日本地球惑星科学連合2023年大会

講演情報

[J] オンラインポスター発表

セッション記号 A (大気水圏科学) » A-CG 大気海洋・環境科学複合領域・一般

[A-CG42] 陸域から沿岸域における水・土砂動態

2023年5月24日(水) 10:45 〜 12:15 オンラインポスターZoom会場 (8) (オンラインポスター)

コンビーナ:有働 恵子(東北大学大学院工学研究科)、浅野 友子(東京大学)、木田 新一郎(九州大学・応用力学研究所)、山崎 大(東京大学生産技術研究所)

現地ポスター発表開催日時 (2023/5/23 17:15-18:45)

10:45 〜 12:15

[ACG42-P09] Effect of spatial resolution on the performance of large-scale hydrodynamic models

*Prakat Modi1Menaka Revel1、Xudong Zhou1Dai Yamazaki1 (1.Institute of Industrial Science, The University of Tokyo, Japan)

キーワード:Hydrodynamic Model, Spatial Resolution, Model Evaluation

Large-scale hydrodynamic modelling is essential in flood prediction for planning adaptation and mitigation measures. Hydrodynamic models' performance gets affected due to spatial resolution and other uncertainties associated with model structure, parameters, and input datasets. Many studies have been performed to quantify the effect of spatial resolution suggested improvement in the accuracy of the river's hydrodynamics with high-resolution simulations. Even though these researchers showed the places of discrepancies by comparing model simulations with the observation, they had limitations in finding out the primary changes in the model's hydrodynamic behaviour due to spatial resolution differences.
Here using the catchment-based macroscale floodplain (CaMa-Flood) hydrodynamic model, we performed the detailed comparison of 01 min, 03 min, 06 min and 15 min spatial resolution simulations to know 1. the places showing discrepancies between simulations, and 2. the leading cause behind these discrepancies. We found that the water surface elevation was more comparable for various spatial resolutions than the discharge. Discharge for the mainstream of the Amazon river showed that the higher resolution has lower high flow and higher low flow compared to the coarser resolution simulation. Even though the discharge magnitude had minor differences, few places showed a significant difference in the peak flow timing causing large discrepancies in the model performance. These places need more careful consideration and further investigations to improve the accuracy of model representability.