The Earthquake‐Source Inversion Validation (SIV) Project

Abstract

Finite-fault earthquake source inversions infer the (time-dependent)
\ndisplacement on the rupture surface from geophysical
\ndata. The resulting earthquake source models document the
\ncomplexity of the rupture process. However, multiple source
\nmodels for the same earthquake, obtained by different research
\nteams, often exhibit remarkable dissimilarities. To address the
\nuncertainties in earthquake-source inversion methods and to
\nunderstand strengths and weaknesses of the various approaches
\nused, the Source Inversion Validation (SIV) project conducts a
\nset of forward-modeling exercises and inversion benchmarks. In
\nthis article, we describe the SIV strategy, the initial benchmarks,
\nand current SIV results. Furthermore, we apply statistical tools
\nfor quantitative waveform comparison and for investigating
\nsource-model (dis)similarities that enable us to rank the solutions,
\nand to identify particularly promising source inversion approaches.
\nAll SIV exercises (with related data and descriptions)
\nand statistical comparison tools are available via an online collaboration
\nplatform, and we encourage source modelers to use
\nthe SIV benchmarks for developing and testing new methods.
\nWe envision that the SIV efforts will lead to new developments
\nfor tackling the earthquake-source imaging problem.