Peter L. Langen

Darkening of the Greenland ice sheet is known to amplify surface melting. Here, analysis of automatic weather station (AWS) data around the ice sheet and satellite‐derived albedo data reveals that the bare ice duration (BID) is a predictor of the average albedo during summer (June through August, JJA, ). Inverse correlation between the two is found in both the AWS ( = 0.74) and satellite ( = 0.55) data, meaning that longer periods of bare ice result in lower average albedo. The average decrease in measured by AWS is −0.0044 ± 0.0002 (mean ± standard error (SE)), associated with an additional ∼9 cm ablation (∼8 cm w.e.) for each 0.01 decrease in . Satellite observations reveal that lowers −0.0032 ± 0.0000 (mean ± SE) across the ice sheet, increasing season‐total melt by ∼4 cm w.e. for each 0.01 decrease in . The simple correlation between BID and offers a convenient expression for current Greenland ice sheet melt modeling, but we advise caution when using this expression for predictive melt modeling because the data scatter in versus BID plots become more dispersed around the best fit linear regression line as BID increases. We contend that this scattering is most likely due to spatial variability in the dominant darkening processes, which differ in their albedo‐lowering capacity, and that this variability will become increasingly important to constrain in a warming climate as larger surface areas of the ice sheet are exposed and the melt seasons lengthen.

Understanding local impacts of extreme weather, from e.g. precipitation or temperature, requires high-resolution downscaling of those atmospheric variables. Deterministic and classical statistical methods fail to adequately represent variability and extremes, motivating the use of probabilistic generative methods. Recent advances in generative deep learning show promise for spatial realism and ensemble generation, but their physical fidelity and limitations remain insufficiently understood. We present CEDDAR v1.0.2 (Controllable Ensemble Diffusion Downscaling for Atmospheric Rainfall), a diffusion-based generative model for daily precipitation downscaling, conditioned on large-scale (∼30 × 30 km) ERA5 precipitation fields, to produce kilometre-scale (2.5 × 2.5 km) precipitation over Denmark, using the DANish ReAnalysis product (DANRA) as reference. The modelling framework is based on an Elucidated Diffusion Model (EDM) backbone and incorporates soft physics-guidance through seasonal (Day-of-Year) and geographic conditioning with a specific land-focused design.We further introduce regularisation terms through a Signed-Distance Function weighted loss, and an auxiliary RainGate head for shaping wet-dry occurrence statistics and improving ensemble calibration. By design, the model is probabilistic and capable of producing ensembles. We further introduce an inference-time control parameter, σ^*, that allows for interpretable adjustment of stochastic scale sensitivity without retraining. For model-behaviour assessment, we introduce a multi-perspective evaluation protocol with emphasis on trade-offs across probabilistic, spatial, climatological, and scale-dependent diagnostics rather than single-metric optimisation. Our results show strong spatial and spectral realism with generally well-calibrated ensembles, but also demonstrate that visually realistic fine-scale structure does not guarantee climatological fidelity. We also find that explicitly introducing σ^* reveals systematic, though non-linear, trade-offs between large-scale coherence, fine-scale variability, and ensemble spread. This highlights the potential and the limitations of controllable diffusion-based downscaling. In conclusion, this work presents a diffusion-based downscaling framework and evaluation suite that works as a diagnostic testbed for understanding how generative models behave in physically constrained settings, and offers guidance on their appropriate use and common failure modes relevant to future operational solutions and climate-impact applications.

