# ArchesWeather & ArchesWeatherGen: a deterministic and generative model for efficient ML weather forecasting

arXiv: https://arxiv.org/abs/2412.12971 · HTML: https://mlsanity.com/paper/2412.12971/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Placed in Flow matching, time series and science, one of 14 directions in Diffusion models.
- First posted on arXiv 17 Dec 2024.

## Placed in

- Diffusion models / [Flow matching, time series and science](https://mlsanity.com/diffusion/flow-matching-time-series-and-science/)

## Built on by (most cited first, 30 of 32)

- [Fixing the Double Penalty in Data-Driven Weather Forecasting Through a Modified Spherical Harmonic Loss Function](https://mlsanity.com/paper/2501.19374/) (2025)
- [Deep Learning and Foundation Models for Weather Prediction: A Survey](https://mlsanity.com/paper/2501.06907/) (2025)
- [Demystifying Data-Driven Probabilistic Medium-Range Weather Forecasting](https://mlsanity.com/paper/2601.18111/) (2026)
- [Elucidated Rolling Diffusion Models for Probabilistic Forecasting of Complex Dynamics](https://mlsanity.com/paper/2506.20024/) (2025)
- [Swift: An Autoregressive Consistency Model for Efficient Weather Forecasting](https://mlsanity.com/paper/2509.25631/) (2025)
- [U-Cast: A Surprisingly Simple and Efficient Frontier Probabilistic AI Weather Forecaster](https://mlsanity.com/paper/2604.09041/) (2026)
- [Scaling Laws of Global Weather Models](https://mlsanity.com/paper/2602.22962/) (2026)
- [AERIS: Argonne Earth Systems Model for Reliable and Skillful Predictions](https://mlsanity.com/paper/2509.13523/) (2025)
- [LaDCast: A Latent Diffusion Model for Medium-Range Ensemble Weather Forecasting](https://mlsanity.com/paper/2506.09193/) (2025)
- [(Sparse) Attention to the Details: Preserving Spectral Fidelity in ML-based Weather Forecasting Models](https://mlsanity.com/paper/2604.16429/) (2026)
- [Generating ensembles of spatially-coherent in-situ forecasts using flow matching](https://mlsanity.com/paper/2504.03463/) (2025)
- [How to use score-based diffusion in earth system science: A satellite nowcasting example](https://mlsanity.com/paper/2505.10432/) (2025)
- [Warm Starts Accelerate Conditional Diffusion](https://mlsanity.com/paper/2507.09212/) (2025)
- [ArchesClimate: Probabilistic Decadal Ensemble Generation With Flow Matching](https://mlsanity.com/paper/2509.15942/) (2025)
- [AI-boosted rare event sampling to characterize extreme weather](https://mlsanity.com/paper/2510.27066/) (2025)
- [Forecast Skill Is Not Decision Skill: Evidence from Weather-Dependent Decision Tasks](https://mlsanity.com/paper/2512.14779/) (2025)
- [Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting](https://mlsanity.com/paper/2605.06916/) (2026)
- [IndiaWeatherBench: A Dataset and Benchmark for Data-Driven Regional Weather Forecasting over India](https://mlsanity.com/paper/2509.00653/) (2025)
- [Probabilistic Transformers for Joint Modeling of Global Weather Dynamics and Decision-Centric Variables](https://mlsanity.com/paper/2601.03753/) (2026)
- [Super-Resolving Coarse-Resolution Weather Forecasts With Flow Matching](https://mlsanity.com/paper/2604.00897/) (2026)
- [Data-driven Mesoscale Weather Forecasting Combining Swin-Unet and Diffusion Models](https://mlsanity.com/paper/2503.19354/) (2025)
- [Generating time-consistent dynamics with discriminator-guided image diffusion models](https://mlsanity.com/paper/2505.09089/) (2025)
- [Synergistic Neural Forecasting of Air Pollution with Stochastic Sampling](https://mlsanity.com/paper/2510.23977/) (2025)
- [STIPP: Space-time in situ postprocessing over the French Alps using proper scoring rules](https://mlsanity.com/paper/2601.02882/) (2026)
- [FlowDA: Accurate, Low-Latency Weather Data Assimilation via Flow Matching](https://mlsanity.com/paper/2602.06800/) (2026)
- [Ensemble Graph Neural Networks for Probabilistic Sea Surface Temperature Forecasting via Input Perturbations](https://mlsanity.com/paper/2603.06153/) (2026)
- [Accurate, Efficient, and Explainable Deep Learning Approaches for Environmental Science Problems](https://mlsanity.com/paper/2605.19366/) (2026)
- [MotifGen: Spatiotemporal interpolation of misaligned satellite images via multi-source generative modeling, in an application to tropical cyclones](https://mlsanity.com/paper/2606.24263/) (2026)
- [Decision-Aware Training for Sample-Based Generative Models](https://mlsanity.com/paper/2607.01171/) (2026)
- [Weather Emulators at the Frontier of Heat Extremes Predictability](https://mlsanity.com/paper/2607.28220/) (2026)

## Abstract (arXiv)

Weather forecasting plays a vital role in today's society, from agriculture and logistics to predicting the output of renewable energies, and preparing for extreme weather events. Deep learning weather forecasting models trained with the next state prediction objective on ERA5 have shown great success compared to numerical global circulation models. However, for a wide range of applications, being able to provide representative samples from the distribution of possible future weather states is critical. In this paper, we propose a methodology to leverage deterministic weather models in the design of probabilistic weather models, leading to improved performance and reduced computing costs. We first introduce ArchesWeather, a transformer-based deterministic model that improves upon Pangu-Weather by removing overrestrictive inductive priors. We then design a probabilistic weather model called ArchesWeatherGen based on flow matching, a modern variant of diffusion models, that is trained to project ArchesWeather's predictions to the distribution of ERA5 weather states. ArchesWeatherGen is a true stochastic emulator of ERA5 and surpasses IFS ENS and NeuralGCM on all WeatherBench headline variables (except for NeuralGCM's geopotential). Our work also aims to democratize the use of deterministic and generative machine learning models in weather forecasting research, with academic computing resources. All models are trained at 1.5deg resolution, with a training budget of ~9 V100 days for ArchesWeather and ~45 V100 days for ArchesWeatherGen. For inference, ArchesWeatherGen generates 15-day weather trajectories at a rate of 1 minute per ensemble member on a A100 GPU card. To make our work fully reproducible, our code and models are open source, including the complete pipeline for data preparation, training, and evaluation, at https://github.com/INRIA/geoarches .