Pith. sign in

REVIEW 4 major objections 5 minor 40 references

LangYa: Revolutionizing Cross-Spatiotemporal Ocean Forecasting

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read LangYa, a single AI model, claims 1–7 day global ocean forecasts at 1/12° resolution with about a 50% thermocline accuracy gain over existing numerical and AI systems.

desk verdict A well-engineered AI ocean forecasting system with honest limitations, but the headline accuracy claims are not yet verified because LangYa's RMSE is computed against the same GLORYS12 reanalysis used for training. read the letter →

arxiv 2412.18097 v3 pith:6KAOYOJ4 submitted 2024-12-24 physics.ao-ph cs.AI

Nan Yang , Chong Wang , Meihua Zhao , Zimeng Zhao , Huiling Zheng , Bin Zhang , Jianing Wang , Xiaofeng Li This is my paper · ORCID
classification physics.ao-phcs.AI
keywords oceanforecastinglargeAImodelscross-spatiotemporalair-seacouplingthermoclineLLMtimeembeddingGLORYS12self-attention
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

LangYa is an attempt to show that one neural network, rather than a stack of single-day or single-variable models, can forecast the global ocean's temperature, salinity, and currents for lead times of 1 to 7 days at 1/12° daily resolution. The paper argues that explicit time information supplied by an LLM-based embedding, a stochastic sampling scheme for atmosphere-ocean coupling, cosine self-attention, and an adaptive thermocline loss together make this possible. A sympathetic reader would care because the claimed result—roughly 50% better thermocline accuracy and lower RMSE than the leading numerical and AI systems, at about 10,000 times lower inference cost—would make AI ocean forecasting a practical alternative to operational numerical models.

What carries the argument

Four mechanisms carry the argument. The LLM-based time embedding encodes the initial date and lead time as $E(t,\tau) = E_Y(t_Y) \oplus E_M(t_M) \oplus E_D(t_D) + E_\tau(\tau)$, with the year coefficients drawn from an instruction-tuned LLM, so a single model can forecast any lead time from 1 to 7 days. The asynchronous cross-iterative random sampling strategy writes the air-sea coupling as $H(X_t, A_{(t-10):t}) \triangleq X_t \oplus SS_t \oplus \sum_{k=t-10}^{t} H(A_k) * A_k$, where each $H(A_k)$ is a learned scalar weight for a past day's atmospheric field; this is meant to capture the stochastic, lagged influence of the atmosphere on the ocean. The ocean self-attention module replaces dot-product attention with cosine attention plus post-normalization in Swin Transformer blocks to stabilize training. The adaptive thermocline loss multiplies the temperature error by a normalized vertical-gradient term $\mathrm{Norm}(\partial [X_T]^*_t/\partial z) + 1$, forcing the network to spend capacity where the thermocline gradient is sharp.

What would settle it

Score LangYa's 2020–2021 forecasts against independent Argo float profiles, not GLORYS12 reanalysis, and check whether the reported ~50% thermocline RMSE improvement over XiHe and the numerical models survives; if it does not, the central accuracy claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that LangYa, a single trained model, produces deterministic global 1/12° daily forecasts of ocean temperature, salinity, and zonal and meridional currents for 1–7 day lead times in one inference pass, avoiding both the cumulative drift of autoregressive forecasting and the cost of training a separate model for each lead time. The paper further claims that its four modules each earn their place: the LLM-based time embedding enables cross-spatiotemporal forecasts; the asynchronous cross-iterative random sampling strategy simulates lagged atmospheric forcing; the ocean self-attention module stabilizes training; and the adaptive thermocline loss yields about a 50% improvement in thermocline forecast accuracy. On the 2020–2021 test period, LangYa is reported to achieve lower RMSE than XiHe, XiHe-AR, PSY4, BLK, GIOPS, and FOAM for temperature and currents, with salinity comparable to the best of them, and to run 10,000 times faster than operational numerical models.

Load-bearing premise

The load-bearing premise is that GLORYS12 reanalysis, the same model-based product used for training, is an adequate stand-in for the true ocean when measuring forecast error, so reanalysis-aligned biases could inflate the reported skill.

