Impacts of Natural Events

General Technical Report PNW-104 April 1980

This file was created by scanning the printed publication. Mis- scans identified by the software have been corrected; however, some errors may IMPACTS OF remain. NATURAL EVENTS

DOUGLAS N. SWANSTON

U.S. Department of Agriculture Forest Service Pacific Northwest Forest and Range Experiment Station Natural events affecting vegetative cover and the hydrology and stability of a stream and its parent watershed are key factors influencing the quality of anadromous fish habitat.

High intensity storms, drought, soil mass movement, and fire have the greatest impacts. Wind, stream icing, and the influence of insects and disease are important locally.

Keywords: Anadromous fish habitat, hydrology, stability. USDA FOREST SERVICE General Technical Report PNW-104

INFLUENCE OF FOREST AND RANGELAND MANAGEMENT ON ANADROMOUS FISH HABITAT IN WESTERN ‘NORTH AMERICA

William R. Meehan, Technical Editor

2. Impacts of Natural Events

Douglas N. Swanston

1980

PACIFIC NORTHWEST FOREST AND RANGE EXPERIMENT STATION Forest Service, U.S. Department of Agriculture Portland, Oregon PREFACE

This is the second in a series of publications summarizing knowledge about the influences of forest and rangeland management on anadromous fish habitat in western North America. This paper addresses the effects on fish habitat of naturally occurring watershed disturbances and sets the scene for future discussions of the influences of human activities. Our intent in presenting the information in these publications is to provide managers and users of the forests and rangelands of western North America with the most complete information available for estimating the consequences of various management alternatives. In this series of papers, we will summarize published and unpublished reports and data as well as the observations of resource scientists and managers developed over years of experience in the West. These compilations will be valuable to resource managers in planning uses of forest and rangeland resources, and to scientists in planning future research. The extensive lists of references will serve as a bibliography on forest and rangeland resources and their uses for western North America. Previous publications in this series include: 1 "Habitat requirements of anadromous salmonids," by D. W. Reiser and T. C. Bjornn. TABLE OF CONTENTS

INTRODUCTION...... ) .. e .... 1 HYDROLOGIC IMPACTS ...... 1 Basic hydrologic processes ...... 2 Minimumflows ...... 5 Maximumflows ...... 6 .. SEDIMENT AND ORGANIC DEBRIS LOADING ...... 6 MASS SOIL MOVEMENT ...... 8 Slump-earthflows ...... 8 Debris avalanches and debris flows 11 Debris torrents ...... 13 WIND ...... 14 INSECTS AND DISEASE ...... 16 FREEZING AND ICE FORMATION ...... 17 FIRE ...... 18 CONCLUSIONS ...... 21 LITERATURE CITED ...... 21 Not all of these impacts are necessarily damaging to the habitat. For example, movement and redistribution of gravel by scour and streambed overturn may flush finer sediment from the gravels, improve access to gravel and intragravel waterflow, and create larger areas of usable spawning gravel, Stream-channel scour and excess streamflow may also remove blocking organic debris and improve access. The actual effect on habitat accepta- bility is .largely dependent on peripheral damage to other habitat variables and on timing. Scour and streambed overturn after spawning can effectively destroy an entire generation by burial and mechanical grinding of eggs and alevins. INTRODUCTION

Acceptable habitat for anadromous salmonids is the product of interaction among the geologic, climatic, and vegetative factors that control channel gradient and configuration, bottom composition, base flow, quantity and timing of maximum discharge, sediment and organic loading, and ground and overstory vegetative cover. Under "normal" or "average" climatic conditions, those factors are in virtual equilibrium, and acceptable habitat is maintained with few or no short-term changes, Any alteration in this normal condition, however, has the potential for initiating events that may drastically alter the state of equilibrium. The HYDROLOGIC IMPACTS result may be removal or redistribution of spawning gravels, Seasonal and short-term addition of substantial quanti- changes in weather and hydrologic ties of sediment and organic conditions and the amount and debris, alteration of access, distribution of vegetative cover destruction of viable eggs and are major factors controlling alevins, removal or redistribu- streamflow .and occurrence of tion of food organisms, and natural events that may be increased temperatures from damaging to anadromous fish removal of streamside vegetation. habitat . BASIC HYDROLOGIC PROCESSES and frequent high-intensity, long- duration storms occur. For The quantity and timing ,of example, in old-growth Douglas-fir streamflow in anadromous fish forests in the western Cascades of spawning channels is largely Oregon, Rothacher (1963) found that determined by water input and nearly 100 percent of storms the hydrologic processes operat- less than 0.13-cm (0.05-in) was ing within the contributing water- intercepted and evaporated, but shed. Harr (1976) clearly de- less than about 5 percent of storms scribes these hydrologic processes of more than 20 cm (8 in) was for small forested watersheds. intercepted and evaporated. In Water is introduced into the the drier interior areas of the hydrologic cycle as rain, atmos- West, annual rainfall is low, and pheric moisture (fog), or snow. high-intensity, long-duration Rain falling on a watershed may storms are infrequent, trees are reach the ground and stream sur- less closely spaced, and total faces directly or be intercepted vegetative cover may be sparse. by vegetation (fig.1) Some of Under these conditions, intercep- . tion losses are substantially ATMOSPHERE less, ranging from 5 to 14 percent of total annual rainfall (Anderson et al. 1976), but are significant in controlling the total amount of water entering a watershed. When major storms do occur, interception losses become negligible.

Snow falling on the watershed is subject to the same interception as rain and in about the same proportions (Dunford and Niederhof 1944, Rowe and Hendrix 1951, Sartz and Trimble DEEPSEEPAGE 1956, Hart 1963). In addition, where substantial snow packs Figure 1.--Diagram of the rain-dominated develop, water may be detained hydrologic cycle (from Harr 1976). for considerable periods in its this intercepted water evaporates; passage through the watershed and the remainder reaches the ground into the stream system. Snow packs as stem-flow and drips through contribute to such "surface storage" the canopy and leaf cover. The both in the frozen phase and as relative amount of rain "lost" to free water held in the pore evaporation or delayed in its spaces (Anderson et al. 1976). passage to the watershed by this The volume and duration of interception process depends on detention depend on such condi- the amount and extent of vegetative tions as snow depth, air tempera- cover and the intensity and dura- ture, pore size and initial free tion of the storm producing the water content (Corps of Engineers rain. In the humid, heavily 1956, Smith and Halverson 1969, vegetated, old-growth forest areas Smith 1974). Greatest detention within the influence of the Pacific is in the warm snow packs of the storm systems, interception losses Sierra Nevada and the Pacific from rainfall are significant in slopes of the northern Rocky the spring, summer, and fall when Mountains where most of the water major storms are infrequent. is detained over the winter and Interception losses are substan- released when the snow melts. In tially reduced during the winter the Cascade and Coast Ranges of when evapotranspiration is low

2 northern California, Oregon, of stored water (melting snow pack) Washington, British Columbia, and coupled with rainfall, which to- Alaska, although most precipita- gether may exceed the infiltration occurs as rain, snow is tion capacity of the soil. common--particularly at higher elevations. Occasionally, a Upon entering the soil, water snow pack may remain for 1 to 3 is subject to gravitational and months, but in most years it capillary forces that cause it to usually melts within 1 to 3 weeks. move and frictional forces that The highest runoff in this region tend to restrict movement. Be- has resulted from rapid snowmelt cause of the slope of most water- during warming trends coupled sheds and because soil conductivity with prolonged heavy rainfall generally decreases with depth, (Waananen et al. 1971, Harr et al. water entering the soil begins to 1979). Watershed vegetation move downslope as it moves deeper delays snowmelt by shading and into the soil. The direction and thus extends the time that snow rate at which the water moves is held in surface storage. It depend on rainfall rates and soil also controls, to a certain ex- properties. Both rate and direc- tent, the rate and timing of tion vary considerably over the snowmelt. Although snow under course of a storm (Harr 1977) . forest cover melts later and Maximum velocities of soil water- persists longer than snow in the flow are low and frequently about open, it may melt more rapidly equal to the average rate of once melting begins, because rainfall during a storm. This melting begins later in the slow-moving soil water is subject season when temperatures may be to evaporation and depletion by much higher (Anderson et al. 1976). plants through the process of A mixture of vegetated and open transpiration. The rate at which areas on a watershed may promote plants withdraw water is largely snowmelt at different times and a function of energy available reduce the total quantity released for water vaporization in leaves at any one time. Topography also and the availability and ease with desynchronizes melting; snow on which water may be withdrawn from southerly aspects may disappear the soil. Thus, during the spring before much of the snow melts on and early summer when soil water northerly aspects. content is high and the growing season is at its peak, evapo- Once the water reaches the transpiration withdrawals are watershed floor, it infiltrates high. If no additional water is into the soil. Depending on the added to the soil as the growing difference between the rate of season progresses, soil moisture water arrival at the soil surface decreases and the remaining water and the soil's infiltration capac- becomes more tightly held by the ity or ability to allow water to soil. The result is a decrease enter, some water enters the sub- in subsurface flow to the channel surface and some may become over- and a net reduction in streamflow. land flow. In undisturbed for- This condition is alleviated after ests, infiltration rates of soils the first storms in the fall re- generally far exceed the maximum charge the soil water deficit. rates of rainfall so that all water enters the soil. Overland In simplified terms, streamflow occurs under such conditions flow--on an annual or longer of vegetation, primarily as the basis--is the difference between result of limited capacity for precipitation and evapotranspira- storage, percolation, and channel tion losses (Harr 1976). Soil expansion. Overland flow may moisture storage changes from also result from rapid addition time to time. Some water may

