DySCo: Dynamically consistent data-driven downscaling of extremes in climate projections

S. Stamatelopoulos; M. Wang; I. Lopez-Gomez; L. Zepeda-Nunez; Z. Y. Wan; R. Carver; F. Sha; T. P. Sapsis

cs.LGAug 22, 2026

DySCo: Dynamically consistent data-driven downscaling of extremes in climate projections

Authors: S. Stamatelopoulos, M. Wang, I. Lopez-Gomez, L. Zepeda-Nunez, Z. Y. Wan, R. Carver, F. Sha, T. P. Sapsis

Organizations: Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA · Department of Mechanical Engineering, City University of Hong Kong, Kowloon, Hong Kong, China · Google Research, Mountain View, CA, USA · Meta, Menlo Park, CA, USA

Abstract

Regional climate risk assessment is critical for applications such as infrastructure design, disaster forecasting, and insurance resource allocation. However, estimating regional (i.e., high-spatial-resolution) risk with global climate models (GCMs) remains computationally prohibitive, which has driven the development of downscaling methods for coarse GCM outputs. Downscaling is vital for rare events, since quantifying their extreme properties requires high spatial resolution and very long GCM simulations. These methods non-intrusively increase GCM resolution while correcting statistical biases from unresolved fine-scale processes, thereby improving the accuracy of extreme event statistics with long return periods. A key challenge is preserving dynamical consistency, as freely evolving GCM trajectories are not expected to track the observational dataset used for training the correction operator. This is critical for causal extreme event analyses, where storyline-based risk assessment, i.e., extreme event catalogs, is necessary for effective planning. We address this challenge by introducing Dynamically and Statistically Consistent downscaling (DySCo), a non-intrusive framework yielding high-resolution climate projections consistent with coarse GCM dynamics. DySCo relies on a data-driven reformulation of nudging to create dynamically paired training trajectories without intrusive GCM modifications. Using these paired trajectories, we train a dynamically and statistically consistent, two-stage operator. We evaluate the method by downscaling the Community Earth System Model v2 Large Ensemble (LENS2) in time and space towards historical reanalysis. Results show DySCo achieves superior dynamical consistency with the coarse GCM trajectories, essentially applying a minimal, causal correction to the GCM, preserving top statistical performance comparable to state-of-the-art unsupervised models.

Figures & tables

Figure 1: Illustration of the generation of paired training trajectories through nudging. Based on the original dynamical system $G$ and a reference trajectory $y$ , a new dynamical system $G^{\rm nudged}$ ( 1 ) is defined to generate trajectories $x^{\rm nudged}$ that track the evolution of $y$ . These trajectories can be used as training pairs $(x^{\rm nudged},y)$ to define a supervised learning problem for the mapping $x^{\rm nudged}\to y$ . Because the fast dynamics of $x^{\rm nudged}$ evolve freely according to $G$ , the learned mapping can be applied to new trajectories $\dot{x}^{\prime}=G(x^{\prime})$ .

Figure 2: Case study of dynamical coherence. (A) the synoptic footprint of a LENS2 extreme temperature event in the Pacific Northwest is shown ( LE2-1251.001 member, 2016 August 6th). The contours indicate mean-sea-level pressure, revealing a surface thermal low like event ( $\sim$ 1008hPa) co-located with extreme surface temperatures. (B) The corresponding output from DySCo, illustrating the high-resolution topographic modulation of the phenomenon, preserves the $T\leftrightarrow P$ relationship while successfully debiasing the field toward the ERA5 distribution. (C) Temperature-pressure spatial correlation time series around the August 6th event, computed using a 1-day rolling average. (D) The corresponding output from GenFocal, which while able to sample a physical trajectory, it does not correspond to the LENS2 extreme event.

Figure 3: Anomaly correlation coefficient ( 12 ) between LENS2 and corrected output, for DySCo and GenFocal. The top four plots for pressure and humidity depict the correlation between LENS2 and the output of the debiasing stage of both models, $G_{\rm deb}(x)$ . The bottom four plots for temperature and wind speed depict the correlation with the fully downscaled output.

Figure 4: MSL temporal coherence between input LENS2 driver and debiased fields. Shaded region represents the 5 th and 95 th percentile range.

Variable/Model	LENS2	DySCo	GenFocal	BCSD
	Mean Anomaly Correlation Coefficient $\uparrow$
$P$ (Pa)	-	0.91	0.12	0.82
$W$ (m/s)	-	0.69	0.12	0.63
$T$ (K)	-	0.77	0.17	0.74
$Q$ (g/kg)	-	0.79	0.23	0.75
$HI$ ( ∘ C)	-	0.51	0.13	0.55

Table 1: Marginal metrics and ACC over CONUS, for various models.

Figure 5: Bias over the 2010-2019 testing domain of the four downscaled climatological variables, for DySCo and the unsupervised GenFocal baseline.

