountain streams in the coniferous forests of the Pacific Northwest are shaped primarily by external factors, such as h i l l s l o p e erosion processes, bedrock control of channel position and geometry, and the role of riparian vegetation in stabilizing stream banks and providing large organic debris (Swanson et al., 1982b). Accelerated rates of stream bank retreat and stream bed aggradation have been associated with an increased sediment supply following land-use changes (O'Loughlin, 1969; Narver, 1972; Madej, 1978, 1982; Bourgeois, 1981; Chamberlin, 1982; Toews and Moore, 1982b; Lyons and Beschta, 1983) and/or major flood events (Kelsey, 1980; L i s l e , 1981, 1982; 185 TABLE A.7: Rates of sediment production by road surface erosion Location Road Type Sediment Production Source (nr/Knr road/yr) Idaho Batholith West Virginia Jammer Skid During Logging 15,120 9,730 First Year Second Year of Dis-use of Dis-use 2,200 970 560 20 Megahan & Kidd (1972a) Hornbeck & Reinhart (1964) Moderate- Light- Abandoned use use Western Haul 10,400 950 120 Reid (1981); Wash. Reid et a l . (1981) 186 Lyons and Beschta, 1983). If most of the sediment i s introduced from point sources (landslides), i t may result i n i t i a l l y in l o c a l , low gradient zones of aggradation and increased channel width (O'Loughlin, 1969; Madej, 1978, 1982; Kelsey, 1980; L i s l e , 1981, 1982; Lyons and Beschta, 1983), due to lateral channel migration and bank undercutting, which may encourage further h i l l s l o p e f a i l u r e and sediment transfer to the stream (Swanson et a l . , 1982b). The coarse sediment fraction may be transported subsequently in a series of slow-moving waves (Madej, 1978, 1982; Kelsey, 1980), whose impacts are greatest near the source and become attenuated as they tr a v e l downstream, unless there i s an increase in sensitivity of the lower reaches to disturbance (Rice, 1981), such as may accompany timber harvesting activities in the riparian zone (Chamberlin, 1982). Logging destabilizes stream banks by causing a loss i n root binding strength after tree removal, surface disturbance during f e l l i n g and/or yarding, and s i t e - s p e c i f i c erosion around in-stream debris (Bourgeois, 1981; Chamberlin, 1982). For example, Narver (1972) found that streamside logging caused stream channel widening and aggradation; Bryant (1980) found that cross-stream yarding resulted in severe bank cutting; and Toews and Moore (1982b) noted higher bank erosion rates with cross-stream f e l l i n g and yarding and/or streamside logging treatments, than with the leave-strip treatment. Large organic debris controls channel form as well as sediment and water routing, and provides food and habitat for aquatic organisms in mountain streams (Heede, 1972, 1981; Swanson et a l . , 1976, 1982b; Swanson and Lienkaemper, 1978; Keller and Swanson, 1979; Keller and Ta l l y , 1979; Bryant, 1980; Mosley, 1981; Chamberlin, 1982; Marston, 1982). The volumes of sediment stored behind in-stream obstructions in forested watersheds are 187 summarized in Table A.8; the anomalous result of P i t l i c k (1982) reflects the greater intensity of h i l l s l o p e erosion processes in northwest California (see Section A.3). By removing stable, in-stream debris and introducing smaller, less stable debris (Swanson et al., 1976; Bryant, 1980; Chamberlin, 1982; Toews and Moore, 1982a, 1982b), timber harvesting a c t i v i t i e s i n the ripar i a n zone may increase the rates of sediment transport (Chamberlin, 1982) and bank cutting (Bryant, 1980). Consequently, rip a r i a n protection measures (e.g. Moore, 1980) emphasize that cross-stream and in-stream act-i v i t i e s should be minimized, by fe l l i n g and yarding away from the stream, and using high deflection cable systems (Chamberlin, 1982). A.6 Hydrologic changes resulting from timber harvesting Streamflow i s determined by the input of water and the hydrologic processes operating within a catchment, and Harr (1976b) describes the hydrologic cycle for small, forested watersheds. Timber harvesting activities can modify streamflow by influencing snow distribution and melt rates; the interception, evapotranspiration, and s o i l storage