Pith. sign in

REVIEW 6 major objections 6 minor 80 references

Deep Learning Weather Models for Subregional Ocean Forecasting: A Case Study on the Canary Current Upwelling System

T0 review · 6 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read An ocean-adapted graph neural network trained on satellite sea surface temperature can beat numerical ocean models at medium-range forecasts in the Canary Current upwelling region.

desk verdict A careful, useful adaptation of GraphCast to a regional ocean domain, with honest self-criticism, but the headline skill numbers are measured against the same L4 product used for training and lack in-situ validation. read the letter →

arxiv 2505.24429 v2 pith:LRCJXCWW submitted 2025-05-30 physics.ao-ph cs.AIcs.LG

Giovanny A. Cuervo-Londoño , Javier Sánchez , Ángel Rodríguez-Santana This is my paper
classification physics.ao-phcs.AIcs.LG
keywords SeasurfacetemperatureforecastingGraphneuralnetworksCanaryCurrentUpwellingSystemData-drivenoceanpredictionOperationaloceanographyConvLSTMGLORYSreanalysisMedium-range
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

This paper tries to establish that a graph neural network originally built for global weather forecasting, retrained on satellite sea surface temperature fields, can beat both a convolutional LSTM baseline and high-resolution numerical ocean products (GLORYS reanalysis and the PSY4V3R1 forecast system) at 1-20 day lead times in the Canary Current upwelling system. The headline results are RMSE reductions of roughly 76% over GLORYS at 5-day lead times, about 48% at 10 days, skill improvements over ConvLSTM of 19-26.5%, and a 20-day forecast produced in about 140 seconds, roughly 100 times faster than a 7-day numerical simulation. The authors argue that this demonstrates fine-scale mesoscale features like upwelling filaments and eddies can be learned directly from observations, without solving physical equations or using explicit wind and bathymetry forcing. If true, it makes fast, cheap, medium-term regional ocean forecasting practical for operational oceanography and for sectors like fisheries and marine conservation.

What carries the argument

The central machinery is the multiscale graph that connects the 300x300 SST grid to a curvilinear triangular mesh refined to three levels ($M^{3}$), with directed grid-to-mesh and mesh-to-grid edges plus bidirectional mesh edges. An interaction network performs six message-passing steps in an 8-dimensional latent space to propagate SST information across scales, and a binary land/ocean mask in the loss function restricts learning to ocean cells. The model is autoregressive: it is trained to map two consecutive SST fields to the next field, then iterated to produce 20-day rollouts.

What would settle it

Compute the same 5- and 10-day RMSE comparisons using independent in-situ SST observations (moorings, drifters, or Argo profiles) in the Canary Current upwelling region over 2017-2020; if the graph model's advantage over GLORYS and ConvLSTM largely disappears, or its errors exceed the L4 instrumental threshold of 0.25 degrees C as often as theirs do, the claimed skill is an artifact of verifying on the training product.

Watch

Extended reading notes

Core claim

The paper claims that an ocean-adapted version of the graph neural network GraphCast can forecast sea surface temperature in the Canary Current upwelling system with higher accuracy than a ConvLSTM baseline and the numerical GLORYS reanalysis, over lead times of 1-20 days. Trained entirely on the L4 satellite SST product (1982-2020, 0.05 degrees resolution) with a spatially masked loss that ignores land, the autoregressive model predicts x_{t+1}=f(x_t,x_{t-1}) through an encoder-processor-decoder over a three-level curvilinear mesh. Verification against the same L4 product for 2017-2020 shows RMSE reductions of about 76% at 5 days and 47-48% at 10 days relative to GLORYS, and year-by-year skill improvements of 19.4-26.5% over ConvLSTM, with the largest gains at Cape Ghir, Cape Bojador, and Cape Blanc. The paper also reports the model's limits: it crosses the L4 instrumental error threshold around day 8, becomes increasingly overactive at longer leads, produces triangular decoder artifacts that raise RMSE variability, and degrades faster than ConvLSTM when initial-condition error is large. The claimed discovery is that a weather GNN, retrained on observations and given a regional mesh, is a viable medium-range SST forecaster for an energetic eastern-boundary upwelling system.

Load-bearing premise

The load-bearing premise is that the CMEMS L4 satellite SST product is an unbiased, gap-free ground truth, because the same product is used to train the model and to score every comparison, and no check against independent in-situ measurements is reported.

Editorial extensions

If this is right

  • At 5-day lead times the model's RMSE is about 76% lower than GLORYS, and at 10 days about 48% lower, so medium-range SST forecasts in this region can be produced with a small fraction of the numerical-model error.
  • Because a 20-day forecast runs in about 140 seconds on one GPU versus roughly 4 hours for a 7-day GLORYS simulation, operational schedules could shift from daily batch runs to on-demand, ensemble-style forecasts.
  • The model retains clear skill at Cape Ghir, Cape Bojador, and Cape Blanc, the three highest-error zones, with annual RMSE improvements over ConvLSTM of 19.4-26.5%, so graph-based models capture at least part of the mesoscale dynamics that smooth convolutional models blur.
  • The same architecture, retrained, is a template for extending data-driven SST forecasting to other variables such as salinity and currents, and to other regional ocean domains.
  • The reported sensitivity to initial-condition errors and the triangular artifacts at long lead times argue for hybrid designs that combine graph spatial precision with convolutional stabilization, as the paper itself proposes.

