Modeling landslide susceptibility from a DTM in Nelson County, Virginia: A remote sensing-GIS approach

This dissertation attempts to model landslide susceptibility in central Nelson County, Virginia, from a DTM using remote sensing and GIS techniques, and to examine the influence of the DTM resolution on the modeling. Landslide paths were interpreted from aerial photographs. Once verified in the field, the landslide locations were digitized and subsequently rasterized. A DTM of the study area was built from elevation data in a 7.5-minute quadrangle. Slope orientation, gradient, and slope morphometry were analyzed from the DTM using computer programs written in C.;The GIS overlay analysis revealed the areas affected by landsliding. Statistical analysis of the overlay result provided an insight into the topographic variables that were statistically significant to the development of landsliding and the topographic conditions conducive to landsliding. A GIS model encompassing the significant variables was constructed.;The results indicate that the spatial distribution of landslides in the study area is statistically dependent upon slope orientation, configuration, elevation, and slope shape at resolutions 72 m and finer. NE-, N-, NW-, and W-facing slopes are more susceptible to landsliding than E-, SE-, S-, and SW-facing slopes. Slopes that are concave in cross section are much more vulnerable to landsliding than other shapes. Landsliding potential varies little with slope profile characteristics. In cross section, however, the potential is much larger for concave than for linear or convex slopes. Slope length does not appear to affect the potential for landsliding. Terrain with a gradient smaller than 31;The spatial pattern of landsliding potential modeled at 24 m and 48 m resolutions are strikingly similar. However, dramatic changes are introduced to the result at the 72 m resolution. Thus, 48 m is the minimum acceptable resolution for the modeling. The integration of remote sensing and GIS, combined with statistical analysis and testing, allowed landsliding potential in the study area to be modeled at a high level of accuracy using topographic variables alone. (Abstract shortened with permission of author.)...