17:15 〜 19:15
[MIS11-P12] Inversion of multi-subfault parameters for the Meiwa and Okinawa Sakishima tsunamis based on the distribution of tsunami deposits in the caves on Ishigaki Island
キーワード：津波堆積物、石垣島、洞窟堆積物、逆解析
The southern Ryukyu Islands, including the Ishigaki Island, have been inundated by large-scale tsunamis. The Meiwa and Okinawa Sakishima tsunamis are the two most considerable events estimated to have been caused by M8 class earthquakes in 1771 CE and ~2000 years ago, respectively. The historical documents implied that the Meiwa tsunami was ~30 m in the maximum run-up height around Ishigaki Island's southern and western regions, and the Okinawa Sakishima tsunami is also supposed to be a scale comparable to the Meiwa tsunami, considering the spatial distribution of the tsunami boulders transported by it. However, the tsunami deposits have recently been discovered in the caves on Ishigaki Island, suggesting that those tsunamis had much larger inundation heights. Radiocarbon dating of those deposits exhibited ages around 1771 CE and 1800 BP, implying that the Meiwa and Okinawa Sakishima tsunamis deposited them. The opening of those caves was 40 meters in maximum elevation, which is higher than the previous estimation of the inundation heights of those tsunamis. Thus, it is necessary to reconsider the scale of the earthquakes that caused these tsunamis. This study aimed at reconstructing the slip distributions of seismic faults of both tsunamis' sources based on the historical and geologic records in the caves in the Ishigaki Islands. We explored the fault slip distributions that best explain the spatial distributions of the inundated caves and regions described in the historical records. To do so, the tsunami simulation model JAGURS was employed as a forward model to predict the extent of tsunami inundation. This study considered two types of fault models with multiple (two and eight) segments exhibiting different slip lengths in the inter-plate fault plane along the southern part of the Ryukyu Trench, whereas the previous studies assumed a single-segment fault model. Random search and Bayesian optimization were used to optimize fault parameters. As a result, it was revealed that the fault model with large localized slips ranging from 45 to 60 m near the trench axis could reproduce the inundation extent of the Meiwa tsunami based on both historical and geological records. Also, the inversion result based solely on tsunami deposits in the caves suggested that the Sakishima tsunami could have been caused by the fault exhibiting localized large slips similar to the Meiwa tsunami's fault model. The slip lengths of the two paleo-tsunami source faults along the Ryukyu Trench estimated in this study were nearly equivalent to the maximum slip lengths in the observational record of the subduction zone earthquakes, such as the 2011 Tohoku earthquake. These results updated the fault model for the Meiwa tsunami and newly proposed a model for the Okinawa Sakishima tsunami. Understanding the slip distribution of the source faults of paleo-tsunamis will contribute to future tsunami disaster prevention for the southern Ryukyu Trench.