CoChat

A machine learning model that outperforms conventional global subseasonal forecast models

Open access

Published version of a preprint

The DOI matches its Crossref record; 102 of the 102 references it cites were found in a registry or an index; 24 of the 74 sources its claims cite were checked for support.

Abstract

Skillful subseasonal forecasts are crucial for various sectors of society but pose a grand scientific challenge. Recently, machine learning-based weather forecasting models outperform the most successful numerical weather predictions generated by the European Centre for Medium-Range Weather Forecasts (ECMWF), but have not yet surpassed conventional models at subseasonal timescales. This paper introduces FuXi Subseasonal-to-Seasonal (FuXi-S2S), a machine learning model that provides global daily mean forecasts up to 42 days, encompassing five upper-air atmospheric variables at 13 pressure levels and 11 surface variables. FuXi-S2S, trained on 72 years of daily statistics from ECMWF ERA5 reanalysis data, outperforms the ECMWF's state-of-the-art Subseasonal-to-Seasonal model in ensemble mean and ensemble forecasts for total precipitation and outgoing longwave radiation, notably enhancing global precipitation forecast. The improved performance of FuXi-S2S can be primarily attributed to its superior capability to capture forecast uncertainty and accurately predict the Madden-Julian Oscillation (MJO), extending the skillful MJO prediction from 30 days to 36 days. Moreover, FuXi-S2S not only captures realistic teleconnections associated with the MJO but also emerges as a valuable tool for discovering precursor signals, offering researchers insights and potentially establishing a new paradigm in Earth system science research.

Overview

CoChat Summary

TL;DR

FuXi-S2S is a 51-member machine learning subseasonal forecasting system trained on ERA5 that extends global daily forecasts to 42 days and is evaluated against ECMWF S2S reforecasts. The paper reports better ensemble-mean and probabilistic skill for total precipitation and outgoing longwave radiation, plus improved Madden–Julian Oscillation prediction and an earlier forecast of the 2022 Pakistan floods.

Key findings

  • FuXi-S2S outperforms ECMWF S2S for ensemble-mean forecasts of total precipitation (TP) and outgoing longwave radiation (OLR), with significantly higher temporal anomaly correlation coefficients (TCC) for TP and OLR across lead times of 3, 4, 5, 6, 3–4, and 5–6 weeks.
  • For MJO prediction, FuXi-S2S keeps bivariate correlation above the skill threshold of 0.5 for up to 36 days, extending skillful prediction from 30 days for ECMWF S2S.
  • For the 2022 Pakistan floods, FuXi-S2S predicted rainfall of at least 4 standard deviations above climatology for forecasts initialized on July 21, 2022, about 4 weeks in advance; the observed rainfall anomaly was about 6 standard deviations above climatology.
  • The paper attributes much of the improved skill to flow-dependent perturbations for ensemble generation and reports that increasing the ensemble size from 51 to 101 members further improves performance.

Methods

  • The model is trained on 67 years of ERA5 daily statistics (1950–2016) and evaluated on 2017–2021 testing data, with 2022 used for the Pakistan flood case study and comparison with ECMWF S2S real-time forecasts.
  • FuXi-S2S is an autoregressive encoder-perturbation-decoder model that generates 51-member ensemble forecasts of 76 variables (5 upper-air variables at 13 pressure levels and 11 surface variables) at 1.5° spatial resolution and daily temporal resolution, out to 42 days.
  • Skill is assessed with detrended TCC for deterministic forecasts, RPSS for probabilistic forecasts, BSS for extreme precipitation, RMM bivariate COR for MJO, and saliency maps to interpret the Pakistan flood prediction.

Limitations

  • The paper states that FuXi-S2S currently has coarser spatial resolution than ECMWF S2S (1.5° vs 36 km) and predicts daily mean values only up to 50 hPa, lacking some useful variables such as daily maximum and minimum temperatures.
  • The authors note that their Pakistan flood interpretability analysis focuses on gradients with respect to SST, and that other precursor signals such as U, V, and Z anomalies are not examined in detail; they say this is for future research.
  • Evident from the text: evaluation is limited to 2017–2021 testing data for most metrics and a single highlighted extreme event in 2022, so broader out-of-sample generalization beyond these periods is not directly shown.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41467-024-50714-1

