pith. sign in

arxiv: 2508.16168 · v2 · submitted 2025-08-22 · ⚛️ physics.ao-ph

FuXi-TC: A generative framework integrating deep learning and physics-based models for improved tropical cyclone forecasts

Shan Guo , Lei Chen , Yangyang Zhao , Yuetan Lin , Zeyi Niu , Xinyan Zhang , Ziyao Sun , Xiaohui Zhong
show 1 more author
Hao Li
This is my paper

Pith reviewed 2026-05-18 21:56 UTC · model grok-4.3

classification ⚛️ physics.ao-ph
keywords tropical cyclone forecastingdiffusion modelsgenerative modelshybrid forecastingprecipitation predictionintensity forecastdownscalingensemble forecasts
0
0 comments X

The pith

FuXi-TC conditions a diffusion model on FuXi large-scale forecasts to match ECMWF tropical cyclone intensity skill while improving precipitation predictions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces FuXi-TC to address the difficulty of predicting tropical cyclone intensity, where numerical weather prediction models are limited by computation and predictability while pure deep learning models underestimate intensity from reanalysis biases. FuXi-TC combines the track strengths of the FuXi model with a diffusion-based generative approach to downscale fine-grained fields. Evaluations for the 2024 Western North Pacific show the method matches ECMWF deterministic intensity forecasts but with better precipitation skill, higher speed, and lower cost. It also generalizes without fine-tuning to North Atlantic hurricanes and supports probabilistic outputs from ensemble inputs.

Core claim

FuXi-TC is a diffusion-based generative forecasting framework that conditions a diffusion model on the large-scale forecasts of the global FuXi model to downscale and deliver higher-accuracy forecasts of fine-grained variable fields such as wind speed and precipitation. In evaluations across the 2024 Western North Pacific, this approach matches the TC intensity forecast skill of the operational ECMWF deterministic model while delivering superior precipitation forecasts, at significantly higher inference speeds and lower computational costs. FuXi-TC also demonstrates robust zero-shot generalization to North Atlantic hurricanes and yields well-dispersed probabilistic forecasts when applied to

What carries the argument

Diffusion model conditioned on FuXi large-scale forecasts for downscaling to accurate fine-grained tropical cyclone fields such as wind speed and precipitation.

If this is right

  • The method produces intensity forecasts comparable to the leading operational deterministic model.
  • Precipitation forecasts exceed the accuracy of the ECMWF baseline in the tested region.
  • Inference runs at much higher speed and lower computational cost than full physics-based simulations.
  • Zero-shot transfer works across basins such as from Western North Pacific to North Atlantic without retraining.
  • Conditioning the same framework on ensemble inputs generates dispersed probabilistic forecasts with refined intensity values.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar conditioning of generative models on large-scale outputs could mitigate intensity underestimation in other deep learning weather systems.
  • Forecast centers might reduce reliance on expensive high-resolution NWP runs by using this downscaling step for targeted variables.
  • The approach invites tests on additional extreme weather types where fine-scale physics are critical.
  • Extending the conditioning to other global deep learning models could test broader applicability beyond the FuXi base.

Load-bearing premise

Conditioning the diffusion model on FuXi large-scale forecasts is sufficient to accurately represent fine-grained fields like wind speed and precipitation without inheriting or amplifying biases from reanalysis-based training data.

What would settle it

Direct side-by-side error statistics for a new season of Western North Pacific or North Atlantic cyclones showing FuXi-TC intensity errors exceeding those of ECMWF or precipitation errors no better than baselines would disprove the claimed matching skill and superiority.

