Pith. sign in

REVIEW 5 major objections 6 minor 4 cited by

GEN2: A Generative Prediction-Correction Framework for Long-time Emulations of Spatially-Resolved Climate Extremes

T0 review · 5 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A two-step generative emulator, trained on one high-warming climate realization, claims to reproduce the statistics of climate extremes under different emissions scenarios without additional expensive runs.

desk verdict A solid cross-scenario climate emulator with a genuine out-of-sample test; the strongest claim is narrower than the abstract suggests, but it deserves peer review. read the letter →

arxiv 2508.15196 v1 pith:7QHTUIRY submitted 2025-08-21 physics.comp-ph cs.LGnlin.CD

Mengze Wang , Benedikt Barthel Sorensen , Themistoklis Sapsis This is my paper
classification physics.comp-phcs.LGnlin.CD
keywords climateemulationgenerativemodelextremeeventsdiffusionnudgingPCApatternscalingscenarioextrapolation
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 sets out to establish that a cheap, two-step generative model can replace large ensembles of Earth-system simulations for the specific purpose of quantifying how extreme weather statistics change across emissions scenarios. The model converts a single scalar input—the trajectory of global mean temperature—into globally resolved fields of wind, temperature, and humidity, with an inexpensive Gaussian emulator producing the large-scale statistics and a diffusion-model debiasing step restoring the non-Gaussian tails. The central demonstration is extrapolation: trained on one realization of one strong-warming scenario, GEN2 is shown to reproduce the ensemble extreme-event statistics of that scenario and of three scenarios the training never saw, with temperature-tail errors under about 0.5 K. If that holds, long-run extreme-event risk assessment—which currently requires thousands of expensive model years—can be done orders of magnitude more cheaply, making it feasible to explore many emissions futures and to generate the large ensembles rare events demand.

What carries the argument

The key mechanism is the pairing of a conditional Gaussian emulator with a nudged diffusion debiaser. The emulator projects fluctuations onto 500 PCA modes, models each mode's seasonal mean and variance as linear regressions on seasonal global mean temperature, and simulates daily fluctuations as a multivariate autoregressive Gaussian process; this is cheap, stable, and scenario-extrapolating, but Gaussian and truncated. The debiaser is a conditional score-based diffusion model (U-Net backbone) that turns emulated snapshots into reference-like snapshots. The enabling trick is nudging: because the free-running emulator diverges chaotically from the reference, training pairs are instead genera

What would settle it

Run GEN2 on an emissions scenario with a global-mean-temperature overshoot—rise, peak, then decline—and compare its predicted 2090-2099 tail quantiles against a multi-member ensemble of that scenario. If the 97.5% temperature quantile RMSE grows beyond the roughly 0.5 K level the paper reports for its extrapolation tests, or if the spatial pattern of the tails is the training scenario's pattern rather than the overshoot scenario's, the linear-in-global-temperature and scenario-invariant assumptions are falsified.

Watch

Extended reading notes

Core claim

On the paper's terms, the discovery is that the intractable full climate trajectory can be split into a cheap Gaussian part and a learned non-Gaussian part, and that the correction learned between nudged trajectories transfers to free-running and out-of-sample scenarios. Concretely, after training on 1979-2018 reanalysis data, the model reproduces single-point statistics (variance, 97.5% quantile, skewness, kurtosis), spatial two-point correlations, cross-variable correlations (temperature-humidity, zonal-meridional wind), and the Wheeler-Kiladis spectrum of zonal wind. Then, trained on a single realization of one high-warming scenario, it reproduces the 10-member ensemble statistics of temp

Load-bearing premise

The model assumes that the way a region's climate responds to warming is fixed by a linear relationship between global mean temperature and the seasonal mean and spread of the leading 500 spatial patterns, with those patterns themselves unchanged in a different emissions scenario.

Editorial extensions

If this is right

  • A new emissions scenario can be evaluated at roughly 720 simulated years per day using only its global mean temperature trajectory, without retraining.
  • The one-member-to-ensemble extrapolation means a single transient run can stand in for a multi-member ensemble when computing tail statistics, if the underlying system is ergodic.
  • Because the debiased output is a joint field of temperature, humidity, and winds, it preserves cross-variable correlations, so compound extremes such as high temperature with low humidity can be quantified in scenarios never seen during training.
  • The same two-step architecture could be retargeted to other coarse-resolution climate variables, with the Gaussian emulator's regressions customized to whatever statistics are available.
  • Feeding GEN2 output into existing downscaling techniques extends the extrapolation from roughly 100-km grid statistics to more localized extremes.

Reading between the lines

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

  • The linear-in-global-temperature assumption would face its hardest test on an overshoot trajectory in which global mean temperature returns to earlier values while regional patterns may not; reproducing the return-path extremes would be a stronger confirmation than the one-scenario-to-another tests reported here.
  • Because debiasing is applied snapshot-by-snapshot, temporal extreme statistics—heatwave duration, storm persistence, consecutive dry days—are the most likely place for the model to fail; a video-diffusion extension is the natural experiment to check.
  • Precipitation is the conspicuous absent variable: its tails are more intermittent and non-Gaussian than temperature, so demonstrating the same cross-scenario tail accuracy for rain would materially extend the claim beyond the present four-variable set.
  • The current scenario transfer is within one Earth-system model's internal logic; the more demanding claim, that one learned debiaser transfers across structurally different models, remains open and would be a direct test of whether the corrected statistics are model-agnostic.
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

5 major / 6 minor

Summary. The paper proposes GEN2, a two-stage generative emulator for spatially resolved climate extremes. Stage 1 is a conditional Gaussian emulator that projects U, V, T, and Q onto the first 500 PCA modes, models the seasonal mean and variance of the coefficients as linear functions of global mean temperature, and represents daily fluctuations by a vector autoregressive process with matched time-lagged covariances. Stage 2 is a conditional score-based diffusion model trained on pairs of nudged emulator trajectories and reference data to correct non-Gaussian biases and tail statistics. The model is first validated on ERA5 1979–2018 and then applied to MPI-ESM1-2-LR CMIP6 data, training on one SSP585 realization and evaluating on 10 ensemble members across SSP126, SSP245, SSP370, and SSP585. Reported results show accurate single-point statistics, spatial and cross-variable correlations, Wheeler–Kiladis spectra, and RMSE below 0.5 K for the 97.5% temperature quantile in scenarios unseen during training.

