a reduced-order physics-based Inversion framework for joint on-fault slip and off-fault DEformation in elasto-plastic Media

Objective

During an earthquake, coseismic deformation is partitioned between slip on the fault surface and distributed plastic deformation in the surrounding volume. Understanding how off-fault plastic processes interact with fault slip is crucial, as they can control whether ruptures reach the surface and how strongly the ground shakes. Yet, current source imaging ignores this plastic component, providing an incomplete and biased picture of the seismic source and limiting our understanding of rupture processes. IDEM aims to fill this gap by developing a new inversion method that explicitly accounts for the Earth’s elasto-plastic crust. I will combine dynamic rupture simulations that include off-fault plasticity with global optimization to retrieve ensembles of preferred solutions. This represents a major challenge, as dynamic rupture simulations are until now prohibitively expensive for use in large-scale global optimization. To overcome this limitation, I will use Reduced Order Modeling (ROM), a powerful but rarely used technique in seismology, which captures the essential behavior of complex models while being fast to evaluate, thereby enabling efficient exploration of parameter space without sacrificing accuracy. The new inversion framework will jointly resolve initial conditions (stress, fault friction, and off-fault plastic properties) as well as resulting slip distributions and plastic deformation. Beyond delivering physics-based and more robust images of earthquake sources, the project will clarify how off-fault processes shape rupture dynamics and interact with fault slip. This, in turn, will shed light on how faults accommodate deformation and improve the interpretation of long-term fault slip rates, thereby contributing to a more accurate assessment of seismic hazard.