Title: Machine Learning for Earthquake Monitoring: From Land to Ocean

Abstract:

In recent years, machine learning-based arrival time picking has demonstrated remarkable accuracy and efficiency in monitoring earthquakes both on land and at sea. It has been widely applied to the rapid construction of high-precision earthquake catalogs, which contribute to a better understanding of detailed earthquake source processes, as well as their associated seismogenic mechanisms. In this talk, I will present our recently developed seamless, machine learning-based earthquake detection and location package — LOC-FLOW — as well as the OBS phase picker — OBSTransformer — and their applications. Examples include earthquake swarms, hydraulic fracturing-induced earthquakes, and marine earthquakes offshore the Cascadia region.