Hybrid Modelling of the Economical Consequences of Extreme Magnitude Earthquakes
Abstract
A hybrid modelling methodology is proposed to estimate the probability of exceedance of the intensities of extreme magnitude earthquakes (PEI) and of their direct economical consequences (PEDEC). The hybrid modeling uses 3D seismic wave propagation (3DWP) combined with empirical Green function (EGF) and Neural Network (NN) techniques in order to estimate the seismic hazard (PEIs) of extreme earthquakes (plausible) scenarios corresponding to synthetic seismic sources. The 3DWP modeling is achieved by using a 3D finite difference code run in the ~100 thousands cores Blue Gene Q supercomputer of the STFC Daresbury Laboratory of UK. The PEDEC are computed by using appropriate vulnerability functions combined with the scenario intensity samples, and Monte Carlo simulation. The methodology is validated for Mw 8 magnitude subduction events, and show examples of its application for the estimation of the hazard and the economical consequences, for extreme Mw 8.5 subduction earthquake scenarios with seismic sources in the Mexican Pacific Coast. The results obtained with the proposed methodology, such as those of the PEDECs in terms of the joint event "damage Cost (C) - maximum ground intensities", of the conditional return period of C given that the maximum intensity exceeds a certain value, could be used by decision makers to allocate funds or to implement policies, to mitigate the impact associated to the plausible occurrence of future extreme magnitude earthquakes.
- Publication:
-
AGU Spring Meeting Abstracts
- Pub Date:
- May 2013
- Bibcode:
- 2013AGUSM.U33A..05C
- Keywords:
-
- 1906 INFORMATICS / Computational models;
- algorithms;
- 0555 COMPUTATIONAL GEOPHYSICS / Neural networks;
- fuzzy logic;
- machine learning;
- 3275 MATHEMATICAL GEOPHYSICS / Uncertainty quantification;
- 7290 SEISMOLOGY / Computational seismology