cs.LGSep 30, 2026

RainAtlas: A Multi-Continental Dataset for Precipitation Downscaling

Authors: Pierre-Louis Lemaire, Luca Schmidt, Wietze Suijker, Alex Hernandez-Garcia, David Rolnick

Organizations: connAIx Research School, Tübingen AI Center, University of Tübingen · Mila - Quebec AI Institute · Cluster of Excellence Machine Learning, University of Tübingen · IVADO · Universit´e de Montr´eal · McGill University

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Extreme rainfall events are increasing in intensity and frequency as climate change accelerates. While kilometer-scale precipitation forecasts are critical for supporting local decision-making, the limited availability of high-resolution precipitation observations hinders their accuracy, especially in under-resourced regions. Machine learning models are widely used to downscale precipitation data to km-scale, but their application to unseen geographies presents challenges. First, processing raw high-resolution precipitation datasets across regions requires significant engineering and domain expertise. Second, generalization across regions remains difficult. To help overcome these barriers, we release RainAtlas, a large-scale, ML-ready and multi-continental dataset for precipitation downscaling. Covering three continents, RainAtlas harmonizes heterogeneous hourly km-scale observations to a common 2-km grid. Each regional partition contains around 210,000 aligned low- and high-resolution precipitation pairs, respectively from ERA5 reanalysis and direct observations. We benchmark state-of-the-art ML-based downscaling models across RainAtlas using a wide range of metrics. Our evaluation reveals substantial variance in out-of-domain generalization depending on the training regions. This underscores the need for cross-regional, multi-source km-scale evaluation, establishing RainAtlas as a well-positioned benchmark for precipitation downscaling research.