3 also seep deep into the subsoil these withdrawals become less and bedrock and not appear as important. Stormflow resulting streamflow in a small headwall from extreme events is minimally basin. Generally, however, water influenced by these withdrawal not removed by plants ultimately processes, although a forest cover moves downslope as saturated or does detain some portion of any unsaturated flow to supply streams. rainstorm and thus somewhat re- duces flood discharge and peak flows. As the watershed responds to rainfall, streamflow increases to Watershed size influences the a maximum known as "peak flows." quantity of streamflow and the size Each storm hydrograph has its own and timing of peak flows during any peak flow, reflecting the inter- particular storm. Generally, the action of rainfall with the physi- smaller the watershed, the more cal characteristics of the water- rapid are the streamflow increases shed. High, sustained rates of in response to rainfall. For rainfall contribute to greater example, in small watersheds on storm runoff and higher peak flow. the H. J. Andrews Experimental The magnitude of the increase in Forest in western Oregon, maximum streamflow between the start of rates of runoff have approached storm runoff and the peak is 80 percent of the average rate of highly variable. It depends on rainfall for the previous 12 to 24 the antecedent moisture content of hours and 75 percent of the maximum the watershed's soils and the 6-hour rainfall (Rothacher et al. characteristics of the storm. In- 1967). As watersheds increase in creases in streamflow of at least size, total water yields increase, two orders of magnitude are not but peak flows and response time to infrequent between the start and storm events become somewhat reduced. the peak of storm runoff. Maxi- mum peak flows have resulted from Quantity and timing of local rain on snow, during which a sub- stormflow are also related to ele- stantial portion of streamflow vation. Higher elevation water- comes from rapid snowmelt con- sheds generally receive a larger current with the downslope move- quantity of water per storm. Be- ment of water. cause these watersheds are small, soil water retention and evapotrans- The quantity and rate at pirational losses are less in- which water reaches the channel fluential and runoff and peak flows and passes through a watershed tend to be higher than in the larger, 'system during a particular hydro- lower watersheds. Studies in North logic event is influenced by Carolina (Hewlett 1967) showed that storm and watershed size and forested primary ridges at 1 524 m certain topographic considera- delivered almost 457.2 mm of direct tions. Obviously, the larger runoff per year, but forest land at a storm, the greater is the lower elevations delivered only amount of water going into the 63.5 mm. Rainfalls of 213.4 to system and the larger the poten- 274.3 mm in a December storm on tial streamflow. The influence three watersheds above 914.4 m of vegetation on streamflow re- produced 127.0 to 223.5 mm of sulting from small storms is direct runoff and maximum peaks greater than that from large of 68-167 m3/s. The rainfall on storms because interception and three watersheds below 91.4 m evapotranspiration account for was 172.7 to 177.8 mm; direct a larger proportion of small runoff was 40.6 mm to 58.4 mm rainfalls, and soil water re- and peaks were 22 to 32 m3/s tained against gravity by plant (Hoover and Hursh 1944) . roots accounts for a greater proportion of water entering the Perhaps the most important soil. As rainfall increases, concept in understanding the

4 hydrology of watersheds is the MINIMUM FLOWS variable source area of storm runoff described by Harr (1976). Monthly precipitation and This concept relates storm runoff corresponding monthly streamflow to a dynamic source area that vary greatly by season in the expands and contracts according mountainous regions of the North- to rainfall characteristics and west. The heaviest rainfall the capacity of the soil mantle usually occurs from high-intensity to store and transmit water storms during the late fall and (Hewlett and Nutter 1970). Thus, winter in the Pacific coastal as a storm progresses, the channel mountains and during the late network expands to many times its spring along the west slope of perennial dimensions (fig. 2) , the Rocky Mountains. Extended dry weather is infrequent in the northern portions of the Pacific coastal mountain system (Washington, British Columbia, and Alaska). In these areas, rainfall is sufficiently frequent throughout th-e year to maintain adequate streamflow even during the summer. Elsewhere in the forested regions of the North- west, extended summer drought is common; most of the annual precipitation occurs during the late fall and early spring storms. In such areas, minimum streamflows occur during the late summer and early fall and may be 1,000 to 5,000 times smaller than maximum peak flows in winter. Flow may cease entirely in many first-order streams (Harr 1976).

Figure 2.--Time lapse view of a small During low-flow periods, soil watershed showing the expansion of the moisture is at its lowest because channel network and source area of water has been removed by evapo- storm runoff (shaded area) (from transpiration and slow subsurface Harr 1976). drainage to streams. Storms are infrequent and small; a large pro- and streams become both longer portion of rainfall, when it occurs, and wider. As a result of the is intercepted by forest vegetation variable source area of stream- and evaporates, with little water flow, both quantity and quality reaching the soil. Any storm run- of streamflow can change drasti- off results almost entirely from cally over a given period because channel interception, and peak the proportion'of a watershed flows are extremely small. The active in streamflow production net result is that streams have changes. For sediment and organic little capacity to move sediment debris, the stream represents a or transport debris. Streamflow depository of variable surface is maintained by slow drainage area as well as a mode of trans- from local ground-water sources portation. During extreme runoff, or isolated saturated zones. sediment stored along the channel Streamflow is at a very low rate, and debris not falling directly and stream network divisions are into the channel of the permanent at annual minimums. First-order stream may be subject to trans- channels have no flow, and higher port as the channel system expands.

5 order channels are short, with produced almost entirely by high- flow occupying only a small part intensity rainstorms, Within the of their widths. Flow may con- snowpack zone, both rain alone sist of a slow trickle between and warm rain on snow associated relatively isolated pools. Under with warming temperatures are these conditions, large areas of important generators of peak flow. stream gravel are exposed; intra- In western Oregon, peak flows dur- gravel water movement is reduced ing major winter storms may be as and limited to considerable depths much as 100 times greater than below the surface, During flows immediately before such droughty periods, .the potential storms.l/ Comparable increases for fire in the surrounding vege- can be postulated for the rest tation is also greatly increased, of the Pacific coastal mountain belt based on the similarity in intensity and frequency of storms throughout the region. Along the west slopes of the Rocky Mountains, Storms of high intensity or peak flows develop during the long duration in undisturbed spring as a result of melting snow watersheds cause increased peak and occasionally as the result of flows and local flooding,, fre- upslope or convective storms, which quently resulting in accelerated dump large quantities of rain in mass soil movement, streambank the foothills during May and June erosion, and surface erosion of (Anderson 1975). If snowmelt and nonvegetated areas with attendant storm rainfall are synchronized, increases in sediment discharge substantial increases in peaks to the channel 1/ (Rothacher 1959, and accompanying flooding are 1973; Anderson 1975: Fredriksen produced (Goodell 1959) et al. 1975). Patton and Baker (1976) have defined flash flood or high peak-flow potential in drainage basins by a regional indexing technique computed as the standard deviations of the annual maximum streamflows, Basins with high flash-flood potential tend toward greater relief and greater drainage densities combined with steep hillslopes and stream-channel gradients, These are characteristics of the majority of the Pacific Coast anadromous fish stream and river systems, In the heavily forested Pacific coastal mountain systems, highest peak flows come during the late fall and winter as a result of heavy rainfall on wet soil mantles, At lower eleva- SEDIMENT AND ORGANIC DEBRIS tions, these peak flows are LOADING 1-/Unpublished report , "Forest Heavy sedimentation of streams practices and streamflow in western and channel damage from flood flow Oregon," by R. D. Harr. Paper pre- and soil mass movements are closely sented at Symposium on Watershed linked to the pulsing action of Management, Am. SOC. Civ. Eng. , Logan, streamflow caused by fall and Utah, 1975. winter storm fronts from the

