Numerical Simulations Using Various Models for Tsunamis Due to a Fluid or Rigid Bodies Falling Down a Uniform Slope

• Published in: Journal of disaster research Vol. 16; no. 7; pp. 994 - 1004
• Main Authors: Kakinuma, Taro, Yanagihara, Mitsuru, Iribe, Tsunakiyo, Nagai, Kuninori, Hara, Chisato, Hamada, Natsuki, Nakagaki, Tatsuya, Sujatmiko, Karina Aprilia, Magdalena, Ikha, Nagai, Kaori, Kannonji, Rika, Chen, Songgui, Shirai, Tomoki, Arikawa, Taro
• Format: Journal Article
• Language: English
• Published: 01.10.2021
• ISSN: 1881-2473; 1883-8030
• DOI: 10.20965/jdr.2021.p0994

Tsunami generation due to a landslide has been simulated using various numerical models, and the resulting water surface displacements from the models, as well as the corresponding experimental data, are compared. The numerical models used in this study are a two-layer long-wave model, a two-level non-hydrostatic model, a three-dimensional model, a lattice-Boltzmann-type model, an SPH-type model, and an MPS-type model. Tsunamis generated by a fluid falling down a uniform slope are accurately reproduced by the models, especially when the wave height of the tsunami is not large. When using the two-layer long-wave model, in which the two layers of a falling fluid and seawater are assumed not to mix, the parameters including the seabed friction coefficient, adjusted in one case, are not appropriate for other mixing conditions. The two-level model with non-hydrostatic pressure exhibits wave disintegration owing to the effects of both nonlinearity and dispersion, although the second wave generated by the reflection of a wave traveling towards the shore is not simulated accurately. Tsunamis caused by a group of rigid cylinders falling down a uniform slope have also been simulated using two Lagrangian models, namely the SPH- and MPS-type models. Although the first peak at the water level is accurately reproduced by both models, the water level at the trough between the first and second crests is overestimated.