Pith. sign in

REVIEW 3 major objections 6 minor 31 references

DLESyM-Ocean: A Deep Learning Probabilistic Global Model for Simulating Present-Day Upper Ocean and Sea Ice

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A compact neural network trained on reanalysis can emulate the global upper ocean and sea ice probabilistically over multi-year autoregressive simulations, reproducing recent extremes and remaining stable.

desk verdict Solid within-subfield advance that overstates calibration in the abstract and misses the 2023 El Niño amplitude, but the core multi-year stability and skill results are credible and worth refereeing. read the letter →

arxiv 2608.11545 v1 pith:MET6G64A submitted 2026-08-12 physics.ao-ph

Zachary I Espinosa , Nathaniel Cresswell-Clay , William Yik , Cecilia M. Bitz , Edward Blanchard-Wrigglesworth , Peter Harrington , David Pruitt , Michael S. Pritchard
show 1 more author
Dale R. Durran
This is my paper · ORCID
classification physics.ao-ph
keywords deeplearningEarthsystemmodelseaiceupperoceanprobabilisticensembleforecastingpatchenergyscoreHEALPixautoregressivestability
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

DLESyM-Ocean is a compact neural network that aims to act as a probabilistic emulator for the global upper ocean and sea ice: given the current and previous ocean state and three atmospheric forcing fields, it predicts states four and eight days ahead, and can be looped autoregressively for years. The paper's central claim is that this trained model produces well-calibrated, spatially coherent ensembles with less error than persistence or ocean climatology at 90-day leads, and that it reproduces the 2023 El Niño transition, the 2019 northeast Pacific marine heatwave, and the 2023 Antarctic sea ice record while remaining stable over five- and twenty-nine-year integrations. If true, the model would supply a fast, probabilistic ocean–sea ice component that could be coupled to AI atmosphere models for subseasonal-to-seasonal forecasting and for generating counterfactual ocean states.

What carries the argument

The load-bearing mechanism is the almost-fair patch energy score (afPES), a proper scoring rule that generalizes CRPS to multivariate joint distributions by evaluating an energy score over localized 3×3 spatial patches on a HEALPix sphere. The patch restriction avoids the high-dimensional distance concentration that makes global energy scores unstable, while the 'almost-fair' parameterization interpolates between the biased empirical estimator and the exactly fair estimator, whose zero-gradient extremes would otherwise detach the most extreme ensemble members from training. Ensemble members are produced by injecting a 32-dimensional Gaussian noise vector into conditional layer norms (CLNs) that apply channel-wise scale and shift; the network learns residuals on top of a global skip connection, with hard clipping to keep sea ice concentration and thickness in physical ranges. Custom isolatitude padding and a smoother-based upsampling scheme remove the face-seam and checkerboard artifacts that the authors identify in prior HEALPix U-Nets.

What would settle it

Take the 2023 El Niño case and rerun the 50-member ensemble with the outgoing longwave radiation forcing replaced by its 1994–2018 climatology, leaving winds and geopotential unchanged. If the underprediction of the Niño3.4 anomaly persists, the paper's proposed explanation (an OLR anomaly that is too weak or model drift) is falsified; if the forecast degrades further, OLR is carrying part of the signal.

Watch

Extended reading notes

Core claim

The paper claims that a single U-Net with roughly 3.1 million trainable parameters, trained on ERA5 and UFS-Replay data, learns sufficient autoregressive ocean and sea ice dynamics to simulate present-day upper-ocean and sea ice states for multiple years when driven by the observed atmospheric state. The central methodological choice is training with an almost-fair patch energy score (afPES) instead of a marginal loss like CRPS, which the authors argue is what lets individual ensemble members remain spatially coherent. The evidence offered includes 90-day ensemble forecasts that beat persistence and climatology baselines at nearly all variables and leads; a 5-year, 50-member integration whose climatology and interannual variability track reanalysis; and 29-year integrations that capture the ENSO cycle. The authors also show that the ensembles bracket observed extremes, including the 2023 Antarctic sea ice minimum and the 2019 Blob 2.0 heatwave, while noting a consistent underprediction of the 2023 El Niño amplitude that they attribute to the OLR forcing or model drift.