6 Pacific Ocean and by the rapid transported for considerable dis- snowmelt and upslope convective tances down the channel to be storms generated in the interior deposited as windrows of small mountain areas. Streams rise and organic detritus and piles or jams fall rapidly in relation to pre- of mixed logs and smaller organic cipitation, storm duration, snow- debris. The point of deposition melt, and the amount of water and the distance of travel depend already in the soil. largely on the originating-process and the volume of streamflow in The suspended sediment con- the channel at the time of depo- tent of stormflow generally lags sition. Streams of all sizes are behind the storm hydrograph, with affected by debris, but loading the highest sediment concentrations tends to decrease as stream width in transport occurring as stream- increases. In small headwater flow increases to its peak. Sedi- channels, trees generally lie ment transported during these where they fall and are only moved stormflow periods can be substan- as a result of decomposition or tial. Mean annual concentrations by catastrophic events, such as of sediment from undisturbed extreme stormflow or debris tor- watersheds are relatively low. rents.2/ As streams within a For example, Fredriksen et al. watershed expand in response to (1975) reported mean annual storm runoff, they become wider concentrations of sediment for and deeper. Large organic material undisturbed watersheds in the may float and accumulate at channel western Cascades of Oregon and in obstructions, forming debris the Oregon Coast Ranges from 1.3 accumulations or debris dams. In to 44.6 mg/l and from 1.4 to 21.4 rivers, where a tree bole of any mg/l, respectively. This contrasts size can be floated, large organic sharply with mean maximum in- debris usually accumulates at creases in concentration occurring bends or along the channel margin primarily during stormflow, which as a result of high flow (Swanson ranged from 11 to 52 times the et al. 1976). mean annual rate for the western Cascades watershed and from 39 to Large jams can block the 83 times the mean annual rate for passage of migrating fish and the Coast Ranges watershed. Much effectively close areas to spawning of this increase is the direct (Meehan 1974). In addition, such result of soil mass movements into jams may form a temporary base the stream channels and mobiliza- level for the affected channel, tion of sediment temporarily stored resulting in deposition on the within or along the channel margins. upstream side and extensive scour downstream. If the jam is large Large organic debris deposited enough, the stream may form a new and redistributed during stormflow channel, causing extensive bank and soil mass movements is a erosion and effectively bypassing common and important channel the jam and sections of spawning feature, with both physical and gravels downstream. In headwater biological consequences to anadro- streams, debris torrents and extreme mous fish habitat. stormflows carry large amounts of Most often debris--including wood through the system causing material resulting from fire, extensive scour, redistribution disease, and decomposition--is of stream gravels, and damage to delivered to the stream by a culverts, roads, and bridges .2/ combination of processes includ- 2/Manuscript in preparation, "some ing windthrow, streambank under- management impacts of organic debris in cutting, and soil mass movements. Pacific Northwest streams," by F. J. Once in the stream, it may be Swanson and G. W. Lienkaemper. USDA deposited almost immediately in For. Serv., Pac. Northwest For. and or along the channel margins or Range Exp. Stn., Portland, Oreg.

7 Debris accumulations in small- periods--produces saturated soil to medium-sized streams (third- and a rising water table on steep and fourth-order) often trap gravel slopes, frequently initiating soil that may provide excellent spawning mass movements that transport habitat for anadromous fish. Also, large volumes of sediment and fine sediment trapped by debris in organic debris into and through headwater streams is routed through the channel systems. Such the system slowly as the wood accu- processes add substantial quantities mulations decompose. Wood in small of sediment and organic debris to streams frequently provides a the stream channels over short "steepened channel profile," in periods (minutes, hours, days) , which much of the drop in the stream causing channel alteration, rapid is over wood-created falls alter- increases in bed and suspended loads, nating with longer, low-gradient siltation of gravels, and partial pools. This gives a large portion or complete blocking of the stream of the stream a lower gradient than channel through debris-dam forma- the overall stream channel. tion. If entry velocities and channel gradient are high enough, The biological community can the tremendous bulking effects of benefit from such stream debris. contained soil, rock, and organic Debris accumulations provide cover debris commonly scour the channel for resident and anadromous fish below the point of entry, removing (Narver 1971, Hall and Baker 1975), and redistributing bottom gravels while serving to retain detritus and destroying substantial portions entering the stream system. Before of the streamside vegetation. detritus feeders find detritus of terrestrial origin palatable, it The mechanics of failure, types must be conditioned by the microbes of soil mass movement, and the fac- in the stream (Triska and Sedell tors controlling and contributing to 1975). Debris accumulations keep development of soil mass movements the conditioned detritus, the in- on forested terrain are well de- sects that eat it, and the fish scribed in the literature (Bishop that eat the insects all in the and Stevens 1964; Swanston 1967, same area. 1969, 1971, 1974, Swanston and 1 Swanson 1976). Three groups of processes have a major impact on habitat .

SLUMP-EARTHFLOWS Earthflow processes (fig. 3) are for the most part slow moving; where they intersect a stream, they provide continuous long-term sources of sediment to the channel. Measured rates of movement in Oregon and northern California, where these processes are most common, range from 2.5 cm/year to as high as 2 720 cm/year with the higher rates predominantly occurring along the major anadromous fish rivers draining the northern California Coast Ranges (table 1). Based on studies of 19 earthflows MASS SOIL MOVEMENT entering the Van Duzen River basin in northern California, Kelsey Increased water in the soil-- (1978) estimated that the total both seasonal and during storm

8 Figure 3.--Diagram of a typical slump-earthflow developed in deeply weathered bedrock and d surficial materials. Slumping is the backward rotation of a block of soil along a curved failure surface with little downward displacement. Earth- flows usually begin with a slump or series of slumps- and-, through a combination of true rheological flow of the clay fraction and slumping and sliding of individual blocks, move downslope with a continu- ity of motion resembling the flow of a viscous fluid.

Table 1--Rates of movement of active earthflows in the western Cascade Range, Oregon (Swanston and Swanson 1976), and Van Duzen River basin, northern California (Kelsey 1978)

Period of Movement Method of Location record rate observation

Years 1 Landes Creek 15 12 Deflection of (Sec. 21, T.22 S, R.4 E) road 1 Boone Creek 2 25 Deflection of (Sec. 17, T.17 S, R.5 E) road Cougar Reservoir 1 2 2.5 Deflection of (Sec. 29, T.17 S, R.5 E) road 1 Lookout Creek 1 7 Strain rhombus (Sec. 30, T.15 S, R.6 E) measurements across active ground breaks 2 Donaker Earthflow 60 Resurvey of stake (Sec. 10, T.1 N, R.3 E) line 2 Chimney Rock Earthflow 530 Resurvey of stake (Sec. 30, T.2 N, R.4 E). line 2 Halloween Earthflow 2 720 Resurvey of stake (Sec. 6, T.1 N, R.5 E) line

'Swanston and Swanson 1976 . 2Kelsey 1978 .

