Enhancing operational wind downscaling capabilities over Canada: Application of a Conditional Wasserstein GAN methodology

Abstract

Abstract Downscaling based on Artificial Intelligence (AI) has been demonstrated to be a powerful tool for improving the spatial resolution of weather forecasts. This study advances wind downscaling by extending the DownGAN framework introduced by Annau et al. to operational Canadian Numerical Weather Predictions (NWP), namely from the Global Deterministic Prediction System (GDPS) to the domain of the High-Resolution Deterministic Prediction System (HRDPS). We enhance the AI model by incorporating high-resolution static covariates, such as topography on the HRDPS grid, into a Conditional Wasserstein Generative Adversarial Network with Gradient Penalty, implemented using a UNET-based generator. Following the DownGAN framework, our methodology integrates low-resolution GDPS forecasts (15 km, 10-day horizon) and high-resolution HRDPS forecasts (2.5 km, 48-hour horizon) with Frequency Separation techniques adapted from computer vision. Through robust training and inference over the whole Canadian region, we demonstrate the operational scalability of our approach, while achieving significant improvements in wind downscaling accuracy. Statistical validation highlights reduction in root mean square error (RMSE) and log spectral distance (LSD) metrics compared to the original DownGAN. High-resolution conditioning covariates and Frequency Separation strategies prove instrumental in enhancing model performance. This work underscores the potential for extending high-resolution wind forecasts beyond the 48-hour horizon, bridging the gap to the 10-day low-resolution global forecast window.