WeatherNext 3: Increasing resolution and performance of global weather models with raw observations

State-of-the-art AI weather models have shown impressive medium-range forecast skill and computational efficiency, but suffer two key shortcomings: their forecasts have lower spatial and temporal resolution than the best physics-based models and they are exclusively initialized with and trained on analysis data. As a result, they cannot directly make use of observations, and any biases in the analysis are inherited by the forecast. WeatherNext 3 addresses these shortcomings and establishes a new state-of-the-art for probabilistic medium-range forecasting skill. First, WeatherNext 3 generates new forecasts every hour (rather than every 6 hours like traditional global models) by ingesting low-latency geostationary satellite data. Second, WeatherNext 3's temporal and spatial resolution are on par with physics-based global models, with hourly time steps and 0.1 degree resolution for single-level variables, including solar radiation and cloud cover. Third, WeatherNext 3 moves beyond traditional analysis variables by learning to predict satellite-derived precipitation estimates, as well as tropical cyclone and station observations. Modelling sparse station data allows WeatherNext 3 to make 2m temperature and dewpoint predictions at any location and time, conditioned on local geographical features, with substantially lower error than competing global models, even when evaluated against unseen stations. Together, WeatherNext 3's capabilities move operational AI-based weather forecasting beyond emulating the traditionally distinct stages of data assimilation, forecasting and post-processing, which helps to further push the frontier of performance and granularity for global weather prediction.

Improving precipitation forecasts in an AI weather model using observational data

Artificial intelligence weather prediction (AIWP) systems now surpass state-of-the-art physical models for medium-range weather forecasting. Current global AIWP models are trained almost exclusively using one reanalysis dataset, ERA5, but it has known biases, particularly for precipitation. Here we fine-tune a graph-transformer architecture with IMERG precipitation data at 0.25° resolution. The resulting model improves medium-range continuous ranked probability scores by up to 19%, while also demonstrating superior skill for tropical storms and drizzle events. Our model exceeds the Brier skill score of state-of-the-art operational models on extreme rainfall prediction by 57% globally; however, a physics-based operational model remains more reliable for the heaviest precipitation events. Our results demonstrate that incorporating observations-based precipitation data directly into training can substantially improve precipitation forecasts.

Uncertainty-Aware End-to-End AI Weather Forecasting: Disentangling Observation and Model Contributions

End-to-end weather forecasting systems produce skillful global gridded and station forecasts directly from raw Earth observations, replacing the numerical weather prediction pipeline, including data assimilation, at a fraction of its cost. These systems are deterministic and issue no uncertainty. Here we render the Aardvark Weather model probabilistic by attaching one stochastic mechanism to each component: learned, input-dependent noise at the observation encoder, capturing aleatoric uncertainty inherited from the observing system, and Monte Carlo dropout in the processor, capturing epistemic uncertainty in the learned dynamics. The resulting nested ensemble attributes forecast spread to the two sources through a law-of-total-variance decomposition, cross-checked by withholding observation streams. Probabilistic finetuning significantly improves the mean forecast, by 4.2% on average across variables and lead times. The ensemble is calibrated against ERA5 through the medium range (spread-skill ratio 0.98), keeps station RMSE within 2.4% of the deterministic model while beating it in CRPS at every lead time, and trails the operational ECMWF ensemble. The encoder branch behaves as observation-driven uncertainty. Component-attributed uncertainty makes end-to-end forecasts more transparent, a step toward observation-driven digital twins of the atmosphere.

Diffusion Distillation for Efficient Weather Ensembles

Diffusion models generate skillful weather ensembles but require costly iterative sampling. We introduce a supervised energy-distance distillation method that compresses a multi-step diffusion teacher into a single-step student by aligning student forecasts with teacher samples and ground-truth observations. Experiments on global forecasting and typhoon-track prediction show that our student outperforms existing distillation methods and preserves skill for extreme events. It matches or surpasses the teacher across key metrics using only one neural function evaluation per autoregressive step.

Climate Physics Dynamic Matching

Deep generative models such as flow matching and diffusion models have shown potential for learning complex dynamical systems, but typically act as black boxes that neglect underlying physical structure, while physics-based models governed by partial differential equations are often incomplete due to missing source terms, or uncertain parametrisations. We present Climate Physics Dynamic Matching (ClimPhyDM), a variational simulation-free dynamics informed framework for weather forecasting that combines an advection-type physics prior with data-driven components in a variational framework. % to capture the stochasticity and multi-modality of unresolved atmospheric dynamics. On the ERA5 benchmark at hourly (42-hour) and monthly (5-month) resolutions, ClimPhyDM outperforms ClimODE, and GB-DM, keeping the lower error at extended horizon, indicating improved temporal stability and resistance to error accumulation, while its simulation-free paradigm also enables training on a single modest 12 GB consumer GPU.

