How Much Does the Seismic Data Tell us About the Sinking of the Kursk?

The Norwegian seismological research group, NORSAR, reported two approximately co-located seismic disturbances in the Barents Sea on 12 August 2000. The NORSAR seismologists and other workers have associated these disturbances with the sinking of the Kursk. Seismologists from the University of Arizona and Los Alamos show that the spectra of seismic data, recorded at regional distances, from the second disturbance show a characteristic ``bubble-pulse'' frequency, diagnostic of an underwater explosion. They argue that the bubble-pulse frequency is consistent with a yield of 3.1-4.5 tonnes equivalent TNT (assuming source depths between 85 and 100 m). They also argue that the high correlation between the seismic signals from the first and second disturbances suggests that the first was also an underwater explosion. We have performed preliminary P-wave modeling, using an impulsive explosion-source within a water layer over a half space, with take-off angles appropriate for regional phases. The results suggest that interference between direct P and multiple water reverberations, due to both upward and downward P radiation (pnwP and PnwP), gives rise to amplitude spectra with peaks and notches that are sensitive to not only the water depth, but also the source depth. We will show the results of model fits to the amplitude spectra from the first and second disturbances recorded at ARCESS in northern Norway. We will also assess to what extent the source depths from the modeling of the seismic data are consistent with depths inferred by other workers from non-seismic evidence.