Regional Modeling of Storm Surges Using Localized Features and Transfer Learning

• Published in: Journal of geophysical research. Machine learning and computation Vol. 2; no. 3
• Main Authors: Nagaraj, Meghana, Rodriguez, Alejandra, Wahl, Thomas
• Format: Journal Article
• Language: English
• Published: 01.09.2025
• ISSN: 2993-5210; 2993-5210
• DOI: 10.1029/2025JH000650

Storm surges induced by low pressure and high winds from tropical or extratropical cyclones are the main driver of major coastal flooding events. While tide gauges provide the most accurate sea level observations, their records are often short, spatially uneven, and contain gaps, posing challenges for a detailed analysis of surge characteristics continuously along the coastline. To perform extreme value analysis for hazard and risk assessment requires long time series of storm surges, which do often not exist. In this study, we use a regional long short‐term memory (LSTM) model to predict storm surges at multiple tide gauges simultaneously, while also accounting for localized features. Storm surge observations from 31 tide gauges across Florida are used as a predictand; atmospheric and oceanic data are used as dynamic predictors; and shelf width, nearshore slope, shelter factor, and tidal range are used as static predictors. Results demonstrate that incorporating local static attributes improves model performance. Pearson correlation increases by 20% and RMSE drops by 22%, on average, in the cross‐validation. The pretrained model captures localized surge patterns at gauged and ungauged locations making it suitable to simulate storm surge while accounting for local bathymetric features. Comparison with a state‐of‐the‐art hydrodynamic numerical model hindcast shows that the regional LSTM approach performs better in most locations. The proposed modeling framework is efficient and can derive hindcasts and future projections of storm surges facilitating robust analysis of trends, decadal variations, and the calculation of return levels with reduced uncertainty. Coastal areas are increasingly at risk due to rising sea levels and extreme weather events, which lead to devastating flooding. Storm surges, caused by high winds and low pressures from cyclones, are a major driver of such flooding events. This study investigates the use of a machine learning model called long short‐term memory (LSTM) to predict storm surges at multiple tide gauges along Florida's coast. We used data from tide gauges, as well as climate and oceanic variables, and local characteristics to improve storm surge predictions. Results showed that by incorporating local features for each tide gauge, such as shelf width and shelter factor, the model's performance improved. This modeling framework can be used for more accurate storm surge predictions in gauged and ungauged locations, helping to improve flood risk assessments. Regional LSTM model predicts storm surges at multiple tide gauges simultaneously, improving performance including local static features Proposed model outperforms hydrodynamic models and at‐site models, capturing surge patterns at both gauged and ungauged locations The regional LSTM model can offer a robust framework for modeling storm surges at various spatial scales