Significance. If the cross-scenario extrapolation claim holds, GEN2 offers a computationally cheap alternative to large ESM ensembles for extreme-event risk assessment, with a reported 720 simulated years per day versus roughly 40 for ESMs. The framework is modular, builds transparently on prior work, and is described in sufficient detail to be reproduced. Strengths include the explicit statement of the modeling assumptions in Appendix A, the use of publicly available ERA5 and CMIP6 data, the inclusion of algorithmic pseudocode, and the systematic comparison of the Gaussian emulator with the full GEN2 model, which demonstrates that the diffusion debiasing step improves skewness, kurtosis, and quantile statistics. The main weakness is that the most consequential evidence—cross-scenario extrapolation—is currently supported by a narrow set of diagnostics, while the ERA5 validation is in-sample.

major comments (5)
  1. [§2.1, Figs. 2–3] The ERA5 validation is performed on the same 1979–2018 period used for training. This is an in-sample test, so the reported accuracy of standard deviation, quantiles, skewness, correlations, and the Wheeler–Kiladis spectrum may reflect memorization of the training distribution rather than genuine predictive skill. Please add a temporal out-of-sample evaluation, e.g., train on 1979–2000 and validate on 2001–2018, or use cross-validation, and report the same metrics for the held-out period.
  2. [§2.2, Fig. 5 and Tables D1–D4] The central cross-scenario claim is supported almost entirely by the global RMSE of the 97.5% temperature quantile (Fig. 5c). No quantitative cross-scenario metrics are reported for U, V, or Q, and no cross-scenario tail metrics (e.g., skewness, kurtosis, 99th percentile, joint T–Q statistics) are given for any variable. Tables D1–D4 are ERA5-only. Please provide per-scenario error tables analogous to D1–D4 for SSP126, SSP245, and SSP370, including all four variables and multiple percentile/statistics levels, so that the claim of accurate extreme-event statistics in unseen scenarios can be evaluated.
  3. [Appendix A, bullets 1–3 and Eq. (A10)] The extrapolation rests on three invariance assumptions: the PCA basis, climatological mean, and global standard deviations are scenario-invariant, and the seasonal mean/variance depend linearly on global mean temperature with coefficients fitted on one SSP585 realization. Figure A1 illustrates linearity for only two PCA coefficients and uses the CNRM-CM6-1-HR dataset, not the MPI-ESM1-2-LR model used in the main cross-scenario test. Please provide evidence for MPI-ESM1-2-LR that the regression slopes and PCA bases are consistent across scenarios, or quantify the sensitivity of the final statistics to violations of these assumptions.
  4. [§2.2, Fig. 5c] The reported RMSE (<0.5 K) is for the full GEN2 output, which conflates the Gaussian emulator's scenario mapping (Eq. A10) with the diffusion debiasing step trained only on nudged SSP585 pairs. A biased emulator for SSP126/245/370 could in principle be partially corrected, or partially masked, by a debiaser that was not trained on such biases. Please report the Gaussian-emulator-only error and the full GEN2 error for each unseen scenario and variable, so that the extrapolation behavior of each component can be separated.
  5. [§2.2, Figs. 4–5] For the SSP585 within-scenario evaluation, the 10 reference members include the single member used for training. This makes the within-scenario comparison more favorable than a true leave-one-out test. Please report the SSP585 metrics with the training member excluded, or use a leave-one-out procedure, to avoid inflating the apparent skill.
minor comments (6)
  1. [Throughout] Typos and inconsistent labels: 'balooned' (Introduction), 'traininig' (§2.1), 'Wienner' (Appendix B), 'Gassuain' (Table D1 caption), 'SSP125' in Fig. 5b versus SSP126, '1.5◦' should be '1.5°', and 'ϕ ∈ [−π/2, π,2]' in Appendix A should be 'ϕ ∈ [−π/2, π/2]'.
  2. [Eq. (A17) and surrounding text] The notation '˙ν = ˙ˆη − 1/τ(ν − η)' is unclear because ˆη is a discrete-time process and its time derivative is not defined in the main text. Please define the intended continuous-time interpretation or rewrite the nudging equation in the discrete-time setting used in the emulator.
  3. [§4.2] The sentence 'the nudging term is an order of magnitude than the other terms' is missing a comparative phrase; clarify whether it is intended to be larger or smaller.
  4. [Appendix C] The notation for coordinates is inconsistent: earlier θ is latitude and φ is longitude, but Appendix C writes 'θ, ϕ, t' with 'ϕ ∈ [−15°, 15°]'. Please align the notation throughout.
  5. [References] Reference [48] is incomplete (no journal or arXiv identifier), and reference [45] is the authors' prior work that this paper extends; please provide complete bibliographic information and clarify the incremental contribution relative to [45] and [48].
  6. [Fig. 5c] Please add confidence intervals or a comparison with the internal variability of the 10-member ensemble for the RMSE curves. Without such context it is unclear whether an RMSE below 0.5 K is statistically distinguishable from zero or from the ensemble spread.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cross-scenario evaluation is an out-of-sample test against held-out CMIP6 realizations.