of water; and the s o i l structure, and thereby, water i n f i l t r a t i o n and transmission rates (Chamberlin, 1982). Changes i n annual water y i e l d , average peak discharge, and average minimum flow in some clearcut and partially cut (i.e. less than 50%) watersheds in the Pacific Northwest are summarized in Table A.9 in percent-age form, or by an x i f the absolute value i s unspecified; further examples are reviewed by Anderson et a l . (1976), Harr (1976a) and Bosch and Hewlett (1982). Many of the studies i n Table A.9 found an increased annual water yield, with the greatest absolute increases during the autumn-winter 188 TABLE A.8: Sediment storage behind in-stream obstructions Volume of Stored Sediment Location m3/30 m Channel Relative to Annual Sediment Yield Source Idaho" Batholith 1.1 15.8 Megahan & Nowlin (1976) Idaho Batholith 2.9 15.5 Megahan (1982) Coast Range, Ore. 3.0 1.2 Marston (1982) Cascade Range, Ore. 6.0 89.9a Swanson et a l . (1982b) New Zealand 0.3 - Mosley (1981) Northwest Calif. 51.7 - P i t l i c k (1982) aCompared to bed load yield. 189 TABLE A.9: Hydrologic changes following clearcut and patchcut logging of watersheds in the Pacific Northwest Logging Site Annual Water Average Peak Yield (%) Discharge (%) Inc. Dec. N i l Inc. Dec. Nil Average Minimum Flow (%) Inc. Dec. N i l Source o Clear-Cut South Coastal B.C. 14 South Coastal B.C. Coast Ranges, Ore. 26 Coast Ranges, Ore. Cascade Ranges, Ore. Cascade Ranges, Ore. 30 Cascade Ranges, Ore. 32 Cascade Ranges, Ore. Southwest Ore. 43 North Carolina 20 30 30 35 7 22 32 78 x x 105 Hetherington (1982) Cheng et a l . (1975) Harris (1971) Harr et a l . (1975) Harr & McCorison (1979) Harr et a l . (1982) Rothacher (1970) Rothacher (1973) Harr et a l . (1979) Hewlett & Helvey (1970) TABLE A.9 (Contd) Annual Water Average Peak Average Minimum Logging Site Yield (%) Discharge (%) Flow (%) Source Inc. Dec. N i l Inc. Dec. N i l Inc. Dec. N i l Patch-Cut South Coastal B.C. Coast Ranges, Ore. Coast Ranges, Ore. Cascade Ranges, Ore. Cascade Ranges, Ore. Cascade Ranges, Ore. Northwest Ore. Southwest Ore. Northern Calif. Northern Calif. 22 12 14 11 x x 30 47 20 Hetherington (1982) Harris (1971) Harr et a l . (1975) Harr et a l . (1982) Rothacher (1970) Rothacher (1973) Harr (1980) Harr et a l . (1979) Rice (1981) Ziemer (1981a) Coast Ranges, Ore. = Alsea Cooperative Watershed Study Cascade Ranges, Ore. = H. J. Andrews Experimental Forest Northwest Ore. = Fox Creek Watershed Study rainy season, and largest relative increases during the summer (Rothacher, 1970, 1971; Harr, 1976a; Harr et a l . , 1979, 1982). However, the observed increases in summer low flows diminish rapidly to pre-logging or lower levels as phreatophytic riparian vegetation becomes established (Harr, 1979, 1983). In order to detect increases in streamflow from partially cut water-sheds, 20 to 40% of the forest cover has to be removed (Rothacher, 1970, 1971; Bosch and Hewlett, 1982). Peak flows may be larger, smaller, or unchanged after logging (see Table A.9), depending on the type and extent of alteration of the hydrologic system, but timber removal results usually in greater increases of autumn and small peak flows, than of winter and large peak flows (Rothacher, 1971, 1973; Harr et a l . , 1975; Ziemer, 1981a). However, some studies have reported changes in the size and timing of even large peak flows, as a result of s o i l disturbance. When roads occupy at least 12% of a watershed, their low i n f i l t r a t i o n capacity and the interception of sub-surface water by road cuts may increase peak flows and decrease their time-to-peak (Harr et a l . , 1975, 1979). These effects also have been attributed to s o i l compaction by tractor yarding and wind throwing of logging slash (Harr, 1976a; Harr et a l . , 1979), although Cheng et a l . (1975) and de Vries and Chow (1978) found reduced peak flows and a delayed time-to-peak, where logging had disrupted the sub-surface channel pathways and forced water to travel more slowly through the s o i l matrix. Hydrologic changes may affect the magnitude and frequency of sediment transport processes, and extreme high flows may create unstable stream banks (Anderson et a l . , 1976). However, the suggestion by Rice (1981) that the f l u v i a l system responds primarily to changes in sediment avail-a b i l i t y , rather than to high flows per se, i s substantiated by the observations of Lyons and Beschta (1983) in the western Oregon Cascades, and 192 of Toews and Moore (1982b) on the west coast of Vancouver Island, B r i t i s h Columbia. Thus, the main effects of a changed hydrologic regime are best interpreted in terms of variations in sediment delivery. 