Bridging short- and medium-range weather forecasting with machine learning

The National Oceanic and Atmospheric Administration (NOAA) employs independent prediction systems for distinct forecast products. While some separation is practical, we argue that combining short- and medium-range weather into a single prediction system would provide the public with a useful distillation of global weather and its impacts. To this end, we present Nested-EAGLE (Experimental Artificial intelligence Global and Limited-area Ensemble): a 0.25° global weather model with a 6 km refinement over the Contiguous United States (CONUS). The model achieves significantly lower mean-squared error in near-surface and low-level quantities over CONUS compared to NOAA's Global Forecast System and High-Resolution Rapid Refresh (HRRR), while remaining competitive throughout the rest of the global atmosphere. We show that the skill gains for near-surface fields stem from incorporating high-resolution regional analysis data into training through the nesting process. Forecasts of precipitation amounts are less skillful than those from HRRR, owing to deterministic training. However, we show that Nested-EAGLE provides the most accurate forecasts of storm locations at longer leads, despite blurred extrema. Our results motivate future work to extend the skill gains beyond CONUS and improve precipitation representation.

Learning Continuous Regional Temperature Fields with Lead-Time and Resolution Queries

Accurate regional near-surface temperature forecasting is fundamental to short-range weather services and downstream risk assessment. Existing deep learning-based regional forecasters commonly produce a fixed set of future frames on a prescribed grid, limiting their use when forecast products must be evaluated at query-dependent lead times or display resolutions. To overcome these fixed-output constraints, we formulate regional T2M forecasting as query-conditioned continuous spatiotemporal temperature field evaluation and propose the Continuous Spatiotemporal Temperature Forecaster (CSTF), a neural field that turns forecast lead time and output resolution into explicit queries when evaluating 2-m temperature (T2M). Specifically, CSTF first encodes multivariable ERA5 histories into latent meteorological states and then decodes T2M as a coordinate-based field. Accordingly, spatial location, forecast lead time, and output resolution are introduced as queries, enabling standard hourly forecasts, intermediate lead-time diagnostics, and resolution-controllable outputs within a unified field-evaluation framework. Furthermore, to maintain coherence across flexible field queries, we design spatial-gradient, temporal-difference, and scale-consistency objectives that regularize regional thermal structures, lead-wise evolution, and cross-resolution agreement. Experiments on the Southeast China 0-6 h ERA5-Land benchmark demonstrate that CSTF achieves the best aggregate deterministic skill, including a 17.0 percent reduction in Bias, with global-scope diagnostics further illustrating flexible lead-time and resolution-controllable inference.

Frequency-aware forecasting for short-term typhoon gust prediction

Accurate gust forecasting under typhoon conditions remains challenging due to the highly non-stationary and multi-scale characteristics of extreme wind fluctuations. Existing deep learning models often struggle to simultaneously capture long-term trends and rapid local variations, resulting in degraded performance during extreme events. We propose WDANet, a frequency-aware forecasting framework that integrates stationary wavelet decomposition, a Feature-wise Linear Modulation (FiLM) strategy, and a dual-branch encoder-decoder architecture, enabling separate modeling of trend and fluctuation components. Taking the offshore regions of the Western Pacific in China as an example, we conduct fine-grid wind gust prediction research. The results demonstrate that WDANet shows advantages for short lead times under the experimental setting across a 24-h forecasting horizon and achieves higher prediction accuracy than ECMWF-HRES within the first 6 h. During extreme wind events, WDANet more accurately captures gust peaks and attains the best RMSE and MAE performance. These results highlight its potential for offshore wind power operation, disaster warning, and risk mitigation.

AFDBench: A Reasoning-First AI Scientist for NationalWeather Service Forecast Discussions

Large language models (LLMs) hallucinate numerical values when generating high-stakes meteorological text, posing risks for weather communication. We present AFDBench, an AI meteorologist that generates professional Area Forecast Discussions (AFDs) by reasoning through structured AI weather forecast data from Google's WeatherNext 2. We introduce AFDBench, the first benchmark for evaluating generative meteorological reasoning, comprising 7,732 expert written discussions from 13 National Weather Service (NWS) offices paired with real AI weather forecast inputs, and three complementary metrics: Met-Align (numerical accuracy), Style-Align (professional dialect adherence), and Input-Grounding (fidelity to source weather data). Zero-shot evaluations reveal that open-source LLMs achieve low Style-Align (~0.33) and moderate Input-Grounding (~0.88), failing to write in the professional NWS register or faithfully use their input data. We apply Group Relative Policy Optimization (GRPO) with domain-specific rewards targeting temperature accuracy, synoptic correctness, and format compliance. On 1,033 held-out samples from two unseen NWS offices, GRPO nearly doubles Style-Align from 0.318 to 0.619 and improves Input-Grounding from 0.881 to 0.940, demonstrating that reinforcement learning teaches a 7B-parameter model to write like a professional meteorologist and faithfully interpret AI weather data.