Editorial extensions

If this is right

  • One consequence is that a single LangYa model can replace the usual suite of one-model-per-lead-time or autoregressive chains for 1–7 day global ocean forecasts.
  • A second is that the adaptive thermocline loss, if it carries over to independent conditions, gives AI forecasts their clearest advantage in the 50–300 m layer where vertical temperature gradients govern mixing and heat exchange.
  • A third is that one-second inference on a single GPU makes global daily ocean-state forecasting cheap enough for real-time operational use and for large ensemble runs.
  • A fourth is that treating forecast lead time as an explicit conditioning variable, rather than as a separate training target, is what allows one model to exhibit cross-spatiotemporal skill.
  • Finally, day-7 forecasts are made in a single pass, so cumulative error from chaining daily forecasts is removed by construction.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: If time embedding is as general as claimed, the same single-model design could extend to lead times beyond 7 days or to other ocean variables such as sea ice and biogeochemical tracers without adding sub-models.
  • Editorial inference: The random-sampling coupling treats air-sea interaction as a learned stochastic perturbation; a natural test the paper does not report is whether the spread of LangYa ensemble members across sampling draws tracks the actual forecast error, which would show the coupling is capturing real uncertainty rather than noise.
  • Editorial inference: Because the headline verification uses the same reanalysis family as the training data, the paper's own stated limitation about cross-data-source adaptability is the key risk; evaluating LangYa initialized from an operational data assimilation system would be a more demanding test than the GLORYS12-based comparison.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This manuscript presents LangYa, a data-driven global ocean forecasting system trained on GLORYS12 and ERA5 reanalyses (1993-2019) and evaluated on 2020-2021. The model forecasts temperature, salinity, and zonal/meridional currents on 32 depth levels at 1/12 degree resolution for lead times of 1-7 days with a single set of weights. The four claimed contributions are an LLM-based time embedding that encodes initial date and lead time, an asynchronous cross-iterative random sampling strategy that ingests 10-day atmospheric histories with learned scalar weights, a cosine-attention ocean self-attention module, and an adaptive thermocline loss that weights temperature error by the vertical temperature gradient. The authors report lower RMSE than XiHe, XiHe-AR, PSY4, BLK, GIOPS, and FOAM, and a roughly 50% improvement in thermocline forecast accuracy.

Significance. If the performance claims were verified against an independent observational reference, LangYa would be a significant contribution to AI-based ocean forecasting: a single model spanning multiple lead times at eddy-resolving resolution, with inference of about one second and a thermocline-focused loss, is exactly the kind of system the community needs. The paper also makes useful design choices that are clearly described, uses publicly available data, compares against several operational systems, and adds MAE and PSNR metrics to RMSE. The strengths are, however, currently offset by the evaluation design: the headline metrics appear to use the same reanalysis product for training and verification, the comparison to numerical models is not on a common reference, and the thermocline result is tied to the loss function in a way that makes the improvement partly by construction. These issues are fixable with additional experiments and reporting, but they are central to the paper's claims.

