統計数理研究所

講演者: Dr. Abhey.R. Bansal
(University of Edinburgh, Edinburgh EH9 3JW, Scotland, UK
National Geophysical Research Institute, Council of Scientific and Industrial Research, Hyderabad, India)

講演題目: Statistical seismology of Sumatra:
before and after the mega-earthquake of 26 December 2004

概要: We examine the effect of the Sumatra-Andaman Islands mega-earthquake of 26 December 2004 on Sumatran seismicity. Irrespective of time window the inter-event time distribution exhibits Omori-Utsu decay at short time intervals, followed by a gamma distribution at intermediate and long time intervals. The mean inter-event time drops significantly after the event, associated with an increase in the fractal dimension of the time series sequence from 0.07 ± 0.01 to 0.19 ± 0.01, and a doubling of the coefficient of variation of inter-event time. Epidemic Type Aftershock Sequence (ETAS) model parameters are used to quantify changes in the statistical properties. These changes can be significant: the background random seismicity rate μ drops to be indistinguishable from zero after the mega-earthquake (indicating a single dominant aftershock sequence) whereas the productivity factor K is ~13.4 times higher. The magnitude sensitivity parameter α increases after the earthquake, indicating larger aftershocks are proportionately easier to trigger, and the Omori-Utsu exponent p decreases (to ~1), consistent with a longer duration of aftershock sequences. When examined in retrospect, a non-stationary ETAS model provides a better fit to the seismicity data prior to the main shock, the best fitting model being one with a change point to enhanced rates 4.3 years before the mega-earthquake, most likely due to a localized swarm at that time.