THE ROLE OF REMOTE SENSING AND GEOSPATIAL ANALYSIS IN THE DEVELOPMENT OF A LANDSLIDE INVENTORY

The Office of the State Geologist (OSG) has been conducting a state-wide landslide inventory across Arkansas since 2016. The inventory initially encompassed areas within Crawford, Franklin, Cleburne, and Van Buren Counties. Most recently, the OSG was awarded two Hazard Mitigation grants in 2021 for the purpose of: conducting a landslide inventory in Newton, Madison, and Washington Counties and updating the landslide portion of the All-Hazard Mitigation Plans for Arkansas and for Newton, Madison, and Washington Counties.

This landslide inventory study follows the methodology from Bulletin 82, Washington Geological Survey Streamlined Landslide Inventory Protocol with minor modifications. This method, a streamlined mapping approach using LiDAR, is an effective method to accurately map landslide landforms. The OSG utilized ArcGIS and the state-wide 2017 1-meter LiDAR to create a digital slope model on which polygons defining landslide features, were digitized. A landslide geodatabase was created that includes such elements as age, strata within which the landslide initiated, movement type, features observed, anthropogenic factors influencing movement, depth to slip plane, volume, spatial area, and other factors. Landslides are classified based on the ability to identify a distinct image of the headscarp, flank, toe and/or other internal features. The classifications range from a high of 40 (landslide features are readily identifiable) to a low of 10 (slopes have very little evidence of mass wasting).

Various geospatial analyses were run on these landslide polygons to: evaluate the ability of certain geospatial analyses to identify landslides, determine rate of movement of a landslide, determine the mean slope of landslide categories, and determine the spatial relationship of the landslides with respect to populated areas. Future research will expand these and other relationships to include slope susceptibility analyses and derivative maps.