Numerical analysis of the paleotsunami sizes through simulation of a large coralline boulder movement at Ishigaki Island, Japan

Sakishima Islands (southern Ryukyu Islands), Japan, have been frequently affected by large tsunamis through the historical ages. The AD1771 Meiwa tsunami was an extreme event and the run-up height was ~30 m in maximum. The source model of the AD1771 Meiwa tsunami is still controversial: (1) earthquake plus landslide, (2) tsunami earthquake, and (3) splay fault models. The islands were also affected by the prehistoric tsunamis according to the geological evidence. For example, numerous coral boulders that were transported by paleotsunamis, including AD1771 event, were reported at the islands. The corals that constitute the boulders had once lived in the coral reefs fringing the islands and died when they were cast ashore to the land. Therefore, it is possible to estimate the timing of their deposition by using radiocarbon dating technique. In fact, the radiocarbon dating of Porites tsunami boulders at Sakishima Islands indicates that tsunamis had occurred repeatedly within a 150-400 years interval. However, the sizes of each tsunami are still uncertain. Hence, an important question has remained: whether there were any tsunami(s) in the past that were equivalent or larger in size comparing with the AD1771 Meiwa tsunami. In this study, we numerically investigated the transportation of a very large paleotsunami boulder at the southeastern part of Ishigaki Island, one of Sakishima Islands, using a boulder transport model in order to evaluate the sizes of paleotsunamis prior to the AD1771 event. The boulder that we used is called 'tsunami ufu-ishi' and it is the largest coralline boulder (>500 t) in Ishigaki Island. It is deposited approx. 100 m from the shoreline and 10 m in elevation. It was probably transported from the coral reef by the tsunami around 2000 years ago according to the radiocarbon dating. However, it is not necessarily that the boulder reached its present position due to one tsunami event 2000 years ago. In fact, according to the paleomagnetic analysis, it is proposed that the boulder was rotated twice possibly by two tsunami events. Alternatively, we cannot fully exclude the possibility that the boulder reached its present position due to one tsunami event. Therefore, we assumed that the boulder was deposited either by the two tsunamis or one tsunami. The result of the calculation showed that if we assumed tsunami ufu-ishi was moved by two tsunami events, the sizes of each tsunami are almost equivalent to the AD1771 Meiwa tsunami. In contrast, if we assume that the boulder was deposited by the single tsunami event, the flow depth at the boulder's position should have been 2-4 times higher than that of the AD1771 Meiwa tsunami. However, it is unlikely that such a size of tsunami was generated because the slip amount of the fault is unusually large in this case. Therefore, we interpreted that the AD1771 Meiwa tsunami was not a special event and similar size of the tsunamis had occurred repeatedly in this region.