Home  /  SCEC Workshops  /  2008 Annual Meeting: Earthquake Source Inversion Workshop

2008 Annual Meeting: Earthquake Source Inversion Workshop

Conveners: Martin Mai, Danijel Schorlemmer, and Morgan Page
Date: September 7, 2008
Location: Plaza Ballroom CD, Hilton Palm Springs Resort

Earthquake source inversions image the spatio-temporal rupture evolution, and hence constitute a critical research tool to develop a better understanding of the details of the earthquake source process. However, finite-source rupture models, developed by different research teams for the same earthquake, exhibit surprisingly large intra-event variability. Moreover, a recent, small-scale “source inversion blind test” revealed the difficulties that several inversion methods had, even for a rather simple test case. Despite a number of new developments in source-inversion approaches, the reliability, resolution, and robustness of these inversion strategies has not been rigorously examined and tested.

This workshop serves as a starting point for launching an Earthquake Source Inversion Validation Exercise. We invite researchers working in earthquake source inversion (and inversion theory and application in general), users of source-rupture models (e.g., rupture-dynamic and ground-motion-simulation people), and statistical analysts (testing-center folks and related programmers). The scope of the workshop is (a) to briefly summarize current source-inversion methods; (b) to define the long-term strategy for conducting such a validation exercise; (c) to discuss the logistical and computational framework needed for this exercise; (d) to setup task groups for the different steps of the validation (model building; ground-motion synthesis; statistical analysis of submitted models; computational infrastructure).


8:00 - 8:10 Introduction: Scope of the workshop & review of an initial blind-test M. Mai
8:10 - 8:25 Strategies for uncertainty assessment in source inversions M. Page
8:25 - 8:40 Importance of covariance components for finite-source inversions Y. Yagi
8:40 - 8:55 Current status of source inversion methods in Japan K. Koketsu
8:55 - 9:10 Inverting for dynamic source parameters R. Madariaga
9:10 - 9:25 Experience from the Dynamic Code Validation Project R. Harris
9:25 - 9:40 An existing testing center: CSEP D. Schorlemmer
9:40 - 10:00 Break  
10:00 - 12:00 OPEN DISCUSSION: Planning the source-inversion validation exercise  
10:00 - 10:40 DISCUSSION 1: Science issues / General questions
  • Source inversion: an under-determined problem
  • How to best quantify waveform fits?
  • How to optimally weight different data sets used in the inversion?
  • Effects of source-receiver constellation?
  • Effects of different stabilizing/smoothing constraints?
10:40 - 11:20 DISCUSSION 2: General set-up of the exercise
  • Greens functions: pre-computed or velocity-density models specified?
  • Greens functions: Simple media or realistic structures?
  • Source geometry: how much information should be provided?
  • What "data" (synthetics) are desired? Statics, strong-motion, teleseismics?
  • "Clean" synthetics or noise-contaminated?
  • How to compare inversion results; what are relevant statistical measures?
  • A fully prospective inversion?
11:20 - 12:00 DISCUSSION 3: Logistics, resource, and management
  • Overall strategy and rough time-frame
  • Formats for submission of models/synthetics?
  • Authorization of data streams?
  • Needed computational/infrastructure resources?
  • Proposal for future funding for these efforts?