full rationale

The paper's central claim is tested on 10 realizations of unseen emission scenarios while training uses one SSP585 realization. The Gaussian emulator's linear regressions (Eq. A10) are fitted to training data, but the predicted extreme-event statistics are emergent outputs of the fitted stochastic process plus a diffusion debiaser, and are evaluated against held-out reference fields. No fitted parameter is renamed as a prediction, and no self-citation supplies the load-bearing conclusion. The nudging framework from Barthel Sorensen et al. [45] is a methodological citation, not an imported uniqueness theorem; the paper explicitly notes alternative generative models (flow matching, stochastic interpolants) could replace the diffusion model, indicating the choice is not forced. The linear-in-Tg and scenario-invariance assumptions (Appendix A, bullets 1-3) are clearly stated modeling assumptions with external support ([31,32]) and an independent illustration (Fig. A1 using CNRM-CM6-1-HR), so they create an extrapolation risk but do not make the derivation circular. The historical ERA5 validation is in-sample, which is a standard fit-quality assessment rather than a disguised prediction. The acknowledged limitations (per-snapshot debiasing, no downscaling) further confirm that the claims are not tautological: the framework could fail if the scenario-invariance or linearity assumptions break, and the authors do not hide that dependence.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The extrapolation claim is carried by a set of explicit but strong assumptions, not by physical equations. The model has many fitted parameters (PCA decomposition, linear response functions, AR matrices, diffusion weights) and several user-chosen hyperparameters (500 modes, tau=6h). The central test is on the same ESM used for training, so the ledger is dominated by assumptions of scenario invariance rather than independent physical support.

free parameters (7)
  • PCA basis, mode shapes phi_i, eigenvalues lambda_i, and per-variable global standard deviations sigma_g,k = computed from training data (ERA5 1979-2018; MPI historical+SSP585 member)
    The spatial structure of the emulator; assumed fixed across scenarios.
  • Linear regression coefficients p_s,i,0, p_s,i,1 and q_s,i,0, q_s,i,1 for seasonal mean and variance vs Tg = least-squares per season and PCA mode (Eq. A10)
    Carries the climate-change signal; if the linear map is wrong, extrapolation fails.
  • Autoregressive coefficient matrices Psi_1..Psi_M and noise covariance R = Yule-Walker solutions (Eqs. A13-A14)
    Controls temporal correlations of daily fluctuations in the Gaussian emulator.
  • Number of retained PCA modes I=500 = 500 (79.6% ERA5 variance, 78.2% MPI variance)
    Truncation level; the omitted 20% variance is left to the diffusion correction.
  • Nudging timescale tau = 6 hours
    User-chosen free parameter separating slow and fast dynamics; inherited from prior work [45].
  • Autoregressive order M
    Maximum time lag in the AR process; not stated in the manuscript, so replication requires guessing.
  • Diffusion model hyperparameters and trained weights = sigma_min=0.01, sigma_max from data, Adam lr=2e-4, batch=8, 200 epochs
    The learned debiasing map is the second stage; its weights are fit to nudged-reference pairs.
assumptions (6)
  • domain assumption The PCA basis computed from the training period remains an efficient basis for other future climate scenarios.
    Appendix A, bullet 1; essential for using the same spatial modes in unseen SSP scenarios.
  • domain assumption The seasonal mean and variance of PCA coefficients vary linearly with global mean temperature.
    Appendix A.2.2, Eq. A10; fitted in one scenario and assumed to hold across all scenarios and ensemble members.
  • domain assumption The statistics of daily fluctuations, given season, are independent of the year and climate scenario.
    Appendix A, bullet 3; this is what allows the model trained on one realization to generate statistics for other scenarios.
  • domain assumption The climatological mean, global standard deviation, and PCA mode shapes are unchanged with time or future scenario.
    Appendix A.1, after Eq. A7; the model only emulates PCA coefficients, so the basis and climatology must transfer.
  • ad hoc to paper A diffusion model trained on nudged emulator-reference pairs generalizes to free-running emulator outputs.
    Section 4.2 and Appendix A.3; the nudging feedback creates a distribution mismatch that is only partially corrected by a mean/variance rescale.
  • domain assumption Daily fluctuations are Gaussian with finite-order autoregressive dynamics.
    Appendix A.2.2, Eq. A12; non-Gaussianity is intentionally excluded from the emulator and assigned to the ML correction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GEN2: A Generative Prediction-Correction Framework for Long-time Emulations of Spatially-Resolved Climate Extremes." pith.science (2026). https://pith.science/paper/7QHTUIRY

@misc{pith2026250815196,
  author       = {Pith},
  title        = {Pith review of: GEN2: A Generative Prediction-Correction Framework for Long-time Emulations of Spatially-Resolved Climate Extremes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7QHTUIRY}},
  note         = {Machine review of arXiv:2508.15196}
}
read the original abstract

Accurately quantifying the increased risks of climate extremes requires generating large ensembles of climate realization across a wide range of emissions scenarios, which is computationally challenging for conventional Earth System Models. We propose GEN2, a generative prediction-correction framework for an efficient and accurate forecast of the extreme event statistics. The prediction step is constructed as a conditional Gaussian emulator, followed by a non-Gaussian machine-learning (ML) correction step. The ML model is trained on pairs of the reference data and the emulated fields nudged towards the reference, to ensure the training is robust to chaos. We first validate the accuracy of our model on historical ERA5 data and then demonstrate the extrapolation capabilities on various future climate change scenarios. When trained on a single realization of one warming scenario, our model accurately predicts the statistics of extreme events in different scenarios, successfully extrapolating beyond the distribution of training data.

Discussion (0). Sign in to comment.