major comments (4)
  1. [Methods, Model Training and Metrics; Experimental Setting and Main Results] The headline comparisons in Figs. 2 and 5 are not defined against a common reference. Equation (10) computes RMSE against X_t^*, and the training/evaluation description indicates that X_t^* is GLORYS12 reanalysis for LangYa, while the numerical models and XiHe are evaluated through the IV-TT Class 4 observational framework. A lower RMSE against reanalysis does not establish lower error against observations, especially because GLORYS12 is itself a model-based product and the Discussion concedes that LangYa is 'trained and tested on reanalysis data.' Please report, for every system and every lead time, RMSE/MAE computed on the same grid and on the same observational points, and provide confidence intervals or significance tests.
  2. [Adaptive Thermocline Loss Function, Eqs. (8)-(9)] The claim of 'approximately a 50% improvement in thermocline forecast accuracy' is not quantified in the main text and is partly by construction. Equation (8) adds a loss term proportional to the vertical temperature gradient of the ground truth, and Eq. (9) trains with that term; therefore a lower thermocline RMSE relative to a model trained without the term is expected. To support the physical claim, evaluate thermocline error against independent Argo profiles, report the exact percentage and how it is computed from Fig. 5, and show that the improvement persists when the thermocline region is defined independently of the loss weighting.
  3. [Atmospheric Forcing Ocean Module, Eq. (6)] The asynchronous cross-iterative random sampling strategy is a load-bearing component, but the paper does not validate that it captures atmosphere-ocean coupling rather than acting as stochastic augmentation. The learned scalar weights H(A_k) and the random sampling over a 10-day window need diagnostics: report the learned weights as a function of lag, compare with a fixed-window or no-atmosphere baseline, and check physical consistency (for example, wind-forced response in ocean currents). Without such evidence, the claim that the module 'effectively simulates air-sea interactions' is unsupported.
  4. [Experimental Setting and Main Results; Comparative Methods] The paper states that LangYa was evaluated 'with both reanalysis and observational datasets,' but no figure or table reports the observational evaluation, and the IV-TT Class 4 framework is not connected to any LangYa curve. Please add a dedicated observational verification section with masks, sample sizes, and a clear statement of which systems are compared at observation points and which are compared on the reanalysis grid. This is necessary to support the abstract's claim of superiority over numerical and AI-based systems.
minor comments (5)
  1. [Methods, Model Training and Metrics, Eq. (10)] The MSE formula has malformed exponent notation and the MAE expression appears to be missing the outer average over grid points; please fix the typography and confirm the metric definitions.
  2. [Fig. 1] The label 'T Yp e K ey M odules' has unintended spacing and should likely read 'Type Key Modules'.
  3. [Eqs. (1) and (6)] The symbol H is used both for the whole air-sea interaction module in Eq. (1) and for the learned scalar weight in Eq. (6); please use distinct notation to avoid ambiguity.
  4. [Comparative Methods] The attribution of FOAM to the Institute of Atmospheric Physics, CAS, conflicts with standard references, which identify FOAM as the UK Met Office's Forecast Ocean Assimilation Model; please verify and cite the correct source.
  5. [Code Availability] The code availability section promises future release of trained models and inference code; for a paper making strong performance claims, please provide the evaluation scripts and trained model at submission, or state explicitly that the results are not yet reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found: LangYa's forecasting claims rest on held-out reanalysis evaluation and external benchmark comparisons, not on equations that reduce to their inputs.

full rationale

The paper's derivation chain is a standard supervised learning pipeline: GLORYS12 and ERA5 inputs are transformed by the architecture in Eqs. 1-7, optimized against the loss in Eqs. 8-9, and then evaluated on reserved 2020-2021 data and compared with XiHe and numerical models (Figs. 2 and 5). The adaptive thermocline loss (Eq. 8) does upweight errors in high-vertical-gradient regions, so improved thermocline RMSE is a plausible empirical consequence of the design; however, the paper reports measured RMSE differences rather than deriving the claimed roughly 50% improvement directly from the loss, so this is not a circular reduction. Self-citations such as [16] and [18] provide general context and are not load-bearing for the central forecast-accuracy claims. The Discussion explicitly concedes that LangYa is trained and tested on reanalysis data, which is a validation-risk and generalizability concern rather than a circularity, and it does not make the headline comparisons equivalent to the model's inputs by construction.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical entities. Its free parameters are mostly architectural hyperparameters and the thermocline-loss weight. The main non-empirical burden is the reliance on reanalysis as truth and the heuristic transfer of LLM embeddings to temporal encoding.