Reading between the lines

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

  • An implication the paper leaves implicit is that its headline skill numbers are measured against the same L4 satellite product used for training; without independent in-situ validation, part of the reported advantage over GLORYS may reflect the model learning the analysis product's own smoothing rather than physical forecast skill.
  • The triangular artifacts and steadily rising overactivity at longer lead times suggest the decoder's mesh-to-grid mapping, not the encoder or processor, is the main bottleneck; an attention-based or convolution-blended decoder is a direct, testable next step.
  • The same recipe, regional curvilinear mesh plus a land/ocean-masked loss, should transfer to other eastern boundary upwelling systems; testing in, say, the Benguela or Humboldt systems would show whether the advantage is generic or specific to the Canary region.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

6 major / 6 minor

Summary. This manuscript adapts the GraphCast graph neural network, originally developed for global weather prediction, to produce daily sea surface temperature (SST) forecasts at 0.05° resolution over the Canary Current Upwelling System (CCUS) out to 20 days. The model is trained on CMEMS L4 reprocessed satellite SST (1982–2020) and evaluated on 2017–2020 against the same L4 product, a ConvLSTM baseline, and the GLORYS12V1 reanalysis and PSY4V3R1 forecast system. The headline claims are a 74–76% RMSE reduction over GLORYS at 5-day lead time, 44–51% at 10 days, up to 26.5% over ConvLSTM, and roughly 100× faster inference. The paper also examines interannual, seasonal, and spatial error patterns and discusses model limitations, including sensitivity to initial condition errors and triangular mesh artifacts.

Significance. If the reported skill held up under independent verification, the paper would provide a useful demonstration that global weather GNNs can be adapted to regional ocean forecasting with large speed gains, making it a relevant contribution to the emerging ML ocean-prediction literature. The authors are to be credited for releasing the source code, for comparing against a traditional numerical product, and for candidly discussing known failure modes such as initial-condition sensitivity and mesh artifacts. However, the evaluation as presented does not yet support the central claim that the model 'surpasses traditional methods' for the real ocean state: the verification is performed against the same L4 analysis used for training, the comparison to GLORYS is a forecast-versus-reanalysis rather than forecast-versus-forecast, the reported headline reductions are numerically inconsistent across sections, and the sub-20 km resolution claim is unsupported by the presented metrics.

major comments (6)
  1. [Appendix C.2, Eq. (17)] The RMSE is defined as sqrt((\hat{x}_t^i - x_t^i)^2), which is a pointwise absolute error, not a root-mean-square over the verification sample. As written, the formula omits the spatial and temporal averaging that the accompanying text ('average magnitude') and all subsequent results require. Please correct the definition to include the averaging operator used in the implementation (e.g., over grid cells and forecast realizations), and recompute the reported numbers if the implemented metric differed.
  2. [Sections 2.2, 4.1, Appendix A.1] The model is trained and verified on the same CMEMS L4 SST product. Because L4 is an analyzed, gap-filled, smoothed product built by merging and intercalibrating multiple satellite sensors, a model trained on L4 can learn to replicate its biases and smoothness, so low RMSE against L4 partially measures self-consistency. The comparison to GLORYS and PSY4V3R1 provides some independent signal, but both are numerical products that assimilate some of the same satellite SST, and no validation against in-situ observations (drifters, moorings, Argo) is provided. Please add an independent validation or explicitly restrict the claims to skill relative to the L4 product.
  3. [Abstract, Sections 4.2–4.4] The 5-day RMSE reduction relative to GLORYS is reported as 76% in the abstract, 75.5% in Section 4.2, 74.2% in Section 4.4, and up to 77.4% in the seasonal analysis of Section 4.3. These numbers are not reconciled; the manuscript does not state whether they refer to different spatial domains (full domain vs. selected capes), different reference periods, or different averaging procedures. Please clarify and ensure a single, reproducible number is used for the headline claim.
  4. [Section 4.2, Figure 5] The comparison of GraphCast 5-day forecasts against the GLORYS reanalysis is a forecast-versus-reanalysis comparison: GLORYS provides a smoothed, data-assimilated estimate of the past ocean state, not a 5-day forecast. This baseline mismatch can inflate the apparent skill improvement, because the reanalysis is not penalized by forecast lead time. A fair assessment of forecast skill would compare against the PSY4V3R1 forecast at the same lead times (currently only shown in Figure 3), or against persistence and climatology baselines.
  5. [Section 4.4 and Abstract] The headline improvements at Cape Ghir, Cape Bojador, and Cape Blanc are based on selecting the highest-error locations after inspecting the RMSE maps. The domain-average reduction (74.2%) is lower than the cape-specific reductions (69.7–78.6%), so reporting 'up to 76%' from post hoc selected points overstates the overall skill. Please state whether these capes were pre-registered as hypotheses, or correct for multiple comparisons, or report the domain-average as the primary metric.
  6. [Section 6, Section 5] The claim that the model resolves 'filaments and eddies below 20 km in scale' is not supported by any analysis in the paper. The evaluation uses spatially averaged RMSE, ACC, bias, and activity; no spectral analysis, feature-tracking, or independent high-resolution comparison is presented, and the L4 training data is itself a smoothed analysis. Please remove the claim or provide direct evidence of resolved sub-20 km structures.
minor comments (6)
  1. [Section 2.1] The study region is described as extending 'from 21◦S to 33◦N' but the domain is in the North Atlantic; this should read '21°N to 33°N'.
  2. [Section 5] The acronym CCUS is used for the Canary Current Upwelling System but is expanded as 'California Current Upwelling System' in the discussion; please correct to 'Canary Current Upwelling System'.
  3. [Section 3.4] The text describes replacing the icosahedral mesh with a 'square curvilinear mesh' but then refers to 'triangular elements' and 'triangular face' in the decoder; please clarify the mesh geometry used.
  4. [Figure 4] The bottom panel y-axis label 'Err(t)/ t [C/day]' and the '1e 16' annotation appear garbled; please check the rendering.
  5. [Table 1] The column headers 'N GraphCast% ConvLSTM%' are ambiguous; please clarify that N is the number of RMSE values and specify the units of the percentages.
  6. [Appendix A.2 and A.4] The hyperparameter search lists mesh refinement levels {2,4,6} but the final model configuration uses M_3; please state whether M_3 was selected from a different set or how it relates to the searched values.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the same L4 SST product is used for training and verification, which is a validity limitation rather than a derivation collapse.

