Exploitation of high-sampling Hi-net data to study seismic energy scaling: The aftershocks of the 2000 Western Tottori, Japan, earthquake

Abstract

High-quality seismic data with broad frequency band are essential for the study of seismic energy, E s , and its scaling with seismic moment, M o . The 2000 Western Tottori earthquake (M w 6.6) and its aftershocks as recorded by NIED Hi-net including undistributed high-sampling data provide an excellent data set for this purpose. In this study we use: 1) regular data sampled at 100 sps of small and intermediate (M 2–4) aftershocks just after the mainshock, and, 2) 100 sps data and high sampling 1000 sps data of small events (M 0–3) about two years after the mainshock. Spectral ratios are calculated between all combinations of events that both occurred close to one another and had similar mechanisms. We calculated seismic energies of P and S waves for each event by fitting omega-square spectral models to the spectral ratios. Analysis of both the high and lower sampling rate data results in statistically significant size dependence of E s /M o ; however, none of these trends can explain overall scaling when all events, including the mainshock, are considered. Artificial size dependence due to band limitation and omega-square assumption may be responsible for the apparent trends.