Forward citations

Cited by 4 Pith papers

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

  1. Optimal scenario design for climate emulation

    physics.ao-ph 2026-06 unverdicted novelty 7.0 of 10

    Optimizing training data via a differentiable SCM yields climate emulators that outperform those trained on six standard ScenarioMIP pathways while using less data and isolating distinct forcing responses.

  2. Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction

    physics.ao-ph 2026-06 unverdicted novelty 6.0 of 10

    Online conformal prediction post-processing guarantees calibrated uncertainty coverage for GenCast, NeuralGCM, and AIFS-ENS forecasts of temperature and precipitation including extremes.

  3. Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    Tyche achieves competitive probabilistic weather forecasting skill and calibration using a single-step flow model with JVP-regularized training and rollout finetuning.

  4. Score-based generative emulation of impact-relevant Earth system model outputs

    physics.ao-ph 2025-10 conditional novelty 5.0 of 10

    A score-based generative emulator on a spherical mesh produces joint distributions of impact-relevant climate variables that closely match three parent ESMs in pre-industrial and forced regimes, with errors small rela...

Reference graph

Works this paper leans on

70 extracted references · 47 canonical work pages · cited by 4 Pith papers

  1. [1]

    Nature Climate Change 10(7), 611–621 (2020) https://doi.org/ 10.1038/s41558-020-0790-4

    Raymond, C., Horton, R.M., Zscheischler, J., Martius, O., AghaKouchak, A., Balch, J., Bowen, S.G., Camargo, S.J., Hess, J., Kornhuber, K., Oppenheimer, M., Ruane, A.C., Wahl, T., White, K.: Understanding and managing connected extreme events. Nature Climate Change 10(7), 611–621 (2020) https://doi.org/ 10.1038/s41558-020-0790-4

    work page doi:10.1038/s41558-020-0790-4 2020

  2. [2]

    npj Climate and Atmospheric Science 4(1), 1–4 (2021) https://doi.org/10.1038/ s41612-021-00202-w

    Robinson, A., Lehmann, J., Barriopedro, D., Rahmstorf, S., Coumou, D.: Increas- ing heat and rainfall extremes now far outside the historical climate. npj Climate and Atmospheric Science 4(1), 1–4 (2021) https://doi.org/10.1038/ s41612-021-00202-w

    work page 2021

  3. [3]

    Nature Climate Change 11(8), 689–695 (2021) https://doi.org/ 10.1038/s41558-021-01092-9

    Fischer, E.M., Sippel, S., Knutti, R.: Increasing probability of record-shattering climate extremes. Nature Climate Change 11(8), 689–695 (2021) https://doi.org/ 10.1038/s41558-021-01092-9

    work page doi:10.1038/s41558-021-01092-9 2021

  4. [4]

    Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change

    Allen, S., Barros, V., (Canada, I., (UK, D., Cardona, O., Cutter, S., Dube, O.P., Ebi, K., (USA, C., Handmer, J., (Australia, P., Lavell, A., (USA, K., Mastrandrea, M., Mcbean, G., Mechler, R., (UK, T., Nicholls, N., (Norway, K., (USA, T.: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. Special Report of Working Gr...

    work page arXiv 2012

  5. [5]

    Columbia University Press, ??? (2015)

    Houser, T., Hsiang, S., Kopp, R., Larsen, K., Delgado, M., Jina, A., Mastran- drea, M., Mohan, S., Muir-Wood, R., Rasmussen, D.J., Rising, J., Wilson, P., Fisher-Vanden, K., Greenstone, M., Heal, G., Oppenheimer, M., Stern, N., Ward, B., Bloomberg, M.R., Paulson, H.M., Steyer, T.F.: Economic Risks of Climate Change: An American Prospectus. Columbia Univer...

    work page doi:10.7312/hous17456 2015

  6. [6]

    AON: 2025 Climate & catastrophe insight (2025)

    work page 2025

  7. [7]

    Nature Climate Change 11(2), 87–94 (2021) https://doi.org/10.1038/ s41558-020-00984-6

    Fiedler, T., Pitman, A.J., Mackenzie, K., Wood, N., Jakob, C., Perkins- Kirkpatrick, S.E.: Business risk and the emergence of climate analyt- ics. Nature Climate Change 11(2), 87–94 (2021) https://doi.org/10.1038/ s41558-020-00984-6

    work page 2021

  8. [8]

    the basic experiment

    Smagorinsky, J.: General circulation experiments with the primitive equations: I. the basic experiment. Monthly Weather Review 91(3), 99–164 (1963) https: //doi.org/10.1175/1520-0493(1963)091⟨0099:GCEWTP⟩2.3.CO;2

    doi:10.1175/1520-0493(1963)091 1963

Show all 70 references
  1. [9]

    Monthly Weather Review 93(12), 727–768 (1965) https://doi.org/10.1175/ 1520-0493(1965)093⟨0727:NRF ANL⟩2.3.CO;2 34

    Smagorinsky, J., Manabe, S., Holloway, J.L.: NUMERICAL RESULTS FROM A NINE-LEVEL GENERAL CIRCULATION MODEL OF THE ATMOSPHERE. Monthly Weather Review 93(12), 727–768 (1965) https://doi.org/10.1175/ 1520-0493(1965)093⟨0727:NRF ANL⟩2.3.CO;2 34

    1965

  2. [10]

    Monthly Weather Review 93(12), 769–798 (1965) https://doi.org/10.1175/ 1520-0493(1965)093⟨0769:SCOAGC⟩2.3.CO;2

    Manabe, S., Smagorinsky, J., Strickler, R.F.: SIMULATED CLIMATOLOGY OF A GENERAL CIRCULATION MODEL WITH A HYDROLOGIC CYCLE. Monthly Weather Review 93(12), 769–798 (1965) https://doi.org/10.1175/ 1520-0493(1965)093⟨0769:SCOAGC⟩2.3.CO;2

    1965

  3. [11]

    Mintz, Y.: Very Long-Term Global Integration of the Primitive Equations of Atmospheric Motion: An Experiment in Climate Simulation. In: Billings, D.E., Broecker, W.S., Bryson, R.A., Cox, A., Damon, P.E., Donn, W.L., Eriksson, E., Ewing, M., Fletcher, J.O., Hamilton, W., Jerzyk...