free parameters (3)
  • Loss weight lambda = 1.0
    Equation 9 sets lambda = 1.0 for the thermocline loss term. This scalar is chosen by hand and directly controls the contribution of the thermocline loss to the overall training objective, affecting the reported thermocline accuracy.
  • Number of atmospheric lookback days (10) = 10
    The atmospheric forcing module uses a fixed 10-day lookback window in Equation 6. This is a design choice not derived from atmospheric decorrelation times and affects the model's effective input.
  • Windowing sizes and patch size = window 5x12, patch 4x4
    These are architecture hyperparameters chosen without a systematic search, but they are not fitted to the target data directly; they affect efficiency and performance.
assumptions (3)
  • domain assumption GLORYS12 reanalysis is a valid ground truth for forecast verification.
    The paper computes RMSE against GLORYS12 for the main skill scores. Reanalysis is itself a model with uncertainties, and training on the same product biases fair comparison to independent observations.
  • ad hoc to paper Swin Transformer with cosine attention can extract stable spatiotemporal features for ocean forecasting.
    The ocean self-attention module is presented as a stability fix, but no rigorous proof or theoretical analysis is given, only empirical claims on this model.
  • domain assumption The LLM-based time embedding captures meaningful temporal context.
    Using Llama3.1-8B hidden states to represent calendar year is a heuristic transfer from language modeling to physical time, with no evidence that this beats simpler sinusoidal embeddings.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LangYa: Revolutionizing Cross-Spatiotemporal Ocean Forecasting." pith.science (2026). https://pith.science/paper/6KAOYOJ4

@misc{pith2026241218097,
  author       = {Pith},
  title        = {Pith review of: LangYa: Revolutionizing Cross-Spatiotemporal Ocean Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6KAOYOJ4}},
  note         = {Machine review of arXiv:2412.18097}
}
read the original abstract

Ocean forecasting is crucial for both scientific research and societal benefits. Currently, the most accurate forecasting systems are global ocean forecasting systems (GOFSs), which represent the ocean state variables (OSVs) as discrete grids and solve partial differential equations (PDEs) governing the transitions of oceanic state variables using numerical methods. However, GOFSs processes are computationally expensive and prone to cumulative errors. Recently, large artificial intelligence (AI)-based models significantly boosted forecasting speed and accuracy. Unfortunately, building a large AI ocean forecasting system that can be considered cross-spatiotemporal and air-sea coupled forecasts remains a significant challenge. Here, we introduce LangYa, a cross-spatiotemporal and air-sea coupled ocean forecasting system. Results demonstrate that the time embedding module in LangYa enables a single model to make forecasts with lead times ranging from 1 to 7 days. The air-sea coupled module effectively simulates air-sea interactions. The ocean self-attention module improves network stability and accelerates convergence during training, and the adaptive thermocline loss function improves the accuracy of thermocline forecasting. Compared to existing numerical and AI-based ocean forecasting systems, LangYa uses 27 years of global ocean data from the Global Ocean Reanalysis and Simulation version 12 (GLORYS12) for training and achieves more reliable deterministic forecasting results for OSVs. LangYa forecasting system provides global ocean researchers with access to a powerful software tool for accurate ocean forecasting and opens a new paradigm for ocean science.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 29 canonical work pages

  1. [1]

    10 Environmental Management 8, 27–43 (1984)

    Teal, J.M., Howarth, R.W.: Oil spill studies: a review of ecological effects. 10 Environmental Management 8, 27–43 (1984)

    work page 1984

  2. [2]

    Journal of Operational Oceanography 13(1), 1–18 (2020)

    Schiller, A., Brassington, G.B., Oke, P., Cahill, M., Divakaran, P., Entel, M., Freeman, J., Griffin, D., Herzfeld, M., Hoeke, R.,et al.: Bluelink ocean forecasting australia: 15 years of operational ocean service delivery with societal, economic and environmental benefits. Journal of Operational Oceanography 13(1), 1–18 (2020)

    work page 2020

  3. [3]

    International Journal of Environmental Science and Technology 16(3), 1789–1806 (2019)

    Zohdi, E., Abbaspour, M.: Harmful algal blooms (red tide): a review of causes, impacts and approaches to monitoring and prediction. International Journal of Environmental Science and Technology 16(3), 1789–1806 (2019)

    work page 2019

  4. [4]

    Ocean Dynamics 63, 83–88 (2013)

    Breivik, Ø., Allen, A.A., Maisondieu, C., Olagnon, M.: Advances in search and rescue at sea. Ocean Dynamics 63, 83–88 (2013)

    work page 2013

  5. [5]

