All Science Fair Projects

Causes of Landslides in the Walla Walla River Watershed and the Effects on Salmonid Health and Habitat

Shauna H. Nyborg

2003

Abstract: The Blue Mountain region, located in the southeastern corner of Washington and northeastern corner of Oregon contains a wide variety of natural resources. Land use activities include agriculture, livestock production, mining, outdoor recreation, and timber harvesting. These activities oftentimes destabilize hillslope soils. Eastern Washington’s dry climate results in soils that are particularly susceptible to failure in intense rainfall events, producing extensive landslides which negatively affect aquatic ecosystems. Throughout 2002, approximately 250 possible landslides within two drainages in the Walla Walla River Watershed were inventoried and ground-truthed by the Washington State Department of Natural Resources to determine their causes. The data reveal that overall, the largest contributors to stream sedimentation are channel-related slides (59%) and natural slides(24.8%). Human-induced destabilization (road building, harvesting, thinning, etc.) account for a total of 13.6% of the landslides. It should be noted that a certain amount of sediment in streams in virtual equilibrium is “normal” considering the maintenance of stream health, but excessive sedimentation and influx of large woody debris may cause the decline of fish in watersheds. Dwindling populations salmonid species indicate that poor water quality and habitat degradation are issues in the Blue Mountains and that the causes are in part related to land use activities.

INTRODUCTION

A landslide generally refers to “the movement of a mass of rock, debris or earth down a slope” (Cruden in Turner and Schuster 1996). The term encompasses a wide variety of mass movement types from debris flows to rock falls. A classification scheme was developed by Varnes (1978) in which a landslide can be categorized and described by two nouns. The first noun indicates the material and the second noun describes the type of movement (e.g. rock fall, debris flow, etc.). Types of material consist of rock, debris, and earth. Movements range from falls (the most rapid), topples, slides, spreads, to flows (the slowest).

Landslides vary in type, as well as in the nature of the material, and can occur within a range of velocities. Planar landslides are typically shallow and fail along a planar surface. Deep-seated slides involve the backward rotation of a block along a curved failure surface. Road-related slides occur due to the undercutting of slopes during road construction or concentration of water in poorly consolidated road fill. Channel-related landslides involve those slides confined to a stream channel. Factors that trigger landslides include change in slope, increased load, vibrations or shocks, change in soil water content, frost action, weathering of rocks, groundwater movement, reduction in lateral support, or removal or change in type of vegetation covering slopes.

Landslides can be triggered by a number of mechanisms, broadly divided into four categories: geological (jointed, weak materials, etc.), morphological (tectonic uplift, wave erosion, etc.), physical (intense rainfall, earthquake, etc.), and human-induced (logging, mining, etc.) (Turner and Schuster 1996). Ultimately, these variables affect the stability of a slope. Factors involved in slope stability include: slope angle, position and form, soil characteristics, bedrock type, groundwater, root strength, elevation, road placement, and stand age (Abramson et al. 1996).

Most landslides in Washington State occur as a result of a combination of poor forest management, intense rainfall, seismic events, and volcanic eruptions. The Cascade Range and the created rain-shadow separate Washington into two distinct climatic regions: the rainy western side and dry eastern side. Despite the lack of precipitation, eastern Washington’s landscape has been significantly shaped by rain and snow. The 1996 rain-on-snow event is the most recent example of how dramatically the hillsides can respond to intense rainfall. In the foothills near Walla Walla, debris flows exceeded 100 failures every 2.5 km2 (Harp 1997). Over the years, landslides in the Walla Walla River Watershed have affected agricultural and residential property, forest ecosystems, and aquatic ecosystems.

Slides are of particular ecological importance with regard to their role in supplying sediment to streams. Excessive landsliding into streams can increase sediment load and detrimentally effect salmonid health and habitat. Because many of the land-use activities (timber harvesting, road construction, etc.) in the Blue Mountains lead to slope destabilization, the stream systems are particularly vulnerable to increased sedimentation.

