Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Generating Unseen Nonlinear Evolution in Sea Surface Temperature Using a Deep Learning-Based Latent Space Data Assimilation Framework

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 DeepDA performs variational data assimilation in a 128-dimensional latent space and reconstructs unseen nonlinear sea surface temperature evolution from 10 percent of the original observation information, with error growth no more than 40…

desk verdict A careful, internally consistent application of latent-space variational DA to ocean SST, but the central 'unseen evolution' claim is undermined by verifying against the same ERA5 product used to train the generative model. read the letter →

arxiv 2412.13477 v1 pith:BUEXL64R submitted 2024-12-18 physics.ao-ph cs.AIcs.CVcs.LGphysics.geo-ph

Qingyu Zheng , Guijun Han , Wei Li , Lige Cao , Gongfu Zhou , Haowen Wu , Qi Shao , Ru Wang
show 4 more authors
Xiaobo Wu Xudong Cui Hong Li Xuan Wang
This is my paper · ORCID
classification physics.ao-phcs.AIcs.CVcs.LGphysics.geo-ph
keywords DeepDAlatentspacedataassimilationseasurfacetemperatureSTAVAEnonlinearevolutionspatio-temporalattentionmulti-sourcefusionENSO
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

DeepDA is a deep-learning data assimilation framework that moves the classical 3D-Var minimization into a low-dimensional latent space learned by a generative autoencoder. The paper argues that this latent-space formulation can fuse sparse observations with background fields to reconstruct nonlinear, multi-scale sea surface temperature (SST) evolution that the observations alone do not resolve. Its headline evidence is an observing-system simulation experiment in which reducing observation resolution from 0.25° to 2.5° (which the paper counts as 10 percent of the observation information) raises analysis RMSE by no more than 40 percent, while daily analyses over four years hold RMSE near 0.3°C and anomaly correlation near 0.8. The authors also test the framework with real blended satellite and in-situ observations plus a 51-member ensemble forecast background, and show that the leading elements of the latent vector track ENSO events, arguing that the learned latent patterns carry physical meaning. If the claim holds, latent-space DA offers a computationally light path to generate physically interpretable ocean analyses from very sparse data.

What carries the argument

The load-bearing object is the 128-dimensional latent space of STAVAE, a convolutional variational autoencoder whose encoder maps a 384,000-pixel SST anomaly field into a mean and log-variance vector from which a latent vector is sampled, and whose decoder maps that vector back to a full field. Around this core DeepDA wraps a 3D-Var cost function in latent coordinates, $$\mathcal{J}(\mathbf{z}) = \frac12 \|\mathbf{z}-\mathbf{z}_b\|^2_{\mathbf{B}$_z^{{-1}}$} + \frac12 \|\mathbf{y} - H(D(\mathbf{z}))\|^2_{\mathbf{R}^{-1}},$$ where $D$ is the decoder and $H$ is bilinear interpolation onto observation points. The STAR module—residual convolutional blocks with a spatio-temporal attention mechanism that replaces CBAM's channel-attention MLP with a sequence-transfer layer—is what lets the VAE extract multi-scale nonlinear features. Because the minimization runs in only 128 dimensions with an empirically estimated, diagonally dominant background error covariance, the whole data assimilation step takes roughly 15 seconds and fits on a personal computer. This small optimization space is what lets the scheme stay stable when observations are extremely sparse.

What would settle it

Take a 2.5°-observation run and compute, for each optimized latent vector, its statistical distance from the center of the training-set latent distribution (for example, the Mahalanobis distance under the estimated latent covariance). If many analyses fall far outside that distribution and their decoded fields disagree with withheld high-resolution observations in the same cases, the generated structures are decoder extrapolation rather than reconstructed physics.

Watch

Extended reading notes

Core claim

The paper's central claim is that data assimilation should be run in the latent space of a generative model rather than in the raw grid space. DeepDA first trains a spatio-temporal attention variational autoencoder (STAVAE) to compress Pacific sea surface temperature anomaly fields into a 128-dimensional latent vector, then minimizes a classical 3D-Var cost function in that latent space with the model weights frozen. The decoder, together with bilinear interpolation onto observation points, plays the role of the observation operator, and the background error covariance matrix estimated in latent space is diagonally dominant. In observing-system simulation experiments the scheme keeps daily analyses at roughly 0.3°C RMSE and 0.8 anomaly correlation over four years; reducing observation resolution from 0.25° to 2.5° (10 percent of the observation information in the paper's counting) raises RMSE by no more than 40 percent. The paper also shows that the first two latent dimensions track ENSO and that their reconstructed spatial patterns resemble the classic El Niño pattern, which it offers as evidence that the latent space is physically meaningful rather than a purely statistical compression.

Load-bearing premise

The load-bearing premise is that the 128-number latent representation learned from three decades of SST reanalysis is smooth and faithful enough that every point the optimizer visits still decodes to a physically realistic ocean pattern.

