Seismic velocity models are essential for many applications including derivation of earthquake properties, studies of tectonic processes, and analysis of earthquake ground motions. In various places there are velocity models of different spatial coverage and resolution. To benefit from the complementary strengths of such models, we aim to develop machine learning methods for enhancing regional velocity models of relatively low-resolution with information contained in local higher-resolution models.
Related paper(s):
- Zhang & Ben-Zion, 2024