Understanding local impacts of extreme weather, from e.g. precipitation or temperature, requires high-resolution downscaling of those atmospheric variables. Deterministic and classical statistical methods fail to adequately represent variability and extremes, motivating the use of probabilistic generative methods. Recent advances in generative deep learning show promise for spatial realism and ensemble generation, but their physical fidelity and limitations remain insufficiently understood. We present CEDDAR v1.0.2 (Controllable Ensemble Diffusion Downscaling for Atmospheric Rainfall), a diffusion-based generative model for daily precipitation downscaling, conditioned on large-scale (∼30 × 30 km) ERA5 precipitation fields, to produce kilometre-scale (2.5 × 2.5 km) precipitation over Denmark, using the DANish ReAnalysis product (DANRA) as reference. The modelling framework is based on an Elucidated Diffusion Model (EDM) backbone and incorporates soft physics-guidance through seasonal (Day-of-Year) and geographic conditioning with a specific land-focused design.We further introduce regularisation terms through a Signed-Distance Function weighted loss, and an auxiliary RainGate head for shaping wet-dry occurrence statistics and improving ensemble calibration. By design, the model is probabilistic and capable of producing ensembles. We further introduce an inference-time control parameter, σ^*, that allows for interpretable adjustment of stochastic scale sensitivity without retraining. For model-behaviour assessment, we introduce a multi-perspective evaluation protocol with emphasis on trade-offs across probabilistic, spatial, climatological, and scale-dependent diagnostics rather than single-metric optimisation. Our results show strong spatial and spectral realism with generally well-calibrated ensembles, but also demonstrate that visually realistic fine-scale structure does not guarantee climatological fidelity. We also find that explicitly introducing σ^* reveals systematic, though non-linear, trade-offs between large-scale coherence, fine-scale variability, and ensemble spread. This highlights the potential and the limitations of controllable diffusion-based downscaling. In conclusion, this work presents a diffusion-based downscaling framework and evaluation suite that works as a diagnostic testbed for understanding how generative models behave in physically constrained settings, and offers guidance on their appropriate use and common failure modes relevant to future operational solutions and climate-impact applications.

Understanding local impacts of extreme weather, from e.g. precipitation or temperature, requires high-resolution downscaling of those atmospheric variables. Deterministic and classical statistical methods fail to adequately represent variability and extremes, motivating the use of probabilistic generative methods. Recent advances in generative deep learning show promise for spatial realism and ensemble generation, but their physical fidelity and limitations remain insufficiently understood. We present CEDDAR v1.0.2 (Controllable Ensemble Diffusion Downscaling for Atmospheric Rainfall), a diffusion-based generative model for daily precipitation downscaling, conditioned on large-scale (∼30 × 30 km) ERA5 precipitation fields, to produce kilometre-scale (2.5 × 2.5 km) precipitation over Denmark, using the DANish ReAnalysis product (DANRA) as reference. The modelling framework is based on an Elucidated Diffusion Model (EDM) backbone and incorporates soft physics-guidance through seasonal (Day-of-Year) and geographic conditioning with a specific land-focused design.We further introduce regularisation terms through a Signed-Distance Function weighted loss, and an auxiliary RainGate head for shaping wet-dry occurrence statistics and improving ensemble calibration. By design, the model is probabilistic and capable of producing ensembles. We further introduce an inference-time control parameter, σ^*, that allows for interpretable adjustment of stochastic scale sensitivity without retraining. For model-behaviour assessment, we introduce a multi-perspective evaluation protocol with emphasis on trade-offs across probabilistic, spatial, climatological, and scale-dependent diagnostics rather than single-metric optimisation. Our results show strong spatial and spectral realism with generally well-calibrated ensembles, but also demonstrate that visually realistic fine-scale structure does not guarantee climatological fidelity. We also find that explicitly introducing σ^* reveals systematic, though non-linear, trade-offs between large-scale coherence, fine-scale variability, and ensemble spread. This highlights the potential and the limitations of controllable diffusion-based downscaling. In conclusion, this work presents a diffusion-based downscaling framework and evaluation suite that works as a diagnostic testbed for understanding how generative models behave in physically constrained settings, and offers guidance on their appropriate use and common failure modes relevant to future operational solutions and climate-impact applications.

Predicting surface melt on the Greenland ice sheet is critical for understanding surface mass balance (SMB) and its sensitivity to a changing climate. Polar regional climate models (RCMs) are the primary tools for simulating melt and projecting future SMB, but different models produce significantly different results. However, they are too computationally expensive to create the large ensembles needed to quantify this uncertainty. We develop a neural network based emulator that predicts daily surface melt from atmospheric variables, trained on output from the polar RCM HIRHAM5 and its firn model DMIHH forced by ERA-Interim. The emulator uses a physics-informed design combining short-term weather with long-term climate memory, capturing both immediate atmospheric forcing and accumulated firn characteristics. Input selection study shows that turbulent heat fluxes, downwelling radiation, and precipitation together with seasonal encoding suffice to reproduce surface melt. The emulator achieves mean absolute error below 0.21 mm w.e. per day relative to the surface melt produced by DMIHH across all six Greenland drainage basins, with the errors primarily attributable to spatial over-smoothing. Our work demonstrates that machine learning can successfully emulate firn model behavior from climate forcing alone with computational costs orders of magnitude lower than traditional simulations. Once retrained for specific climate forcings, the emulator thus enables extensive ensemble projections. Furthermore, the modular architecture can be readily adapted to emulate other SMB quantities such as runoff. This represents a crucial first step toward computationally efficient emulation of polar regional climate models and surrogate modeling of SMB components in Earth system modeling.