Cites

102 of 102 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Comparison of globally-averaged and latitude-weighted temporal anomaly correlation coefficient (TCC) of the ensemble mean between ECMWF subseasonal-to-seasonal (S2S) reforecasts (in blue) and FuXi-S2S forecasts (in red) for total precipitation (TP), 2 m temperature (T2M), geopotential at 500 hPa (Z500), and outgoing longwave radiation (OLR). Rows 1 and 2 represent the performance across these variables, utilizing all testing data from the period spanning from 2017 to 2021. A bootstrapping approach, repeated 1000 times, is used for significance testing. When the FuXi-S2S forecasts fail to show a statistically significant improvement over the ECMWF S2S reforecasts at the 97.5% confidence level, a pale color scheme is used to denote these results.
Fig. 1 Comparison of globally-averaged and latitude-weighted temporal anomaly correlation coefficient (TCC) of the ensemble mean between ECMWF subseasonal-to-seasonal (S2S) reforecasts (in blue) and FuXi-S2S forecasts (in red) for total precipitation (TP), 2 m temperature (T2M), geopotential at 500 hPa (Z500), and outgoing longwave radiation (OLR). Rows 1 and 2 represent the performance across these variables, utilizing all testing data from the period spanning from 2017 to 2021. A bootstrapping approach, repeated 1000 times, is used for significance testing. When the FuXi-S2S forecasts fail to show a statistically significant improvement over the ECMWF S2S reforecasts at the 97.5% confidence level, a pale color scheme is used to denote these results.CC BY-NC-ND · PubMed Central
Fig. 2 Maps displaying the average Ranked Probability skill Score (RPSS) (first and second rows) and Brier Skill Score (BSS) (third and fourth rows) without latitude weighting, comparing ECMWF subseasonal-to-seasonal (S2S) (first column) and FuXi-S2S (second column) forecasts. Additionally, the third column depicts the difference in RPSS and BSS between FuXi-S2S and ECMWF S2S for total precipitation (TP) at forecast lead times of weeks 3–4 (first and third rows) and weeks 5–6 (second and fourth rows), utilizing all testing data from 2017 to 2021. Red contour lines in the first and second columns indicate areas with positive values of RPSS and BSS. Stippling on the map denotes areas where the skill score is statistically significant at the 97.5% confidence level. Specifically, in columns 1 and 2, stippling indicates regions where the skill scores of the ECMWF S2S and FuXi-S2S models significantly surpass those of climatology. In column 3, stippling highlights areas where the FuXi-S2S model significantly outperforms the ECMWF S2S.
Fig. 2 Maps displaying the average Ranked Probability skill Score (RPSS) (first and second rows) and Brier Skill Score (BSS) (third and fourth rows) without latitude weighting, comparing ECMWF subseasonal-to-seasonal (S2S) (first column) and FuXi-S2S (second column) forecasts. Additionally, the third column depicts the difference in RPSS and BSS between FuXi-S2S and ECMWF S2S for total precipitation (TP) at forecast lead times of weeks 3–4 (first and third rows) and weeks 5–6 (second and fourth rows), utilizing all testing data from 2017 to 2021. Red contour lines in the first and second columns indicate areas with positive values of RPSS and BSS. Stippling on the map denotes areas where the skill score is statistically significant at the 97.5% confidence level. Specifically, in columns 1 and 2, stippling indicates regions where the skill scores of the ECMWF S2S and FuXi-S2S models significantly surpass those of climatology. In column 3, stippling highlights areas where the FuXi-S2S model significantly outperforms the ECMWF S2S.CC BY-NC-ND · PubMed Central
Fig. 3 Comparison of real-time multivariate Madden–Julian Oscillation (MJO) (RMM) bivariate Correlation (COR) of the ensemble means between ECMWF subseasonal-to-seasonal (S2S) reforecasts (in blue) and FuXi-S2S forecasts (in red) using all testing data from 2017 to 2021. a Comparison of RMM bivariate COR as a function of forecast lead times. Dashed black line signifies the prediction skill threshold of COR = 0.5. b The RMM bivariate COR is depicted as a function of the month of initialization (x-axis) and forecast lead time (y-axis), with red and blue lines indicating the skillful MJO prediction days of ECMWF S2S (in blue) and FuXi-S2S (in red), respectively.
Fig. 3 Comparison of real-time multivariate Madden–Julian Oscillation (MJO) (RMM) bivariate Correlation (COR) of the ensemble means between ECMWF subseasonal-to-seasonal (S2S) reforecasts (in blue) and FuXi-S2S forecasts (in red) using all testing data from 2017 to 2021. a Comparison of RMM bivariate COR as a function of forecast lead times. Dashed black line signifies the prediction skill threshold of COR = 0.5. b The RMM bivariate COR is depicted as a function of the month of initialization (x-axis) and forecast lead time (y-axis), with red and blue lines indicating the skillful MJO prediction days of ECMWF S2S (in blue) and FuXi-S2S (in red), respectively.CC BY-NC-ND · PubMed Central
