A Probabilistic U-Net Approach to Downscaling Climate Simulations
This addresses the computational bottleneck in climate modeling for researchers needing high-resolution data, but it is an incremental adaptation of existing methods.
The paper tackled the problem of downscaling coarse climate simulations to finer resolutions needed for impact studies by adapting a probabilistic U-Net approach. The result showed that WMSE-MS-SSIM training objective performed well for extremes under certain settings, while afCRPS better captured spatial variability across scales.
Climate models are limited by heavy computational costs, often producing outputs at coarse spatial resolutions, while many climate change impact studies require finer scales. Statistical downscaling bridges this gap, and we adapt the probabilistic U-Net for this task, combining a deterministic U-Net backbone with a variational latent space to capture aleatoric uncertainty. We evaluate four training objectives, afCRPS and WMSE-MS-SSIM with three settings for downscaling precipitation and temperature from $16\times$ coarser resolution. Our main finding is that WMSE-MS-SSIM performs well for extremes under certain settings, whereas afCRPS better captures spatial variability across scales.