5:15 PM - 7:15 PM
[SSS12-P07] Spatiotemporal analysis of seismicity in Niigata Chuetsu region using the Epidemic-Type Aftershock Sequence model (2)
Keywords:Epidemic-Type Aftershock Sequence model, earthquake, seismicity
The Epidemic-Type Aftershock Sequence (ETAS) model (Ogata, 1988) is a statistical model that shows the seismicity rate in target region over a given period, under the assumption that all earthquakes generate aftershocks. The Chuetsu region of Niigata Prefecture (mid-Niigata region) is located in the eastern margin of the Japan Sea tectonic zone and Niigata-Kobe Tectonic Zone, where large earthquakes occurred frequently. In recent years, 2004 Mid-Niigata earthquake (M6.8) and 2007 Niigata Prefecture Chuetsu-Oki earthquake (M6.8) were occurred in the area.
In this study, we conducted a spatiotemporal analysis of seismicity in the Chuetsu region of Niigata Prefecture from 2000 to 2020 using the spatiotemporal ETAS model (Ogata, 1998). We use Japan Meteorological Agency’s unified epicenter catalog data (M3 or greater). The ETAS parameters are estimated using the package developed by A. Jalilian (2019).
The ETAS models fitted to some time series that include the 2004 Mid-Niigata earthquake and the 2007 Niigata Chuetsu-Oki earthquake overestimate the actual seismicity (Fig.). This gap between the two may be attributed to the fact that multiple aftershocks of M6 or greater occurred following Mid-Niigata earthquake mainshock, then b-value declined temporarily and leaded ETAS model to overestimate the actual seismicity.
Furthermore, a decrease in background seismicity rate mu is shown in these time series compared to other periods. We estimated the ETAS parameters for the entire study area of Chuetsu region and for source region of the Mid-Niigata earthquake to examine this background seismicity rate decrease in more detail. The results showed that the estimated background seismicity rate mu for the entire study area was more than an order of magnitude lower than mu estimated in other periods for the same area or for the source region in the same period. We considered this decrease of mu may be influenced by Coulomb failure stress change (dCFF) caused by slip oh the Mid-Niigata earthquake, so we calculated delta dCFF to mainshock fault. The results indicated that dCFF increased only in source area, and in surrounding the source region, we found the decrease of Coulomb failure stress over a relatively wide area. This suggests that seismicity quieted due to a decrease in Coulomb failure stress attributed to the slip of the Mid-Niigata earthquake and bring the decrease of mu.
Additionally, the Niigata Prefecture Chuetsu-Oki earthquake had its mainshock hypocenter located near the edge of the aftershock zone. Since the ETAS model assumes that aftershocks spread concentrically from the mainshock, we modified the mainshock hypocenter to a centroid location that is closer to the center of the aftershock zone and refitted the ETAS model to the time series that included the Niigata Prefecture Chuetsu-Oki earthquake. As a result, the model fitted the actual seismic activity more accurately when the mainshock hypocenter was replaced with the centroid. Furthermore, the ETAS parameter A, which represents the productivity of aftershocks, was found to be smaller when the centroid was used compared to the time series including the Mid-Niigata earthquake. The reason of the decrease of productivity of aftershocks in time series including the Niigata Prefecture Chuetsu-Oki earthquake may be the difference of maturity between source faults of Niigata Prefecture Chuetsu-Oki earthquake and Mid-Niigata earthquake, and the fact that its source region was in an area where dCFF had decreased due to the Niigata Chuetsu earthquake.
In this study, we conducted a spatiotemporal analysis of seismicity in the Chuetsu region of Niigata Prefecture from 2000 to 2020 using the spatiotemporal ETAS model (Ogata, 1998). We use Japan Meteorological Agency’s unified epicenter catalog data (M3 or greater). The ETAS parameters are estimated using the package developed by A. Jalilian (2019).
The ETAS models fitted to some time series that include the 2004 Mid-Niigata earthquake and the 2007 Niigata Chuetsu-Oki earthquake overestimate the actual seismicity (Fig.). This gap between the two may be attributed to the fact that multiple aftershocks of M6 or greater occurred following Mid-Niigata earthquake mainshock, then b-value declined temporarily and leaded ETAS model to overestimate the actual seismicity.
Furthermore, a decrease in background seismicity rate mu is shown in these time series compared to other periods. We estimated the ETAS parameters for the entire study area of Chuetsu region and for source region of the Mid-Niigata earthquake to examine this background seismicity rate decrease in more detail. The results showed that the estimated background seismicity rate mu for the entire study area was more than an order of magnitude lower than mu estimated in other periods for the same area or for the source region in the same period. We considered this decrease of mu may be influenced by Coulomb failure stress change (dCFF) caused by slip oh the Mid-Niigata earthquake, so we calculated delta dCFF to mainshock fault. The results indicated that dCFF increased only in source area, and in surrounding the source region, we found the decrease of Coulomb failure stress over a relatively wide area. This suggests that seismicity quieted due to a decrease in Coulomb failure stress attributed to the slip of the Mid-Niigata earthquake and bring the decrease of mu.
Additionally, the Niigata Prefecture Chuetsu-Oki earthquake had its mainshock hypocenter located near the edge of the aftershock zone. Since the ETAS model assumes that aftershocks spread concentrically from the mainshock, we modified the mainshock hypocenter to a centroid location that is closer to the center of the aftershock zone and refitted the ETAS model to the time series that included the Niigata Prefecture Chuetsu-Oki earthquake. As a result, the model fitted the actual seismic activity more accurately when the mainshock hypocenter was replaced with the centroid. Furthermore, the ETAS parameter A, which represents the productivity of aftershocks, was found to be smaller when the centroid was used compared to the time series including the Mid-Niigata earthquake. The reason of the decrease of productivity of aftershocks in time series including the Niigata Prefecture Chuetsu-Oki earthquake may be the difference of maturity between source faults of Niigata Prefecture Chuetsu-Oki earthquake and Mid-Niigata earthquake, and the fact that its source region was in an area where dCFF had decreased due to the Niigata Chuetsu earthquake.