WNP-SWH reproducibility package: code, WAVEWATCH III configuration, and data information

Reproducibility package for the study 'Enhancing Significant Wave Height Simulations Using Distribution-Aware Deep Learning and Environmental Predictors' (JGR: Machine Learning and Computation). It contains the complete preprocessing, training, inference, evaluation, and figure/table plotting scripts; the trained model weights (best_model.pth); the study-specific WAVEWATCH III (version 6.07) coarse and fine grid configuration; model configuration and column mapping; bathymetry grid inputs; and data availability information. The preprocessing script randomly divides the 2018-2021 samples into training and validation sets at an 8:2 ratio, with 2022-2023 as the independent test set, consistent with the manuscript. The WAVEWATCH III source is not included because the official NOAA release does not permit third-party redistribution.