full rationale

The paper's quantitative chain is empirical rather than definitional: an ocean-adapted GraphCast is trained on CMEMS L4 SST (Appendix A.1) and its skill is measured by RMSE, ACC, RA, and bias against the same L4 product (Sections 3, 4.1, and Eq. 17). This does not make any reported quantity equal to an input by construction: the test years (2017-2020) are out-of-sample, forecasts are autoregressive rollouts rather than fitted values, and the headline comparisons are made against independent benchmarks (GLORYS, PSY4V3R1, ConvLSTM) that were not trained on the GraphCast objective. The same-L4 training/verification choice is a genuine limitation for physical validity, since L4 is an analyzed, gap-filled product and no in-situ validation is provided; the paper itself acknowledges that the L4 resolution cannot capture submesoscale features (Section 5). However, this is an external-validity or correctness concern, not a circular derivation. The few self-citations (e.g., Estrada-Allis et al. 2023, Sangrà et al. 2009, 2015) are used only for oceanographic context and are not load-bearing for the forecasting claim. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' own prior work, and no ansatz is smuggled in via citation. Therefore, no significant circularity is present.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central evaluation rests on the L4 SST product and the GraphCast architecture; no new physical entities are introduced. Hyperparameters were selected on a validation set. The key domain assumptions are that L4 SST is accurate enough as ground truth and that SST alone suffices to forecast SST in the CCUS region.

free parameters (6)
  • number of message-passing steps = 6
    Selected via grid search among {2,6,7}; affects model expressiveness and cost.
  • latent size (MLP hidden dimension) = 8
    Selected via grid search among {2,4,8,16}; much smaller than GraphCast's 512 due to single-variable output.
  • mesh refinement level (M_r) = 3
    Selected via grid search among {2,4,6}; higher levels add resolution at greater cost.
  • neighborhood radius multiplier = 0.6
    Controls encoder connectivity from grid to mesh nodes; chosen by hand and not optimized.
  • batch size = 8
    Preliminary experiments indicated negligible impact among tested values.
  • learning rate schedule = 1e-3 to 0 (half-cosine)
    Standard choice from Lam et al. 2023, not tuned.
assumptions (5)
  • domain assumption GraphCast architecture (Lam et al. 2023) provides a suitable inductive bias for regional SST forecasting.
    Invoked throughout Section 3.4; the model is a direct adaptation of the global weather GNN.
  • domain assumption The L4 SST product is an accurate gap-free representation of true SST, suitable as both training target and verification reference.
    Section 2.2 describes the product; all metrics in Section 4 verify against this same product.
  • domain assumption SST evolution in the CCUS is predictable from two previous SST snapshots plus time-of-day/year features, without wind or bathymetric forcing.
    Model input (Appendix B, Eq. for v_g) includes only SST, time features, and a land-sea mask; no atmospheric or bathymetric variables.
  • domain assumption No temporal leakage between train (1982-2012), validation (2013-2016), and test (2017-2020) despite daily overlapping windows.
    Appendix A.1 describes year-based split; assumed to prevent the model from seeing test-period data during training.
  • ad hoc to paper L4 satellite instrumental error of ±0.25°C is a meaningful threshold for evaluating forecast skill.
    Used in Sections 4.1-4.3 to define 'days exceeding the threshold'; not a statistical significance bound.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep Learning Weather Models for Subregional Ocean Forecasting: A Case Study on the Canary Current Upwelling System." pith.science (2026). https://pith.science/paper/LRCJXCWW

@misc{pith2026250524429,
  author       = {Pith},
  title        = {Pith review of: Deep Learning Weather Models for Subregional Ocean Forecasting: A Case Study on the Canary Current Upwelling System},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LRCJXCWW}},
  note         = {Machine review of arXiv:2505.24429}
}
read the original abstract

