Projects
Research statement
Earthquakes and volcanic eruptions are among the most devastating events. The signatures of these events in geophysical data allow us to understand only partially their mechanism and cycles. During quiescent periods, weaker signals, such as slow earthquakes, are regularly observed, whose roles in the seismic and volcanic cycle remain poorly understood. My research focuses on identifying and modeling these signatures in a data-driven manner by viewing geophysical data through the lens of artificial intelligence. I aim to objectively define these observables and highlight the physical processes that link them to significant events in active geological systems.
Current projects
Enhancing earthquake location with domain adaptation
Part of my research is funded by the Data Intelligence Institute of Paris (diip) which is an interdisciplinary research laboratory at Université Paris Cité. The institute supports the development of interdisciplinary practices in the fields of data science and data intelligence, and brings together researchers from a variety of scientific fields, including formal sciences, physical sciences, life sciences, and social sciences.
The current funded project aims to correct the systematically biased hypocenters obtained with a permanent seismic array from the hypocenters inferred with a temporary array with an adequate geometry, as illustrated in the figure below. We consider the case of Mayotte to develop the method and show the potential outcomes on other datasets of interest. We will learn the catalog bias from the events detected with the trusted array over five weeks and test the prediction quality over one week. Once successful, we will deploy the technique over several years of continuous data at Mayotte and other contexts.
Future project
If you are interested in building a collaboration with me, feel free to contact me at my email address below.