The Washington State Department of Natural Resources (WDNR) inventoried possible landslide areas in the South Fork Touchet River and Wolf Creek drainage basins with the goal of ground-truthing them to obtain further information into their causes and other characteristics. This study focuses on examining what activities are responsible for landslides in these two drainages and the subsequent effect increased sedimentation from the slides likely have on salmonid habitat and survival.

THE WALLA WALLA RIVER WATERSHED

Physiographic Setting

The Walla Walla River Watershed occupies 4550 km2 in southeastern Washington and northeastern Oregon (Fig. 1). The portion of the watershed located in Washington State is 3310 km2 and bounded by the Columbia River on the west, the Blue Mountains on the east, the state line to the south, and the Snake River Basin to the north (Economic and Engineering Services, Inc. 2002). The Washington section of the Walla Walla River Watershed includes three main subbasins: Mill Creek, Touchet River, and Walla Walla River.

Figure 1. Location map of Walla Walla River Watershed (after Economic and Engineering Services, Inc. 2002)

Climate

The climate of the Walla Walla River Watershed varies from warm and semiarid in the western regions of the Columbia Basin, to cooler and wetter conditions in the Blue Mountains. The study area was located in the marine zone of the Blues which is characterized by a relatively cool, moist climate (Jaindl and Quigley 1996). Average temperatures range from -7 to -4° C in the winter to 32 to 35° C in the summer, with the summer highs peaking in July and decreasing in late August. With regards to precipitation in the basin, the lower west end averages less than 25 cm of rain per year while the higher east end of the basin averages 100-152 cm of both rain and snow per year (Economic and Engineering Services, Inc. 2002).

Geology

The bedrock in the Walla Walla River Watershed is Columbia River Basalt (CRB) up to 3 km thick. These basalts were generated for over three million years beginning 17 million years ago in the Miocene forming an extensive basalt plateau that covers over 155400 km2. Approximately 10 million years ago, the area was folded, generating an anticline which became the Blue Mountains (Jaindl and Quigley 1996).

In this section of the Blues, watershed topography reflects the lithologic characteristics of the underlying jointed CRB bedrock and the anticline. Canyons of the main streams run north-south, and plateau tops are generally inclined to the north. Canyon walls average 300 m high and have been eroded to relatively uniform and steep slopes averaging in gradient between 45% and 70% (South Fork Touchet Watershed Analysis 1997). Tributaries of the main streams are typically steep, first or second order drainages. The valley bottoms are generally straight and narrow. Rapid shifting of the valley streams results in little development of alluvial fans or tributary drainages (South Fork Touchet Watershed Analysis 1997).

In the Touchet-Wolf Creek study area, the maximum elevation is about 1300 m along the ridgeline between the Wolf Fork and North Fork Touchet River drainages, while a minimum elevation is about 600 m in the South Fork Touchet River drainage.
The soils of the Blue Mountains unconformably overlie the CRBs and older rocks and consist of Pleistocene to Holocene-aged alluvium, colluvium, fanglomerates, terrace gravels, volcanic ash, and loess (Walker 1997). In the Touchet-Wolf Creek study area, soils of primarily loess and volcanic ash cover the bedrock. The ash is from the eruption of Mount Mazama approximately 6,800 B.P. and ranges up to 30 cm within the loess soil (South Fork Touchet Watershed Analysis 1997).

Landslides in the watershed

Being that the soils vary from sandy loam with volcanic ash and clays, to gravelly loam with prevalent volcanic ash, their compositions make them naturally quite permeable but quasi-stable (Clifton 2002). However, their high permeability causes them to be especially prone to landsliding in periods of intense rainfall. Review of the Umatilla National Forest Soil Resource Inventory, completed in 1978, showed a low frequency of natural landslides and debris flows in average years of precipitation (Clifton 2002). However, during periods of record floods, the frequency of mass wasting was high (Clifton 2002).

