Papers for

weather forecasting teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Models improve forecasts when text agrees with sensor data trends

Aligned, Conflicting, or Unspecified? A Diagnostic Benchmark for Multimodal Time Series Forecasting

Abstract: Artificial Intelligence of Things (AIoT) systems increasingly combine sensed time series with external text such as expert reports and news, yet how reliably models use text that agrees or conflicts with the sensed history remains unclear. We introduce the first real-world diagnostic benchmark organized by text-history relationship: 1,762 text-window pairs across sensor telemetry (Weather), web-service traffic (Wikipedia), and market data streams (Finance), each labeled Aligned, Conflicting, or Unspecified by whether the text's direction agrees with the historical trend, opposes it, or is not stated. Across ten pretrained models evaluated zero-shot, we investigate text selection and combination under text-history relationships (RQ1) and error diagnosis and horizon-scaled revision control (RQ2). For RQ1, we compare two types of text on paired Weather windows: one directly forecasts temperature changes, while the other describes related weather conditions, such as winds and fronts. Under Aligned, the first type yields the lowest error for most models. Under Conflicting, it increases most models' error above the no-text baseline; we therefore add only its stated magnitude and timing to the second type. Compared with either type alone, this combination lowers mean absolute scaled error (MASE) for six models and raises directional accuracy (DA) for seven. For RQ2, an error decomposition attributes the error increase to oversized revisions, and a training-free rule that limits them outperforms forecasts both with and without text for the four models with this trade-off, lowering MASE by 4.49%-21.98% and raising DA by 2.99%-9.71%. Our dataset is available at https://github.com/Shelton1202/ACU-Bench.

Fri 13 NovArtificial Intelligence
The gist
Forecasting systems often combine sensor data with related text like weather reports or news, but it is unclear how well they handle situations when the text agrees or disagrees with the sensor data. The authors created a benchmark with examples labeled as aligned, conflicting, or unspecified based on how the text relates to the sensor history. They found that when text aligns with the data, models forecast better, but conflicting text can harm accuracy unless carefully combined. They also designed a simple rule to limit excessive forecast changes caused by conflicting text, which improved accuracy for some models.
Open → 2611.17680v1

Diagonal spherical neural operators enable precise Earth system forecasts

DIAGNO: Diagonal Spherical Neural Operators for Heterogeneous Earth Dynamics Modeling

Abstract: Accurate forecasting of Earth's dynamic systems is vital in geophysics, where neural operators offer a powerful data-driven paradigm. Standard operators often treat the Earth as a flat Euclidean plane, introducing geometric artifacts. In contrast, Spherical Neural Operators (SNOs) resolve this mismatch by operating within the spherical spectral space, naturally adapting to spherical geometries. However, for computational efficiency, existing SNOs typically rely on the rotation equivariance assumption. This assumption simplifies the transitions of physical states into isolated spectral modal evolutions, thereby over-idealizing the Earth as a homogeneous system. To solve this problem, we no longer rely on the rotation equivariance assumption that isolates spectral modal evolution, but directly derive an explicit modal interaction kernel and propose Diagonal Spherical Neural Operators named DIAGNO. DIAGNO explicitly parameterizes cross-modal interactions through spectral modal decomposition, thereby preserving the inherent heterogeneity of Earth dynamics. Furthermore, leveraging the unique diagonal structure of the spherical spectrum, we design inter- and intra-diagonal interaction mechanisms to natively capture essential zonal and meridional dynamics, respecting intrinsic geophysical regularities. Extensive experiments across simulated fluid dynamics and real-world Earth system reanalysis (including atmosphere and ocean) datasets demonstrate DIAGNO's superior capability, consistently achieving state-of-the-art performance in multi-step forecasting. Beyond this, we validate DIAGNO's practical analytical value in professional geophysical domains by assessing its fidelity in capturing climatological anomalies and maintaining physical consistency.

Thu 12 NovArtificial IntelligenceMachine Learning
The gist
Earth’s weather and ocean systems are complex and happen on a round planet, but many computer models simplify Earth as flat, which causes errors. The authors created a new type of neural operator called DIAGNO that works directly with spherical data and considers different ways Earth's systems vary across the globe. This approach captures important geographical patterns and is better at predicting future changes in Earth systems. Their tests show DIAGNO outperforms previous models and stays consistent with known weather and ocean behaviors.
Open → 2611.16356v1

Multigrid neural operator improves fluid flow prediction accuracy

SUMNO: Spatially Unrolled Multigrid Neural Operator for Fluid Flow Surrogates

Abstract: Data-driven surrogate modeling has emerged as a powerful paradigm for accelerating computational fluid dynamics, and hierarchical U-shaped networks, which resemble the structure of classical Geometric Multigrid (GMG) methods, have become prevalent designs. Modern surrogates built on these designs achieve competitive accuracy at a fraction of the cost of numerical solvers; however, accurate prediction of complex flows remains challenging, and no existing architecture leads across them. In such networks, the multigrid analogy mainly concerns the multi-resolution topology, while their components largely follow common vision practice. To address this, this study introduces SUMNO, a multigrid neural operator for fluid flow fields on regular grids that spatially unrolls a V-cycle into a feed-forward U-shaped network. Parameter-free down- and upsampling stand in for restriction and prolongation, and additive skip connections for coarse-grid correction, with normalization-free transfer and skip pathways. Within this macro-architecture, convolutions are used at the fine and intermediate resolutions and shifted-window attention at the coarsest, whose window partitioning is generalized to handle periodicity independently per axis. Evaluated on unseen test configurations across four unsteady and steady flow benchmarks against five representative baselines, SUMNO is best in 5 of 7 settings and ranks second in the remaining two. It reduces the 10-step rollout MSE by about 30% relative to the strongest baseline, and outperforms three baselines with over 130 million parameters using 9.4 million parameters.

Thu 12 NovArtificial Intelligence
The gist
Simulating fluid flows like air or water is usually very slow and costly. The authors developed a new neural network called SUMNO that mimics traditional multigrid methods but processes fluid data differently to predict flow patterns faster and more accurately. They tested it on several fluid flow problems and found it often outperforms existing models while using fewer parameters. This could help make fluid simulations more efficient for various applications.
Open → 2611.16341v1

Climate temperature forecasting improved with efficient deep learning model

Climate Temporal Temperature Prediction via an Adaptive Stacked Bidirectional Residual Mamba Network

Abstract: Accurate temperatures forecasting relies on traditional meteorological parameters that are essential for monitoring weather informatics and guiding forecasting efforts. This study presents Deep DualMamba, a novel deep learning framework for temporal temperature prediction, built upon an enhanced variant of the recently proposed Mamba network. Unlike conventional Transformer-based architectures, which dominate foundation models but suffer from computational inefficiencies on long sequences, Deep DualMamba leverages selective structured state space models (SSMs) to achieve both scalability and performance. We identify a core limitation of prior subquadratic-time models-namely, their inability to perform long-term and robust forecasting and introduce two key innovations. First, we propose an adaptive bidirectional Mamba as encoders to learn complex and efficient features. Second, we develop a hardware-aware, parallel recurrent algorithm to compensate for the loss of convolutional efficiency, ensuring practical viability at scale. Last, the Deep DualMamba's approximation ability and convergence is theoretical analyzed. These design choices culminate in a simplified, attention- and MLP-free neural architecture that exhibits linear complexity with sequence length and significantly improved inference throughput. To demonstrate its applicability in climate informatics, we implement a full forecasting pipeline using historical meteorological data from UK monitoring stations.