Figures & tables

Figure 1: Global distribution of ground-based radar stations (blue dots) and their available coverage (white transparent circles) . Some regions have excellent coverage (USA, Europe, East-Asia), while other densely populated regions have little to no precipitation observations available (Africa, South America, Central Asia). RainAtlas consists of the three multi-source precipitation products highlighted in the figure, covering three Northern Hemisphere regions.
VariableDescriptionUnitLevel
tpTotal precipitationmmsurface
cpConvective precipitationmmsurface
capeConvective potential energyJ ⋅\cdot kg -1surface
spSurface pressurePasurface
tisrTop-of-the-atmosphere incident solar radiationJ ⋅\cdot m -2surface
tcwTotal column waterkg ⋅\cdot m -2total column
Table 1: ERA5 and static input data variables.
Figure 2: Diffusion-based models outperform other baselines. In-domain performance evaluation on importance-subsampled test sets. MAE/CRPS, and bias are computed without artificial coarsening. MAE/CRPS, and variogram rows have region-specific yy -axis ranges to facilitate the visualization. Arrows indicate the desired direction for each metric. Numerical results are detailed in Table 4 .
Figure 3: Models trained on MRMS show more robust geographical generalization. Relative MAE, CRPS and VARIOGRAM degradation ( %\% ) with respect to in-domain training. Additional results are detailed in Figures 15 and 16 .
Appendix figures & tables18 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 4: Comparison of MRMS and ERA5 precipitation samples. The high-resolution MRMS crops (top) are compared with the low-resolution ERA5 crops (bottom) for two different locations. In Location A , the MRMS data appears to be upsampled from a coarser grid. In contrast, Location B shows the typical fine-scale details expected from MRMS.
Figure 5: (left) Raw maximum hourly precipitation per grid-cell for EURADCLIM (2013 to 2023), and (right) same using a 125 mm threshold for easier identification of problematic grid-cells.
 MRMSEURADCLIMEA
 RandomImportanceRandomImportanceRandomImportance
≤0\leq 0 mm (%)95.8072.7682.8155.7796.6880.33
0−10-1 mm (%)2.3812.6315.3437.181.015.05
1−41-4 mm (%)1.4711.721.657.131.7710.58
4−54-5 mm (%)0.100.910.080.360.151.03
5−85-8 mm (%)0.131.140.080.390.191.37
Appendix
Table 2: Statistics of RainAtlas ’s training datasets with 150150 k samples, generated with random or importance subsampling. The first section contains grid cell occurrences of multiple precipitation intensities, while the second section gives information on crops: average precipitation, average maximum intensity, and proportion of crops with more than 10%,25%10\%,25\% and 50%50\% non-zero precipitation grid cells.
Figure 6: Train, validation, and test temporal splits. Because MRMS only partially overlap with EURADLCIM and EASTASIA, we reserve 2021 and 2022 for subsampling validation and test datasets. Since atmospheric predictability is limited to approximately 14 days, training on data from later periods does not induce data leakage.
 UNetDiffusionCorrective DiffusionSerpentFlowConsistency
LossL2L_{2}EDM denoisingEDM denoisingL2L_{2}Pseudo-Huber
OptimizerAdamWAdamWAdamWAdamWAdamW
Learning rate2×10−42{\times}10^{-4}10−410^{-4}10−410^{-4}10−410^{-4}10−410^{-4}
Weight decay10−410^{-4}10−210^{-2}10−210^{-2}10−210^{-2}10−210^{-2}
Residual✓—✓(UNet)——
Dropout0.10.10.10.10.1
Appendix
Table 3: Training and architecture hyperparemeters for every model. Additional hyperparameters are specified under each model specification when they are not shared with others. We note that most of these hyperparameters are certainly sub-optimals as we didn’t perform any extensive hyperparameters seach. Most values are taken from previous work ( Watt and Mansfield, 2024 ; Karras et al., 2022 ; Song and Dhariwal, 2024 ; Keisler et al., 2026 ) .
  MAE / CRPS (↓)(\downarrow)BIAS (→0)(\rightarrow 0)LHD (↓)(\downarrow)SSR (→1)(\rightarrow 1)VARIOGRAM (↓)(\downarrow)
MRMSBilinear0.7288 -0.0490 19.9497 --
UNet0.5750±0.0026-0.4033±0.006921.0799±0.1626--
Diffusion0.4593±0.0014-0.0472±0.04134.4261±0.67630.9437±0.01030.1671±0.0004
Corrective Diffusion0.4639±0.0031-0.0341±0.03583.3621±1.09400.8373±0.04360.1662±0.0007
SerpentFlow0.5031±0.00050.0353±0.02718.5822±0.95191.0268±0.06160.1798±0.0012
Consistency0.4867±0.01070.2825±0.06237.0539±0.20341.2294±0.02250.1825±0.0027
Appendix
Table 4: In-domain performance evaluation on importance-subsampled test datasets. MAE, CRPS, and BIAS are computed on the original resolution without artificial coarsening. Probabilistic metrics (SSR and VARIOGRAM) are evaluated exclusively for generative models, omitting the deterministic Bilinear (ERA5) and UNet baselines. Arrows in the header indicate the optimal direction for each metric. The best and second-best results are bolded and underlined , respectively. Subscripts denote standard deviations.
