The Omori-Utsu law is a fundamental empirical law descirbing the behaviors of aftershocks' seismicity rate. However, the mechanism behind the 'c' parameter in this law is under debate. The incompleteness of earthquake catalog following the mainshock obscure the attempts to understand the 'c' parameter. In this project, I aim to develop or employ state-of-the-art machine learning model for detecting early aftershocks that are buried inside the mainshock's coda wave or elevated noise level.
The extensive extraction of ground water in the Central Valley has caused tremendous ground subsidence within 70 years. Our fellow graduate student Matt Lees did fanscinating work to model the clay compaction due to ground water extraction in 1D case. Under the supervision of Prof. Eric Dunham, I am working on a second project to model the subsidence in 3D scenario to provide comprehensive evaluation about current status about subsidence risks.
With the advent of machine learning based earthquake catalog workflow, numerous catalogs are built, together with template matching method. While earthquake relocation is often the last step for building catalog,s it's time to rethink the difference in each method. In this project, we compare the 8 widly used earthquake location methods with synethetic data based on 2019 Ridgecrest setting (Fast Marching Method, 3D velocity structure, elevation effect, realistic phase error).Â
More detail can be found in the Seismological Research Letters: https://doi.org/10.1785/0220240354 or my Github repo here ,