Global Models¶
-
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...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. more
-
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...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. more
-
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...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. more
-
Prithvi-Precip: Integrating Satellite Observations into an Atmospheric AI Foundation Model for Precipitation Forecasting¶
Accurate precipitation forecasting remains one of the most challenging problems in weather prediction. While recent AI weather prediction (AIWP) systems have achieved substantial improvements in...Accurate precipitation forecasting remains one of the most challenging problems in weather prediction. While recent AI weather prediction (AIWP) systems have achieved substantial improvements in medium-range forecasting skill, precipitation often remains a secondary target and is commonly learned from reanalysis datasets that contain considerable uncertainty. In this work, we investigate two complementary strategies for improving AI-based precipitation forecasts. Building on the Prithvi-WxC foundation model, we develop Prithvi-Precip, a global precipitation forecasting system, and examine (1) the impact of training targets derived from satellite-based precipitation estimates rather than reanalysis fields and (2) the direct assimilation of satellite observations into the forecasting model. We systematically evaluate key design choices for finetuning the Prithvi-WxC AI foundation model for precipitation forecasting. We find that autoregressive rollout training produces substantially more accurate forecasts than direct conditioning on forecast lead time. Using independent radar-based precipitation estimates for evaluation, we show that training on satellite-derived precipitation targets yields improved forecast accuracy relative to training on MERRA-2 precipitation fields. Furthermore, direct ingestion of satellite observations provides additional improvements at short lead times, with the largest gains occurring in tropical and subtropical regions. Together, these advances enable Prithvi-Precip to substantially improve upon directly comparable precipitation forecasts from the Goddard Earth Observing System. Our results highlight the potential of improved precipitation targets and the direct integration of satellite observations as promising pathways for advancing medium-range AI precipitation forecasting. more
-
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,...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. more
-
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...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 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. more
-
Spatial Generalization Tests for Machine Learning-based Weather Models to Assess Physical Consistency¶
Machine learning-based weather prediction is revolutionizing weather forecasting by learning from weather data in present-day climate. However, generalization to other climates remains a major...Machine learning-based weather prediction is revolutionizing weather forecasting by learning from weather data in present-day climate. However, generalization to other climates remains a major challenge. With melting sea ice, land-use change, and increasing ocean temperatures, boundary conditions are changing. Therefore, generalization in time depends on generalization in space. Here, we present three test cases to evaluate whether machine learning-based weather and climate models generalize in space and apply them to GraphCast and NeuralGCM. We reverse or rotate the planet in longitude or latitude under the model's coordinate system and adapt all boundary conditions and forcings accordingly. Physics-based general circulation models simulate a rotated/reversed planet with only rounding errors, but GraphCast and NeuralGCM fail these tests. The analyses furthermore revealed unphysical variable mappings based on correlation rather than causation. We argue that machine learning-based climate models should be designed to pass generalization tests to prevent overfitting on present-day regional climate. more
-
Aircast-Mars: A Mars Foundation Model for Global Weather Forecasting with HEALPix-Aware Convolutions¶
Foundation models for planetary atmospheres promise fast, lightweight surrogates of expensive general circulation models (GCMs) for mission planning and scientific inquiry. Here we present...Foundation models for planetary atmospheres promise fast, lightweight surrogates of expensive general circulation models (GCMs) for mission planning and scientific inquiry. Here we present Aircast-Mars, a deep-learning weather prediction system for Mars trained on the Ensemble Mars Atmosphere Reanalysis System (EMARS) v1.0. We regrid temperature, zonal wind, and meridional wind fields across 28 vertical levels onto a hierarchical equal-area isolatitude pixelization (HEALPix) mesh at Nside = 64 (~110 km resolution) and train a HEALPix-aware 2D U-Net inspired by the DLESyM architecture to predict the next hourly