VeinCast: Physics-Guided Dynamic Field Graphs with Graph-Conditioned Fusion for Global Medium-Range Weather Forecasting

Global medium-range weather forecasting requires modeling structured yet state-dependent interactions among heterogeneous atmospheric fields. Existing data-driven models largely learn these interactions implicitly, whereas equation-level physical constraints may inherit approximation and model-form biases. We present VeinCast, a physics-guided dynamic field graph and graph-conditioned fusion framework that jointly forecasts 69 surface and upper-air fields. Within each local window, its Physics-Guided Dynamic Field Graph combines predefined atmospheric relations with state-dependent Top-K residual edges and adapts Earth-window attention using the resulting graph context. Graph-Conditioned Latent Fusion further employs graph context and source-node centrality to guide field-to-latent aggregation, while bounded feedback preserves field-specific information. On the $1.5^\circ$ ERA5 benchmark, VeinCast demonstrates competitive forecasting performance across all 69 meteorological fields at lead times of up to 14 days, compared with representative global weather forecasting models including FuXi, Pangu-Weather, GraphCast, FengWu, and ARROW. Ablations confirm that the two modules provide complementary gains, demonstrating the effectiveness of relational-level physical guidance for data-driven weather forecasting.

FreCast: Refining Radar Echo Intensity via Phase-Preserving Amplitude Residual Diffusion for Precipitation Nowcasting

Precipitation nowcasting predicts the spatiotemporal evolution of future radar echoes from historical radar echo sequences, thereby estimating the occurrence, development, and movement of precipitation over the near term. In recent years, deep learning has become an important approach to precipitation nowcasting. Although state-of-the-art models can generally capture the overall spatial distribution of future precipitation, their predictions still exhibit substantial biases in radar echo intensity at individual locations. This observation motivates a more targeted strategy for reducing forecast errors. Instead of regenerating an entire radar echo sequence without spatial constraints, the predicted precipitation structure can be used to guide the refinement of echo intensities at individual locations. This structure-guided refinement directly targets echo intensity biases. Accordingly, we propose FreCast, a two-stage framework for radar echo prediction. The first stage generates an initial forecast of future radar echoes. The second stage uses the spatial structure of the initial forecast as a constraint to further correct intensity biases at individual locations in the first-stage prediction. Experiments on three datasets demonstrate that FreCast achieves consistent improvements across forecast skill metrics. Qualitative results further show that FreCast better preserves rainband continuity and intense precipitation structures at longer lead times.

Tropical Cyclone Forecasting via Latent Rectified Flow using Satellite Imagery and Atmospheric Fields

Tropical cyclones are growing more destructive in a changing climate, and efficient forecasting of their structure and track has become a necessity. Deep generative models promise an alternative to computationally expensive numerical weather prediction (NWP), yet current systems produce either satellite imagery or atmospheric fields, never both; they need many sampling steps, putting them out of reach of modest hardware; and their storm tracks come from regression heads with no physical link to the generated atmosphere. This work presents a single-pass model that jointly forecasts GRIDSAT-B1 infrared imagery and four ERA5 atmospheric fields (U-wind, V-wind, air temperature, and surface pressure) out to nine hours. A five-channel variational autoencoder compresses each 5 x 256 x 256 frame to a 4 x 64 x 64 latent, and a conditional rectified-flow UNet with a factorized temporal-attention module predicts the next three frames from three past frames, their best-track coordinates, and timestamps. The model is then reward-fine-tuned (DRaFT) against a differentiable track error derived from the predicted winds through a steering-flow calculation. On held-out 2022 storms the model reaches 16.35 dB PSNR and 0.759 SSIM, ahead of a reproduced cascaded-diffusion baseline at every lead time (+0.84 dB at +9 h) while sampling ~30x faster (56 ms vs. 1673 ms). Track error at +9 h is 62.4 km, 15% below the baseline, and a reward fine-tuning study demonstrates a further 8-11% track-error reduction across sampler budgets.