    1965

  4. [12]

    In: International Conference on Computational Science, pp

    Taylor, M.A., Cyr, A.S., Fournier, A.: A non-oscillatory advection operator for the compatible spectral element method. In: International Conference on Computational Science, pp. 273–282 (2009). Springer

    2009

  5. [13]

    The International Journal of High Performance Computing Applications 26(1), 74–89 (2012)

    Dennis, J.M., Edwards, J., Evans, K.J., Guba, O., Lauritzen, P.H., Mirin, A.A., St-Cyr, A., Taylor, M.A., Worley, P.H.: Cam-se: A scalable spectral element dynamical core for the community atmosphere model. The International Journal of High Performance Computing Applications 2...

    2012

  6. [14]

    Journal of Advances in Modeling Earth Systems 14(12) (2022)

    Golaz, J.-C., Van Roekel, L.P., Zheng, X., Roberts, A.F., Wolfe, J.D., Lin, W., Bradley, A.M., Tang, Q., Maltrud, M.E., Forsyth, R.M., et al.: The doe e3sm model version 2: overview of the physical model and initial model evaluation. Journal of Advances in Modeling Earth Syste...

    2022

  7. [15]

    Cambridge University Press, Cambridge (2007)

    Stensrud, D.J.: Parameterization Schemes: Keys to Under- standing Numerical Weather Prediction Models. Cambridge University Press, Cambridge (2007). https://doi.org/10.1017/ CBO9780511812590 . https://www.cambridge.org/core/books/parameterization- schemes/C7C8EC8901957314433BE...

    2007

  8. [16]

    Journal of the Atmospheric Sciences 66(6), 1665– 1683 (2009) https://doi.org/10.1175/2008JAS2806.1

    Holloway, C.E., Neelin, J.D.: Moisture Vertical Structure, Column Water Vapor, and Tropical Deep Convection. Journal of the Atmospheric Sciences 66(6), 1665– 1683 (2009) https://doi.org/10.1175/2008JAS2806.1

    2009 doi

  9. [17]

    Proceedings of the National Academy of Sciences 111(9), 35 3280–3285 (2014) https://doi.org/10.1073/pnas.1222477110

    Friend, A.D., Lucht, W., Rademacher, T.T., Keribin, R., Betts, R., Cadule, P., Ciais, P., Clark, D.B., Dankers, R., Falloon, P.D., Ito, A., Kahana, R., Kleidon, A., Lomas, M.R., Nishina, K., Ostberg, S., Pavlick, R., Peylin, P., Schaphoff, S., Vuichard, N., Warszawski, L., Wil...

    2014

  10. [18]

    Proceedings of the National Academy of Sciences 113(5), 1285–1290 (2016) https://doi.org/10.1073/pnas.1515160113

    Bloom, A.A., Exbrayat, J.-F., Velde, I.R., Feng, L., Williams, M.: The decadal state of the terrestrial carbon cycle: Global retrievals of terrestrial carbon alloca- tion, pools, and residence times. Proceedings of the National Academy of Sciences 113(5), 1285–1290 (2016) http...

    2016 doi

  11. [19]

    Proceedings of the National Academy of Sciences 115(39), 9684–9689 (2018) https://doi.org/10.1073/pnas.1810286115

    Rasp, S., Pritchard, M.S., Gentine, P.: Deep learning to represent subgrid pro- cesses in climate models. Proceedings of the National Academy of Sciences 115(39), 9684–9689 (2018) https://doi.org/10.1073/pnas.1810286115

    2018 doi

  12. [20]

    Journal of Advances in Modeling Earth Systems 11(8), 2728–2744 (2019) https://doi.org/10.1029/2019MS001711

    Brenowitz, N.D., Bretherton, C.S.: Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining. Journal of Advances in Modeling Earth Systems 11(8), 2728–2744 (2019) https://doi.org/10.1029/2019MS001711

    2019 doi

  13. [21]

    Geophysical Research Letters 48(6), 2020–091363 (2021) https://doi.org/10.1029/2020GL091363

    Yuval, J., O’Gorman, P.A., Hill, C.N.: Use of Neural Networks for Stable, Accurate and Physically Consistent Parameterization of Subgrid Atmospheric Processes With Good Performance at Reduced Precision. Geophysical Research Letters 48(6), 2020–091363 (2021) https://doi.org/10....

    2020 doi

  14. [22]

    Geophysical Research Letters 48(15), 2021–092555 (2021) https://doi.org/10.1029/2021GL092555

    Watt-Meyer, O., Brenowitz, N.D., Clark, S.K., Henn, B., Kwa, A., McGibbon, J., Perkins, W.A., Bretherton, C.S.: Correcting Weather and Climate Models by Machine Learning Nudged Historical Simulations. Geophysical Research Letters 48(15), 2021–092555 (2021) https://doi.org/10.1...

    2021 doi

  15. [23]

    Journal of Advances in Modeling Earth Systems 14(2), 2021–002794 (2022) https://doi.org/10.1029/2021MS002794

    Bretherton, C.S., Henn, B., Kwa, A., Brenowitz, N.D., Watt-Meyer, O., McGib- bon, J., Perkins, W.A., Clark, S.K., Harris, L.: Correcting Coarse-Grid Weather and Climate Models by Machine Learning From Global Storm-Resolving Simula- tions. Journal of Advances in Modeling Earth ...

