Ground Penetrating Radar (GPR) is shown to be a successful tool in detecting tunnels and voids. Lava tubes are tunnel-like features in volcanic terrains that can be potential safe places for human crews and equipment on the Moon and Mars. We utilize GPR to detect and map lava tubes (Valentines cave, Skull cave and Hercules Leg cave) in Lava Beds National Monument, CA. Our preliminary results show that the ceiling of the lava tubes are readily detectable by GPR. However, due to the strong radar velocity contrast between lava and the air-filled tubes, accurate recovery of the position of the lava tube floor is much more challenging. Careful migration of the GPR data is required to resolve the floor signature and create an image with the tube floor restored to its correct depth. We are developing an optimal workflow for recovering complete lava tube geometries. We can do this because we have collected centimeter-scale LiDAR data from the interior of tubes as well as on the surface along GPR transect lines. Thus we can test the accuracy of GPR migration methods against the LiDAR-measured tube geometry. We are testing conventional 2D migration techniques as well as topographic migration. At selected field sites we have limited 3D ‘grids’ of data. We expect to compare the results of different migration techniques to identify optimal methods for this problem. As a part of this project, we also seek to develop a library of different lava tube geometries and their corresponding GPR image from their migrated sections. The GPR image library will encompass a range lava tube geometries, including tubes of different heights, widths, shapes, and structures (e.g., pillars), plus a variety of floor textures (e.g., smooth, ropey, rubble) and overhead thickness. This library will be an asset for determining the utility of deploying GPR technology in mapping a tube-rich environment.