The Effect of Varying Soil and hydrologic conditions on GPR interpretations to detect unmarked graves in Colonial Williamsburg

Ground-Penetrating Radar (GPR) is a powerful geophysical tool commonly used to non-invasively locate grave shafts. Accurate interpretations of GPR radargrams, and the resulting depth slices, can be challenging due to differing conductivities in soil because of material composition and spatially variable soil moisture gradients. GPR transmits an electromagnetic wave into the soil which reflects off subsurface anomalies and returns to the receiver. The time it takes to do so is converted into a depth based on soil properties. Particularly conductive or insulative materials can attenuate the signal, influencing travel time and depth of penetration. This project aims to refine future gravesite detection protocol with implications extending to identifying subsurface features in general. First, GPR usage was explored in a survey of the P4 Parking Lot Cemetery in Colonial Williamsburg comparing depth slices produced from two different systems using different antenna frequencies. Later, 10x10 m and 10.5x9 m GPR transect and soil moisture grids were taken using a 200 MHz antenna and processed as radargrams and depth slices in EKKO_Project 6. Previously exposed (but never excavated) grave shaft locations were overlaid onto three slices representing different environmental conditions. Accuracy of each depth slice was approximated by calculating the distance between the centroid of the known grave shafts and centroid of the GPR-derived hotspots in ArcGIS. In the P4 Parking Lot Cemetery, higher soil moisture conditions produced more accurate results but only when considering resulting hotspots as the boundaries of the grave shafts and the void spaces between hotspots (anti-hotspots) as the true location of subsurface features