    Research activities in Earth system modelling (2020)

    Garraffo, Z.D., Cummings, J.A., Paturi, S., Hao, Y., Iredell, D., Spindler, T., et al.: Rtofs-da: real time ocean-sea ice coupled three dimensional variational global data assimilative ocean forecast system. Research activities in Earth system modelling (2020)

    work page 2020

  6. [6]

    Bulletin of the American Meteorological Society 101(8), 1340–1356 (2020)

    Francis, P., Jithin, A., Effy, J., Chatterjee, A., Chakraborty, K., Paul, A., Balaji, B., Shenoi, S., Biswamoy, P., Mukherjee, A., et al.: High-resolution oper- ational ocean forecast and reanalysis system for the indian ocean. Bulletin of the American Meteorological Society 101(8), 1340–1356 (2020)

    work page 2020

  7. [7]

    New Frontiers in Operational Oceanography 1, 563–592 (2018)

    Lellouche, J.-M., Greiner, E., Le Galloudec, O., Regnier, C., Benkiran, M., Testut, C.-E., Bourdalle-Badie, R., Drevillon, M., Garric, G., Drillet, Y.: The mercator ocean global high-resolution monitoring and forecasting system. New Frontiers in Operational Oceanography 1, 563–592 (2018)

    work page 2018

  8. [8]

    Journal of Operational Oceanography 8(sup2), 201–220 (2015)

    Tonani, M., Balmaseda, M., Bertino, L., Blockley, E., Brassington, G., Davidson, F., Drillet, Y., Hogan, P., Kuragano, T., Lee, T.,et al.: Status and future of global and regional ocean prediction systems. Journal of Operational Oceanography 8(sup2), 201–220 (2015)

    work page 2015

Show all 40 references
  1. [9]

    Nature 525(7567), 47–55 (2015)

    Bauer, P., Thorpe, A., Brunet, G.: The quiet revolution of numerical weather prediction. Nature 525(7567), 47–55 (2015)

    2015

  2. [10]

    Quarterly Journal of the Royal Meteorological Society 142(695), 659–671 (2016)

    Smith, G.C., Roy, F., Reszka, M., Surcel Colan, D., He, Z., Deacu, D., Belanger, J.-M., Skachko, S., Liu, Y., Dupont, F., et al.: Sea ice forecast verification in the canadian global ice ocean prediction system. Quarterly Journal of the Royal Meteorological Society 142(695), 6...

    2016

  3. [11]

    Frontiers in Earth Science 9, 698876 (2021) 11

    Jean-Michel, L., Eric, G., Romain, B.-B., Gilles, G., Ang´ elique, M., Marie, D., Cl´ ement, B., Mathieu, H., Olivier, L.G., Charly, R.,et al.: The copernicus global 1/12 oceanic and sea ice glorys12 reanalysis. Frontiers in Earth Science 9, 698876 (2021) 11

    2021

  4. [12]

    Nature 566(7743), 195–204 (2019)

    Reichstein, M., Camps-Valls, G., Stevens, B., Jung, M., Denzler, J., Carvalhais, N., Prabhat, F.: Deep learning and process understanding for data-driven earth system science. Nature 566(7743), 195–204 (2019)

    2019

  5. [13]

    ECMWF Technical Memoranda878, 24 (2021)

    D¨ uben, P., Modigliani, U., Geer, A., Siemen, S., Pappenberger, F., Bauer, P., Brown, A., Palkovic, M., Raoult, B., Wedi, N., et al.: Machine learning at ecmwf: A roadmap for the next 10 years. ECMWF Technical Memoranda878, 24 (2021)

    2021

  6. [14]

    Nature Communications 16(1), 1919 (2025)

    Camps-Valls, G., Fern´ andez-Torres, M.-´A., Cohrs, K.-H., H¨ ohl, A., Castelletti, A., Pacal, A., Robin, C., Martinuzzi, F., Papoutsis, I., Prapas, I., et al.: Artificial intelligence for modeling and understanding extreme weather and climate events. Nature Communications 16(...

