ASSESSING THE ACCURACY OF ELECTRICAL RESISTIVITY TOMOGRAPHY AT MODELING SUBSURFACE VOID SPACES OF VARYING DEPTHS AND SURROUNDINGS

Electrical resistivity tomography (ERT) offers a robust, nonintrusive method for modeling subsurface features in karst terrains such as cave passages, water conduits, and sinkholes. Competitors to ERT such as boreholing and microgravity are either too costly or impractical in karst landscapes, making it ever more crucial to fully understand and calibrate ERT. Despite this, several problems remain with employing ERT, which are rooted in ERT’s limited resolution, the non-uniqueness and instability of solutions to inversion models, and the inclusion of a smoothness parameter in modeling that blurs void-bedrock interfaces. The present research studies how detection and resolution of void spaces are affected by their depth and surrounding void spaces and hydrogeology.

To evaluate its reliability, ERT was performed on the caves of Burnsville Cove in Virginia. Three 335 meter long 56 electrode lines were set up atop deep maze passages, rooms, and streams in Helictite Cave, as well as over shallower maze passages in Wishing Well Cave. The apparent resistivity datasets collected in the field were analyzed with EarthImager2D, which uses a linear least squares finite element method to generate a subsurface resistivity structure that best fits the measured dataset. Since Helictite and Wishing Well are well-mapped, the generated inversion models were tested for accuracy against the real-life subsurface structure. Preliminary results show that while shallow, adjacent passages can be distinguished from one another, ERT blurs deep cave maze passages into giant, amorphous resistive blobs that might be mistaken for caverns. Moreover, the void spaces around subterranean streams don’t appear in inversion models, which could confuse the observer into thinking that no caves or conduits are present in the region at all. The final results of the study should pave the way towards better characterizing new, unmapped subsurface features in karst terrains.