Topics may include:
- AI based Nowcasting methods and systems, use of observations and weather analysis
- Physics and AI driven Mesoscale and convection permitting modelling
- Development on AI for Ensemble prediction techniques and products
- AI for weather forecasting application
- AI for Seamless prediction and application
- Statistical and AI NWP Post-processing
- Use of machine learning, data mining and other advanced analytical techniques
- Presentation of results from relevant international research projects of EU, WMO, and EUMETNET etc.
Traditional weather forecasting relies on large scale numerical simulations that run on high-performance computing systems. These methods require substantial computational resources, involve complex workflows, and generate large volumes of data that often exceed individual user needs. Forecast-in-a-Box leverages advances in data-driven modelling to greatly reduce computational and energy costs while delivering tailored forecast products directly to users. Partly funded from the European Commission’s Destination Earth initiative, it packages the entire forecasting chain into a simple and user-friendly application. Built on the open-source Anemoi1 and Earthkit2 projects, it offers a reproducible and modular environment that integrates data access, model execution, and visualisation. This enables accurate forecasts that can be run locally on user desktops, on premise computing infrastructure, or in the cloud.
The approach is being evaluated through a World Meteorological Organization (WMO) Integrated Processing and Prediction System (WIPPS) pilot project led by the Norwegian Meteorological Institute (MET Norway). In this project, a fully packaged forecasting system based on affordable hardware is provided to the Malawi Department of Climate Change and Meteorological Services (DCCMS). The forecasting system is driven by Forecast-in-a-Box and leverages MET Norway’s Bris3 model (Norwegian word for “light wind), a high-resolution data driven weather forecasting model built using the Anemoi framework. The solution is designed to be largely self-contained, with the only external dependency being the retrieval of ECMWF analysis dataset for forecast initialisation.
1https://anemoi.readthedocs.io/en/latest/
2https://earthkit.ecmwf.int
3https://lumi-supercomputer.eu/data-driven-weather-forecasting-model/
How to cite: Carton de Wiart, C., Cook, H., Tuma, V., Wong, J., Futsæter, H. A., Østvand, L., Bønes, V., Moe, B., Kristiansen, J., Hawkes, J., Sandu, I., and Quintino, T.: Forecast-in-a-Box: AI weather forecasting, easy to run and simple to deploy, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20013, https://doi.org/10.5194/egusphere-egu26-20013, 2026.
Recent advancements in machine learning based weather prediction (MLWP) present novel opportunities for downstream applications like forecasting of renewable energy production from intermittent sources, like wind and solar. MLWP models guarantee shorter simulation run times and lower computational costs, allowing faster updates of downstream models and greater flexibility in the generation of weather scenarios.
Forecasting renewable energy generation critically depends on available weather forecast data at adequate temporal and spatial resolution. Using MLWP weather data in energy system modelling and forecasting has been limited by the coarse temporal resolution of the current generation of models (e.g. ECMWF’s AIFS Ensemble model runs at 6-hour time steps).
In Europe, power market participants are increasingly exposed to weather forecast inaccuracies. This is due to the combined effect of how the power price is calculated for each price area, and the recent increase in intermittent renewable installed capacities. In detail, power prices are set each day for the following day by balancing supply and demand for each Market Time Unit (MTU), which are now 15 minutes long. It is then massively important to benchmark weather forecasts on a time resolution closer to the power market MTU, to properly assess which period will potentially be oversupplied, or undersupplied from intermittent renewable sources. In this context, the 6-hour time resolution of current MLWP models becomes a significant limiting factor for their usefulness.
Using NVIDIA’s Earth2Studio framework, we demonstrate an efficient, integrated MLWP pipeline combining the [open source] AIFS model with the ModAFNO time interpolation model to provide 1-hourly time-resolution MLWP data. This interpolated data is applied to our intermittent renewable energy production models to assess the interpolation quality compared the uninterpolated AIFS data and the best-in-class numerical weather prediction data provided by ECMWF’s IFS Ensemble forecast.
How to cite: Brenna Schjønberg, H., Parviero, R., Koch, M., and Carpentieri, A.: ML-based time interpolation of AIFS Ensemble for renewable energy forecasting, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11158, https://doi.org/10.5194/egusphere-egu26-11158, 2026.
