AGU Session: Earthquake Source Inversion Under Scrutiny

Please consider the following session if you are working towards an
understanding of inverting data to infer source parameters.

Earthquake Source Inversion Under Scrutiny: Validation, Resolution, Robustness

Sponsor: Seismology
CoSponsor: Geodesy

Conveners: Paul Martin Mai
ETH Zurich, Institute of Geophysics
Schafmattstrasse 30
Zurich, ZH, CHE 8093
mai@sed.ethz.ch

Gaetano Festa
University of Naples, Dept. of Physics, ITA
festa@na.infn.it

Cotton Fabrice
LGIT Grenoble, FRA
fabrice.cotton@obs.ujf-grenoble.fr

Hiroe Miyake
University of Tokyo, Earthquake Res. Inst., JPN
hiroe@eri.u-tokyo.ac.jp

Index Terms: 7200 7209 7215 7215 7260.

Description: Although a number of inversion methods have been developed to image the time-space evolution of earthquake ruptures, the reliability and robustness of the resulting earthquake source models is in most cases still inadequately assessed. Comparing source models inverted using similar data, but different inversion techniques, applied by different research teams, often show dissimilarities beyond what a standard resolution analysis would have predicted. The effects of Greens function's uncertainty, data distribution and data quality, and subjective choices of model parameterization are hardly ever properly accounted for in the reported rupture models. To better understand the quality of earthquake source images and to what extent we can "trust" the results -- which are an integral ingredient for studying earthquake dynamics, for ground-motion simulation and seismic hazard assessment -- we urgently need a critical evaluation of current (and past) source-inversion methods.

This session solicits contributions on recent advancement in earthquake source-inversion technologies, addressing in particular the a postiori model uncertainty, source-model resolution and robustness. We invite theoretical work on optimal source inversion strategies and data requirements (data distribution, combination of different data sets), as well as studies on innovative procedures to evaluate the robustness of inverse methods or to quantify similarities and discrepancies between inverted models. We also solicit contributions that address new ways to image the space-time evolution of earthquake rupture, potentially at multiple space-time scales simultaneously. Contributions presenting finite-source inversions for recent earthquakes are welcome as well.

Furthermore, we promote a "blind test" on earthquake-source inversion (http://www.seismo.ethz.ch/staff/martin/SPICE.html, conducted within EU Marie Curie Research Training Network). Synthetic data have been generated which are currently inverted by a number of researchers, but we motivate any interested scientists to participate in this source-inversion validation exercise. The session will gather initial results from this blind test to stimulate discussion on reliability, robustness and resolution of inverted rupture models.