Editorial extensions

If this is right

  • Reducing observation resolution from 0.25° to 2.5° raises the analysis RMSE by at most 40 percent, and DeepDA still reconstructs large-scale spatial structure, including a tropical instability wave that is aliased with noise in the 2.5° observation.
  • Over 1,461 daily assimilation experiments from 2020 to 2023, analysis RMSE remains below 0.4°C and anomaly correlation is stable around 0.8, so the scheme's performance is stable across seasons and years.
  • Using an ensemble background widens the latent optimization space; ensemble averaging reduces RMSE by about 30 percent relative to the best single member, and more ensemble members improve the ensemble mean.
  • Fusing real observations with a 51-member seasonal forecast ensemble background yields analyses whose anomaly correlation differs from the reference reanalysis by about 3 percent, versus about 10 percent for the ensemble mean, indicating robust multi-source fusion.
  • The first two latent elements of STAVAE track ENSO, and the latent-pattern phase space separates the 1997/98 eastern-Pacific El Niño from the 2015/16 mixed El Niño, suggesting that the latent space is physically interpretable.

Reading between the lines

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

  • The paper explicitly leaves a physical validation of the latent-pattern connection to future work, so the ENSO interpretability claim currently rests on correlation between latent time series and known indices rather than on demonstrated causality.
  • A natural test the paper does not run is to check whether optimized latent vectors stay inside the training distribution under 2.5° observations; if they systematically leave it, the reconstructed structure would be decoder extrapolation rather than reconstructed physics.
  • Because the framework is variable-agnostic, the same latent-space 3D-Var recipe could be applied to other ocean fields such as sea surface height or subsurface temperature, or to radiance observations with nonlinear observation operators; the 40-percent error-growth bound would then have to be re-measured per variable.
  • The term 'unseen nonlinear evolution' here refers to spatial patterns that sparse observations alias or blur, not to temporal forecasting; whether DeepDA analyses are dynamically consistent across consecutive times would require a temporal consistency experiment that the paper does not report.
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. The paper proposes DeepDA, a latent-space variational data assimilation framework in which a spatio-temporal attention variational autoencoder (STAVAE) maps sea surface temperature anomaly fields into a 128-dimensional latent space, and a 3D-Var-style cost function is minimized in that space to fuse observations and background information. The generative proxy model is trained on ERA5 SSTA from 1989–2018. In observing system simulation experiments, ERA5 serves as ground truth, persistence of ERA5 as background, and interpolated ERA5 plus Gaussian noise as pseudo-observations; the authors report stable RMSE/ACC over 1461 days and argue that the error increase stays below 40% when the observation resolution coarsens from 0.25° to 2.5°. Real-data experiments fuse OISST observations with SEAS5 ensemble forecasts, again benchmarked against ERA5. The paper also includes an ablation of the generative model and a pattern-based analysis linking latent dimensions to ENSO.

Significance. If the results hold, DeepDA is a computationally attractive route to nonlinear data assimilation: it avoids full-state covariance matrices, fuses heterogeneous data in under a minute per analysis, and its latent patterns show physically interpretable ENSO structure. Strengths include a systematic architecture ablation, frozen generator weights during DA, a clearly stated variational objective, and a long OSSE evaluation period. The significance is conditional, however, because the headline claim is currently verified against the same reanalysis product used to train the generative model, and no classical DA baseline is reported. The paper is of interest to the physics-of-the-ocean and machine-learning-for-Earth-systems communities, but the evidence for "generating unseen nonlinear evolution" needs to be made independent of the training distribution.