Figure 7: Effect of spatial coarsening on pointwise error and bias (importance-subsampled test sets). MAE / CRPS ( top , ↓\downarrow ) and bias ( bottom , →0\rightarrow 0 ) evaluated across four effective spatial resolutions (native 22 km, 88 km, 1616 km, and 2424 km) obtained by average pooling. Shaded bands denote ±\pm standard deviation across three random seeds. Pointwise errors systematically decrease with spatial aggregation due to reduced sensitivity to small-scale displacement errors (double-penalty effect), while bias remains scale-invariant. Relative performance rankings remain unchanged across resolutions.
Figure 8: In-domain intensity distribution and spectral fidelity (importance-subsampled test sets). (Top) Precipitation intensity histograms comparing model predictions against ground truth up to 300300 mm h -1 on a logarithmic scale. (Bottom) Radially Averaged Power Spectral Density (RAPSD) normalized by observed power spectrum across radial wavenumbers kk . The horizontal dashed line at y=1.0y=1.0 indicates perfect spectral agreement with observations. Shaded bands represent ±\pm standard deviation across evaluation crops and ensemble members.
Figure 9: Fractions Skill Score across precipitation thresholds (importance-subsampled test sets). Fractions Skill Score (FSS, ↑1\uparrow 1 ) evaluated across neighborhood scales ( 4,8,16,32,644,8,16,32,64 km) for thresholds q∈{1,5,10,20}q\in\{1,5,10,20\} mm h -1 ( left to right ). Shaded bands denote ±\pm standard deviation across three random seeds. Generative baselines, especially diffusion and corrective Diffusion, retain high skill for extreme precipitation ( q≥5q\geq 5 mm h -1 ), whereas deterministic baselines degrade sharply.
Figure 10: Ensemble calibration via rank histograms (importance-subsampled test sets). Talagrand rank histograms for generative models evaluated using an 88 -member ensemble ( 99 rank bins). The horizontal dashed line at y≈0.111y\approx 0.111 marks the ideal uniform distribution corresponding to perfect ensemble calibration. Random tie-breaking is applied to values below 0.10.1 mm h -1 .
  MAE / CRPS (↓)(\downarrow)BIAS (→0)(\rightarrow 0)LHD (↓)(\downarrow)SSR (→1)(\rightarrow 1)VARIOGRAM (↓)(\downarrow)
MRMSBilinear0.1115 0.0144 18.7129 --
UNet0.0719±0.0001-0.0513±0.001119.8704±0.1937--
Diffusion0.0609±0.00030.0323±0.00635.0397±0.65271.2478±0.01580.0314±0.0004
Corrective Diffusion0.0616±0.00080.0224±0.00753.6442±1.65671.0549±0.07880.0304±0.0006
SerpentFlow0.0742±0.00020.0349±0.00289.2778±0.54591.2290±0.05200.0340±0.0005
Consistency0.0659±0.00170.0716±0.01097.3231±0.23791.4877±0.03180.0347±0.0008
Appendix
Table 5: In-domain performance evaluation on randomly-subsampled test datasets. Details regarding metrics and formatting are identical to Table 4 . Probabilistic metrics are evaluated exclusively for generative models. The best and second-best results are bolded and underlined , respectively. Subscripts denote standard deviations.
Figure 11: Effect of spatial coarsening on pointwise error and bias (randomly-subsampled test sets). Settings and results are similar to Figure 7 .
Figure 12: In-domain intensity distribution and spectral fidelity (randomly-subsampled test sets). Settings are similar to Figure 8 . Generative baselines show much larger RAPSD, with high variance on EA, compared to evaluation on importance-subsampled test datasets.
Figure 13: Fractions Skill Score across precipitation thresholds (randomly-subsampled test sets). Settings are similar to Figure 9 . Generative models maintain notable skill for higher intensity thresholds even under climatologically unweighted random sampling.
Figure 14: Ensemble calibration via rank histograms (randomly-subsampled test sets). Settings are similar to Figure 10 . Ensemble calibration patterns remain consistent with the importance-subsampled results.
Figure 15: Out-of-domain generalization relative degradation on importance-subsampled test sets. Relative degradation ( %\% ) with respect to in-domain training across five evaluation metrics (MAE/CRPS, BIAS, LHD, SSR, and VARIOGRAM). Best-performing out-of-domain training domains per column are boxed in black. Scores are averaged over three random seeds, except for configurations with runs still in progress at submission, which are averaged over the seeds available. Full results will be updated during the rebuttal period.
Figure 16: Out-of-domain generalization relative degradation on randomly-subsampled test sets. Relative degradation ( %\% ) with respect to in-domain training across five evaluation metrics (MAE/CRPS, BIAS, LHD, SSR, and VARIOGRAM). Scores are averaged over three random seeds, except for configurations with runs still in progress at submission, which are averaged over the seeds available. Full results will be updated during the rebuttal period.
Figure 17: Intensity histograms and relative RAPSD under geographical generalization (importance-subsampled test sets). Line colors distinguish the different training domains (MRMS, EURADCLIM, EASTASIA, EUR+EA, MRMS+EA, MRMS+EUR) against ground truth (black).

