Japan Geoscience Union Meeting 2023

Presentation information

[J] Oral

H (Human Geosciences ) » H-DS Disaster geosciences

[H-DS06] Tsunami and tsunami forecast

Tue. May 23, 2023 10:45 AM - 12:00 PM 106 (International Conference Hall, Makuhari Messe)

convener:Satoko Murotani(National Museum of Nature and Science), Toshitaka Baba(Graduate School of Science and Technology, Tokushima University), Chairperson:Hiroaki Tsushima(Meteorological Research Institute, Japan Meteorological Agency), Satoko Murotani(National Museum of Nature and Science)

11:30 AM - 11:45 AM

[HDS06-14] Tsunami forecasting method using time-averaged ocean bottom pressure data

*Oishi Tatsuto1, Yuichiro Tanioka1, Yusuke Yamanaka1 (1.Hokkaido University)


Keywords:Real time tsunami forecasting, S-net

The 2011 Tohoku earthquake caused devastating tsunami damage in Japan. Thus, a dense cabled observation network, called the seafloor observation network for earthquakes and tsunami along the Japan Trench (S-net), was installed in 2016. Since then, the tsunami forecasting methods using S-net data have been developed vigorously. Especially, tsunami assimilation techniques or waveform inversion techniques to obtain initial tsunami conditions were newly developed. However, the ocean bottom pressure data are always contaminated by non-tsunami components such as sea-bottom acceleration change, ocean acoustic wave or P wave due to the earthquake. Therefore, it is necessary to separate tsunami components from those non-tsunami components before those recent tsunami forecasting methods are applied.
In this study, we developed a new tsunami forecasting method using the time-averaged ocean bottom pressure data. Hence, a pre-data process to eliminate non-tsunami components from original data is unnecessary. To test our method, tsunamis from the various fault models distributed along the plate interface along the Japan Trench were first computed based on a linear theory. Then, time averages of computed ocean bottom pressure data at S-net stations were calculated by changing a length of time average window. Second, initial tsunami conditions, initial ocean surface deformations, were estimated using interpolation of those time averaged ocean pressure data at S-net stations. Third, forecasted tsunamis were computed from those estimated initial tsunami conditions. Finally, forecasted tsunami waveforms along coasts were compared with the tsunami waveforms computed from the fault model at first by using the maximum tsunami height ratio and a variance reduction (VRC).
We found that this tsunami prediction method, while generally performing well, slightly underestimates the tsunami heights along coasts. In particular, for tsunamis generated near the Japan Trench, where the observation stations are sparse, the maximum tsunami height along coasts was underestimated by a factor of two, and the VRC was ~40-60%. However, for tsunamis generated near coasts, which can be observed by a sufficient number of S-net, the maximum tsunami height along coasts was approximately coincidence, and the VRC was ~80%, which was generally a reasonable reproduction of the tsunami waveforms along coasts.
Therefore, although this method needs to be improved for tsunamis in areas near the Japan Trench, where S-net stations are a few, it can be used for tsunamis in areas with enough S-net stations. Thus, we will correct the underestimation of tsunamis generated in areas with sparse S-net stations and develop an immediate tsunami prediction method for a wider area, including the Japan Trench.