Computationally Efficient Hybrid Downscaling of Surf Zone Hydrodynamics: Methodology and Evaluation

• Published in: Journal of geophysical research. Machine learning and computation Vol. 2; no. 2
• Main Authors: Echevarria, E. R., Contardo, S., Pérez‐Díaz, B., Hoeke, R. K., Leighton, B., Trenham, C., Cagigal, L., Méndez, F. J.
• Format: Journal Article
• Language: English
• Published: Wiley 01.06.2025
• Subjects: hybrid downscaling; machine learning; nearshore; surf‐zone; XBeach
• DOI: 10.1029/2024JH000523
• ISSN: 2993-5210

We present a hybrid surf‐zone model that combines numerical simulations and statistical/machine learning techniques, enabling accurate calculations of nearshore wave and hydrodynamic parameters with high computational efficiency. The approach involves defining representative forcing conditions, carrying out numerical model (XBeach) simulations for these cases, and training machine learning models capable of predicting selected model output variables. Data decomposition via Empirical Orthogonal Function analysis further simplifies the process, reducing the output data dimensionality, with minimal accuracy loss (with exception of certain wetting‐drying processes). Three machine learning approaches of increasing complexity are compared: a multi‐variate linear regression (LR), a Radial Basis Functions (RBF) interpolator and a Deep Neural Network (DNN). The LR model fails to account for the complex non‐linearities in coastal wave dynamics, which warrants the use of more complex machine learning techniques. Both the RBF interpolator and the DNN models demonstrate high levels of accuracy in the prediction of short wave heights, mean wavelength, and depth‐averaged currents, with slightly lower accuracy for long (infragravity) wave heights and fraction of breaking waves. The proposed surrogate model thus offers an efficient alternative to computationally expensive numerical model simulations, enabling rapid and reliable long‐period deterministic simulations (multi‐decadal hindcasts) and/or multi‐ensemble probabilistic scenario simulations of nearshore hydrodynamic conditions. We provide a comprehensive description of the implementation details and assess the surrogate model's performance in representing various wave and hydrodynamic parameters. We discuss potential use cases and limitations, noting that this hybrid modeling technique can be adapted for use with other numerical models in various settings. Plain Language Summary This study presents an efficient and accurate approach to modeling nearshore wave conditions, water levels and currents, by combining numerical model simulations with machine learning techniques. Traditional numerical models are computationally intensive, while the hybrid approach presented here is a much faster alternative. The study provides a comprehensive, step‐by‐step guide for building this surrogate model, which involves defining a set of representative conditions and training a machine learning model to capture the relationship between input forcings and output variables (gridded fields of selected parameters). To improve efficiency, an Empirical Orthogonal Function analysis is applied as a data reduction step, simplifying the process without substantially compromising accuracy. This surrogate model demonstrates high precision, and can be used to calculate long‐term climatologies of nearshore wave and hydrodynamic conditions (something that would be prohibitively expensive with the traditional numerical model), or the almost instantaneous production of nearshore forecasts. Key Points A hybrid XBeach model integrates numerical simulations with machine learning for efficient coastal predictions The surrogate model efficiently and accurately predicts nearshore wave and hydrodynamic parameters Data reduction via Empirical Orthogonal Functions provides enhanced efficiency with minimal accuracy loss