Thu 12 NovMachine Learning
The gist
Predicting temperature changes over time is important for weather monitoring but can be challenging due to complex data and long sequences. The study introduces Deep DualMamba, a new deep learning model that can handle long and complicated climate data efficiently without the heavy computing costs of other models. The authors improve previous methods by designing a structure that learns from both past and future data points and runs faster on typical hardware. They also provide a theoretical basis for why their approach works well. Their system was tested using historical temperature data from UK weather stations to show its usefulness.
Open → 2611.16322v1

Improving high-resolution future weather forecasts with low-rank adaptation

PROLA: Principal-Orthogonal Low-rank Adaptation for Predictive Spatiotemporal Weather Downscaling

Abstract: High-resolution weather prediction is important for resolving local atmospheric patterns missed by coarse global forecasts. Large pretrained weather backbones have improved global forecasting, but adapting them to produce future high-resolution fields from recent coarse states remains challenging and costly. We formulate this setting as Predictive Spatiotemporal Weather Downscaling (PSWD), a task that jointly refines spatial and temporal resolution while predicting future high-resolution trajectories. To this end, we propose a framework that decomposes prediction into spatial and temporal stages using a shared pretrained backbone, fixed numerical scaffolds, and lightweight residual heads. For backbone adaptation, we introduce PROLA (PRincipal-Orthogonal Low-rank Adaptation), which allocates a fixed low-rank budget between pretrained weights' principal directions and their orthogonal complement. PROLA further rescales rank-1 optimizer updates using gradient signals, improving adaptation without increasing trainable rank. Experiments in the PSWD setting show that PROLA outperforms representative low-rank adaptation baselines. Further experiments across diverse datasets and tasks support the broader applicability of PROLA, showing that its benefits extend beyond the primary weather downscaling setting.

Sun 8 NovMachine LearningArtificial Intelligence
The gist
Predicting detailed, high-resolution weather in the future is hard when starting from rough, low-resolution weather data. The paper’s authors created a new way to improve these forecasts by breaking down the problem into spatial and temporal parts, using an existing weather model with small, smart adjustments. Their method, called PROLA, changes only a small part of the model’s settings but does so more effectively by focusing on key directions. Tests show this approach works better than other techniques for improving weather forecasts, and it might help in other areas too.
Open → 2611.11318v1

Multi-link satellite signals improve rainfall detection accuracy

Evaluation of a Multi-Link Machine Learning Approach for Rainfall Detection Using Satellite Microwave Links

Abstract: The impacts of climate change have increased the need for reliable real-time rainfall monitoring systems. This study investigates rainfall detection using a network of opportunistic sensors based on satellite microwave links (SMLs) combined with machine learning (ML) classifiers. Two models, a random forest (RF) and a convolutional neural network (CNN), were trained on aggregated SML data, including information from neighboring links. The results demonstrate that leveraging information from multiple links substantially enhances the detection of both wet and dry periods compared to single-link approaches. Moreover, the RF model consistently provides a more stable and reliable trade-off than CNN, improving rainfall detection and reducing misclassifications overall, with particularly strong performance during harsh weather conditions. The study spanned up to 32 rainy days between 2021 and 2024.

Fri 6 NovArtificial IntelligenceMachine Learning
The gist
Reliable rainfall detection is important, especially as weather changes become more extreme. This study shows that using data from multiple satellite microwave links together helps computers better tell when it is raining or dry. The researchers tested two types of machine learning models and found that the random forest model gave more dependable results, especially in bad weather. This approach could make rain monitoring faster and more accurate in real time.
Open → 2611.09156v1

Stable neural dynamics benefit from unified gradient regularization methods

Unrolled gradients in disguise: bridging interpolation-based and Jacobian regularization for stable neural dynamics

Abstract: Hybrid models and neural differential equations (NDEs) are becoming increasingly important for modelling physical systems; however, they often encounter stability and accuracy issues during long-term integration. Training on unrolled trajectories is known to limit these divergences but quickly becomes too computationally expensive due to the need to compute gradients over an iterative process, making this method impractical for large-scale physical systems (e.g. weather and climate forecasting). Two popular paradigms for improving the stability of learned dynamics are to train on interpolated or noisy samples or to regularize the Jacobian of the learned representation. In this work, we demonstrate that these paradigms can be understood under a common lens and that they both correspond to an approximation of the gradients obtained in an unrolled learning strategy. Building on this understanding, we derive a principled and complete approach to regularize training and demonstrate its impact on the stability of learned dynamical systems.

Mon 2 NovMachine Learning
The gist
Models that simulate physical systems often run into problems when predicting too far ahead, causing mistakes to grow. The authors found that two ways to keep these models stable—training on in-between or noisy data and controlling how sensitive the model is—actually share a similar underlying trick. They showed how these tricks approximate a more expensive training method that looks ahead step-by-step. Using this insight, the authors developed a new, more complete way to train these models so they stay stable longer.
Open → 2611.03461v1

Machine learning ensemble improves weather forecast uncertainty estimates

Sonny-ENS: demystifying machine-learning ensemble weather forecasting

Abstract: Every weather forecast is uncertain, and decisions from flood warnings to energy trading depend on how well that uncertainty is quantified. Machine-learning (ML) ensembles now quantify it better than the best physics-based ensemble, yet two things about them remain opaque: whether their skill requires hundreds to thousands of accelerator-days of training, and how a network with a simple source of randomness produces a spread that follows the flow. Here we introduce Sonny-ENS, an ensemble of three 68.9-million-parameter networks with Monte Carlo dropout, trained on the continuous ranked probability score (CRPS) in 2.7 A40 GPU-days. On 2022, it has a lower CRPS than the European Centre for Medium-Range Weather Forecasts ensemble (ENS) on 96% of 570 targets and an 8.3% lower mean CRPS at lead times of 1-10 days. It outperforms the operational GenCast on 69% of the targets, and a 51-member, 15-day forecast takes 69 s on one GPU. Ablations identify the ingredients of this skill, and a Kolmogorov-Arnold sparse autoencoder shows that dropout injects uncertainty mainly in the tropics, whereas the spread grows where the atmosphere is baroclinically unstable (spatial correlation 0.91 with the Eady growth rate). Skillful probabilistic weather forecasts are therefore within reach of a single GPU server.

Sun 1 NovMachine Learning
The gist
Weather forecasts are never certain, and knowing the range of possible outcomes is crucial for decisions like flood warnings and energy trading. The authors introduce Sonny-ENS, a new way to use machine learning that creates multiple weather predictions quickly and with better accuracy than current top models. It uses fewer computing resources than expected and shows that its uncertainty estimates closely follow real atmospheric conditions. This means useful, reliable probabilistic forecasts can now be run on a single GPU server.
Open → 2611.01614v1

Wavelet loss improves high resolution weather forecast accuracy and detail

Scale-aware and spatially-localised optimisation of kilometre-scale MLWP models

Abstract: Preserving fine-scale structure without degrading spatial accuracy is a central challenge for kilometre-scale machine-learning weather prediction (MLWP). We introduce a loss based on the dual-tree complex wavelet transform that constrains atmospheric fields jointly by scale, location, and orientation. We evaluate its performance using an experimental joint global--regional FastNet configuration, which produces \SI{1.5}{km} forecasts over the UK alongside $\sim$\SI{31}{km} global forecasts. Relative to both mean-squared error (MSE) and an FFT-based spectral objective, the wavelet loss substantially improves spectral fidelity. Compared with MSE training, it maintains similar pointwise performance, improves neighbourhood-based verification, and produces sharper precipitation fields with more realistic intensity distributions. These results demonstrate that spatially localised wavelet constraints can preserve multiscale atmospheric structure while largely avoiding the loss of pointwise accuracy associated with globally supported spectral objectives.

