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

TL;DR

U-Cast is a simple, efficient probabilistic weather forecaster using a standard U-Net, achieving state-of-the-art performance with significantly reduced computational costs and training time.

Contribution

The paper introduces U-Cast, a straightforward probabilistic weather forecasting model that rivals complex models while being computationally cheaper and easier to train.

Findings

U-Cast matches or exceeds the probabilistic skill of leading models.

It reduces training compute by over 10 times.

It trains in under 12 GPU-days and forecasts in 11 seconds.

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.