Load-bearing premise

The model's skill rests on the idea that three weather variables—the height of the 1000-hPa pressure surface, wind speed 10 meters above the surface, and outgoing heat radiation—plus the initial ocean state are enough to predict how the upper ocean and sea ice will evolve; if the real ocean needs more information from the atmosphere, or if the relationships learned from 1994–2018 change over time, the skill claim breaks.

Editorial extensions

If this is right

  • If DLESyM-Ocean's skill holds when coupled to forecast atmospheric fields rather than perfect reanalysis, it would give AI Earth system models a multi-year-stable ocean and ice component for the first time.
  • The afPES objective provides a template for training other spatially coherent probabilistic emulators without the Fourier spectral losses used by atmospheric models, which are hard to apply across ocean coastlines.
  • The ensemble spread scaling with RMSE in eddy-active regions (r=0.85 for SST) implies the model has learned a useful uncertainty map for upper-ocean predictability, not just a mean climatology.
  • The model's ability to produce diverse subsurface trajectories from identical atmospheric forcing suggests it can generate counterfactual ocean heat-content states for marine heatwave studies, though the authors stop short of claiming those are dynamically equivalent to NWP ensemble members.

Reading between the lines

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

  • The 2023 El Niño amplitude error is the most informative failure mode: if it stems from the OLR forcing, then the model's three-variable forcing set is insufficient for capturing the energy balance of rapid ENSO growth, and adding surface flux or momentum predictors should be tested.
  • Because the model is trained on reanalysis, its 'internal variability' is really the statistical spread of learned ocean-atmosphere correlations; whether that spread generalizes to non-stationary climates (for example, a warming Arctic with thinner ice) can only be tested with out-of-sample decades.
  • One could extend the same architecture to predict additional ocean variables (e.g., biogeochemical tracers) by adding channels and reweighting the afPES, since the loss is agnostic to variable type.
  • If coupled to an AI atmosphere model, the 4-day and 8-day lead structure could enable online data assimilation, since the residual formulation gives a natural prior for state updates.
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

3 major / 6 minor

Summary. DLESyM-Ocean is a global, probabilistic deep-learning emulator for upper-ocean and sea-ice fields, built on a ConvNeXt U-Net with HEALPix tiling and an almost-fair patch energy score (afPES) loss. The model ingests current and previous ocean/ice states plus three atmospheric forcing fields (1000-hPa height, 10-m wind speed, and OLR) and outputs 4- and 8-day residuals, with ensemble spread generated by conditioning-layer norm noise. Training uses ERA5 and UFS-Replay data from 1994-2018; validation is 2019-2021 and the test period 2022-2023. The paper reports 90-day forecast skill against persistence and climatology, a 5-year stable climatological integration, case studies of the 2019 Blob, the 2023 El Nino transition, and the 2023 Antarctic sea-ice minimum, and a 29-year ENSO integration. The central claims are that the model is stable, skillful, and well-calibrated, with minimal bias relative to reanalyses.

Significance. If the claims are upheld, the paper would demonstrate a fast (3.1M-parameter), stable, probabilistic ocean/sea-ice emulator that can be driven by atmospheric forcing and potentially coupled to AI atmosphere models. The afPES loss and the careful treatment of fair-score degeneracies are a useful contribution to probabilistic ML for Earth systems. The multi-year stability and the diverse ensemble case studies are encouraging. However, the calibration claim is contradicted by the paper's own diagnostics, and the 2023 El Nino underprediction raises a substantial question about the sufficiency of the chosen forcing set. These issues currently limit the strength of the central contribution.