9 sediment discharge to the river protrudes into the channel and by earthflow processes between is gradually eroded away by high 1941 and 1975 was 1 409 500 m3. winter flows, Slumping (fig. 4), This is equivalent to an annual from the undercutting of the yield of 41 455 m3 or about protruded material by stormflow, 2 182 m3 per failure. In con- may abruptly add large quantities trast, studies at Lookout Creek of soil, rock, and organic debris in the western Cascades of Oregon to the channel. If flows are high showed annual movement rates of enough, this soil and some of the only 10 cm/year (Swanston and finer organic debris may be trans- Swanson 1976), with estimated ported out of the system almost annual yields from a single immediately to be distributed earthflow of only 340 m3. downstream as blankets and windrows of sediment and organic Earthflow movement is pre- material. The larger rock and dominantly seasonal--with most organic debris may remain behind movement occurring after fall and as a residual ''lag," forming a winter rains have thoroughly tangled mass of rocks and logs wetted the slopes--although behind which gravels and sediment movement may be continuous in from up-channel may accumulate. areas where groundwater maintains This lag frequently creates a the water content of the moving sharp increase in channel gradient mass through the accumulation zone. During periods of movement, the individual earthflow toe

Figure 4.--The "Drift Creek Slide" in the west-central Coast Ranges of Oregon, a slump in nearly horizontal sandstone and siltstone. The lower end has been transformed into an earthflow and has dammed Drift Creek to form a lake more than 12 meters deep.

10 DEBRIS AVALANCHES AND Debris avalanches have rather DEBRIS FLOWS consistent characteristics in the various geologic and geomorphic Debris avalanches, flows, settings from northern California and torrents constitute the most to southeast Alaska and eastward damaging of the soil mass-move- into the Idaho batholith country ment processes to anadromous fish (Bishop and Stevens 1964; Dyrness habitat. Debris avalanches and 1967; Swanston 1970, 1974; Gonsior debris flows (fig. 5A and B) are and Gardiner 1971; O'Loughlin rapid, shallow mass movements 1972; Colman 1973; Fiksdal 1974). that develop on steep hillsides In all these areas, debris ava- and generally contribute 60 per- lanches are usually triggered cent or more of their initial during high-intensity storms. failure volume almost immediately For example, in the H. J. Andrews to the channel. Debris avalanches Experimental Forest, Oregon, generally grade into debris flows storms of a 7-year or less return as water content of the sliding period initiate debris avalanch- mass increases . ing in forest areas. Swanston (1969) has correlated storms with A DEBRIS AVALANCHE- DEBRIS FLOW a 5-year return interval with accelerated debris avalanching in coastal Alaska.

Debris avalanches leave scars as spoon-shaped depres- sions with long tails extendinq downslope toward the channel, from which less than 10 to as much as 10 000 m3 of soil and organic debris have been transported. DEBRIS AVALA Average volumes of individual debris avalanches in forested B areas in the Pacific Northwest range from about 1 540 to 4 600 m3. Areas prone to debris avalanches are typified by shallow, low-cohe- sion soils on steep slopes where subsurface water may be concentra- ted by subtle topography on impermeable bedrock or glacial till surfaces. Because debris avalanches are shallow failures, factors such as timber and other vegetation (which control root- anchoring effects and transfer of wind stress to the soil mantle, Figure 5.--Debris-avalanche debris-flow as well as rate of water supply failures. Debris avalanches are to the soil during rainfall and rapid downslope movements of soil, snowmelt) have significant effect rock, and organic debris, which gen- on where and when debris avalanches erally grade into debris flows as occur. An added factor in areas water content of the sliding mass of even occasional seismic activity increases: A, Diagram of typical is the lateral stress applied to debris avalanche or debris flow in the soil mantle by ground shaking shallow, semi-cohesionless soils; during earthquakes. Seismic B, massive debris avalanche developed activity has been noted as a fac- in granite-derived soils near Wrangell, tor contributing to soil mass- Alaska. movement initiation in Alaska

11 (Bishop and Stevens 1964) and in The rate of occurrence is con- . the central Rocky Mountains trolled by stability of the land- (Bailey 1971) . In the Queen scape and the frequency of storms Charlotte Islands, British Columbia, severe enough to trigger them. which is noted as the most seismic- Swanston and Swanson (1976 and ally active area in Canada table 2) report annual rates of (Sutherland-Brown 1968), a direct debris-avalanche erosion from six correlation with high soil mass- forested study sites in Oregon, movement activity has been postu- Washington, and British Columbia, lated by Alley and Tomson (1978). ranging from 11 to 72 m3/km2 per year .

Table 2--Debris-avalanche erosion in forest, clearcut, and roaded areas (Swanston and Swanson 1976)

Debris- Rate of debris- Site Period of ----- avalanche erosion record Area- - - -- Slides avalanche relative to erosion forested areas Years Percent Km2 Number M3/(km2.yr) Stequaleho Creek, Olympic Peninsula, Washington (Fiksdal 1974) Forest 84 79.0 19.3 25 71.8 1.0 Clearcut 6 18.0 4.4 ------Road 6 3.0 .7 83 11 825.0 165.0 24.4 108 Alder Creek, western Cascade Range, Oregon (Morrison 1975) Forest 25 70.5 12.3 7 45.3 1.0 Clearcut 15 26.0 4.5 18 117 .1 2.6 Road 15 3.5 -6 75 15 565.0 344.0 17.4 100 Selected drainages, Coast Mountains, S.W. British Columbia Forest 32 83.9 246.1 29 11.2 1.0 Clearcut 32 9.5 26.4 18 24.5 2.2 Road 32 1.5 4.2 -11 l282 .5 25.2 276.7 58

H. J. Andrews Experimental Forest, western Cascade Range, Oregon (Swanson and Dyrness 1975) Forest 25 77.5 49.8 31 35.9 1.0 Clearcut 25 19.3 12.4 30 132.2 3.7 Road 25 3.2 2.0 69 1 772.0 49.0 64.2 130

1Calculated from O'Loughlin (1972, and personal communication) , assuming that area of road construction in and outside clearcuttings is 16 percent of the area clearcut. Colin L. O'Loughlin is now at Forest Research Institute, New Zealand Forest Service, Rangiora, New Zealand.

12 When a debris avalanche occurs, large quantities of soil, rocks, and organic debris are dumped directly into the channel. Because debris avalanches occur primarily during high stormflow, a large part of the soil and organic debris is transported almost immediately away from the entry point to be deposited downstream as blankets and windrows. Boulders and large organic debris remaining at the entry point may temporarily dam the channel and cause an increase in channel gradient for a variable distance downstream.

DEBRIS TORRENTS Debris torrents are the rapid movement of water-charged soil, rock, and organic debris down steep stream channels Figure 6.--Debris torrent developed (fig. 6). Debris torrents typi- in pumice on the north slope of cally occur in steep, intermittent, Entiat Valley, east-central Cascades, first- and second-order channels. Washington. Torrent developed as These events are triggered dur- the result of temporary damming of ing extreme stormflow by debris the stream at near peak flow by avalanches from adjacent hill- debris avalanches along the channel slopes, which enter a channel slopes. and move directly downstream, or by the breakup and mobilization of debris accumulations in the channel. The initial slurry of water and associated debris commonly entrains large quantities of additional inorganic and living and dead organic material from the streambed and banks. Some torrents are triggered by debris avalanches of less than 100 m3, but may ultimately include 10 000 m3 of debris entrained along the track of the torrent. As the torrent moves downstream, hundreds of meters of channel may be scoured to bedrock (fig. 7). When the torrent loses momentum, a tangled mass of large organic debris is deposited in a matrix of sediment and fine organic material covering areas up to several hectares (Swanston and Figure 7.--Channel scoured to bedrock by debris torrent passage in the Oregon Coast Ranges.

13 Swanson 1976, and fig. 8) The of Oregon (Morrison 1975, Swanston main factors controlling the and Swanson 1976). In these studies, occurrence of debris torrents are rates of debris-torrent occurrences the quantity and stability of were observed to be 0.005 and 0.008 debris in channels (supplied by event per square kilometer per earlier debris avalanches and year for forested areas (table 3) . debris torrents), steepness of Torrent tracts initiated in forested channel, stability of adjacent areas ranged from 100 to 2 280 m hillslopes, and peak-discharge and averaged 610 m of channel length. characteristics of the channel, Debris avalanches played a dominant The concentration and stability role in triggering 83 percent of of debris in channels reflect inventoried torrents (Greswell et al, the history of stream flushing 1979). Mobilization of stream debris and the health and stage of not immediately related to debris development of the surrounding avalanches (debris-dam failure within timber stand (Froelich 1973), the channel) has been an important local factor in initiating debris torrents in headwater streams (Swanston 1969) .

