2:15 PM - 2:30 PM
[MIS16-03] The 2011 Tohoku-Oki Tsunami Source Model Estimated from Tsunami Trace Heights and Sediment Thickness of Tsunami Deposits
Keywords:2011 Tohoku-Oki earthquake, Tsunami trace height, Tsunami deposit
Historical records without instrumental observation records, especially those before the Middle Ages, are unreliable, and it is not sometimes possible to measure the tsunami run-up limits and arrival points at the time of tsunamis. Although tsunami deposits provide geological evidence of tsunami attacks, the relationship between sediment thickness and tsunami inundation height is complex, and there are many problems in estimating tsunami magnitude from tsunami deposits alone. Therefore, in this study, using the 2011 Tohoku-Oki earthquake tsunami, we examined whether it is possible to estimate the tsunami source model based on numerical simulations of tsunami propagation and tsunami sediment movement, and comparisons of tsunami inundation height, sediment thickness distribution of tsunami deposits and the computations, respectively.
According to the tsunami trace points investigated by the Tohoku Earthquake Tsunami Joint Survey Group, the tsunami inundation heights with trace reliability A in the range of 35 to 41 degrees latitude are 2364 points (Mori et al., 2013). This is too dense for the minimum computational grid intervals of the coast (2 arc-sec; Chikasada, 2020), therefore it was averaged at 0.1 degree steps and set to 60 points. The initial tsunami source of 8 models with different slip amounts and rupture areas (Models 2~9) were prepared based on the 55-subfault model with delayed slip (Model 1) (Satake et al., 2013). JAGURS (e.g., Baba et al. 2015, 2017) and TUNAMI-STM (e.g., Takahashi et al. 1999, Sugawara et al., 2019) were used for numerical simulations of tsunami propagation and sediment transport, respectively. The study sites of the tsunami deposits are the Numanohama marsh, Miyako City, Iwate Prefecture (e.g., Goto et al., 2015), the Sendai plain, Sendai to Yamamoto City, Miyagi Prefecture (e.g., Abe et al., 2012), and the Idagawa lowland, Minamisoma City, Fukushima Prefecture (Kusumoto et al., 2018).
Comparison of tsunami inundation height and the simulated tsunami height shows that Models 1, 4, and 5 are consistent with observations. Models 4 and 5 are the models of 1.5 and 2 times the slip amounts of subfaults without the delayed slip, respectively, suggesting that there is a possibility of overestimating the tsunami magnitude. Next, we compared the observed and calculated thicknesses of tsunami deposits. Models 1 and 2 for the Idagawa lowland and Models 1, 2 and 6 for the Sendai plain are consistent with observations. Here, Model 2 is the 55-subfaults model without the delayed slip and Model 6 is a model without fault rupture in the shallow part along the trench axis, meaning that large slip areas near the trench axis may be missed. Conversely, the Numanohama marsh underestimated from the observation even though numerical simulations were performed with all models. This is probably because the tsunami deposits at Numanohama marsh are mainly composed of gravels, which is inconsistent with the formula of sand movement for suspended load and bedload layers adopted in TUNAMI-STM.
As mentioned above, the magnitude of tsunami can be partially estimated from either the tsunami inundation heights or the tsunami deposits. However, if the initial tsunami wave source is Model 1, both the tsunami inundation heights and the sediment thickness distribution of the tsunami deposits can be successfully explained. Combining numerical simulations of tsunami propagation and sediment transport with tsunami inundation heights and sediment thicknesses of tsunami deposits would improve the accuracy of estimating initial tsunami sources. It should be noted that both the inundation height and tsunami deposits do not include information on the delayed slip of the fault, and the sediment thickness obtained from the sediment transport simulation does not have centimeter-order resolution. Although there are many issues, such as the need to consider that the sediment thickness the tsunami deposits depend on the state of preservation and the advancement of sediment transport simulation, this method can be used to estimate historical earthquakes and tsunamis in the Nankai Trough.
According to the tsunami trace points investigated by the Tohoku Earthquake Tsunami Joint Survey Group, the tsunami inundation heights with trace reliability A in the range of 35 to 41 degrees latitude are 2364 points (Mori et al., 2013). This is too dense for the minimum computational grid intervals of the coast (2 arc-sec; Chikasada, 2020), therefore it was averaged at 0.1 degree steps and set to 60 points. The initial tsunami source of 8 models with different slip amounts and rupture areas (Models 2~9) were prepared based on the 55-subfault model with delayed slip (Model 1) (Satake et al., 2013). JAGURS (e.g., Baba et al. 2015, 2017) and TUNAMI-STM (e.g., Takahashi et al. 1999, Sugawara et al., 2019) were used for numerical simulations of tsunami propagation and sediment transport, respectively. The study sites of the tsunami deposits are the Numanohama marsh, Miyako City, Iwate Prefecture (e.g., Goto et al., 2015), the Sendai plain, Sendai to Yamamoto City, Miyagi Prefecture (e.g., Abe et al., 2012), and the Idagawa lowland, Minamisoma City, Fukushima Prefecture (Kusumoto et al., 2018).
Comparison of tsunami inundation height and the simulated tsunami height shows that Models 1, 4, and 5 are consistent with observations. Models 4 and 5 are the models of 1.5 and 2 times the slip amounts of subfaults without the delayed slip, respectively, suggesting that there is a possibility of overestimating the tsunami magnitude. Next, we compared the observed and calculated thicknesses of tsunami deposits. Models 1 and 2 for the Idagawa lowland and Models 1, 2 and 6 for the Sendai plain are consistent with observations. Here, Model 2 is the 55-subfaults model without the delayed slip and Model 6 is a model without fault rupture in the shallow part along the trench axis, meaning that large slip areas near the trench axis may be missed. Conversely, the Numanohama marsh underestimated from the observation even though numerical simulations were performed with all models. This is probably because the tsunami deposits at Numanohama marsh are mainly composed of gravels, which is inconsistent with the formula of sand movement for suspended load and bedload layers adopted in TUNAMI-STM.
As mentioned above, the magnitude of tsunami can be partially estimated from either the tsunami inundation heights or the tsunami deposits. However, if the initial tsunami wave source is Model 1, both the tsunami inundation heights and the sediment thickness distribution of the tsunami deposits can be successfully explained. Combining numerical simulations of tsunami propagation and sediment transport with tsunami inundation heights and sediment thicknesses of tsunami deposits would improve the accuracy of estimating initial tsunami sources. It should be noted that both the inundation height and tsunami deposits do not include information on the delayed slip of the fault, and the sediment thickness obtained from the sediment transport simulation does not have centimeter-order resolution. Although there are many issues, such as the need to consider that the sediment thickness the tsunami deposits depend on the state of preservation and the advancement of sediment transport simulation, this method can be used to estimate historical earthquakes and tsunamis in the Nankai Trough.