Oceanographic forecasting impacts various sectors of society by supporting environmental conservation and economic activities. Based on global circulation models, traditional forecasting methods are computationally expensive and slow, limiting their ability to provide rapid forecasts. Recent advances in deep learning offer faster and more accurate predictions, although these data-driven models are often trained with global data from numerical simulations, which may not reflect reality. The emergence of such models presents great potential for improving ocean prediction at a subregional domain. However, their ability to predict fine-scale ocean processes, like mesoscale structures, remains largely unknown. This work aims to adapt a graph neural network initially developed for global weather forecasting to improve subregional ocean prediction, specifically focusing on the Canary Current upwelling system. The model is trained with satellite data and compared to state-of-the-art physical ocean models to assess its performance in capturing ocean dynamics. Our results show that the deep learning model surpasses traditional methods in precision despite some challenges in upwelling areas. It demonstrated superior performance in reducing RMSE errors compared to ConvLSTM and the GLORYS reanalysis, particularly in regions with complex oceanic dynamics such as Cape Ghir, Cape Bojador, and Cape Blanc. The model achieved improvements of up to 26.5% relative to ConvLSTM and error reductions of up to 76% in 5-day forecasts compared to the GLORYS reanalysis at these critical locations, highlighting its enhanced capability to capture spatial variability and improve predictive accuracy in complex areas. These findings suggest the viability of adapting meteorological data-driven models for improving subregional medium-term ocean forecasting.

Figures

Figures reproduced from arXiv: 2505.24429 by the authors.

Figure 1
Figure 1. Summer (JJA) climatology of sea surface temperature ( [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Representation of a multi-mesh with a refinement factor of [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Performance evaluation of the models for SST, verified against satellite L4 observations. The plots display four [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Predictability barrier plots of daily RMSE for SST forecasts at lead times of 1–20 days from 2017 to 2020, [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Time series of RMSE for SST predictions over the test period. RMSE values from GraphCast forecasts at [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Seasonal predictability barrier plots showing daily RMSE for SST forecasts at lead times of 1–20 days, [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: Point-wise RMSE average for 20 lead times forecasts of the GraphCast model. Each map represents the [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Point-wise RMSE average for 20 lead times forecasts of the ConvLSTM model. Each map represents the [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

80 extracted references · 66 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    P., Samuels, B

    Adcroft, A., Hallberg, R., Dunne, J. P., Samuels, B. L., Galt, J. A., Barker, C. H., and Payton, D. (2010). Simulations of underwater plumes of dissolved oil in the Gulf of Mexico . Geophysical Research Letters , 37(18)

    work page 2010

  3. [3]

    D., Álvarez Salgado, X

    Arístegui, J., Barton, E. D., Álvarez Salgado, X. A., Santos, A. M. P., Figueiras, F. G., Kifani, S., Hernández-León, S., Mason, E., Machú, E., and Demarcq, H. (2009). Sub-regional ecosystem variability in the Canary Current upwelling . Progress in Oceanography , 83(1):33--48. Eastern Boundary Upwelling Ecosystems: Integrative and Comparative Approaches

    work page 2009

  4. [4]

    Arístegui, J., Sangrá, P., Hernández-León, S., Cantón, M., Hernández-Guerra, A., and Kerling, J. (1994). Island-induced eddies in the Canary Islands . Deep Sea Research Part I: Oceanographic Research Papers , 41(10):1509--1525

    work page 1994

  5. [5]

    Barton, E., Arístegui, J., Tett, P., Cantón, M., García-Braun, J., Hernández-León, S., Nykjaer, L., Almeida, C., Almunia, J., Ballesteros, S., Basterretxea, G., Escánez, J., García-Weill, L., Hernández-Guerra, A., López-Laatzen, F., Molina, R., Montero, M., Navarro-Pérez, E., Rodríguez, J., van Lenning , K., Vélez, H., and Wild, K. (1998). The transition ...

    work page 1998

  6. [6]

    W., Hamrick , J

    Battaglia , P. W., Hamrick , J. B., Bapst , V., Sánchez-González , A., Zambaldi , V., Malinowski , M., Tacchetti , A., Raposo , D., Santoro , A., Faulkner , R., Gulcehre , C., Song , F., Ballard , A., Gilmer , J., Dahl , G., Vaswani , A., Allen , K., Nash , C., Langston , V., Dyer , C., Heess , N., Wierstra , D., Kohli , P., Botvinick , M., Vinyals , O., ...

    arXiv 2018

  7. [7]

    W., Pascanu, R., Lai, M., Rezende, D., and Kavukcuoglu, K

    Battaglia, P. W., Pascanu, R., Lai, M., Rezende, D., and Kavukcuoglu, K. (2016). Interaction Networks for learning about objects, relations and physics. Technical report, Cornell University

    work page 2016

  8. [8]

    Bechtold, P., K \"o hler, M., Jung, T., Leutbecher, M., Rodwell, M., and Vitart, F. (2008). Advances in simulating atmospheric variability with IFS cycle 32r3. ECMWF Newsletter No. 114 - Winter 2007/08 , 114:29--38

    work page 2008

Show all 80 references
  1. [9]

    J., Lefèbvre, M., Traon, P.-Y

    Bell, M. J., Lefèbvre, M., Traon, P.-Y. L., Smith, N., and Wilmer-Becker, K. (2009). GODAE : the global ocean data assimilation experiment. Oceanography , 22(3):14--21

    2009

  2. [10]

    Bengio, Y., Simard, P., and Frasconi, P. (1994). Learning long-term dependencies with gradient descent is difficult. IEEE Transactions on Neural Networks , 5(2):157--166

    1994

  3. [11]

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

    2023

  4. [12]

    L., Cheung, W

    Bindoff, N. L., Cheung, W. W. L., Kairo, J. G., Arístegui, J., Guinder, V. A., Hallberg, R., Hilmi, N., Jiao, N., Karim, M. S., Levin, L., O'Donoghue, S., Cuicapusa, S. R. P., Rinkevich, B., Suga, T., Tagliabue, A., and Williamson, P. (2019). Changing ocean, marine ecosystems,...