read the original abstract

Tropical cyclones (TCs) are among the most devastating natural hazards, yet their intensity remains notoriously difficult to predict. NWP models are constrained by both computational demands and intrinsic predictability, while state-of-the-art deep learning-based weather forecasting models tend to underestimate TC intensity due to biases in reanalysis-based training data. Here, we present FuXi-TC, a diffusion-based generative forecasting framework that combines the track prediction strength of the FuXi model with the intensity representation of NWP simulations. By conditioning a diffusion model on the large-scale forecasts of the global FuXi model, FuXi-TC effectively downscales and delivers higher-accuracy forecasts of fine-grained variable fields such as wind speed and precipitation. In evaluations across the 2024 Western North Pacific, our approach matches the TC intensity forecast skill of the operational ECMWF deterministic model while delivering superior precipitation forecasts. Meanwhile this is achieved with significantly higher inference speeds and lower computational costs. Moreover, FuXi-TC demonstrates robust zero-shot generalization directly when applied to North Atlantic hurricanes without any fine-tuning. When applied to the FuXi ensemble model, this framework effectively yields well-dispersed probabilistic forecasts and refines the ensemble intensity predictions.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The manuscript introduces FuXi-TC, a diffusion-based generative framework that conditions a diffusion model on large-scale forecasts from the pre-trained FuXi model to downscale fine-scale tropical cyclone fields such as wind speed and precipitation, while incorporating NWP outputs for intensity representation. On 2024 Western North Pacific cases, the approach is reported to match the intensity forecast skill of the operational ECMWF deterministic model, deliver superior precipitation forecasts, operate at higher inference speeds with lower costs, exhibit zero-shot generalization to North Atlantic hurricanes, and produce improved probabilistic forecasts when applied to the FuXi ensemble.

Significance. If the reported skill matches and generalization hold under rigorous validation, this hybrid DL-physics approach would represent a meaningful advance in tropical cyclone forecasting by addressing intensity underestimation in pure data-driven models while retaining computational efficiency. The zero-shot cross-basin performance and ensemble refinement are notable strengths that could influence operational hybrid modeling strategies.

major comments (1)
  1. The central claim that FuXi-TC matches ECMWF TC intensity skill while improving precipitation forecasts rests on the assumption that conditioning the diffusion model on FuXi large-scale outputs suffices to recover accurate fine-scale fields without inheriting or amplifying reanalysis biases; this requires explicit bias diagnostics and sensitivity tests in the results to be load-bearing for the hybrid framework's validity.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive comment, which helps strengthen the validation of our hybrid approach. We address the point below and have incorporated revisions to provide the requested diagnostics.

read point-by-point responses

  1. Referee: The central claim that FuXi-TC matches ECMWF TC intensity skill while improving precipitation forecasts rests on the assumption that conditioning the diffusion model on FuXi large-scale outputs suffices to recover accurate fine-scale fields without inheriting or amplifying reanalysis biases; this requires explicit bias diagnostics and sensitivity tests in the results to be load-bearing for the hybrid framework's validity.

    Authors: We agree that direct bias diagnostics and sensitivity tests would make the hybrid framework's validity more robust. While the original manuscript demonstrates that FuXi-TC matches ECMWF intensity skill (a physics-based reference free of the same reanalysis biases) and outperforms on precipitation, this provides indirect evidence against bias amplification. In the revised version we add: (i) spatial bias maps of 10-m wind and precipitation for FuXi-TC versus FuXi and ECMWF relative to reanalysis during 2024 WNP TCs; (ii) sensitivity experiments that modulate the strength of FuXi conditioning and track resulting changes in intensity and precipitation metrics. These analyses show that the diffusion step systematically reduces FuXi's intensity underestimation without introducing or amplifying reanalysis biases, thereby supporting the central claim. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The described framework conditions a diffusion model on large-scale forecasts from the existing FuXi model to downscale fine-grained TC fields such as wind speed and precipitation. This is a standard hybrid generative approach that integrates pre-trained components without any derivation step that reduces by construction to fitted inputs or self-referential definitions. Evaluation claims (matching ECMWF intensity skill on 2024 WNP cases, superior precipitation, zero-shot NA generalization) are presented as empirical outcomes on held-out data rather than tautological predictions. No load-bearing self-citation chains, uniqueness theorems imported from prior author work, or ansatzes smuggled via citation appear in the abstract or stated central claim. The derivation remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities; the framework implicitly depends on the accuracy of the upstream FuXi model and NWP simulations.

pith-pipeline@v0.9.0 · 5769 in / 1036 out tokens · 35550 ms · 2026-05-18T21:56:47.629910+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

54 extracted references · 54 canonical work pages · 1 internal anchor

  1. [1]