Tue 27 OctMachine Learning
The gist
Predicting weather accurately at very fine scales is tough because weather changes at many sizes and places. The authors developed a new way to train weather models using a wavelet-based loss function that looks at weather patterns by their size, location, and direction. This helps keep important details and sharp features in forecasts, like rain intensity, without losing accuracy at specific points. They showed this method works better than older ways on a combined global and regional weather prediction model.
Open → 2610.37325v1

Generative model reduces error in estimating missing scientific data steps

Generative Refinement Learning for Temporal Interpolation in Scientific Machine Learning

Abstract: Computationally expensive simulations and real-world observations often provide data at a lower temporal resolution than downstream tasks require. Averaging observations at fixed grid points can introduce errors, and motion estimates alone do not determine how the physical field values change. In this work, we study whether learning to correct a simple numerical interpolation improves the reconstruction of intermediate scientific fields. Our model, interpolative FLow EXpert (iFLEX), first estimates the missing field by taking a weighted average of the two observations at each grid point, giving more weight to the observation closer to the requested time. During training, subtracting this estimate from an observed intermediate field gives the residual that a conditional diffusion model learns to generate. Both observations condition the model, together with the target time, the interval between observations, and available physical parameters. To reconstruct an unobserved intermediate field, the generated residual is added to the weighted average. We evaluate iFLEX on Navier-Stokes Kraichnan turbulence, Shanghai radar precipitation, and sea surface temperature. Across these datasets, iFLEX achieves lower reconstruction error than the evaluated baselines, including optical-flow and diffusion models developed for video frame interpolation in computer vision. In a controlled turbulence comparison at a Reynolds number of 36,000, learning the residual reduces relative Frobenius norm error from about 13.7 for direct field generation to 10.5, with the diffusion backbone and training setup held fixed. Code is available at https://github.com/paulpuren/iflex.

Tue 27 OctMachine LearningArtificial Intelligence
The gist
Some scientific data come at slow speeds, so we don’t see what happens in between moments. The authors created a method called iFLEX to better guess what the data looks like in these missing times. Their method starts with a simple guess based on nearby moments and then uses a special AI model to improve that guess by learning the difference between the guess and true data. This approach works better than older methods when tested on things like weather and fluid flow data.
Open → 2610.36751v1

Vision transformers estimate sunlight from limited sky images

Vision Transformers for Estimating Irradiance Using Data Scarce Sky Images

Abstract: Accurate estimation of diffuse horizontal irradiance (DHI) is critical for optimising photovoltaic system performance and energy forecasting yet remains challenging in regions lacking comprehensive ground-based instrumentation. Recent advancements using Vision Transformers (ViTs) trained on extensive sky image datasets have shown promise in replacing costly irradiance measurement equipment, but the scarcity of long-term, high-quality sky imagery significantly restricts practical implementation. Addressing this critical gap, this study proposes a novel dual-framework approach designed for data-scarce scenarios. First, calculated atmospheric parameters, including extraterrestrial irradiance and cyclic time encodings, are integrated to represent sky conditions without utilising any instrumentation. Next, a sequential pipeline initially predicts synthetic global horizontal irradiance (GHI) and uses it as a feature, to refine DHI estimation. Finally, a dual-parallel architecture simultaneously processes raw and overlay-enhanced fisheye sky images. Overlays are generated through unsupervised, physics-informed cloud segmentation to highlight dynamic sky features. Empirical validation is performed using data from the Chilbolton Observatory, chosen for its temperate climate and frequent cloud variability. To simulate data-scarce conditions, models are trained on a single month (e.g., January) and evaluated across a temporally disjoint, full-year test set. Under this setup, the sequential and dual-parallel frameworks achieve RMSE values within 2-3 W/m^2 and 1-6 W/m^2, respectively, of a state-of-the-art ViT trained on the complete dataset. By combining physics-informed modelling with unsupervised segmentation, the proposed method provides a scalable and cost-effective solution for DHI estimation, advancing solar resource assessment in data-constrained environments.

Thu 22 OctArtificial IntelligenceComputer Vision and Pattern RecognitionMachine Learning
The gist
Measuring sunlight accurately is important for solar power but hard in places without many instruments. The authors developed a new method using vision transformers, a type of AI model, to estimate sunlight from a small number of sky pictures. They include calculations about the sun’s position and cloud patterns to help the model work better with little data. This method was tested in a cloudy climate and gave results close to those from larger data-based models, making it useful for places with limited measurement tools.
Open → 2610.30940v1

Observation assimilation transformer improves weather forecasts from sparse stations

Observation Assimilation Transformer: Probabilistic Weather Analysis from Sparse Station Observations

Abstract: We present the Observation Assimilation Transformer (OAT), a geometry-aware neural process that corrects and refines a prior numerical weather prediction (NWP) field into a high-resolution posterior field by assimilating sparse observations. OAT is an encoder-decoder transformer that considers a set of observed locations and their NWP prior to condition predictions at a set of arbitrary unobserved query locations. Its sparse attention mechanism makes it scalable to thousands of station observations and depends on relative positions between locations, never on the absolute coordinates. The model allows for partially missing observations, predictions at arbitrary continuous locations without grid projection, and outputs are parametric distributions. A station-randomized training algorithm is developed with continuous ranked probability score (CRPS) loss. We test OAT on two datasets - one at regional scale with hundreds of stations, and one at continental scale with thousands of stations. On these datasets OAT outperforms several baselines and shows well calibrated predictions. In an illustrative case study, OAT's downscaling capabilities are demonstrated by downscaling ERA5 (~31km) to COSMO (1km), where OAT achieves lower mean absolute error (MAE) and CRPS for temperature and precipitation on test stations compared to COSMO.

Tue 20 OctMachine Learning
The gist
Weather forecasts use data from many weather stations, but sometimes these observations are few and spread out. The authors created the Observation Assimilation Transformer (OAT), a model that improves weather predictions by smartly combining weather data and prior forecast information. It can make detailed, location-specific predictions even where there are no direct observations. Tests showed OAT gives more accurate and reliable weather details, for example turning coarse weather data into finer, more precise forecasts.
Open → 2610.28327v1

Precipitation forecasts can mislead when trained on wet-biased data

Wet-Biased Data Pipelines Mislead the Training and Evaluation of Generative Precipitation Nowcasting

Abstract: Generative models for precipitation nowcasting are trained and evaluated predominantly on wet-biased datasets, which over-represent precipitation events to emphasize rare extremes. Selecting test cases based on target observations is known to favor forecasts that match the selected distribution rather than the intended deployment distribution. We show analytically how this target-dependent sampling also shifts the conditional distribution learned by generative models. When training and test sampling match, even an ideal model learns precisely the distribution favored by evaluation, allowing the evaluation bias to conceal the training bias. Using a fixed satellite-to-radar diffusion model, we demonstrate that target-dependent test sampling can reverse the calibration assessment. The model appears under-predictive on a wet-biased test set but over-predictive on a deployment-representative test set. This result is replicated outside the model's training domain after a seven-year gap. Utilizing classifier-free guidance as an example, we demonstrate how the contradictory calibration assessment can misdirect downstream model development.

Mon 19 OctMachine Learning
The gist
Training artificial intelligence models to predict rain often uses datasets with too many rainy examples. The authors show this can make the models learn and be tested on an unrealistic view of weather, causing evaluations to be misleading. When tested on data that matches real-world conditions, the same model can seem to predict too much rain instead of too little. This means forecast developers might be led in the wrong direction when improving models based on biased tests.
Open → 2610.26486v1

Weather model speeds up forecasting by choosing tokens wisely

UtoMe: Observation-Uncertainty-Guided Token Merging for Weather Foundation Models

Abstract: Weather transformers process dense reanalysis fields whose reliability varies across variables, locations, and weather regimes. Token merging can reduce computation, but feature similarity alone does not account for this heterogeneous input uncertainty. We introduce UtoMe, an observation-uncertainty-guided token merging framework that combines dynamics saliency with ensemble data assimilation spread as a relative input-reliability proxy. The resulting importance score protects selected tokens as anchors and calibrates bipartite matching between the remaining tokens. Three variants use external uncertainty at both training and inference, during training only, or neither stage. In ERA5 72-hour forecasting at two spatial resolutions, UtoMe improves Overall error over train-time ToMe at matched token budgets. At $1.40625^\circ$ and 50% token reduction, UtoMe-E lowers Overall error from 67.32 to 53.00 at the same 0.55x encoder FLOPs. Small reductions remain close to the full-token baseline; larger reductions incur measurable accuracy losses. Component ablations and spatial uncertainty controls support combining dynamics with aligned uncertainty when allocating a weather transformer's token budget.