In addition to high rainfall, the jointed and fractured nature of the bedrock also contributes to hillsides’ propensities to fail. Fractures, joints, bed attitudes, and faults control the movement and distribution of water and represent areas of weakness within the parent rock (Swanston 1978). Joints parallel to the dip of the slope create paths for concentrated water movement and transport it downslope where excess hydrostatic pressure develops at the slope’s base. Release of pressure may cause a landslide to occur (Swanston 1978).

Forest practices can affect hillslope stability by reducing root strength either by wood or root deterioration, site disturbance, change in vegetative cover, temporary increase in the water content of the soil, road construction, or reduction in the soil’s capacity to absorb precipitation due to compaction from use of heavy machinery. Landslide inventories reveal that slides are most frequent four to ten years after logging because this is when the slopes are least stable and vegetation is regenerating. After 20 years, the frequency of landslides returns to pre-harvesting levels (Sidle et al. 1985).

WALLA WALLA RIVER WATERSHED SALMONID SPECIES

The stream systems of the Blue Mountains are home to a number of fish species belonging to the family Salmonidae. Salmonid species are referred to as “cold water fish” because they prefer cooler water than do “warm water” fish such as catfish, bass, suckers, crappies, etc. (Moore and Flaherty 1996). Warm water fish species are not native to the Pacific Northwest and due to their less stringent environmental requirements, are becoming successful in replacing cold water species as aquatic habitats warm up due to human-induced alterations (Moore and Flaherty 1996).

Salmonids include fish that are both anadromous and resident. Anadromous fish spend most of their adult life in saltwater, but return to freshwater where they reproduce, and where juveniles spend anywhere from several months to a few years before heading to the ocean to further develop (Moore and Flaherty 1996). Resident species spend their entire lives in freshwater. The salmonids in the Walla Walla River Watershed include: Chinook salmon (Oncorhynchus tshawytscha) re-introduced in 2000 to the Walla Walla River, non-native brown trout (Salmo trutta), native steelhead/rainbow trout (Salmo gairdneri), and native bull trout (Salvelinus confluentus).

Within the Blue Mountains, bull trout populations have noticeably declined. They have been referred to as indicator species for poor water quality. Like other salmonid species, they require cool, well-oxygenated water, a clean gravel substrate, and tree cover and shade. Bull trout populations are so low in the Walla Walla River that they were listed as a Threatened Species by the U.S. Fish and Wildlife Service in June of 1998. The Walla Walla steelhead trout were also listed as a Threatened Species by the National Marine Fisheries Service in 1999. Being particularly sensitive to water temperature and stream sedimentation, their decline suggests that streams are being damaged or polluted to a greater extent by human activities than nature would do alone.

METHODS

A landslide inventory of the Touchet subbasin of the Walla Walla River Watershed was completed in 2002 by the WDNR. Using a combination of 1:24,000 scale orthophotos from 1996 and aerial photographs, nearly 250 possible landslides were located in the total 93 km2 (234 were accessed by road or trail; gates, road closures, or other obstacles prevented ground-truthing of the remaining 15 or so). The sites were then transferred to a topographic map to be used for field reconnaissance. Landslides are recognized on aerial photos and orthophotos by their hummocky terrain, light-colored material, lobe shape, etc. Each landslide was given a code depending on which specific drainage basin it was associated with. Slope angle, slide dimensions, and type (planar, deep seated, etc.) were estimated on the orthophotos, and then ground-truthing determined the actual characteristics.

The slides were located in the South Fork Touchet and Wolf Creek drainages located about seven miles southeast of Dayton, Washington. Twenty-five landslides of the accessible 234 were ground-truthed in the months of October and November, 2002, using the landslide-highlighted topographic map and a GPS receiver. The other 209 landslides had been ground-truthed in the months prior.