atmospheric state. The model employs custom inter-face padding that respects the topology of the 12-face HEALPix sphere and modern ConvNeXt residual blocks with capped Gaussian Error Linear Unit (GELU) activations. While containing 4.3 million trainable parameters, a compact size compared to terrestrial weather foundation models, the network achieves a best validation Mean Squared Error (MSE) of 1.58e-5 in normalized units. Recursive autoregressive rollouts remain stable and physically coherent for 25 hours (one Martian sol), with Root Mean Square Error (RMSE) growing monotonically from ~0.004 at t + 1 h to ~0.031 at t + 25 h without divergence. Compared to a baseline 3D U-Net, the HEALPix-aware architecture reduces validation loss by more than an order of magnitude while using fewer parameters. The model generates a one-hour forecast in approximately 0.5 seconds on a single GPU, offering several orders-of-magnitude speedup over traditional numerical GCMs. These results demonstrate that parsimonious, geometry-respecting neural architectures can capture synoptic-scale Martian atmospheric dynamics and provide a foundation for planetary-scale weather forecasting. more
-
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...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. more
-
Hard conservation correctors can hide a degrading model when training autoregressive emulators¶
AI weather and climate emulators increasingly incorporate physical principles into their formulation. One approach is to apply hard correctors that modify network outputs so that global mass, water,...AI weather and climate emulators increasingly incorporate physical principles into their formulation. One approach is to apply hard correctors that modify network outputs so that global mass, water, or energy budgets close. Prior work introduced such training-time correctors in the CREDIT framework and reported reduced precipitation bias and improved stability. Motivated by those results, we fine-tuned a global atmosphere emulator with a water-budget corrector, using the corrected prediction in the supervised loss and evaluating through post-correction budget closure. By that measure, training appeared successful. Every delivered field closed the moisture budget to machine precision. However, raw precipitation developed a growing global low bias over 18 training epochs, while the required correction increased from about 2% to roughly 24%. The cause is a scale degeneracy. A uniform change in raw precipitation amplitude is offset by a compensating change in the correction factor, leaving the corrected field, and therefore the supervised loss, unchanged. This invariance removes the restoring force on raw precipitation amplitude, allowing other training pressures to drive drift. Two changes recovered stable behavior. We supervised the pre-correction prediction and penalized its raw budget imbalance, while the hard correction remained in place for the delivered field. The required correction returned to less than 1% within the next epoch. A controlled 2x2 ablation showed that the runaway occurred only when corrected-output supervision was combined with no imbalance penalty. Exact post-correction closure therefore says little about whether the raw model has learned the budget. When a corrector removes information from the loss, the raw fields and the applied correction need to be tracked. more
-
Fourier Geometric Wind Power Forecasting with Numerical Weather Prediction¶
Accurate short-term wind power forecasting is essential for grid stability and operational planning, yet remains challenging due to the complex interactions between atmospheric conditions and turbine...Accurate short-term wind power forecasting is essential for grid stability and operational planning, yet remains challenging due to the complex interactions between atmospheric conditions and turbine dynamics. However, existing methods fail to effectively incorporate weather forecasting with wind turbine data (i.e., SCADA), leading to suboptimal solutions. To address this, we introduce a multimodal framework that integrates historical point-based SCADA data with grid-based Numerical Weather Prediction (NWP) forecasts, which is challenging due to heterogeneous input and the complex physical wind-turbine interactions. Our approach first explicitly decomposes inputs into scalar and vector features to better capture both site-specific and geometric dependencies and then incorporates a geometric encoder to extract rotation-invariant features from wind vectors. We further leverages a Fourier Neural Operator (FNO) architecture, which performs global convolutions in the frequency domain to efficiently model long-range spatiotemporal relationships. Extensive experiments on three real-world wind farms, with weather forecasting data, demonstrate that our model consistently outperforms state-of-the-art baselines, highlighting the effectiveness of its physically-informed design. The core implementation of our method is publicly available at: https://github.com/shawn-sypiao/GWPF. more
-
