APPLICATION OF BACKPROPAGATION NEURAL NETWORKS TO IDENTIFICATION OF SEISMIC ARRIVAL TYPES

A back-propagation neural network (BPNN) approach is developed to iden tify P- and S-arrivals from three-component recordings of local earthq uake data. The BPNN is trained by selecting trace segments of P- and S -waves and noise bursts converted into an attribute space based on the degree of polarization (DOP). After training, the network can automat ically identify the type of arrival on earthquake recordings. Compared with manual analysis, a BPNN trained with nine groups of DOP segments can correctly identify 82.3% of the P-arrivals and 62.6% of the S-ani vals from one seismic station, and when trained with five groups from a training dataset selected from another seismic station, it can corre ctly identify 76.6% of the P-arrivals and 60.5% of S-arrivals. This ap proach is adaptive and needs only the onset time of arrivals as input, although its performance cannot be improved by simply adding more tra ining datasets due to the complexity of DOP patterns. Our experience s uggests that other information or another network may be necessary to improve its performance. (C) 1997 Elsevier Science B.V.