PRELIMINARY EVALUATION OF OHIO STATEWIDE IMAGERY PROGRAM AIRBORNE LIDAR FOR ABANDONED UNDERGROUND COAL MINE DETECTION, MINERAL RIDGE AREA, TRUMBULL COUNTY, OHIO

Land subsidence and collapse in areas of abandoned underground coal mines pose a significant safety hazard throughout Appalachia. Prediction of areas prone to subsidence or collapse, however, is complicated by the fact that many abandoned mines are either unmapped or inadequately mapped. We used publicly available reconnaissance level (typical 6.6 ft or 2 m ground strike spacing) airborne LiDAR from the Ohio Statewide Imagery Program (OSIP) to evaluate its potential for the detection of subtle topographic anomalies associated with abandoned underground mines in a portion of the Mineral Ridge area about 14 km NW of Youngstown, Ohio. The area was selected because 1) it has documented mine subsidence problems, 2) a reliable mine map exists for a portion of the pilot area, and 3) it has a variety of land use and ground cover. Subsidence features observed in the field range from small pits 1 m to 2 m in diameter to broad swales tens of meters in length. Areas identified as closed depressions on smoothed 2-foot (0.6 m) LiDAR digital elevation models (DEMs) show a qualitative correspondence to known underground mine workings and, in some cases, subsidence problem areas. In other cases, subsidence problem areas did not correspond with calculated DEM depressions. Conversely, closed depressions are generally absent over sandstone bodies known as horsebacks, where coal is known to be absent. DEM smoothing, performed using a moving window median filter, was necessary in order to remove fine scale depressions, for example roadside ditches, unrelated to mine subsidence. We found moving windows between 11 by 11 and 21 by 21 rasters to be the most useful sizes for the delineation of inferred subsidence-related topographic anomalies in the pilot area. Subtraction of a smoothed DEM from a depressionless DEM created using standard hydrologic modeling algorithms seemed to provide more useful maps than identification of pits based on surface curvature criteria. Future work is anticipated to include the evaluation of more sophisticated spectral techniques such as wavelet transforms and more complete integration of available historical mine maps and modern geotechnical data.