General circulation models (GCMs) are widely used for future climate projections, but their coarse spatial resolution and systematic biases limit their direct use for impact studies. This limitation is particularly critical for wind-related applications, such as wind energy, which require spatially coherent, multivariate, and physically plausible near-surface wind fields. Classical statistical downscaling and bias correction methods partly address this issue. Still, they struggle to preserve spatial structure, intervariable consistency, and robustness under climate change, especially in high-dimensional settings. Recent advances in generative machine learning offer new opportunities for downscaling and bias correction, eliminating the need for explicitly paired low- and high-resolution datasets. However, many existing approaches remain difficult to interpret and challenging to deploy in operational climate impact studies. In this work, we apply SerpentFlow, an interpretable, generative, domain alignment framework, to the multivariate downscaling and bias correction of wind variables from GCM outputs. This is a method that generates low-resolution/high-resolution training data pairs by separating large-scale spatial patterns from small-scale variability. Large-scale components are aligned across climate model and observational domains. Conditional fine-scale variability is then learned using a flow-matching generative model. We apply the approach to multiple wind variables downscaling, including average and maximal wind speed, zonal and meridional components, and compare it with widely used multivariate bias correction methods. Results show improved spatial coherence, intervariable consistency, and robustness under future climate conditions, highlighting the potential of interpretable generative models for wind and energy applications.

Subseasonal forecasts provide valuable guidance for decision-making but remain limited by systematic biases and coarse spatial resolution, particularly over regions with complex orography. Quantile mapping (QM) is widely used for post-processing because it enables effective bias correction and downscaling, yet it applies a distributional correction and therefore neglects the case-specific dependence between forecasts and observations. To move toward a more integrated approach, we explore a diffusion-based generative framework for simultaneous probabilistic bias correction and high-resolution downscaling, with lead time explicitly included as a conditioning variable. We employ a conditional denoising diffusion model (DM) trained on high-resolution gridded observations (2 km) and conditioned on coarse-resolution ensemble-mean subseasonal forecasts (36 km) from the European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS) subseasonal system. We focus here on a univariate temperature model and on training strategies. By learning conditional distributions of observations given forecasts and lead time, diffusion-based generative models enable case-dependent corrections that exploit forecast–observation relationships. Two training configurations are examined to assess how lead-time conditioning and data availability influence the learning of biases. Results show that the DM recovers realistic fine-scale temperature patterns and achieves meaningful lead-time-dependent bias correction across subseasonal lead times, even when trained on relatively limited datasets. However, compared to QM, the diffusion-based forecasts exhibit lower skill in probabilistic metrics such as the continuous ranked probability score (CRPS) and discrimination for extremes. Overall, this work represents a first step toward lead-time-aware subseasonal post-processing using DMs and highlights directions for future methodological development.

Multi-model ensembles are widely used in climate science, yet Coupled Model Intercomparison Project Phase 6 (CMIP6) models are neither independent nor equally skillful. We present a framework for interpretable ensemble learning that improves prediction while making model contributions explicit. Using monthly CMIP6 near-surface temperature and precipitation (1948–2014) against ERA5, we compare arithmetic multi-model ensemble (AMME) averaging with Ridge regression, random forest, and a continuous gating mixture that adaptively interpolates between them. On held-out test years (2006–2014), gating achieves the best temperature performance and improves precipitation distributional behavior relative to both AMME and individual learners. We trace these gains back to individual models and space: dominant-cluster and dominant-model maps show that gating reallocates trust in regionally coherent patterns rather than collapsing to a single best model, while entropy-based diagnostics quantify where trust is concentrated versus distributed. These analyses provide a practical pathway for transparent ensemble learning that supports both predictive skill and scientific accountability.

Next-generation high-resolution (km-scale) climate models promise unprecedented accuracy in climate projections, but realizing their potential requires robust methods to quantify how well simulations align with real-world observations. Average-based metrics conventionally used for climate model evaluation ignore the physics encoded in the fine-scale structures of km-scale simulations. To overcome this limitation, we propose a novel, statistically principled evaluation methodology based on the likelihood function of a generative image model. Our method provides a continuous similarity metric derived from the likelihood distribution of observation and simulation snapshots, which can redefine the evaluation, intercomparison, and parameter tuning of high-resolution climate models. We demonstrate the applicability and interpretability of this method by evaluating convective clouds simulated by two state-of-the-art global km-scale models, using their outgoing infrared radiation fields. This work establishes a scalable pathway toward observation-based evaluation of next-generation climate simulations.