Explore similar work

CardsList
Sep 23, 2026cs.LG

Evaluating Cross-region Generalization for Wavelet-Diffusion Precipitation Downscaling

Diffusion models have shown strong potential for kilometer-scale precipitation downscaling, but their performance in geographically unseen regions and event regimes remains insufficiently understood. Building on the wavelet diffusion model (WDM) framework, this study evaluates cross-region and cross-event generalization. Six 3 x 3 deg U.S. regions represent convective, winter, tropical, and atmospheric-river precipitation regimes. Low-resolution inputs are generated by block averaging NOAA Multi-Radar/Multi-Sensor (MRMS) composite reflectivity fields. A WDM trained only on Oklahoma (OK) samples and a WDM trained on all six regions are compared with nearest-neighbor and Bicubic interpolation. Model performance is evaluated using three metric families that measure image-domain reconstruction, spectral and distributional fidelity, and bin-wise precipitation detection. The OK-trained WDM remains competitive outside OK. Although the all-region WDM delivers the best and most consistent overall image-domain and detection performance, its gains are uneven across precipitation intensities. Bin-wise critical success index (CSI) over 5-dBZ reflectivity bins shows that WDM improvements concentrate in localized higher-reflectivity structures, which image-domain metrics partly obscure. In addition, the performance differences among samples are strongly associated with the spatial organization of the precipitation field, quantified by Moran's I as the spatial autocorrelation of each reflectivity bin. The sample-level Moran's I-CSI correlation stratified by sample intensity reaches 0.901 in all six regions, including regions unseen during training. Overall, these findings support future efforts to transfer downscaling models to regions with limited local training data and to generate globally consistent, high-resolution precipitation products.

Weikang Qian, Yixin Wen, Chugang Yi +3

University of Florida, Department of Geography, Gainesville, FL, USA · School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY, USA · Institute for Advanced Computational Science, Stony Brook University, NY, USA +5

Dynamical DownscalingRobust Cross-Domain GeneralizationWavelets
AskPDF
May 17, 2026physics.ao-ph

Longwang: Zero-Shot Global Spatiotemporal Precipitation Downscaling with a Latent Generative Prior

High-resolution precipitation information is essential for climate impact assessment, yet global climate models remain too coarse to resolve key small-scale processes. Existing machine learning downscaling methods often require paired low- and high-resolution data for supervised learning, are tied to fixed regions or scale factors during inference, and can be computationally expensive to train and run in physical space. Here we introduce Longwang, a zero-shot latent generative framework for global spatiotemporal precipitation downscaling. Longwang learns a context-conditioned latent generative prior and combines it with a physically informed observation operator through posterior sampling, enabling daily O(10 km) precipitation fields to be generated from monthly O(100 km) inputs. On ERA5 reanalysis, Longwang outperforms standard posterior sampling with an unconditional generative prior in reconstructing fine-scale spatial patterns, preserving temporal coherence, and recovering extreme precipitation intensities. The framework further generalizes to historical climate simulations and future climate projections under substantial distribution shift.

Yue Wang, Daniele Visioni

1*Department of Earth and Atmospheric Sciences, Cornell University, 112 Hollister Drive, Ithaca, 14853, NY, USA.

Dynamical DownscalingPrecipitation NowcastingGlobal Climate Models
AskPDF
Sep 24, 2026cs.LG

Lightweight Probabilistic Downscaling from a Deterministic Base Model

Climate data downscaling is the task of increasing the spatial resolution of climate data, typically by generating fine-resolution regional climate data from coarse global model output. Recent machine learning (ML) work in the related task of weather forecasting has seen significant improvements due to newly devised training methods and architectural components, but these have not yet benefited downscaling. We adapt two of these methods to create a family of lightweight probabilistic ML downscaling models built on a modified U-Net backbone and evaluate them on the CORDEX-ML-Bench suite for daily maximum temperature and precipitation across three geographic regions: the Alps, New Zealand and South Africa. We find that a two-stage training curriculum, combining deterministic pretraining with probabilistic tuning, transfers well to downscaling, beating the state-of-the-art for RMSE. Our work provides an advancement towards lightweight, probabilistic downscaling models, reducing the current trade-off between computational intensity and distributional fit.

Joseph McLean, Tiffany Vlaar, Sigrid Passano Hellan +1

University of Glasgow · Climate System and Climate Services, NORCE Research AS and Bjerknes Centre for Climate Research, Bergen, Norway

Dynamical DownscalingProbabilistic ForecastingClimateLightweight
AskPDF