Key research themes
1. How can modular and scalable inversion frameworks improve the computational efficiency and flexibility of 3D electromagnetic data inversion?
This research area focuses on the development of modular, reusable, and extensible computational frameworks for inversion of electromagnetic geophysical data. It highlights the benefits of fine-grained modularization, parallelization, and code reuse to enable flexible application to diverse EM data types and to facilitate rapid prototyping of new inversion algorithms. This is critical given the computational intensity and complexity of 3D EM inversion problems across different geophysical techniques.
2. What novel parameterization and regularization strategies can enhance the stability and resolution of 2D/2.5D potential field data inversion?
This area investigates innovative parameterization techniques such as series expansions, constraint formulations based on known a priori information, and sparsity-promoting regularizations that improve both the numerical stability and physical plausibility of inverted potential field models. It addresses the challenges posed by non-uniqueness and underdetermination by reducing model dimensionality or enforcing physically meaningful constraints, thus enabling more accurate recovery of subsurface structures from gravity, self-potential, and resistivity data.
3. How can advanced inversion methods address ill-posedness and non-uniqueness in potential field data inversion using sparsity-promoting and variational approaches?
This theme explores mathematical and computational developments to overcome the fundamental ill-posedness and non-uniqueness of potential field inverse problems, particularly those formulated as measure-theoretic or constrained optimization problems. It includes sparse recovery concepts in infinite-dimensional settings, regularized nonlinear inversion formulations via penalty methods and constrained searches, and uncertainty quantification through systematic estimation of parameter uncertainties and inversion error bounds.