Tools
Tools
Mahesh, Ankur, Collins, William D. 
ORCID: https://orcid.org/0000-0002-4463-9848, Bonev, Boris, Brenowitz, Noah, Cohen, Yair, Elms, Joshua, Harrington, Peter, Kashinath, Karthik, Kurth, Thorsten, North, Joshua, O'Brien, Travis, Pritchard, Michael, Pruitt, David, Risser, Mark, Subramanian, Shashank and Willard, Jared
(2025)
Huge ensembles – Part 1: Design of ensemble weather forecasts using spherical Fourier neural operators.
Geoscientific Model Development, 18 (17).
pp. 5575-5603.
ISSN 1991-9603
Preview |
PDF (gmd-18-5575-2025)
- Published Version
Available under License Creative Commons Attribution. Download (9MB) | Preview |
Abstract
Simulating low-likelihood high-impact extreme weather events in a warming world is a significant and challenging task for current ensemble forecasting systems. While these systems presently use up to 100 members, larger ensembles could enrich the sampling of internal variability. They may capture the long tails associated with climate hazards better than traditional ensemble sizes. Due to computational constraints, it is infeasible to generate huge ensembles (comprised of 1000–10 000 members) with traditional, physics-based numerical models. In this two-part paper, we replace traditional numerical simulations with machine learning (ML) to generate hindcasts of huge ensembles. In Part 1, we construct an ensemble weather forecasting system based on spherical Fourier neural operators (SFNOs), and we discuss important design decisions for constructing such an ensemble. The ensemble represents model uncertainty through perturbed-parameter techniques, and it represents initial condition uncertainty through bred vectors, which sample the fastest-growing modes of the forecast. Using the European Centre for Medium-Range Weather Forecasts Integrated Forecasting System (IFS) as a baseline, we develop an evaluation pipeline composed of mean, spectral, and extreme diagnostics. With large-scale, distributed SFNOs with 1.1 billion learned parameters, we achieve calibrated probabilistic forecasts. As the trajectories of the individual members diverge, the ML ensemble mean spectra degrade with lead time, consistent with physical expectations. However, the individual ensemble members' spectra stay constant with lead time. Therefore, these members simulate realistic weather states during the rollout, and the ML ensemble passes a crucial spectral test in the literature. The IFS and ML ensembles have similar extreme forecast indices, and we show that the ML extreme weather forecasts are reliable and discriminating. These diagnostics ensure that the ensemble can reliably simulate the time evolution of the atmosphere, including low-likelihood high-impact extremes. In Part 2, we generate a huge ensemble initialized each day in summer 2023, and we characterize the simulations of extremes.
| Item Type: | Article |
|---|---|
| Additional Information: | Code and data availability: The code, datasets, and models used to produce the results used in this paper are archived on DataDryad under https://doi.org/10.5061/dryad.2rbnzs80n (Mahesh et al., 2025b). The code is integrated with Zenodo at the aforementioned DOI, and it is also available at https://github.com/ankurmahesh/earth2mip-fork (Mahesh et al., 2025c) as an additional download location. We include the code to train SFNO, conduct ensemble inference with bred vectors and multiple checkpoints, and scoring and analysis code. We also open-source the model weights of the trained SFNO. See the README of the DOI for information on how to use the code base and for the permission license associated with the code and data. The code is available via the Lawrence Berkeley Lab BSD variant license, and the data are available with a CC0 license. To run the ensemble for inference, a current version of the project is available from the project website at https://github.com/NVIDIA/earth2studio (NVIDIA, 2025) under the Apache-2.0 license. |
| Faculty \ School: | Faculty of Science > School of Environmental Sciences |
| Related URLs: | |
| Depositing User: | LivePure Connector |
| Date Deposited: | 21 Aug 2026 15:45 |
| Last Modified: | 23 Aug 2026 05:37 |
| URI: | https://ueaeprints.uea.ac.uk/id/eprint/104295 |
| DOI: | 10.5194/gmd-18-5575-2025 |
Downloads
Downloads per month over past year
Actions (login required)
 |
View Item |