Timestep-Conditioned Transformers for Global Weather Forecasting

Existing machine-learning weather forecasting models rely on predetermined and fixed autoregressive timesteps. The choice of model timestep involves a fundamental trade-off: shorter timesteps (e.g. 1 to 6 hours) finely resolve atmospheric dynamics within the diurnal cycle but increase error accumulation for a given forecast horizon, while longer timesteps (e.g. 24 hours) reduce error accumulation but limit the usability of short-range forecasts where sub-daily predictability is high. In this work, we present GEM-3, a probabilistic global weather model that addresses this trade-off through explicit multi-timestep inference. With a single set of trained weights, the model timestep can be configured at inference time to balance predictability and usability across a broad forecast horizon. Additionally, we find that mixed-timestep training consistently improves rollout stability relative to timestep-specialist models. Under the hood, GEM-3 is a lightweight neighborhood-attention transformer with ~134M parameters on an equirectangular grid with a number of architectural advancements beyond its predecessor GEM-2. The result is a practical forecasting system that couples near-SOTA medium-range probabilistic skill, stable extended-range rollouts, efficient training and inference, and decision-relevant diagnostics.

MarsCast: Transfer Learning of AI Weather Foundation Models to Planetary Atmospheres

We investigate the transferability of Earth weather foundation models to planetary atmospheres by adapting the GraphCast graph neural weather forecasting model to Mars. While GraphCast achieves state-of-the-art performance for terrestrial forecasting, its applicability to non-Earth environments remains unexplored. Using the Mars Climate Database (MCD), which provides global atmospheric fields across vertical altitude levels (similar to Earth pressure levels), we evaluate zero-shot and fine-tuned GraphCast predictions of Martian temperature and wind fields. Zero-shot forecasts produce a surprisingly accurate depiction of current conditions but fail to reproduce diurnal variability and rapidly decay toward climatological mean states. To address this limitation, we fine-tune GraphCast using MCD variables and top-of-atmosphere solar radiation forcing while holding humidity constant. Fine-tuning enables rapid learning of Martian thermal variability. Within as few as 10 training epochs, the model begins to capture the diurnal cycle and forecasts up to 10 days reproduce seasonal and vertical temperature structure. Prediction quality improves with training sample size and exhibits sensitivity to seasonal initialization. These results demonstrate that Earth-trained AI weather models can be adapted to simulate Martian atmospheric dynamics, providing a pathway toward rapid planetary weather prediction to support mission operations, dust storm risk mitigation, and future human exploration.

Do AI weather models miss extremes?

First-generation AI weather models are often reported to underperform at extremes, mostly in reanalysis-based evaluations of deterministic regression systems. We verify eleven physical and AI forecast systems against European synoptic, solar, and rain-gauge stations over ten months for 10 m wind, 2 m temperature, hourly shortwave accumulation, and hourly precipitation, scoring mean absolute error (MAE) against ECMWF IFS in ERA5 1991-2020 climatological regimes. Among these systems, AI models do not show a uniform relative-skill deficit in the tails. Jua EPT-2.1 Europa leads all-conditions wind (+8.4%), while Jua EPT-2 HRRR leads temperature overall (+12.1%) and in the heat regime (+19.6 +/- 2.2%). EPT-2.1 Europa and DWD ICON Global lead at gale-force wind. Jua EPT-2.1 Helios leads solar overall (+10.2 +/- 1.7%), in overcast conditions (+16.4 +/- 3.4%), and in the clear-sky tail (+24.8 +/- 5.4%). For precipitation, three Jua models gain 14-15% at moderate intensity and 9-11% at P75-P95; EPT-2 Reasoning remains ahead above P95 (+1.7 +/- 0.5%). Failures are model-specific: ECMWF AIFS loses 4.9 +/- 2.0% in the heat tail, while NOAA GFS loses 22.8 +/- 2.0% there. Every model, including numerical weather prediction systems, shows a shared conditional bias toward the centre of the observed distribution, with an inter-model spread several times smaller than the shared signal. Missing relative skill at extremes is therefore not a property of AI weather models as a class, but of particular AI and physical models.

Pushing the Limits of High-Resolution Weather Forecasting through Data Scaling

The development of 0.1$^{\circ}$ global weather forecasting models based on machine learning (ML) is constrained by the limited availability of high-resolution data, as decades of reanalysis are only available at 0.25$^{\circ}$ resolution. While existing approaches fine-tune 0.25$^{\circ}$ forecast models on limited 0.1$^{\circ}$ samples, we show that this transfer is hindered by the irreversible information loss inherent in coarse-resolution forecasting. Therefore, we propose BaguanHR, a framework that shifts the focus from transferring models to transferring data. We first show that super-resolution (SR) has lower conditional entropy and input amplification than forecasting, making it a more robust vehicle for resolution transfer. By leveraging this advantage through variable-wise SR, we synthesize extensive 0.1$^{\circ}$ data from ERA5. BaguanHR's performance on the synthetic-plus-real dataset exceeds both ML-based methods and IFS-HRES, achieving superior performance across over 85% of the lead times within 72 hours. Furthermore, our findings highlight a power-law scaling effect, as a twofold increase in data reduces RMSE by 4.6% for 72-hour forecasting and 4.9% for 120-hour forecasting. Our results demonstrate that scaling high resolution ML-based forecasting is primarily a data bottleneck, and that variable-wise super-resolution provides a simple yet general solution to unlock long coarse-resolution reanalyses for high-resolution training.

