Efficient 3D airborne GPR simulation and migration imaging for landslide investigation

This paper investigates methodologies for 3D numerical simulation and imaging in airborne ground-penetrating radar (GPR) applications for landslide surveys. We propose a novel subgridding FDTD(2,4) method that enables memory-efficient 3D simulation of arbitrarily complex landslide models on standard workstations. The proposed approach implements high-order FDTD(2,4) operators consistently across both coarse and fine grid regions, successfully simulating challenging scenarios including rugged terrain and vegetation-covered surfaces, while reducing memory requirements by 60-90% compared to conventional FDTD methods. Combined with optimized migration algorithms, we have established a complete “simulation-to-imaging” workflow capable of effectively detecting bedrock-overburden interfaces in landslides. Comprehensive validation through theoretical model tests and field-scale landslide model experiments confirms the method’s effectiveness and demonstrates its practical value for rapid landslide assessment and disaster prevention applications. The developed techniques bridge critical gaps between theoretical modeling and field deployment of airborne GPR systems.