Sun 18 OctMachine LearningArtificial IntelligenceComputer Vision and Pattern Recognition
The gist
Weather forecasts use big computer models that look at lots of weather data, but not all data points are equally reliable. The paper introduces UtoMe, a method that helps the computer model pick important pieces of data (tokens) based on how certain the observations are and their weather dynamics. This smarter selection reduces the computing work while keeping forecast accuracy good, especially for big token reductions. The authors tested this on a common weather dataset and found their method improved forecast errors compared to earlier token merging approaches.
Open → 2610.25001v1

Spherical neural operator improves accuracy of meteorological downscaling

Spherical Physics-informed Neural Operator with Multi-scale Coupling for Meteorological Downscaling

Abstract: Meteorological downscaling is crucial for high-resolution regional climate forecasting and disaster early warning. While neural operators have emerged as a promising paradigm for modeling complex spatiotemporal mappings, existing frameworks often struggle with spherical manifold geometric distortions, inherent atmospheric multi-scale coupling mismatches, and lack of explicit atmospheric laws. We propose the Spherical Physics-informed Neural Operator, which utilizes a Spherical Laplacian Decomposition to partition atmospheric fields into hierarchical frequency components, maintaining exact point-wise correspondence across scales. To evaluate these representations at arbitrary locations, we introduce a localized spherical integral operator that approximates continuous kernel transforms via geometry-aware attention. Dynamical consistency is further enforced by embedding differentiable constraints into the learning process. Extensive experiments demonstrate that our framework attains superior accuracy and zero-shot generalization across various meteorological variables and unseen queries, representing a robust and interpretable solution for global-to-regional meteorological downscaling.

Sat 17 OctArtificial IntelligenceMachine Learning
The gist
High-resolution weather forecasts require translating global climate data to detailed local predictions, which is a tough problem especially over the curved Earth. The authors created a new neural network method that respects the spherical shape of the Earth and breaks down atmospheric data into different scales for better understanding. Their approach also includes physical laws to keep predictions realistic and works well even in areas not seen during training. This method helps provide more accurate and reliable local weather forecasts.
Open → 2610.24708v1

Ai models outperform traditional systems in tracking tropical cyclones

Can AI Forecast the Extremes? Evaluating Probabilistic Weather Models on Tropical Cyclones

Abstract: Weather prediction is undergoing an important shift, with data-driven AI models outperforming leading numerical weather prediction systems on average at a fraction of the computational cost. Yet, whether this success extends to extreme events remains contested. Earlier evaluations have exposed important shortcomings, but have largely focused on deterministic models and metrics, leaving the capabilities of the latest probabilistic models less understood. We conduct an extensive benchmark of tropical cyclone forecasting, comparing GenCast, FourCastNet 3, and FGN with ECMWF's operational physics-based ensemble, IFS-ENS. We develop a suite of probabilistic metrics and diagnostic analyses to assess cyclone tracking, intensity biases, uncertainty, and the wind-pressure relationship that characterizes storm structure. We find that, despite remaining limitations in intensity and ensemble spread, leading AI models substantially outperform IFS-ENS in cyclone tracking: FGN achieves better probabilistic location scores for 90% of evaluated storms in 2025. Further analyses reveal model-specific failure modes and suggest a lingering influence of pretraining data, pointing to opportunities for targeted improvement. Nonetheless, these results suggest that AI models are capable of capturing physical relationships that remain predictive well beyond typical weather conditions, motivating further investigation into how these capabilities emerge and how to strengthen them.

Thu 15 OctMachine Learning
The gist
Predicting weather storms like tropical cyclones is difficult, especially knowing exactly where they will go. The authors compared new AI models with a traditional physics-based weather system to see which predicts storm paths better. They found AI models can track storms more accurately and with better uncertainty estimates than the traditional system, though there are still challenges predicting storm intensity. These results suggest AI can learn important physical patterns in extreme weather conditions, offering new ways to improve forecasts.
Open → 2610.22077v1

Post-training fix improves extreme value modeling in diffusion models

Radial Tail Repair: Post-Training Correction for Diffusion and Flow Models

Abstract: Diffusion and flow-matching models may struggle to accurately represent distributional tails, especially for heavy-tailed data. Existing approaches often use heavy-tailed noise or modify the generative dynamics, requiring dedicated retraining and more difficult estimation. We take a simpler approach: rather than relearning the full distribution, we correct only the model's behavior in the extreme regime. Our method builds on the radial-angular decomposition of multivariate heavy tails, where radius and direction become asymptotically independent. This allows us to preserve the directional structure learned by the pretrained model and correct only its radial tail. We estimate the tail onset from reference and generated samples, then rescale generated extremes so that their radial survival matches the target distribution. The resulting method is a lightweight post-training correction applied to a frozen diffusion or flow-matching model, requiring neither retraining nor changes to its sampling dynamics. Experiments on challenging synthetic heavy-tailed distributions and real weather and wildfire datasets show that this simple correction substantially improves extreme-tail coverage using only a small number of fitting samples.

Mon 12 OctMachine Learning
The gist
Modeling rare extreme events is hard for diffusion and flow-based AI models, especially when data has heavy tails, which means extreme values happen more often than usual. The authors propose a simple fix that adjusts the model only in the extreme regions without retraining the whole model. Their method separates the direction and strength of data points and corrects just the strength for the extremes. This improves how well the model matches rare events in tests on synthetic and real weather and wildfire data.
Open → 2610.16851v1

Sensitivity fidelity reveals true physical learning beyond forecast accuracy

Sensitivity Fidelity Reveals Physical Dependencies Beyond Forecast Accuracy in Deep-Learning Earth-System Models

Abstract: Deep-learning models increasingly rival numerical simulation in Earth-system forecasting, yet forecast error alone cannot determine whether their learned input-output dependencies are physically consistent. Here we introduce sensitivity fidelity as an additional criterion: alongside prediction error, model responses to input perturbations should be examined against a trusted reference or established physical knowledge. In an idealized Rayleigh-Bénard convection benchmark, four surrogate architectures achieved low forecast errors yet encoded markedly different sensitivities. Quantitative comparison with a matched adjoint distinguished the models far more sharply than forecast accuracy and showed that predictive skill and sensitivity fidelity need not coincide, while finite-perturbation experiments further linked these differences to the propagation of initial-condition uncertainty. We then applied this strategy to El Niño-Southern Oscillation prediction, where no adjoint reference exists and learned sensitivities showed event-wise robustness, validity under finite perturbations, and consistency with established precursors. These results identify gradient-based evaluation as a discriminating complement to forecast-error metrics, providing a practical test of whether data-driven Earth-system models learn physically credible dependencies.