    Nature 436(7051), 686–688 (2005) https://doi.org/10.1038/nature03906

    Emanuel, K.: Increasing destructiveness of tropical cyclones over the past 30 years. Nature 436(7051), 686–688 (2005) https://doi.org/10.1038/nature03906

    work page doi:10.1038/nature03906 2005

  2. [2]

    Nature Climate Change 2 (2012) https://doi.org/10.1038/nclimate1357

    Emanuel, K., Chonabayashi, S., Bakkensen, L., Mendelsohn, R.: The impact of climate change on global tropical cyclone damage. Nature Climate Change 2 (2012) https://doi.org/10.1038/nclimate1357

    work page doi:10.1038/nclimate1357 2012

  3. [3]

    Bulletin of the American Meteorological Society 90(4), 489–496 (2009) https://doi.org/ 10.1175/2008BAMS2631.1

    Zhang, Q., Wu, L., Liu, Q.: Tropical cyclone damages in china 1983–2006. Bulletin of the American Meteorological Society 90(4), 489–496 (2009) https://doi.org/ 10.1175/2008BAMS2631.1

    work page doi:10.1175/2008bams2631.1 1983

  4. [4]

    Global and Planetary Change 150, 81–93 (2017) https://doi.org/10.1016/j.gloplacha.2017

    Zhang, Q., Gu, X., Shi, P., Singh, V.P.: Impact of tropical cyclones on flood risk in southeastern china: Spatial patterns, causes and implications. Global and Planetary Change 150, 81–93 (2017) https://doi.org/10.1016/j.gloplacha.2017. 02.004

    work page doi:10.1016/j.gloplacha.2017 2017

  5. [5]

    Nature Climate Change 2(4), 289–294 (2012) https://doi.org/10.1038/nclimate1410

    Peduzzi, P., Chatenoux, B., Dao, H., De Bono, A., Herold, C., Kossin, J., Mouton, F., Nordbeck, O.: Global trends in tropical cyclone risk. Nature Climate Change 2(4), 289–294 (2012) https://doi.org/10.1038/nclimate1410

    work page doi:10.1038/nclimate1410 2012

  6. [6]

    Meteorology and Atmospheric Physics 137(2), 21 (2025)

    He, J., Lau, T., Chan, Y., Cheung, P., Lam, C., Choy, C., Chan, P.: An observa- tional analysis of super typhoon yagi (2024) over the south china sea. Meteorology and Atmospheric Physics 137(2), 21 (2025)

    work page 2024

  7. [7]

    Remote Sensing 17(9), 1598 (2025)

    Nguyen, V.D., Rouzegari, N., Dao, V., Almutlaq, F., Nguyen, P., Sorooshian, S.: Comparative analysis of satellite-based precipitation products during extreme 15 rainfall from super typhoon yagi in hanoi, vietnam (september 2024). Remote Sensing 17(9), 1598 (2025)

    work page 2024

  8. [8]

    (eds.): Climate Change 2021: The Physical Science Basis

    Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., P´ ean, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M.I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J.B.R., Maycock, T.K., Waterfield, T., Yelek¸ ci, O., Yu, R., Zhou, B. (eds.): Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessme...

    work page 2021

  9. [9]

    Science 309, 1844– 1846 (2005) https://doi.org/10.1126/science.1116448

    Webster, P., Holland, G., Curry, J., Chang, H.-R.: Changes in tropical cyclone number, duration, and intensity in a warming environment. Science 309, 1844– 1846 (2005) https://doi.org/10.1126/science.1116448

    work page doi:10.1126/science.1116448 2005

  10. [10]

    (eds.) Overview of the NOAA/NCEP Operational Hurricane Weather Research and Forecast (HWRF) Modelling System, pp

    Tallapragada, V.: In: Mohanty, U.C., Gopalakrishnan, S.G. (eds.) Overview of the NOAA/NCEP Operational Hurricane Weather Research and Forecast (HWRF) Modelling System, pp. 51–106. Springer, ??? (2016). https://doi.org/10.5822/ 978-94-024-0896-6 3 . https : //doi.org/10.5822/978 − 94 − 024 − 0896 − 63

    work page doi:10.5822/978 2016

  11. [11]