Weather Emulators at the Frontier of Heat Extremes Predictability

Atmospheric predictability declines rapidly beyond the next ten days, such that forecasts at longer lead times primarily convey large-scale trends rather than specific states. Yet in a warming world, improving early warnings of extreme heat is an increasingly critical challenge. Here we evaluate six state-of-the-art deep learning weather emulators - Pangu-Weather, FuXi, ArchesWeather, AIFS, GraphCast and Aurora - alongside leading dynamical systems and statistical baselines in forecasting global near-surface temperature and extreme heat at lead times of 10-15 days. We find that several emulators rival or even surpass physics-based forecasts in deterministic temperature skill, but do so at the cost of reduced spectral fidelity, in a process widely known as blurring. While all models show some degree of predictive skill for extreme heat, most emulators under-represent peak intensities, and IFS recall is greater than that of any of the emulators. These results highlight both the emerging potential of AI to enhance extended range temperature prediction, and the remaining challenges in delivering reliable, actionable early warnings in a changing climate.

Rescene: band-limited stochastic forcing turns a frozen neural weather operator into a climate emulator

Over the past few years, the rapid development of machine learning (ML) models for weather forecasting has produced deterministic models whose medium-range skill matches or exceeds that of the European Centre for Medium-Range Weather Forecasts (ECMWF)'s high-resolution forecast (HRES). However, when these models are integrated freely beyond the horizon they were trained for, they blow up, drift, or lose their seasonal cycle, and retraining them for stability is expensive. We therefore ask what can be recovered from a strictly frozen backbone. We present Rescene, a 0.4 M-parameter wrapper around a frozen 1.5 degree, 6-hourly vision-transformer operator, developed using ERA5 reanalysis data and comprising a deterministic "slow clock" (0.33 M) that blends the forecast toward a lead-aware day-of-year climatology and a generative head (0.06 M) that adds a spectrally shaped stochastic perturbation at every step. The performance evaluation demonstrates that the deterministic wrapper alone is stable for decades but collapses daily variability to 40% of ERA5. Adding the generative head restores 126% (Z500) and 130% (MSLP) of the observed daily variability with pattern correlations of 0.89 and 0.92, recovers 82% of the observed blocking frequency, keeps the ensemble calibrated (spread-skill ratio 0.78-0.97 from day 7 to day 90), and integrates for 100 years with no detectable drift (+0.008 +/- 0.014 K per century). Moreover, because the perturbation is band-limited to total wavenumber $k \le 20$, the small scales are never forced, yet realistic $k \ge 20$ power is sustained: a direct decomposition of the 6-hourly energy budget shows that the frozen operator supplies 28 times more energy than the perturbation at $k \ge 40$, with a fractional growth rate 247 times larger at the grid scale than at planetary scales.

AIFS-TC: A simple correction competitive with the operational frontier for tropical cyclone intensity forecasting

AI weather models are in the process of revolutionising weather forecasting. While these models have been shown to achieve superior performance to physics-based NWP in forecasting tropical cyclone (TC) tracks, they tend to dramatically underestimate intensity. Here we present AIFS-TC, a simple correction to the AIFS-Single model that is competitive with the operational state-of-the-art for forecasting maximum wind speed and minimum central pressure at lead times of 12 h to seven days. This performance also holds for rapid intensification events. Notably, the entire system was autonomously designed and built by a large language model (Claude Fable 5) in a few hours, directed through a small number of natural-language prompts by a single domain scientist. That the operational frontier can be reached with an open-source AI forecast model (AIFS-Single) and relatively simple, cheap post-processing is significant for TC science, and points to agentic coding as a route to rapid exploration and progress in life-saving early-warning systems in other domains.

Explaining Weather Bulletins via ILP

Inductive Logic Programming (ILP) originated within the Logic Programming community in the Nineties as a framework for combining symbolic learning with declarative knowledge representation. Nowadays, mature ILP frameworks exist and they are capable of learning complex, non-monotonic hypotheses, thus broadening both the modeling capabilities and the scope of real-world applications of ILP. This work is primarily based on the FastLAS2 framework and aims to generate simple, interpretable hypotheses to help clarify the weather bulletins issued by OSMER FVG, the Regional Meteorological Observatory of the Italian region of Friuli Venezia-Giulia. In this paper we present a pipeline that, starting from simulated meteorological raw data and from OSMERs' bulletins (used as ground truth), extracts data as ASP facts and generates ILP examples. From such examples an explanatory hypothesis is then inferred via FastLAS2. Such a hypothesis (translated into natural language) explains the weather forecast issued by human experts, and in particular the rationale behind experts' choices of specific symbols in the bulletin pictogram (the symbol-annotated meteorological map of the forecast). The proposed approach is general, not specific to any particular region and it can equally be applied to bulletins from other sources and to different regions.