At each landslide site a number of details were recorded which accompanied the identifying subbasin code. First, was the feature an actual landslide? Sometimes “landslides” in orthophotos turned out to be areas of aspen trees or low-lying brush. Second, was the landslide deep-seated or planar? Third, the width and length were estimated using a range finder, and depth was roughly estimated by sight. Fourth, was the shape of the slope concave or convex? Fifth, the gradient of slope on which the landslide occurred was measured with a clinometer. Sixth, location of the landslide (latitude and longitude) and elevation were determined using the GPS receiver. Lastly, the probable cause of the landslide was noted. Causes were classified as either road-related (oversteepened slopes or unstable road fill), harvest-related (clearcut slopes or skid trails), channel (occurring in a channel), natural (naturally occurring on a slope without obvious human inducement), or other feature (if they were very old landslides or not slides at all). Digital pictures also were taken for later use in comparing different sites. All the data collected were then entered into a spreadsheet to be compared to other landslides previously ground-truthed by other WDNR employees.

RESULTS

Figure 2 shows the percentage distribution of the 234 landslides according to associated land use in the Touchet-Wolf Creek study. The greatest percentage of landslides (59%) occurred in stream channels. Natural landslides—slides resulting from intense rainfall for example—accounted for 24.8% of the landslides. Of human-induced slumps, roads were the largest cause at 11%. The landslides related to timber harvesting totaled 2.6%. Other feature slides amounted to 2.6%.

In Figure 3, the slope gradient is compared to the total percent of slides. The greatest number of slides (49%) occur on slopes between 41-50° and the least number (1%) of slides are on slopes 11-20°. Roughly 4% of slumps occur on slopes 1-10°. About 17% of landslides are on slopes 31-40°, 25% on slopes 51-60°, and 3% on slopes 61-70°. Approximately 4% occur on very steep slopes ranging from 71-80°.

Figure 2. Distribution of mass wasting features and association with land use.

Figure 3. Slope gradient and percent total slides

Table 1 shows causes and the distribution of landslides according to the slope gradient. The greatest number of road, channel, natural, and other features landslides occurred on slopes between 41-50°. Timber-related slumps mostly occurred on slopes 31-40°. Timber, channel, and natural landslides had little or no failures below 31°, and road, timber, and channel, landslides did not occur on slopes greater than 60°. Other features spanned a range of gradients from flat to very steep. Natural landslides had the greatest concentration of failures above 41°, whereas road-related slides had the greatest concentration of failures below 41°.

DISCUSSION

The Touchet-Wolf Creek data reflect that of a related study (Fig. 4). A 2002 Phase II Flood Assessment and mass wasting inventory study completed by Caty Clifton of the Umatilla National Forest Service in Pendleton, Oregon evaluated the stream systems of the Umatilla River, Tiger Canyon, Mill Creek, and Tucannon River (hereafter the “Umatilla study”) of the Blue Mountains to determine the cause and frequency of mass wasting features triggered by the intense rainfall during 1995-1996. Only road, timber, and channel-related slides were studied. Of the 66 inventoried and ground-truthed landslides, 68% were channel-related, 21% were caused by road failures and 9% were associated with timber harvest areas.

Figure 4. Distribution of mass wasting features and association with land use (after Clifton 2002)

Both the Touchet-Wolf Creek data and the Umatilla data reflect that channel-related debris flows are the greatest causes of sedimentation in streams. The high number of channel landslides can possibly be explained by the fact that these slopes receive a concentrated quantity of precipitation which can lead to frequent failure. Many of the lower elevation, lower order streams originate in unforested, grassy areas. This can increase their propensity to fail due to the lack of soil strengthening tree roots and dense under-story cover. A factor not noted in channel-related slides was the presence or absence of clearcut slopes along the channel. That is, some of the stream-side slopes may in fact have been harvested for timber and could have been responsible for landslides, but that element was not considered because along the length of channels there are a relatively significant number of variables affecting slope stability. Slopes outside of channels and on typical hillsides can be treated as isolated cases having only one cause.