Sun 11 OctMachine Learning
The gist
Deep-learning models can predict events like weather patterns well, but having a low error in predictions doesn't mean the model understands the true physical relationships involved. The authors introduced a way called sensitivity fidelity that checks how models respond to small changes in input compared to trusted references or known physics. They found models with similar prediction scores could behave very differently in these sensitivity checks. This method helps to verify if AI models learn meaningful physical dependencies and can improve trust in their forecasts.
Open → 2610.15540v1

Weatherbridge improves forecasts by adapting multiple weather data sources

WeatherBridge: Forecast-Aware Multi-Source Input Adaptation for Pretrained Weather Forecasting Models

Abstract: Data-driven weather forecasting models are commonly trained on ERA5, yet operational deployment must use timely initial states from physics-based numerical weather prediction (NWP) systems whose distributions differ substantially from ERA5. Direct initialization with these states can degrade autoregressive forecasts, motivating input adaptation that keeps the pretrained forecasting model unchanged. Here, we introduce WeatherBridge, a unified, source-conditioned adapter that maps heterogeneous operational NWP initial states to ERA5 reference states without updating the pretrained forecasting model. WeatherBridge predicts the structured residual corrections with a Swin-UNet and injects source embeddings across multiple spatial scales, allowing the model to capture shared atmospheric structure and source-specific discrepancies. Its training objective combines latitude-weighted ERA5 reconstruction with One-step Rollout Alignment (ORA), which backpropagates one-step forecast error through a frozen FuXi model to optimize the adapted states for both initial state fidelity and downstream forecast compatibility. We evaluate WeatherBridge using daily GFS, HRES, and CMA initializations from 1 January to 22 November 2025 across six variables and lead times up to 240 hours. Relative to raw source initialization of FuXi, WeatherBridge reduces the mean RMSE of FuXi over the 6- 240-hour rollout for 17 of 18 source-variable combinations, with a median reduction of 10.2% and a maximum reduction of 46.3%. The adapter trained with frozen FuXi also improves FengWu forecasts on GFS inputs without retraining either the adapter or FengWu. These results demonstrate that forecast-aware input adaptation can improve forecasts from heterogeneous operational NWP inputs and transfer across pretrained ERA5-based forecasting models.

Sat 10 OctArtificial Intelligence
The gist
Weather forecasting models often use a specific type of data called ERA5 to make predictions. However, in real-world use, they get data from other weather systems that don’t match ERA5 perfectly, which can hurt forecast accuracy. The authors created WeatherBridge, a tool that adjusts different weather data to look like ERA5 data without changing the forecasting model itself. This adjustment helps the model make better forecasts across various weather variables and time frames. WeatherBridge also works well with different pretrained forecasting models, making it flexible and useful for improving weather predictions.
Open → 2610.14863v1

Just weather scoring improves fast and accurate radar precipitation forecasts

Just Weather Scoring: Efficient End-to-end Nowcasting with Distributional Diffusion

Abstract: Generative diffusion models are well-suited for probabilistic precipitation nowcasting, but existing approaches often rely on separately trained compression or deterministic forecasting components and remain costly at inference due to iterative denoising. We introduce Just Weather Scoring (JWS), a single-stage, end-to-end diffusion model which addresses both issues by forecasting directly in radar space and enabling few-step generation. Radar-space modeling greatly simplifies training and inference and eliminates uncertainty arising from lossy compression. JWS combines Masked Asynchronous Diffusion, a timestep-sampling scheme that preserves clean context while adapting diffusion training to high-dimensional spatio-temporal data, with a simple scoring-rule objective that aligns training with probabilistic forecasting and unlocks few-step generation. On the SEVIR and MeteoNet benchmarks, JWS achieves state-of-the-art probabilistic forecasting performance at reduced training and inference cost. Even our smallest model remains competitive using substantially fewer parameters and more than 17x faster inference.

Thu 8 OctMachine LearningComputer Vision and Pattern Recognition
The gist
Predicting short-term rainfall is important but challenging and slow with current AI methods. The authors introduce Just Weather Scoring, a new model that learns to predict future radar images of rain directly, without extra compression steps, making it faster and simpler. Their method also trains to match probability forecasts better and uses fewer steps to make predictions. Tests show it achieves top accuracy while running much faster and using fewer resources.
Open → 2610.12189v1

Low-rank tensor model improves satellite rainfall estimates across the US

Low-rank tensor structure of precipitation and its application to satellite-reference merging

Abstract: The intermittent and variable nature of precipitation makes its accurate estimation over extended domains difficult, yet its spatiotemporal structure suggests that a low-rank representation may be possible. This work represents daily precipitation over the contiguous United States (CONUS) as spatiotemporal tensors and applies CANDECOMP/PARAFAC factorization, showing that preserving the native spatial and temporal modes yields more accurate reconstruction than factorizing independent daily fields or unfolded space--time matrices. Building on this finding, this work presents TMerge, a tensor-based framework that integrates satellite precipitation with sparse reference observations through shared low-rank spatial and temporal factors. TMerge was applied to correct the IMERG Final Run product with climate prediction center reference observations over CONUS. During 2019-2022, TMerge increased correlation from 0.53 to 0.85 and reduced root-mean-square error and mean absolute error by 48.2% and 29.3%, respectively. TMerge consistently outperformed linear bias correction, quantile mapping, and neural networks across seasons, precipitation-intensity regimes, and regions. Improvements were spatially coherent and largest in coastal regions where IMERG errors were greatest. These results demonstrate that low-rank tensor structure parsimoniously approximates the dominant spatiotemporal variability of precipitation and provides a practical mechanism for improving satellite estimates under limited reference observations over extended domains.

Wed 7 OctMachine Learning
The gist
Rainfall varies a lot in space and time, making it hard to measure accurately everywhere. The authors showed that using a special math tool called a low-rank tensor can capture rainfall patterns more efficiently by keeping time and space data separate. They built a method called TMerge that combines satellite rainfall data with ground measurements and drastically improves accuracy across the US. This approach worked better than previous correction methods, especially in coastal areas where satellite errors were largest.
Open → 2610.11000v1

SoftSEEPS enhances precipitation forecasts with differentiable scoring

SoftSEEPS improves ML-based precipitation forecasting

Abstract: In this paper we have developed a differentiable approximation of the well-known SEEPS score, which we name SoftSEEPS. This allows the training of a Machine Learning model to forecast precipitation directly. We test SoftSEEPS on the IMERG dataset (0.1 degree resolution) by training a decoder for precipitation on the latent space of a pre-trained low-resolution forecasting model. Combining SoftSEEPS and RMSE in a joint objective is possible with marginal trade-offs in either metric.

Wed 7 OctMachine Learning
The gist
Forecasting rainfall is tricky, and standard ways to measure the accuracy of forecasts are often not easy for computers to use when learning. The authors created SoftSEEPS, a new method that smooths out a popular evaluation score so a machine learning model can train directly with it. They tested this on weather data to improve rain predictions from a simpler forecasting model. Their approach works well when combined with traditional error measures without losing much accuracy.
Open → 2610.09752v1

Xaurora turns weather models into fast probabilistic forecasts

Xaurora: Generative Weather Forecasting with Denoising Stochastic Interpolants from a Foundation Model Prior

Abstract: Deep learning has revolutionised weather forecasting in recent years, especially through atmospheric foundation models, which offer competitive skill for a fraction of the computational costs of classic physics-based models. However, most existing foundation models are deterministic, limiting the generation of large ensembles for accurate uncertainty quantification, extreme weather risk assessment, and long-range weather forecasting. Furthermore, these models incur a large, often prohibitive, computational overhead to train from scratch. To address these shortcomings, we turn a pretrained deterministic prior model, namely the Aurora foundation model, into a generative ensemble-prediction model. To that end, we introduce a novel generative method, Denoising Stochastic Interpolants, combined with a replay buffer for Stochastic Differential Equation (SDE) rollout, enabling probabilistic training of SDE trajectories. Our stochastic foundation model, Xaurora, is finetuned from the small Aurora version, yet it approaches the state-of-the-art on global ensemble metrics and is competitive with the large version of Aurora. Our method is parameter and sample efficient, and generates skilful 15-day forecasts in 13 minutes. Our results demonstrate that deterministic foundation models can be efficiently extended into even stronger stochastic models.