Minimizing pixel-wise errors in precipitation nowcasting inherently biases models toward smooth predictions, causing failures in resolving extreme convective events. To address this, we propose IMPA-Net, a meteorology-aware framework centered on spectral consistency. The architecture integrates three innovations: a parameter-free Spatial Mixer to encode multi-variate physical interactions (e.g., terrain-wind coupling); an Integrated Multi-scale Predictive Attention (IMPA) module to capture dynamics from Meso-β to Meso-γ scales; and a Meteorology-Aware Dynamic Loss (MAD-Loss) that employs asymmetric penalties to counteract regression-to-the-mean. Experiments demonstrate a 37.3% relative improvement in HSS for severe convection (≥45 dBZ). Crucially, RAPSD analysis confirms that IMPA-Net maintains spectral energy consistency across high-frequency bands, enabling it to successfully simulate the complex "dissipation-initiation" lifecycle that existing baselines fail to capture. These findings validate that integrating domain knowledge advances the physical plausibility of data-driven forecasting.
How to cite: He, G. and Cui, H.: IMPA-Net: Meteorology-Aware Multi-Scale Fusion and Dynamic Loss for Extreme Radar-Based Precipitation Nowcasting, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3224, https://doi.org/10.5194/egusphere-egu26-3224, 2026.
Machine learning-based global weather forecasts often suffer from coarse spatial resolution, limiting their ability to capture fine-scale temperature variability in regions with complex terrain or strong urban–rural gradients. We present SR-Weather, a two-stage deep learning framework that downscales coarse 0.25° forecasts into 1 km air temperature fields. Our model is trained using ERA5 and MODIS-derived temperature data, and leverages high-resolution auxiliary inputs, including elevation, impervious surface fraction, and spatial information–normalized air temperature to enhance spatial fidelity. Applied to 7-day lead forecasts from the FuXi model, SR-Weather consistently outperforms FuXi’s own 1-day lead predictions, indicating strong capabilities in both resolution enhancement and bias correction. The model also exhibits robustness under cloud-contaminated MODIS observations by reconstructing missing temperature values using auxiliary data. While developed and validated over South Korea, SR-Weather is region-agnostic and applicable globally due to the availability of MODIS inputs and minimal reliance on localized data. These results position SR-Weather as a scalable solution for high-resolution, ML-based weather forecasting.
How to cite: Park, H., Park, S., Kang, D., and Kim, J.-H.: SR-Weather: Super-Resolution Machine Learning Weather Forecast for 1-km Air Temperature Prediction, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-9075, https://doi.org/10.5194/egusphere-egu26-9075, 2026.
Extreme precipitation poses significant risks to society and infrastructure, highlighting the urgent need for accurate short-term nowcasting. While deep learning models have shown promise in precipitation forecasting, they often lack integration with physical principles, leading to inconsistencies and limited skill in predicting convective evolution. In this study, we introduce RainCast—a novel generative nowcasting framework that synergistically combines deterministic physical modeling with stochastic generative networks to improve the accuracy and physical consistency of extreme rainfall forecasts.
RainCast integrates a deterministic branch based on Neural Ordinary Differential Equations (Neural ODE) to simulate large-scale advective processes and a generative branch built upon a conditional diffusion model to capture fine-scale stochastic variability. The model is guided by key physical features such as flow fields, vorticity, and divergence derived from dual-polarization radar observations, which provide essential dynamical information about convective systems. We train and evaluate the framework using vertically integrated liquid water (VIL) data from dual-polarization radars in China (GD-SPOL) and North America (SEVIR).
Quantitative assessments demonstrate that RainCast significantly outperforms existing nowcasting methods such as SimVP, SwinLSTM, and NowcastNet. On the GD-SPOL dataset, RainCast improves the Critical Success Index (CSI) for intense convection (VIL ≥ 160) by up to 14.1% at 90-minute lead times. Structural similarity metrics also show substantial gains, with reductions in Fréchet Video Distance (FVD) by 25.4% and Learned Perceptual Image Patch Similarity (LPIPS) by 44.6%. Case studies further illustrate RainCast’s ability to realistically simulate the evolution of organized convective systems, including squall lines and multicell storms, while maintaining physical coherence in wind field retrievals.
Our results underscore the value of embedding physical guidance into generative deep learning architectures for convective nowcasting. The RainCast framework represents a meaningful step toward more reliable, interpretable, and physically consistent nowcasting of extreme precipitation, with potential applications in operational meteorology and disaster preparedness.
How to cite: Pan, X. and Zhao, K.: Physics-Guided Generative Nowcasting of Extreme Precipitation with Dual-Polarization Radar and Neural ODE-Diffusion Framework, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11714, https://doi.org/10.5194/egusphere-egu26-11714, 2026.
