GIS/EM4 - Ecological change under pressure of urban development

Landslide and debris flow influences on aquatic habitat conditions

GIS/EM4 No. 35

Michael G. Wing

Abstract

Landslides and debris flows may play an important role in fostering quality aquatic habitat for salmonid and other fish species. To provide a coarse scale examination of this theory, a geographic information system (GIS) is used to process landslide and debris flow models on a digital elevation model (DEM) from a coastal Oregon sub-basin. Using a spatial database that contains aquatic habitat information from sub-basin streams, model results are examined to evaluate whether relationship exists between the spatial distribution of debris flows and the amount of large woody debris, large boulders, and gravel in streams. Results indicated that greater numbers of large boulders existed in streams that were influenced by debris flows than in streams not influenced by debris flows. No statistically significant differences were detected in large woody debris and gravel measurements in streams that were influenced by debris flows than in streams not influenced by debris flows. These findings encourage additional research into landslide and debris flow influences on aquatic habitat conditions using finer resolution spatial data.

Keywords

Landslides, debris flows, aquatic habitat, spatial modeling, geographic information systems.

Introduction

Landslides are mass movements of debris, soil, or rock down hill and are typically triggered by large amounts of rainfall or alterations to land cover. Debris flows can quickly form from shallow landslides in steep terrain and occur when landslides liquefy and continue to move down hill. Landslides and debris flows in forested areas have received increasing attention in the U.S. Pacific Northwest as several large landslide-producing storms have occurred within the past five years. The impacts of these events can drastically alter surrounding terrestrial and aquatic natural resources and can threaten human structures and safety.
The processes and conditions related to landslides in forested areas are multifaceted. The impacts of timber harvesting, including vegetation removal, soil compaction, and road building, have been associated with increased landslide occurrence (Swanson and Dryness 1975, Swanston and Swanson 1976). Within forest operations, activities associated with roads have the largest impact in altering slope stability on a unit area basis (Sidle et al. 1985). Other prominent factors related to landslides include precipitation, geology, vegetation, soil characteristics, and slope gradient (Wu and Sidle 1995).
The short-term consequences of landslides and resulting debris flows on channel networks are generally considered to negatively affect aquatic habitat conditions, particularly in smaller streams (Swanson et al. 1987). The intensity of debris flow impacts varies in response to changes in channel gradient, surrounding landforms, and the relative location of a channel in a stream network (Fannin and Rollerson 1993). Debris flows in high gradient streams may scour the channel to its bedrock and remove riparian vegetation and large woody debris (Swanson et al. 1998). These events may result in simplified channel characteristics and elevated sediment loads. Deposition of materials carried by debris flows typically occurs in lower gradient channels or at tributary junctions and can greatly influence subsequent channel morphology and streamside vegetation (Benda and Cundy 1990).
Despite the recognition of the immediate negative influences of debris flows on stream channel characteristics, debris flow impacts may provide several long-term benefits in terms of aquatic habitat quality. The inputs of large wood, gravel, boulders, and flood plain sediment from debris flows over broad time scales may provide an important ingredient in maintaining productive aquatic habitat conditions (Reeves et al. 1995, Bisson et al. 1997). The presence of these materials may play a critical role in creating complex instream habitat by creating debris jams that can modify high flows, create pools, and lead to additional areas where spawning gravel can accumulate.
The objective of this study is to provide a coarse scale examination of landslide and debris flow influences on selected aquatic habitat variables in a forested setting. To this end, spatial landslide and debris flow models that rely on elevation data were used to identify potentially high risk land slide areas and to create debris flow paths in a coastal Oregon sub-basin. The debris flow path results were used to identify streams in a spatial aquatic habitat database that were potentially subject to debris flow influences. A Wilcoxon rank-sum test was used to test whether the amount of large woody debris, large boulders, and gravel in streams influenced by debris flows differed from streams not influenced by debris flows.

Study Area

The Siletz sub-basin is located in the western Oregon Coast Range and occupies approximately 195,000 ha (Figure 1). Land cover in the Siletz sub-basin is predominately forested by coniferous species (Douglas-fir, western hemlock, and lodgepole pine) but red alder presence is abundant in some riparian and upslope areas. Land owners include federal, tribal, industrial forest, and private non-industrial entities. The highly dissected mountainous topography and sandstone- interspersed by siltstone- bedrock of the Siletz sub-basin is typical of the Oregon coast range. Elevation in the Siletz sub-basin ranges from sea level to 1048 m and slopes vary between 0 and 49 degrees. Precipitation averages about 1500 mm annually.

