Discrimination of chemical explosions and earthquakes in central Europe—a case study

Both natural and artificial seismicity is observed in the Vogtland area at near regional distance (180 km) from the GERESS array. This setting offers the opportunity to test various discrimination criteria using a uniform data set of 39 earthquakes and 22 chemical explosions (strip mine and quarry blasts) from a single epicentral region, mostly eliminating wave-path and distance effects. Discriminants used include L_g/P_g and L_g/R_g maximum amplitude ratios, spectral peaks and slopes extracted by autoregressive / moving average (ARMA)-modeling, master-event correlation, and the search for time-independent frequency structures in sonograms. Although no single discrimi- nant achieves a perfect separation of the two populations, a joint discrimination scheme employing simple majority voting reduces misclassifications to 2%. Future research will focus on possible generalization and the physical basis of the discriminants.