Thunderstorms pose significant risks to society and the economy due to hazards such as heavy precipitation, hail, and strong winds, which is why accurate forecasts are required to mitigate their impacts. Convection-permitting numerical weather prediction (NWP) models can explicitly resolve convective processes, but predicting thunderstorms from their output remains challenging since there is no obvious state variable that directly indicates thunderstorm occurrence. Instead, many approaches rely on combining multiple convective indices, such as convective available potential energy (CAPE), which are derived from state variables like temperature, pressure, and specific humidity, and act as surrogates for thunderstorms.
In this study, we present a deep neural network model that bypasses surrogate variables and instead directly processes the vertical profiles of state variables provided by convection-permitting forecasts. Our model, SALAMA 1D, analyzes ten different NWP output fields, such as wind velocity, temperature, and ice particle mixing ratios, across the vertical dimension, to produce the corresponding probability of thunderstorm occurrence. The model’s architecture is motivated by physics-based considerations and symmetry principles, combining sparse and dense layers to produce well-calibrated, pointwise probabilities of thunderstorm occurrence, while remaining lightweight.
We trained our model on two summers of forecast data from ICON-D2-EPS, a convection-permitting ensemble weather model for Central Europe operationally run by the German Meteorological Service (DWD), using the lightning detection network LINET as the ground truth for thunderstorm occurrences. Our results demonstrate that, up to lead times of (at least) 11 hours, SALAMA 1D outperforms a comparable machine learning model that relies solely on thunderstorm surrogate variables. Additionally, a sensitivity analysis using saliency maps indicates that the patterns learnt by our model are to a considerable extent physically interpretable. Finally, we show that spatial coverage can be extended to all of Europe by retraining on ICON-EU reanalysis data. Our work advances NWP-based thunderstorm forecasting by demonstrating the potential of deep learning to extract predictive information from high-dimensional NWP data—without sacrificing model interpretability.
How to cite: Vahid Yousefnia, K., Metzl, C., and Bölle, T.: SALAMA 1D: Deep-learning-based identification of thunderstorm occurrence in NWP forecasts without relying on convective indices, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13981, https://doi.org/10.5194/egusphere-egu26-13981, 2026.
Accurate real-time prediction of heavy precipitation is essential for disaster prevention. It remains a challenge for operational meteorology, especially for sudden localized convective storms for which traditional radar and observation extrapolation methods struggle to capture their rapid vertical development, which typically originate at altitudes of 4--8 km before descending to the surface in about 10 minutes.
In Japan, three Multi-Parameter Phased Array Weather Radars (MP-PAWR) generating 3D data every 30 seconds with high vertical resolution have been deployed. Leveraging these dense 4D observations, an AI-based model produces real-time nowcasts (very short-term forecasts) with high-resolution of 500 m and 10-minute lead time. Updated every 30 seconds, our nowcasts outperform traditional methods for predicting the onset and the dissipation of localized convective precipitation. However, performance is degraded during the mature phase of the storm when its structure becomes more complex (e.g., overlapping convective cells in different lifecycle states, domination of horizontal motion in radar pattern changes) (Baron et al., 2025a).
Two major improvements are currently being investigated: 1) a Quantile Regression Neural Network (QRNN) technique has been integrated to assess the probability distribution of possible nowcasts and thus provide credible intervals (Baron et al., 2025b), and 2) a better representation of 3D motion is being implemented, as it plays a critical role during the mature phase of storms. The new version of the model will integrate two separate modules: one specialized for capturing 3D-motion vectors, while the second predicts rainfall intensity with motion guidance. Both modules use the current nowcast model architecture which has demonstrated solid performance. The motion module is trained using 3D motion vectors derived directly from the radar observations through a 3D Tracking Radar Echoes by Correlation (TREC) method originally designed for PAWR extrapolation (Otsuka et al., 2016).
This study will present these developments with a special focus on the motion guidance module that is being implemented. The limitations of our approach will also be discussed (e.g., QRNN vs diffusion model, TREC limitation for weak gradient cases, no information on rain precursors and mesoscale scales).
Baron et al., 2025a: “Real-time nowcasting of sudden heavy rainfall using artificial neural network and multi-parameter phased array radar”, SOLA, https://doi.org/10.2151/sola.2025-039
Baron et al., 2025b: “3D Precipitation Nowcasting from Phased Array Radar with Uncertainty Estimation Using a Quantile Regression Neural Network”, IEEE RadarConf25, 10.1109/RadarConf2559087.2025.11204931
Otsuka et al., 2016: Precipitation nowcasting with three-dimensional space–time extrapolation of dense and frequent phased-array weather radar observations. Wea. Forecasting, 31, 329–340.
