ИСТИНА

This poster describes results of work to develop an adaptive algorithm for identification of main seismic phases based on time analysis (wavelet analysis) of seismograms in real time. We provide the main methods for dealing with this problem in case there is considerable manmade noise. We demonstrate results from testing the algorithm on actual records of recent large earthquakes. The algorithm was estimated for its efficiency. It is shown that the probability of an error of the first kind (failure to detect an earthquake) is on the order of 6%, while that of an error of the second kind (false alarm) is about 20%. Recommendations are provided as to further improvement of the algorithm.