Authors: Bridgette Ingram Fritz*, University of Tennessee, Liem Tran, University of Tennessee, Sally Horn, University of Tennessee, Larry McKay, University of Tennessee
Topics: Paleoenvironmental Change, Geographic Information Science and Systems, Geomorphology
Keywords: geomorphon, paleofloods, sinkhole detection, GIS
Session Type: Paper
Start / End Time: 1:20 PM / 3:00 PM
Room: Napoleon B2, Sheraton 3rd Floor
Presentation File: No File Uploaded
This study is part of a larger paleoflood desktop survey for the Upper Tennessee River Basin in which we employ hydraulic modeling, multi-level landscape/landform classification, and geographic information systems (GIS) overlay analysis to develop a desktop model which predicts the presence of Holocene-age paleoflood deposits in different geomorphic settings. For this particular study, we focus on sinkholes which serve as repositories for paleoflood deposits. We first applied the landform classification method geomorphon using US Geological Survey digital elevation models at 1m and 10m resolutions. The geomorphon method, which is based on the principle of pattern recognition, delineates the ten most common landforms present across our study area (e.g., ridge, shoulder, spur, hollow, etc.). Next, we extracted the pit landforms from the geomorphon classification. Then, we applied a multi-criteria algorithm that evaluates the geological setting, spatial arrangement, and geometric ratios of these pits to decide whether they are true sinkholes or not. We calibrated the algorithm by comparing classified sinkholes with mapped sinkholes on USGS topographic maps as well as with sinkholes detected from aerial imagery. We will conduct field work in the next phase of the desktop survey to explore the sinkholes with high potential for the presence of paleoflood deposits.