Since the Great East Japan Earthquake in 2011, we have studied stable structures against tsunami overflows and developed a numerical simulation model to predict the drifting of debris generated by the destruction of buildings. However, complex behaviors of tsunami run-up on land and the resulting damages are not fully understood and modeled yet. In addition, regarding a technique for real-time inundation forecast, which is expected to be utilized for evacuation and other purposes, only data from GPS-mounted wave buoys is used; other valuable data is not fully utilized yet.


Therefore, we are aimed for changing disaster prevention and reduction countermeasures into a pre-disaster stage from post-disaster one, and will conduct research to establish resilient coastal zones which can withstand the most severe tsunamis, i.e., to protect lives against the most severe tsunami, to avoid catastrophic socioeconomic damages, and to enable early recovery and reconstruction.

As for the development of a 3-D simulation model for drift behavior and its application to a 2-D simulation model, we conducted the experiments on drift objects in a coastal area including a port and studied the methods of coupling the fluid model and drift object model that are integrated in the 3-D simulation model. In these experiments, we acquired 3-D data using acceleration sensors embedded in drift objects of various shapes and analyzed plane behavior of the drift objects by image analysis.

Aiming to build a multi-observation based tsunami forecasting method, we developed a new tsunami source inversion method that utilizes data-driven bases created by time-reversal imaging, which improved the accuracy of the tsunami source estimation by 14%. In addition, we succeeded in developing a reciprocal-theorem-based technique that reduced the cost of calculating Green's functions, which is necessary for the tsunami source estimation, to approximately 1/100. These new technological inventions allow us to perform the tsunami source analyses with ultrahigh resolution, which are expected to improve the accuracy of tsunami waveform prediction.

Regarding the development of a method for estimating local scour that a tsunami causes around structures, we conducted large-scale hydraulic experiments and performed numerical calculations on local scour that occurs near the breakwater head, and identified the characteristics of the large eddy that occurs around the head. In addition, we conducted experiments on the scour that occurs behind the breakwater due to tsunami overflow and clarified that the rubble mound suppresses overflow-induced scour.

Concerning the application of particle methods to tsunami-induced large deformation of port structures, we developed the multi-physics model based on the Discrete Element Method and a porous model to simulate the behavior of various rigid objects, e.g. mounds, blocks, caissons. In addition, we have developed a novel wave-generation model that considers the incident wave height as the only parameter and built a framework in which the model can be easily connected to 2-D planar simulation models.