April 22 at 16:00 - 18:10, 2021 (JST)
  • Dr. Iyan Mulia (Research Scientist, Prediction Science Laboratory, RIKEN Cluster for Pioneering Research (CPR))
  • via Zoom

Dedicated tsunami observing systems are mostly expensive and are often not sustainable. Therefore, alternative approaches should be implemented to overcome the issues. We introduced innovative ways to observe tsunamis using existing instrumentation available on unconventional platforms such as commercial vessels and airplanes. Our study demonstrated that the accuracy of the proposed observing systems is adequate for detecting large tsunamis offshore. The use of such systems is expected to provide more cost-effective and sustainable observations for the future. Additionally, we also developed a tsunami forecasting system based on machine learning to improve or complement the conventional methods that typically require considerable computational resources. On the contrary, the main appealing feature of the machine learning is the computational speed that would be suitable for a real-time prediction of tsunami inundation or flooding. We found that the application of machine learning can significantly improve the computing time without sacrificing the accuracy compared to the conventional methods.

*Please contact Keita Mikami's mail address to get access to the Zoom meeting room.

Related News