    2025

  7. [15]

    arXiv preprint arXiv:2409.04109 (2024)

    Si, C., Yang, D., Hashimoto, T.: Can llms generate novel research ideas? a large- scale human study with 100+ nlp researchers. arXiv preprint arXiv:2409.04109 (2024)

    2024 arXiv

  8. [16]

    Proceedings of the National Academy of Sciences 122(4), 2415501122 (2025)

    Wang, C., Yang, N., Li, X.: Advancing forecasting capabilities: A contrastive learning model for forecasting tropical cyclone rapid intensification. Proceedings of the National Academy of Sciences 122(4), 2415501122 (2025)

    2025

  9. [17]

    Remote Sens

    Wang, H., Li, X.: DeepBlue: Advanced convolutional neural network applications for ocean remote sensing, IEEE Geosci. Remote Sens. Mag., 12, 138–161 (2023)

    2023

  10. [18]

    Springer, ??? (2023)

    Li, X., Wang, F.: Artificial Intelligence Oceanography. Springer, ??? (2023)

    2023

  11. [19]

    National Science Review 7(10), 1584–1605 (2020)

    Li, X., Liu, B., Zheng, G., Ren, Y., Zhang, S., Liu, Y., Gao, L., Liu, Y., Zhang, B., Wang, F.: Deep-learning-based information mining from ocean remote-sensing imagery. National Science Review 7(10), 1584–1605 (2020)

    2020

  12. [20]

    Nature 597(7878), 672–677 (2021)

    Ravuri, S., Lenc, K., Willson, M., Kangin, D., Lam, R., Mirowski, P., Fitzsi- mons, M., Athanassiadou, M., Kashem, S., Madge, S., et al.: Skilful precipitation nowcasting using deep generative models of radar. Nature 597(7878), 672–677 (2021)

    2021

  13. [21]

    arXiv preprint arXiv:2202.11214 (2022)

    Pathak, J., Subramanian, S., Harrington, P., Raja, S., Chattopadhyay, A., Mar- dani, M., Kurth, T., Hall, D., Li, Z., Azizzadenesheli, K., et al.: Fourcastnet: A global data-driven high-resolution weather model using adaptive fourier neural operators. arXiv preprint arXiv:2202...

    2022 arXiv

  14. [22]

    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)

    2023

  15. [23]

    Science 382(6677), 1416–1421 (2023) 12

    Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W., et al.: Learning skillful medium-range global weather forecasting. Science 382(6677), 1416–1421 (2023) 12

    2023

  16. [24]

    arXiv preprint arXiv:2304.02948 (2023)

    Chen, K., Han, T., Gong, J., Bai, L., Ling, F., Luo, J.-J., Chen, X., Ma, L., Zhang, T., Su, R., et al.: Fengwu: Pushing the skillful global medium-range weather forecast beyond 10 days lead. arXiv preprint arXiv:2304.02948 (2023)

    2023

  17. [25]

    Nature 632(8027), 1060–1066 (2024)

    Kochkov, D., Yuval, J., Langmore, I., Norgaard, P., Smith, J., Mooers, G., Kl¨ ower, M., Lottes, J., Rasp, S., D¨ uben, P.,et al.: Neural general circulation models for weather and climate. Nature 632(8027), 1060–1066 (2024)

    2024

  18. [26]

    arXiv preprint arXiv:2308.03152 (2023)

    Xiong, W., Xiang, Y., Wu, H., Zhou, S., Sun, Y., Ma, M., Huang, X.: Ai-goms: Large ai-driven global ocean modeling system. arXiv preprint arXiv:2308.03152 (2023)

    2023 arXiv

  19. [27]

    arXiv preprint arXiv:2402.02995 (2024)

    Wang, X., Wang, R., Hu, N., Wang, P., Huo, P., Wang, G., Wang, H., Wang, S., Zhu, J., Xu, J., et al.: Xihe: A data-driven model for global ocean eddy-resolving forecasting. arXiv preprint arXiv:2402.02995 (2024)

    2024 arXiv

  20. [28]

    arXiv preprint arXiv:2412.05454 (2024)

