# WeatherDreamer: Learning a World Model from Multimodal Earth Observations

DevFeed: [WeatherDreamer: Learning a World Model from Multimodal Earth Observations](<https://devfeed.tech/articles/weatherdreamer-learning-a-world-model-from-multimodal-earth-observations-135895.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.18393>)

Author: Kate Duffy, Elise Penn, Tristan Ballard, August Posch, Thomas J. Vandal

Published: 2026-11-16T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [World models](<https://devfeed.tech/topics/world-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>)

Tags: [analysis](<https://devfeed.tech/tags/analysis.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [earth](<https://devfeed.tech/tags/earth.md>), [forecast](<https://devfeed.tech/tags/forecast.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [global](<https://devfeed.tech/tags/global.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [inference](<https://devfeed.tech/tags/inference.md>), [irradiance](<https://devfeed.tech/tags/irradiance.md>), [irregular](<https://devfeed.tech/tags/irregular.md>), [learning](<https://devfeed.tech/tags/learning.md>), [location](<https://devfeed.tech/tags/location.md>), [matching](<https://devfeed.tech/tags/matching.md>), [microwave](<https://devfeed.tech/tags/microwave.md>), [minutes](<https://devfeed.tech/tags/minutes.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [operational](<https://devfeed.tech/tags/operational.md>), [physics-ao-ph](<https://devfeed.tech/tags/physics-ao-ph.md>), [world-model](<https://devfeed.tech/tags/world-model.md>)

## AI overview

WeatherDreamer learns data assimilation and forecasting together from observations across 18 global observing systems, without reanalysis. It carries a latent atmospheric state forward hourly, updates it with new observations, and predicts measurements for any observing system and location. At held-out stations, its ensembles outperform NOAA's GEFS for temperature, dew point, and wind through 12 hours and are better calibrated than GEFS and ECMWF's ensemble. Short-lead estimates of precipitation, cloud-top height, and solar irradiance--variables not used as inference inputs--match or exceed operational products. Its departures from NOAA's GSI first guess are close for surface stations and moisture-sensitive microwave channels, but up to five times larger for temperature-sounding channels. The model is not given observation errors; its learned updates assume errors close to known errors for microwave sounders and scatterometer winds. It produces an hourly global analysis in about 1.5 minutes on one GPU.

## Source excerpt

arXiv:2611.18393v1 Announce Type: new Abstract: Machine learning weather models now rival operational global forecasts but still start from analyses produced by separate, expensive data assimilation (DA) systems, which for global models are typically updated every six hours. Observation-to-observation models avoid this dependence, but most omit the recursive filtering cycle that lets DA systems carry a background state through sparse, irregular observations. We introduce WeatherDreamer, a world model that learns this cycle end-to-end from 18 global observing systems, without reanalysis. It infers a latent state from observations, carries it forward hour by hour with learned dynamics, corrects it each hour with new observations, and can predict what any of its observing systems would measure at any location. At held-out stations, its ensembles beat NOAA's Global Ensemble Forecast System (GEFS) for temperature, dew point and wind out to 12 h and are better calibrated than GEFS and the ECMWF ensemble. Precipitation, cloud-top height and solar irradiance, never inputs at inference, are inferred with skill matching or exceeding operational products at short leads. Against the first guess of NOAA's Gridpoint Statistical Interpolation (GSI) DA system, WeatherDreamer's short-range departures are within 10% of GSI's at surface stations and within 25% for moisture-sensitive microwave channels, but up to five times larger for temperature-sounding channels. Never given observation errors, its learned update assumes errors close to known instrument and retrieval errors of microwave sounders and scatterometer winds. WeatherDreamer also separates viewing geometry from atmospheric state, recovering 83-96% of observed limb darkening and producing seamless nadir-view imagery. Producing an hourly global analysis in about 1.5 minutes on one GPU, WeatherDreamer shows that assimilation and forecasting can be learned together directly from observations.