Slip Tendency Analysis using Physics-Based Machine Learning

A slip tendency analysis is a powerful approach to estimate how close a fault is from reactivating, based on its local stress and orientation. However, recovering the full tensorial stress state usually requires quite some time and resources to derive a geomechanical model of the area.

In our recent study published in Geophysical Research Letters, https://doi.org/10.1029/2024GL109524, we use a Physics-Based Machine Learning approach to estimate the spatial distribution of 2D stress tensors, based on stress orientation from the World Stress Map and displacements from Global Navigation Satellite System (GNSS) observations. This regional-scale stress estimation allows us to derive rapid slip tendency analyses on fault networks in Australia, as demonstrated in an analysis on the Eyre Peninsula.