major comments (4)
  1. [§2.3.1–2.3.2, §3.3] The verification design does not support the "unseen" claim as stated. STAVAE is trained on ERA5 SSTA from 1989–2018 (§2.3.1), while the OSSE uses ERA5 as ground truth, derives pseudo-observations from ERA5 via Eq. (12), and computes all RMSE/ACC metrics against ERA5 (§2.3.2, §3.3). The test years 2020–2023 are temporally outside the training interval, so this is not simple memorization, but the decoder's training distribution still coincides with the verification distribution. The reported RMSE and the ≤40% degradation could therefore reflect the prior's tendency to decode into ERA5-like states rather than a genuinely independent reconstruction of nonlinear evolution. Please repeat the headline experiment with a truth product not used to train GenPM (e.g., OISST or an independent reanalysis), and verify the Section 4.1 real-data fusion against OISST as well as ERA5.
  2. [Abstract, §3.3, Figs. 7–8] The "10% of observation information" statement is numerically inaccurate. Reducing the observation grid from 0.25° to 2.5° reduces the number of observation points by a factor of 100 in two dimensions (to 1%, not 10%); if the intended meaning is that the linear resolution is 10% of the original, that should be stated explicitly. Since this number appears in the Abstract and in Section 3.3 as the paper's headline claim, the wording should be corrected or carefully qualified.
  3. [§2.3.2, Eq. (11), Eq. (12)] The observation error specification is incomplete. Pseudo-observations are generated with "random Gaussian perturbations" in Eq. (12), but no noise variance is reported, and the observation error covariance R in the cost function Eq. (11) is never defined. Because the resolution-sensitivity experiment in Section 3.3 is designed to isolate the effect of observation resolution, the assumed R must be specified and held fixed across all resolution cases for the comparison to be meaningful.
  4. [§3.1–3.3] No standard data assimilation baseline is reported. Section 3 reports DeepDA analyses and their RMSE/ACC, but the claim that latent-space generative fusion adds skill is not compared with a same-configuration 3D-Var, optimal interpolation, or ensemble Kalman filter. Without such a baseline, it is unclear whether the robustness and the ≤40% degradation result are specific to DeepDA or would also be achieved by a conventional smoother using the same sparse observations. Please add at least one classical baseline to the OSSE in Section 2.3.2 and report its RMSE/ACC alongside DeepDA in Sections 3.1–3.3.
minor comments (5)
  1. [Table 1] The row numbering in Table 1 is inconsistent: the STAVAE row is labelled "6" after row 7; please renumber the models consecutively.
  2. [§2.2.3, Eq. (11)] In Eq. (11), the symbol D is used both for the decoder and for its tangential operator; please use separate symbols, for example a calligraphic D for the decoder and a bold D_z for the tangent.
  3. [§3.2, Text S1] The construction of the ensemble background members is described only in the supplementary Text S1; a brief description should appear in Section 2.3.2 because it is central to the ensemble-average claim and to the sensitivity analysis in Table S1.
  4. [§4.2, Fig. 10] The comparison of latent vector elements Z1 and Z2 to EOF principal components is not apples-to-apples because VAE latent dimensions have no inherent variance ordering; please state how the first two elements were selected or explicitly note that they are arbitrary coordinates of the learned latent space.
  5. [Appendix A] Please state explicitly that the validation set used to estimate B_z is the year 2019 and is disjoint from the 2020–2023 OSSE evaluation period; this is currently implied but not stated.

Circularity Check

1 steps flagged · score 4.0 of 10

The sparse-observation robustness claim is verified against ERA5, the same reanalysis product used to train the generative model, making the 'unseen nonlinear evolution' evaluation partly self-referential.

  1. fitted input called prediction [Section 2.3.1 (Data Preparation), Section 2.3.2 (OSSE, Eqs. 12-13), Section 3.1 (Overall Performance)]
    "data from 1989 to 2018 are selected to train the GenPM. ... In Fig.3(b), we show the details and flowchart of OSSE, where observations and background are derived from the ERA5 ground truth. ... a total of 1461(4 years) DA experiments are performed and the RMSE and ACC relative to ERA5 are calculated respectively."

    The GenPM is fitted to ERA5 SSTA (1989-2018), and then every OSSE truth, pseudo-observation (Eq. 12) and persistence background (Eq. 13) is built from ERA5, with the final RMSE/ACC computed against that same ERA5 field. The decoder is therefore rewarded for staying on the ERA5-like manifold it was trained to reproduce; the 'generated unseen nonlinear evolution' is, at least in part, interpolation within the training distribution rather than an out-of-sample physical reconstruction. The 40% error-increase claim measures self-consistency inside the training product, so the sparse-observation robustness result is a fitted-input evaluation rather than a test on an independent ground truth.

full rationale

The variational cost function (Eq. 11) is a standard 3D-Var objective in latent space and is not derived from the target result; the architecture and STAR module are substantive. There is no load-bearing self-citation chain or imported uniqueness theorem; self-citations [24,25,30] are background only. However, the central evaluation is partially circular: the generative prior is trained on ERA5 and the verification benchmark is also ERA5, so the sparse-observation 'unseen evolution' claim is not an independent physical test. This raises the score above the 0-2 range but does not reach 6+ because the DA minimization could still fail within the ERA5 manifold and the method has independent algorithmic content.

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

The framework rests on the variational autoencoder assumption, the persistence background, and the transferability of the validation-estimated covariance. No new physical entities are introduced.