    2019

  5. [13]

    P., Lucic, A., Stanley, M., Brandstetter, J., Garvan, P., Riechert, M., Weyn, J., Dong, H., Vaughan, A., Gupta, J

    Bodnar, C., Bruinsma, W. P., Lucic, A., Stanley, M., Brandstetter, J., Garvan, P., Riechert, M., Weyn, J., Dong, H., Vaughan, A., Gupta, J. K., Tambiratnam, K., Archibald, A., Heider, E., Welling, M., Turner, R. E., and Perdikaris, P. (2024). A foundation model of the atmosphe...

    2024

  6. [14]

    B., Clare, M

    Bouallègue, Z. B., Clare, M. C. A., Magnusson, L., Gascón, E., Maier-Gerber, M., Janoušek, M., Rodwell, M., Pinault, F., Dramsch, J. S., Lang, S. T. K., Raoult, B., Rabier, F., Chevallier, M., Sandu, I., Dueben, P., Chantry, M., and Pappenberger, F. (2024). The rise of data-dr...

    2024

  7. [15]

    and Verron, J

    Brasseur, P. and Verron, J. (2006). The SEEK filter method for data assimilation in oceanography: a synthesis. Ocean Dynamics , 56:650--661

    2006

  8. [16]

    Cabanes, C., Grouazel, A., von Schuckmann, K., Hamon, M., Turpin, V., Coatanoan, C., Paris, F., Guinehut, S., Boone, C., Ferry, N., de Boyer Mont\'egut, C., Carval, T., Reverdin, G., Pouliquen, S., and Le Traon, P.-Y. (2013). The CORA dataset: validation and diagnostics of in-...

    2013

  9. [17]

    P., Cazenave, A., Champollion, N., Dieng, H., Llovel, W., Forsberg, R., von Schuckmann, K., and Wada, Y

    Chambers, D. P., Cazenave, A., Champollion, N., Dieng, H., Llovel, W., Forsberg, R., von Schuckmann, K., and Wada, Y. (2017). Evaluation of the global mean sea level budget between 1993 and 2014. Integrative study of the mean sea level and its components , pages 315--333

    2017

  10. [18]

    B., and He, R

    Chattopadhyay, A., Gray, M., Wu, T., Lowe, A. B., and He, R. (2024). OceanNet : a principled neural operator-based digital twin for regional oceans. Scientific Reports , 14(1):21181

    2024

  11. [19]

    Chattopadhyay, A., Mustafa, M., Hassanzadeh, P., and Kashinath, K. (2021). Deep spatial transformers for autoregressive data-driven forecasting of geophysical turbulence. In Proceedings of the 10th International Conference on Climate Informatics , CI2020, page 106–112, New Yor...

    2021

  12. [20]

    Q., and Hassanzadeh , P

    Chattopadhyay , A., Sun , Y. Q., and Hassanzadeh , P. (2023). Challenges of learning multi-scale dynamics with AI weather models: Implications for stability and one solution . arXiv e-prints , page arXiv:2304.07029

    2023 arXiv

  13. [21]

    European North West Shelf/Iberia Biscay Irish seas - high resolution l4 sea surface temperature reprocessed

    CMEMS (2024). European North West Shelf/Iberia Biscay Irish seas - high resolution l4 sea surface temperature reprocessed

    2024

  14. [22]

    Couairon, G., Lessig, C., Charantonis, A., and Monteleoni, C. (2024). ArchesWeather : An efficient AI weather forecasting model at 1.5 ^o resolution. Technical Report 2405.14527, Cornell University

    2024 arXiv

  15. [23]

    E., Hanna, E., and Bigg, G

    Cropper, T. E., Hanna, E., and Bigg, G. R. (2014). Spatial and temporal seasonal trends in coastal upwelling off Northwest Africa , 1981–2012. Deep Sea Research Part I: Oceanographic Research Papers , 86:94--111

    2014

  16. [24]

    Cushman-Roisin, B., Tang, B., and Chassignet, E. P. (1990). Westward motion of mesoscale eddies. Journal of Physical Oceanography , 20(5):758 -- 768

    1990

  17. [25]

    P., Uppala, S

    Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, ...

    2011

  18. [26]

    P., Roberts, M

    Docquier, D., Grist, J. P., Roberts, M. J., Roberts, C. D., Semmler, T., Ponsoni, L., Massonnet, F., Sidorenko, D., Sein, D. V., Iovino, D., et al. (2019). Impact of model resolution on Arctic sea ice and North Atlantic ocean heat transport. Climate Dynamics , 53:4989--5017

    2019

  19. [27]

    Dueben, P. D. and Bauer, P. (2018). Challenges and design choices for global weather and climate models based on machine learning. Geoscientific Model Development , 11(10):3999--4009

    2018

  20. [28]

    N., Rodríguez-Santana, \'A ., Naveira-Garabato, A

    Estrada-Allis, S. N., Rodríguez-Santana, \'A ., Naveira-Garabato, A. C., García-Weil, L., Arcos-Pulido, M., and Emelianov, M. (2023). Enhancement of turbulence and nutrient fluxes within an Eastern Boundary Upwelling Filament : a diapycnal entrainment approach. Frontiers in Ma...

