Skip to the content.

KAUST-IAMCS Workshop on Modeling and Simulation of Wave Propagation and Applications

Tarje Nissen-Meyer, Swiss Federal Institute of Technology (Switzerland)
Computable Seismology: Perspectives for Efficient Global Wave Propagation and Inversion

Abstract

Numerical methods for seismic wave propagation have come a long way in the past decades and are now capable of replicating actual seismic observations to astonishing agreement. However, two fundamental bottlenecks are still at large: the reliability and accuracy of a priori assumptions on structure and source, as well as the immense computational cost to solve realistic systems especially for the inverse problem. This is most visible at two end-member scales: global-scale tomography and exploration-scale imaging. I will quantify these problems and propose avenues away from Moore's Law to optimally circumvent these problems depending on the geophysical question at hand. In particular, I will focus on an efficient 3D global-scale wave propagation method that collapses the computational domain to 2D, thus allowing for a maximal exploitation of the notoriously sparse data space for both forward and inverse modeling, represented by various geophysical questions. Finally, I will sketch perspectives of applying scattering theory to vastly accelerate 3D wave propagation at the global scale, possibly enabling iterative or probabilistic inverse approaches with typical or moderate HPC resources.