Development of an Ensemble Regional Tide and Storm Surge Model (ETSM) for the Coasts of Korea

The Korean peninsula's intricate coastline, distinguished by its numerous islands and extreme tidal variations, is particularly susceptible to storm surges-especially when typhoons coincide with high tides, posing significant threats to both life and property. Accurate prediction of storm surges is crucial for mitigating these potential disasters. Current operational forecasting systems, implemented by various agencies, typically rely on single atmospheric model inputs as their forcing mechanism. However, such deterministic approaches often exhibit considerable variability in their predictions due to inherent uncertainties in atmospheric modeling processes.

This study focuses on developing an Ensemble Regional Tidal and Storm Surge Model (ETSM), integrating forecast outputs from 26 atmospheric ensemble members to enhance storm surge predictions along the coasts of Korea. The model was evaluated through case studies of major typhoons, including HINNAMNOR (2022) and KHANUN (2023), by comparing observed water levels against both deterministic and ensemble model predictions. The deterministic model tended to both overestimate and underestimate, whereas the ensemble spread encompassed the observed water levels, demonstrating that the ensemble model provided better predictions in representing actual storm surge events.

Additionally, the study analyzed the characteristics of storm surge heights observed during the summer of 2024, establishing threshold values. The analysis revealed significant regional variations in storm surge height distributions across the West, South, and East Coasts of Korea. To assess the model's predictive performance, probabilistic validation using Brier Scores and ROC (Receiver Operating Characteristic) metrics was performed. The results indicate reliable predictive performance while also revealing for further improvement.