5:15 PM - 6:45 PM
[HDS11-P06] Estimation of fault rupture zone and submarine landslides scale for the 1596 CE Keicho Bungo earthquake by numerical tsunami simulations
Keywords:The 1596 CE Keicho Bungo earthquake, Numerical tsunami simulation, Submarine active fault, Submarine landslide
Inland crustal earthquakes can generate tsunamis in cases where the epicenter is located on the seafloor. However, the tsunamis generated by the inland crustal earthquakes are observed less frequently than those generated by the plate boundary earthquakes, and uncertainties remain about the source and scale of such tsunamis. The 1596 CE Keicho Bungo earthquake is the inland crustal earthquake with M 7.0±1/4 that occurred in the submarine active faults distributed in Beppu Bay in southwest Japan (Usami, 1987). According to the historical records, this earthquake generated tsunami that inundated the entire coastal area of Beppu Bay. The tsunami heights are estimated to have been 4 to 5 m inside Beppu Bay, about 6 m at the bay mouth, and about 2 m at Usuki outside the bay (Matsusaki et al., 2022). Ishibe and Shimazaki (2005) conducted the numerical tsunami simulations based on the normal fault models distributed in Beppu Bay. The results could not explain the tsunami heights in the historical records. They proposed the possibilities that the fault rupture zone was not located only on the normal faults, but also extended to the right-lateral strike-slip faults distributed in the northeast direction from Beppu Bay, and the tsunami was generated by secondary factors such as a submarine landslide. However, these possibilities have not yet been studied. We attempted to estimate the fault rupture zone and the submarine landslide scale of the 1596 CE Keicho Bungo earthquake by the numerical tsunami simulations using the fault models and the submarine landslide models.
The tsunami simulation code JAGURS (Baba et al., 2015) was used for this study. The fault models were based on the model the same scale as the 1596 CE Keicho Bungo earthquake (Oita Prefecture, 2013) and the fault locations from the active fault database (AIST, 2021). We computed 12 simulations by dividing the normal faults distributed around Beppu Bay into 5 segments and the right-lateral strike-slip faults distributed in the northeast direction from Beppu Bay into 4 segments, and combining the various segments with any slip amount. The submarine landslide models were set up based on the landslide topography identified on the seafloor of Beppu Bay by MEXT and Kyoto University (2015) and applied to the semiempirical equation of Sabeti and Heidarzadeh (2020) to obtain the initial water heights. The topography were computed, off Beppu (600 m wide, 1000-1200 m long, 2-6 m thick) and off Oita (3000 m wide, 1300 m long, 10 m thick), with the initial water heights set at 4 m and 8 m.
We compared the tsunami heights computed with those estimated from historical records. The normal fault models of Mw 7.2, which was inundated the entire coastal area of Beppu Bay, but the tsunami heights was 2 to 3 m, which was lower than the historical record. On the other hand, the models actived at Mw 7.3, generated tsunami heights of 3 to 6 m, thus reproducing better. The models with active normal and right-lateral strike-slip faults, most of the tsunamis generated by the right-lateral strike-slip faults propagated out of Beppu Bay, thus the tsunami heights inside Beppu Bay were almost the same as those computed only with normal faults. As for the submarine landslide models, the landslide tsunamis did not cause much inundation. The landslide tsunami generation with a time lag of 30 to 60 minutes between the tsunami generated by the fault movement, the tsunami was amplified in the south coast of Beppu Bay. In the south coast, the results of the fault models alone were more than 1 m below the tsunami height estimated from the historical record, and the combination of the submarine landslide models reproduced the historical record better. These results suggested the possibility that the tsunami generated by the 1596 CE Keicho Bungo earthquake was mainly due to the fault movement and that the tsunami caused by the submarine landslide amplified the tsunami locally.
The tsunami simulation code JAGURS (Baba et al., 2015) was used for this study. The fault models were based on the model the same scale as the 1596 CE Keicho Bungo earthquake (Oita Prefecture, 2013) and the fault locations from the active fault database (AIST, 2021). We computed 12 simulations by dividing the normal faults distributed around Beppu Bay into 5 segments and the right-lateral strike-slip faults distributed in the northeast direction from Beppu Bay into 4 segments, and combining the various segments with any slip amount. The submarine landslide models were set up based on the landslide topography identified on the seafloor of Beppu Bay by MEXT and Kyoto University (2015) and applied to the semiempirical equation of Sabeti and Heidarzadeh (2020) to obtain the initial water heights. The topography were computed, off Beppu (600 m wide, 1000-1200 m long, 2-6 m thick) and off Oita (3000 m wide, 1300 m long, 10 m thick), with the initial water heights set at 4 m and 8 m.
We compared the tsunami heights computed with those estimated from historical records. The normal fault models of Mw 7.2, which was inundated the entire coastal area of Beppu Bay, but the tsunami heights was 2 to 3 m, which was lower than the historical record. On the other hand, the models actived at Mw 7.3, generated tsunami heights of 3 to 6 m, thus reproducing better. The models with active normal and right-lateral strike-slip faults, most of the tsunamis generated by the right-lateral strike-slip faults propagated out of Beppu Bay, thus the tsunami heights inside Beppu Bay were almost the same as those computed only with normal faults. As for the submarine landslide models, the landslide tsunamis did not cause much inundation. The landslide tsunami generation with a time lag of 30 to 60 minutes between the tsunami generated by the fault movement, the tsunami was amplified in the south coast of Beppu Bay. In the south coast, the results of the fault models alone were more than 1 m below the tsunami height estimated from the historical record, and the combination of the submarine landslide models reproduced the historical record better. These results suggested the possibility that the tsunami generated by the 1596 CE Keicho Bungo earthquake was mainly due to the fault movement and that the tsunami caused by the submarine landslide amplified the tsunami locally.