2:00 PM - 2:15 PM
[MIS16-02] Numerical estimation of tsunami sources of tsunami deposits in Onuma, Minamisanriku Town, Miyagi Prefecture
Keywords:tsunami deposit, numerical simulation, Sanriku coast, Jogan tsunami
The Sanriku coast has been repeatedly inundated by large-scale tsunamis, as exemplified by the 2011 CE Tohoku-oki tsunami, 1933 CE Showa-sanriku tsunami, and 1896 CE Meiji-sanriku tsunami. In addition to these historical earthquake tsunamis, prehistoric tsunami deposits have been reported in some areas (e.g., Goto et al., 2015; Takada et al., 2016; Ishimura and Miyauchi, 2017), but their sources have not been discussed. Ishimura and Miyauchi (2017) conducted trench and geoslicer surveys in the coastal lowland of Onuma, Minami-Sanriku Town, Miyagi Prefecture, on the southern Sanriku coast. This paper reports six tsunami deposits in the mud sequence, divided into gravelly tsunami deposits (G1–G3) composed of rounded slate and sandy tsunami deposits (S1–S3) containing shell fragments. In this study area, beaches are composed of gravel in the north and sand in the south, implying that each tsunami deposit was transported from different sources. Given that the gravelly tsunami deposits were sourced from the north gravelly beach, they were formed by a tsunami with a maximum run-up height of more than 12 m, which exceeded the hill located north side of the trench site. In contrast, the sandy tsunami deposits were formed by a tsunami with a run-up height of 5–12 m, which would have overtopped the southern beach ridge and dune, but not the north hill. The grain size of tsunami deposits probably reflects differences in tsunami size. The ages of the tsunami deposits are 680–1010 cal. yr BP (S1), 1540–2350 cal. yr BP (G1), 2730–3630 cal. yr BP (G2), 3850–4080 cal. yr BP (G3), 4450–4780 cal. yr BP (S2), and 5540–5730 cal. yr BP (S3) (Ishimura, 2019). If the tsunami sources of these tsunami deposits are clarified, it will better understand the size and sources of prehistoric tsunamis on the Sanriku coast.
In this study, we conducted numerical tsunami simulations by JAGURS (Baba et al., 2015) using the following existing fault models: the 2011 CE Tohoku-oki tsunami (Satake et al., 2011), the 1933 CE Sanriku tsunami (Tanioka and Satake, 1996), the 1896 CE Sanriku tsunami (Aida, 1977), and the 869 CE Jogan tsunami (Namegaya et al., 2010; Namegaya and Satake, 2014). The tsunami impact on Onuma was evaluated based on whether it inundated the coastal lowland and crossed the southern beach ridge and dune or the north hill. The 2011 CE Tohoku-oki tsunami model inundated the lowland over the northern hill, whereas the 1933 CE Showa-sanriku and the 1896 CE Meiji-sanriku models resulted in no inundation of the lowland. The 869 CE Jogan tsunami model (Namegaya and Satake, 2014), which was proposed with reference to the inundation limit of the 2011 CE Tohoku-oki tsunami, showed that the tsunami slightly crossed the northern hill and inundated the lowland at least from the south for any fault width model . This result is inconsistent with the lack of Jogan tsunami deposits between To-a (915 CE) and G1 (1540–2350 cal. yr BP) in the sediment core at Onuma, suggesting two possibilities: (1) the Jogan tsunami inundated Onuma but did not form tsunami deposit and (2) the proposed Jogan tsunami model (Namegaya and Satake, 2014) is an overestimation. It is noted the current numerical estimation in this study used modern topographic data, which may not reproduce the topography at the time. In this area, the 2011 Tohoku-oki tsunami eroded the southern beach and formed a channel extending inland, implying the calculations are likely based on topographic conditions more prone to inundation. In addition, the elevation in this area is likely to have been higher than present because of the long-term subsidence trend (Ishimura and Miyauchi, 2017). Therefore, it is necessary to perform numerical tsunami simulations again using the topographic data reproducing the original situation to examine the cause of the above discrepancy.
In this study, we conducted numerical tsunami simulations by JAGURS (Baba et al., 2015) using the following existing fault models: the 2011 CE Tohoku-oki tsunami (Satake et al., 2011), the 1933 CE Sanriku tsunami (Tanioka and Satake, 1996), the 1896 CE Sanriku tsunami (Aida, 1977), and the 869 CE Jogan tsunami (Namegaya et al., 2010; Namegaya and Satake, 2014). The tsunami impact on Onuma was evaluated based on whether it inundated the coastal lowland and crossed the southern beach ridge and dune or the north hill. The 2011 CE Tohoku-oki tsunami model inundated the lowland over the northern hill, whereas the 1933 CE Showa-sanriku and the 1896 CE Meiji-sanriku models resulted in no inundation of the lowland. The 869 CE Jogan tsunami model (Namegaya and Satake, 2014), which was proposed with reference to the inundation limit of the 2011 CE Tohoku-oki tsunami, showed that the tsunami slightly crossed the northern hill and inundated the lowland at least from the south for any fault width model . This result is inconsistent with the lack of Jogan tsunami deposits between To-a (915 CE) and G1 (1540–2350 cal. yr BP) in the sediment core at Onuma, suggesting two possibilities: (1) the Jogan tsunami inundated Onuma but did not form tsunami deposit and (2) the proposed Jogan tsunami model (Namegaya and Satake, 2014) is an overestimation. It is noted the current numerical estimation in this study used modern topographic data, which may not reproduce the topography at the time. In this area, the 2011 Tohoku-oki tsunami eroded the southern beach and formed a channel extending inland, implying the calculations are likely based on topographic conditions more prone to inundation. In addition, the elevation in this area is likely to have been higher than present because of the long-term subsidence trend (Ishimura and Miyauchi, 2017). Therefore, it is necessary to perform numerical tsunami simulations again using the topographic data reproducing the original situation to examine the cause of the above discrepancy.