IAG-IASPEI 2017

Presentation information

Oral

IASPEI Symposia » S13. Earthquake source mechanics

[S13-3] Earthquake source mechanics III

Thu. Aug 3, 2017 1:30 PM - 3:00 PM Main Hall (Kobe International Conference Center 1F)

Chairs: Yuji Yagi (Graduate School of Life and Environmental Sciences) , Satoshi Ide (University of Tokyo)

1:45 PM - 2:00 PM

[S13-3-02] A statistical characterization of earthquake initiation and its implication

Shunta Noda1, William Ellsworth2 (1.Railway Technical Research Institute, Tokyo, Japan, 2.Stanford University, CA, USA)

Rapid magnitude (M) determination is critical for earthquake early warning (EEW). A number of methods have been proposed to do this. For example, the use of the frequency content of initial P wave (e.g., Allen & Kanamori, 2003), or a combination of displacement amplitude and a ground motion prediction equation (e.g., Odaka et al., 2003) or from the time between the onset of a body wave and the peak amplitude arrival of that wave (e.g., Noda et al., 2016). It is an essential question for these techniques whether or not the final M is deterministic at the time of initiation (e.g., Olson & Allen, 2005). To investigate this issue, Noda & Ellsworth (2016, GRL) used Japanese K-NET data from 150 events with 4.5 <= Mw <= 9.0 and hypocentral distance (R) less than 200 km, and examined the P-wave absolute displacements. They found that the P-wave displacement began in a similar way at the onset and departed from the similarity earlier for smaller events. In the magnitude range up to Mw 7, the scaling relation between the departure time (Tdp) and the final Mw resulted in Mw = 2.29 * logTdp + 5.95 which suggests that Tdp occurs before the completion of rupture in a statistical sense. Noda & Ellsworth indicated that the scaling relation was consistent with what Olson & Allen (2005) called the deterministic property of earthquakes in which the P-wave frequency contents (tau-p) scales with the final M even when the time window is significantly shorter than the typical source duration. Here we discuss how such a deterministic property could be explained by the model of Murphy & Nielsen (2009) where the available elastic energy at the initial rupture patch must exceed a certain threshold in order for the rupture to propagate spontaneously far beyond the initial patch. When the energy lies below the threshold, the final M is statistically deterministic at the time of initiation. When the energy exceeds the threshold, final M cannot be reliably determined from initial P wave alone.