Discriminating a Nuclear Blast from A Natural Seismic Event Using A Full Moment-Tensor Inversion Method

Abstract

Identification of seismic signals if they are generated by tectonic or volcanic origin or underground nuclear tests is of primary importance for impacts on environmental issues. The recent blast on September 3, 2017 has been in question due to its potential destruction on the global environment. Here, we applied a method of full moment-tensor inversions to analyse broadband waveforms generated by such an explosion. The signals were recorded by a network of 4 regional seismic stations (MAJO, INCN, MDJ, BJT), situated between 360 and 1,100 km away from the explosion epicenter. The results showed that a variance reduction of 72.6%, providing high confidence for all values obtained from the method used: a seismic moment of Mo = 6.25 · 1024 dyne cm, corresponding to a moment magnitude of Mw = 5.8, with a dominant volumetric component (ISO) of 50% relatively compared to a 6.8% double-couple (DC) part and a 43.2% compensated linear vector dipole (CLVD) part, and a centroid-depth of 1.0 km. The relative dominance of the ISO seismic component over the other two components, along with the shallow source extracted from the cross-correlation and source mechanism analysis, suggests that the September 3, 2017 explosion was induced by a suspected nuclear test.