| Abstract: | A landslide susceptibility model, employing a digital elevation model (DEM) and geological data, was used in a GIS to predict slope stability in a region of the H J Andrews Long-Term Research Forest, located in the Western Cascade Range in Oregon, USA. To evaluate the contribution of error in elevation to the uncertainty of the final output of the model, several different, but equally probable, versions of the input DEM were created through the addition of random, spatially autocorrelated noise (error) files. The realized DEMs were then processed to produce a family of slope stability maps from which the uncertainty effects of elevation error upon landslide susceptibility could be assessed. The ability to assess this uncertainty has the potential to help us better understand the inherent strengths and weaknesses of applying digital data and spatial information systems to this application, and to facilitate improved natural resource management decisions in relation to timber harvesting and slope stability problems. |