Figure 8.--Debris jam in an anadromous fish stream in the Oregon Coast Ranges. The jam resulted from a debris torrent entering the stream from an adjacent slope. WIND Although debris torrents Strong winds, primarily during pose significant hazards to anad- storms, frequently uproot trees, romous fish habitat, they have re- which drastically disturbs forest ceived little study (Fredriksen soils and may effectively alter the 1963, 1965; Morrison 1975; immediate channel environment. Swanson et al. 1976). Velocities of debris torrents, estimated to Windthrow has been recognized be up to several tens of meters for many years as a widespread per second, are known from only natural phenomenon in forested a few spoken and written accounts. regions throughout the United The rates of occurrence of tor- States (Shaler 1891, Holmes 1893, rents' have been systematically Van Hise 1904, Lutz and Griswell documented in only two small 1939, Stephens 1956). Storms with areas of the Pacific Northwest, winds of hurricane force may cause both in the western Cascade Range an entire stand to be uprooted; more commonly, however, scattered

14 Table 3--Debris-torrent occurrence for selected areas in western Oregon (Swanston and Swanson 19 7 6)

Rate of debris- Area Period Debris torrents Debris torrents Total tr iggered with no torrent occurrence Site of of relative to watershed by debris associated record avalanches debris avalanche No . No ./km2/yr forested areas

H . J. Andrews Experimental Forest, western Cascade Range I Oregonl/ Forest 49.8 25 9 1 10 0.008 . 1.0 Clearcut 12.4 25 5 6 11 .036 4.5 Road 2.0 25 -17 --- -17 .340 42.0 64.2 31 7 38

Alder Creek drainage, western Cascade Range, Oregon Forest 12 .3 90 5 1 6 .005 . 1.0 Clearcut 4.5 15 2 1 3 .044 8.8 Road .6 15. 6 -- 6 .667 133 .4 , 17.4 13 2 15

1/Frederick J. Swanson, unpublished data, on file at Forestry Sciences Laboratory, USDA Forest Service Pacific Northwest Forest and Range Experiment Station, Corvallis, Oregon. individuals or groups of trees probable initiating cause of are knocked down (fig. 9). In debris avalanches in coastal the Pacific Northwest and along Alaska (Swanston 1967) and the the North Pacific coast, wind- central Oregon Coast Ranges, and throw is an important initiator it is believed to be'a contributor of soil mass movement and, ex- to debris-avalanche development cluding soil mass-movement activ- in the western Cascades (Swanson ity, is probably the most signifi- and Dyrness 1975). cant natural phenomenon providing organic materials to the stream Windthrow adjacent to channels system. adds large organic debris directly to them and may open large sections of the stream to direct sunlight. Large organic debris in streams (logs and branches) may create debris dams, allowing gravel and fine sediment to accumulate from upstream, at least as long as the dam remains in place. Such dams may be removed during stormflows or, if the dam remains in place, may cause local flooding or alteration of the channel to bypass the obstruction. The exposure of the channel to direct sunlight may increase water temperatures in that section during the summer, and the lack of insulating canopy may decrease winter temperatures.

Figure 9.--Natural windthrow along Kook Creek, Chichagof Island, southeast Alaska.

Studies of windthrow in virgin forests of the Oregon Coast Ranges (Ruth and Yoder 1953) 1953) indicate it is most severe on areas with a high water table or very shallow soils. Windthrow on such forested slopes tends to open the mineral soil to direct ingress of water, destroys the anchoring and reinforcing effect of tree roots on unstable sites, and--if windthrow occurs on a section of slope at or near INSECTS AND DISEASE saturation--may initiate soil mass movement directly through The killing of trees by insects impact and instantaneous pore- and disease affects the forest and water pressure development ultimately the stream environment (Swanston 1967). Windthrow has much the same as windthrow. Inter- been identified directly as the ception and evapotranspiration are

16 reduced or stopped, but infiltra- 25 years. Peak flows were 27 per- tion is usually not affected cent higher. The actual impacts (Anderson et al. 1976). Because on the channel system are unknown. of the loss of overhead cover near stream channels, water temperature may be locally increased and water chemistry altered temporarily by leaf fall and accumulation of organic vegetation in channels. The weakened root systems in dead trees also make the affected timber much more susceptible to windthrow (with its accompanying soil dis- turbance and accumulation of large organic debris in channels). Although no direct data are available, the destruction of viable root systems on unstable slopes by these epidemics would greatly reduce the slopes' ability to resist failure when soil water content is high and in periods of stormflow. Forest Service data (USDA FREEZING AND ICE FORMATION Forest Service 1958, 1975) indicate that insects kill far more In northern latitudes, freez- trees than disease does. Damage ing temperatures and development of from insect and disease infesta- ice on hillslopes and within stream tion is scattered. Mortality in channels may substantially reduce any given area is usually con- rate of streamflow and increase fined to a single tree species sediment contributions from bare and not all the individuals of hillslope areas under certain that species are infested. In conditions. If the channel freezes mixed forest types, this has over, ice formation and subsequent practically no impact on erosion melting and breakup may result in and streamflow because living flooding and extensive bank and forest cover and its attendant channel erosion by mechanical plow- root systems are maintained. ing and formation of "anchor ice. '' Under those conditions, the only damage to the stream environment In areas where extensive might be from death of streamside channel-ice formation is rare, trees, which would open the freezing temperatures tend to have channel to direct sunlight and their greatest impact on accelerated increase organic debris loading. discharge of surface sediment to the channel. Occasionally, large areas do become infested, and substantial "Concrete frost"--wet soil portions of the timber cover are solidly frozen--probably occurs destroyed. On the White River only sporadically on forested drainage in Colorado (Love 1955, slopes in the interior areas of Bethlahmy 1974), bark beetles the Northwest (Anderson et al. 1976). killed most of the trees in a In the Pacific coastal mountains, 1 974-km2 area with a resultant it is generally absent. Where increase in total streamflow of present, it may prevent infiltration about 22 percent over the next and cause local overland flow; its

17 frequency is so low, however, that downstream transport of bottom it probably has very little effect materials. on water in the channel. Much more important in terms of soil In small streams used by freezing effects is the develop- anadromous fish in the arctic and ment of "needle ice." Needle ice subarctic, such ice formation is produced by the growth of frost diverts a substantial volume of crystals beneath pebbles and soil water from winter streamflow and particles on unvegetated slopes results in rapid transport of during diurnal cycles of freezing channel sediment downstream, Kane and thawing (Sharpe 1960) . The and Slaughter (1973) have estimated particles are lifted perpendicular that winter icing of Gold Stream-- to the slope surface: when the a stream near Fairbanks, Alaska, needle ice begins to melt, the ice that is used by anadromous fish-- crystals and their load of earth, accounts for nearly 40 percent of pebbles, and organic debris fall the winter streamflow, downslope and may continue to slide and roll for some distance. The breakup of ice cover in Such "surface creep" is an im- the spring generally follows portant local contributor to ablation of the seasonal snowpack sediment transport from bare when rising water in the channel mineral soil areas to stream causes cracking of the ice from channels throughout the mountain- vertical hydrostatic pressures, ous areas of Western North America, and the resultant blocks and plates of ice are carried downstream as In areas where extensive ice floes. The movement of these channel ice is formed, freezing floes is intermittent and jerky, results in supercooling of the resulting in periodic damming and water, nucleation of "frazil ice" flooding of low areas near the particles (spicules and thin channel, extensive gouging and plates of ice formed in super- mechanical erosion of the channel cooled, turbulent waters) and banks, and sedimentation and formation of anchor ice around redistribution of bottom gravels, stones and gravel particles along the channel bottom (Michel 1973, 0 a Gillfilain et al. 1973). Static * 0 . ice begins to form along the & stream banks in areas of non- * turbulent flow. Through accumulation of frazil ice along the rough streamward edges of the static ice, slush and ice flows eventually form a continuous ice cover. 'Anchor ice forms along the channel bottom from the accumulation of frazil ice particles on the rough surfaces of coarse bottom sediments and on the lee sides of pebbles, rocks, and boulders. During ice formation, anchor ice frequently breaks loose from the bottom and is carried to the surface or downstream, with gravel and coarse bottom sediments still adhering. The result may be an FIR E extensive redistribution and Burning of forests destroys the covering vegetation on slopes

18 and along stream channels and may locally alter the physical properties of the surface layers of soil. The immediate impact is to increase both water yield and stormflow discharge from the watershed. Fire also exposes the bare mineral soil to increased surface runoff and erosion, Surface runoff from burned areas generally increases dissolved nutrient transport and loading of the channel systems. Surface erosion processes--including raindrop and rill erosion and dry ravel during wetting/ drying and freezing/thawing cycles--transport large volumes of sediment and debris from the watershed slopes to the channel (fig. 10). From 1 to 5 years afterward, fire may increase the potential for accelerated landslide activity through decay of anchoring and reinforcing root systems (Bishop and Stevens 1964, Swanston 1974, Ziemer and Swanston 1977.