    2021 doi

  16. [24]

    Journal of Advances in Modeling Earth Systems 14(9), 2022–003219 (2022) https://doi.org/10.1029/2022MS003219

    Clark, S.K., Brenowitz, N.D., Henn, B., Kwa, A., McGibbon, J., Perkins, W.A., Watt-Meyer, O., Bretherton, C.S., Harris, L.M.: Correcting a 200 km Resolution Climate Model in Multiple Climates by Machine Learning From 25 km Resolution Simulations. Journal of Advances in Modelin...

    2022 doi

  17. [25]

    arXiv (2022)

    Pathak, J., Subramanian, S., Harrington, P., Raja, S., Chattopadhyay, A., Mar- dani, M., Kurth, T., Hall, D., Li, Z., Azizzadenesheli, K., Hassanzadeh, P., Kashinath, K., Anandkumar, A.: FourCastNet: A Global Data-driven High- resolution Weather Model using Adaptive Fourier Ne...

    2022 doi arXiv

  18. [26]

    Science 382(6677), 1416–1421 (2023) 36

    Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W., et al.: Learning skillful medium-range global weather forecasting. Science 382(6677), 1416–1421 (2023) 36

    2023

  19. [27]

    https://arxiv.org/abs/2405

    Bodnar, C., Bruinsma, W.P., Lucic, A., Stanley, M., Vaughan, A., Brandstetter, J., Garvan, P., Riechert, M., Weyn, J.A., Dong, H., Gupta, J.K., Thambiratnam, K., Archibald, A.T., Wu, C.-C., Heider, E., Welling, M., Turner, R.E., Perdikaris, P.: A Foundation Model for the Earth...

    2024

  20. [28]

    Nature 619(7970), 533–538 (2023)

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

    2023

  21. [29]

    arXiv preprint arXiv:2310.02074 (2023)

    Watt-Meyer, O., Dresdner, G., McGibbon, J., Clark, S.K., Henn, B., Duncan, J., Brenowitz, N.D., Kashinath, K., Pritchard, M.S., Bonev, B., et al.: Ace: A fast, skillful learned global atmospheric model for climate prediction. arXiv preprint arXiv:2310.02074 (2023)

    2023 arXiv

  22. [30]

    Climatic change 60(3), 217–242 (2003)

    Mitchell, T.D.: Pattern scaling: an examination of the accuracy of the technique for describing future climates. Climatic change 60(3), 217–242 (2003)

    2003

  23. [31]

    Climatic Change 122(3), 459–471 (2014) https://doi.org/10.1007/s10584-013-1032-9

    Tebaldi, C., Arblaster, J.M.: Pattern scaling: Its strengths and limitations, and an update on the latest model simulations. Climatic Change 122(3), 459–471 (2014) https://doi.org/10.1007/s10584-013-1032-9

    2014 doi

  24. [32]

    Part I: Surface Air Temperature over Land (2018) https://doi.org/10.1175/JCLI-D-17-0780.1

    Osborn, T.J., Wallace, C.J., Lowe, J.A., Bernie, D.: Performance of Pattern- Scaled Climate Projections under High-End Warming. Part I: Surface Air Temperature over Land (2018) https://doi.org/10.1175/JCLI-D-17-0780.1

    2018 doi

  25. [33]

    Journal of Climate 27(5), 1829–1844 (2014)

    Castruccio, S., McInerney, D.J., Stein, M.L., Crouch, F.L., Jacob, R.L., Moyer, E.J.: Statistical emulation of climate model projections based on precomputed gcm runs. Journal of Climate 27(5), 1829–1844 (2014)

    2014

  26. [34]

    Authorea Preprints (2024)

    Freese, L.M., Fiore, A.M., Selin, N.E.: Spatially resolved temperature response functions to co2 emissions. Authorea Preprints (2024)

    2024

  27. [35]

    Earth System Dynamics 11(1), 139–159 (2020)

    Beusch, L., Gudmundsson, L., Seneviratne, S.I.: Emulating earth system model temperatures with mesmer: from global mean temperature trajectories to grid- point-level realizations on land. Earth System Dynamics 11(1), 139–159 (2020)

    2020

  28. [36]

    0: an emulator with internal variability and space–time correlation for earth system models

    Link, R., Snyder, A., Lynch, C., Hartin, C., Kravitz, B., Bond-Lamberty, B.: Fldgen v1. 0: an emulator with internal variability and space–time correlation for earth system models. Geoscientific Model Development 12(4), 1477–1489 (2019)

    2019

  29. [37]

    Climatic Change 146, 319–333 (2018)

    Alexeeff, S.E., Nychka, D., Sain, S.R., Tebaldi, C.: Emulating mean patterns and variability of temperature across and within scenarios in anthropogenic climate change experiments. Climatic Change 146, 319–333 (2018)

    2018

  30. [38]

    Jour- nal of Advances in Modeling Earth Systems 14(10), 2021–002954 (2022) https: //doi.org/10.1029/2021MS002954

    Watson-Parris, D., Rao, Y., Olivi´ e, D., Seland, Ø., Nowack, P., Camps-Valls, 37 G., Stier, P., Bouabid, S., Dewey, M., Fons, E., Gonzalez, J., Harder, P., Jeg- gle, K., Lenhardt, J., Manshausen, P., Novitasari, M., Ricard, L., Roesch, C.: ClimateBench v1.0: A Benchmark for D...

    2021 doi

  31. [39]

    In: Oh, A., Neumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S

    Kaltenborn, J., Lange, C., Ramesh, V., Brouillard, P., Gurwicz, Y., Nagda, C., Runge, J., Nowack, P., Rolnick, D.: Climateset: A large-scale climate model dataset for machine learning. In: Oh, A., Neumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S. (eds.) Advances in...