    2023

  21. [29]

    Ezraty, R., Girard-Ardhuin, F., Pioll \'e , J.-F., Kaleschke, L., and Heygster, G. (2007). Arctic and Antarctic sea ice concentration and arctic sea ice drift estimated from special sensor microwave data. D \'e partement d’Oc \'e anographie Physique et Spatiale, IFREMER, Brest...

    2007

  22. [30]

    G., Ziemann, D., Kolber, Z., and Bienfang, P

    Falkowski, P. G., Ziemann, D., Kolber, Z., and Bienfang, P. K. (1991). Role of eddy pumping in enhancing primary production in the ocean. Nature , 352(6330):55--58

    1991

  23. [31]

    L., Chune, S

    Galloudec, O. L., Chune, S. L., Nouel, L., Fernandez, E., Derval, C., Tressol, M., Dussurget, R., Biardeau, A., and Tonani, M. (2024). Product User Manual for Global Ocean Physics Analysis and Forecasting Product . Copernicus Marine Environment Monitoring Service (CMEMS). CMEM...

    2024

  24. [32]

    Geer, A. J. (2016). Significance of changes in medium-range forecast scores. Tellus A: Dynamic Meteorology and Oceanography , 68(1):30229

    2016

  25. [33]

    and Schmidhuber, J

    Gers, F. and Schmidhuber, J. (2000). Recurrent nets that time and count. In Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium , volume 3, pages 189--194 vol.3

    2000

  26. [34]

    Gers, F., Schmidhuber, J., and Cummins, F. (1999). Learning to forget: continual prediction with LSTM . In 1999 Ninth International Conference on Artificial Neural Networks ICANN 99. (Conf. Publ. No. 470) , volume 2, pages 850--855 vol.2

    1999

  27. [35]

    K., Koutník, J., Steunebrink, B

    Greff, K., Srivastava, R. K., Koutník, J., Steunebrink, B. R., and Schmidhuber, J. (2017). LSTM : A search space odyssey. IEEE Transactions on Neural Networks and Learning Systems , 28(10):2222--2232

    2017

  28. [36]

    Hagen, E., Zulicke, C., and Feistel, R. (1996). Near-surface structures in the Cape Ghir filament off Morocco . Oceanologica Acta , 19(6):577--598

    1996

  29. [37]

    and Czaja, A

    Hausmann, U. and Czaja, A. (2012). The observed signature of mesoscale eddies in sea surface temperature and the associated heat transport. Deep Sea Research Part I: Oceanographic Research Papers , 70:60--72

    2012

  30. [38]

    and Schmidhuber, J

    Hochreiter, S. and Schmidhuber, J. (1997). Long Short-Term Memory . Neural Computation , 9(8):1735--1780

    1997

  31. [39]

    Holmberg , D., Clementi , E., and Roos , T. (2024). Regional Ocean Forecasting with Hierarchical Graph Neural Networks . arXiv e-prints , page arXiv:2410.11807

    2024 arXiv

  32. [40]

    and Hamman, J

    Hoyer, S. and Hamman, J. (2017). xarray: N-D labeled arrays and datasets in Python . Journal of Open Research Software , 5(1)

    2017

  33. [41]

    G., Charly, R., Tony, C., Charles-Emmanuel, T., Florent, G., Giovanni, R., Mounir, B., Yann, D., and Pierre-Yves, L

    Jean-Michel, L., Eric, G., Romain, B.-B., Gilles, G., Angélique, M., Marie, D., Clément, B., Mathieu, H., Olivier, L. G., Charly, R., Tony, C., Charles-Emmanuel, T., Florent, G., Giovanni, R., Mounir, B., Yann, D., and Pierre-Yves, L. T. (2021). The Copernicus global 1/12^ oce...

    2021

  34. [42]

    Keisler , R. (2022). Forecasting Global Weather with Graph Neural Networks . arXiv e-prints , page arXiv:2202.07575

    2022 arXiv

  35. [43]

    o wer, M., Lottes, J., Rasp, S., D \

    Kochkov, D., Yuval, J., Langmore, I., Norgaard, P. C., Smith, J. A., Mooers, G., Kl \"o wer, M., Lottes, J., Rasp, S., D \"u ben, P. D., Hatfield, S., Battaglia, P. W., Sánchez-González, A., Willson, M., Brenner, M. P., and Hoyer, S. (2023). Neural general circulation models f...

    2023

  36. [44]

    Lam, R., Sánchez-González, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W., Merose, A., Hoyer, S., Holland, G., Vinyals, O., Stott, J., Pritzel, A., Mohamed, S., and Battaglia, P. (2023). Learning skillful medium-range...

    2023

  37. [45]

    Lang , S., Alexe , M., Chantry , M., Dramsch , J., Pinault , F., Raoult , B., Clare , M. C. A., Lessig , C., Maier-Gerber , M., Magnusson , L., Ben Bouall \`e gue , Z., Prieto Nemesio , A., Dueben , P. D., Brown , A., Pappenberger , F., and Rabier , F. (2024). AIFS -- ECMWF 's...

    2024 arXiv

  38. [46]

    Lellouche, J.-M., Greiner, E., Le Galloudec, O., Garric, G., Regnier, C., Drevillon, M., Benkiran, M., Testut, C.-E., Bourdalle-Badie, R., Gasparin, F., Hernandez, O., Levier, B., Drillet, Y., Remy, E., and Le Traon, P.-Y. (2018). Recent updates to the Copernicus Marine Servic...

