11:00 AM - 11:15 AM
[SSS14-08] Estimation of earthquake history along the Sagami Trough using marine event deposits
Keywords:turbidite, plate subduction zone, Kanto earthquakes
Event deposits such as turbidites on the deep seafloor are used to estimate the timing and interval of earthquakes. However, turbidite deposition is not necessarily caused by gravity flows accompanying landslides that occur during earthquakes, but can also be formed by the supply of suspended matter from rivers during floods and by waves. In this study, piston core samples were obtained from the slope base west of the Miura Knoll in Sagami Bay (PC01 site) and the trench landward slope off Kujukuri beach (PC02 site), where there is little possibility of flood-origin inflow, during the Shinsei Maru KS-13-T5 cruise. Both samples consist of olivegreen hemipelagic silty mud containing 2 to 5 fine to coarse sand layers per meter. The analysis was conducted using an X-ray CT scanner, a magnetic susceptibility meter, and non-destructive analysis using ITRAX, and the age was estimated using the radiocarbon age of planktonic foraminifera. The event layers were identified based on core observations, CT images, and changes in element concentrations in ITRAX, and the deposition age of each event layer was estimated using radiocarbon dating. At depths of 153 cm and 304 cm in PC01, layers containing pumice were observed. Based on the ages, they are estimated to be the Kozushima-Tenjosan tephra (AD 838; Kobayashi et al., 2007, Bull. Geol. Survey Japan, etc.) and the Amagi-Kawagodaira tephra (3,137-3,160 cal BP; Tani, 2013, Radiocarbon), respectively. Event layers with normal grading were recognized in 9 layers of the PC01 core and 26 layers of the PC02 core. The ages of these event layers were compared with the Kanto earthquake history estimated by the Headquarters for Earthquake Research Promotion (2014) and Komori (2023). The event layers in the PC01 core correspond mostly to the earthquakes occurred around Sagami Bay, and in the stratum where the existence of an earthquake is not inferred from the land data, no event layer is also recognized in the core sample. For the earthquakes possibly off Boso, the event layer corresponding to the earthquakes about 3000 years ago is not recognized in the PC02 core, but the event layers correspond to the other earthquakes in general.