HR: 16:45h
AN: S32F-04    [PDF]
TI: The 1997 Kagoshima (Japan) Earthquake Doublet: A Quantitative Analysis of Stress Interaction
AU: * Woessner, J
EM: woessner@seismo.ifg.ethz.ch
AF: ETH Zurich, Institute of Geophysics, HPP O11, Zurich, 8093 Switzerland
AU: Hauksson, E
EM: hauksson@gps.caltech.edu
AF: California Institute of Technology, Seismological Laboratory MS 252-21, Pasadena, CA 91125 United States
AU: Wiemer, S
EM: stefan@seismo.ifg.ethz.ch
AF: ETH Zurich, Institute of Geophysics, HPP O11, Zurich, 8093 Switzerland
AU: Neukomm, S
EM: samuel.neukomm@seismo.ifg.ethz
AF: ETH Zurich, Institute of Geophysics, HPP O11, Zurich, 8093 Switzerland
AB: Understanding how the nucleation of earthquakes is affected by sudden changes in the state of stress in their immediate vicinity may provide insight into the elusive relationship between static stress changes and earthquake occurrence. As working hypothesis, we assume that if aftershocks are in part caused by stress changes from their mainshock, changes in their decay rate may reflect changes in the state of stress induced by nearby large earthquakes. The 1997 Kagoshima (Japan) earthquake doublet provides a unique opportunity to analyze this hypothesis for two moderate M6 events that occurred on adjacent faults with the epicenters located in a distance of about 5 km from each other. We map the Omori law parameters on an equally spaced grid for the time period between the two mainshocks (the learning period of 47.8 days) using four Omori law type models with increasing complexity. The best fitting model is found using the corrected Akaike Criterion Information. We then forecast the number of aftershocks for the next 50 days and compare it to the actual observed number. Uncertainties in the rate forecasts are obtained by a bootstrap approach, allowing us to compute a detailed map of the significance of the relative rate changes. We find four regions with highly significant relative rate changes, three negative and one positive, which are a consequence of the activation and deactivation of aftershock activity due to the second mainshock, respectively. While our results confirm a relative rate drop in the Western part of the aftershock sequence that is in agreement with the stress triggering hypothesis (Stein, 2003), the other changes cannot be readily explained. Because of the importance of rate changes for the evaluation of the stress triggering hypothesis as well as for rate and state friction models, we consider our quantitative analysis technique introduced here important and a step forward in the process of understanding the behavior of aftershocks.
DE: 7200 SEISMOLOGY
DE: 7215 Earthquake parameters
DE: 7223 Seismic hazard assessment and prediction
DE: 7230 Seismicity and seismotectonics
SC: Seismology [S]
MN: 2003 Fall Meeting