Intensive drying of soil, Figure 10.--Dry ravel, or dry creep combustion of organic matter and sliding, of surface materials that binds soil aggregates, loss downslope after a fire along the of litter cover, and strong con- Klamath River in northern California. vective winds produced by the fire’s heat all contribute to on increased rill, sheet, and debris movement down steep slopes soil mass-movement erosion commonly during hot fires. In steep observed after intense fire has not terrain, rolling rocks and logs been well documented, Hot ground released by burning of roots and fire can reduce water storage other supportive organic matter capacity of surface organic matter trigger downslope movement of (Dyrness et al. 1957). Reduced additional material, greatly interception and evapotranspira- increasing concentration of tion may result in decreased sum- large woody debris on slopes and mer drawdown of soil water by vege- within stream channels .3/ tation (Klock and Helvey 1976a) , Hydrophobic (water repelling) although sometimes, as in heath soils have been reported in a vegetation, water loss can increase great variety of ecosystems when the soil surface is exposed after wildfire and slash fires (C. H. Grimingham, Univ. of (DeBano 1969, DeByle 1973, Megahan Aberdeen, personal communication). .and Molitor 1975, Dyrness 1976, Effects of reduced interception Campbell et al. 1977). The rela- and evapotranspiration may be tive importance of this phenomenon offset in part by increased overland flow in response to reduced 3/Manuscript in preparation, “Fire infiltration from loss of litter and geomorphological processes , by layer, development of hydrophobic F. J. Swanson. USDA For. Serv., Pac. soil, compaction by raindrops, Northwest For. and Range Exp. Stn., plugging of pores by fine soil Portland , Oreg.

19 material, and sometimes actual erosion depends on the intensity, fusing of the soil surface severity, and frequency of burning (Dyrness et al. 1957; Ahlgren and how much of a particular water- and Ahlgren 1960; Brown 1972: shed burned (Anderson et al. 1976). Helvey 1972, 1973; Rice 1973; If much foliage is destroyed,. Anderson et al. 1976; Campbell interception and evapotranspiration et al. 1977). In general, these are reduced, and the potential for factors lead to increases in both accelerated landslide activity soil water-storage and runoff from from root deterioration is in- burned sites. creased. Where the organic layers of the forest floor are also con- Contrasts in snow hydrology sumed and mineral soil is exposed, of burned and unburned ecosystems infiltration and soil water-storage capacities are reduced, greatly have received little study, par- increasing surface erosion and ticularly in terms of fire-in- duced changes in groundwater runoff potential. 3) regime (see footnote . Snow Massive wildfires in western accumulation and melt in open forests have accelerated both ero- (clearcut or naturally treeless) and forested areas have been the sion and sedimentation as a result subject of extensive research; of these processes. Studies of a stand of blackened snags pre- fire-denuded watersheds in west- central Washington (Klock and sents a vastly different environ- Helvey 1976a and b) showed that ment from either forested or maximum streamflows were double treeless areas. Speculation on the rate of flows before the fire. snow hydrology of burned areas is In addition, as the result of complicated by the great con- combined rapid snowmelt, high trasts between cold/dry and warm/ intensity rainstorms, and the wet snow types and between snow- destruction of covering and pack and multiple accumulation/ anchoring vegetation, massive melt seasonal regimes. Work on debris torrents occurred 2 years warm snowpacks by Smith (1974) and after the burn with a frequency others does suggest, however, that 10 to 28 times greater than before formation of melt zones around the fire. Noble and Lundeen (1971) blackened snags and rapid-conden- report a postfire erosion rate of sation melting may add more melt- 413.3 m3/km2 per year for a water to the soil in burned areas portion of the South Fork Salmon than in forests and snag-free open River in the Idaho Batholith. areas. Forests may have greater This amounts to an acceleration loss by evaporation; snowpacks in seven times greater than sediment open areas may contain continuous, yield for similar, but unburned, relatively impermeable horizons lands in the vicinity (Megahan and that carry meltwater directly to Molitor 1975) streams. . In northern California, Wallis Destruction of covering and Anderson (1965) reported sedi- vegetation by fire opens the ment discharge 2.3 times greater channel to direct sunlight, in- from burned than from unburned areaS. creasing water temperatures and Seventeen years after the Tillamook perhaps producing substantial Burn in the Wilson River watershed increases in organic debris of Oregon, the annual rate of sedi- loading from falling limbs, ment discharge was 175.3 t/km2, litter, and trees. The increase 5-8 times that of nearby unburned in stormflow and acceleration of forested watersheds with similar geology (Anderson 1954) .

20 Fire affects nutrient avail- CONCLUSIONS ability and subsequent nutrient loading of streams in several The natural events described ways. Nutrients incorporated in may operate separately or in com- vegetation, litter, and soil can bination to create limiting habi- be volatilized during pyrolysis tat characteristics in a particular and combustion, mineralized dur- stream section or system. Human ing oxidation, or lost by ash activities in the stream and its convection (Grier 1975). After parent watershed may profoundly the fire is out, nutrients can affect these events, their fre- then be redistributed by leaching quency, and magnitude. A firm of the ash layer and soil and understanding of the natural transported to the stream by processes is thus essential for surface erosion, soil mass move- a clear understanding of the ment, or solution transport. effects of forest and rangeland Studies on nutrient transfer to management on the habitat of anad- streams after wildfire and con- romous salmonids. trolled fire in the western Cas- cades of Oregon (R. L. Fredriksen, Forestry Sciences Laboratory, Corvallis, personal communication) and in north-central Washington (Tiedemann 1973, Grier 1975) , indicate that after a fire, free nitrogen concentrations were lower than in nonburned areas, nitrate increased, and phosphorus may have increased slightly. In western Oregon, Fredriksen (personal communication) reported that pH changes generally last less than 1 month, increased phosphorus concentrations no more than 2 years, and increased nitrate concentrations from 1 to 10 years. The impact of these changes in nutrient concentration on salmonid productivity is not well LITERATURE CITED known. The levels of increased nutrients reported in streams Ahlgren, I. F., and G. E. Ahlgren.

after fire appear to be below 1960. Ecological effects of ' toxic thresholds for aquatic forest fires. Bot. Rev. 26(4): organisms and dissipate rapidly 483-533. with stream dilution and flushing. Anderson, H. Gibbons and Salo (1973) 1975. Relative contributions of pointed out that the addition of sediment from source areas and nutrients to a stream may be bene- transport processes. p. 66-73. ficial, especially to relatively In Present and Prospective sterile streams, by supporting Technology for Predicting additional plant and animal life; Sediment Yields and Sources. such results remain difficult to Proc. Sediment Yield Workshop, predict, however, and excessive USDA Sediment Lab, [Oxford Miss., nutrient loading may result in Nov. 28-30, 1972] U.S. Agric. eutrophication. Res. Serv. Rep. ARS-S-40.

