Probabilistic Tsunami Hazard Assessment along Nankai Trough (1) An assessment based on the information of the forthcoming earthquake that Earthquake Research Committee(2013) evaluated

Monday, 14 December 2015
Poster Hall (Moscone South)
Kenji Hirata1, Hiroyuki Fujiwara1, Hiromitsu Nakamura2, Masaki Osada3, Nobuyuki Morikawa3, Shin'ichi Kawai2, Tatsuo Ohsumi2, Shin Aoi1, Naotaka Yamamoto1, Hisanori Matsuyama4, Nobuhiko Toyama5, Tadashi Kito5, Yoichi Murashima6, Yasuhiro Murata7, Takuya Inoue7, Ryu Saito7, Jyunpei Takayama7, Shinichi Akiyama8, Mariko Korenaga8, Yuta Abe8 and Norihiko Hashimoto8, (1)National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan, (2)NIED, Tsukuba, Japan, (3)NIED National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan, (4)OYO Corporation, Tokyo, Japan, (5)OYO Corporation, Tsukuba, Japan, (6)KKC, Tokyo, Japan, (7)KOKUSAI KOGYO CO., LTD, Tokyo, Japan, (8)ITOCHU Techno-Solutions Corporation, Chiyoda-ku, Japan
The Earthquake Research Committee(ERC)/HERP, Government of Japan (2013) revised their long-term evaluation of the forthcoming large earthquake along the Nankai Trough; the next earthquake is estimated M8 to 9 class, and the probability (P30) that the next earthquake will occur within the next 30 years (from Jan. 1, 2013) is 60% to 70%.

In this study, we assess tsunami hazards (maximum coastal tsunami heights) in the near future, in terms of a probabilistic approach, from the next earthquake along Nankai Trough, on the basis of ERC(2013)’s report. The probabilistic tsunami hazard assessment that we applied is as follows;

(1) Characterized earthquake fault models (CEFMs) are constructed on each of the 15 hypothetical source areas (HSA) that ERC(2013) showed. The characterization rule follows Toyama et al.(2015, JpGU). As results, we obtained total of 1441 CEFMs.

(2) We calculate tsunamis due to CEFMs by solving nonlinear, finite-amplitude, long-wave equations with advection and bottom friction terms by finite-difference method. Run-up computation on land is included.

(3) A time predictable model predicts the recurrent interval of the present seismic cycle is T=88.2 years (ERC,2013). We fix P30 = 67% by applying the renewal process based on BPT distribution with T and alpha=0.24 as its aperiodicity.

(4) We divide the probability P30 into P30(i) for i-th subgroup consisting of the earthquakes occurring in each of 15 HSA by following a probability re-distribution concept (ERC,2014). Then each earthquake (CEFM) in i-th subgroup is assigned a probability P30(i)/N where N is the number of CEFMs in each sub-group. Note that such re-distribution concept of the probability is nothing but tentative because the present seismology cannot give deep knowledge enough to do it. Epistemic logic-tree approach may be required in future.

(5) We synthesize a number of tsunami hazard curves at every evaluation points on coasts by integrating the information about 30 years occurrence probabilities P30(i) for all earthquakes (CEFMs) and calculated maximum coastal tsunami heights. In the synthesis, aleatory uncertainties relating to incompleteness of governing equations, CEFM modeling, bathymetry and topography data, etc, are modeled assuming a log-normal probabilistic distribution.

Examples of tsunami hazard curves will be presented.