    Bulletin of the American Meteorological Society 105(6), 932–961 (2024) https://doi.org/10.1175/BAMS-D-23-0139.1

    Alaka, G.J., Sippel, J.A., Zhang, Z., Kim, H.-S., Marks, F.D., Tallapragada, V., Mehra, A., Zhang, X., Poyer, A., Gopalakrishnan, S.G.: Lifetime performance of the operational hurricane weather research and forecasting model (hwrf) for north atlantic tropical cyclones. Bulletin of the American Meteorological Society 105(6), 932–961 (2024) https://doi.org/...

    work page doi:10.1175/bams-d-23-0139.1 2024

  12. [12]

    Weather and Forecasting 38(10), 2057–2075 (2023) https://doi.org/10.1175/WAF-D-23-0041.1

    Wang, W., Han, J., Yang, F., Steffen, J., Liu, B., Zhang, Z., Mehra, A., Tallapra- gada, V.: Improving the intensity forecast of tropical cyclones in the hurricane analysis and forecast system. Weather and Forecasting 38(10), 2057–2075 (2023) https://doi.org/10.1175/WAF-D-23-0041.1

    work page doi:10.1175/waf-d-23-0041.1 2057

  13. [13]

    Acta Meteorologica Sinica 79(1), 94–103 (2021) https://doi.org/10.11676/qxxb2020.067

    Ma, S., Zhang, J., Qu, A., Wang, D., Shen, X.: Impacts to tropical cyclone prediction of grapes tym from increasing of model vertical levels and enlarge- ment of model forecast domain. Acta Meteorologica Sinica 79(1), 94–103 (2021) https://doi.org/10.11676/qxxb2020.067

    work page doi:10.11676/qxxb2020.067 2021

  14. [14]

    Weather and Forecasting 35(5), 1913–1922 (2020) https://doi.org/10.1175/WAF-D-20-0059.1

    Cangialosi, J.P., Blake, E., DeMaria, M., Penny, A., Latto, A., Rappaport, E., Tallapragada, V.: Recent progress in tropical cyclone intensity forecasting at the national hurricane center. Weather and Forecasting 35(5), 1913–1922 (2020) https://doi.org/10.1175/WAF-D-20-0059.1

    work page doi:10.1175/waf-d-20-0059.1 1913

  15. [15]

    npj Climate and Atmospheric Science 8(1), 38 (2025)

    Xu, H., Zhao, Y., Zhao, D., Duan, Y., Xu, X.: Exploring the typhoon intensity forecasting through integrating ai weather forecasting with regional numerical weather model. npj Climate and Atmospheric Science 8(1), 38 (2025)

    work page 2025

  16. [16]

    Tropical Cyclone Research and Review (2025) 16

    Xinyuan, B., Jinping, L., Yihong, D.: Review of artificial intelligence application in typhoon forecasting. Tropical Cyclone Research and Review (2025) 16

    work page 2025

  17. [17]

    Nature 619(7970), 533–538 (2023)

    Bi, K., Xie, L., Zhang, H., Chen, X., Gu, X., Tian, Q.: Accurate medium-range global weather forecasting with 3d neural networks. Nature 619(7970), 533–538 (2023)

    work page 2023

  18. [18]

    Science China Earth Sciences, 1–13 (2024)

    Zhong, X., Chen, L., Liu, J., Lin, C., Qi, Y., Li, H.: Fuxi-extreme: Improving extreme rainfall and wind forecasts with diffusion model. Science China Earth Sciences, 1–13 (2024)

    work page 2024

  19. [19]

    : Evaluating data- driven forecasts of extreme weather

    Mahesh, A., Cohen, Y., Brenowitz, N., Elms, J., Subramanian, S., Harrington, P., Anandkumar, A., Pathak, J., Kurth, T., Bonev, B., et al. : Evaluating data- driven forecasts of extreme weather. In: AGU Fall Meeting Abstracts, vol. 2023, pp. 31–2513 (2023)

    work page 2023

  20. [20]