    2023

  32. [40]

    arXiv (2024)

    L¨ utjens, B., Ferrari, R., Watson-Parris, D., Selin, N.: The impact of internal variability on benchmarking deep learning climate emulators. arXiv (2024). https: //doi.org/10.48550/arXiv.2408.05288 . http://arxiv.org/abs/2408.05288

    2024 doi

  33. [41]

    0: A benchmark for data-driven climate projections

    Watson-Parris, D., Rao, Y., Olivi´ e, D., Seland, Ø., Nowack, P., Camps-Valls, G., Stier, P., Bouabid, S., Dewey, M., Fons, E., et al.: Climatebench v1. 0: A benchmark for data-driven climate projections. Journal of Advances in Modeling Earth Systems 14(10), 2021–002954 (2022)

    2021

  34. [42]

    Geophysical Research Letters 49(17), 2022–099012 (2022)

    Quilcaille, Y., Gudmundsson, L., Beusch, L., Hauser, M., Seneviratne, S.I.: Showcasing mesmer-x: Spatially resolved emulation of annual maximum tempera- tures of earth system models. Geophysical Research Letters 49(17), 2022–099012 (2022)

    2022

  35. [43]

    Environmental Research Letters 15(7), 074006 (2020)

    Tebaldi, C., Armbruster, A., Engler, H., Link, R.: Emulating climate extreme indices. Environmental Research Letters 15(7), 074006 (2020)

    2020

  36. [44]

    arXiv (2022)

    Wikner, A., Harvey, J., Girvan, M., Hunt, B.R., Pomerance, A., Anton- sen, T., Ott, E.: Stabilizing Machine Learning Prediction of Dynamics: Noise and Noise-inspired Regularization. arXiv (2022). https://doi.org/10.48550/arXiv. 2211.05262 . http://arxiv.org/abs/2211.05262

    2022 doi arXiv

  37. [45]

    Journal of Advances in Modeling Earth Systems 16(3), 2023–004122 (2024) https://doi.org/10.1029/2023MS004122

    Barthel Sorensen, B., Charalampopoulos, A., Zhang, S., Harrop, B.E., Leung, L.R., Sapsis, T.P.: A Non-Intrusive Machine Learning Framework for Debiasing Long-Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics. Journal of Advances in Modeling Eart...

    2023 doi

  38. [46]

    arXiv (2024)

    Barthel Sorensen, B., Zepeda-N´ u˜ nez, L., Lopez-Gomez, I., Wan, Z.Y., Carver, R., Sha, F., Sapsis, T.: A probabilistic framework for learning non-intrusive cor- rections to long-time climate simulations from short-time training data. arXiv (2024). https://doi.org/10.48550/ar...

    2024 doi arXiv

  39. [47]

    Journal of Advances in Modeling Earth Systems 16(8), 2023–004138 (2024) https://doi.org/10.1029/2023MS004138

    Zhang, S., Harrop, B., Leung, L.R., Charalampopoulos, A.-T., Barthel Sorensen, 38 B., Xu, W., Sapsis, T.: A Machine Learning Bias Correction on Large-Scale Environment of High-Impact Weather Systems in E3SM Atmosphere Model. Journal of Advances in Modeling Earth Systems 16(8),...

    2023 doi

  40. [48]

    Wang, M., Souza, A., Ferrari, R., Sapsis, T.: Spatially-resolved emulation of climate extremes via machine learning stochastic models (2023)

    2023

  41. [49]

    arXiv (2021)

    Song, Y., Sohl-Dickstein, J., Kingma, D.P., Kumar, A., Ermon, S., Poole, B.: Score-Based Generative Modeling through Stochastic Differential Equations. arXiv (2021). https://doi.org/10.48550/arXiv.2011.13456 . http://arxiv.org/abs/ 2011.13456

    2021 doi arXiv

  42. [50]

    arXiv (2023)

    Bischoff, T., Deck, K.: Unpaired Downscaling of Fluid Flows with Diffusion Bridges. arXiv (2023). https://doi.org/10.48550/arXiv.2305.01822 . http://arxiv. org/abs/2305.01822

    2023 doi arXiv

  43. [51]

    arXiv (2015)

    Ronneberger, O., Fischer, P., Brox, T.: U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv (2015). https://doi.org/10.48550/arXiv. 1505.04597 . http://arxiv.org/abs/1505.04597

    2015 doi arXiv

  44. [52]

    arXiv (2023)

    Wan, Z.Y., Baptista, R., Chen, Y.-f., Anderson, J., Boral, A., Sha, F., Zepeda- N´ u˜ nez, L.: Debias Coarsely, Sample Conditionally: Statistical Downscaling through Optimal Transport and Probabilistic Diffusion Models. arXiv (2023). https://doi.org/10.48550/arXiv.2305.15618 ....

    2023 doi arXiv

  45. [53]

    arXiv (2024)

    Lopez-Gomez, I., Wan, Z.Y., Zepeda-N´ u˜ nez, L., Schneider, T., Anderson, J., Sha, F.: Dynamical-generative downscaling of climate model ensembles. arXiv (2024). https://doi.org/10.48550/arXiv.2410.01776 . http://arxiv.org/abs/2410.01776

    2024 doi arXiv

  46. [54]

    arXiv (2024)

    Wan, Z.Y., Lopez-Gomez, I., Carver, R., Schneider, T., Anderson, J., Sha, F., Zepeda-N´ u˜ nez, L.: Statistical Downscaling via High-Dimensional Distribution Matching with Generative Models. arXiv (2024). https://doi.org/10.48550/arXiv. 2412.08079 . http://arxiv.org/abs/2412.08079

    2024 doi arXiv

  47. [55]

