Boxing Earthquakes and Faults in ACtive Tectonics

Large-magnitude earthquakes accommodate most of the tectonic deformations of the Earth along active fault systems.

However, despite the occurrence of numerous earthquakes every year, our knowledge of the physical processes governing earthquake ruptures, the relation between rupture propagation, slip distribution and fault geometry, and the evolution of the fault geometry through successive earthquake cycles is still extremely limited, hindering significant progress in earthquake hazard mitigation.

Although an increasing number of observations points to a key role of the fault geometry and its evolution on the way rupture propagates and ends through successive earthquake cycles, understanding the 3D geometry of fault systems and its dynamics during earthquakes from natural data alone remains difficult and fraud with problems.In this project I propose a new approach to address those pending questions by generating my own earthquakes from a combination of lab experiments and numerical simulations, to make a major step forward in the understanding of natural observations of earthquake ruptures and fault systems.

These experimental earthquakes will provide original data to study simultaneously rupture processes and fault geometry, and its evolution in 3D.

The new dataset will be used to train neural networks designed to solve for earthquake source parameters, including 3D rupture geometry and finite slip distribution.

Eventually, the neural networks will analyze real earthquake ruptures, incorporating remote sensing, field, and seismological data, to produce 3D earthquake rupture models.

The project BE_FACT will thus produce an integrated view of the earthquake fault systems that will answer these long-lasting questions about the intimate relations between earthquake ruptures and fault system geometry, providing a new stepping-stone toward a more earthquake-resilient society.