Paper No. 220-11
Presentation Time: 10:50 AM
ADVANCING LANDSLIDE SUSCEPTIBILITY AND RISK MAPPING THROUGH FEMA HAZARD MITIGATION PROJECTS IN EASTERN KENTUCKY (Invited Presentation)
CRAWFORD, Matthew, Kentucky Geological Survey, University of Kentucky, 228 Mining and Mineral Resources Bldg., Lexington, KY 40506-0107, DORTCH, Jason M., Kentucky Geological Survey, University of Kentucky, 228 Mining and Mineral Resources Building, 310 Columbia Ave, Lexington, KY 40506-0107, KOCH, Hudson, University of Kentucky, Kentucky Geological Survey, 228 Mining and Minerals Resources Bldg., 310 Columbia Ave, Lexington, KY 40506 and HANEBERG, William, Kentucky Geological Survey, University of Kentucky, 504 Rose Street, Lexington, KY 40506
FEMA Hazard Mitigation Assistance (HMA) grant programs aim to reduce losses and establish long-term solutions to costly hazards, making them a valuable tool for government entities, state geological surveys, and universities developing innovative research or mitigation plans. In 2021, the Kentucky Geological Survey (KGS) completed a FEMA Pre-Disaster Mitigation grant-funded assessment of landslide susceptibility and risk for five counties in eastern Kentucky. We based this landslide susceptibility model on statistics of hillslope morphological features such as slope, curvature, roughness, aspect, and elevation. Using a dual machine learning approach, we modeled the probability of landslide occurrence, defined as an area that is occupied, or might be occupied in the future, by a landslide. We also developed the risk assessment using a static, socioeconomic approach that includes landslide effects on population, roads, railroads, buildings, and land class. incorporated hazard (susceptibility results), vulnerability, and consequences to produce each risk map.
In early 2022, KGS received approval for a new FEMA HMA grant program project assessing landslide susceptibility and risk for additional eastern Kentucky counties. A key component of the new project will be addressing the most effective ways to improve model accuracy; therefore, we have posed several questions. What additional inputs may contribute to increased model accuracy? Is there a ceiling on model accuracy? Does independent landslide inventory mapping help to constrain a reasonable threshold on model accuracy, whereby spending significant time for higher accuracy is unnecessary? We will assess additional landslide variables such as topographic wetness, flow accumulation, root strength, proximity to roads and streams, geology, and soil erodibility that were not in the initial modeling. We aim to build on our modeling approaches that balance innovative techniques and statistical improvement with the need for practical hazard mitigation and safety solutions for communities.