    Aouni, A.E., Gaudel, Q., Regnier, C., Van Gennip, S., Drevillon, M., Drillet, Y., Lellouche, J.-M.: Glonet: Mercator’s end-to-end neural forecasting system. arXiv preprint arXiv:2412.05454 (2024)

    2024

  21. [29]

    Nature Communications 16(1), 2268 (2025)

    Cui, Y., Wu, R., Zhang, X., Zhu, Z., Liu, B., Shi, J., Chen, J., Liu, H., Zhou, S., Su, L., et al.: Forecasting the eddying ocean with a deep neural network. Nature Communications 16(1), 2268 (2025)

    2025

  22. [30]

    arXiv preprint arXiv:1607.06450 (2016)

    Ba, J.L., Kiros, J.R., Hinton, G.E.: Layer normalization. arXiv preprint arXiv:1607.06450 (2016)

    2016 arXiv

  23. [31]

    In: Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pp

    Chen, C.-F.R., Fan, Q., Panda, R.: Crossvit: Cross-attention multi-scale vision transformer for image classification. In: Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision, pp. 357–366 (2021)

    2021

  24. [32]

    In: European Conference on Computer Vision, pp

    Touvron, H., Cord, M., J´ egou, H.: Deit iii: Revenge of the vit. In: European Conference on Computer Vision, pp. 516–533 (2022). Springer

    2022

  25. [33]

    Journal of Geophysical Research: Oceans 125(6), 2019–014978 (2020)

    Xue, P., Malanotte-Rizzoli, P., Wei, J., Eltahir, E.A.: Coupled ocean-atmosphere modeling over the maritime continent: A review. Journal of Geophysical Research: Oceans 125(6), 2019–014978 (2020)

    2020

  26. [34]

    Quarterly Journal of the Royal Meteorological Society 147(741), 4186–4227 (2021)

    Bell, B., Hersbach, H., Simmons, A., Berrisford, P., Dahlgren, P., Hor´ anyi, A., Mu˜ noz-Sabater, J., Nicolas, J., Radu, R., Schepers, D., et al.: The era5 global reanalysis: Preliminary extension to 1950. Quarterly Journal of the Royal Meteorological Society 147(741), 4186–4...

    2021

  27. [35]

    Computational and Information Systems Labora- tory https://climatedataguide

    Saha, S., et al.: Research data archive at the national center for atmospheric research. Computational and Information Systems Labora- tory https://climatedataguide. ucar. edu/climate-data/climate-forecast-system- reanalysis-cfsr (last access: 2 January 2023) (2010) 13

    2010

  28. [36]

    Journal of Operational Oceanography 8(sup1), 98–111 (2015)

    Ryan, A., Regnier, C., Divakaran, P., Spindler, T., Mehra, A., Smith, G., David- son, F., Hernandez, F., Maksymczuk, J., Liu, Y.: Godae oceanview class 4 forecast verification framework: global ocean inter-comparison. Journal of Operational Oceanography 8(sup1), 98–111 (2015)

    2015

  29. [37]

    Journal of oceanography 73, 503–526 (2017)

    Chu, P.C., Fan, C.: Exponential leap-forward gradient scheme for determining the isothermal layer depth from profile data. Journal of oceanography 73, 503–526 (2017)

    2017

  30. [38]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., Guo, B.: Swin trans- former: Hierarchical vision transformer using shifted windows. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 10012–10022 (2021)

    2021

  31. [39]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Liu, Z., Hu, H., Lin, Y., Yao, Z., Xie, Z., Wei, Y., Ning, J., Cao, Y., Zhang, Z., Dong, L., et al.: Swin transformer v2: Scaling up capacity and resolution. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 12009–12019 (2022)

    2022

  32. [40]

    Advances in neural information processing systems 33, 6840–6851 (2020) Methods Dataset The OSVs data used in LangYa are derived from the GLORYS12 reanalysis data

    Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. Advances in neural information processing systems 33, 6840–6851 (2020) Methods Dataset The OSVs data used in LangYa are derived from the GLORYS12 reanalysis data. GLORYS12 was developed and implemented und...

    2020

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.