Recent advances in visual-language models (VLMs) have led to significant improvements in a plethora of complex multimodal tasks like image captioning, report generation, and visual perception. However, generating text from meteorological data is highly challenging because the atmosphere is a chaotic system that is rapidly changing at various spatial and temporal scales. Given the complexity of atmospheric phenomena, it is critical to verifiably quantify the effectiveness of existing VLMs on weather forecasting data. In this work, we present SynopticBench, a high-quality dataset consisting of 1,367,041 text samples of Area Forecast Discussions created by the National Weather Service over the continental United States paired to images of 500 mb geopotential height, 2 m temperature, and 850 mb wind velocity in weather forecasts. We also present Synoptic Phenomena Alignment and Coverage Evaluation (Space), a novel evaluation framework that can be used to effectively estimate the quality of text descriptions of synoptic weather phenomena. Extensive experiments on generating forecast discussions using state-of-the-art VLMs show the sensitivity of existing evaluation metrics in this domain and enable further exploration into synoptic weather and climate text generation.

In recent years, deep learning has transformed data processing, justifying its use in the current study. These models consist of multiple layers that calibrate weights to optimize accuracy. By mimicking neural processes, these networks facilitate high-performance predictive modeling. This study focused on the application of convolutional neural networks for forecasting Sargassum seaweed arrivals in Guadeloupe, using varied climatic data such as winds, surface currents, satellite images, and observations of seaweed arrivals on the Guadeloupe coast. We tested two different approaches: classification models for forecasting up to 14 days and zone regression models for forecasting up to 14 days, also indicating a probability of algae arrival. The results showed that the classification models achieved a mean accuracy of 94.75%, a mean specificity of 97.42%, and a mean sensitivity of 98.78%. Zone regression models achieved a mean accuracy of around 90%, a mean specificity of around 97%, and a mean sensitivity of around 85%, with a mean MAE of 0.13 and a mean RMSE of 0.18. These models perform better than the decision tree proposed in the state of the art for the same problem.

The “differentiability gap” presents a primary bottleneck in Earth system deep learning: Since models cannot be trained directly on non-differentiable scientific metrics and must rely on smooth proxies (e.g., MSE), they often fail to capture high-frequency details, yielding “blurry” outputs. We develop a framework that bridges this gap using two different methods to deal with non-differentiable functions: The first is to analytically approximate the original non-differentiable function into a differentiable equivalent one; the second is to learn differentiable surrogates for scientific functionals. We formulate the analytical approximation by relaxing discrete topological operations using temperature-controlled sigmoids and continuous logical operators. Conversely, our neural emulator uses Lipschitz-convolutional neural networks to stabilize gradient learning via: (1) spectral normalization to bound the Lipschitz constant and (2) hard architectural constraints enforcing geometric principles. We demonstrate this framework’s utility by developing the Minkowski image loss, a differentiable equivalent for the integral-geometric measures of surface precipitation fields (area, perimeter, and connected components). Validated on the EUMETNET OPERA dataset, our constrained neural surrogate achieves high emulation accuracy, completely eliminating the geometric violations observed in unconstrained baselines, which generate physically impossible precipitation fields in up to 7.8% of cases. However, applying these differentiable surrogates to a deterministic super-resolution task reveals a fundamental trade-off: While strict Lipschitz regularization ensures optimization stability, it inherently over-smooths gradient signals, restricting the recovery of highly localized convective textures. This work highlights the necessity of coupling such topological constraints with stochastic generative architectures to achieve full morphological realism.

Wind energy stands out as a promising clean and renewable energy alternative, not only for its potential to combat global warming but also for its capacity to meet the ever-growing demand for energy. However, analysis of wind data to fully harness the benefits of wind energy demands tackling several related challenges: (1) Current data resolution is inadequate for capturing the detailed information needed across diverse climatic conditions; (2) Efficient management and storage of real-time measurements are currently lacking; (3) Extrapolating wind data across spatial specifications enables analysis at costly-to-measure, unobserved points is necessary. In response to these challenges, we introduce the One Stone Three Bird model, a modality-agnostic learning framework utilizing Implicit Neural Network. Our model effectively compresses a large volume of climate data into a manageable latent codec. It also learns underlying continuous climate patterns, enabling reconstruction at any scale and supporting modality transfer and fusion. Extensive experimental results show consistent performance improvements over existing baselines in both (1) continuous super-resolution reconstruction and (2) data compression tasks for different cross-altitude prediction scenarios. Through systematic ablation studies, we demonstrate the effectiveness of each core component, quantifying its individual contribution to the overall performance of the proposed design.

