EGS Meeting

Grenoble, August 29 - September 1, 2006

Several SPICE scientists (Prof. Moczo, Prof. Igel, Prof. Vilotte, Dr. Mai), were directly involved in the organization of this workshop in Grenoble, that had a a central goal to investigate, how well one can predict the ground motion following a large earthquake. This problem is directly related to the technical problem of wave propagation in 3D which is a central theme of the SPICE project. The meeting was centered around a "benchmark" exercise: a 3D model of the Grenoble Valley and the seismic source specifications were handed out to the participants several months in advance. The groups calculated waveforms for this model with various approximations and submitted the seismograms and specific waveform characteristics (e.g., the peak ground velocity in the whole area) to the organizers who presented the comparative results at the meeting. These results were to some extent shocking. Despite the fact that basically everybody was solving the same physical forward problem, the results showed quite a strong divergence. This was in part due to the fact that the problem was quite complex in terms of structure and source and there were many potential sources of errors. This was in fact an important exercise in the sense that it highlighted the necessity to have benchmarking structure just like SPICE is developing (see part on research achievements). In the meantime, several more iterations led to converging results, which is encouraging. An important consequence of the meeting is that the Grenoble benchmark data will be merged with the SPICE benchmarking exercises in the near future. This will allow anybody developing forward modelling algorithms to test under quite complex situations. Another outcome of the workshop has demonstrated the problems related with grid generation. This has led partly to the collaboration of SPICE scientists with the US community. Visits to the US (funded by Sandia Labs) are scheduled for 2007 to address the issues involved.