major comments (3)
  1. [Abstract; Section 3.1, Figures 1I-P, S5, S6] The abstract claims DLESyM-Ocean 'produces a well-calibrated, spatially coherent, and skillful ensemble,' but the paper's own calibration diagnostics show otherwise: Section 3.1 reports that the spread-skill ratio is 'modestly underdispersive for most variables and lead times' (Figures 1I-P, S5), and the rank histograms in Figure S6 are U-shaped, which the caption identifies as underdispersion. Underdispersion means the ensemble is not well calibrated in the standard probabilistic sense; the spatial correlation between spread and RMSE (r=0.85, Figure 2) does not establish calibration. This is a load-bearing claim because the abstract's headline is a well-calibrated ensemble. Please either revise the abstract and conclusions to say 'slightly underdispersive but spatially coherent,' or provide a different calibration metric that supports the original wording. The distinction matters for users who will interpret ensemble spread as predictive uncertainty.
  2. [Section 4 (Conclusion) and Section 3.3 (Figure 5D, 6)] The 2023 El Nino is the most energetic interannual event in the test set, and no ensemble member reaches the observed Nino-3.4 amplitude by year-end (Figure 5D). The paper attributes this in Section 4 to a 'possible mismatch between observed OLR and surface warming' or 'some drift,' but offers no quantitative test of either explanation. This matters because the central skill claim rests on the sufficiency of the three atmospheric forcings (z1000, w10, OLR) chosen in Section 2.4; if OLR does not capture the air-sea fluxes that drive ENSO growth, the 90-day skill shown for 2022 may not generalize. Please either add a test (e.g., recompute the 2023 case with additional or alternative forcing fields, or compare the model's implied surface heat fluxes against reanalysis) or explicitly narrow the skill claim to the variables and periods where the forcing assumption is supported.
  3. [Section 3.3 (Figure 5A)] The 29-year ENSO evaluation (Figure 5A) includes the training period (1994-2018), and the paper acknowledges this. The reproduction of the 1997/98 and 2015/16 El Nino events is therefore not an out-of-sample test. The only out-of-sample interannual event, 2023, is underpredicted. To support the claim that the model 'reproduces ENSO variability,' please separate training-period from validation/test-period skill (e.g., show skill only for 2019-2023 for the 2019-initialized run) or discuss the training-period results explicitly as a consistency check rather than predictive evidence. As it stands, the evidence for interannual predictive skill outside the training distribution is limited to the underpredicted 2023 event.
minor comments (6)
  1. [Section 2.1 and Figure S1] The text states channel depths D1=136, D2=64, D3=34, but Figure S1 labels D2=68. Please reconcile the two values.
  2. [Section 3.1 vs SI Figures S2-S6] The main text describes 50-member forecasts initialized weekly from January 2022 through December 2022, while the SI captions refer to 25-member forecasts initialized weekly from 2021-01 through 2023-12. Please clarify which configuration underlies each figure and ensure the calibration discussion is consistent with the ensemble size actually used.
  3. [Equation (2)] The variables Z in MLP_gamma(Z) and MLP_beta(Z) are not defined in the text; presumably they should be the conditioning noise vector nv. Please clarify.
  4. [Section 2.3] The phrase '600 epochs using 4-10 AR steps (100 epoch per AR step)' should read '100 epochs per AR-step length' for consistency with Table 2.
  5. [Figure 8 caption] The caption refers to '(K-T) Hovmoller diagrams,' but only panels K and L are shown in the figure; please correct the panel range.
  6. [Data Availability] The paper states that all data are publicly available, but no code or model checkpoints are listed. Given the novelty of the afPES loss and the architecture details, releasing code or weights would substantially strengthen reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central skill claim is tested on held-out periods against external baselines; self-citations are architectural and not load-bearing.

full rationale

The paper's central claim is that DLESyM-Ocean produces well-calibrated, spatially coherent, skillful ensembles of upper-ocean and sea-ice conditions. This claim is evaluated on genuinely held-out data: training uses 1994-2018, validation 2019-2021, and the headline 90-day forecast skill is computed on 2022, with the 2023 El Nino case lying entirely in the test period. The model under-predicts the 2023 ENSO amplitude (Fig. 5D), which is the opposite of what a circularly constructed evaluation would produce. The 29-year ENSO rollout does include training-period forcing, but the authors disclose this and use it to assess long-term stability and climatology, not to claim out-of-sample forecast skill. Skill metrics are benchmarked against persistence and a 25-member probabilistic ocean climatology, which are external baselines independent of the model's fitted parameters. The paper does not fit a parameter to a subset of data and then rename it a prediction: the afPES loss hyperparameters (alpha = 0.95, ensemble size M = 2) are fixed training choices, not fitted to the evaluation targets. Self-citations to Cresswell-Clay et al. (2025) and Karlbauer et al. (2024) provide the architecture, HEALPix mesh, and training conventions, but the paper independently modifies these components (isolatitude padding, custom upsampling, global residual connection, conditional layer norms) and does not invoke those papers to justify forecast skill. No uniqueness theorem is imported, and no ansatz is smuggled in via citation. The assumed sufficiency of the three atmospheric forcing fields is a substantive external-validity assumption, and the paper itself flags the 2023 OLR mismatch and possible drift as limitations; these are robustness concerns, not circular reductions. The SI contains an omitted proof that afCRPS breaks the fair-score degeneracies, which is a missing derivation rather than a circular step. No equation in the paper defines a predicted quantity in terms of the target by construction, and no fitted constant is relabeled as a forecast. The derivation chain is therefore self-contained with respect to circularity.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The model is an emulator built on a standard U-Net over HEALPix, so the central claim does not introduce new physical entities or forces. Its novel content is the afPES loss and ensemble mechanism; the free parameters listed are training choices that influence the calibration and skill metrics. The assumptions concern the fidelity of the training target and the sufficiency of the chosen atmospheric predictors.