    In: EGU General Assembly Conference Abstracts, p

    Wang, J., Tabas, S., Yang, F., Levit, J., Stajner, I., Montuoro, R., Tallapragada, V., Gross, B.: Machine learning weather prediction model development for global ensemble forecasts at ncep. In: EGU General Assembly Conference Abstracts, p. 11707 (2024)

    work page 2024

  21. [21]

    Quarterly journal of the royal meteorological society 146(730), 1999–2049 (2020)

    Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Hor´ anyi, A., Mu˜ noz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., et al.: The era5 global reanal- ysis. Quarterly journal of the royal meteorological society 146(730), 1999–2049 (2020)

    work page 1999

  22. [22]

    org/10.1175/JCLI-D-16-0557.1

    Hodges, K., Cobb, A., Vidale, P.L.: How well are tropical cyclones represented in reanalysis datasets? Journal of Climate 30(14), 5243–5264 (2017) https://doi. org/10.1175/JCLI-D-16-0557.1

    work page doi:10.1175/jcli-d-16-0557.1 2017

  23. [23]

    Journal of Marine Science and Engineering 12(11), 2099 (2024)

    Zhang, X., Zuo, C., Wang, Z., Tao, C., Han, Y., Zuo, J.: Typhoon storm surge simulation study based on reconstructed era5 wind fields—a case study of typhoon “muifa”, the 12th typhoon of 2022. Journal of Marine Science and Engineering 12(11), 2099 (2024)

    work page 2022

  24. [24]

    Bulletin of the American Meteorological Society 91(3), 363–376 (2010)

    Knapp, K.R., Kruk, M.C., Levinson, D.H., Diamond, H.J., Neumann, C.J.: The international best track archive for climate stewardship (ibtracs) unifying tropical cyclone data. Bulletin of the American Meteorological Society 91(3), 363–376 (2010)

    work page 2010

  25. [25]

    arXiv preprint arXiv:2501.18122 (2025)

    Wang, X., Liu, L., Chen, K., Han, T., Li, B., Bai, L.: Vqlti: Long-term tropical cyclone intensity forecasting with physical constraints. arXiv preprint arXiv:2501.18122 (2025)

    work page arXiv 2025

  26. [26]

    Nature Communications 16(1), 5923 (2025)

    Huang, C., Mu, P., Zhang, J., Chan, S., Zhang, S., Yan, H., Chen, S., Bai, C.: Benchmark dataset and deep learning method for global tropical cyclone forecasting. Nature Communications 16(1), 5923 (2025)

    work page 2025

  27. [27]

    Journal of Geophysical Research: Machine Learning and Computation 1(3), 2024–000207 (2024)

    Liu, H.-Y., Tan, Z.-M., Wang, Y., Tang, J., Satoh, M., Lei, L., Gu, J.-F., Zhang, 17 Y., Nie, G.-Z., Chen, Q.-Z.: A hybrid machine learning/physics-based model- ing framework for 2-week extended prediction of tropical cyclones. Journal of Geophysical Research: Machine Learning and Computation 1(3), 2024–000207 (2024)

    work page 2024

  28. [28]

    Earth and Space Science 12(2), 2024–003952 (2025)

    Niu, Z., Huang, W., Zhang, L., Deng, L., Wang, H., Yang, Y., Wang, D., Li, H.: Improving typhoon predictions by integrating data-driven machine learning model with physics model based on the spectral nudging and data assimilation. Earth and Space Science 12(2), 2024–003952 (2025)

    work page 2024

  29. [29]

    arXiv preprint arXiv:2407.06100 (2024)

    Husain, S.Z., Separovic, L., Caron, J.-F., Aider, R., Buehner, M., Chamberland, S., Lapalme, E., McTaggart-Cowan, R., Subich, C., Vaillancourt, P.A., et al.: Leveraging data-driven weather models for improving numerical weather predic- tion skill through large-scale spectral nudging. arXiv preprint arXiv:2407.06100 (2024)

    work page arXiv 2024

  30. [30]

    In: International Confer- ence on Machine Learning, pp

    Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., Ganguli, S.: Deep unsuper- vised learning using nonequilibrium thermodynamics. In: International Confer- ence on Machine Learning, pp. 2256–2265 (2015). PMLR

    work page 2015

  31. [31]

    Advances in Neural Information Processing Systems 33, 6840–6851 (2020)

