Modeling slope stability uncertainty: a case study at the H.J. Andrews Long-Term Ecological Research Site, Oregon

A slope stability model employing Digital ElevationModel (DEM) data was used in a Geographic Information System(GIS) to predict landslide susceptibility in a large region of the H.J.Andrews Long-Term Ecological Research (LTER) site, located in theWestern Cascades of Oregon. To assess the uncertainty of the finaloutput, several different, but equally probable, versions of the inputDEM were created through the addition of random, spatiallyautocorrelated noise (error) files. The DEMs were then processed toproduce a family of slope stability maps from which the uncertaintyeffects of DEM elevation error upon the final susceptibility indexcould be assessed. The slope stability model itself involved theoverlay of rock strengths with slope classes to establish landslidesusceptibility indices. The inherent nature of error in DEMs used formodeling slope stability, coupled with much of the Pacific Northwest'scommercial timber being located in mountainous terrain, makes thisan intriguing problem to solve. The ability to understand the resultantuncertainty due to elevation error when applied to the model has thepotential to facilitate improved natural resource managementdecisions in relation to harvesting and subsequent slope stability.