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...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. more
-
Robustness of Deep Learning Models for PV Power Forecasting under NWP Forecast Errors: A Spatiotemporal and Physically Interpretable Analysis¶
Engineering use of AI forecasting models requires not only high nominal accuracy but also predictable behavior under uncertain inputs. In photovoltaic (PV) forecasting, this requirement is especially...Engineering use of AI forecasting models requires not only high nominal accuracy but also predictable behavior under uncertain inputs. In photovoltaic (PV) forecasting, this requirement is especially challenging because numerical weather prediction (NWP) errors are temporally correlated, state dependent, and physically coupled across variables. Existing evaluations, however, often rely on perfect forecast assumptions or simplistic perturbations that do not reflect these characteristics. This study presents a physically constrained robustness evaluation framework based on simulation, using virtual PV power as a controlled response variable to isolate the propagation of input uncertainty from confounders at the plant level. Six representative machine learning and deep sequence models, including PatchTST, GRU, N-HITS, and LightGBM, are evaluated under dynamic NWP perturbations with heteroscedasticity modulated by clear-sky conditions and Erbs reconstruction that preserves radiation consistency. The results show that sequence models provide stronger noise filtering and temporal resilience than a strong tabular baseline under medium to high disturbance regimes. SHapley Additive exPlanations (SHAP) and Integrated Gradients (IG) further support a feature reallocation tendency at the case level, in which predictive reliance shifts from corrupted future forecasts toward more stable historical observations and deterministic physical priors. A Pareto analysis of accuracy under clean conditions, robustness, and computational latency then translates these findings into engineering implications for robustness assessment and model selection under forecast uncertainty. more
-
Evaluating the Fidelity of GraphCast AI Forecasts for the Indian Summer Monsoon: A Climatological Assessment Against ERA-5 Reanalysis and IMERG Observations¶
The Indian Summer Monsoon (ISM) represents one of the most consequential and dynamically complex phenomena in the global climate system, yet its prediction remains challenging for both physics-based...The Indian Summer Monsoon (ISM) represents one of the most consequential and dynamically complex phenomena in the global climate system, yet its prediction remains challenging for both physics-based and data-driven models. This study evaluates Google's GraphCast, a machine-learning-based global weather prediction system, against ERA5 reanalysis and IMERG precipitation observations for the boreal summer season (June-September, JJAS) over 2021-2024. We analyze deterministic GraphCast forecasts initialized at 00 UTC using four 6-hourly lead times, composited for +24 h, +48 h, and +72 h lead performance across the ISM domain. The evaluation includes climatological mean state, rainfall intensity distribution, thermodynamic drivers (DTT, Q1, Q2), monsoon dynamics, and variability across multiple timescales. Results show GraphCast reproduces the broad spatial pattern of monsoon rainfall and the annual cycle with good fidelity at short leads, but exhibits a domain-averaged wet bias over the core monsoon region and strong suppression of rainfall variance across nearly all timescales (regional power-spectrum variance ratio ~0.14 relative to IMERG). The rainfall intensity distribution is right-shifted and compressed, with moderate-heavy rainfall (95th percentile) biased wet and extreme events underrepresented. These biases are accompanied by a deficient lower-tropospheric Q1 profile and degraded northward-propagating intraseasonal variability. Overall, GraphCast shows a deterministic smoothing signature in precipitation, providing key diagnostics for next-generation AI weather models and their use in tropical extended-range forecasting. more
-
Global reanalysis from observations alone with machine learning¶
Earth system reanalysis datasets are foundational for weather and climate research and provide the gridded training data used by most machine learning weather prediction systems. Here we show results...Earth system reanalysis datasets are foundational for weather and climate research and provide the gridded training data used by most machine learning weather prediction systems. Here we show results from a prototype system that suggest that machine learning models trained only on Earth system observations can potentially be used to generate multi-decade global reanalyses without using physics-based numerical models. The resulting gridded fields capture large-scale atmospheric structure and variability across multiple timescales, while exhibiting signs of physical coherence in several key dynamical diagnostics. Evaluations of the prototype against held-out independent atmospheric observations indicate that the root mean square vector error of upper-level winds is close to that of ERA5 when compared at a consistent resolution, and that the standard deviation of the error at the surface is between that of 4th- and 5th-generation ECMWF reanalyses (ERA-Interim and ERA5). Furthermore, while traditional reanalysis production is computationally expensive, typically taking several years to produce, the reanalysis presented here was generated during the course of a single working day. These results suggest that observation-trained machine learning models offer a promising new approach for reanalysis production from observations alone. more
