企画セッション: 計算力学×データサイエンス

[2B13-18-03] Improving sequential Bayesian update for tsunami scenario detection by using geodetic data learning

*Reika Nomura¹, Yu Otake¹, Shuji Moriguchi¹, Diego Melgar², Randall LeVeque^3,1, Kenjiro Terada¹ (1. Tohoku University, 2. University of Oregon, 3. University of Washington)

Keywords:Tsunami scenario detection, Bayesian update, GNSS geodetic data

In this study, we investigate the effectiveness of geodetic data monitored by GNSS (Global Navigation Satellite System) as the prior learning to observational ocean data for the improved tsunami scenario detection. For the case study targeting Nankai-trough, 600 earthquake/tsunami scenarios are generated by GeoClaw and fakequake software. With the synthetically generated geodetic displacements by fakequake software, we examine the reasonable initial probability setting which is prior to the learning on the ocean wave data.