Mon 5 OctMachine Learning
The gist
Weather forecasts often need to show possible different outcomes, not just one prediction, to help with things like extreme weather warnings. Most advanced AI weather models only give one answer, making it hard to understand uncertainty. The authors turned an existing AI weather model into one that can create many possible forecasts quickly and accurately. Their approach, called Xaurora, can produce detailed uncertainty estimates while running much faster than traditional methods.
Open → 2610.06509v1

Latent space filtering improves forecasting with sparse noisy data

LD-EnFF: Latent-Dynamics Ensemble Flow Filtering for Data Assimilation with Sparse Observations

Abstract: Data assimilation combines model forecasts with noisy, incomplete observations to estimate the evolving state of a dynamical system. Existing methods face two compounding challenges: high-dimensional nonlinear dynamics make repeated forward simulation computationally expensive, while sparse observations provide limited direct information about the full state. To address these challenges, we propose the Latent-Dynamics Ensemble Flow Filter (LD-EnFF), a sequential Bayesian filtering framework that performs both forecast propagation and filtering updates in a compact latent space. LD-EnFF combines a latent dynamics surrogate for ensemble propagation with a variational autoencoder (VAE)-based observation model that evaluates a state-dependent observation likelihood in latent space. At each assimilation step, an ensemble filtering update based on flow matching uses the forecast ensemble and this likelihood to generate posterior samples, jointly updating latent states and uncertain parameters. This design avoids repeated full-state simulation during forecasting and full-field reconstruction during likelihood evaluation. LD-EnFF substantially outperforms a broad range of data assimilation algorithms on benchmarks spanning Kolmogorov flow, tsunami propagation, and atmospheric modeling, all featuring complex dynamics and sparse, noisy observations.

Fri 2 OctMachine Learning
The gist
Forecasting how complex systems change is hard because simulations take a lot of time and observations are limited and noisy. The authors developed a new method called LD-EnFF that works in a simpler, compressed version of the system, called a latent space, to speed up forecasting and make better guesses. LD-EnFF uses a special neural network to understand how observations relate to this latent space and updates its predictions more efficiently. Tests show it outperforms many existing methods on problems like fluid flow, tsunamis, and weather with sparse and noisy data.
Open → 2610.04034v1

AI models emulate sudden shifts in polar vortex weather patterns

AI Emulation of Stochastic Sudden Stratospheric Warming with Interpretable Latent Structure

Abstract: Rare weather regime transitions pose a challenge for data-driven modeling due to class imbalance. In this study, we develop a probabilistic deep learning emulator for a prototypical system with regime transitions, the stochastic Holton--Mass model of stratospheric variability, and analyze the structure of its learned latent space. The Holton--Mass model exhibits two metastable regimes, a strong and a weak polar vortex, maintained by nonlinear wave--mean flow interactions, with weak stochastic forcing intermittently triggering rare transitions between these regimes that qualitatively represent SSW events. We employ a ResNet-inspired Conditional Variational Autoencoder with six-layer encoder and decoder layers and explicit current-state conditioning to model the distribution of the system's state at the next time step (one day). The emulator accurately reproduces short-term dynamics, steady-state probability distributions, regime persistence statistics, rare transition rates, the transition committor function, and the transition expected lead time of the physical model. Beyond emulation fidelity, we interrogate the learned latent representation to understand how the model internalizes the underlying metastable structure of the dynamics. Principal Component Analysis of the 32-dimensional latent space reveals a clear and unsupervised separation into four physically interpretable clusters corresponding to strong versus weak vortex regimes and stable versus transition-prone configurations. Such emergent regime separation in latent space is hard to identify for deep generative models applied to high-dimensional stochastic systems. Our results show that carefully designed probabilistic emulators can uncover physically meaningful manifolds governing extreme-event dynamics, potentially aiding the development of improved operational advanced warning systems.

Thu 1 OctMachine Learning
The gist
Sudden Stratospheric Warming (SSW) events are rare shifts in the Earth's polar vortex that impact weather but are hard to predict due to their rarity. The authors created an AI model that can imitate these sudden weather changes in a simplified system, learning to capture both common and rare weather states. Their AI not only predicts short-term weather patterns but also organizes its internal understanding into clear groups that reflect underlying physical weather conditions. This approach could help improve forecasting systems by better spotting when rare events are likely to happen.
Open → 2610.02069v1

Climate downscaling methods vary in preserving spatial accuracy and detail

Beyond Pointwise Error: A Multi-Metric Evaluation of Spatial Climate Downscaling

Abstract: Climate downscaling aims to reconstruct fine scale spatial fields from coarse resolution inputs. Evaluating the quality of these reconstructions is challenging: low pointwise error can come at the cost of fine scale variability, while realistic spatial variability can be achieved with inaccurate local structures. The evaluation metric can therefore change which method appears to perform best. This work presents a multi metric benchmark comparing five spatial downscaling methods on ERA5 temperature, wind, and precipitation fields. Five criteria assess complementary properties: pointwise error, structural similarity, distribution error, spectral error, and gradient error. The results reveal a systematic trade off between spatial fidelity and fine scale variability. Some methods perform best on pointwise and spatially aligned metrics, but lose high frequency content, while others preserve substantially more spectral variability at the cost of less accurately positioned local structures. Consequently, method rankings change across metrics and variables. These results show that there is no single best downscaling method. Multi metric evaluation is therefore essential for assessing which properties of a climate field are preserved.

Thu 1 OctMachine LearningArtificial Intelligence
The gist
Climate downscaling tries to make detailed local weather information from rough, large-area data. The authors show that different ways of doing this have trade-offs: some get local details right but miss subtle patterns, while others keep more natural variation but aren’t perfectly accurate at specific spots. They tested five methods on temperature, wind, and rain data and found no single best solution. They suggest using multiple evaluation methods to judge which downscaling approach works best for different needs.
Open → 2610.01579v1

PrecipJEPA improves future rain forecast accuracy using motion-based modeling

PrecipJEPA: JEPA-Regularized Future-State Prediction with Motion-Source Rendering for Precipitation Nowcasting

Abstract: Long-term precipitation nowcasting requires modeling radar-echo evolution while preserving localized high-intensity structures. Recent radar-specific studies motivate location-aware prediction and separating echo displacement from intensity change. However existing encoders learn historical representations mainly from final forecast errors. We propose PrecipJEPA, which couples a structured forecasting path with an auxiliary path that enriches its encoder from observed radar history. In the forecasting path, an online encoder first converts the observations into spatiotemporal tokens. The Task-Driven Future-State Predictor (TFP) combines these tokens with a recent-dynamics summary and spatiotemporal queries to construct future radar states. The Parallel Motion-Source Renderer (PMSR) decodes these states into motion and source-sink fields that transform the latest observation into future frames. During joint training, the History-Masked JEPA (H-JEPA) operates on the auxiliary path to predict masked historical features from visible context, directly supervising the same online encoder from the observed sequence. Experiments on SEVIR and MeteoNet show that PrecipJEPA improves highest-threshold CSI by 118.6% and 35.1%, respectively, over the strongest baselines, while maintaining the highest mean CSI throughout the 3-hour forecast.

Wed 30 SeptMultimediaMachine Learning
The gist
Forecasting where and how much it will rain in the next few hours is hard because rain patterns move and change intensity in complex ways. The authors propose PrecipJEPA, a new method that learns from radar images over time to better predict future rain by understanding both motion and intensity changes. It uses two paths: one that predicts future radar images and another that learns from past data by filling in missing parts, helping the model get better at understanding rain patterns. Tests show it significantly improves accuracy for predicting heavy rainfall over three hours compared to previous methods.
Open → 2609.38926v1

Transversal pooling networks improve stability under input transformations

Transversal Pooling Neural Networks

Abstract: Many learning tasks require stability to small transformations while retaining sensitivity to larger ones. We introduce \emph{transversal pooling neural networks}, which generalize spatial max pooling to affine group actions. We establish equivariance to a chosen subgroup and derive explicit stability bounds for individual pooled wavelet coefficients under affine perturbations of the input. Experimentally, we demonstrate the utility of our networks in low-data environments and for predicting tropical cyclone intensification.