free parameters (5)
  • alpha (afPES interpolation parameter) = 0.95
    Chosen by hand for both training stages; interpolates between biased and fair energy score estimators, affecting ensemble spread and spatial coherence (Section 2.2).
  • per-variable loss weights w_i = Varies; e.g., sic=2.000, sst=0.400, sit=5.882, ssh=0.333 (Table 1)
    Computed as 0.001/V_i from 10-epoch validation losses so each variable contributes roughly equally to the afPES loss; this is fitting to validation data and shapes the model's error priorities (Section 2.3).
  • ocean fraction threshold for evaluation mask = >= 0.50
    Binary land-sea mask applied after inference uses this threshold; changing it alters all reported RMSE and climatology metrics (Section 2.4).
  • latent noise vector dimension = 32
    The global noise vector injected into conditional layer norms has dimension 32, controlling the degrees of freedom of ensemble diversity; chosen by architecture design, not derived (Section 2.1).
  • training ensemble size M = 2
    The afPES loss is evaluated with a two-member ensemble during training, which interacts with the almost-fair correction and the degeneracy analysis in the supplement (Section 2.3).
assumptions (4)
  • domain assumption ERA5 and UFS-Replay reanalyses are treated as ground truth for both training and verification.
    The model is trained to reproduce these products and skill is measured against them; errors in the reanalyses become errors in the model. Invoked throughout Section 2.4 and Section 3.
  • domain assumption The three atmospheric forcing fields (z1000, 10-m windspeed, OLR) plus the ocean and ice initial state are sufficient to determine future upper-ocean and sea ice evolution.
    The forcing set is fixed in Section 2.4, and the paper attributes the 2023 El Nino underprediction to OLR behavior in Section 4, making this sufficiency assumption load-bearing.
  • ad hoc to paper The almost-fair correction eliminates the zero-gradient degeneracies of the fair PES for the training configuration (M=2, 3x3 patches, multivariate channels).
    The supplement derives degeneracies only for scalar M=2 patches; the paper states the pathologies persist for P>1 and M>2 and that afPES breaks them, with the derivation 'omitted for brevity'.
  • domain assumption Ensemble spread from CLN noise represents internal ocean-sea ice variability, not just injected noise.
    The paper interprets diverse members as 'learned autoregressive ocean dynamics' (Sections 3.3 and 4) but validates this only via correlation between spread and RMSE (r=0.85), not via physical consistency of individual members.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DLESyM-Ocean: A Deep Learning Probabilistic Global Model for Simulating Present-Day Upper Ocean and Sea Ice." pith.science (2026). https://pith.science/paper/MET6G64A

@misc{pith2026260811545,
  author       = {Pith},
  title        = {Pith review of: DLESyM-Ocean: A Deep Learning Probabilistic Global Model for Simulating Present-Day Upper Ocean and Sea Ice},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MET6G64A}},
  note         = {Machine review of arXiv:2608.11545}
}
read the original abstract