free parameters (5)
  • Latent dimension = 128
    Chosen by hand; the latent vector size directly controls the Bz matrix dimension and the capacity of the generative model (Section 2.2.3).
  • Huber loss delta = 1
    Set to 1 in Eq. (8); controls the reconstruction loss shape in the generative proxy model.
  • Persistence lead time tau = 15 days
    Used to generate OSSE background fields from persistence; affects the background error covariance and the difficulty of the assimilation task (Section 2.3.2).
  • Background error covariance Bz = Estimated from validation set 2019
    Inferred from Eq. A1 using the validation period; assumed representative of persistence forecast errors during 2020-2023 (Appendix A).
  • DA learning rate range = 0.001 to 0.0001
    Adam optimizer range for latent space optimization, chosen empirically (Section 2.4).
assumptions (4)
  • domain assumption SST anomalies in the Pacific are well approximated by a 128-dimensional multivariate Gaussian latent distribution.
    The variational autoencoder assumes a Gaussian prior (Section 2.2.1) and the DA cost function assumes Gaussian background and observation errors.
  • domain assumption Persistence of the SST field for 15 days is a valid background that shares error statistics with the validation period.
    OSSE backgrounds are generated from persistence (Eq. 13) but Bz is estimated from a different period, which assumes stationarity of the error covariance.
  • domain assumption ERA5 reanalysis is an acceptable ground truth for evaluating analysis fields.
    Used as truth in OSSE and as benchmark in the multi-source fusion experiments (Sections 2.3.2 and 2.3.3).
  • ad hoc to paper The decoder generalizes to latent vectors reached during DA optimization even if those vectors are outside the training distribution.
    The claim of generating unseen nonlinear evolution requires that the trained decoder produces realistic SST states for any latent vector encountered during minimization (Section 2.2.2).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Generating Unseen Nonlinear Evolution in Sea Surface Temperature Using a Deep Learning-Based Latent Space Data Assimilation Framework." pith.science (2026). https://pith.science/paper/BUEXL64R

@misc{pith2026241213477,
  author       = {Pith},
  title        = {Pith review of: Generating Unseen Nonlinear Evolution in Sea Surface Temperature Using a Deep Learning-Based Latent Space Data Assimilation Framework},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BUEXL64R}},
  note         = {Machine review of arXiv:2412.13477}
}
read the original abstract

Advances in data assimilation (DA) methods have greatly improved the accuracy of Earth system predictions. To fuse multi-source data and reconstruct the nonlinear evolution missing from observations, geoscientists are developing future-oriented DA methods. In this paper, we redesign a purely data-driven latent space DA framework (DeepDA) that employs a generative artificial intelligence model to capture the nonlinear evolution in sea surface temperature. Under variational constraints, DeepDA embedded with nonlinear features can effectively fuse heterogeneous data. The results show that DeepDA remains highly stable in capturing and generating nonlinear evolutions even when a large amount of observational information is missing. It can be found that when only 10% of the observation information is available, the error increase of DeepDA does not exceed 40%. Furthermore, DeepDA has been shown to be robust in the fusion of real observations and ensemble simulations. In particular, this paper provides a mechanism analysis of the nonlinear evolution generated by DeepDA from the perspective of physical patterns, which reveals the inherent explainability of our DL model in capturing multi-scale ocean signals.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ReconMOST: Multi-Layer Sea Temperature Reconstruction with Observations-Guided Diffusion

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A guided diffusion model pre-trained on climate simulations reconstructs multi-layer global ocean temperature from sparse observations, reporting low MSE on CMIP6 and EN4 data.

Reference graph

Works this paper leans on

56 extracted references · 28 canonical work pages · cited by 1 Pith paper

  1. [1]

    Oh, J.-S

    J.-H. Oh, J.-S. Kug, S.-I. An, F.-F. Jin, M.J. McPhaden, J. Shin, Emergent climate change patterns originating from deep ocean warming in climate mitigation scenarios, Nat. Clim. Change 14 (2024) 260–266.https://doi.org/10.1038/s41558-024-01928-0

    work page doi:10.1038/s41558-024-01928-0 2024

  2. [2]

    Deser, F

    C. Deser, F. Lehner, K.B. Rodgers, T.Ault, T.L. Delworth, P.N. DiNezio,A. Fiore, C. Frankignoul, J.C. Fyfe, D.E. Horton, J.E. Kay, R. Knutti, N.S. Lovenduski, J. Marotzke, K.A. McKinnon, S. Minobe, J. Randerson, J.A. Screen, I.R. Simpson, M. Ting, Insights from Earth system model initial-condition large ensembles and future prospects, Nat. Clim. Change 10...

    doi:10.1038/s41558-020-0731-2 2020

  3. [3]

    Gustafsson, T

    N. Gustafsson, T. Janjić, C. Schraff, D. Leuenberger, M. Weissmann, H. Reich, P. Brousseau, T. Montmerle,E.Wattrelot,A.Bučánek,M.Mile,R.Hamdi,M.Lindskog,J.Barkmeijer,M.Dahlbom, B. Macpherson, S. Ballard, G. Inverarity, J. Carley, C. Alexander, D. Dowell, S. Liu, Y. Ikuta, T. Fujita, Survey of data assimilation methods for convective-scale numerical weathe...

    doi:10.1002/qj.3179 2018

  4. [4]