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems 33, 6840–6851 (2020)

    work page 2020

  32. [32]

    Springer (2025)

    Wu, L., Yu, R., Xiang, C., Yu, H., Feng, Y., Zhou, X.: Extreme Impacts of Four Landfalling Tropical Cyclones in China during the 2024 Peak Season. Springer (2025)

    work page 2024

  33. [33]

    In: Forty-first International Conference on Machine Learning (2024)

    Villalobos, P., Ho, A., Sevilla, J., Besiroglu, T., Heim, L., Hobbhahn, M.: Position: Will we run out of data? limits of LLM scaling based on human- generated data. In: Forty-first International Conference on Machine Learning (2024). https://openreview.net/forum?id=ViZcgDQjyG

    work page 2024

  34. [34]

    doi:10.1038/s41586-024-07566-y , url =

    Shumailov, I., Shumaylov, Z., Zhao, Y., Papernot, N., Anderson, R., Gal, Y.: Ai models collapse when trained on recursively generated data. Nature 631(8022), 755–759 (2024) https://doi.org/10.1038/s41586-024-07566-y

    work page doi:10.1038/s41586-024-07566-y 2024

  35. [35]

    Is model collapse inevitable? Breaking the curse of recursion by accumulating real and synthetic data,

    Gerstgrasser, M., Schaeffer, R., Dey, A., Rafailov, R., Sleight, H., Hughes, J., Korbak, T., Agrawal, R., Pai, D., Gromov, A., et al.: Is model collapse inevitable? breaking the curse of recursion by accumulating real and synthetic data. Preprint at https://arxiv.org/abs/2404.01413 (2024)

    work page arXiv 2024

  36. [36]

    Journal of the European Meteorological Society 1, 100002 (2024)

    Bauer, P.: What if? numerical weather prediction at the crossroads. Journal of the European Meteorological Society 1, 100002 (2024)

    work page 2024

  37. [37]

    (No Title) (2019) 18

    Kenneth, R., Howard, J., James, P., Michael, C., Carl, J.: International best track archive for climate stewardship (ibtracs) project, version 4. (No Title) (2019) 18

    work page 2019

  38. [38]

    Advances in Atmospheric Sciences 38(4), 690–699 (2021)

    Lu, X., Yu, H., Ying, M., Zhao, B., Zhang, S., Lin, L., Bai, L., Wan, R.: West- ern north pacific tropical cyclone database created by the china meteorological administration. Advances in Atmospheric Sciences 38(4), 690–699 (2021)

    work page 2021

  39. [39]

    Journal of Atmospheric and Oceanic Technology 31(2), 287–301 (2014)

    Ying, M., Zhang, W., Yu, H., Lu, X., Feng, J., Fan, Y., Zhu, Y., Chen, D.: An overview of the china meteorological administration tropical cyclone database. Journal of Atmospheric and Oceanic Technology 31(2), 287–301 (2014)

    work page 2014

  40. [40]

    arXiv preprint arXiv:2409.07188 (2024)

    Zhong, X., Chen, L., Fan, X., Qian, W., Liu, J., Li, H.: Fuxi-2.0: Advancing machine learning weather forecasting model for practical applications. arXiv preprint arXiv:2409.07188 (2024)

    work page arXiv 2024

  41. [41]

    npj climate and atmospheric science 6(1), 190 (2023)

    Chen, L., Zhong, X., Zhang, F., Cheng, Y., Xu, Y., Qi, Y., Li, H.: Fuxi: a cas- cade machine learning forecasting system for 15-day global weather forecast. npj climate and atmospheric science 6(1), 190 (2023)

    work page 2023

  42. [42]

    A Description of the Advanced Research WRF Model Version 4.3 (2021) NCAR/TN-556+ STR 556 (2021)

    Jensen, A.A., Gill, D.O., Powers, J.G., Duda, M.G.: A description of the advanced research wrf model version 4.3. A Description of the Advanced Research WRF Model Version 4.3 (2021) NCAR/TN-556+ STR 556 (2021)

    work page 2021

  43. [43]

    Geophysical Research Letters 52(15), 2025–115926 (2025)

