Synthetic Modeling of Multi-Offset GPR and Seismic Data for Void and Buried Rock Detection on the Moon

We seek to utilize geophysical methods on Earth's moon to characterize the lunar regolith and to locate lava tubes, which can provide shelter for astronauts and their equipment against harsh surface conditions. A collocated GPR and seismic survey would help disambiguate some of the uncertainties in interpreting results of each individual method. To assess the benefits of the joint method, we carry out 2D simulations containing voids within a rectangular prism of basalt, and explore how different geometries and frequency filtering affect the GPR and seismic inversions. The GPR forward model uses a finite difference time-domain (FDTD) approach (gprMax) while the seismic model uses a spectral element approach (SpecFem). The two datasets underwent a common midpoint stack with normal moveout correction and a 2D Kirchhoff migration. Results of the migration show similar estimated depth to the ceiling of the void between the two models. A well-defined floor is apparent in the migrated GPR image which is not apparent in the seismic image. Planned future work includes testing the method with more complex simulated geometries and eventually real data, with plans to develop a method for joint inversion.