    Choi, D.-H

    Y. Choi, D.-H. Cha, M.-I. Lee, J. Kim, C.-S. Jin, S.-H. Park, M.-S. Joh, Satellite radiance data assimilation for binary tropical cyclone cases over the western North Pacific, J. Adv. Model. Earth Syst.9(2017)832–853.https://doi.org/10.1002/2016MS000826

    work page doi:10.1002/2016ms000826 2017

  5. [5]

    J.R. Eyre, W. Bell, J. Cotton, S.J. English, M. Forsythe, S.B. Healy, E.G. Pavelin, Assimilation of satellite data in numerical weather prediction. Part II: Recent years, Q. J. R. Meteorol. Soc. 148 (2022)521–556.https://doi.org/10.1002/qj.4228

    doi:10.1002/qj.4228 2022

  6. [6]

    Bannister, A review of operational methods of variational and ensemble-variational data assimilation,Q.J.R.Meteorol.Soc.143(2017)607–633.https://doi.org/10.1002/qj.2982

    R.N. Bannister, A review of operational methods of variational and ensemble-variational data assimilation,Q.J.R.Meteorol.Soc.143(2017)607–633.https://doi.org/10.1002/qj.2982

    doi:10.1002/qj.2982 2017

  7. [7]

    Barton, A.J

    P.G.Thoppil,S.Frolov,C.D.Rowley,C.A.Reynolds,G.A.Jacobs,E.JosephMetzger,P.J.Hogan,N. Barton, A.J. Wallcraft, O.M. Smedstad, J.F. Shriver, Ensemble forecasting greatly expands the prediction horizon for ocean mesoscale variability, Commun. Earth Environ. 2 (2021) 1–9. https://doi.org/10.1038/s43247-021-00151-5

    doi:10.1038/s43247-021-00151-5 2021

  8. [8]

    Fairbairn, S.R

    D. Fairbairn, S.R. Pring, A.C. Lorenc, I. Roulstone, A comparison of 4DVar with ensemble data assimilationmethods,Q.J.R.Meteorol.Soc.140(2014)281–294.https://doi.org/10.1002/qj.2135

    work page doi:10.1002/qj.2135 2014

Show all 56 references
  1. [9]

    Z.He,D.Yang,Y.Wang, B.Yin,Impactof4D-VardataassimilationonmodellingoftheEast China Seadynamics,OceanModel.176(2022)102044.https://doi.org/10.1016/j.ocemod.2022.102044

    2022

  2. [10]

    N.Y. Ko, G. Song, W. Youn, I.H. Choi, T.S. Kim, Improvement of Extended Kalman Filter Using Invariant Extended Kalman Filter, in: 2018 18th Int. Conf. ControlAutom. Syst. ICCAS, 2018: pp. 948–950.https://ieeexplore.ieee.org/document/8571929(accessedMay24,2024)

    2018

  3. [11]

    H. He, L. Lei, J.S. Whitaker, Z.-M. Tan, Impacts of Assimilation Frequency on Ensemble Kalman Filter Data Assimilation and Imbalances, J. Adv. Model. Earth Syst. 12 (2020) e2020MS002187. https://doi.org/10.1029/2020MS002187

    2020 doi

  4. [12]

    Part I: Observing System Simulation Experiment, Mon

    X.Wang,D.M.Barker,C.Snyder,T.M.Hamill,AHybridETKF–3DVARDataAssimilationScheme for the WRF Model. Part I: Observing System Simulation Experiment, Mon. Weather Rev. 136 (2008)5116–5131.https://doi.org/10.1175/2008MWR2444.1

    2008 doi

  5. [13]

    Moradkhani, K.-L

    H. Moradkhani, K.-L. Hsu, H. Gupta, S. Sorooshian, Uncertainty assessment of hydrologic model states and parameters: Sequential data assimilation using the particle filter, Water Resour. Res. 41 (2005).https://doi.org/10.1029/2004WR003604

    2005 doi

  6. [14]

    Hoteit, D.-T

    I. Hoteit, D.-T. Pham, G. Triantafyllou, G. Korres, A New Approximate Solution of the Optimal Nonlinear Filter for Data Assimilation in Meteorology and Oceanography, Mon. Weather Rev. 136 26 (2008)317–334.https://doi.org/10.1175/2007MWR1927.1

    2008 doi

  7. [15]

    Evensen, The Ensemble Kalman Filter: theoretical formulation and practical implementation, OceanDyn.53(2003)343–367.https://doi.org/10.1007/s10236-003-0036-9

    G. Evensen, The Ensemble Kalman Filter: theoretical formulation and practical implementation, OceanDyn.53(2003)343–367.https://doi.org/10.1007/s10236-003-0036-9

    2003 doi

  8. [16]

    de Moraes, H

    R.J. de Moraes, H. Hajibeygi, J.D. Jansen, A multiscale method for data assimilation, Comput. Geosci.24(2020)425–442.https://doi.org/10.1007/s10596-019-09839-2

    2020 doi

  9. [17]