Accurate projections of ice sheet surface mass balance (SMB) are critical for sea-level rise estimates. Polar regional climate models (RCMs) coupled with firn models are the primary tools for simulating melt and projecting future SMB. However, the projections from different RCMs significantly deviate from each other. Additionally, the computational costs of polar RCMs and their firn models limit the creation of large ensembles needed to statistically assess the likely range in future melt and runoff.Machine learning emulators offer a promising solution by enabling rapid predictions at a fraction of the computational cost. We therefore present machine learning emulators that predict daily surface melt and runoff from atmospheric variables from their associated polar RCM over Greenland. The emulators use a novel physics-informed, modular architecture that combines short-term weather patterns with long-term climate memory, capturing both immediate atmospheric forcing and accumulated firn characteristics.Our work demonstrates that machine learning can successfully emulate firn model behavior from climate forcing alone. This represents a crucial first step toward computationally efficient emulation of polar RCMs, facilitating generation of large ensembles, sensitivity analysis, and potential integration as surrogate models within Earth system models.

To mitigate agricultural greenhouse gas emissions, Danish ministerial agreements have initiated a land-use transformation of historical dimensions, focusing on restoring and rewetting extensive peatland areas currently used for agriculture. In addition, a CO2-equivalent tax on emissions from organic peatlands is scheduled to be implemented from 2028. This study informs discussions on requirements and best practices for rewetting and peatland restoration and highlights the importance of including changing climate conditions and rewetting management scenarios in future peatland management strategies.The study integrates process-based hydrological modeling and empirical CO2 flux modeling at a daily temporal resolution to evaluate how peatland hydrology influences CO2 emissions under scenarios of rewetting and climate change.Following the calibration of a three-dimensional transient distributed hydrological model for a peat-dominated catchment, daily groundwater table dynamics were simulated to represent hydrological conditions in drained peat soils. These simulations were coupled with an empirical CO2 flux model, developed from a comprehensive daily dataset of groundwater table depth, temperature, and soil CO2 flux measurements. The novel empirical CO2 flux model captures a clear temperature-dependent response of soil CO2 emissions to variations in groundwater table depth.By applying this coupled modeling framework, we quantified CO2 emissions at daily timescales. The results demonstrate that incorporating both temperature sensitivity and high-resolution temporal variability in water level significantly influence projections of CO2 fluxes. In particular, high CO2 emissions are expected in cases of co-occurrence of elevated air temperature and low groundwater tables. Using 17 different climate projections from the Euro-CORDEX regional climate modeling project, we simulated future groundwater table depth and temperature-dependent CO2 emissions. We find increased emissions due to increased temperatures, which, however, can be counter-balanced (in the Danish case) or amplified depending on the future trend in groundwater table depth.Our results further demonstrate that rewetting strategies that achieve near-surface groundwater tables mainly during winter result in only marginal emission reductions compared to drained conditions. Conversely, near-surface groundwater tables in summer offer more effective reductions (up to 50%).The study illustrates the value of combining detailed hydrological simulations with emission models.

