2:30 PM - 2:45 PM
[SCG50-04] Tsunami simulation for the Hualien, Taiwan earthquake (M7.4) on April 3, 2024
★Invited Papers
Keywords:Taiwan, 2024 Hualien earthquake, Tsunami simulation
On April 3, 2024 at 7:58 (local time), which is around 8:58 in Japan (hereinafter described in JST), an earthquake of M7.4 occurred at a depth of 40 km near Hualien City, Taiwan (23.836°N, 121.598°E) according to the USGS. The maximum seismic intensity was 6+ (on Taiwan scale which is similar to Japanese scale) in Hualien City, and 5 was observed at many locations in Taiwan, according to Central Weather Bureau. The maximum intensity in Japan was 4 on Yonaguni Island according to JMA. The earthquake caused significant building and human damage in Hualien City and its vicinity. The focal mechanism of this earthquake was a reverse fault type with a pressure axis in the WNW-ESE direction.
The JMA issued a tsunami warning for Okinawa Island, Miyako Island, and the Yaeyama region at 9:01 on the same day, downgraded to a tsunami advisory at 10:40, and finally canceled it at 12:00. This earthquake caused tsunamis of 27 cm in Kubura, Yonaguni Island, 25 cm in Hirara, Miyako Island, and 17 cm in Ishigaki Port, Ishigaki Island. The tsunami amplitude at nearby DART stations were up to 1 cm (52404 and 52405).
In this study, we simulated the tsunami caused by the 2024 earthquake, assuming a rectangular fault model. The length and width of the fault are 60 km and 30 km, respectively. Following the W-phase moment tensor solution (NP2) of the USGS, we set the fault parameters as, strike: 26°, dip: 58°, rake: 81°, top depth:15 km, average slip: 1.5 m. Assuming a rigidity of 5 x 1010 N/m2, the seismic moment is 1.35 x 1020 Nm (Mw=7.4). The initial condition for the tsunami propagation was calculated using a rectangular fault model (Okada, 1985, BSSA), which is considered as the same with the crustal deformation on the seafloor. The bathymetry data used were resampled to the 12 arc-sec grid data from the GEBCO_2023 15 arc-sec grid data. To calculate tsunami propagation, the linear shallow water equations or the long wave equations were numerically solved using the finite difference method (Satake, 1995, PAGEOPH) with GPGPU. The DART and tide gauge data were downloaded from the NOAA and UNESCO/IOC websites, respectively, and we compared the calculated tsunami waveforms obtained from the simulation with the observed ones. The calculated tsunami waveforms generally reproduced their amplitudes and arrival times well. Since there is a slight phase delay (the observed waveform arrives later than the calculated one) at the somewhat distant DART, it may be necessary to apply phase correction using the method of Watada et al. (2014, JGR).
When an event accompanied by a tsunami occurs, such as the recent Taiwan earthquake, preliminary tsunami simulations are made available on the website (https://iisee.kenken.go.jp/staff/fujii/TsunamiTop.html). On each event page, the tsunami source model, maximum height of simulated tsunami and tsunami waveforms, and the animation video of tsunami propagation are posted.
The JMA issued a tsunami warning for Okinawa Island, Miyako Island, and the Yaeyama region at 9:01 on the same day, downgraded to a tsunami advisory at 10:40, and finally canceled it at 12:00. This earthquake caused tsunamis of 27 cm in Kubura, Yonaguni Island, 25 cm in Hirara, Miyako Island, and 17 cm in Ishigaki Port, Ishigaki Island. The tsunami amplitude at nearby DART stations were up to 1 cm (52404 and 52405).
In this study, we simulated the tsunami caused by the 2024 earthquake, assuming a rectangular fault model. The length and width of the fault are 60 km and 30 km, respectively. Following the W-phase moment tensor solution (NP2) of the USGS, we set the fault parameters as, strike: 26°, dip: 58°, rake: 81°, top depth:15 km, average slip: 1.5 m. Assuming a rigidity of 5 x 1010 N/m2, the seismic moment is 1.35 x 1020 Nm (Mw=7.4). The initial condition for the tsunami propagation was calculated using a rectangular fault model (Okada, 1985, BSSA), which is considered as the same with the crustal deformation on the seafloor. The bathymetry data used were resampled to the 12 arc-sec grid data from the GEBCO_2023 15 arc-sec grid data. To calculate tsunami propagation, the linear shallow water equations or the long wave equations were numerically solved using the finite difference method (Satake, 1995, PAGEOPH) with GPGPU. The DART and tide gauge data were downloaded from the NOAA and UNESCO/IOC websites, respectively, and we compared the calculated tsunami waveforms obtained from the simulation with the observed ones. The calculated tsunami waveforms generally reproduced their amplitudes and arrival times well. Since there is a slight phase delay (the observed waveform arrives later than the calculated one) at the somewhat distant DART, it may be necessary to apply phase correction using the method of Watada et al. (2014, JGR).
When an event accompanied by a tsunami occurs, such as the recent Taiwan earthquake, preliminary tsunami simulations are made available on the website (https://iisee.kenken.go.jp/staff/fujii/TsunamiTop.html). On each event page, the tsunami source model, maximum height of simulated tsunami and tsunami waveforms, and the animation video of tsunami propagation are posted.