Uncertainties in finite-fault slip inversions: how critical is the forward model?

Théa Ragon  - CNRS, ISTerre (introduction 15min: Hugo Sanchez-Reyes)
--- 
We try to image physical mechanisms ongoing at depth, on faults, and responsible for the tectonic deformations we measure at the Earth’s surface. One of the goal of these models is to better understand why, and how, earthquakes start, propagate and stop. Yet, deformation on faults is complex and mildly understood beyond first order characteristics, one reason being the uncertainty of inferred finite-fault models. 20 years ago, Igor Beresnev published an article entitled “Uncertainties in Finite-Fault Slip Inversions: To What Extent to Believe? (A Critical Review)”. In the last 20 years, methodological developments, computing resources, and data quality and quantity have increased drastically. Can we now believe our slip models? Did the way we handle uncertainties improve drastically as well? One prerequisite in slip/source inversions is to assume that the source and measurements are related via a known forward model, or Green’s functions. But the forward model is often inaccurate and ignores complexities that are either hard to capture or simply unknown. In this talk, I will use three examples to explore how we can overcome our poor knowledge of the forward problem, and what is the impact on our models. In particular, with a synthetic setting and the case of the 2010 Mw 8.8 Maule (Chile) earthquake, we will review how accounting for inaccurate assumptions in the forward model can improve the robustness of our models and help us optimize the use of our data. We will then examine an ill-posed alternative: to jointly solve for both the forward and inverse problems, as applied to the empirical Green’s functions approach. I will conclude with some perspectives on practical frameworks to systematically describe inaccuracies in the forward problem, and key questions that could benefit from it.