COMPARISON OF NED, SRTM, AND LIDAR DATA IN KARST GIS FOR NORTHERN ALABAMA

Three elevation datasets were used to assess strengths and weaknesses of geospatial data and analytical methods of karst mapping in Madison County, northern Alabama. As the Highland Rim Physiographic Section in northern Alabama is the most heavily karstic region in the state, this is a priority area for karst mapping and research. Madison County, part of the Highland Rim, is underlain by Mississippian age carbonate units, including the Bangor, Monteagle, and Tuscumbia Limestones. These units in the county are associated with karst features including 460 moderate to large sinkholes and 547 caves.

For this karst GIS project, three data types were used: USGS NED 30-meter grid data; SRTM 30-meter grid data; and high resolution LiDAR point data. Noted data limitations include older dates of acquisition (NED and SRTM), lower spatial resolution (NED and SRTM), slower processing speeds (LiDAR), and higher noise levels (LiDAR). Alternatively, strengths of different datasets include potential for multitemporal change detection (NED vs. SRTM vs. LiDAR), high resolution (LiDAR), and more recent acquisition dates (SRTM and LiDAR).

Analytical methods incorporated hydrology tools (flow direction and sinks) and contouring tools in ArcGIS software. Data were used and resampled at multiple resolutions. Results were compared with USGS 1:24,000-scale topographic maps and 1-meter aerial photography. Spatial resolution had significant influence on accuracy of sinkhole identification. Resolutions less than 5 meters produced the highest noise levels, while resolutions at 10 to 30 meters more successfully identified sinkholes present on 1:24,000-scale topographic maps and aerial photos. That some sinkholes on topographic maps were not identified in the processed data may reflect multitemporal changes in land-cover or land-use. Although not included in this study due to time and budget constraints, ground truthing may be included in future research to help clarify apparent sinkhole absences and causes thereof.