-
Integrating GNSS-Derived Zenith Wet Delay into a Weather Foundation Model Improves Precipitation Forecasting¶
Global Navigation Satellite Systems (GNSS), best known for positioning, also serve weather science, as atmospheric water vapour delays their signals. This delay, the Zenith Wet Delay (ZWD), is a...Global Navigation Satellite Systems (GNSS), best known for positioning, also serve weather science, as atmospheric water vapour delays their signals. This delay, the Zenith Wet Delay (ZWD), is a direct, all-weather measure of column moisture. Although assimilated into numerical weather prediction for decades, ZWD is not yet used by leading machine learning weather models (MLWM), despite addressing a known deficiency: the underestimation of severe precipitation. Here we present the first integration of GNSS-derived ZWD into Aurora, a state-of-the-art weather foundation model. Our extended Aurora learns ZWD with skill comparable to its pretrained variables. More importantly, including ZWD systematically improves forecasts when fine-tuning for six-hour accumulated precipitation. Gains grow with severity, reaching an 8.8% increase in Equitable Threat Score at the 99th percentile, while the precipitation power spectrum becomes more realistic at synoptic and planetary scales. Direct GNSS observations therefore encode information that MLWM can exploit for high-impact precipitation. more
-
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...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. more
-
On the Genealogy of Machine Learning Weather Prediction¶
Modern machine-learning weather prediction (MLWP) has largely inherited the initial-value-problem (IVP) framing of numerical weather prediction (NWP). This inheritance leads to a dominant paradigm of...Modern machine-learning weather prediction (MLWP) has largely inherited the initial-value-problem (IVP) framing of numerical weather prediction (NWP). This inheritance leads to a dominant paradigm of learned autoregressive time-stepping and constrains how the learning problem is defined and architectures are favored. In this study we make the inheritance explicit, contrast two philosophical traditions: "scientific surrogate modeling," where machine learning (ML) is embedded within a physical system and must respect its structural constraints, and "free-form data-driven modeling," where atmospheric fields are treated as spatiotemporal sequences and models learn latent dynamics without explicit physical constraints. By reviewing the governing primitive equations, surveying recent literature, and analyzing concrete physical examples, we map each modeling paradigm to either a state-conditioned or evolution operator formulation. We conclude that principled model selection requires explicitly aligning architecture and training objectives with either the physical system structure or the statistical structure of the data. more
-
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...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. more
-
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...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°{} 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°{}, 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. more
-
Enhancing a high resolution data-driven weather prediction model with surface descriptors¶
We study the importance of surface characteristics when forecasting near-surface variables with a data-driven weather prediction model. To target the challenge of predicting small-scale weather...We study the importance of surface characteristics when forecasting near-surface variables with a data-driven weather prediction model. To target the challenge of predicting small-scale weather conditions at high resolution, we introduce a range of surface descriptors in the training of a state-of-the-art data-driven model. The input data includes surface descriptors inherited from the numerical weather prediction model used to produce the training dataset and topographic neighbourhood indices. We found that errors of 2-metre temperature and 10-metre wind speed forecasts were reduced by 1.9% and 3.0% respectively compared to a baseline model over the model domain. Over certain surfaces, the improvements were significantly larger. For example, we found a 12% reduction of temperature mean absolute errors over urban areas when the urban fraction was included in the model input. Furthermore, we investigated how the model responded to removal of glaciers, resulting in an increase of temperature. This indicates that 1) the model produce a physically reasonable response and 2) input datasets can be updated without the need to retrain the model. The latter suggests a great benefit for operational systems as training is expensive compared to running these models. This study highlights the importance of including surface conditions in the prediction of near-surface variables. more