Mon 28 SeptMachine Learning
The gist
Many computer learning tasks need to recognize objects even when they are slightly moved or transformed, but still detect bigger changes. The authors introduce transversal pooling neural networks, which extend a common image-processing step called max pooling to more complex transformations like rotations and stretches. They prove these networks behave predictably when inputs are changed slightly and test them on small datasets and predicting hurricane strength changes. This new approach helps computer systems better manage images or data that are shifted or transformed.
Open → 2609.36237v1

Adaptive score filtering improves high-dimensional nonlinear data assimilation

AECSF: Adaptive Ensemble Conditional Score Filtering for High-Dimensional Nonlinear Data Assimilation

Abstract: Bayesian state estimation for high-dimensional nonlinear dynamical systems entails a fundamental tension between statistical fidelity and computational tractability, as particle weights can collapse, while Gaussian ensemble updates can miss non-Gaussian posterior structure. Score-based diffusion filters offer a sampling-based alternative, but existing training-free score filters often rely on heuristic likelihood corrections, which can compromise posterior accuracy by neglecting uncertainty about the system state associated with each noisy reverse particle. To address these issues, we propose AECSF, a training-free adaptive ensemble conditional score filter. AECSF constructs an analytically tractable score estimator from the conditional Tweedie identity, which recasts noisy posterior score estimation as estimating the conditional mean of the system state given a noisy reverse particle and the observation. To estimate these conditional means efficiently, AECSF employs a shared adaptive weighted proposal ensemble, while particle-specific conditional weights yield an estimate for each noisy reverse particle without separate proposal sampling. The proposal ensemble is updated using reverse-particle information within the same reverse-diffusion run to improve conditional-mean estimation. Theoretically, we characterize when a fixed weighted proposal measure yields the exact noisy posterior score. Under stated assumptions, we establish a bound relating conditional-mean estimation errors to reverse-sampling endpoint error. Numerical experiments demonstrate that AECSF improves the accuracy of posterior sampling and nonlinear filtering in high-dimensional problems with limited forecast ensembles.

Sat 26 SeptMachine Learning
The gist
Estimating the true state of complex systems with many variables is hard because current methods either oversimplify or become unreliable. The authors propose AECSF, a new way to better estimate these states without needing training data. AECSF uses a clever statistical identity to improve accuracy by adaptively weighting multiple guesses about the system’s state. Their approach helps better capture uncertainty and improves predictions in challenging, high-dimensional problems with limited data.
Open → 2609.32411v1

Extreme event prediction improves with local instability sensing

Mechanism-Aware Ensemble Conditioning for Data-Limited Emulation of Extreme Events

Abstract: Extreme events in chaotic systems are difficult to learn from short trajectories because they are controlled by transient finite-time instability rather than by frequently observed bulk dynamics. We propose a mechanism-aware conditioning plug-in framework that turns a nudged coarse ensemble into a non-intrusive sensor of local instability geometry. In the small-noise regime, the ensemble covariance aggregates the same finite-time deformation kernels that govern local instability, providing a Jacobian-free proxy for the local amplification structure around a synchronized coarse trajectory. A small FiLM module injects statistics of this ensemble geometry into an otherwise unchanged backbone while leaving the coarse simulator unchanged. We demonstrate this interface in two distinct pipelines: a Transformer-style residual-attention corrector for a controlled low-dimensional chaotic system and a probabilistic recurrent STORN corrector for topographic two-layer quasi-geostrophic (QG) flow. In the low-dimensional benchmark, ensemble covariance directions co-activate with OTD modes and FiLM conditioning improves 99th-percentile exceedance-frequency errors over an identical no-context Transformer baseline. In QG, a fixed ensemble-conditioned FiLM-STORN model trained on only \(50\) time units substantially improves long-horizon rare-event statistics in the data-limited regime, including density-tail errors, exceedance frequencies, and spatial exceedance-area distributions relative to an unconditioned STORN trained on the same data; on averaged high-threshold exceedance diagnostics, it also outperforms the baseline STORN trained with $20$ times more high-resolution data. These results show that local instability geometry is not merely interpretable post hoc, but an actionable conditioning signal for data-efficient rare-event emulation.

Fri 25 SeptMachine Learning
The gist
Predicting rare, extreme events in chaotic systems is tough because these events depend on short-term instabilities that are not common in usual data. The authors developed a method that uses a group of slightly different simulations to detect these local instabilities without complex calculations. By feeding this information into machine learning models, the method improves the prediction of rare events, even with limited training data. They tested their approach on both simple chaotic systems and more complex fluid models, showing better accuracy compared to traditional methods.
Open → 2609.30746v1

WeatherDiagFlow improves radar forecasts with evidence and audits

WeatherDiagFlow: Evidence-Grounded Radar Nowcasting with Diagnostic Flow Refinement

Abstract: Radar nowcasting is essential for short-term warning and emergency response, yet conventional systems mainly return future radar fields and provide limited support for operational communication and post-event verification. We formulate radar nowcasting as an evidence-grounded forecast--bulletin--audit task, in which a numerical forecaster produces both future radar fields and structured diagnostic evidence. Forecast-time bulletins use only model-available evidence, whereas post-event audits incorporate future radar truth only after the forecast horizon is observed. Based on this task formulation, WeatherDiagFlow predicts motion, growth and decay, heavy-echo risk, and uncertainty to condition rolling flow refinement, while frozen-scaffold residual calibration improves long-lead strong-echo preservation. A multi-agent layer converts the structured evidence into operational bulletins and independently generates verification audits without feeding textual outputs back into the forecaster. Experiments on FJRADAR demonstrate competitive overall performance and improved strong-echo event skill. WeatherDiagFlow therefore connects numerical prediction, evidence-grounded reporting, and auditable verification under a leakage-controlled protocol.

Thu 24 SeptMachine LearningMultimedia
The gist
Short-term weather forecasts from radar help with emergency warnings but usually only show future radar images without much explanation. The authors present WeatherDiagFlow, which not only predicts future radar but also provides clear evidence and reports about what the forecast means and how confident it is. This system also checks its own forecasts against real radar data afterward to verify accuracy. By combining predictions, evidence, and audits, WeatherDiagFlow aims to make radar forecasts easier to understand and trust.
Open → 2609.29772v1

Wavelet-diffusion models improve precipitation detail across US regions

Evaluating Cross-region Generalization for Wavelet-Diffusion Precipitation Downscaling

Abstract: 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.

Wed 23 SeptMachine Learning
The gist
Predicting detailed rainfall patterns from coarse weather data is hard, especially in places where models weren’t trained. This paper looks at a special AI method called wavelet diffusion models (WDMs) that can create detailed precipitation maps from smooth inputs. The authors test how well these models work in different US climate regions, including places they haven’t seen during training. They find that WDMs trained on one region can still do a good job elsewhere, and models trained on all regions work even better in spotting intense rain areas. This could help make weather predictions more detailed and accurate in many places, even with limited local data.
Open → 2609.28749v1

Transformer model improves planetary boundary layer height estimates from satellite data

PBLH Estimation from Satellite Radiances via a Dual-Encoder Transformer

Abstract: Estimating the Planetary Boundary Layer Height (PBLH) from satellite observations is a challenging regression problem due to the indirect relationship between top-of-atmosphere radiances and near-surface atmospheric structure. Progress has been limited both by the lack of architectures capable of handling the multimodal, spatially incomplete nature of satellite overpasses, and by the scarcity of suitable datasets. In this paper, we build upon the large-scale dataset pairing MetOp radiances with ERA5 PBLH labels that we introduced in our previous work, making three contributions. First, we establish a benchmark across eight approaches spanning pixel-wise regression, swath-wise sequence models, and convolutional and Transformer models operating on the full orbital passage. Second, we quantify what the resulting model actually relies on, using grouped Shapley decomposition over the input blocks. Third, we present the best-performing architecture found: a dual-encoder Transformer whose masked-input handling lets it operate in all weather conditions. The proposed model achieves MAE = 155.8 m on the held-out global test set, outperforming all baselines on every evaluation subset. On 30 out-of-distribution granules acquired on two days overlapping the TEAMx observational campaign, it achieves MAE = 165.3 m, outperforming a pixel-wise baseline trained on the same data (MAE = 197 m).