    Quarterly Journal of the Royal Meteorological Society 146(730), 1999–2049 (2020)

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

    1999

  48. [56]

    Wheeler, M., Kiladis, G.N.: Convectively Coupled Equatorial Waves: Analysis of Clouds and Temperature in the Wavenumber–Frequency Domain (1999)

    1999

  49. [57]

    Reviews of Geophysics 47(2) (2009) https: //doi.org/10.1029/2008RG000266 39

    Kiladis, G.N., Wheeler, M.C., Haertel, P.T., Straub, K.H., Roundy, P.E.: Con- vectively coupled equatorial waves. Reviews of Geophysics 47(2) (2009) https: //doi.org/10.1029/2008RG000266 39

    2009 doi

  50. [58]

    Nature Communications 14(1), 2145 (2023) https://doi.org/10.1038/s41467-023-37847-5

    Bevacqua, E., Suarez-Gutierrez, L., J´ ez´ equel, A., Lehner, F., Vrac, M., Yiou, P., Zscheischler, J.: Advancing research on compound weather and climate events via large ensemble model simulations. Nature Communications 14(1), 2145 (2023) https://doi.org/10.1038/s41467-023-37847-5

    2023 doi

  51. [59]

    Nature Climate Change 8(6), 469–477 (2018) https://doi.org/10.1038/s41558-018-0156-3

    Zscheischler, J., Westra, S., Hurk, B.J.J.M., Seneviratne, S.I., Ward, P.J., Pitman, A., AghaKouchak, A., Bresch, D.N., Leonard, M., Wahl, T., Zhang, X.: Future climate risk from compound events. Nature Climate Change 8(6), 469–477 (2018) https://doi.org/10.1038/s41558-018-0156-3

    2018 doi

  52. [60]

    Weather and Climate Extremes 30, 100283 (2020)

    Wehner, M., Gleckler, P., Lee, J.: Characterization of long period return values of extreme daily temperature and precipitation in the cmip6 models: Part 1, model evaluation. Weather and Climate Extremes 30, 100283 (2020)

    2020

  53. [61]

    Prentice Hall, Englewood Cliffs, N.J

    Box, G.E.P., Jenkins, G.M., Reinsel, G.C.: Time Series Analysis: Forecasting and Control, Third edition edn. Prentice Hall, Englewood Cliffs, N.J. (1994). http://www.gbv.de/dms/bowker/toc/9780130607744.pdf

    1994

  54. [62]

    Monthly Weather Review 128(10), 3664–3673 (2000) https://doi.org/10.1175/1520-0493(2000)128⟨3664:ASNTFD⟩2.0.CO;2

    Storch, H.v., Langenberg, H., Feser, F.: A Spectral Nudging Technique for Dynamical Downscaling Purposes. Monthly Weather Review 128(10), 3664–3673 (2000) https://doi.org/10.1175/1520-0493(2000)128⟨3664:ASNTFD⟩2.0.CO;2

    2000 doi

  55. [63]

    Journal of Climate 18(8), 1227–1246 (2005) https://doi.org/10.1175/ JCLI3369.1

    Miguez-Macho, G., Stenchikov, G.L., Robock, A.: Regional Climate Simulations over North America: Interaction of Local Processes with Improved Large-Scale Flow. Journal of Climate 18(8), 1227–1246 (2005) https://doi.org/10.1175/ JCLI3369.1

    2005

  56. [64]

    Journal of Advances in Modeling Earth Systems 11 (2019) https://doi.org/10.1029/2019MS001831

    Sun, J., Zhang, K., Wan, H., Ma, P.-L., Tang, Q., ZHANG, S.: Impact of Nudging Strategy on the Climate Representativeness and Hindcast Skill of Constrained EAMv1 Simulations. Journal of Advances in Modeling Earth Systems 11 (2019) https://doi.org/10.1029/2019MS001831

    2019 doi

  57. [65]

    Geoscientific Model Development 14(5), 2827–2841 (2021) https://doi.org/10.5194/gmd-14-2827-2021

    Huang, Z., Zhong, L., Ma, Y., Fu, Y.: Development and evaluation of spectral nudging strategy for the simulation of summer precipitation over the Tibetan Plateau using WRF (v4.0). Geoscientific Model Development 14(5), 2827–2841 (2021) https://doi.org/10.5194/gmd-14-2827-2021

    2021 doi

  58. [66]

    arXiv preprint arXiv:2111.13606 (2021)

    Batzolis, G., Stanczuk, J., Sch¨ onlieb, C.-B., Etmann, C.: Conditional image generation with score-based diffusion models. arXiv preprint arXiv:2111.13606 (2021)

    2021 arXiv

  59. [67]

    arXiv preprint arXiv:2210.02747 (2022)

    Lipman, Y., Chen, R.T., Ben-Hamu, H., Nickel, M., Le, M.: Flow matching for generative modeling. arXiv preprint arXiv:2210.02747 (2022)

    2022 arXiv

  60. [68]

    arXiv preprint arXiv:2303.08797 (2023) 40

    Albergo, M.S., Boffi, N.M., Vanden-Eijnden, E.: Stochastic interpolants: A unifying framework for flows and diffusions. arXiv preprint arXiv:2303.08797 (2023) 40

    2023 arXiv

  61. [69]

    Stochastic Processes and their Applications 12(3), 313–326 (1982)

    Anderson, B.D.: Reverse-time diffusion equation models. Stochastic Processes and their Applications 12(3), 313–326 (1982)

    1982

  62. [70]

    arXiv (2015)

    He, K., Zhang, X., Ren, S., Sun, J.: Deep Residual Learning for Image Recog- nition. arXiv (2015). https://doi.org/10.48550/arXiv.1512.03385 . http://arxiv. org/abs/1512.03385 41

    2015 doi arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.