Test operation of a real-time tsunami inundation forecast system using actual data observed by S-net
Abstract
If the tsunami inundation information can be rapidly and stably forecast before the large tsunami attacks, the information would have effectively people realize the impeding danger and necessity of evacuation. Toward that goal, we have developed a prototype system to perform the real-time tsunami inundation forecast for Chiba prefecture, eastern Japan, using off-shore ocean bottom pressure data observed by the seafloor observation network for earthquakes and tsunamis along the Japan Trench (S-net) (Aoi et al., 2015, AGU). Because tsunami inundation simulation requires a large computation cost, we employ a database approach searching the pre-calculated tsunami scenarios that reasonably explain the observed S-net pressure data based on the multi-index method (Yamamoto et al., 2016, EPS). The scenario search is regularly repeated, not triggered by the occurrence of the tsunami event, and the forecast information is generated from the selected scenarios that meet the criterion. Test operation of the prototype system using the actual observation data started in April, 2017 and the performance and behavior of the system during non-tsunami event periods have been examined. It is found that the treatment of the noises affecting the observed data is the main issue to be solved toward the improvement of the system. Even if the observed pressure data are filtered to extract the tsunami signals, the noises in ordinary times or unusually large noises like high ocean waves due to storm affect the comparison between the observed and scenario data. Due to the noises, the tsunami scenarios are selected and the tsunami is forecast although any tsunami event does not actually occur. In most cases, the selected scenarios due to the noises have the fault models in the region along the Kurile or Izu-Bonin Trenches, far from the S-net region, or the fault models below the land. Based on the parallel operation of the forecast system with a different scenario search condition and examination of the fault models, we improve the stability and performance of the forecast system.This work was supported by Council for Science, Technology and Innovation(CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP), "Enhancement of societal resiliency against natural disasters"(Funding agency: JST).
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFMNH23A0252S
- Keywords:
-
- 3225 Numerical approximations and analysis;
- MATHEMATICAL GEOPHYSICS;
- 4332 Disaster resilience;
- NATURAL HAZARDS;
- 4341 Early warning systems;
- NATURAL HAZARDS;
- 4564 Tsunamis and storm surges;
- OCEANOGRAPHY: PHYSICAL