LANDSLIDE-SUSCEPTIBILITY MAPPING AND RISK ASSESSMENT, EASTERN KENTUCKY

The Kentucky Geological Survey is assessing landslide susceptibility and risk for the Big Sandy Area Development District, a five-county area in eastern Kentucky. Landslide-susceptibility maps were developed from hillslope geomorphic variables from 1,054 landslides in one county. A dual machine-learning approach was used to select variables and create the maps. Bagged trees, a machine-learning random-forest classifier, was used to evaluate geomorphic variables, and 12 were identified as important: standard deviation of plan curvature, standard deviation of elevation, sum of plan curvature, minimum slope, mean plan curvature, range of elevation, sum of roughness, mean curvature, sum of curvature, mean roughness, minimum curvature, and standard deviation of curvature. These variables were further evaluated using logistic regression modeling to determine the probability of landslide occurrence and, in turn, to create a landslide-susceptibility map. The performance of the logistic-regression model was evaluated by the receiver operating characteristic curve, area under the curve, which was 0.83, indicating strong model performance. The model was validated on a separate area and then applied to the remainder of the project area. The resulting maps are divided into five classes of landslide susceptibility: low, low–moderate, moderate, moderate–high, and high. The risk assessments ranged from simple intersections of hazard and exposure data to a more comprehensive analysis of landslide susceptibility and exposures (i.e., population, roads, railroads, and buildings). The physical and monetary vulnerabilities of the exposures were also considered in the risk calculations. The results will be incorporated into the existing Big Sandy Area Development District mitigation plans and strategies as part of an inclusive planning process that gages stakeholder needs.