RECOGNIZING EXPLOSION SITES WITHOUT SEISMOGRAM READINGS - NEURAL-NETWORK ANALYSIS OF ENVELOPE-TRANSFORMED MULTISTATION SP RECORDINGS 3-6 HZ

Seismic-waveform similarities for closely spaced earthquakes and explo sions in particular are well established observationally. In many indu strialized countries of low seismicity more than 90 per cent of seismi c event recordings stem from chemical explosions and thus contribute s ignificantly to the daily analyst workload. In this study we explore t he possibility of using envelope waveforms from a priori known explosi on sites (learning) for recognizing subsequent explosions from the sam e site excluding any analyst interference. To ensure high signal corre lation while retaining good SNRs we used envelope-transformed waveform s, including both the P and Lg arrivals. To ensure good spatial resolu tion we used multistation (network) recordings. The interpolation and approximation neural network (IANN) of Winston (1993) was used for tea ching the computer to recognize new explosion recordings from a specif ic site using detector output event files of waveforms only. The IANN output is a single number between 0 and 1, and on this scale an accept ance threshold of 0.4 proved appropriate. We obtained 100 per cent cor rect decisions between two sets of 'site explosions' and hundreds of ' non-site' explosions/earthquakes using data files from the Norwegian S eismograph Network.