Fig. 4 Comparative analysis for the 2022 Pakistan floods predictions between the ECMWF subseasonal-to-seasonal (S2S) and FuXi-S2S models as well as the precursor signals that contributed to accurate predictions by the FuXi-S2S model. Comparison of spatially and temporally averaged standardized total precipitation (TP) anomaly (a) over the 2 weeks from August 16th to August 31st, 2022, showcasing GPCP observations (in black) alongside predictions from ECMWF S2S real-time forecasts (in blue) and FuXi-S2S forecasts (in red), with initialization dates: August 11th (08-11, MM-DD), August 8th (08-08), August 4th (08-04), August 1st (08-01), July 28th (07-28), July 25th (07-25), and July 21st (07-21). The black lines on the bar of ECMWF S2S and FuXi-S2S forecasts represent the 25th and 75th percentiles. For the comparison of temporally averaged standardized TP anomaly maps (b), the first column represents GPCP observations, while the second and third columns display predictions from ECMWF S2S and FuXi-S2S, respectively, both initialized on July 28th, and the fourth and fifth columns correspond to predictions from ECMWF S2S and FuXi-S2S, respectively, with an initialization date of July 21st. Green contour indicates the border line of Pakistan. The saliency maps (c) were generated using the gradient of the negative standardized TP anomaly, averaged over the Pakistan region, in relation to the input SST. These maps correspond to forecasts initialized on July 28th (07-28, first column) and July 21st (07-21, second column). Here, the red and blue colors indicate the positive and negative correlations between the negative of standardized TP and variations in SST. The black lines on the bars in this figure represent the 25th and 75th percentiles of the ensemble forecasts for each start date for both ECMWF and FuXi-S2S models.
Fig. 4 Comparative analysis for the 2022 Pakistan floods predictions between the ECMWF subseasonal-to-seasonal (S2S) and FuXi-S2S models as well as the precursor signals that contributed to accurate predictions by the FuXi-S2S model. Comparison of spatially and temporally averaged standardized total precipitation (TP) anomaly (a) over the 2 weeks from August 16th to August 31st, 2022, showcasing GPCP observations (in black) alongside predictions from ECMWF S2S real-time forecasts (in blue) and FuXi-S2S forecasts (in red), with initialization dates: August 11th (08-11, MM-DD), August 8th (08-08), August 4th (08-04), August 1st (08-01), July 28th (07-28), July 25th (07-25), and July 21st (07-21). The black lines on the bar of ECMWF S2S and FuXi-S2S forecasts represent the 25th and 75th percentiles. For the comparison of temporally averaged standardized TP anomaly maps (b), the first column represents GPCP observations, while the second and third columns display predictions from ECMWF S2S and FuXi-S2S, respectively, both initialized on July 28th, and the fourth and fifth columns correspond to predictions from ECMWF S2S and FuXi-S2S, respectively, with an initialization date of July 21st. Green contour indicates the border line of Pakistan. The saliency maps (c) were generated using the gradient of the negative standardized TP anomaly, averaged over the Pakistan region, in relation to the input SST. These maps correspond to forecasts initialized on July 28th (07-28, first column) and July 21st (07-21, second column). Here, the red and blue colors indicate the positive and negative correlations between the negative of standardized TP and variations in SST. The black lines on the bars in this figure represent the 25th and 75th percentiles of the ensemble forecasts for each start date for both ECMWF and FuXi-S2S models.CC BY-NC-ND · PubMed Central
Fig. 5 Schematic diagram of the structures of the FuXi Subseasonal-to-Seasonal (FuXi-S2S) model. a Inference stage of the FuXi-S2S model. ht represents the hidden feature generated by the Encoder from the input data. The perturbation vector zt is generated by the perturbation module, resulting in the perturbed hidden feature (h~)^t. b Training stage of the FuXi-S2S model. N(Θp^t) and N(Θq^t) are the low-rank multivariate Gaussian distributions generated by encoders P and Q, respectively. The Kullback–Leibler (KL) divergence loss measures the discrepancy between the distributions predicted by both encoders, N(Θp^t) and N(Θq^t).
Fig. 5 Schematic diagram of the structures of the FuXi Subseasonal-to-Seasonal (FuXi-S2S) model. a Inference stage of the FuXi-S2S model. ht represents the hidden feature generated by the Encoder from the input data. The perturbation vector zt is generated by the perturbation module, resulting in the perturbed hidden feature (h~)^t. b Training stage of the FuXi-S2S model. N(Θp^t) and N(Θq^t) are the low-rank multivariate Gaussian distributions generated by encoders P and Q, respectively. The Kullback–Leibler (KL) divergence loss measures the discrepancy between the distributions predicted by both encoders, N(Θp^t) and N(Θq^t).CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