-
Modelling convective cell occurrence in proximity to cold fronts using extreme gradient boosting¶
Machine learning is emerging as a valuable tool for convection-related applications such as post-processing numerical weather prediction output, improving understanding of convective storm...Machine learning is emerging as a valuable tool for convection-related applications such as post-processing numerical weather prediction output, improving understanding of convective storm climatology and potentially improving existing convective parameterization schemes. In a rapidly developing field, it is vital to assess the strengths and limitations of machine learning approaches across different applications. Here, a probabilistic model is developed using a convective cell dataset as ground truth and predictors primarily from ERA5. The model's ability to reproduce the convective cell climatology at different regions relative to cold fronts (i.e. post-frontal and pre-frontal) is assessed during the warm-season in Germany. The optimal number of features (predictors) is selected using a feature elimination strategy. Overall, the optimised model exhibits high skill in reproducing the spatial and temporal cell frequency at different regions relative to the front. While the highest cell frequency is correctly identified near the surface front, the model underestimates the actual cell count in this region. Feature importance analysis shows that the model depends most heavily on CAPE to make its predictions. Additionally, the time of day predictor is key for accurately capturing the diurnal cycle of convective cells on both sides of the cold front. The study highlights both the advantages and the limitations of data-driven models, offering valuable insights for future data-driven climate and weather prediction models. more
-
Otter Weather: Skillful and Computationally Efficient Medium-Range Weather Forecasting¶
State-of-the-art medium-range AI weather models can outperform traditional Numerical Weather Prediction (NWP) but require massive training budgets. This restricts usage for under-resourced groups and...State-of-the-art medium-range AI weather models can outperform traditional Numerical Weather Prediction (NWP) but require massive training budgets. This restricts usage for under-resourced groups and severely limits fast model iteration. Here we develop Otter Weather, a highly efficient spatiotemporal forecasting model designed to democratise high-performance weather prediction with AI. Evaluated on ERA5 reanalysis data at 1.5° resolution using standard WeatherBench protocols, the Otter family significantly advances the skill-compute Pareto frontier. The deterministic version outperforms the best NWP baseline by 9.6% at a 24-hour lead time while requiring fewer than 3.5 A100-days for training. It provides a 2x efficiency gain over lightweight AI models and a 100-fold reduction in compute compared to resource-intensive frontier architectures. We extend these efficiency gains into probabilistic forecasting by training via the Continuous Ranked Probability Score (CRPS). Scaling to a larger architecture, Otter-XL achieves a 9.7% CRPS improvement over the IFS ENS baseline. This yields an almost two-fold increase in predictive skill over comparable lightweight models at similar compute budgets. Otter-XL also outperforms frontier architectures like GenCast by over 2%, while using an order of magnitude less compute. Finally, Otter is applied out-of-the-box to a complex acoustic scattering PDE task where it outperforms a state-of-the-art foundation modelling approach, suggesting that the advances made here might apply across a range of scientific domains. more
-
Event-Aware Loss Design for Forecasting of Convective Precipitation and Lightning¶
Accurate forecasting of high-impact weather, specifically extreme precipitation and lightning, remains a significant challenge in numerical weather prediction (NWP) due to the complexity of...Accurate forecasting of high-impact weather, specifically extreme precipitation and lightning, remains a significant challenge in numerical weather prediction (NWP) due to the complexity of atmospheric microphysics. While deep-learning models have shown promise in large-scale forecasting, they often suffer from systematic under-prediction of rare, high-intensity events and localized convective showers when optimized with conventional loss functions like Mean Squared Error (MSE). This study proposes an Event-Aware multi-task deep-learning post-processing framework designed to improve the representation of convective processes by leveraging lightning observations. The model jointly predicts precipitation amount, rainfall probability, and lightning occurrence using a