While AI has shown remarkable promise in atmospheric and meteorological forecasting, accurately simulating other components of the Earth system with AI remains an active frontier. We present DLESyM-Ocean, a Deep Learning Earth System Model that simulates global present-day sea ice and upper ocean conditions. Unlike conventional probabilistic models optimized via diffusion objectives or losses such as continuous-ranked probability score, DLESyM-Ocean is trained using a patch energy score loss. When driven by atmospheric forcing, DLESyM-Ocean produces a well-calibrated, spatially coherent, and skillful ensemble of sea ice and upper ocean conditions with minimal bias relative to reanalysis products. DLESyM-Ocean is stable when autoregressively run for multi-year simulations and produces a climatology and variability with minimal bias compared with reanalysis. We evaluate case studies including a recent sea ice extreme, a severe marine heatwave, the 2023 El Ni\~no transition, and the 2023 spike in global mean temperature. In all of these case studies, DLESyM-Ocean produces realistic surface and subsurface trajectories and ample ensemble diversity in response to common atmospheric forcing, suggestive of learned autoregressive ocean dynamics. When coupled with other Earth system components, such as the atmosphere, the computational efficiency of DLESyM-Ocean makes it a promising tool for subseasonal to seasonal forecasting.

Figures

Figures reproduced from arXiv: 2608.11545 by the authors.

Figure 1
Figure 1. Forecast Skill: (A–H) Ensemble spread (dashed lines) and root-mean-square error (RMSE) of the ensemble mean (solid lines) and a single ensemble member (dotted lines) for 50-member, 90-day DLESyM-Ocean forecasts initialized weekly from 2022-01 through 2022-12. Forecasts are shown for two configurations: DLESyM-Ocean driven by concurrent ERA5 atmospheric forcing (blue) and DLESyM-Ocean-AC (orange). Comparison benchmar… view at source ↗
Figure 2
Figure 2. Spread & Calibration: (A–C) Spatial distributions of (A) ensemble spread, (B) ensemble-mean RMSE (RMSE_ENS), and (C) the unbiased spread-skill ratio for SST, averaged across all initialization times at an 80-day lead time. (D–F) As in (A–C), but for SSH. As in [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Global Climatology: (A) Time series of the global mean sea surface temperature (SST) anomaly for ERA5 (solid black line), DLESyM-Ocean ensemble mean (solid magenta line), and DLESyM-Ocean ensemble members (solid blue lines; N = 50) from January 2019 through December 2023, computed relative to a 1994–2018 monthly ERA5 climatology. (B–D) Global 5-year mean spatial distributions of SST over the 2019–2023 period for (B)… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Sea Ice Mean Climatology and Monthly Variability: Annual cycle of Arctic (A) and Antarctic (B) sea ice extent (SIE) averaged over 2019-01 to 2023-12 for ERA5 and DLESyM-Ocean. (C) Time series of Arctic SIE anomaly for ERA5 (solid black line), DLESyM-Ocean ensemble mean…
Figure 5
Figure 5. Figure 5: El Niño: (A) Time series of the sea surface temperature (SST) anomaly in the Niño-3.4 region (5◦S–5◦N, 170◦W–120◦W) from 1994 through 2023 for ERA5 (solid black line), DLESyM-Ocean mean of ensemble simulations initialized on January 2, 1994 and forced with concurrent E…
Figure 6
Figure 6. Figure 6: El Niño 2023: (A–C) Mean SST anomalies for September–December 2023 from (A) ERA5, (B) the maximum of a 50-member DLESyM-Ocean ensemble (the member with the largest positive area-averaged Niño-3.4 anomaly), and (C) the ensemble mean. (D) Standard deviation of SST anomal…
Figure 7
Figure 7. Figure 7: Marine Heatwave: (A) Time series of sea surface temperature (SST) anomaly averaged over the domain indicated by the black bounding box in (B) from January 2019 through September 2019. Data are shown for ERA5 (solid black line), the 50-member DLESyM-Ocean ensemble mean …
Figure 8
Figure 8. Figure 8: Antarctic Sea Ice: (A–E) Spatial distributions of austral winter (June–August [JJA] 2023) sea ice con￾centration (SIC) anomalies for (A) ERA5, (B) a 50-member DLESyM-Ocean ensemble driven by concurrent ERA5 atmospheric forcing, (C) CESM2 nudged to concurrent ERA5 winds…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 18 canonical work pages

