U-Cast: A Surprisingly Simple and Efficient Frontier Probabilistic AI Weather Forecaster

2026-04-10

AI Summary

This work addresses the high computational demands and reliance on specialized architectures in state-of-the-art probabilistic weather forecasting models by proposing an efficient approach based on a standard U-Net. The method first employs deterministic pretraining using a Masked Autoencoder (MAE), followed by short-horizon probabilistic fine-tuning with the Continuous Ranked Probability Score (CRPS) loss, and leverages Monte Carlo Dropout to generate ensemble forecasts. Evaluated at 1.5° resolution, the model matches or exceeds the forecast skill of GenCast and the IFS Ensemble Prediction System (ENS), while requiring less than 12 H200 GPU-days for training. Notably, generating a 60-step ensemble forecast takes only 11 seconds, substantially lowering the computational barrier and resource requirements for operational probabilistic weather prediction.

Technology Category

• Machine Learning: Probabilistic Circuits and Graphical Models
• Reasoning under Uncertainty: Probabilistic Programming
• Computer Vision: Diffusion Models for Vision

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

AI-based weather forecasting now rivals traditional physics-based ensembles, but state-of-the-art (SOTA) models rely on specialized architectures and massive computational budgets, creating a high barrier to entry. We demonstrate that such complexity is unnecessary for frontier performance. We introduce U-Cast, a probabilistic forecaster built on a standard U-Net backbone trained with a simple recipe: deterministic pre-training on Mean Absolute Error followed by short probabilistic fine-tuning on the Continuous Ranked Probability Score (CRPS) using Monte Carlo Dropout for stochasticity. As a result, our model matches or exceeds the probabilistic skill of GenCast and IFS ENS at 1.5$^\circ\$ resolution while reducing training compute by over 10$\times$ compared to leading CRPS-based models and inference latency by over 10$\times$ compared to diffusion-based models. U-Cast trains in under 12 H200 GPU-days and generates a 60-step ensemble forecast in 11 seconds. These results suggest that scalable, general-purpose architectures paired with efficient training curricula can match complex domain-specific designs at a fraction of the cost, opening the training of frontier probabilistic weather models to the broader community. Our code is available at: https://github.com/Rose-STL-Lab/u-cast.

Problem

• probabilistic weather forecasting
• computational efficiency
• barrier to entry
• AI-based forecasting
• training cost

Innovation

• U-Net
• probabilistic forecasting
• CRPS
• Monte Carlo Dropout
• efficient training