Uncertainty-Aware End-to-End AI Weather Forecasting: Disentangling Observation and Model Contributions
Summary
This paper introduces a probabilistic approach to end-to-end AI weather forecasting by enhancing the Aardvark Weather model with learned observation noise and Monte Carlo dropout, improving mean forecasts by 4.2% on average and separating aleatoric and epistemic uncertainty.
View Cached Full Text
Cached at: 09/01/26, 11:41 AM
Paper page - Uncertainty-Aware End-to-End AI Weather Forecasting: Disentangling Observation and Model Contributions
Source: https://huggingface.co/papers/2608.30795
Abstract
The Aardvark Weather model is made probabilistic by adding learned observation noise and Monte Carlo dropout to its encoder and processor, yielding a calibrated nested ensemble that improves mean forecasts and separates aleatoric and epistemic uncertainty.
End-to-end weather forecasting systems produce skillful global gridded and station forecasts directly from raw Earth observations, replacing the numerical weather prediction pipeline, including data assimilation, at a fraction of its cost. These systems are deterministic and issue no uncertainty. Here we render the Aardvark Weather model probabilistic by attaching one stochastic mechanism to each component: learned, input-dependent noise at theobservation encoder, capturingaleatoric uncertaintyinherited from the observing system, andMonte Carlo dropoutin the processor, capturingepistemic uncertaintyin the learned dynamics. The resultingnested ensembleattributes forecast spread to the two sources through alaw-of-total-variance decomposition, cross-checked by withholding observation streams.Probabilistic finetuningsignificantly improves the mean forecast, by 4.2% on average across variables and lead times. The ensemble is calibrated against ERA5 through the medium range (spread-skill ratio 0.98), keeps station RMSE within 2.4% of the deterministic model while beating it in CRPS at every lead time, and trails the operational ECMWF ensemble. The encoder branch behaves as observation-driven uncertainty. Component-attributed uncertainty makes end-to-end forecasts more transparent, a step toward observation-driven digital twins of the atmosphere.
View arXiv pageView PDFProject pageAdd to collection
Get this paper in your agent:
hf papers read 2608\.30795
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2608.30795 in a model README.md to link it from this page.
Datasets citing this paper1
#### rodrigoalmeida1994/uqe2e Updatedabout 2 hours ago • 50
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2608.30795 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Otter Weather: Skillful and Computationally Efficient Medium-Range Weather Forecasting
Otter Weather is a computationally efficient AI model for medium-range weather forecasting that outperforms numerical weather prediction baselines and frontier AI models while requiring significantly less training compute, aiming to democratize high-performance weather prediction.
AdaWeather: Adaptively Mixing Probabilistic Weather Forecasts with Logarithmic Regret
Introduces AdaWeather, an adaptive framework that combines multiple probabilistic weather forecasts using machine learning and mixture of experts, achieving logarithmic regret compared to the best static mixture of experts and showing empirical improvements in temperature forecasting.
Scalable Uncertainty Quantification for Extreme Weather Forecasting via Empirical Neural Tangent Kernels
The paper proposes Neural Tangent Kernel-based uncertainty quantification for deterministic deep learning weather models, achieving sharper adaptive prediction intervals during extreme events without retraining.
Probabilistic bias adjustment of seasonal forecasts using generative machine learning: A case study of Arctic sea ice predictions
This paper presents a probabilistic post-processing framework using conditional variational autoencoders (cVAEs) to bias-adjust seasonal forecasts of Arctic sea ice, improving calibration, sharpness, and spectral power over standard methods.
The Frictions That Make AI Forecasting Hard (8 minute read)
An article discussing why AI forecasting is difficult, emphasizing that capability gains often fail to translate into end-to-end impact due to social, institutional, and tacit bottlenecks, using radiology as a case study.