# A Probabilistic U-Net Approach to Downscaling Climate Simulations

arXiv: https://arxiv.org/abs/2511.03197 · HTML: https://mlsanity.com/paper/2511.03197/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 16 references that are in the map.
- First posted on arXiv 5 Nov 2025.

## Builds on (most cited first, 7 of 7)

- [Elucidating the Design Space of Diffusion-Based Generative Models](https://mlsanity.com/paper/2206.00364/) (2022)
- [A Probabilistic U-Net for Segmentation of Ambiguous Images](https://mlsanity.com/paper/1806.05034/) (2018)
- [Generative Diffusion-based Downscaling for Climate](https://mlsanity.com/paper/2404.17752/) (2024)
- [DiffESM: Conditional Emulation of Temperature and Precipitation in Earth System Models with 3D Diffusion Models](https://mlsanity.com/paper/2409.11601/) (2024)
- [Algorithmic Hallucinations of Near-Surface Winds: Statistical Downscaling with Generative Adversarial Networks to Convection-Permitting Scales](https://mlsanity.com/paper/2302.08720/) (2023)
- [Multi-variable Hard Physical Constraints for Climate Model Downscaling](https://mlsanity.com/paper/2308.01868/) (2023)
- [Climate Variable Downscaling with Conditional Normalizing Flows](https://mlsanity.com/paper/2405.20719/) (2024)

## Abstract (arXiv)

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× 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.