    Niu, Z., Wang, D., Mu, M., Huang, W., Fan, X., Yang, M., Qin, B.: Machine- learning (ml)-physics fusion model accelerates the paradigm shift in typhoon forecasting with a cnop-based assimilation framework. Geophysical Research Letters 52(15), 2025–115926 (2025)

    work page 2025

  44. [44]

    part i: Description and sensitivity analysis

    Thompson, G., Rasmussen, R.M., Manning, K.: Explicit forecasts of winter pre- cipitation using an improved bulk microphysics scheme. part i: Description and sensitivity analysis. Monthly Weather Review 132(2), 519–542 (2004)

    work page 2004

  45. [45]

    Monthly Weather Review 144(3), 833–860 (2016)

    Zheng, Y., Alapaty, K., Herwehe, J.A., Del Genio, A.D., Niyogi, D.: Improving high-resolution weather forecasts using the weather research and forecasting (wrf) model with an updated kain–fritsch scheme. Monthly Weather Review 144(3), 833–860 (2016)

    work page 2016

  46. [46]

    Journal of Geophysical Research: Atmospheres 113(D13) (2008)

    Iacono, M.J., Delamere, J.S., Mlawer, E.J., Shephard, M.W., Clough, S.A., Collins, W.D.: Radiative forcing by long-lived greenhouse gases: Calculations with the aer radiative transfer models. Journal of Geophysical Research: Atmospheres 113(D13) (2008)

    work page 2008

  47. [47]

    Journal of Geophysical Research: Atmospheres 108(D22) (2003)

    Ek, M., Mitchell, K., Lin, Y., Rogers, E., Grunmann, P., Koren, V., Gayno, G., Tarpley, J.: Implementation of noah land surface model advances in the national centers for environmental prediction operational mesoscale eta model. Journal of Geophysical Research: Atmospheres 108(D22) (2003)

    work page 2003

  48. [48]

    Journal of Geophysical Research: Atmospheres 118(18), 10–490 (2013) 19

    Hu, X.-M., Klein, P.M., Xue, M.: Evaluation of the updated ysu planetary bound- ary layer scheme within wrf for wind resource and air quality assessments. Journal of Geophysical Research: Atmospheres 118(18), 10–490 (2013) 19

    work page 2013

  49. [49]

    Weather and Forecasting 1(aop) (2025)

    Husain, S.Z., Separovic, L., Caron, J.-F., Aider, R., Buehner, M., Chamberland, S., Lapalme, E., McTaggart-Cowan, R., Subich, C., Vaillancourt, P.A., et al.: Leveraging data-driven weather models for improving numerical weather predic- tion skill through large-scale spectral nudging. Weather and Forecasting 1(aop) (2025)

    work page 2025

  50. [50]

    In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed- ical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pp. 234–241. Springer, Cham (2015)

    work page 2015

  51. [51]

    In: NIPS 2017 Workshop on Autodiff (2017)

    Paszke, A., Gross, S., Chintala, S., Chanan, G., Yang, E., DeVito, Z., Lin, Z., Desmaison, A., Antiga, L., Lerer, A.: Automatic Differentiation in PyTorch. In: NIPS 2017 Workshop on Autodiff (2017)

    work page 2017

  52. [52]

    Adam: A Method for Stochastic Optimization

    Kingma, D.P., Ba, J.: Adam: A Method for Stochastic Optimization. Preprint at https://arxiv.org/abs/1412.6980 (2017)

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  53. [53]

    Loshchilov, I., Hutter, F.: Decoupled Weight Decay Regularization. In: Int. Conf. Learn. Represent. (2017)

    work page 2017

  54. [54]

    P., Lucic, A., Stanley, M., Allen, A., Brandstetter, J., Garvan, P., Riechert, M., Weyn, J

    Bodnar, C., Bruinsma, W.P., Lucic, A., Stanley, M., Allen, A., Brandstetter, J., Garvan, P., Riechert, M., Weyn, J.A., Dong, H., et al.: A foundation model for the earth system. Nature, 1–8 (2025) https://doi.org/10.1038/s41586-025-09005-y 20

    work page doi:10.1038/s41586-025-09005-y 2025