Personal tools
You are here: Home Science & Projects Deisa Extreme Computing Initiative Projects 2009 - 2010 SEismic Tomographic Inversion and Verification of Earth Models

SEismic Tomographic Inversion and Verification of Earth Models

Research Area Earth Sciences
Principal Investigator(s) Dr. Josep de la Puente
  • Instituto de Ciencias del Espacio, CSIC-IEEC, Barcelona, Spain
  • Ludwig-Maximilians-Universit√§t M√ľnchen, Department of Earth and Environmental Sciences, Geophysics, Germany


The SETIVEM project aims at improving significantly our knowledge of the subsurface structure by minimizing the discrepancies between observed seismograms from the field and the numerically predicted ground motion. The project uses the most recent 3D seismic simulation tools 1) to enhance the accuracy of forward predictions using precise representations of the geology and the source physics and 2) to apply the adjoint method to produce improved tomographic images of the subsurface. The most advanced tomographic tools nowadays rely on the solutions of the whole wave field through iteratively improved geological models. The size of the models considered and the high-frequency phenomena involved in the wave fields make such computations a challenge for state of the art computing platforms. On the other hand, forward solvers can use those tomographic models, combined with further a priori information on the geological setup, for the purpose of accurately predicting the ground motion induced by given seismic sources. The quality and physical interpretability of such predictions depend on extremely different scales, ranging from meters to capture the source physics or small-scale geological features to tens of kilometers to propagate the waves from the source to the receivers through strongly heterogeneous material. SETIVEM focuses on problems in exploration seismology which, due to their size and frequency content, are unsolvable on local medium-size computing systems but require exceptional computational resources and expertise. With the results of SETIVEM using a European framework of large computational power it is possible to solve current problems in seismology that in turn will set future challenges in numerical seismology for the upcoming years.

Document Actions