USING RADAR INTERFEROMETRY TO ASSESS MINE SUBSIDENCE IN CENTRAL OHIO

The Ohio Department of Natural Resources (ODNR) geological division estimates that roughly 4,300 abandoned mining sites exist today, covering an approximated area of 800mi² throughout 43 counties in Ohio; a result of greater than 200 years of mining. The risks associated with mine subsidence are well documented in Ohio: as of 2005 the Ohio Department of Transportation had spent an estimated $14.3 million to repair state and federal highway damage. For example, the Ohio Department of transportation has estimated that it will need to spend an additional $26,163,544 over 960 days on mine remediation below SR-2 near Toledo, which would in total cost $169,238,295 to motorists forced to take the 19 mile detour.

The advent of Differential Interferometric images derived from repeat-pass Synthetic Aperture Radar (SAR) technology provides a cost effective method for quantifying subsidence at a large scale (250km²), with extreme precision (<0.1mm). Also this technique has been successfully used to map deformation from mine subsidence.

We test the feasibility of measuring subsidence using the 3-Pass INSAR and Persistent Scatterer (PS) techniques to measure subsidence over mines in eastern Ohio. Results from interferometry are overlain over existing abandoned mine maps and documented subsidence locations This allows interferometry results to be analyzed in an integrated GIS environment and aids in understanding, identifying, and quantifying where subsidence has occurred. These are used in a model to help identify causes for subsidences (rapid or gradual), with the potential of identifying locations at high risk of future subsidences, developing a robust predictive model for determining where mine subsidence will occur in the future, and making recommendations to mitigate the areas of high and extreme risk hazard.

Relevant geologic variables included in the predictive model are, height of mined out area, width of unsupported mine roof, thickness of overburden, bedrock competency, pillar dimensions, fractures and/or jointing, and time. Thus far, a weighted raster multivariable model is used to model location of actual subsidence claims