  1. [1]

    top of atmosphere

    experiences a zero gradient and detaches. • Bracketing (x1≤y≤x M ): Both extremes experience a zero gradient simultaneously, leaving the ensemble boundaries completely unconstrained. Unlike theM= 2 bracketing case where this cancellation reduces the total loss identically to zero (fCRPS M=2 = 0), forM >2 the total fPES remains strictly positive due to the...

    work page 1994

  2. [3]

    Gregory J. Hakim. Dynamical Tests of a Deep Learning Weather Prediction Model in: Artificial Intelligence for the Earth Systems V olume 3 Issue 3 (2024). https://journals.ametsoc.org/view/journals/aies/3/3/AIES-D-23-0090.1.xml,

    work page 2024

  3. [5]

    Aimip phase 1: systematic evaluations of ai weather and climate models.arXiv preprint arXiv:2605.06944,

    Brian Henn, Christopher S Bretherton, Nikolay Koldunov, Christian Lessig, Maria J Molina, Troy Arcomano, Oliver Watt-Meyer, Guillaume Couairon, Renu Singh, Robert Brunstein, et al. Aimip phase 1: systematic evaluations of ai weather and climate models.arXiv preprint arXiv:2605.06944,

    arXiv

  4. [7]

    Daniel Holmberg, Emanuela Clementi, and Teemu Roos

    doi:10.1029/2023MS00413710.22541/essoar.170110658.85641696/v1. Daniel Holmberg, Emanuela Clementi, and Teemu Roos. Regional Ocean Forecasting with Hierarchical Graph Neural Networks, November

    work page arXiv

  5. [8]

    Njord: A Probabilistic Graph Neural Network for Ensemble Ocean Forecasting

    Daniel Holmberg, Joel Oskarsson, Erik Wikingsson, Fredrik Lindsten, and Teemu Roos. Njord: A probabilistic graph neural network for ensemble ocean forecasting.arXiv preprint arXiv:2605.15470,

    work page Pith review arXiv

  6. [10]

    doi:10.1029/2024JH000433

    ISSN 2993-5210. doi:10.1029/2024JH000433. Spencer K. Clark, Oliver Watt-Meyer, Anna Kwa, Jeremy McGibbon, Brian Henn, W. Andre Perkins, Elynn Wu, Christopher S. Bretherton, and Lucas M. Harris. ACE2-SOM: Coupling to a slab ocean and learning the sensitivity of climate to changes in CO$_2$, December

    work page doi:10.1029/2024jh000433

  7. [11]

    Representing the surface ocean in ecmwf’s data-driven forecasting system aifs.arXiv preprint arXiv:2604.25559,

    Sara Hahner, Lorenzo Zampieri, Jean-Raymond Bidlot, Philip Browne, Matthew Chantry, Mariana CA Clare, Harrison Cook, Peter Dueben, Rachel Furner, Sarah Keeley, et al. Representing the surface ocean in ecmwf’s data-driven forecasting system aifs.arXiv preprint arXiv:2604.25559,

    arXiv

  8. [18]

    WeatherBench 2: A benchmark for the next generation of data-driven global weather models

    Stephan Rasp, Stephan Hoyer, Alexander Merose, Ian Langmore, Peter Battaglia, Tyler Russel, Alvaro Sanchez- Gonzalez, Vivian Yang, Rob Carver, Shreya Agrawal, Matthew Chantry, Zied Ben Bouallegue, Peter Dueben, Carla Bromberg, Jared Sisk, Luke Barrington, Aaron Bell, and Fei Sha. WeatherBench 2: A benchmark for the next generation of data-driven global we...

    arXiv

Show all 31 references
  1. [19]

    Youmin Tang, Richard Kleeman, and Sonya Miller

    doi:10.1175/1520-0493(2000)128<1835:ACOPFF>2.0.CO;2. Youmin Tang, Richard Kleeman, and Sonya Miller. ENSO Predictability of a Fully Coupled GCM Model Using Singular Vector Analysis.Journal of Climate, 19(14):3361–3377, July

    2000 doi

  2. [21]

    The 2023 global warming spike was driven by the el niño–southern oscillation.Atmospheric chemistry and physics, 24 (19):11275–11283,