21 Anderson, H. W. California. M.S. thesis, Pa. 1954. Suspended sediment dis- State Univ., University Park. charge as related to stream- 151 p. flow, topography, soil and land-use. Trans. Am. Geophys. Corps of Engineers. Union 35 (2):268-281. 1956. Snow hydrology: North Pacific Div., Corps of Engineers, Anderson, H. W., M. D. Hoover, U.S. Army, Portland, Oreg. 437 p. and K. G. Reinhart. 1976. Forests and water: De Bano, L. F. effects of forest management 1969. Observations on water repel- on floods, sedimentation and lent soils in the Western United water supply. USDA For. Serv. States. In Symp. on Water Repel- Gen. Tech. Rep. PSW-18, 115 p. lent Soils. Univ. Calif., Pac. Southwest For. and Range Riverside. p. 17-30. Exp. Stn., Berkeley, Calif. De Byle, N. V. Bailey, R. G. 1973. Broadcast burning of log- 1971. Landslide hazards related ging residue and water repellency to land use planning on the of soils: Northwest Sci. 47(3) : Teton National Forest, north- 77-87. west Wyoming. USDA For. Serv. , Intermt. Reg., Ogden, Utah. Dunford, E. G., and C. H. Niederhof. 131 p. 1944. Influence of aspen, young lodgepole pine and open grassland Bethlahmy, N. upon factors affecting water 1974. More streamflow after a yield. J. For. 42(9):673-677. bark beetle epidemic. J. Hydrol. 23 (3/4) : 185-189. Dyrness, C. T. 1967. Mass soil movement in the Bishop, D. M., and M. E. Stevens. H. J. Andrews Experimental Forest. 1964. Landslides on logged USDA For. Serv. Res. Pap. PNW-42, areas in southeast Alaska. 12 p. Pac. Northwest For. and USDA For. Serv. Res. Pap. Range Exp. Stn., Portland, Oreg. NOR-1, 18 p. Northern For. Exp. Stn., Juneau, Alaska. Dyrness, C. T. 1976. Effects of wildfire on soil Brown, J. A. H. wettability in the high Cascades 1972. Hydrologic effects of a of Oregon: USDA For. Serv. Res. brushfire in a catchment in Pap. PNW-202, 18 p. Pac. North- southeastern New South Wales. west For. and Range Exp. Stn., J. Hydrol. 15(1) :77-96. Portland, Oreg.

Campbell, R. E., M. B. Baker, Jr., Dyrness, C. T., C. T. Youngberg, P. F. Ffolliott, F. R. Larson, and and R. H. Ruth. C. C. Avery. 1957. Some effects of logging 1977. Wildfire effects on a and slash burning on physical ponderosa pine ecosystem: an soil properties in the Corvallis Arizona case study. USDA For. watershed. USDA For. Serv. Res. Serv. Res. Pap. RM-191, 12 p. Pap. PNW-19. 15 p. Pac. North- Rocky Mountain For. and Range west For. and Range Exp. Stn., Exp. Stn., Fort Collins, Colo. Portland, Oreg.

Colman, S. M. Fiksdal, A. J. 1973. The history of mass move- 1974. A landslide survey of the ment processes in the Redwood Stequaleho Creek Watershed. Creek Basin, Humboldt County, Supplement to final report, FRI-UW-7404. Fish. Res. Inst., Univ. Wash., Seattle. 8 p.

22 Fredriksen, R. L. USDA For. Serv. Res. Pap. INT-97, 1963. A case history of a mud 34 p. Intermt. For. and Range and rock slide on an experi- Exp. Stn., Ogden, Utah. mental watershed. USDA For. Serv. Res. Note PNW-1, 4 p. Goodell, B. C. Pac. Northwest For. and Range 1959. Management of forest stands Exp. Stn., Portland, Oreg. in Western United States to influence the flow of snow-fed Fredriksen, R. L. streams. p. 49-58. In Int. 1965. Christmas storm damage Assoc. Sci. Hydrol. Publ. 48. on the H. J. Andrews Experi- Gentbrugge, Belgium. mental Forest. USDA For. Serv. Res. Note PNW-29, 11 p. Pac. Greswell, S., D. Heller, and Northwest For. and Range Exp. D. N. Swanston. Stn., Portland, Oreg. 1979. Mass movement response to forest management in the central Fredriksen, R. L., D. G. Moore, Oregon Coast Range. USDA For. and L. A. Norris. Serv. Resour. Bull. PNW-84, 26 p., 1975. The impact of timber Pac. Northwest For. and Range Exp. harvest, fertilization and Stn., Portland, Oreg. herbicide treatment on stream- water quality in western Oregon Grier, C. C. and Washington. p. 283-311. 1975. Wildfire effects on nutrient In Forest Soils and Forest Land distribution and leaching in a Management, Proc. Fourth North coniferous ecosystem. Can. J. Am. For. Soils Conf., Les For. Res. 5 (4): 599 -607 . Presses de L'Univ. Laval, Quebec . Hall, J. D., and C. 0. Baker. 1975. Biological impacts of Froelich, H. A. organic debris in Pacific 1973. Natural and man-caused Northwest streams. p. 1-13. slash in headwater streams. In Logging Debris in Streams Loggers Handb. Vol. 33. 8 p. Workshop. Oreg. State Univ., Pacific Logging Congress, Corvallis. Portland, Oreg. Harr, R. D. Gibbons, D. R., and E. 0. Salo. 1976. Hydrology of small forest 1973, An annotated bibliography streams in western Oregon. USDA of the effects of logging on For. Serv. Gen. Tech. Rep. PNW-55, fish of the Western United 15 p. Pac. Northwest For. and States and Canada: USDA For. Range Exp. Stn., Portland, Oreg. Serv. Gen. Tech. Rep. PNW-10, 145 p. Pac. Northwest For. and Harr, R. D. Range Exp. Stn., Portland, Oreg. 1977. Water flux in soil and subsoil on a steep forested Gillfilain, R. E., W. L. Kline, slope. J. Hydrol. 33(1):37-58. T. E. Osterkamp, and C. S. Benson. 1973. Ice formation in a small Harr, R. D., R. L. Fredriksen, Alaska stream. p. 505-573. and J. Rothacher. In The Role of Snow and Ice in 1979. Changes in streamflow. Hydrology. Proc. Banff Symp., USDA For. Serv. Res, Pap. PNW-249, Int. Assoc. Sci. Hydrol. Publ. 21 p. Pac. Northwest For. and 107. Gentbrugge, Belgium. Range Exp. Stn., Portland, Oreg.

Gonsior, M. J., and R. B. Gardiner. Hart, G. 1971. Investigation of slope 1963. Snow and frost conditions failure in the Idaho Ratholith. in New Hampshire, under hardwoods and pine and in the open. J. For. 61(4) :287-289.

23 Helvey, J. D. wildfire in north central Wash- 1972. First year effects of ington. p. 91-98. Proc. wildfire on water yield and Third Inter-Agency Sediment Conf. stream temperature in north- Symp. 1, Sediment Yield and Sources.\ central Washington. p. 308- Water Resour. Comm. 317. In Watersheds in Transition. Proc. Am. Water Klock, G. O., and J. D. Helvey. Resour. Assoc.. 1976b. Soil water trends following wildfire on the Entiat Helvey, J. D. Experimental Forest. In Proc. 1973. Watershed behavior after Tall Timbers Fire Ecology Conf. forest fire in Washington. 15:193-200. p. 403-422. In Proc. Irrig. and Drain. Div., Spec. Conf., Love, L. D. Am. Soc. Civ. Eng. 1955. The effect of streamflow on the killing of spruce and Hewlett, J. D. pine by the Englemann spruce 1967. A hydrologic response map beetle. Trans. Am. Geophys. for the state of Georgia. Union 36 (1): 113-118 Water Resour. Bull. 3 (3): 4 -20. Lutz, H. J., and F. S. Griswell. Hewlett, J. D.,and W. L. Nutter. 1939. The influence of tree 1970. The varying source area roots on soil morphology. Am. J. of streamflow from upland Sci. 237 (5): 389 -400 . basins. p. 65-68. In Proc. Symp. Interdisciplinary Aspects Meehan, W. R. of Watershed Manage. Am. Soc. 1974. The forest ecosystem of Eng., New York. southeast Alaska. 3. Fish habitats. USDA For. Serv. Gen. Holmes, W. H. Tech. Rep. PNW-15, 41 p. Pac. 1893. Vestiges of early man in Northwest For. and Range Exp. Minnesota. Am. Geol. 2(4): Stn., Portland, Oreg. 219-240. Megahan, W. F., and D. E. Molitor. Hoover, M. D., and C. R. Hursh. 1975. Erosional effects of 1943. Influence of topography wildfire and logging in Idaho. and soil depth on runoff from p. 423-445. In Watershed Manage. forest land. Trans. Amer. Symp., Am. Soc. Civ. Eng., Logan, Geophys. Union 25(2):693-698. Utah. Kane, D. N., and C. W. Slaughter. Michel, B. 1973. Seasonal regime and 1973. Properties and processes hydrological significance of of river and lake ice. p. 454- stream icings in central Alaska. 481. In The Role of Snow and p. 528-540. In The Role of Ice in Hydrology. Proc. Banff Ice and Snow in Hydrology. Symp., Int. Assoc. Hydrol. Sci. Proc. Banff Symp., Int. Assoc. Publ. 107. Gentbrugge, Belgium. Sci. Hydrol. Publ. 107. Gentbrugge, Belgium. Morrison, P. H. 1975. Ecological and geomor- Kelsey, H. M. phological consequences of mass 1978. Earthflows in the movements in the Alder Creek Franciscan Melange, California. Watershed and implications for Geology 6 (6): 361 -365. forest land management. B.A. thesis, Univ. Oreg., Eugene. Klock, G. O., and J. D. Helvey. 102 p. 1976a. Debris flows following