193 APPENDIX B VALUES OF SEDIMENT PRODUCTION AND DELIVERY BY LANDSLIDES 194 TABLE B.1: Armentieres Creek watershed: Landslides Slide frequency (events/Km /yr) Soil transfer (m /Km /yr) Watershed Condition 3 Period of Record Total Production Transfer to Creek and Footslopes Delivery to Creek % b Total Production Transfer to Creek and Footslopes % b Delivery to Creek % b Unlogged 1964-1975 0.21 0.13 62 0.08 38 401 369 92 307 77 1976-1982 0.81 0.40 50 0.31 39 961 777 81 701 73 1964-1982° 0.43 0.23 54 0.17 39 607 519 86 452 74 Logged 1964-1975 1.11 0.44 40 Ni l N i l 444 189 43 Nil N i l 1976-1982 0.76 0.38 50 0.19 25 305 267 88 57 19 1964-1982° 0.98 0.42 43 0.07 7 393 218 55 21 5 Ratio of Logged to Unlogged 1964-1982 2.3 1.8 0.8 0.4 0.2 0.6 0.4 0.6 0.1 0.1 aUnlogged area: 3.18 KmZ Logged area (including roads): 0.75 Km bPercentage of total production. cWeighted by the length of record of each constituent period. TABLE B.2: Mosquito Creek tributary watershed: Landslides Slide frequency (events/Km /yr) Soil transfer (m /Km /yr) Watershed Condition 3 Period of Record Total Production Transfer to Creek and Footslopes % b Delivery to Creek % b Total Production Transfer to Creek and Footslopes % b Deliver to Creek y % b Unlogged 1955-1963 0.06 0.02 34 Nil N i l 95 37 39 Nil Nil 1964-1975 0.04 0.02 49 Nil N i l 85 72 85 Nil N i l 1976-1982 0.03 0.03 100 0.03 100 113 113 100 113 100 1955-1982° 0.05 0.02 51 0.01 18 95 71 75 28 29 Logged 1955-1963 Nil Nil - N i l - N i l N i l - N i l -1964-1975 0.23 0.08 33 Nil N i l 100 46 46 Nil Nil 1976-1982 0.27 0.13 50 Nil N i l 132 93 70 Nil Nil 1955-1982° 0.24 0.10 40 Nil N i l 112 63 56 Nil Nil Ratio of Logged to Unlogged 1955-1982 5.4 4.2 0.8 Nil N i l 1.2 0.9 0.7 Nil Nil 3Unlogged area: 4.30 Km Logged area (including roads): 1.08 Km Percentage of total production. cWeighted by the length of record of each constituent period. TABLE B.3: Mountain Creek watershed: Landslides Period Watershed of Condition 3 Record 9 o o Slide frequency (events/Km /yr) Soil transfer (m /Km /yr) Transfer to Delivery Transfer to Delivery Total Creek and to Total Creek and to Production Footslopes % Creek % Production Footslopes % Creek ) Unlogged Logged Ratio of Logged tc Unlogged0 1934-1954 0.01 0.01 100 0.01 100 31 31 100 31 100 1955-1965 0.02 0.01 50 0.01 50 15 10 67 10 67 1966-1975 0.10 0.04 36 0.03 27 182 47 26 33 18 1976-1982 0.19 0.04 20 0.01 7 172 99 58 13 8 1934-1982° 0.05 0.02 34 0.01 23 78 39 50 24 31 1934-1954 Nnr 1955-1965 0.15 Nil N i l N i l N i l 315 Ni l N i l N i l Nil 1966-1975 0.40 0.16 40 0.08 20 355 210 59 8 2 1976-1982 1.50 0.81 54 Nil N i l 300 138 46 Nil N i l 1955-1982° 0.58 0.26 45 0.03 5 326 110 34 3 1 1934/1955 -1982 10.9 14.4 1.3 2.4 0.2 4.2 2.8 0.7 0.1 0.1 /~2 Unlogged area: 11.40 Km Logged area (including roads): 1.24 Km"1 Percentage of total production. 'Weighted by the length of record of each constituent period, ^ a t i o of 1955-1982 logged values to 1934-1982 unlogged values. TABLE B.4: Lagins Creek tributary watershed: Landslides 3 9 o o Slide frequency (events/Km /yr) Soil transfer (m /Km /yr) Period Transfer to Delivery Transfer to Delivery Watershed of Total Creek and to Total Creek and to Condition Record Production Footslopes % c Creek % c Production Footslopes % c Creek % c 1934-1954 0.02 0.01 33 0.01 33 14 6 43 6 43 1955-1965 0.06 0.05 75 0.03 51 46 43 93 34 74 1966-1975 0.06 0.04 67 Nil N i l 23 19 83 Ni l Nil 1976-1982 Nil N i l - N i l - N i l N i l - N i l -1934-1982d 0.04 0.02 61 0.01 28 21 16 76 10 48 a A l l landslides were initiated in unlogged terrain. Unlogged area: 5.29 Km Logged area (including roads): 0.63 Km cPercentage of total production. ^Weighted by the length of record of each constituent period.