    Shiv Priyam Raghuraman, Brian Soden, Amy Clement, Gabriel Vecchi, Sofia Menemenlis, and Wenchang Yang. The 2023 global warming spike was driven by the el niño–southern oscillation.Atmospheric chemistry and physics, 24 (19):11275–11283,

    2023

  3. [22]

    doi:10.1029/1999GL011047. R. Justin Small, Shang-Ping Xie, and Yuqing Wang. Numerical simulation of atmospheric response to Pacific tropical instability waves.Journal of Climate, 16(22):3723–3741,

    doi

  4. [25]

    Era5 hourly data on single levels from 1940 to present.Copernicus climate change service (c3s) climate data store (cds), 10(1.24381):24381,

    Hans Hersbach, Bill Bell, Paul Berrisford, Gionata Biavati, András Horányi, Joaquín Muñoz Sabater, Julien Nicolas, Carole Peubey, Raluca Radu, Iryna Rozum, et al. Era5 hourly data on single levels from 1940 to present.Copernicus climate change service (c3s) climate data store ...

    1940

  5. [27]

    showing the La Niña to El Niño transition. Model comparisons between UFS-Replay (A, C, E, G) and DLESyM-Ocean ensemble mean (50-members; B, D, F, H) are presented at four depth levels: T0m (A–B), T50m (C–D), T100m (E–F), and T200m (G–H). All DLESyM-Ocean simulations are initia...

    2023

  6. [28]

    Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean

    Panels B-J are as in (A) but for 9 randomly chosen DLESyM-Ocean ensemble members. Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean. The RMSE relative to UFS-Replay is displa...

    1994

  7. [29]

    Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean

    Panels B-J are as in (A) but for 9 randomly chosen DLESyM-Ocean ensemble members. Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean. The RMSE relative to UFS-Replay is displa...

    1994

  8. [30]

    Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean

    Panels B-J are as in (A) but for 9 randomly chosen DLESyM-Ocean ensemble members. Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel P shows the 50-member ensemble mean. The RMSE relative to UFS-Replay is displa...

    1994

  9. [31]

    Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel L shows the 50-member ensemble mean

    Panels B-J are as in (A) but for 9 randomly chosen DLESyM-Ocean ensemble members. Panel K shows the ensemble member with the lowest RMSE over the full Hovmöller with respect to UFS-Replay, and panel L shows the 50-member ensemble mean. The RMSE relative to UFS-Replay is displa...

    1994

  10. [2000]

    Fourcastnet 3: A geometric approach to probabilistic machine-learning weather forecasting at scale.arXiv preprint arXiv:2507.12144,

    Boris Bonev, Thorsten Kurth, Ankur Mahesh, Mauro Bisson, Jean Kossaifi, Karthik Kashinath, Anima Anandkumar, William D Collins, Michael S Pritchard, and Alexander Keller. Fourcastnet 3: A geometric approach to probabilistic machine-learning weather forecasting at scale.arXiv p...

    arXiv

  11. [2001]

    Fair scores for ensemble forecasts.Quarterly Journal of the Royal Meteorological Society, 140(683): 1917–1923,

    CAT Ferro. Fair scores for ensemble forecasts.Quarterly Journal of the Royal Meteorological Society, 140(683): 1917–1923,

    1917

  12. [2003]

    Physical drivers of the summer 2019 north pacific marine heatwave.Nature communications, 11(1):1903,

    Dillon J Amaya, Arthur J Miller, Shang-Ping Xie, and Yu Kosaka. Physical drivers of the summer 2019 north pacific marine heatwave.Nature communications, 11(1):1903,

    2019

  13. [2005]

    doi:10.1086/427976

    ISSN 0004-637X. doi:10.1086/427976. Andrea Zonca, Leo Singer, Daniel Lenz, Martin Reinecke, Cyrille Rosset, Eric Hivon, and Krzysztof Gorski. healpy: equal area pixelization and spherical harmonics transforms for data on the sphere in python.Journal of Open Source Software, 4(...

    doi

  14. [2006]

    Eduardo Blanchard-Wrigglesworth, Roberto Bilbao, Aaron Donohoe, and Stefano Materia

    doi:10.1175/JCLI3771.1. Eduardo Blanchard-Wrigglesworth, Roberto Bilbao, Aaron Donohoe, and Stefano Materia. Record warmth of 2023 and 2024 was highly predictable and resulted from enso transition and northern hemisphere absorbed shortwave anomalies.Geophysical Research Letter...