Boreal forests play a critical role in the global carbon cycle as they are one of the largest terrestrial carbon sinks globally. In this study, we employ explainable machine learning (ML) techniques to investigate the influence of environmental and vegetation variables on net ecosystem exchange (NEE), focusing specifically on the effects of diffuse radiation. We utilize a sub-hourly resolution data set including satellite and in-situ observations from three boreal or hemiboreal forest research stations across the latitudes 58–68° N to capture latitudinal variability in forest carbon uptake. Using SHAP (Shapley Additive Explanations) values, we identify key drivers influencing NEE and quantify their importance across various ML model architectures. Photosynthetically active radiation (PAR), diffuse radiation, normalized difference vegetation index, and soil temperature were identified as the variables having the largest explanatory power for NEE across the ML models. ML models using only these variables result in $ {R}^2\approx 0.8 $ and RMSE$ \approx 2.3 $ $ \mu \mathrm{mol}\;{\mathrm{m}}^{-2}\hskip0.1em {\mathrm{s}}^{-1} $. Further analysis of SHAP values indicates that higher diffuse radiation (DiffPAR) is associated with more negative NEE (stronger carbon sink). SHAP analysis highlights this effect much more clearly than raw DiffPAR measurements, because it accounts for interactions with other environmental factors. This suggests that the diffuse radiation effect emerges from interactions between DiffPAR and co-varying environmental factors (such as cloudiness and total PAR). Cases identified by SHAP ($ \mathrm{DiffPAR}\ \mathrm{SHAP}<-2 $) have a median NEE $ 1.55\mu \mathrm{mol},{\mathrm{m}}^{-2},{\mathrm{s}}^{-1} $ more negative than cases with raw $ \mathrm{DiffPAR}\ge 400 $. When applied to ML models, SHAP uncovers nonlinear, context-dependent interactions between diffuse radiation and other drivers of NEE without assuming a priori relationships.

Conformal prediction (CP) can yield statistically valid prediction intervals for any regression model, with no model modifications and small computational costs. To assess its practical value, we apply conformal methods to quantify uncertainty in machine learning emulators of six microphysical process rates (MPRs). MPRs describe small-scale processes in atmospheric clouds such as precipitation formation and aerosol–cloud interactions and help understand weather and climate. The emulators are trained on simulation output from the ICOsahedral Nonhydrostatic (ICON) model in a limited-area numerical weather prediction configuration. We compare split CP for deterministic emulators with conformalized quantile regression (CQR) for quantile regression (QR) emulators. Both CP methods yield well-calibrated and sharp prediction intervals on average, but CQR provides more consistent intervals across several orders of magnitude, making it preferable for the uncertainty quantification of climate variables.

Accurate short-term precipitation forecasts predominantly rely on dense weather-radar networks, limiting operational value in places most exposed to climate extremes. We present TUPANN (Transferable and Universal Physics-Aligned Nowcasting Network), a satellite-only model trained on GOES-16 RRQPE. Unlike most deep learning models for nowcasting, TUPANN decomposes the forecast into physically meaningful components: a variational encoder–decoder infers motion and intensity fields from recent imagery under optical-flow supervision, a lead-time-conditioned MaxViT evolves the latent state, and a differentiable advection operator reconstructs future frames. We evaluate TUPANN on both GOES-16 and IMERG data, in up to four distinct climates (Rio de Janeiro, Manaus, Miami, La Paz) at 10–180-min lead times using the CSI and HSS metrics over 4–64 mm/h thresholds. Comparisons against optical-flow, deep learning, and hybrid baselines show that TUPANN achieves the best or second-best skill in most settings, with pronounced gains at higher thresholds. Training in multiple cities further improves performance, while cross-city experiments show modest degradation and occasional gains for rare heavy-rain regimes. The model produces smooth, interpretable motion fields aligned with numerical optical flow and runs in near real time due to the low latency of GOES-16. These results indicate that physically aligned learning can provide nowcasts that are skillful, transferable, and global.

