# GraphDOP: Towards skilful data-driven medium-range weather forecasts learnt and initialised directly from observations

arXiv: https://arxiv.org/abs/2412.15687 · HTML: https://mlsanity.com/paper/2412.15687/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Dec 2024.

## Built on by (most cited first, 27 of 27)

- [Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation](https://mlsanity.com/paper/2504.18720/) (2025)
- [Deep Learning and Foundation Models for Weather Prediction: A Survey](https://mlsanity.com/paper/2501.06907/) (2025)
- [Observation-guided Interpolation Using Graph Neural Networks for High-Resolution Nowcasting in Switzerland](https://mlsanity.com/paper/2509.00017/) (2025)
- [Huracan: A skillful end-to-end data-driven system for ensemble data assimilation and weather prediction](https://mlsanity.com/paper/2508.18486/) (2025)
- [Learning Coupled Earth System Dynamics with GraphDOP](https://mlsanity.com/paper/2510.20416/) (2025)
- [Towards accurate extreme event likelihoods from diffusion model climate emulators](https://mlsanity.com/paper/2605.03802/) (2026)
- [Skillful high-resolution weather forecasting independent of physical models](https://mlsanity.com/paper/2605.28153/) (2026)
- [XiChen: A global weather observation-to-forecast machine learning system via four-dimensional variational gradient-guided flexible assimilation](https://mlsanity.com/paper/2507.09202/) (2025)
- [Learning more physically realistic dynamics in machine-learning based weather forecasting with latent-space constraints](https://mlsanity.com/paper/2510.04006/) (2025)
- [DAWP: A framework for global observation forecasting via Data Assimilation and Weather Prediction in satellite observation space](https://mlsanity.com/paper/2510.15978/) (2025)
- [Probabilistic Transformers for Joint Modeling of Global Weather Dynamics and Decision-Centric Variables](https://mlsanity.com/paper/2601.03753/) (2026)
- [Operator Learning for Smoothing and Forecasting](https://mlsanity.com/paper/2603.20359/) (2026)
- [Earth-o1: A Grid-free Observation-native Atmospheric World Model](https://mlsanity.com/paper/2605.06337/) (2026)
- [Towards a Foundation Model for the Martian Atmosphere](https://mlsanity.com/paper/2605.28851/) (2026)
- [On the use of Graphs for Satellite Image Time Series](https://mlsanity.com/paper/2505.16685/) (2025)
- [FuXiWeather2: Learning accurate atmospheric state estimation for operational global weather forecasting](https://mlsanity.com/paper/2603.15358/) (2026)
- [Toward Artificial Intelligence Enabled Earth System Coupling](https://mlsanity.com/paper/2604.03289/) (2026)
- [AIFS-COMPO: A Global Data-Driven Atmospheric Composition Forecasting System](https://mlsanity.com/paper/2604.03300/) (2026)
- [Earth Science Foundation Models: From Perception to Reasoning and Discovery](https://mlsanity.com/paper/2605.12542/) (2026)
- [Accurate, Efficient, and Explainable Deep Learning Approaches for Environmental Science Problems](https://mlsanity.com/paper/2605.19366/) (2026)
- [Multimodal Spatiotemporal Atmospheric Data Assimilation with Latent Video Flow-matching](https://mlsanity.com/paper/2608.05103/) (2026)
- [Timestep-Conditioned Transformers for Global Weather Forecasting](https://mlsanity.com/paper/2608.06241/) (2026)
- [Earth observation embeddings are effective sub-grid descriptors for probabilistic weather downscaling](https://mlsanity.com/paper/2608.12271/) (2026)
- [OceanDepths: A Global Dataset of Paired Subsurface and Surface Ocean Observations](https://mlsanity.com/paper/2608.16373/) (2026)
- [Uncertainty-Aware End-to-End AI Weather Forecasting: Disentangling Observation and Model Contributions](https://mlsanity.com/paper/2608.30795/) (2026)
- [The interface of data assimilation and machine learning](https://mlsanity.com/paper/2610.07496/) (2026)
- [Artificial intelligence pathways from weather to climate](https://mlsanity.com/paper/2610.09770/) (2026)

## Abstract (arXiv)

We introduce GraphDOP, a new data-driven, end-to-end forecast system developed at the European Centre for Medium-Range Weather Forecasts (ECMWF) that is trained and initialised exclusively from Earth System observations, with no physics-based (re)analysis inputs or feedbacks. GraphDOP learns the correlations between observed quantities - such as brightness temperatures from polar orbiters and geostationary satellites - and geophysical quantities of interest (that are measured by conventional observations), to form a coherent latent representation of Earth System state dynamics and physical processes, and is capable of producing skilful predictions of relevant weather parameters up to five days into the future.