Figure 6: Spatial correlation of near-surface temperature with respect to Miami, showing the LENS2 $1.5^{\circ}$ input, the ERA5 $0.25^{\circ}$ target and the DySCo $0.25^{\circ}$ output.

Panel of Figure 7 — Figure 7: $4^{\circ}\times 4^{\circ}$ local PDFs and correlation plots, across four different cities. For each city, top right plot is correlation of ERA5 with respect to the city location and bottom right plot is the correlation reproduced with DySCo. Top and bottom left plots are the PDFs for that variable and that region, for various models.

Figure 8: Probability density of cyclone trajectories of the ERA5 trajectory and the ensembles generated through DySCo, BCSD and GenFocal.

Figure 9: East-coast cyclone landfall distribution of DySCo, GenFocal and BCSD, over the testing period for 100 members. The points are the ERA5 single-trajectory landfall locations.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 10: Dependence of the debiasing operator on the

\lambda

parameter that control the supervised versus unsupervised signal learned. Marginal performance with respect to the metrics ( 18 , 20 , 22 ) is reported, for the debiased output

G_{\rm deb}(x;\lambda)

in comparison to the coarse scale ERA5 data

y

, for

G_{\rm deb}

trained with varying values of

\lambda

Architecture
Sample length	2 days
Residual blocks	[3,3,3]
Downsampling ratios	[2,2,2]
Hidden channels	[64,64,64]
Spatial attention	[ False , True , True ]
Temporal attention	[ False , True , True ]

Appendix

Table 2: Debiasing model architecture and training hyperparameters.

Architecture
Sample length	7 days
Residual blocks	[4,4,4,4]
Downsampling ratios	[3,2,2,2]
Hidden channels	[128, 256, 384, 512]
Spatial attention	[ False , False , False , True ]
Temporal attention	[ False , False , False , True ]

Appendix

Table 3: Super-resolution model architecture and training hyperparameters.

Figure 11: Per-model calibration of $K$ via an ablation study over $1/K\in\{0.1,0.2,...,1\}$ .

Figure 12: Temporal coherence between input LENS2 driver and debiased fields. Shaded region represents the 5 th and 95 th percentile range.

Climate Downscaling

Climate downscaling translates coarse climate simulations and observations into finer regional detail for local impact and risk assessment.

See all 27 papers

Generative climate downscaling enables high-resolution compound risk assessment by preserving multivariate dependencies
May 12, 2026Takuro Kutsuna, Noriko N. Ishizaki, Norihiro Oyama +1Climate Science Climate Modeling
Toyota Central R&D Labs., Inc. · National Institute for Environmental Studies
Temporal Coverage over Density: Parsimonious Training-Set Design for ML Climate Downscaling
Jun 5, 2026Karandeep Singh, Stefan Rahimi, Chad W. Thackeray +2Surrogate Modeling Climate Science
Atmospheric and Oceanic Sciences, University of California, Los Angeles, CA, USA · Department of Atmospheric Science, University of Wyoming, WY, USA
Physics-Constrained Adaptive Flow Matching for Climate Downscaling
Apr 3, 2026Kevin Debeire, Aytaç Paçal, Pierre Gentine +3Climate Downscaling Physics-Informed Generative Modeling
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany · School of Computation, Information and Technology, Technical University of Munich, Germany · LEAP Center, Columbia University, New York, NY, 10027, USA +1

May 12, 2026J/K move · Enter open · S save

Generative climate downscaling enables high-resolution compound risk assessment by preserving multivariate dependencies

Takuro Kutsuna, Noriko N. Ishizaki, Norihiro Oyama, Hiroaki Yoshida

Toyota Central R&D Labs., Inc. · National Institute for Environmental Studies

Climate Science Climate Modeling Generative Modeling Climate Downscaling Diffusion Models

Open paper AskPDF arXiv

Physics-based climate projections using general circulation models are essential for assessing future risks, but their coarse resolution limits regional decision-making. Statistical downscaling can efficiently add detail, yet many methods treat variables independently, degrading inter-variable relationships that govern compound hazards such as heat stress, drought, and wildfire. Here we show that a diffusion-based multivariate generative framework, combined with bias correction, recovers degraded inter-variable correlations even under a 50

\times

increase in linear resolution. When applied to five meteorological variables over Japan, the framework reduces inter-variable correlation errors by more than fourfold relative to existing baselines while improving both univariate and spatial accuracy, leading to more accurate detection of severe drought. These results demonstrate that multivariate generative downscaling improves the reliability of compound risk assessment under large resolution gaps.

Abstract

Climate Downscaling

Explore similar work

Generative climate downscaling enables high-resolution compound risk assessment by preserving multivariate dependencies

Temporal Coverage over Density: Parsimonious Training-Set Design for ML Climate Downscaling

Physics-Constrained Adaptive Flow Matching for Climate Downscaling