Journal article · 2026 · Open access

AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score

Simon T. K. Lang, Mihai Alexe, Mariana Clare, Christopher David Roberts, Rilwan A. Adewoyin, Zied Ben Bouallègue, Matthew Chantry, Jesper Sören Dramsch, Peter Dominik Dueben, Sara Hahner, Pedro Maciel, Ana Prieto Nemesio and 6 more

npj Artificial Intelligence, 2(1), 18 · Published 2 February 2026 · DOI 10.1038/s44387-026-00073-7

Open PDFView at the publisherCite this paper

In brief

“AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score” is a 2026 journal article by Simon T. K. Lang and 17 co-authors, published in npj Artificial Intelligence. OpenAlex counts 225 citations as of 22 November 2026. It is open access under a CC BY licence: the full text is free to read at npj Artificial Intelligence.

Citations
225
Most cited in
2026 (187)
Field-weighted impact
255.18
References
59
Authors
18

Abstract

Ensemble weather forecasts provide a probabilistic description of the future state of the atmosphere and give users flow-dependent estimates of forecast uncertainty. Here, we introduce AIFS-CRPS, an ensemble variant of the machine-learned Artificial Intelligence Forecasting System (AIFS) developed at ECMWF. Its loss function is the almost fair Continuous Ranked Probability Score (afCRPS). It is based on a proper score, the CRPS, but approximately removes the bias in the score due to finite ensemble size yet avoids a degeneracy of the fair CRPS. The trained model is stochastic and can generate as many exchangeable members as desired. For medium-range forecasts AIFS-CRPS outperforms the physics-based Integrated Forecasting System (IFS) ensemble for the majority of variables and lead times. For subseasonal forecasts, AIFS-CRPS outperforms the IFS ensemble before calibration and is competitive with the IFS ensemble when forecasts are evaluated as anomalies to remove the influence of model biases.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Lang, S. T. K., Alexe, M., Clare, M., Roberts, C. D., Adewoyin, R. A., Bouallègue, Z. B., Chantry, M., Dramsch, J. S., Dueben, P. D., Hahner, S., Maciel, P., Nemesio, A. P., O’Brien, C., Pinault, F., Polster, J., Raoult, B., Tietsche, S., & Leutbecher, M. (2026). AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score. npj Artificial Intelligence, 2(1), 18. https://doi.org/10.1038/s44387-026-00073-7

MLA 9

Lang, Simon T. K., et al. “AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score.” npj Artificial Intelligence, vol. 2, no. 1, 2026, p. 18. https://doi.org/10.1038/s44387-026-00073-7.

Chicago

Lang, Simon T. K., Mihai Alexe, Mariana Clare, Christopher David Roberts, Rilwan A. Adewoyin, Zied Ben Bouallègue, Matthew Chantry, et al. 2026. “AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score.” npj Artificial Intelligence 2 (1): 18. https://doi.org/10.1038/s44387-026-00073-7.

Harvard

Lang, S. T. K. et al. (2026) ‘AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score’, npj Artificial Intelligence, 2(1), p. 18. doi:10.1038/s44387-026-00073-7.

IEEE

S. T. K. Lang et al., “AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score,” npj Artificial Intelligence, vol. 2, no. 1, p. 18, 2026, doi: 10.1038/s44387-026-00073-7.

BibTeX

@article{lang2026aifs,
  title     = {{AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score}},
  author    = {Lang, Simon T. K. and Alexe, Mihai and Clare, Mariana and Roberts, Christopher David and Adewoyin, Rilwan A. and Bouallègue, Zied Ben and Chantry, Matthew and Dramsch, Jesper Sören and Dueben, Peter Dominik and Hahner, Sara and Maciel, Pedro and Nemesio, Ana Prieto and O’Brien, Cathal and Pinault, Florian and Polster, Jan and Raoult, Baudouin and Tietsche, Steffen and Leutbecher, Martin},
  journal   = {npj Artificial Intelligence},
  year      = {2026},
  volume    = {2},
  number    = {1},
  pages     = {18},
  doi       = {10.1038/s44387-026-00073-7},
  url       = {https://doi.org/10.1038/s44387-026-00073-7}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Citations by year

YearCitationsShare of the best year
202411
20253737
2026187187

Its field-weighted citation impact is 255.18: it has been cited 255.18 times as often as the average work of the same field, type and year (1.00 is the world average).

Topic and keywords

Physical Sciences › Earth and Planetary Sciences › Atmospheric Science › Meteorological Phenomena and Simulations

  • ensemble forecasting
  • continuous ranked probability score
  • proper scoring rules
  • forecast uncertainty
  • medium-range forecast
  • forecast calibration
  • ECMWF Integrated Forecasting System

Work with this paper

Most-cited papers in Earth and Planetary Sciences

All highly cited Earth and Planetary Sciences papers

Questions about this paper

How do I cite this paper in APA style?

Lang, S. T. K., Alexe, M., Clare, M., Roberts, C. D., Adewoyin, R. A., Bouallègue, Z. B., Chantry, M., Dramsch, J. S., Dueben, P. D., Hahner, S., Maciel, P., Nemesio, A. P., O’Brien, C., Pinault, F., Polster, J., Raoult, B., Tietsche, S., & Leutbecher, M. (2026). AIFS-CRPS: ensemble forecasting using a model trained with a loss function based on the continuous ranked probability score. npj Artificial Intelligence, 2(1), 18. https://doi.org/10.1038/s44387-026-00073-7

How many times has this paper been cited?

225 times, according to OpenAlex on 22 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at npj Artificial Intelligence: sign in and open the PDF from this page.

Who wrote this paper?

Simon T. K. Lang, Mihai Alexe, Mariana Clare, Christopher David Roberts, Rilwan A. Adewoyin, Zied Ben Bouallègue, Matthew Chantry, Jesper Sören Dramsch, Peter Dominik Dueben, Sara Hahner, Pedro Maciel, Ana Prieto Nemesio and 6 more.

What is the DOI of this paper?

10.1038/s44387-026-00073-7, permanently at https://doi.org/10.1038/s44387-026-00073-7. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 22 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.