Earthquake Prediction and Seismicity Patterns

Abstract

The occurrence rate of the characteristic earthquake (CE) was compared with seismicity in nine source regions of interplate earthquake (i.e., the regions off Shikotan Island, off Nemuro, off Tokachi, off Northern Sanriku, off Miyagi, far off Miyagi, and the Tonankai, Nankai, and Kanto regions) by combining instrumental data and the recurrence interval of CEs evaluated by the Headquarters for Earthquake Research Promotion (HERP) in Japan. The recurrence interval of the Nankai earthquake was estimated on the basis of long historical records and found to be one of the least uncertain ones. We used the unified catalog of earthquakes obtained by the Japan Meteorological Agency (JMA) on the basis of a recently improved seismic network, together with the old JMA catalog. Seismicity of all the interplate source regions indicated that the number of observed events was much less than the number of events predicted from the Gutenberg-Richter (G-R) relation and the occurrence rate of the CE. In all regions except for far off Miyagi, the CEs occurred during the interval of the earthquake catalog. Thus, our data set included the highest seismicity period during an earthquake cycle. In the region off Tokachi where the 1952 and 2003 Tokachi-Oki earthquakes occurred, the magnitude frequency distribution (MFD) during one seismic cycle exhibited a magnitude gap of 1.1 between the CE and the other events. Therefore, our results favored the CE model. Moreover, we conversely estimated an average recurrence interval of the CEs in each region, based on the assumption that the G-R relation holds. Most estimated recurrence intervals were far longer than the evaluation given by HERP.