Accurately downscaling precipitation from coarse to high spatial resolutions remains a critical challenge in climate and hydrometeorological modeling. A key limitation is the frequent misclassification of dry and wet regions, which compromises the realism and reliability of high-resolution outputs. To address this, we propose a deep learning-based downscaling framework that explicitly models dry/wet classification by transforming low-resolution 6×6 precipitation inputs into high-resolution 60 × 60 binary classification fields. We evaluate two architectures, a convolutional encoder-decoder and a conditional Wasserstein generative adversarial network (WGAN), utilizing three training strategies: (1) using binary wet/dry inputs, (2) using precipitation intensity inputs, and (3) using precipitation intensity inputs and adding physical constraints. Models are trained and validated on both synthetically generated precipitation fields and real radar-estimated hourly precipitation data over the contiguous United States. Performance is assessed using metrics including the overall probability of zero ($ {\mathrm{P}}_0 $) and spatial autocorrelations. Results show that incorporating intensity information improves dry/wet classification, while adding physical constraints further enhances accuracy, generalization, and physical consistency, especially for WGAN models. The convolutional encoder-decoder produces smoother outputs with stable performance regarding marginal statistics, whereas the WGAN generates sharper boundaries and more realistic dry/wet fields, improving on the spatial dependence structure. Furthermore, we demonstrate that the derived dry/wet classification fields can be used as binary masks to bias correct downscaled precipitation fields, enhancing both statistical fidelity and spatial realism. These findings highlight the value of a physically informed bias-correction strategy for improving the spatial realism of high-resolution precipitation fields from coarse-scale inputs.

The European electricity power grid is transitioning toward renewable energy sources, characterized by an increasing share of offshoreand onshore wind and solar power. However, the weather dependency of these energy sources poses a challenge to grid stability, with so-called Dunkelflaute events—periods of low wind and solar power generation—being of particular concern due to their potential to cause electricity supply shortages. In this study, we investigate the impact of these events on the German electricity production in the years and decades to come. For this purpose, we adapt a recently developed generative deep learning framework to downscale climate simulations from the CMIP6 ensemble. We first compare their statistics to the historical record taken from ERA5 data. Next, we use these downscaled simulations to assess plausible future occurrences of Dunkelflaute events in Germany under the optimistic low (SSP2–4.5) and high (SSP5–8.5) emission scenarios. Our analysis indicates that both the frequency and duration of Dunkelflaute events in Germany in the ensemble mean are projected to remain largely unchanged compared to the historical period. This suggests that, under the considered climate scenarios, the associated risk is expected to remain stable throughout the century.

This paper evaluates the performance of baseline and domain-augmented ChatGPT models for literature-based knowledge support in flood susceptibility mapping (FSM) using machine Learning approaches. To assess this, we designed five key questions related to FSM, with benchmark responses derived from our comprehensive review article (Pourzangbar et al., Journal of Flood Risk Management 18, e70042), which analyzed 100 studies on ML applications in FSM. The same questions were posed (i) to standard ChatGPT-4 and ChatGPT-4o models without additional contextual material, and (ii) to a domain-augmented GPT-4 configuration (Chat-FSM) equipped with retrieval access to the 100 reviewed articles. The comparison highlights that GPT-based models can reasonably reproduce frequently reported machine learning models and conditioning factors from the reviewed literature, but show weaker consistency in feature selection methods, often suggesting less relevant techniques. Among the models, ChatGPT-4o demonstrated the weakest alignment with benchmark data, while Chat-FSM demonstrated the highest agreement with the benchmark dataset across most evaluated questions. In terms of application-level efficiency, GPT models required substantially less time and computational effort compared to manual literature synthesis under the defined experimental setup. While ChatGPT-based systems can support literature-informed exploration in FSM, human expertise remains essential for critical reasoning, methodological design, and application to novel or context-specific scenarios.

Meteotsunamis—tsunami-like sea level oscillations generated by atmospheric disturbances—pose underestimated risks to coastal regions worldwide. Despite growing evidence of their frequency and impact, limited offshore observations and forecasting capabilities hinder effective monitoring and early detection. Here, we present a data-driven framework for identifying and characterizing meteotsunami dynamics using sparse observational data. Leveraging dynamic mode decomposition and clustering techniques, we extract dominant spatiotemporal patterns and optimize the placement of offshore monitoring stations. We demonstrate the effectiveness of this approach using high-resolution simulations of the 2022 Ireland meteotsunami, a well-documented event exhibiting clear atmospheric forcing and sea-level response. Our results show that a minimal network of five strategically positioned sensors can accurately capture the essential dynamics of the event. This framework establishes a scalable methodology for designing cost-effective monitoring systems, enhancing our ability to detect and understand meteotsunamis under data-scarce conditions.