    Gettelman, A.J

    A. Gettelman, A.J. Geer, R.M. Forbes, G.R. Carmichael, G. Feingold, D.J. Posselt, G.L. Stephens, S.C. van den Heever,A.C. Varble, P. Zuidema, The future of Earth system prediction:Advances in model-datafusion,Sci.Adv.8(2022)eabn3488.https://doi.org/10.1126/sciadv.abn3488

    2022 doi

  10. [18]

    Bonavita, P

    M. Bonavita, P. Laloyaux, Machine Learning for Model Error Inference and Correction, J. Adv. Model.EarthSyst.12(2020)e2020MS002232.https://doi.org/10.1029/2020MS002232

    2020 doi

  11. [19]

    Huang, H

    L. Huang, H. Leng, X. Li, K. Ren, J. Song, D. Wang, A Data-Driven Method for Hybrid Data Assimilation with Multilayer Perceptron, Big Data Res. 23 (2021) 100179. https://doi.org/10.1016/j.bdr.2020.100179

    2021

  12. [20]

    S.Wang, C.Li,R.Wang, Z.Liu,M.Wang,H.Tan,Y.Wu,X.Liu,H.Sun,R.Yang, X.Liu,J.Chen, H. Zhou, I. Ben Ayed, H. Zheng, Annotation-efficient deep learning for automatic medical image segmentation,Nat.Commun.12(2021)5915.https://doi.org/10.1038/s41467-021-26216-9

    2021 doi

  13. [21]

    Lauriola,A

    I. Lauriola,A. Lavelli, F.Aiolli,An introduction to Deep Learning in Natural Language Processing: Models, techniques, and tools, Neurocomputing 470 (2022) 443–456. https://doi.org/10.1016/j.neucom.2021.05.103

    2022 doi

  14. [22]

    S.Frolov,T.Hinz,F.Raue,J.Hees,A.Dengel,Adversarialtext-to-imagesynthesis:Areview,Neural Netw.144(2021)187–209.https://doi.org/10.1016/j.neunet.2021.07.019

    2021 doi

  15. [23]

    Reichstein, G

    M. Reichstein, G. Camps-Valls, B. Stevens, M. Jung, J. Denzler, N. Carvalhais, Prabhat, Deep learning and process understanding for data-driven Earth system science, Nature 566 (2019) 195–204.https://doi.org/10.1038/s41586-019-0912-1

    2019 doi

  16. [24]

    Zheng, X

    G. Zheng, X. Li, R.-H. Zhang, B. Liu, Purely satellite data–driven deep learning forecast of complicated tropical instability waves, Sci. Adv. 6 (2020) eaba1482. https://doi.org/10.1126/sciadv.aba1482

    2020 doi

  17. [25]

    Zheng, R

    Q. Zheng, R. Wang, G. Han, W. Li, X. Wang, Q. Shao, X. Wu, L. Cao, G. Zhou, S. Hu, A Spatiotemporal MultiscaleDeep LearningModel forSubseasonalPredictionofArctic Sea Ice,IEEE Trans.Geosci.RemoteSens.62(2024)1–22.https://doi.org/10.1109/TGRS.2024.3355238

    2024

  18. [26]

    L. Li, R. Carver, I. Lopez-Gomez, F. Sha, J. Anderson, Generative emulation of weather forecast ensembles with diffusion models, Sci. Adv. 10 (n.d.) eadk4489. https://doi.org/10.1126/sciadv.adk4489

    doi

  19. [27]

    Bocquet, Surrogate modeling for the climate sciences dynamics with machine learning and data assimilation,Front.Appl.Math.Stat.9(2023).https://doi.org/10.3389/fams.2023.1133226

    M. Bocquet, Surrogate modeling for the climate sciences dynamics with machine learning and data assimilation,Front.Appl.Math.Stat.9(2023).https://doi.org/10.3389/fams.2023.1133226

    2023

  20. [28]

    S.Cheng,C.Quilodrán-Casas,S.Ouala,A.Farchi,C.Liu,P.Tandeo,R.Fablet,D.Lucor,B.Iooss,J. Brajard,D.Xiao,T.Janjic,W.Ding,Y.Guo,A.Carrassi,M.Bocquet,R.Arcucci,MachineLearning With Data Assimilation and Uncertainty Quantification for Dynamical Systems: A Review, IEEECAAJ.Autom.Sin.1...