Natural landslides are the second largest group making up 24.8% of total slides. Natural landslides do not include channel slides because, as stated above, channel slides may have some land use component and natural slides do not necessarily occur strictly within channels. As Table 1 shows, only natural slides occur on the steepest slopes (71-80°). This could be explained by the fact that land use activities do not utilize such steep slopes, so seeing such landslides is not expected. A number of natural slides are located at the beginning of valleys where smaller drainages originate and the basalt bedrock is exposed creating steep headwalls. The majority of natural slides occur on slopes 41-50° probably as a result of soil saturation and destabilization in intense rainfall events.

Of the human-induced activities, road construction (typically built for timber harvesting) is the most significant contributor to stream sedimentation. In the Touchet-Wolf Creek study, 11% of slumps were road-related and 21% were related in the Umatilla study. Roads expose areas where erosion and mass wasting processes originate. Road construction oftentimes also involves using some sort of fill material that is less compacted then the surrounding soils. During heavy rains, the fill material may weaken then fail and slide down a slope. Additionally, roads oversteepen hillsides reducing lateral support, making them more susceptible to failure. Roads usually persist long after an area has been harvested thus allowing other activities to occur, such as off-roading, which can cause severe soil degradation. Level ground in valley bottoms are sometimes utilized for road construction which can have impacts on riparian zones. In the South Fork Touchet River there are at least four areas where the road crosses the stream, and at least two places where the road is built in riparian areas. Riparian areas are essential in buffering sediment transport to the channel, and without them, streams are left vulnerable to extreme sediment influxes (Moore and Flaherty 1996).

A number of slides are considered “other features” because they are either not actual landslides or are landslides, but that failed a very long time ago. From an orthophoto, these sites appear as hummocky ground—deceivingly similar to a landslide. Some sites are located on relatively flat ground, or are different colored rocks scattered across the hillside (such as basalt boulder outcrops), and on an orthophoto these features look like some sort of mass movement. These lower gradient (1-10°) non-landslides account for 24% of the total of “other features.” On a number of slopes, jumbled rocks and soil suggest that mass wasting has occurred, but decades ago given the presence of large shrubs and meter-plus high trees growing on the surface. These slumps give little indication to their cause, but are likely natural slides because of their locations and absence of land use activities at the probable time of failure.

Even though only 2.6% of the Touchet-Wolf Creek landslides appeared to be timber or clearcut related, clearcuts have a much more significant impact on landslide occurrence. In the slide-prone Mapleton area of the Oregon Coast Ranges, Swanson (1977) found that the frequency of landslides and soil transport were 1.9 and 4.0 times greater in clearcuts than in undisturbed forest. As clearcut patches increase in a given watershed, the cumulative effects will become more significant in riparian and aquatic ecosystems. This is particularly relevant considering that flood events and their intensities will increase as the amount of clearcutting increases (Moore and Flaherty 1996). The effect of clearcuts on the occurrence of landslides is also related to climate. In an area such as the Walla Walla River Watershed where the annual precipitation is relatively low, the total of clearcut-related landslides will likely be less than that of areas west of the Cascades where annual precipitation is greater. More precipitation increases water content in the soil intensifying its propensity to slide (Moore and Flaherty 1996).
The slope gradient graph (Fig. 3) gives an idea as to the steepness of the slopes on which landslides are occur. It needs to be noted that these data include the “other feature” gradients. So the relatively high occurrence of slumps on 1-10° slopes are likely “other features” consisting of flat brush, or something similar.