24 Narver, D. W. Rothacher, J., C. T. Dyrness, and 1971. Effects of logging debris R. L. Fredriksen. on fish production. p. 100-111. 1967. Hydrologic and related In J. T. Krygier and J. D. Hall characteristics of three small (eds.), Proc. Symp. Forest watersheds in the Oregon Cascades. Land Uses and Stream Environ- USDA For. Serv. 54 p. Pac. ment. Oreg. State Univ., Northwest For. and Range Exp. Corvallis. Stn., Portland, Oreg.

Nobel, E. L., and L. J. Lundeen. Rowe, P. B., and T. M. Hendrix. 1971. Analysis and rehabilita- 1951. Interception of rain and tion treatment alternatives snow by second-growth ponderosa for sediment control. p. 86- pine. Trans. Am. Geophys. 95. In J. T. Krygier and Union 32(6) :903-908. J. D. Hall (eds.), Proc. Symp. Forest Land Uses and Stream Envi- Ruth, R. H., and R. A. Yoder. ronment. Oreg. State Univ. 1953. Reducing wind damage in the forests of the Oregon Coast O'Loughlin, C. L. Range. USDA For. Serv. Res. 1972. An investigation of the Pap. PNW-7, 29 p. Pac. Northwest stability of steepland forest For. and Range Exp. Stn., Portland, soils in the Coast Mountains of Oreg . Southwest British Columbia. Ph.D. thesis, Univ. B. C., Sartz, R. S., and G. R. Trimble. Vancouver. 147 p. 1956. Snow storage and melt in a northern hardwoods forest. J. Patton, P. C., and V. R. Baker. For. 54(8) :499-502. 1976. Morphometry and floods in small drainage basins sub- Shaler, N. S. ject to diverse hydrogeomorphic 1891. The origin and nature of controls. Water Resour. Res. soils. 12th Annu. Rep., U.S. 12(5) :941-952. Geol. Surv. 1890-91, Part I, Geology. p. 213-345. Rice, R. M. 1973. The hydrology of chaparral Sharpe, C. F. S. watersheds. p. 27-33. In 1960. Landslides and related Proc. Symp., Living with phenomena. Pageant Books, Inc. Chaparral. Sierra Club, N. J. 137 p. Riverside, California. Smith, J. L. Rothacher, J. 1974. Hydrology of warm snowpacks 1959. How much debris down the and their effects upon water drainage? The Timberman 60(6): delivery ...some new concepts. 75-76. p. 76-89. In Adv. Concepts and Techniques in the Study of Snow Rothacher, J. and Ice Resour. Natl. Acad. Sci., 1963. New precipitation under Washington, D.C. a Douglas-fir forest. For. Sci. 9 (4): 423-429. Smith, J. L., and H. G. Halverson. 1969. Hydrology of snow profiles Rothacher, J. outlined with the profiling snow 197'3. Does harvest in west slope gage. p. 41-48. In Proc. 37th Douglas-fir increase peakflow Annu. West. Snow conf. in small forest streams? USDA For. Serv. Res. Pap. PNW-163, Stephens, E. P. 13 p. Pac. Northwest For. and 1956. The uprooting of trees: A Range Exp. S'tn., Portland, Oreg. forest process. Soil Sci. Soc. Am. Proc . 20 (1): 113-116.

25 Sutherland-Brown, A. Swanston, D. N. 1968. Geology of the Queen 1974. Slope stability problems Charlotte Islands, British associated with timber harvesting Columbia. B.C. Dep. Mines in mountainous regions of the and Pet. Resour. Bull. 54, western United States. USDA For. 226 p. Serv. Gen. Tech. Rep. PNW-21, 14 p. Pac. Northwest For. and Range Exp. Swanson, F. J., and C. T. Dyrness. Stn., Portland, Oreg. 1975. Impact of clearcutting and road construction on soil Swanston, D. N., and F. J. Swanson. erosion by landslides in the 1976. Timber harvesting, mass western Cascade Range, Oregon. erosion and steepland forest Geology 3:393-396. geomorphology in the Pacific Northwest. p. 199-220. In Swanson, F. J., G. W. Lienkaemper, D. R. Coates (ed.) , Geomorphology and J. R. Sedell. and Engineering. Hutchinson and 1976. History, physical effects Ross, Inc., Stroudsburg, Pa. and management implications of large organic debris in western Tiedemann, A. R. Oregon streams. USDA For. Serv. 1973. Stream chemistry following Gen. Tech. Rep. PNW-56, 15 p. a forest fire and urea fertiliza- Pac. Northwest For. and Range tion in north central Washington. Exp. Stn., Portland, Oreg. USDA For. Serv. Res. Note PNW-203, 20 p. Pac. Northwest For. and Swanston, D. N. Range Exp. Stn., Portland, Oreg. 1967. Debris avalaching in thin soils derived from bedrock. Triska, F. J., and J. R. Sedell. USDA For. Serv. Res. Note PNW-64, 1975. Accumulation and processing 7 p. Pac. Northwest For. and of fine organic ‘debris. In Log- Range Exp. Stn., Portland, Oreg. ging Debris in Streams Workshop. Oreg. State Univ., Corvallis. 6 p. Swanston, D. N. of Agriculture, 1969. Mass wasting in coastal U.S. Department ’ Alaska. USDA For. Serv. Res. Forest Service. Pap. PNW-83, 15 p. Pac. 1958. Timber resources for Northwest For. and Range Exp. America’s future. USDA For. Serv. Stn., Portland, Oreg. Resour. Rep. 14, Washington, D.C. 713 p. Swanston, D. N. U.S. Department of Agriculture, 1970. Mechanics of debris Forest Service. avalanching in shallow till 1975. The Nation’s renewable soils of southeast Alaska. resources - an assessment. USDA For. Serv. Res. Pap. U.S. Dep. Agric., Washington, D.C. PNW-103, 17 p. Pac. Northwest 385 p. For. and Range Exp. Stn., Portland, Oreg. Van Hise, C. R. 1904. A treatise on metamorphism. Swanston, D. N. U.S. Dep. Int. Geol. Surv. Monogr. 1971. Principal mass movement 47 p. processes influenced by logging, road building and fire. p. 29-40. Waananen, A. O., D. D. Harris, In J. T. Krygier and J. D. Hall and R. C. Williams. (eds.), Proc. Symp. Forest Land 1971. Floods of December 1964 Uses and Stream Environment. Oreg. and January 1965 in the far State Univ., Corvallis. Western states; Part I. Descrip- tion. U.S. Dep. Int. Geol. Surv. Water-Supply Pap. 1866-A. 265 p.

26 Wallis, J. R., and H. W. Anderson. 1965. An application of multi- variate analysis to sediment network design. Int. Assoc. Sci. Hydrol. Publ. 67, p. 357- 378 . Ziemer, R. R., and D. N. Swanston. 1977. Root strength changes after logging in southeast Alaska. USDA For. Serv. Res. Note PNW-306. 10 p. Pac. Northwest For. and Range Exp. Stn., Portland, Oreg.

27 The FOREST SERVICE of the U.S. Department of Agriculture is dedicated to the principle of multiple use management of the Nation’s forest resources for sustained yields of wood, water, forage, wildlife, and recreation. Through forestry research, cooperation with the States and private forest owners, and management of the National Forests and National Grasslands, it strives - as directed by Congress - to provide increasingly greater service to a growing Nation. The U.S. Department of Agriculture is an Equal Opportunity Employer. Applicants for all Department programs will be given equal consideration without regard to age, race, color, sex, religion, or national origin.