WindHCT: context and local hybrid CNN Transformer for wind speed super-resolution
Global climate models operate at coarse spatial resolutions that limit their ability to represent localized atmospheric phenomena, motivating deep learning–based super-resolution (SR) for meteorological fields. However, most existing SR models are designed for natural images and fail to fully capture the characteristic...