10 Valid · 8 Not valid · 21 Not checked · 92 Own results and interpretations

51 more claims

Create a free account to see all 131 claims and to verify the 28 references still open.

See all claims — free account
  • It may state two separate claims as one. 47
  • Probably not a substantial claim. 33
  • Borderline: one of the tests for a substantial claim is unclear. 1
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

1 of 1

Citations

Cites

1 retracted

Show all 102 references

Research works printed but not linked: 8

  1. National Academies of Sciences. Next generation earth system prediction: strategies for subseasonal to seasonal forecasts (National Academies Press, Washington, DC, 2016).
  2. Nowak, K., Webb, R., Cifelli, R. & Brekke, L. Sub-seasonal climate forecast rodeo. In Proc. 2017 AGU Fall Meeting, 11–15 (New Orleans, LA, 2017).
  3. Domeisen, D. I. et al. Advances in the subseasonal prediction of extreme events: relevant case studies across the globe. Bull. Am. Meteorol Soc. 103, 1473–1501 (2022). · DOI · Paper page
  4. Simonyan, K., Vedaldi, A. & Zisserman, A. Deep inside convolutional networks: Visualising image classification models and saliency maps. In Proc. International Conference on Learning Representations (2013).
  5. Chattopadhyay, A. & Hassanzadeh, P. Long-term instability of deep learning-based digital twins of the climate system: Cause and solution. In Proc. APS March Meeting Abstracts (2023).
  6. Walker, G. T. Correlations in seasonal variations of weather. viii, a further study of world weather. Men. Indian Meteor. Dept. 24, 275–332 (1924).
  7. Huang, R. H. Influence of the heat source anomaly over the western tropical Pacific for the subtropical high over East Asia. In Proc. International Conference on the General Circulation of East Asia. Chendu, China, April 10–15, 1987, 40–50 (1987).
  8. Kingma, D. P. & Ba, J. Adam: A method for stochastic optimization. In Proc. International Conference on Learning Representations (2015).

Cited by

See all 105

Metadata

Authors
Chen, L., Zhong, X., Li, H., Wu, J., Lu, B., Chen, D., Xie, S.-P., Wu, L., Chao, Q., Lin, C., Hu, Z., & Qi, Y.
Affiliations
  • Fudan University (CN)
  • China Meteorological Administration (CN)
  • University of Gothenburg (SE)
  • Scripps Institution of Oceanography (US)
Corresponding author
  • Li, Hao
  • Qi, Yuan
Venue
Nature Communications
Published
2024-07-30
Volume
15
Issue
1
Pages
6425-6425
ISSN
2041-1723 (ISSN-L)
OpenAlex
W4401112323
Cited by
105
References
89
Citations by year
2026: 41 · 2025: 56 · 2024: 8
FWCI
19.17
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by-nc-nd)
  • Research Square (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by-nc-nd)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Natural Science Foundation of China (42175052)
Topics
  • Physical Sciences › Earth and Planetary Sciences › Atmospheric Science › Meteorological Phenomena and Simulations
  • Physical Sciences › Environmental Science › Global and Planetary Change › Climate variability and models
  • Physical Sciences › Environmental Science › Environmental Engineering › Hydrological Forecasting Using AI
Keywords
subseasonal-to-seasonal forecasting, Madden-Julian Oscillation, MJO prediction, precipitation forecasting, outgoing longwave radiation, forecast uncertainty, teleconnections, precursor signals, ERA5 reanalysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A machine learning model that outperforms conventional global subseasonal forecast models. https://app.cochat.ai/papers/10.1038/s41467-024-50714-1
The paper
Chen, L., Zhong, X., Li, H., Wu, J., Lu, B., Chen, D., Xie, S.-P., Wu, L., Chao, Q., Lin, C., Hu, Z., & Qi, Y. (2024). A machine learning model that outperforms conventional global subseasonal forecast models. Nature Communications. https://doi.org/10.1038/s41467-024-50714-1

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.