    2018

  39. [47]

    Lellouche, J.-M., Le Galloudec, O., Dr\'evillon, M., R\'egnier, C., Greiner, E., Garric, G., Ferry, N., Desportes, C., Testut, C.-E., Bricaud, C., Bourdall\'e-Badie, R., Tranchant, B., Benkiran, M., Drillet, Y., Daudin, A., and De Nicola, C. (2013). Evaluation of global monito...

    2013

  40. [48]

    Li , S., Zhao , Y., Varma , R., Salpekar , O., Noordhuis , P., Li , T., Paszke , A., Smith , J., Vaughan , B., Damania , P., and Chintala , S. (2020). PyTorch Distributed: Experiences on Accelerating Data Parallel Training . arXiv e-prints , page arXiv:2006.15704

    2020 arXiv

  41. [49]

    Liu, Y., MacFadyen, A., Ji, Z.-G., and Weisberg, R. H. (2013). Monitoring and modeling the deepwater horizon oil spill: a record breaking enterprise . John Wiley & Sons

    2013

  42. [50]

    Madec, G., Bell, M., Benshila, R., Blaker, A., Boudrallé-Badie, R., Bricaud, C., Bruciaferri, D., Carneiro, D., Castrillo, M., Calvert, D., Chanut, J., Clementi, E., Coward, A., de Lavergne, C., Dobricic, S., Epicoco, I., Éthé, C., Fiedler, E., Ford, D., Furner, R., Ganderton,...

    2024

  43. [51]

    Madec, G. et al. (2008). NEMO ocean engine. note du P \^o le de mod \'e lisation. Institut Pierre-Simon Laplace (IPSL), France , 27:1288--1619

    2008

  44. [52]

    Mourre, B., Aguiar, E., Juz \`a , M., Hern \'a ndez-Lasheras, J., Reyes, E., Heslop, E., Escudier, R., Cutolo, E., Ruiz, S., Mason, E., Pascual, A., and Tintor \'e , J. (2018). Assessment of high-resolution regional ocean prediction systems using multi-platform observations: I...

    2018

  45. [53]

    Mu, B., Peng, C., Yuan, S., and Chen, L. (2019). ENSO forecasting over multiple time horizons using ConvLSTM network and rolling mechanism. In 2019 International Joint Conference on Neural Networks (IJCNN) , pages 1--8

    2019

  46. [54]

    Mu, B., Qin, B., and Yuan, S. (2021). ENSO-ASC 1.0.0: ENSO deep learning forecast model with a multivariate air--sea coupler. Geoscientific Model Development , 14(11):6977--6999

    2021

  47. [55]

    Oskarsson , J., Landelius , T., and Lindsten , F. (2023). Graph-based Neural Weather Prediction for Limited Area Modeling . arXiv e-prints , page arXiv:2309.17370

    2023 arXiv

  48. [56]

    Pelegrí, J., Arístegui, J., Cana, L., González-Dávila, M., Hernández-Guerra, A., Hernández-León, S., Marrero-Díaz, A., Montero, M., Sangrà, P., and Santana-Casiano, M. (2005). Coupling between the open ocean and the coastal upwelling region off Northwest Africa : water recircu...

    2005

  49. [57]

    Pfaff , T., Fortunato , M., Sánchez-González , A., and Battaglia , P. W. (2020). Learning Mesh-Based Simulation with Graph Networks . arXiv e-prints , page arXiv:2010.03409

    2020 arXiv

  50. [58]

    and Autret, E

    Pioll \'e , J.-F. and Autret, E. (2023). QUID for SST TAC Products SST \_ GLO \_ SST \_ L3S \_ NRT \_ OBSERVATIONS \_010\_010. Quality Information Document 2.1, E.U. Copernicus Marine Service Information (CMEMS)

    2023

  51. [59]

    R., El-Kadi, A., Masters, D., Ewalds, T., Stott, J., Mohamed, S., Battaglia, P

    Price, I., Sánchez-González, A., Alet, F., Andersson, T. R., El-Kadi, A., Masters, D., Ewalds, T., Stott, J., Mohamed, S., Battaglia, P. W., Lam, R., and Willson, M. (2024). Probabilistic weather forecasting with machine learning. Nature , 637:84 -- 90

    2024

  52. [60]

    Pujol, M.-I., Faug\`ere, Y., Taburet, G., Dupuy, S., Pelloquin, C., Ablain, M., and Picot, N. (2016). DUACS DT2014 : the new multi-mission altimeter data set reprocessed over 20 years. Ocean Science , 12(5):1067--1090

    2016

  53. [61]

    D., Scher, S., Weyn, J

    Rasp, S., Dueben, P. D., Scher, S., Weyn, J. A., Mouatadid, S., and Thuerey, N. (2020). WeatherBench : a benchmark data set for data-driven weather forecasting. Journal of Advances in Modeling Earth Systems , 12(11):e2020MS002203

    2020

  54. [62]

    B., Dueben, P., Bromberg, C., Sisk, J., Barrington, L., Bell, A., and Sha, F

    Rasp, S., Hoyer, S., Merose, A., Langmore, I., Battaglia, P., Russel, T., Sánchez-González, A., Yang, V., Carver, R., Agrawal, S., Chantry, M., Bouallegue, Z. B., Dueben, P., Bromberg, C., Sisk, J., Barrington, L., Bell, A., and Sha, F. (2024). WeatherBench 2: A benchmark for ...