Nipping the Butterfly Effect in the Bud: Self-Output Fine-Tuning for Autoregressive Weather Prediction

Long-horizon weather forecasting is a fundamental challenge in atmospheric science, for which autoregressive Deep Learning Weather Prediction (DLWP) has emerged as the primary paradigm. Although the autoregressive pipeline is highly scalable and flexible, its prediction errors grow rapidly over long forecasting horizons. In this work, we study this error growth phenomenon from both theoretical and empirical perspectives. Our analysis reveals that the growth is driven by a feedback loop between output errors and input distribution shifts. Specifically, the autoregressive process amplifies small initial output errors, which progressively corrupt subsequent input distributions, echoing the butterfly effect in atmospheric science and ultimately deteriorating forecasting accuracy over longer horizons. Furthermore, we show that this distributional shift originates at the earliest stage of inference, with out-of-distribution signatures detectable as early as the first autoregressive step. To mitigate this issue, we propose \textbf{Self-Output Fine-Tuning (SOFT)}, a plug-and-play strategy that leverages the model's own one-step predictions to calibrate the biased input distribution encountered at the first step. Extensive experiments demonstrate that, despite its simplicity, SOFT achieves state-of-the-art performance on long-horizon forecasting tasks and substantially reduces both prediction errors and distributional discrepancy. The success of SOFT highlights the importance of reexamining the fundamental pipeline of deep learning weather prediction, representing a critical pipeline advance for atmospheric science.

Toward Mechanistic Interpretability of an AI Foundation Model Fine-Tuned for Atmospheric Chemistry

Weather forecasting foundation models (FMs) are increasingly fine-tuned to predict air quality, offering fast global pollution forecasts at lower computational cost than conventional chemical transport models. These FMs are typically trained on reanalysis data and generate forecasts through autoregressive rollout. They do not explicitly represent governing physical or chemical processes. Therefore, high forecast skill does not reveal whether a model has learned physical mechanisms or exploits statistical regularities in its training data. Here, we present the first study of what a FM fine-tuned for atmospheric chemistry has learned by examining Microsoft's Aurora model. We impose controlled chemical perturbations on its forecasts and test them against known photochemical relationships. We then examine the internal representations that generate these forecasts. We find that Aurora captures a first-order ozone response to reactive nitrogen but does not enforce the chemical constraints that a process-based model encodes. It generates chemically inconsistent combinations of related species and relaxes localized emission features such as wildfire plumes toward background. Internally, its representations remain largely organized around the meteorology inherited during pretraining, with little structure specific to chemistry. Using sparse autoencoders, we identify internal components that causally control the chemical forecast but do not map cleanly onto individual atmospheric processes. This work provides a framework for testing whether AI forecasting systems learn atmospheric chemistry from reanalysis data. As these models are increasingly positioned to inform environmental policy decisions, we argue that composition forecasts should also be judged by their internal mechanisms rather than by benchmark skill alone.

Geospatial Diffusion-based Evolution Synthesis (GeoDES) for Storm-Centered Weather Augmentation

While machine learning-based weather models hold significant promise, they struggle to predict the detailed structure of large-scale weather systems such as cyclonic storms. Regional models are constrained by limited historical records within fixed geographic boundaries, while global models are computationally expensive and often operate at resolutions too coarse to capture fine-grained storm dynamics. To bridge this gap, we introduce the Geospatial Diffusion-based Evolution Synthesis (GeoDES) model, a custom image-to-video diffusion model. By focusing generation strictly on the evolving storm structure, GeoDES synthesizes physically consistent, high-fidelity weather events suitable for stress-testing forecast models and expanding meteorological datasets. Evaluations demonstrate that GeoDES outperforms prior methods on key metrics, achieving $52\%$ lower Peak Vorticity Error and $8\%$ higher Anomaly Correlation Coefficient than the next strongest methods on the North Atlantic test set.

