Search

We installed five broadband seismometers around Mt. Usu just before the first eruption in 2000 and the observation continued over 8 months. Among the total of 590 relocated earthquakes that occurred from March 31 12:58 p.m. to April 5, 2000, 41 low-frequency earthquakes were identified. Low-frequency earthquakes, based on waveforms and S-wave spectra, can be classified into several families. We investigated the source process of the low-frequency earthquakes classified as “type 1” for which the similarity among events was the highest. The character of waveforms and S-wave spectra of type 1 earthquakes at Mt. Usu are quite similar to the “hybrid” and “LP” earthquakes at Mt. Redoubt, Alaska, whose source process was explained well by the resonance of a fluid-filled crack. Assuming that they have similar source processes, we estimated several parameters such as crack size by forward modeling. We obtained the following parameters that explain main features (e.g., the frequencies of major peaks) of the observed S-wave spectra at three stations (HGSN, SHNZN and ZNKJI) in the frequency range of 0.5–5 Hz, assuming a vertical crack. Crack length is 200 m, 100 m for its width, 0.14 m for its thickness, 100 for its stiffness, 0.5 for the ratio of bulk modulus and rigidity, and 0.07 for viscous damping loss. The pressure pulse triggering the crack oscillation has a rise time of 0.25 s. From the difference in characteristics of waveforms among stations of various azimuths, the strike of the crack was likely to be northeast-to-southwest. The use of the site-corrected spectra enables us to pin-point the value of viscous damping loss, so that we can conclude that water-steam mixture was responsible for the low-frequency events at Mt. Usu. [Copyright &y& Elsevier]