Author(s): Kenichiro Kobayashi

Linked Author(s): Kenichiro Kobayashi

Keywords: Ensembles; High-resolution; HPC; Shallow water equation; Tokyo

Abstract: In computational engineering research, complex numerical modeling and simulations are often considered essential for academic recognition. Top-tier journals tend to favor work that appears rigorous, even if reviewers cannot sometimes fully assess its merit. Yet, truth often emerges from deceptively simple practices -such as conducting high-resolution simulations repeatedly. Higher resolution reveals previously unseen phenomena, and extensive iterations yield statistically meaningful results. This principle underpins technological advances in satellite imagery, microscopy, MRI, and CT scans. Similarly, statistical analysis relies on expert judgment and large datasets; without sufficient data, reliability declines. This study embraces that perspective, firstly demonstrating the power of high-resolution computation. We present a megacity-scale simulation of Tokyo’s 23 wards, focusing on inundation depth. One of the case studies includes subsidence caused by earthquake-induced liquefaction. Additional case studies include integration with traffic simulations, evaluation of rice paddy dam effects, coupling with basin runoff models, and nesting techniques. We also showcase mega ensemble simulations of the Kumagawa River in Japan. Grounded in high-performance computing (HPC) skills partially acquired during doctoral studies in Germany, this work highlights the precision-driven simulations that have been conducted in Japan for the past 20 years.

DOI:

Year: 2026