On the sensitivity of machine-learned probabilistic weather forecast models to scale-aware scoring rules

Probabilistic forecast models can be machine-learned from data using loss functions based on scoring rules such as the Continuous Ranked Probability Score (CRPS). This note summarises a preliminary study comparing versions of AIFS-CRPS, a global weather forecast model, trained with different univariate and multivariate scoring rules that aim to explicitly represent scale-awareness in the loss function. In the first part, we compare the (almost) fair CRPS, a fair global energy score, and a graph energy score based on node neighbourhoods. Across standard verification metrics, forecast skill is broadly similar. In the extratropics we find only small differences, while in the tropics the graph energy score setup performs somewhat better and the global energy score shows some degradation. These results suggest that multivariate scores are a viable alternative to CRPS-based training for global machine-learned weather forecasting. In the second part of the study, we analyse how different scoring rules and scale-aware loss constraints shape the spectra of forecast fields. It is apparent that any form of explicit scale-awareness improves realism. Here, the largest differences are likely associated with different effective weights per scale.

Apeliotes: A Diffusion-Based Modeling Framework for km-scale Multi-Level Atmospheric Fields

High-resolution atmospheric data are required to resolve mesoscale and localized meteorological structures, however such datasets remain limited in many regions of the world. Existing high-resolution weather products are typically produced through dynamical downscaling, which is computationally expensive and difficult to scale across locations, variables, and forecast scenarios. These limitations motivate machine-learning-based downscaling systems that can generate multiple weather variables stochastically while producing new high-resolution fields directly. In this paper we present Apeliotes, a framework for high-resolution weather forecasting. Built on the global re-analysis atmospheric data, a pre-trained global weather foundation model, and a regionally trained generative diffusion model, Apeliotes not only provides accurate kilometer-scale weather variables, but also multi-level atmospheric fields which are not directly available in the existing global atmospheric data. Our comprehensive evaluation demonstrates that Apeliotes achieves highly competitive performance. The model predicts vertical wind profile with less than 3\% error between truth and predicted fields, achieving correlations of 0.91 for 10-m wind speed and 0.99 for 2-m temperature, with NRMSE values of 0.42 and 0.17, respectively.

TSSM: Triaxial State Space Model for Global Station Weather Forecasting with Temporal-Variable-Historical Modeling

Global Station Weather Forecasting (GSWF) is pivotal for localized and extreme weather prediction over key regions. Despite efforts to exploit look-back windows, existing methods show limited accuracy gains and struggle with extreme events and error accumulation. These limitations stem from overreliance on short-term patterns, which are insufficient to capture chaotic weather dynamics, especially under partial observations. To address this problem, we propose a novel Triaxial State Space Model (TSSM) with a history-enhanced Temporal-VariableHistorical paradigm, which incorporates period-aligned historical weather data to compensate for long-term, large-scale periodic, and full-window weather patterns beyond the temporal lookback window. Specifically, TSSM stacks historical samples into period-aligned batches, where forecasting is causally supported by historical and current observations. Temporal, variable, and historical scanning are designed to capture axial temporal dependencies, variable correlations, and historical evolution. This structure is hierarchically shared to model seasonal to extreme events while alleviating misalignment across historical patterns. TSSM achieves SOTA performance on Weather-5K, the largest station weather dataset to date, with 10% and 61% gains in accuracy and extreme event metrics, and obtains 95% best or second-best results on human-involved datasets. Its advantages are more pronounced in long-horizon and iterative forecasting, reaching a 37.5% gain at 240h and up to 103.5% under a 48h times 5 iterative setting. Moreover, TSSM retains > 90% performance under up to 80% missing observations, compared with < 43% for baselines, demonstrating robustness and practical potential for reliable GSWF in global in-situ observation networks.

An adaptive and evolvable deep reinforcement learning framework for weather prediction

No single AI weather model excels at all variables, pressure levels, and lead times. Rather than building yet another architecture, we reframe the forecasting problem as one of coordination. Here we present Feitian Adaptive Ensemble Weather (FTAE-Weather), a lightweight framework that learns, through deep reinforcement learning, when and where to trust each member of an open pool of pretrained forecasters. A tactical Weight-Agent reads the current atmospheric state and assigns variable- and horizon-specific fusion weights, while a strategic Evolve-Agent periodically prunes underperforming models and absorbs newly released ones. Asynchronous prediction caching keeps training cost independent of the slowest constituent model. Adding fewer than 0.01 percent extra parameters, FTAE-Weather reduces RMSE by from 17.2 percent to 78.3 percent over the best individual model in 10 atmospheric variables and outperforms conventional ensemble baselines across lead times from 72 to 360 hours. The framework thus converts a growing, fragmented inventory of specialist models into a single prediction system that strengthens as the field of AI weather forecasting releases new architectures-turning model diversity from a coordination challenge into a compounding scientific advantage.

