[S09-1] Open session: Earthquake generation process – physics, modeling and monitoring for forecast I

[S09-1-04] Observations and modeling of short-term phenomena in the preparatory stage of large earthquakes

^○Kiyoshi Suyehiro^{1, 2}, Selwyn Sacks², Paul Rydelek², Deborah Smith², Tetsuo Takanami² (1.JAMSTEC, Kanagawa, Japan, 2.DTM, Carnegie Institution for Science, Washington DC, USA)

Past large earthquakes have ample but somewhat inconsistent short-term precursor reports. Utilising these information as prediction before a big event has proved to be formidably difficult. Meanwhile the data quantity has increased rapidly requiring high performance computers to analyse seismic or geodetic data. This trend has greatly contributed in understanding the details of how earthquakes rupture, but not very much on how they prepare themselves. We now know that slow slips contribute to stress re-distribution to complicate the matter.
We refer to a simple cellular automaton model of earthquake adopting Coulomb's failure criterion (Sacks and Rydelek, 1995) and incorporate the effect of dilatancy hardening. This numerical experiment reproduces Gutenberg-Richter's law, b-value decrease with tectonic stress increase and cut off of larger magnitude events after dilatancy hardening sets in. All these are observed in nature.
We speculate based on observations that fluid redistributions become evident when dilatancy breakdown starts to occur leading to triggering a large earthquake. Currently, this phase of the earthquake preparation process is not well established by observations, because they tend to appear sparsely and weakly. We review the Tokai area of the Nankai Trough where an M8 earthquake may be pending considering 1891 Nobi earthquake possibly affecting the sequence of >500 year historical recurrence of M8 events. We note that magnitude dependent seismic quiescence may be evident above the subjecting plate in Tokay area. This suggests that the ongoing preparatory stage is advancing involving a volume outside the plate subduction boundary where the eventual M8 may occur. We propose that the detection and understanding the current phase of preparation need to consider all possible models with physical plausibilities.