Unstructured Grid-Based River-Coastal Ocean Circulation Modeling towards a Digital Twin of the Seto Inland Sea

As computational techniques advance, the scope of digital twins (DTs) is expanding to encompass entire cities, oceans, or even the Earth. Digital twins of oceans can provide highly comprehensive insights and predictions, thus enabling better-informed decision-making regarding ocean-related activities and management. Here, a numerical model of the Seto Inland Sea (SIS), Japan, was built as a basis to establish a digital twin of the SIS. Spatially varying filtering parameters and grid resolutions were applied to facilitate the robust and accurate simulation of coastal and oceanic processes even under varying extreme conditions. The modeling results were validated using observational datasets from forty-two tidal stations, one mooring system, and thirteen water thermometers. The results represented tidal variations, with NRMSE values below 0.15 and R-2 values exceeding 0.87 at all tidal stations. The NRMSE and R-2 values for currents were approximately 0.14 and 0.76, respectively. The model reproduced the extreme storm surge event causing a sea level rise of 1.5 m near Osaka City resulting from Typhoon Jebi in 2018. The model was shown to enable analyses of complex circulations and hazards in the SIS by accurately replicating barotropic and baroclinic processes. After additional modules are added, this model will serve as a basis for constructing a digital twin of the SIS.

Automated summary

As computational techniques advance, digital twins (DTs) are expanding their scope to include entire cities, oceans, or even the Earth. Digital twins of oceans provide comprehensive insights and predictions, enabling better-informed decision-making in ocean-related activities and management. In this study, a numerical model of the Seto Inland Sea (SIS) in Japan was built to establish a digital twin of the SIS. This model accurately simulated coastal and oceanic processes under extreme conditions, validated by observational datasets and successfully reproducing extreme storm surge events. By adding additional modules, this model will serve as the basis for constructing a digital twin of the SIS.