    2024 arXiv

  55. [63]

    H., Guinehut, S., and Larnicol, G

    Rio, M. H., Guinehut, S., and Larnicol, G. (2011). New CNES-CLS09 global mean dynamic topography computed from the combination of GRACE data, altimetry, and in situ measurements. Journal of Geophysical Research: Oceans , 116(C7)

    2011

  56. [64]

    Roberts, M., Vidale, P., Senior, C., Hewitt, H., Bates, C., Berthou, S., Chang, P., Christensen, H., Danilov, S., Demory, M.-E., et al. (2018). The benefits of global high resolution for climate simulation: process understanding and the enabling of stakeholder decisions at the...

    2018

  57. [65]

    C., Pelegrí, J

    Sangrà, P., Pascual, A., Ángel Rodríguez-Santana, Machín, F., Mason, E., McWilliams, J. C., Pelegrí, J. L., Dong, C., Rubio, A., Arístegui, J., Ángeles Marrero-Díaz, Hernández-Guerra, A., Martínez-Marrero, A., and Auladell, M. (2009). The Canary Eddy Corridor : A major pathway...

    2009

  58. [66]

    Sangrà, P., Troupin, C., Barreiro-González, B., Desmond Barton, E., Orbi, A., and Arístegui, J. (2015). The cape G hir filament system in A ugust 2009 ( NW A frica). Journal of Geophysical Research: Oceans , 120(6):4516--4533

    2015

  59. [67]

    Scher, S. (2018). Toward data-driven weather and climate forecasting: Approximating a simple general circulation model with deep learning. Geophysical Research Letters , 45(22):12,616--12,622

    2018

  60. [68]

    Shi , X., Chen , Z., Wang , H., Yeung , D.-Y., Wong , W.-k., and Woo , W.-c. (2015). Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting . arXiv e-prints , page arXiv:1506.04214

    2015 arXiv

  61. [69]

    L., Chassignet, E

    Sommer, J. L., Chassignet, E. P., and Wallcraft, A. J. (2018). Ocean circulation modeling for operational oceanography: Current status and future challenges. New Frontiers in Operational Oceanography

    2018

  62. [70]

    Szekely, T., Gourrion, J., Pouliquen, S., Reverdin, G., and Merceur, F. (2019). CORA , coriolis ocean dataset for reanalysis. Sea Scientific Open Data Publication

    2019

  63. [71]

    Thompson, B., Sánchez, C., Heng, B. C. P., Kumar, R., Liu, J., Huang, X.-Y., and Tkalich, P. (2021). Development of a MetUM (v 11.1) and NEMO (v 3.6) coupled operational forecast model for the maritime continent -- Part 1: Evaluation of ocean forecasts. Geoscientific Model Dev...

    2021

  64. [72]

    Thorpe, A., Bauer, P., Magnusson, L., and Richardson, D. (2013). An evaluation of recent performance of ECMWF ’s forecasts. ECMWF Newsletter No. 137 - autumn 2013 , pages 15--18

    2013

  65. [73]

    weather of the ocean

    Treguier, A.-M., Chassignet, E. p., Le Boyer, A., and Pinardi, N. (2017). Modeling and forecasting the "weather of the ocean" at the mesoscale. Journal Of Marine Research , 75(3):301--329

    2017

  66. [74]

    Troupin, C., Mason, E., Beckers, J.-M., and Sangrà, P. (2012). Generation of the Cape Ghir upwelling filament: A numerical study. Ocean Modelling , 41:1--15

    2012

  67. [75]

    Wang, X., Wang, R., Hu, N., Wang, P., Huo, P., Wang, G., Wang, H., Wang, S., Zhu, J., Xu, J., Yin, J., Bao, S., Luo, C., Zu, Z., Han, Y., Zhang, W., Ren, K., Deng, K., and Song, J. (2024). XiHe : A Data-Driven Model for Global Ocean Eddy-Resolving Forecasting . Technical Repor...

    2024 arXiv

  68. [76]

    Watters , N., Tacchetti , A., Weber , T., Pascanu , R., Battaglia , P., and Zoran , D. (2017). Visual Interaction Networks . arXiv e-prints , page arXiv:1706.01433

    2017 arXiv

  69. [77]

    S., Bakun, A., and McLain, D

    Wooster, W. S., Bakun, A., and McLain, D. R. (1976). The seasonal upwelling cycle along the eastern boundary of the N orth A tlantic. Journal of Marine Research , 34(2):131--141

    1976

  70. [78]

    Yang, J., Zhang, T., Zhang, J., Lin, X., Wang, H., and Feng, T. (2024). A ConvLSTM nearshore water level prediction model with integrated attention mechanism. Frontiers in Marine Science , 11

    2024

  71. [79]

    Yang, X., Zhang, F., Sun, P., Li, X., Du, Z., and Liu, R. (2022). A spatio-temporal graph-guided convolutional LSTM for tropical cyclones precipitation nowcasting. Applied Soft Computing , 124:109003

    2022

  72. [80]

    Zeng, X., Li, Y., and He, R. (2015). Predictability of the loop current variation and eddy shedding process in the Gulf of Mexico using an artificial neural network approach. Journal of Atmospheric and Oceanic Technology , 32(5):1098 -- 1111

    2015

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.