The NAO’s correlation with precipitation in Norway and the Iberian Peninsula is well established, yet its explanatory power diminishes across much of Europe. Other patterns may drive precipitation variability in these regions, but traditional methods for identifying circulation-precipitation relationships have limitations. Prescribed indices assume a causal link between specific sites and physically coherent structures, whereas EOF methods impose linearity and orthogonality constraints that atmospheric circulation does not obey. This study identifies weather regimes associated with precipitation extremes across representative European regions using a non-linear, non-orthogonal, data-driven approach.Specifically, we employ a machine learning approach that discovers weather regimes directly from mean sea level pressure fields, without prescribing their structure a priori. The method builds upon Spuler et al. 2024 & 2025, with changes that allow for larger, higher-resolution input domains. It identifies distinct atmospheric states associated with different precipitation intensities at target locations, linking discovered patterns directly to their impacts. Importantly, once regimes are identified, indices analogous to traditional teleconnection indices can be derived, enabling comparison with established frameworks while capturing dynamics they may miss.We apply this method to daily ERA5 fields, targeting precipitation in selected European regions with contrasting dynamical drivers, including Bergen, the Iberian Peninsula, and Copenhagen. This allows us to present teleconnection patterns identified through this approach over the entire Northern Hemisphere as well as relevant sub-regions, including the North Atlantic and Arctic, focusing on extreme precipitation drivers. We find multiple regimes that resemble different flavors of the well-known NAO pattern, alongside circulation states consistent with blocking-like structures. Comparisons with traditional EOF analysis highlight the effects of relaxing linearity and orthogonality constraints. Correlation maps are produced for both methods, enabling direct evaluation of how the data-driven regimes compare to established EOF-based patterns.The non-linear, data-driven framework remains physically interpretable and avoids the limitations of linear orthogonal decomposition. Though currently applied to ERA5, the approach transfers directly to CMIP6 historical and scenario runs, enabling assessment of how regime frequencies and precipitation associations may shift under climate change. Overall, this study illustrates how ML-based approaches can complement traditional synoptic climatology by allowing circulation–impact relationships to emerge directly from the data.BibliographySpuler, Fiona R. et al. (2024): Identifying probabilistic weather regimes targeted to a local-scale impact variable. Environmental Data Science, 3, e25.Spuler, Fiona R. et al. (2025): Learning predictable and informative dynamical drivers of extreme precipitation using variational autoencoders. Weather and Climate Dynamics, 6, 995-1014.

Predicting surface melt on the Greenland ice sheet is critical for understanding surface mass balance (SMB) and sensitivity to climate change. Polar regional climate models are the primary tools for simulating melt and projecting future SMB, but different models produce significantly different results. However, they are too computationally expensive to create the large ensembles needed to quantify this uncertainty. We develop a neural network based emulator that predicts daily surface melt from atmospheric variables, trained on output from the polar regional climate model HIRHAM5 and its firn model DMIHH forced by ERA-Interim reanalysis. The emulator uses a physics-informed design combining short-term weather patterns with long-term climate memory, capturing both immediate atmospheric forcing and accumulated firn characteristics. The emulator achieves mean absolute error below 0.23 mm w.e. per day across all six Greenland drainage basins, with the errors primarily attributable to spatial over-smoothing. Our work demonstrates that machine learning can successfully emulate firn model behavior from climate forcing alone with computational costs orders of magnitude lower than traditional simulations. Once retrained for specific climate forcings, the emulator thus enables extensive ensemble projections. Furthermore, the modular architecture can be readily adapted to emulate other SMB quantities such as runoff. This represents a crucial first step toward computationally efficient emulation of polar regional climate models and surrogate modeling of SMB components in Earth system modeling.

Predicting surface melt on the Greenland ice sheet is critical for understanding surface mass balance (SMB) and sensitivity to climate change. Polar regional climate models are the primary tools for simulating melt and projecting future SMB, but different models produce significantly different results. However, they are too computationally expensive to create the large ensembles needed to quantify this uncertainty. We develop a neural network based emulator that predicts daily surface melt from atmospheric variables, trained on output from the polar regional climate model HIRHAM5 and its firn model DMIHH forced by ERA-Interim reanalysis. The emulator uses a physics-informed design combining short-term weather patterns with long-term climate memory, capturing both immediate atmospheric forcing and accumulated firn characteristics. The emulator achieves mean absolute error below 0.23 mm w.e. per day across all six Greenland drainage basins, with the errors primarily attributable to spatial over-smoothing. Our work demonstrates that machine learning can successfully emulate firn model behavior from climate forcing alone with computational costs orders of magnitude lower than traditional simulations. Once retrained for specific climate forcings, the emulator thus enables extensive ensemble projections. Furthermore, the modular architecture can be readily adapted to emulate other SMB quantities such as runoff. This represents a crucial first step toward computationally efficient emulation of polar regional climate models and surrogate modeling of SMB components in Earth system modeling.