Japan Geoscience Union Meeting 2024

Presentation information

[J] Poster

S (Solid Earth Sciences ) » S-TT Technology & Techniques

[S-TT39] Creating future of solid Earth science with high performance computing (HPC)

Sun. May 26, 2024 5:15 PM - 6:45 PM Poster Hall (Exhibition Hall 6, Makuhari Messe)

convener:Takane Hori(Japan Agency for Marine-Earth Science and Technology), Yuji Yagi(Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba), Katsuhiko Shiomi(National Research Institute for Earth Science and Disaster Resilience), Takanori Matsuzawa(National Research Institute for Earth Science and Disaster Resilience)

5:15 PM - 6:45 PM

[STT39-P02] Viscoelastic stress loading in southwest Japan due to the 1944 Tonankai and 1946 Nankai earthquakes based on a 3-D heterogeneous rheological model

*Akinori Hashima1, Takane Hori2, Takeshi Iinuma2, Sota Murakami3, Kohei Fujita3, Tsuyoshi Ichimura3 (1.Tokyo Gakugei University, 2.Japan Agency for Marine-Earth Science and Technology, 3.Earthquake Research Institute, University of Tokyo)

Keywords:Stress loading, Nankai trough, Philippine Sea plate, Viscoelasticity, Finite element method, Intra-plate earthquake

Under the Nankai Trough, M8-class earthquakes have occurred repeatedly at intervals of 100 to 200 years, and there is concern about their occurrence in the near future. From the historical documents on earthquake occurrence, seismic activity in the inland areas of southwestern Japan increases from 50 years before to 10 years after the occurrence of a Nankai megathrust earthquake. In order to assess such inland seismic risk, it is important to show the mechanical linkage between the occurrence of a megathrust earthquake and the occurrence of inland earthquakes. For this purpose, we perform stress loading calculations for the inland source faults of the 1944 Tonankai and 1946 Nankai earthquakes. As for the postseismic deformation, viscoelastic relaxation in the asthenosphere has been proved dominant, as observed aftter the 2011 Tohoku earthquake. The studies on its postseismic deformation have shown the need to consider the low-viscosity layer below the slab as well as the mantle wedge under the overriding plate. Therefore, the stress calculations in this study use a large-scale 3-D finite element model that incorporates the plausible viscosity structure. Such a large-scale model has a large number of degrees of freedom and long computation time. In this study, we attempted to reduce the computation time by efficiently using a recent large-scale computing system for crustal deformation calculations. The calculations showed that stress changes of more than 0.1 MPa occurred over a wide area of inland western Japan. However, the effect of viscoelastic relaxation was hardly observed for 4 years. On the other hand, the stress distribution within the slab was significantly affected by the viscous relaxation of the low-viscous layer below the slab. Using the computed stress field, we compute Coulomb rupture stress (ΔCFS) on the source faults in southwest Japan. ΔCFS was positive in the Chubu and Kinki regions after the Tonankai earthquake and positive in the Kinki region and faults around the Median Tectonic Line after the Nankai earthquake. Faults in Kyushu were consistently negative. The relationship of ΔCFS with faults that caused past damaging earthquakes showed that the damaging earthquakes on faults with positive ΔCFS occurred within about 10 years after the Tonankai–Nankai megathrust events in 1854 or in 1944 and 1946. On the other hand, the damaging earthquakes on faults with negative ΔCFS occurred in the 50 years before the 1944 and 1946 events or in recent years (since 1995). The stress loading on the fault due to the Tonankai and Nankai earthquakes is at most 0.1 MPa, which is more than an order of magnitude smaller than the usual stress drop values. Nevertheless, we can see the clear relationship stated above. This might be because the stress accumulation process throughout the Nankai seismic cycle and hence the timing of reaching the rupture strength are clearly different for faults with positive and negative ΔCFS.