APPLICATION OF PATTERN-RECOGNITION TECHNIQUES TO LONG-TERM EARTHQUAKEPREDICTION IN CENTRAL COSTA-RICA

The expert system GEO 2.5 has been applied to the analysis of geophysi cal and geological data from Costa Pica. Geophysical data consist of s eismic tomography, Bouguer anomalies, seismic catalogue (RSN-ICE, 1984 -1993) and seismic historical catalogue. Among the geological data, fa ults, lineaments, topography and a structural zonation of the region c oncerned have been used. Many other features have been obtained throug h the calculation of statistical functions of the original data such a s nearness, density, length and b-value, other functions were secondar y derivatives and modulus of horizontal gradient. A first analysis, ma inly based on seismic tomography data and seismic catalogue, has led t o the identification of a few seismogenic structures in the upper crus t characterized by a high statistical correlation coefficient between seismic activity and positive velocity anomalies. Pattern recognition techniques have been applied to the long-term earthquake prediction by the definition, in the feature space, of non-linear regression functi ons between geophysical/geological data and a preliminary estimate of expected maximum magnitude. This preliminary estimate has been formula ted on the basis of both the historical seismic catalogue and the stru ctural zonation. This approach gives the empirical relationships among different geophysical and geological features, which are potentially related to the phenomenon of stress release. Due to the high non-uniqu eness of such an approach, the results have to be physically understan dable and each function has to be interpreted. The obtained multi-dime nsional function has been applied to the calculation of a forecast max imum magnitude field for central Costa Pica. The results are to be con sidered preliminary. An improvement of such a forecast could be achiev ed by new data, such as heat flow, depth of Mohorovicic surface, activ e faults, vertical movement velocities, etc. (C) 1997 Elsevier Science B.V.