[S-SS08] Statistical seismology and underlying physical processes

[SSS08-01] Inverse analysis of aftershock activities by non-stationary ETAS models based on inhomogeneous seismic catalogs immediately following major earthquakes

*Yosihiko Ogata¹ (1.Research Organization of Information and Systems, The Institute of Statistical Mathematics)

Keywords:detection probability of each earthquake, stationary and nonstationary ETAS model, early aftershocks, M7.6 Noto Peninsula earthquake and the 2023 M6.5 northern Noto Peninsula earthquake

The ETAS model is the recommended model for short-term prediction of aftershocks and earthquake swarms, but due to the large variability in seismic activity, the model parameters must be adjusted to the seismic activity in order to provide accurate predictions. However, adjustment for aftershocks of large earthquakes is hampered by the extreme non-uniformity of aftershock data in the immediate aftershock period. During this period, many aftershocks, even relatively large ones, go undetected.

In this report, we venture to consider all detected aftershock data for the purpose of early on-line forecasting. This requires estimating the detection probability of each earthquake in the catalog and modeling it in conjunction with a stationary or nonstationary ETAS model.

As examples of this analysis, we include the aftershock activity of the 2024 M7.6 Noto Peninsula earthquake and the 2023 M6.5 northern Noto Peninsula earthquake, but I find differences in the results of the two inverse analyses.