In general, slopes between 25-55° are the most prone to fail, and over 30° slopes are considered oversteepened (Abramson 1996). The data in Fig. 3 reflect this generalization because the majority of landslides are concentrated between 31° and 60°. Table 1 shows the relative distribution of landslides within each cause and it too indicates that the majority occur within 31-60° with the greatest number of landslides happening on slopes 41-50°. Roads caused slides on the greatest number of lower gradient slopes between 1-30° than other causes probably due to the failure of road fill material in heavy rains. Timber and channel-related landslides and natural slides barely occur on slopes 1-30° possibly because the slopes are simply not steep enough to initiate a failure or because they are heavily vegetated. Only natural landslides and other features occurred on slopes 61-70°. And even then, they accounted for less than 10% of the total slides for each cause.

Landslide effects on salmonids and streams

Landslide effects on fish and streams include: elevated water temperature, increased vulnerability to predators, decreased dissolved oxygen, damage to riparian habitat, decreased availability of food, too much or too little large woody debris, suspended sediment which can trap emerging fry and inhibit the removal of metabolic waste from the salmonid eggs, and blocked migration, among other things (Meehan and Swanston 1977, Swanson 1980). One or a combination of effects could occur in a stream depending on the extent of the sediment transported from a landslide.

Removal of streamside forest cover results in greater solar radiation reaching the stream which can increase the temperature of the water. Salmonids prefer water between 12 and 14°C. They avoid areas above 15°C and can die at temperatures reaching 24-25°C (British Columbia Ministry of Energy and Mines 2002). Elevated water temperatures affect the fish’s physiological functions. If temperatures are too warm, their metabolic rates may increase to the point at which their energy intake cannot maintain basic physiological functions. High temperatures can increase vulnerability to disease and susceptibility to toxins such as nickel and chlorine, disrupt metabolism, and reduce dissolved oxygen causing the fish to suffocate (Washington State Department of Ecology 2000). Increases in maximum stream temperatures from June to August ranging from 2 to 10°C have been noted in the Northwest after harvesting (Moore and Flaherty 1996).

High levels of suspended sediment (>300 ppm) can damage fragile fish gills by the accumulation of sediment on the gill filaments preventing the fish from being able to aerate their blood (Moore and Flaherty 1996). The fish may eventually die due to anoxemia (acute respiratory failure) and carbon dioxide retention (Meehan et al. 1977). Salmonids also feed by sight, so turbid water may inhibit their ability to find food (British Columbia Ministry of Energy and Mines 2002). Since suspended sediment can also decrease the amount of light and reduce the depth of photosynthesis, primary producers such as algae, may cease to produce oxygen. Fine sediment also can accumulate in spawning gravels and reduce the transport of oxygen to eggs and inhibit the removal of waste products that build up as the embryos develop (Meehan et al. 1977). The sediment in the gravels may also trap emerging fry by decreasing pore space between gravels (Phillips 1971).

On a larger scale, landslides may entirely block stream channels and prevent fish from migrating (Meehan et al. 1977). If riparian areas are damaged or reduced to a small fraction of their previous size, the fish could be more vulnerable to predation. Landslides also reshape stream channels and affect the movement and redistribution of spawning gravels and accessibility to potential spawning habitats (Swanson 1980).

The value of large woody debris (tree branches, tree trunks, and root wads) has been argued, but recent studies have shown that this material provide both food and a substrate for attachment for aquatic invertebrates which are in turn, food sources for fish (Meehan et al. 1997, Conway 1995). The debris also stabilizes stream channels and provides habitat by generating pool environments for salmonids and other organisms (Swanston 1980). Timber harvesting can both prevent the regular deposition of woody debris from entering the stream channels, and increase debris concentration to a detrimental effect. Large log jams can block fish passages up rivers as well as forcing the river to create a new channel around the jam causing extensive bank erosion (Swanston 1980).

It would be logical to consider landslides as having undeniably negative effects on stream systems. However, research done over the past ten years in California has shown that the densities and biomass of salmon in logged watersheds are the same as those of unlogged areas. Interestingly enough, the Coho salmon population fared better in the logged watershed than the population in the unlogged, control watershed of the same study (CPFC 1999). The salmon which fared the best were those in the watershed with small clear cuts and very narrow stream-protecting buffers. Studies done in the early 1970s by the California Department of Fish and Game (CDFG) showed that while some types of logging and road construction increased the fine sediment in spawning gravels, salmon density tended to increase after logging (CPFC 1999).

