Reservoir-Triggered Seismicity in Brazil: Characteristics and new cases

Prompt and accurate tsunami forecast is essentially important to reduce the fatalities of onshore residences due to tsunami, and the 2011 megathrust Tohoku-oki earthquake reinforced the importance of offshore observation for warning of near-field tsunami. Under the sponsorship of MEXT, Japanese government, NIED is now constructing S-net (Seafloor Observation Network for Earthquakes and Tsunamis along the Japan Trench) where offshore observations were insufficient at the time of the Tohoku-oki earthquake. This network consists of 150 ocean bottom observatories linked by fiber optic cables and covers the wide area along the Japan Trench and the southernmost Kuril Trench from Kanto to Hokkaido. In this presentation, we develop a new methodology of real-time tsunami inundation forecast. We also construct tsunami scenario bank (TSB) prior to earthquakes which contains thousands of inundation scenarios generated using various earthquake source models. When tsunami occurs, forecast is carried out by comparing observed offshore ocean bottom pressure data and pre-calculated data. An advantage of our method is that tsunami inundation is estimated without any earthquake source information which is difficult to correctly estimate in real-time. Due to the high-cost calculation for the inundation simulations, the number of the tsunami scenarios stored in TSB is limited. We therefore perform sensitivity analyses to select the tsunami scenarios which efficiently cover possible tsunami scenarios affecting the coastal region where the real-time tsunami forecast will be issued. To select the appropriate scenarios from TSB, we propose an evaluation algorithm based on the combination of three different indexes calculated from maximum absolute values of continuously observed pressure changes. As S-net is under construction and not in operation, performance of our algorithm is examined using simulated tsunami waveforms.