Universal Diffusion-Based Probabilistic Downscaling

2026-02-12

AI Summary

This study addresses the limitations of low-resolution deterministic weather forecasts, which lack fine-scale spatial detail and uncertainty quantification. To overcome this, the authors propose a general-purpose conditional diffusion model trained on paired coarse-resolution forecasts and high-resolution reanalysis data (~25 km → ~5 km) to enable zero-shot probabilistic downscaling. Notably, the framework operates without requiring model-specific fine-tuning for different upstream systems—whether AI-based or numerical weather prediction models—thus establishing the first unified probabilistic downscaling approach. Within a 90-hour forecast horizon, the ensemble mean of the downscaled outputs consistently outperforms the original deterministic forecasts, achieving significantly improved Continuous Ranked Probability Scores (CRPS) and effectively enhancing both spatial detail and probabilistic forecasting skill.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Diffusion Models for Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

We introduce a universal diffusion-based downscaling framework that lifts deterministic low-resolution weather forecasts into probabilistic high-resolution predictions without any model-specific fine-tuning. A single conditional diffusion model is trained on paired coarse-resolution inputs (~25 km resolution) and high-resolution regional reanalysis targets (~5 km resolution), and is applied in a fully zero-shot manner to deterministic forecasts from heterogeneous upstream weather models. Focusing on near-surface variables, we evaluate probabilistic forecasts against independent in situ station observations over lead times up to 90 h. Across a diverse set of AI-based and numerical weather prediction (NWP) systems, the ensemble mean of the downscaled forecasts consistently improves upon each model's own raw deterministic forecast, and substantially larger gains are observed in probabilistic skill as measured by CRPS. These results demonstrate that diffusion-based downscaling provides a scalable, model-agnostic probabilistic interface for enhancing spatial resolution and uncertainty representation in operational weather forecasting pipelines.

Problem

• probabilistic downscaling
• weather forecasting
• diffusion model
• spatial resolution
• uncertainty quantification

Innovation

• diffusion-based downscaling
• probabilistic forecasting
• zero-shot transfer
• model-agnostic
• weather prediction