It might also be noted that although debris flows may seem disastrous immediately after they occur, they may be aiding in the long-term health of the stream as “system resetting” events, or events that revitalize the system (Benda in Trotter 2000). If a stream remains undisturbed for an extended period of time it is eventually depleted of bed material and the large woody debris necessary for fish habitat, and over time the fish populations will decline. New material transported from slopes replenishes the stream and creates new habitat which begins the cycle once again (Trotter 2000).

The fact that studies on the effects of land use activities and fish health vary in their conclusions could be explained by vague definitions of water quality, by use of different evaluation standards, and by differences in timing of sampling and measurement. Also, variables such as differences in site climate, soil, geology, hydrology, and vegetation may give conflicting results. Regardless, the decline in salmonid populations in the Walla Walla River Watershed indicates degradation of habitat is occurring.

CONCLUSION

Despite the fact that our data demonstrates that the majority of landslides are occurring in stream channels, land use activities also contributed to excess sedimentation in streams. Road building is the main culprit of human-induced landslides. Many of these roads have been constructed for purposes of hauling timber out of the forests. Ultimately, the effect of timber harvesting is difficult to quantify and is estimated only by comparing changes in stream habitat. Quantifying the change in fish populations is costly, difficult, and time-intensive (Moore and Flaherty 1996).

Salmonid populations are also affected by changing oceanic conditions, dams, agricultural practices, fish harvesting techniques and levels, pollution, predators, loss of habitat, hatchery practices, etc. Additionally, other human activities such as recreation, livestock grazing, farming, development, etc. have effects on fish habitats and water quality. Thus, forest practices, destabilization of slopes, and possible channel damage are not the only variables affecting the survival of salmonids in the Walla Walla River Watershed drainages. In order to maintain self-sustaining ecosystems, it is important to protect all aspects of those systems from the ridgetops to the valley floors.

ACKNOWLEDGEMENTS

I would like to thank Charles Chesney, forest practitioner for the Washington State Department of Natural Resources for taking me into the field and introducing me to “landslide hunting.” I want to also thank Caty Clifton, forest hydrologist for the U.S. Forest Service, for supplying her landslide inventory information. I would also like to express my thanks to Dr. Bruce E. Howlett for his invaluable assistance in completing my thesis. And to my advisor, Dr. Robert J. Carson for providing information about landslides and suggestions for my thesis.

LITERATURE CITED

Abramson, L.W., et al. 1996. Slope stability and stabilization methods. Wiley-Interscience, New York. 629 p.

Benda, L. E. 1994. Stochastic geomorphology in a humid mountain landscape. Ph.D. dissertation. In Headwater Fishes and their Uppermost Habitats: A Reivew as Background for Stream Typing (P.C. Trotter). Fish, Timber, and Wildlife, Olympia. 69 p.

British Columbia Ministry of Energy and Mines. Aggregate operators best management practices handbook for British Columbia. 2002. http://www.em.gov.bc.ca/Mining/ MiningStats/Aggregate%20BMP%20Handbook/Vol-2-AOH-BMP.pdf. (29 April 2003). 89p.

(CFPC) California Forest Products Commission. 1999. Forests and salmon. http://www.calforests.org/foundation/salmon_paper/toc.html (30 April 2003).

Clifton, C. 2002. Phase II flood assessment, mass wasting inventory summary report. Pendleton, Oregon. 9 p.

Conway, F.D.L. 1995. Forest practices and salmon: what’s the connection? Oregon State University Extension Service, Corvallis, Oregon: 2-4.

Cruden, D. M. and D. J. Varnes. 1996. Landslide types and processes. In Landslides
Investigation and Mitigation (A.K. Turner and R.L. Schuster, eds.), TRB, National Research Council, Washington D.C.: 36.

