Interpolation-Free Deep Learning for Meteorological Downscaling on Unaligned Grids across Multiple Domains with Application to Wind Power

Abstract As climate change intensifies, the shift to cleaner energy sources becomes increasingly urgent. With wind energy production set to accelerate, reliable wind probabilistic forecasts are essential to ensure its efficient use. However, since numerical weather prediction models are computationally expensive, probabilistic forecasts are produced at resolutions too coarse to capture all mesoscale wind behaviors. Statistical downscaling, typically applied to enhance the resolution of climate model simulations, presents a viable solution with lower computational costs by learning a mapping from low-resolution (LR) variables to high-resolution (HR) meteorological variables. Leveraging deep learning, we evaluate a downscaling model based on a state-of-the-art U-Net architecture, applied to an ensemble member from a coarse-scale probabilistic forecast of wind velocity. The architecture is modified to incorporate 1) a learned grid alignment strategy to resolve LR–HR grid mismatches and 2) a processing module for multilevel atmospheric predictors. To extend the downscaling model’s applicability from fixed spatial domains to the entire Canadian region, we assess a transfer learning approach. Our results show that the learned grid alignment strategy performs as well as conventional preprocessing interpolation steps and that LR wind speed at multiple levels is sufficient as a predictor, enabling a more compact architecture. Additionally, they suggest that extending to new spatial domains using transfer learning is promising and that downscaled wind velocities demonstrate potential in improving the detection of wind power ramps, a critical phenomenon for wind energy. Significance Statement We use deep learning algorithms, a subfield of artificial intelligence, to mimic high-resolution wind forecasts using low-resolution meteorological forecasts. The low- and high-resolution forecasts lie on different unaligned grids. As an alternative to conventional numerical weather forecasting systems, our approach uses less than 5 times the memory of state-of-the-art deep learning algorithms by relying on a learned grid alignment procedure that does not compromise performance. We also demonstrate the ability to apply our model across multiple scattered domains in Canada without performance loss and extend its use to any region in Canada through transfer learning. We believe this research paves the way for models covering multiple countries (i.e., foundation models) without requiring excessive training data.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.