Storm Surge in Color: RGB-Encoded Physics-Aware Deep Learning for Storm Surge Forecasting

TL;DR

This paper introduces a novel RGB-encoded deep learning approach using ConvLSTM networks for storm surge forecasting, integrating physical drivers and structured data to improve accuracy and generalization over large coastal regions.

Contribution

It presents a new method that converts unstructured water elevation data into RGB images, enabling deep learning models to better forecast storm surges with physical context.

Findings

Achieves robust 48-hour surge forecasts across Texas coast

Demonstrates strong spatial extensibility to other coastal regions

Enhances forecast accuracy by integrating wind and bathymetry data

Abstract

Storm surge forecasting plays a crucial role in coastal disaster preparedness, yet existing machine learning approaches often suffer from limited spatial resolution, reliance on coastal station data, and poor generalization. Moreover, many prior models operate directly on unstructured spatial data, making them incompatible with modern deep learning architectures. In this work, we introduce a novel approach that projects unstructured water elevation fields onto structured Red Green Blue (RGB)-encoded image representations, enabling the application of Convolutional Long Short Term Memory (ConvLSTM) networks for end-to-end spatiotemporal surge forecasting. Our model further integrates ground-truth wind fields as dynamic conditioning signals and topo-bathymetry as a static input, capturing physically meaningful drivers of surge evolution. Evaluated on a large-scale dataset of synthetic storms in the Gulf of Mexico, our method demonstrates robust 48-hour forecasting performance across multiple regions along the Texas coast and exhibits strong spatial extensibility to other coastal areas. By combining structured representation, physically grounded forcings, and scalable deep learning, this study advances the frontier of storm surge forecasting in usability, adaptability, and interpretability.