Wed 23 SeptComputer Vision and Pattern RecognitionMachine Learning
The gist
Estimating how high the lowest part of the atmosphere reaches (called the boundary layer) using satellite data is tricky because satellites see indirect signals. The authors worked with a big dataset matching satellite readings to known boundary heights and tested multiple methods to find the best one. They designed a special artificial intelligence model that looks at the data in two ways and can handle missing or unclear satellite readings caused by clouds or weather. Their model made more accurate height predictions worldwide than previous methods, including on new data taken during a recent weather campaign.
Open → 2609.28286v1

Sea ice forecast errors corrected using adaptive sparse observations

Taking a Second Look: Correcting Sea Ice Forecasts with Sparse Observations

Abstract: Sea ice forecasts are issued several days ahead, allowing errors to accumulate while new, often sparse sea ice concentration (SIC) observations become available. We find that fixed-propagation errors concentrate near structured, high-gradient ice edges, whereas homogeneous interiors require limited propagation, suggesting that propagation distance should be state dependent. We therefore introduce ECHO (Evidence-guided Correction with Heterogeneous prOpagation), where ECHO-Scale adapts propagation distance while preserving correction geometry, and ECHO-Delta learns a bounded residual around fixed propagation. Across all 96 standard evaluation settings spanning diverse priors, observation times, sparsity levels, geometries, and noise conditions, both outperform fixed propagation. ECHO-Delta achieves the best average accuracy, while ECHO-Scale is more robust to geometry shifts. Code is available at https://github.com/yingtian22/TAKING-A-SECOND-LOOK.

Mon 21 SeptMachine Learning
The gist
Forecasts of sea ice several days ahead can have errors, especially near ice edges where conditions change quickly. The authors found that these errors spread differently across ice areas, depending on whether the ice is stable or near borders. They created ECHO, a method that adjusts how far correction information travels based on ice conditions, improving forecast accuracy. Their approach works better than traditional methods across many different scenarios, handling sparse and noisy data more effectively.
Open → 2609.24591v1

Recursive quantum lstm improves temperature prediction accuracy and stability

Recursive Quantum Long Short-Term Memory for Stable Short-Horizon Temperature Forecasting

Abstract: Quantum long short-term memory (QLSTM) models extend recurrent sequence learning with variational quantum circuits, but their optimization behavior can vary substantially across random initializations and temporal contexts. This paper evaluates a recursive QLSTM architecture against a standard QLSTM for one-step-ahead prediction of daily minimum and maximum temperature. Using daily weather observations from Toronto and identical training settings, we compare convergence, predictive accuracy, and generalization across input windows of 8, 16, and 32 days over 20 random seeds. The recursive model consistently reaches a near-optimal test loss earlier, reduces mean absolute error and root mean squared error, and exhibits a smaller generalization gap. These results indicate that recursive quantum feature transformations can improve stability and out-of-sample performance for compact hybrid quantum--classical temporal models.

Thu 17 SeptMachine Learning
The gist
Predicting daily temperatures can be tricky, especially when using advanced computer models that mix quantum computing ideas with traditional methods. The authors studied two versions of these models to see which predicts daily highs and lows better for Toronto's weather. They found that a recursive version, which reuses information in a special way, was more stable and made more accurate forecasts than the regular version. This suggests new ways to improve weather predictions using hybrid quantum-classical techniques.
Open → 2609.20594v1

Double U-shaped neural operator improves wave modeling accuracy and efficiency

DU-NO: A Parameter-Efficient Double U-Shaped Neural Operator for Phase-Resolving Wave Modeling

Abstract: Phase-resolving wave models such as FUNWAVE-TVD are the accuracy standard for nearshore dynamics, resolving the shoaling, refraction, and breaking of individual waves, but their cost rules them out for the ensembles, uncertainty quantification, and real-time warning that operational forecasting demands. Neural operators promise solver-level accuracy at a fraction of that cost, yet on wave-dominated fields the accurate ones are large: hybrid spectral-convolutional operators such as U-FNO (the strongest baseline in our study after DU-NO) buy their fidelity with tens of millions of parameters. We introduce DU-NO (Double U-shaped Neural Operator), a multiscale U-shaped spectral operator that attaches lightweight convolutional U-Net branches only at its two shallowest encoder and decoder levels. The placement follows a sampling argument: high-wavenumber content exists only on fine grids, so the local, full-band pathways go where that content lives, while the coarse, band-limited levels stay purely spectral. A depth-decaying mode schedule holds the model to 3.64M parameters, an order of magnitude below U-FNO. On our publicly released FUNWAVE-TVD benchmark, DU-NO attains the best autoregressive rollout error of six identically trained architectures, improving on U-FNO by 14.9% with 10.8x fewer parameters, and a frequency-band analysis shows the gain holds across all bands, including the high-wavenumber band where truncated-spectral operators collapse. Parameter-matched controls confirm the gain is architectural: rescaled to the same 3.6M budget, the best baseline still trails DU-NO by 28.6%. The advantage carries beyond nearshore waves: DU-NO matches the strongest baselines on 2D Navier-Stokes and wins clearly on PDEBench shallow-water rollouts. Code, trained models, and evaluation artifacts are available at https://anonymous.4open.science/r/duno-code-5A7B/.

Thu 10 SeptArtificial Intelligence
The gist
Modeling how waves behave near shores can be very accurate but usually requires slow and expensive computer simulations. The authors developed a new neural network design called DU-NO that is much smaller and faster but still very accurate at predicting waves. DU-NO uses a clever structure that focuses on fine details where needed, reducing the number of parameters by about ten times compared to strong previous methods. It also performs well on other fluid dynamics problems, suggesting broad usefulness.
Open → 2609.12115v1

Post-processing methods improve temperature and wind forecasts at new locations

Statistical versus machine learning-based spatial interpolation of post-processed ensemble weather forecasts

Abstract: Statistical post-processing improves ensemble weather forecasts, but generating calibrated predictions at locations without observations remains challenging. This study compares statistical and machine-learning-based methods for post-processing ECMWF 2-m temperature and 10-m wind speed forecasts at observed and unobserved stations in Germany. We consider EMOS-based approaches, distributional regression networks, Transformers, and graph neural networks under both limited and extended predictor settings. For temperature, we also investigate linear forecast combinations and propose an altitude-aware linear pool (ALP). The results show that post-processing improves upon the raw ensemble in most settings, but no single method performs best across all variables, station groups, and evaluation metrics. The proposed ALP provides a small but significant improvement over the standard linear pool at unobserved locations.

Mon 7 SeptMachine Learning
The gist
Weather forecasts often use groups of predictions that need adjusting to be more accurate. The authors compared traditional statistical methods to newer machine learning techniques to improve forecasts of temperature and wind speed in Germany, especially at places without weather stations. They found that while no single approach worked best for every situation, adjusting forecasts this way generally made them better than raw predictions. The authors also introduced a new way to combine forecasts that takes altitude into account, which slightly improved temperature predictions at unobserved locations.
Open → 2609.07512v1