Figure 1. Shaded relief map, surveyed stream reaches, and location of the Siletz sub-basin, Oregon, USA.

Methods

A 30-meter resolution digital elevation model (DEM) was created for the Siletz sub-basin from USGS elevation data using an ArcInfo GIS. A shallow landslide model (Montgomery and Dietrich 1994) was used to identify areas in the DEM that were of greatest risk for landslides. The shallow landslide model draws upon the TOPOG hydrologic model (O'Loughlin 1986) which calculates upslope contributing areas in determining shallow subsurface flow convergence.
A debris flow routing algorithm was applied to the DEM and used the areas identified as high risk by the shallow landslide model as starting points for debris flows. The algorithm created down slope debris flow routes based on DEM slope gradients. The routing process terminated when slope gradients no longer exceeded 3.5 degrees. Previous research has suggested that debris flows often stop in stream channels or valley bottoms at this threshold (Benda and Cundy 1990). The debris flow routing results were converted into an ArcInfo raster-based GRID.
The GRID representing debris flow routes was overlaid on a 1:100,000 scale stream network cover containing aquatic habitat data to identify the sub-population of stream reaches potentially impacted by debris flows. The aquatic habitat data was collected by field crews following protocols developed by the Oregon Department of Fish and Wildlife (ODFW) for Oregon streams (Moore et al. 1998). An ArcInfo GIS was used to associate the aquatic habitat data with the stream network to create a vector coverage of aquatic habitat conditions for surveyed stream reaches (Wing and Skaugset 1998). Within the Siletz sub-basin, aquatic habitat data were collected from 298 stream reaches representing 524 km of stream length (Figure 1).
Mean values were generated for both sets of stream reaches for three aquatic habitat variables: LWD volume per 100 m of stream length, number of large boulders (greater than .5 m in diameter) per 100 m of stream length, and percent of gravel in each stream reach. A nonparametric statistical procedure, the Wilcoxon rank-sum test, was used to test whether significant differences existed between the habitat variables in debris flow impacted stream reaches versus non-impacted stream reaches. The Wilcoxon rank-sum test was used as each of the habitat variables was non-normally distributed and this statistical test is relatively robust when normality assumptions are not met.

Results

Approximately 0.3 percent (5,082 out of 1,760,606 cells) of the total area in the 30 m DEM was classified as potentially unstable or landslide prone according to the landslide model results (Figure 2). The debris flow routing and subsequent overlay process identified 94 out of the 298 stream reaches that were surveyed for aquatic habitat data as being potentially impacted by debris flows (Figure 3).

Figure 2. Location of potentially unstable GRID cells.

Figure 3. Debris flow locations and impacted streams.

Means for the each of the aquatic habitat variables were lower for the debris-flow influenced streams reaches than in the non-influenced stream reaches with the exception of the number of large boulders per 100 m of stream length (Table 1).

Debris Flow Stream Influence	Mean Aquatic Habitat Values
Debris Flow Stream Influence	Large Woody Debris	Large Boulders	Percent Gravel
Absent	19.19	12.88	31.01
Present	17.86	55.70	29.45

Table 1. Mean values of large woody debris, large boulders, and percent gravel.

To test whether the central tendencies of the habitat variables were different in debris flow influenced than non-debris flow influenced stream reaches, variable distributions were examined using the Wilcoxon rank-sum test. Statistically significant differences were detected in the sum of the ranks for the number of large boulders per 100 m of stream length but not for large woody debris volume or percent gravel (Table 2). We can infer from these results that there were significantly more large boulders in stream reaches that were influenced by debris flows than in stream reaches that were not.

	Aquatic Habitat Variables
	Large Woody Debris	Large Boulders	Percent Gravel
p-value	0.57	0.01	0.65

Table 2. Wilcoxon rank-sum results for differences in aquatic habitat
variable distributions between debris flow influenced and non-debris flow
influenced stream reaches.

