Joint Inversion of Surface and Borehole Seismic Datasets Using Adjoint-State Traveltime Tomography for Karst Exploration

Karst formations pose critical safety risks in engineering projects within limestone regions. Conventional geophysical prospecting often struggles to achieve the necessary accuracy and efficiency for such such complex structures. In this study, we present a novel computational physics framework using adjoint-state traveltime tomography to characterize subsurface karst features. This approach offers significant advantages, including independence from ray tracing, high fidelity in highly heterogeneous media, and superior computational efficiency. We achieve superior illumination of the subsurface and mitigate the “low-velocity shadow” effects inherent in singular acquisition geometries. Field experimental successfully identified three horizontally adjacent but disconnected karst caves between 40 and 50 m depth. Each cave is linked to a subvertical low-velocity channel extending upward to the base of the silty clay layer. The direct correlation between channel width and cave size suggests these structures function as runoff passages connecting the caves to the overlying permeable gravel layer, furthe implying that the karst system remains in an active stage of morphological evolution.