LANDSLIDE SUSCEPTIBILITY INDEX MODELING IN ALLEGHENY COUNTY, PENNSYLVANIA

The landscape of southwestern Pennsylvania exemplifies the interplay of surficial processes and underlying geologic structures. Deeply incised stream valleys are prevalent throughout much of the region and are usually controlled by structurally derived fracture lineaments in the sedimentary bedrock. Due to the deep, steep-walled nature of many of these valleys, and the historical land use and development in the region, slope failure is a relatively common occurrence. These mass-wasting events can range from small slumps on the order of a few cubic meters to large landslides up to thousands of cubic meters. One area that is particularly prone to these events is Allegheny County. With the city of Pittsburgh at its center, development in the county is pervasive and continues today. The purpose of this study is to assess landslide susceptibility for Allegheny County, Pennsylvania using a weighted overlay analysis method in GIS. Nine controlling factors of landslide occurrence are considered in the construction of an indexing model. The factors include: slope, aspect, lithology, land use, soil texture, precipitation, distance from streams, distance from roads, and distance from mapped fractures and faults. Parameters are weighted according to an impact score assigned based on its perceived significance for landslide occurrence. Due to subjectivity in weighting the parameters, a model calibration strategy is used to match existing landslide data to predicted high-susceptibility areas. An iterative procedure to compare landslide frequency measures with high susceptibility areas produces an error map, whereby a measure of difference indicates poor model performance. Parameter weights are manipulated to optimize the indexing model and reduce these differences. Completion of the landslide susceptibility map within the GIS framework allows for updates to current data and improvement to the model as conditions within the county change. Additionally, the indexing model output may be used for up-to-date analysis of susceptibility based on land use changes, informing stakeholders and decision-makers of the implications of those changes.