Discussion

Support for the contribution of debris flows to quality aquatic habitat conditions was present in the significantly higher number of large boulders found in debris flow impacted stream reaches than in non-impacted stream reaches. Support for the contribution of debris flows to large woody debris and gravel quantities was absent, however, given the insignificant differences detected in the variable distributions.
The absence of support for debris flow influences on large woody debris and gravel quantities was unexpected but some explanations are possible. The influences of debris flows on aquatic habitat are thought to be functions of broad time scales. In contrast to large boulders, the large woody debris and gravel data used in this study may have been inadequate in providing broad time scale metrics of aquatic habitat conditions. Of the variables that were tested, large boulders may possess a longer instream residence time than either large woody debris or gravel since boulders are more resilient to decay than large woody debris and are not as easily transported as gravel.
In addition, the DEM used in this study had a ground resolution of 30-meters and was likely unable to accurately capture fine scale landscape elevation changes. The landslide and debris flow models used in this study relied considerably on DEM data. Using a finer resolution DEM to more accurately represent topography may have produced different modeling results.
This study did find that large boulder counts were higher in debris flow influenced stream reaches. This positive result encourages further investigation into the role of landslide and debris flow influences on aquatic habitat conditions. One of the challenges for researchers in this area will be to develop spatial databases that accurately describe aquatic habitat conditions over broad temporal scales. Future efforts may also want to extend the analysis scale to include multiple sub-basins in order to assess regional landslide and debris flow influences on aquatic habitat characteristics.

References

Benda, L. E. and T. W. Cundy. 1990. Predicting deposition of debris flows in mountain channels. Canadian Geotechnical Journal 27:409-417.

Bisson, P.A., G.H. Reeves, R.E. Bilby, and R.J. Naiman. 1997. Watershed management and Pacific salmon: desired future conditions. In: Pacific Salmon and Their Ecosystems: Status and Future Options. D.J. Stouder, P.A. Bisson, and R.J. Naiman, Editors. Chapman & Hall, p. 447-474.

Fannin, R.J. and T.P. Rollerson. 1993. Debris flows: some physical characteristics and behavior. Canadian Geotechnical Journal 30:71-81.

Montgomery, D.R. and W.E. Dietrich, 1994. A physically based model for the topographic control on shallow landsliding. Water Resources Research 30:1153-1171.

Moore, K., K. Jones, and J. Dambacher. 1998. Aquatic Inventory Project: methods for stream habitat surveys. Oregon Department of Fish and Wildlife, Research and Development Section, Corvallis, Oregon.

O'Loughlin, E.M. 1986. Prediction of surface saturation zones in natural catchments by topographic analysis. Water Resources Research 22: 794-804.

Reeves, G.H., L.E. Benda, K.M. Burnett, P.A. Bisson, and J.R. Sedell. 1995. A disturbance-based ecosystem approach to maintaining and restoring freshwater habitats of evolutionarily significant units of anadromous salmonids in the Pacific Northwest. American Fisheries Society Symposium 17:334-349.

Sidle, R.C., A.J. Pearce, and C.L. O'Loughlin. 1985. Hillslope stability and land use. American Geophysical Union, Water Resources Monograph No. 11, 140 p.

Swanson, F.J. and C.T. Dryness. 1975. Impact of clear-cutting and road construction of soil erosion by landslides in the western Cascade Range, Oregon. Geology 3:393-396.

Swanson, F.J., S.L. Johnson, S.V. Gregory, and S.A. Acker. 1998. Flood disturbance in a forested mountain landscape. BioScience 48(9):681-689.

Swanson, F.J., L.E. Benda, S.H. Duncan, G.E. Grant, W.F. Megahan, L.M. Reid, and R.R. Ziemer. 1987. Mass failures and other processes of sediment production in Pacific Northwest forest landscapes. In: Streamside management: forestry and fishery interactions. E. O. Salo and T. W. Cundy, Editors. University of Washington, Institute of Forest Resources, Contribution No. 57, p. 9-38.

Swanston, D.N. and Swanson, F.J. 1976. Timber harvesting, mass erosion, and steepland forest geomorphology in the Pacific Northwest. In: Geomorphology and Engineering. Coates, D. R., Editor. Dowden, Hutchinson, and Ross, Inc., p. 199-221.

Wing, M. and A. Skaugset. 1998. GIS casts a line: Examining salmon habitat in Oregon streams. Geo Info Systems 8(7): 36-41.

Wu, W. and R.C. Sidle. 1995. A distributed slope stability model for steep, forested basins. Water Resources Research. 31(8): 2097-2110.

Author

Michael G. Wing, Assistant Professor, Forest Engineering Department
Oregon State University, 215 Peavy Hall, Corvallis, Oregon, USA 97331.
Email: michael.wing@orst.edu, Tel: +1-541-737-4009, Fax: +1-541-737-4316.