shared-backbone Patch-cGAN (Conditional Generative Adversarial Network) architecture. To address the rare event problem, we introduce a lightning-informed loss-weighting strategy that element-wisely multiplies the MSE component by a spatial weight map derived from observed lightning strikes, forcing the model to prioritize accuracy in convective regions during training. Evaluations conducted over the Korean Peninsula during the 2025 Summer demonstrate that our framework outperforms standard AI benchmarks and conventional NWP models, particularly at intense rainfall thresholds (40 mm/6 h). Furthermore, the model exhibits superior skill in predicting lightning compared to conventional lightning parameterization and instability-index-based methods. These results indicate that integrating physical event indicators into the loss formulation effectively guides models to learn the meteorological signatures of deep convection, offering a pathway toward more reliable extreme weather forecasting. more
-
Evaluation of medium range machine learning models for sub-seasonal prediction¶
The performance of two machine learning (ML) atmosphere models - GraphCast and FourCastNetV2 - is evaluated in the context of sub-seasonal prediction, including their ability to represent key climate...The performance of two machine learning (ML) atmosphere models - GraphCast and FourCastNetV2 - is evaluated in the context of sub-seasonal prediction, including their ability to represent key climate drivers of variability, namely the Madden-Julian Oscillation and the Southern Annular Mode. Model skill is assessed over both a 38-year hindcast period and a 2.5-year hindcast period. The longer period overlaps with the training windows of the ML models but provides a larger sample for robust evaluation, while the shorter period is independent of the ML model training period. This dual evaluation illustrates a compromise approach to the problem of insufficient independent data for evaluation of the models for sub-seasonal prediction. The ML models are compared against the Bureau of Meteorology's physics-based seasonal prediction system, ACCESS-S2, for the 38-year period, and a more recent physics-based coupled model for the shorter hindcast period. Across the two evaluation periods, both ML models have surprisingly good skill for sub-seasonal timescales, given they were designed for forecasting on medium range timescales. In general, the ML models are as skilful as the physical model ensemble mean at shorter lead times and comparable to the physical model ensemble members at longer lead times. more
-
Machine learning is revolutionizing weather forecasting -- the next step is a change in how we work¶
Following the success of machine learning in producing weather predictions with competitive skill compared to complex traditional systems, this article shifts attention from forecast output to the...Following the success of machine learning in producing weather predictions with competitive skill compared to complex traditional systems, this article shifts attention from forecast output to the working practices that make prediction systems possible. We argue that machine learning and recent digital technologies will reshape the forecasting value chain: how models are coded and developed, how observations and Earth-system data are exploited, how data and computing are managed, how systems are verified, and how information is created, evaluated and turned into services. We discuss six non-exhaustive areas in which agentic software engineering, open and compressed data, shared verification workflows, interactive computing and generative methods may make modelling, evaluation and service creation faster, more interactive and more widely accessible. These changes will require weather and climate centres to adapt their infrastructures, data stewardship, trust and quality-assurance frameworks, skills and service delivery while maintaining scientific understanding, operational reliability, human expertise and their public-service role. more
-
ARCO-Mars: A Unified Cloud-Optimized Archive of Mars Atmosphere Reanalysis¶
Long-term records of the Martian atmosphere based on general circulation models and reanalysis of atmospheric state variables are important to understand the diurnal, seasonal, and climatological...Long-term records of the Martian atmosphere based on general circulation models and reanalysis of atmospheric state variables are important to understand the diurnal, seasonal, and climatological changes of the planet. Atmospheric dynamics of the Martian atmosphere are strongly influenced by the characterization of dust lifting, solar insolation, and spatial variations in topography. We present ARCO-Mars, a unified Analysis-Ready Cloud-Optimized dataset providing integrated access to three independent Mars atmospheric reanalysis products: EMARS, MACDA, and OpenMARS spanning over Mars Years 24-35. These reanalyses assimilate thermal infrared retrievals from the MGS/TES, ODY/THEMIS, and MRO/MCS instruments, providing both two and three-dimensional surface and atmospheric state variables, including temperature, winds, surface pressure, and dust optical depth. The dataset is stored in Zarr v3 format and hosted on HuggingFace, enabling efficient cloud-based access without requiring local storage of the full archive. We compare the state variables between the three reanalysis products to identify systematic differences, attributed to differences in data assimilation and general circulation models. ARCO-Mars provides a community resource for Mars atmospheric science, numerical weather prediction validation, and machine learning applications, including weather forecasting and data assimilation. more