This study investigates the use of topological data maps for extracting unique tropical cyclone (TC) wind field features. These maps are presented as graphs generated through a sequence of steps that filter, cluster, and identify data structure, and are used to characterize topological properties and shape in the data. The objective and scope of the method is explored through application to wind fields from the HURDAT2 data set, and its viability for detecting anomalous behavior in TCs is considered. We refer to the resulting graphs as wind field connectivity signatures (WFCS) or collective wind field connectivity map (CWFCM), depending on the data set. Our focus is Hurricane Sandy, where the method successfully identifies a complete 360-degree rotation of the high wind speed radii. This cyclical example of phase rotation of wind speed asymmetries corresponds to a distinct structural property of the graph. These methods have not been previously applied to wind field data and have only seen limited use in atmospheric sciences.

Forest fires alter soil organic carbon and suppress soil respiration for decades following disturbance. However, uncertainties in model parameterization and sensitivity hinder robust predictions of autotrophic and heterotrophic soil respiration responses. We addressed this challenge using a novel dataset from a fire chronosequence in the Yukon and Northwest Territories of Canada. The dataset included field measurements of total soil respiration at four sites with varying time since fire, supplemented by field measurements of soil temperature at two depths, remote sensing data on aboveground productivity, and soil moisture at two depths. We evaluated a suite of soil respiration models, ranging from exponential $ {Q}_{10} $ formulations to heterotrophic respiration models using Michaelis–Menten kinetics. To estimate parameters efficiently, we (1) derived algebraic expressions for soil respiration components assuming quasi-steady state dynamics and (2) applied a Markov Chain Monte Carlo (MCMC) approach for parameter estimation. The resulting parameter estimates revealed which parameters were well-constrained and where uncertainty remained. Modeled respiration agreed with established empirical relationships and highlighted two key findings: (1) all chronosequence sites favored models that explicitly included microbial carbon as a distinct pool, and (2) parameters related to aboveground litter inputs were better constrained than those for root turnover. These results held regardless of soil depth or the form of the autotrophic respiration moisture response. These findings indicate that direct field measurements of litterfall rates would reduce model uncertainty, and that targeted sampling during seasonal transitions (e.g., freeze–thaw periods) would provide critical constraints on microbial activity when respiration dynamics are most variable.

Marine heat waves (MHWs) are prolonged periods of elevated ocean temperatures that can devastate marine ecosystems, fisheries, and coastal communities. Skillfully predicting these events with sufficient lead time is crucial for mitigating their adverse effects. This study presents a probabilistic subseasonal MHW forecast tool using a U-Net-based neural network architecture, with a focus on the Northern Indian Ocean and the Arabian Sea. The model was trained using sea surface temperature and sea surface height reanalysis data. The U-Net-based forecast tool demonstrated significant predictive skill up to 10 weeks in advance across various deterministic and probabilistic skill metrics. The model outperformed persistence and climatology-based benchmarks, especially in the tropical warm pool. Future applications of explainable artificial intelligence (XAI) methods have the potential to identify the sources of predictive skill, inform understanding of underlying dynamics, and improve dynamic subseasonal to seasonal forecast models.

This study aims to evaluate the thermal behaviors of surface materials in arid climates to enhance environmental sustainability and energy efficiency. Conducted over 1 year at Dokumapark in Antalya, Turkey, it examines surface temperatures of asphalt, concrete, granite, wood, grass, and soil using thermal using a FLIR-C5 thermal camera. Measurements were taken in the morning, noon, and evening, capturing images from sunny and shaded areas, which were processed with custom Python software. A total of 1728 temperature values were statistically and visually analyzed based on surface–air temperature differences.

Seven machine learning models were used for evaluation, with the neural network model achieving the highest accuracy (R2: 0.9848) and minimal error. The model assessed thermal variations across different periods. Grass and wood exhibited low heat retention, while asphalt and brick reached higher temperatures, with asphalt predicted to exceed 50 oC in summer, potentially impacting thermal comfort. Grass was the most efficient material with minimal temperature fluctuations.

This study highlights the importance of thermal properties in enhancing energy efficiency and user comfort, as well as the necessity of selecting materials for sustainable cities. It suggests that combining artificial intelligence and thermal imaging techniques can be a beneficial tool for ecological and sustainable architectural design.