REGIONAL LEDGE RECOGNITION FROM LIDAR IN STRATIGRAPHY FOR DETAILED GEOLOGIC MAPPING, PARTICULARLY IN THE LOWER ATOKA FORMATION OF CENTRAL AND NORTHWEST ARKANSAS

LiDAR imagery is invaluable to remote geologic mapping. LiDAR data are now widely available, and streamlining mapping approaches that utilize these data thus increasingly important. We introduce a tool that uses LiDAR data to quickly trace topographic ledges in layered geologic exposures, thereby helping to characterize regional stratigraphic variability. We apply this approach to the Pennsylvanian Atoka Formation, which is exposed across central and northwestern Arkansas. In this distal foreland basin region, the Atoka Formation exhibits laterally continuous, thin sandstone ledges that are interbedded throughout thicker units of shale as part of a transgressive-regressive marine shelf sequence. Over a century of hydrocarbon exploration has allowed geologists to develop a stratigraphic framework and detailed naming scheme for many of these Atoka Formation sandstones in the subsurface. However, the application of this naming scheme to surface geology is difficult due to vegetative cover. The Atoka Formation is instead commonly lumped into “Atoka undifferentiated” or simply upper, middle, and lower units in many geologic maps. LiDAR data allows the mapper to visualize these sandstone ledges in outcrop, and our tool assists the mapper in quickly identifying their stratal relations. We present two ledge-mapping methods and develop Python based routines for them. Both methods use topographic derivatives to create raster masks that highlight ledge-like topography. The first method creates a histogram from the masked elevations as a pseudo-stratigraphic column, and then applies gaussian mixture modelling to identify elevations that ledges appear most often within the area of interest. These elevations are then used to create contours that attempt to map the stratigraphy. The second method transforms the masked, ledge-like regions into polygons, and then creates a centerline through those polygons that match expected characteristics of long, narrow ledges. We provide examples of where these algorithms excel and struggle and conclude with the results from the Atoka Formation that capture the stratigraphic complexity and aide in the identification of regionally correlative sandstone intervals.