    2023

  21. [29]

    Maddy,K.Ide,S.Boukabara,ADeep-Learning-Based Microwave Radiative Transfer Emulator for Data Assimilation and Remote Sensing, IEEE J

    X.Liang,K.Garrett,Q.Liu,E.S. Maddy,K.Ide,S.Boukabara,ADeep-Learning-Based Microwave Radiative Transfer Emulator for Data Assimilation and Remote Sensing, IEEE J. Sel. Top. Appl. EarthObs.RemoteSens.15(2022)8819–8833.https://doi.org/10.1109/JSTARS.2022.3210491

    2022

  22. [30]

    Zheng, W

    Q. Zheng, W. Li, Q. Shao, G. Han, X. Wang, AMid- and Long-TermArctic Sea Ice Concentration Prediction Model Based on Deep Learning Technology, Remote Sens. 14 (2022) 2889. 27 https://doi.org/10.3390/rs14122889

    2022 doi

  23. [31]

    de Almeida, H.F

    V.A. de Almeida, H.F. de Campos Velho, G.B. França, N.F.F. Ebecken, Neural networks for data assimilation of surface and upper-air data in Rio de Janeiro, Geosci. Model Dev. Discuss. (2022) 1–23.https://doi.org/10.5194/gmd-2022-50

    2022 doi

  24. [32]

    Laloyaux, T

    P. Laloyaux, T. Kurth, P.D. Dueben, D. Hall, Deep Learning to Estimate Model Biases in an Operational NWP Assimilation System, J. Adv. Model. Earth Syst. 14 (2022) e2022MS003016. https://doi.org/10.1029/2022MS003016

    2022 doi

  25. [33]

    Hatfield, M

    S. Hatfield, M. Chantry, P. Dueben, P. Lopez, A. Geer, T. Palmer, Building Tangent-Linear and Adjoint Models for DataAssimilation With Neural Networks, J.Adv. Model. Earth Syst. 13 (2021) e2021MS002521.https://doi.org/10.1029/2021MS002521

    2021 doi

  26. [34]

    P.Wu,X.Chang,W.Yuan,J.Sun,W.Zhang,R.Arcucci,Y.Guo,Fastdataassimilation(FDA):Data assimilation by machine learning for faster optimize model state, J. Comput. Sci. 51(2021) 101323. https://doi.org/10.1016/j.jocs.2021.101323

    2021

  27. [35]

    R.Arcucci, J. Zhu, S. Hu,Y.-K. Guo, Deep DataAssimilation: Integrating Deep Learning with Data Assimilation,Appl.Sci.11(2021)1114.https://doi.org/10.3390/app11031114

    2021 doi

  28. [36]

    W. Wang, K. Ren, B. Duan, J. Zhu, X. Li, W. Ni, J. Lu, T. Yuan, A Four-Dimensional Variational ConstrainedNeuralNetwork-BasedDataAssimilationMethod,J.Adv.Model. EarthSyst. 16(2024) e2023MS003687.https://doi.org/10.1029/2023MS003687

    2024 doi

  29. [37]

    Kadow, D.M

    C. Kadow, D.M. Hall, U. Ulbrich, Artificial intelligence reconstructs missing climate information, Nat.Geosci.13(2020)408–413.https://doi.org/10.1038/s41561-020-0582-5

    2020 doi

  30. [38]

    H. Lin, J. Tang, S. Wang, S. Wang, G. Dong, Deep learning downscaled high-resolution daily near surface meteorological datasets over East Asia, Sci. Data 10 (2023) 890. https://doi.org/10.1038/s41597-023-02805-9

    2023 doi

  31. [39]

    Fablet, B

    R. Fablet, B. Chapron, L. Drumetz, E. Mémin, O. Pannekoucke, F. Rousseau, Learning Variational Data Assimilation Models and Solvers, J. Adv. Model. Earth Syst. 13 (2021) e2021MS002572. https://doi.org/10.1029/2021MS002572

    2021 doi

  32. [40]

    Melinc, Ž

    B. Melinc, Ž. Zaplotnik, 3D-Var Data Assimilation using a Variational Autoencoder, Q. J. R. Meteorol.Soc.150(2024)2273–2295.https://doi.org/10.1002/qj.4708

    2024 doi

  33. [41]

    Robert, S

    C. Robert, S. Durbiano, E. Blayo, J. Verron, J. Blum, F.-X. Le Dimet,Areduced-order strategy for 4D-Var data assimilation, J. Mar. Syst. 57 (2005) 70–82. https://doi.org/10.1016/j.jmarsys.2005.04.003

    2005 doi

  34. [42]

    Cheng, M

    H. Cheng, M. Jardak, M. Alexe, A. Sandu, A hybrid approach to estimating error covariances in variational data assimilation, Tellus A 62 (2010) 288–297. https://doi.org/10.1111/j.1600-0870.2010.00442.x

    2010

  35. [43]

    Peyron, A

    M. Peyron, A. Fillion, S. Gürol, V. Marchais, S. Gratton, P. Boudier, G. Goret, Latent space data assimilation by using deep learning, Q. J. R. Meteorol. Soc. 147 (2021) 3759–3777. https://doi.org/10.1002/qj.4153

    2021 doi

  36. [44]

    W. Cai, A. Santoso, M. Collins, B. Dewitte, C. Karamperidou, J.-S. Kug, M. Lengaigne, M.J. McPhaden, M.F. Stuecker,A.S. Taschetto,A. Timmermann, L. Wu, S.-W. Yeh, G. Wang, B. Ng, F. Jia,Y.Yang, J.Ying, X.-T.Zheng,T.Bayr,J.R. Brown,A. Capotondi,K.M. Cobb, B. Gan,T.Geng, Y.-G. H...