Economic and Engineering Services, Inc. 2002. Walla Walla watershed WRIA 32 level 1 assessment. http://www.co.walla-walla.wa.us/Departments/planning/Watershed plunit.htm (4 May 2003).

Harp, E. L., et al. 1997. Landslides and landslide hazards in Washington state due to February 5-9, 1996 storm. U.S. Geological Survey. http://landslides.usgs.gov/ html_files/nlic/ Wash_hrp/Wash_hrp.pdf (5 May 2003): 2.

Jaindl, R. G., T. M. Quigley, and R. Harrod. 1996. Forest health issues and resources of the Blue Mountains. Search for a Solution: Sustaining the Land, People, and Economy of the Blue Mountains (R.G. Jaindl and T.M. Quigley, eds.), American Forests, Washington, D.C.: 1-8.

Meehan, W. R. and D. N. Swanston. 1977. Effects of gravel morphology on fine sediment accumulation and survival of incubating salmon eggs. USDA For. Serv. Res. Pap. PNW-220. Pacific Northwest Forest and Range Experiment Station, Portland, Oregon. 16 p.

Moore, B. C. and L. Flaherty. 1996. Fish, riparian, and water quality issues in the Blue Mountains. Search for a Solution: Sustaining the Land, People, and Economy of the Blue Mountains (R.G. Jaindl and T.M. Quigley, eds.), American Forests, Washington, D.C.: 203-232.

Phillips, R.W. 1971. Effects of sediment on the gravel environment and fish production, in Forest Land Use and Stream Environment. Oregon State University, Corvallis: 64-74.

Sidle, R. C., A. J. Pearce, and C. L. O’Loughlin. 1985. Hillslope stability and land use. American Geophysical Union, Washington, D.C. 140 p.

South Fork Touchet watershed analysis, appendix a. 1997. Washington State Department of Natural Resources, Olympia. 14 p.

Swanston, D. N. 1978. Effect of geology on soil mass movement activity in the Pacific Northwest. Forest Service, U.S. Department of Agriculture.: 105, 107.

Swanston, D. N. 1980. Influence of forest and rangeland management on anadromous fish habitat in western North America. USDA For. Serv. Res. Pap. PNW-104. Pacific Northwest Forest and Range Experiment Station, Portland, Oregon. 27 p.

Swanson, F. J., M. M. Swanson, and C. Woods. 1977. Inventory of mass erosion in the Mapleton Ranger District, Suislaw National Forest. For. Sci. Lab, Corvallis, Oregon. 62 p.

Trotter, P. C. 2000. Headwater fishes and their uppermost habitats: a review as background for stream typing. Timber, Fish, and Wildlife, Olympia.: 47-48.

Turner, K. A. and R. L. Schuster, eds. 1996. Landslides: investigation and mitigation. TRB, National Research Council, Washington, D.C. 673 p.

Varnes, D. J. 1978. Slope movement types and processes. In Special Report 29: Landslides: Analysis and Control (R.L. Schuster and R.J. Krizek, eds.), TRB, National Research Council, Washington, D.C.: 11-33.

Walker, G. W. ed. 1990. Cenozoic geology of the Blue Mountain region. USGS Professional Paper 1437. U.S. Government Printing Office, Washington. 135 p.

Washington State Department of Ecology. 2000. Effects of elevated water temperatures on salmonids. http://www.ecy.wa.gov/pubs/0010046.pdf (29 April 2003). 2 p.

Weaver, W. E., D. K. Hagans, and J. H. Popenoe. 1995. Magnitude and causes of gully erosion in the Lower Redwood Creek Basin, northwestern California. In Nolan, K.M., H.M. Kelsey, and D.C. Marron, eds. Geomorphic Processes and Aquatic Habitat in the Redwood Creek Basin, Northwestern California. USGS Professional Paper 1454-A. United States Geological Survey.: 11