    2023 doi

  15. [2016]

    Climate in a bottle: Towards a generative foundation model for the kilometer-scale global atmosphere.arXiv preprint arXiv:2505.06474,

    Noah D Brenowitz, Tao Ge, Akshay Subramaniam, Peter Manshausen, Aayush Gupta, David M Hall, Morteza Mardani, Arash Vahdat, Karthik Kashinath, and Michael S Pritchard. Climate in a bottle: Towards a generative foundation model for the kilometer-scale global atmosphere.arXiv pre...

    arXiv

  16. [2017]

    The era5 global reanalysis.Quarterly journal of the royal meteorological society, 146(730):1999–2049,

    Hans Hersbach, Bill Bell, Paul Berrisford, Shoji Hirahara, András Horányi, Joaquín Muñoz-Sabater, Julien Nicolas, Carole Peubey, Raluca Radu, Dinand Schepers, et al. The era5 global reanalysis.Quarterly journal of the royal meteorological society, 146(730):1999–2049,

    1999

  17. [2019]

    doi:10.1029/2019MS001705

    ISSN 1942-2466. doi:10.1029/2019MS001705. Ryan Keisler. Forecasting global weather with graph neural networks.arXiv preprint arXiv:2202.07575,

    1942 arXiv

  18. [2020]

    Huge ensembles–part 1: Design of ensemble weather forecasts using spherical fourier neural operators.Geoscientific Model Development, 18(17):5575–5603, 2025a

    19 arXivTemplateA PREPRINT Ankur Mahesh, William D Collins, Boris Bonev, Noah Brenowitz, Yair Cohen, Joshua Elms, Peter Harrington, Karthik Kashinath, Thorsten Kurth, Joshua North, et al. Huge ensembles–part 1: Design of ensemble weather forecasts using spherical fourier neura...

    2023

  19. [2021]

    doi:10.1038/s41467-021-25257-4

    ISSN 2041-1723. doi:10.1038/s41467-021-25257-4. Yongcheng Lin, Qinghua Yang, Xuewei Li, Xiaoran Dong, Hao Luo, Yafei Nie, Jiuke Wang, Yiguo Wang, and Chao Min. Ice-kNN-South: A Lightweight Machine Learning Model for Antarctic Sea Ice Prediction.Journal of Geophysical Research:...

    doi

  20. [2023]

    doi:10.1029/2022GL102649

    ISSN 1944-8007. doi:10.1029/2022GL102649. Wei Xiong, Yanfei Xiang, Hao Wu, Shuyi Zhou, Yuze Sun, Muyuan Ma, and Xiaomeng Huang. AI-GOMS: Large AI-Driven Global Ocean Modeling System, August

    1944 doi

  21. [2024]

    Can ai weather models predict beyond two weeks? a quantitative benchmark and analysis of long rollouts.arXiv preprint arXiv:2605.30184,

    Fanny Lehmann, Firat Ozdemir, Yun Cheng, Torsten Hoefler, Sebastian Schemm, Benedikt Soja, and Siddhartha Mishra. Can ai weather models predict beyond two weeks? a quantitative benchmark and analysis of long rollouts.arXiv preprint arXiv:2605.30184,

    arXiv

  22. [2025]

    Testing the limit of atmospheric predictability with a machine learning weather model.arXiv preprint arXiv:2504.20238,

    P Trent V onich and Gregory J Hakim. Testing the limit of atmospheric predictability with a machine learning weather model.arXiv preprint arXiv:2504.20238,

    arXiv

  23. [2026]

    Samudra 2: Scaling ocean emulators across resolutions.arXiv preprint arXiv:2606.02610,

    Yuan Yuan, Jesse Rusak, Alexander Merose, Adam Subel, Pavel Perezhogin, Alistair Adcroft, Carlos Fernandez-Granda, and Laure Zanna. Samudra 2: Scaling ocean emulators across resolutions.arXiv preprint arXiv:2606.02610,

    arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.