-
Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction¶
Probabilistic weather forecasting is undergoing rapid transformation with artificial intelligence (AI). In traditional numerical weather prediction, computing power can limit how well ensemble...Probabilistic weather forecasting is undergoing rapid transformation with artificial intelligence (AI). In traditional numerical weather prediction, computing power can limit how well ensemble forecasts approximate the unknown statistical distribution of future states. AI models facilitate larger ensembles and are trained with probabilistic considerations, ideally leading to better uncertainty quantification. Forecasts from these state-of-the-art models are often considered well-calibrated. However, here we show that the statistical coverage of such models, the ultimate measure of calibration, can struggle, especially on extreme events. To address this shortcoming, we employ conformal prediction, a class of statistical methods that mathematically guarantees coverage under no distributional assumptions, unlike previous post-processing techniques. We apply online conformal prediction to temperature and precipitation forecasts (including extremes) of three leading global weather models, GenCast, NeuralGCM, and AIFS-ENS, ensuring calibrated uncertainty at no expense to other probabilistic metrics. This post-processing method can be applied to any forecasting model. more
-
AIFS-DOP: End-to-End Medium-Range Weather Prediction from Observations Alone with Machine Learning¶
We introduce the Artificial Intelligence Forecasting System for Direct Observation Prediction (AIFS-DOP). AIFS-DOP is trained on a 40-year harmonized dataset of gridded observations, without using...We introduce the Artificial Intelligence Forecasting System for Direct Observation Prediction (AIFS-DOP). AIFS-DOP is trained on a 40-year harmonized dataset of gridded observations, without using numerical weather prediction (NWP) reanalysis or model data. The resulting model is competitive with ECMWF's Integrated Forecasting System (IFS) when scored on a one year period of forecasts across 2021/2022. This progress on Direct Observation Prediction represents the first time that a data-driven model, trained solely on observations, is competitive with the IFS at medium ranges for several key upper-air and surface headline scores, when verified against observation data. more
-
A Hybrid LSTM--Vision Transformer Architecture for Predicting HRRR Forecast Errors¶
Forecast errors in high-resolution numerical weather prediction (NWP) systems are often linked to unresolved planetary boundary layer (PBL) processes, convection, terrain-induced circulations, and...Forecast errors in high-resolution numerical weather prediction (NWP) systems are often linked to unresolved planetary boundary layer (PBL) processes, convection, terrain-induced circulations, and other vertically structured atmospheric phenomena. Previous work demonstrated that Long Short-Term Memory (LSTM) networks can successfully predict forecast errors in the High-Resolution Rapid Refresh (HRRR) model using mesonet observations, but we believe performance degradation is linked to periods of complex vertical atmospheric evolution. To address this limitation, we develop a hybrid LSTM-Vision Transformer (LSTM-ViT) framework that combines temporal sequence learning from surface observations with atmospheric profiles from the New York State Mesonet profiler network. The LSTM-ViT framework is trained to predict HRRR hourly precipitation, 10 m wind speed, and 2 m temperature forecast errors at individual mesonet stations. Across all three predictors, incorporation of profiler-derived atmospheric structure improves forecast error prediction skill relative to the baseline LSTM architecture, with the largest gains occurring at shorter forecast lead times and during periods of enhanced PBL activity. Improvements are particularly pronounced for precipitation forecast error, where the LSTM-ViT framework achieves approximately a twofold increase in predictive skill relative to the baseline LSTM while better capturing convectively driven error evolution and reducing degradation associated with PBL processes. These results demonstrate that combining temporal sequence learning with vertically informed attention mechanisms provides a physically meaningful pathway for improving forecast error prediction in operational NWP systems. Our research offers forecasters enhanced guidance regarding model bias and forecast confidence. more