# Data driven weather forecasts trained and initialised directly from observations

arXiv: https://arxiv.org/abs/2407.15586 · HTML: https://mlsanity.com/paper/2407.15586/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 18 references that are in the map.
- First posted on arXiv 22 Jul 2024.

## Builds on (most cited first, 6 of 6)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Skillful Precipitation Nowcasting using Deep Generative Models of Radar](https://mlsanity.com/paper/2104.00954/) (2021)
- [GenCast: Diffusion-based ensemble forecasting for medium-range weather](https://mlsanity.com/paper/2312.15796/) (2023)
- [Deep Learning for Day Forecasts from Sparse Observations](https://mlsanity.com/paper/2306.06079/) (2023)
- [DiffDA: a Diffusion Model for Weather-scale Data Assimilation](https://mlsanity.com/paper/2401.05932/) (2024)
- [FengWu-4DVar: Coupling the Data-driven Weather Forecasting Model with 4D Variational Assimilation](https://mlsanity.com/paper/2312.12455/) (2023)

## Built on by (most cited first, 19 of 19)

- [Prithvi WxC: Foundation Model for Weather and Climate](https://mlsanity.com/paper/2409.13598/) (2024)
- [Can AI weather models predict out-of-distribution gray swan tropical cyclones?](https://mlsanity.com/paper/2410.14932/) (2024)
- [FuXi Weather: A data-to-forecast machine learning system for global weather](https://mlsanity.com/paper/2408.05472/) (2024)
- [Diffusion-LAM: Probabilistic Limited Area Weather Forecasting with Diffusion](https://mlsanity.com/paper/2502.07532/) (2025)
- [Satellite Observations Guided Diffusion Model for Accurate Meteorological States at Arbitrary Resolution](https://mlsanity.com/paper/2502.07814/) (2025)
- [WeatherReal: A Benchmark Based on In-Situ Observations for Evaluating Weather Models](https://mlsanity.com/paper/2409.09371/) (2024)
- [Accurate deep learning-based filtering for chaotic dynamics by identifying instabilities without an ensemble](https://mlsanity.com/paper/2408.04739/) (2024)
- [4D-Var using Hessian approximation and backpropagation applied to automatically-differentiable numerical and machine learning models](https://mlsanity.com/paper/2408.02767/) (2024)
- [OMG-HD: A High-Resolution AI Weather Model for End-to-End Forecasts from Observations](https://mlsanity.com/paper/2412.18239/) (2024)
- [Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data](https://mlsanity.com/paper/2410.06731/) (2024)
- [CERA: A Framework for Improved Generalization of Machine Learning Models to Changed Climates](https://mlsanity.com/paper/2509.00010/) (2025)
- [Skillful high-resolution weather forecasting independent of physical models](https://mlsanity.com/paper/2605.28153/) (2026)
- [Automated Spatio-Temporal Weather Modeling for Load Forecasting](https://mlsanity.com/paper/2409.16326/) (2024)
- [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)
- [Probability calibration for precipitation nowcasting](https://mlsanity.com/paper/2510.00594/) (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)
- [Earth-o1: A Grid-free Observation-native Atmospheric World Model](https://mlsanity.com/paper/2605.06337/) (2026)
- [PeakWeather: MeteoSwiss Weather Station Measurements for Spatiotemporal Deep Learning](https://mlsanity.com/paper/2506.13652/) (2025)
- [From Drops to Grid: Noise-Aware Spatio-Temporal Neural Process for Rainfall Estimation](https://mlsanity.com/paper/2605.05912/) (2026)

## Abstract (arXiv)

Skilful Machine Learned weather forecasts have challenged our approach to numerical weather prediction, demonstrating competitive performance compared to traditional physics-based approaches. Data-driven systems have been trained to forecast future weather by learning from long historical records of past weather such as the ECMWF ERA5. These datasets have been made freely available to the wider research community, including the commercial sector, which has been a major factor in the rapid rise of ML forecast systems and the levels of accuracy they have achieved. However, historical reanalyses used for training and real-time analyses used for initial conditions are produced by data assimilation, an optimal blending of observations with a physics-based forecast model. As such, many ML forecast systems have an implicit and unquantified dependence on the physics-based models they seek to challenge. Here we propose a new approach, training a neural network to predict future weather purely from historical observations with no dependence on reanalyses. We use raw observations to initialise a model of the atmosphere (in observation space) learned directly from the observations themselves. Forecasts of crucial weather parameters (such as surface temperature and wind) are obtained by predicting weather parameter observations (e.g. SYNOP surface data) at future times and arbitrary locations. We present preliminary results on forecasting observations 12-hours into the future. These already demonstrate successful learning of time evolutions of the physical processes captured in real observations. We argue that this new approach, by staying purely in observation space, avoids many of the challenges of traditional data assimilation, can exploit a wider range of observations and is readily expanded to simultaneous forecasting of the full Earth system (atmosphere, land, ocean and composition).