AIFS-SUBS: Extending Data-Driven Forecasting to Sub-Seasonal Timescales

Data-driven models now rival numerical weather prediction in the medium range, but extending them to sub-seasonal lead times raises challenges absent at shorter horizons. Errors accumulate over long autoregressive rollouts, systematic biases grow with lead time, and several years of data must be held out for independent verification, even though machine-learning models otherwise benefit from longer training records. To address these challenges, we adapt ECMWF's AIFS-CRPS medium-range model. AIFS-SUBS adopts a 24h autoregressive time step to reduce error accumulation, adds stratospheric levels and top-of-atmosphere thermal radiation as predictors, and reserves 2007--2011 as an independent verification window. We evaluate two config-durations: AIFS-SUBS, fine-tuned on operational analyses, and AIFS-SUBS-ERA5, trained on ERA5 alone. Across weeks 2--6, AIFS-SUBS matches the operational Integrated Forecasting System (IFS) in probabilistic skill while reducing systematic biases. For the convective (OLR) component of the Madden--Julian Oscillation (MJO), AIFS-SUBS extends skilful forecasts (correlation > 0.5) by eight days relative to the IFS, while matching or exceeding the IFS for the full multivariate RMM index. AIFS-SUBS also reproduces the observed MJO modulation of tropical cyclone activity comparably. Stratospheric skill is particularly strong with AIFS-SUBS reproducing sudden stratospheric warming (SSW) frequency and surface impact. In the AI Weather Quest, AIFS-SUBS-ERA5 attains a variable-averaged ranked probability skill score slightly ahead of the IFS at weeks 3 and 4. At inference, AIFS-SUBS uses about 200 times less energy than the IFS, opening the door to much larger real-time ensembles. AIFS-SUBS is ECMWF's first machine-learning model targeted at sub-seasonal time-scales.

Enhancing the Forecasting Capability of Multi-Model Blending Algorithms for Extreme Precipitation via Joint Use of Station and Gridded Observations

Accurate extreme precipitation forecasting is critical for disaster mitigation but remains challenging for numerical weather prediction (NWP) models due to systemic intensity underestimation and spatial displacement. Traditional precipitation multi-model blending algorithms perform pixel-by-pixel blending on the forecast field based on weights, which may lead to the expansion of precipitation areas and the smoothing of extreme values. This study proposes an U-Net based two-stage framework: probability classification followed by value reconstruction, to blend forecasts from six major NWP models. A novel station-grid joint supervision mechanism is introduced by integrating observations from 2411 national meteorological stations in China into the loss function, simultaneously constraining spatial structures and peak intensities. Evaluations using independent samples from the 2025 flood season demonstrate that our model significantly outperforms both individual NWPs and current operational products. For rainstorms (>=50 mm), the Threat Score (TS) improved by 38.4% compared to the best NWP. Notably, for extreme events (>=100 mm) driven by extratropical cyclones and the subtropical high, the model successfully elevated the TS to above 0.1, transforming forecasts from having negligible reference value into those with certain operational utility. Furthermore, the model exhibits data-driven spatial correction capabilities, effectively realigning systematic rainbelt displacements with actual precipitation centers. The inclusion of station observations specifically enhanced the TS for rainstorms by 10.4% and effectively balanced the Bias. These results highlight the efficacy of multi-source joint supervision in enhancing the capture of extreme precipitation events.

Less Tokens, Better Forecasts: Sparse Residual Routing for Efficient Weather Prediction

Existing ViT-based weather forecasting models apply uniform computation across all spatial tokens, even though nearby atmospheric grid points often contain similar values and large regions evolve smoothly over time. This makes much of the intermediate per-token computation redundant. Standard token-efficiency methods, such as pruning or merging, reduce cost by removing or fusing tokens. However, weather forecasting is a spatiotemporal dense prediction problem in which a history of atmospheric states must be mapped to future values on the original latitude-longitude grid. Thus, every grid cell must retain a physically meaningful representation, especially under autoregressive rollout. We introduce Sparse-Reslim, a parameter-free plug-in routing module that makes sparse token processing compatible with this fixed-grid requirement. Sparse-Reslim routes only 25% of spatial tokens through the expensive middle transformer blocks and treats those blocks as residual updates: it computes the change produced for the routed tokens and scatters only this delta back to the full sequence. Unselected tokens keep their pre-routing representations exactly, so no grid cell is dropped or replaced by a mask token, and no fusion layer or additional parameters are introduced. Across ERA5 resolutions up to the operational 0.25\textdegree{} standard and two model families, a deterministic Transformer and a diffusion model, Sparse-Reslim improves forecast accuracy on every evaluated variable while substantially reducing cost: training is about 2.5x faster in the main settings and reaches 3.18x speedup at 0.25\textdegree{}, with over 2.2x lower peak memory. A controlled decomposition shows that the accuracy gain comes primarily from sparse routing itself, while random token selection provides an additional regularization benefit without selector overhead.