地球与行星物理学术报告通知（202618期）- 童平

报告人简介：

Ping Tong is an Associate Professor in the School of Physical and Mathematical Sciences at Nanyang Technological University (NTU) and a Principal Investigator at the Earth Observatory of Singapore. His research focuses on inverse problem theory, numerical modelling, and geophysical imaging, with applications to energy and resource exploration, and deep Earth structure and dynamics. He has published over 90 papers in journals such as PNAS, Nature Geoscience, Earth and Planetary Science Letters, Geophysical Research Letters, Journal of Geophysical Research: Solid Earth. He currently leads projects on advanced geophysical imaging technologies, geothermal resource assessment, and AI-driven methods for high-resolution subsurface imaging.

报告内容摘要：

In this presentation, I will discuss our ongoing study on a unified framework for multimodal geophysical inversions developed for multiscale subsurface imaging. The methodology integrates seismic, electrical resistivity, electromagnetic, gravity, and magnetic data within a either fully or partially coupled inversion architecture. Forward modeling and inverse problems are formulated using adjoint-state methods. This formulation enables efficient gradient computation in high-dimensional parameter spaces. Modern AI techniques are incorporated to accelerate convergence, strengthen regularization, and capture nonlinear correlations across modalities. By leveraging the complementary sensitivities of diverse geophysical measurements, the framework mitigates non-uniqueness and enhances resolution across scales, from shallow engineering structures to deep crustal and uppermost mantle features. Applications include geothermal resource assessment in Singapore, cavity detection, and train-induced vibration monitoring. The objective of this ongoing work is to rigorously test and validate the robustness and versatility of multimodal inversion strategies for complex subsurface systems.