    2021 doi

  37. [45]

    Cinelli, M.A

    L.PinheiroCinelli,M.AraújoMarins,E.A.BarrosdaSilva,S.LimaNetto,VariationalAutoencoder, 28 in: L.P. Cinelli, M.A. Marins, E.A. Barros da Silva, S.L. Netto (Eds.), Var. Methods Mach. Learn. Appl. Deep Netw., Springer International Publishing, Cham, 2021: pp. 111–149. https://doi...

    2021 doi

  38. [46]

    Hebert, C

    S.Woo,J.Park,J.-Y.Lee,I.S.Kweon,CBAM:ConvolutionalBlockAttentionModule,in:V.Ferrari, M. Hebert, C. Sminchisescu, Y. Weiss (Eds.), Comput. Vis. – ECCV 2018, Springer International Publishing,Cham,2018:pp.3–19.https://doi.org/10.1007/978-3-030-01234-2_1

    2018 doi

  39. [47]

    G. Liu, J. Wang, Dendrite Net: A White-Box Module for Classification, Regression, and System Identification, IEEE Trans. Cybern. 52 (2022) 13774–13787. https://doi.org/10.1109/TCYB.2021.3124328

    2022

  40. [48]

    H.Hersbach,B.Bell,P.Berrisford,S.Hirahara,A.Horányi,J.Muñoz-Sabater,J.Nicolas,C.Peubey, R. Radu, D. Schepers, A. Simmons, C. Soci, S. Abdalla, X. Abellan, G. Balsamo, P. Bechtold, G. Biavati, J. Bidlot, M. Bonavita, G. De Chiara, P. Dahlgren, D. Dee, M. Diamantakis, R. Dragani...

    2020

  41. [49]

    Decremer, A

    S.J.Johnson,T.N.Stockdale,L.Ferranti,M.A.Balmaseda,F.Molteni,L.Magnusson,S.Tietsche,D. Decremer, A. Weisheimer, G. Balsamo, S.P.E. Keeley, K. Mogensen, H. Zuo, B.M. Monge-Sanz, SEAS5: the new ECMWF seasonal forecast system, Geosci. Model Dev. 12 (2019) 1087–1117. https://doi.o...

    2019 doi

  42. [50]

    Sci.10(2020)1073.https://doi.org/10.3390/app10031073

    D.Yi,J.Ahn,S.Ji,AnEffectiveOptimizationMethodforMachineLearningBasedonADAM,Appl. Sci.10(2020)1073.https://doi.org/10.3390/app10031073

    2020 doi

  43. [51]

    Castro, A

    E. Castro, A. Godavarthi, J. Rubinfien, K. Givechian, D. Bhaskar, S. Krishnaswamy, Transformer-basedproteingenerationwithregularizedlatentspaceoptimization,Nat.Mach.Intell.4 (2022)840–851.https://doi.org/10.1038/s42256-022-00532-1

    2022 doi

  44. [52]

    A.Hannachi,I.T.Jolliffe,D.B.Stephenson,Empiricalorthogonalfunctionsandrelatedtechniquesin atmospheric science: A review, Int. J. Climatol. 27 (2007) 1119–1152. https://doi.org/10.1002/joc.1499

    2007 doi

  45. [53]

    McPhaden, Genesis and Evolution of the 1997-98 El Niño, Science 283 (1999) 950–954

    M.J. McPhaden, Genesis and Evolution of the 1997-98 El Niño, Science 283 (1999) 950–954. https://doi.org/10.1126/science.283.5404.950

    1999 doi

  46. [54]

    S. Hu, A.V. Fedorov, The extreme El Niño of 2015–2016 and the end of global warming hiatus, Geophys.Res.Lett.44(2017)3816–3824.https://doi.org/10.1002/2017GL072908

    2017 doi

  47. [55]

    Paek, J.-Y

    H. Paek, J.-Y. Yu, C. Qian, Why were the 2015/2016 and 1997/1998 extreme El Niños different?, Geophys.Res.Lett.44(2017)1848–1856.https://doi.org/10.1002/2016GL071515

    2017 doi

  48. [56]

    L. Chen, T. Li, B. Wang, L. Wang, Formation Mechanism for 2015/16 Super El Niño, Sci. Rep. 7 (2017)2975.https://doi.org/10.1038/s41598-017-02926-3. 1 Supporting Information for Generating Unseen Nonlinear Evolution in Sea Surface Temperature Using a Deep Learning-Based Latent ...

    2017 doi

Pith tools

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