How to cite: Baron, P., Otsuka, S., Amell, A., Kawamura, S., Satoh, S., and Ushio, T.: AI nowcasting of localized heavy precipitation from fast-scanning radar with probabilistic and 3D motion guided prediction, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15508, https://doi.org/10.5194/egusphere-egu26-15508, 2026.
Satellite brightness temperature (BT) observations contain rich information about the horizontal distributions of cloud and rainfall structures; while radiosonde observations provide high-vertical-resolution measurements of temperature, moisture, and wind in the atmosphere. Beyond their traditional use in assimilation and retrieval, this study demonstrates innovative quantitative uses of BT and radiosonde observations for evaluating high-resolution numerical weather prediction (NWP) simulations of tropical cyclones (TCs) and Southwest Vortices (SWVs).
First, we apply BT observations to document the structural evolution of TCs and SWVs and to directly compare simulated hydrometeor distributions with satellite-observed cloud and precipitation features. These BT-based diagnostics provide objective constraints on model representation of convective initiation and development as well as the impact of diurnal variability.
Second, a BT-based threat-score (BT-TS) framework is introduced to assess the skill of rainfall forecasts with respect to satellite BT observations instead of rainfall observations traditionally used in TS evaluation. Using microwave humidity-sounder channels, the BT-TS metric performs well for assessing rainfall forecast in regions where precipitation observations are sparse or unavailable. The BT-TS forecast results highlight model deficiencies in timing, extent, and intensity of SWV-induced convective rainfall.
Third, radiosonde profiles are used to investigate lower-tropospheric processes critical for vortex evolution, focusing on planetary boundary layer (PBL) height and vertical variability under different vertical-resolution configurations. Verification with high-vertical-resolution (~5–6 m) profiles from 119 Chinese radiosonde stations during the summers of 2021–23 shows that accurately representing PBL height and lower-tropospheric thermodynamic variability requires approximately doubling the number of ERA5 vertical levels.
Together, these BT- and radiosonde-based diagnostics provide a comprehensive observational framework for evaluating the structural evolution of TCs and mesoscale SWVs. Future work will leverage these insights to refine cloud microphysics schemes, optimize model vertical-resolution design, and enhance the predictability of convection-permitting NWP systems.
How to cite: Zou, X.: Besides Assimilation and Retrieval: Innovative Quantitative Uses of Satellite Brightness Temperatures and Radiosonde Observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1397, https://doi.org/10.5194/egusphere-egu26-1397, 2026.
The development of various AI models in recent years has been very promising; the models’ ability to train from reanalysis datasets and evaluate on various metrics opened the door for a variety of new applications. However, the real stress test of any new model is its operational performance - applying predictions to data that weren't available during the model development and assessing the model’s capabilities for predicting real-world scenarios previously unseen.
In August 2025, we deployed our first high-resolution AI-based model for Iceland and it has been providing us with continuous predictions since then. Here we evaluate forecast skill against surface observations and benchmark against NWP models from the United Weather Centres (UWC) in Denmark and Iceland and our local operational NWP model for Iceland. We analyze the model’s strengths and weaknesses in predicting various weather events and discuss how these characteristics may influence the future model design.
How to cite: Stanisławska, K. and Rögnvaldsson, Ó.: AI model in the real world - analysis of the operational performance of a high-resolution AI weather model for Iceland, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18305, https://doi.org/10.5194/egusphere-egu26-18305, 2026.
Renewable energy sources have an increasingly pivotal role in global electricity generation, which poses challenges to the accurate and efficient meteorological forecasting (such as solar irradiance and hub-height wind speed). The development of AI large models has significantly shortened the time required for medium-range global weather forecast. However, their outputs typically lack high-temporal-resolution solar irradiance (e.g., provided only at 6-hour intervals or not at all), which cannot be directly applied to renewable energy forecasting.
In this work, we propose a machine learning framework to integrate the output variables from AI large models with high-resolution solar irradiance forecasting. Specifically, we train XGBoost models at 15 sites in eastern China using ERA5 reanalysis variables (2020–2023) as inputs and hourly surface solar irradiance derived from Himawari-8/9 satellite as targets. The trained models are evaluated on a 2024 test set driven by ERA5, achieving an annual mean hourly RMSE of 88.5 W m-2.
To assess the performance of this approach in medium range forecasting, we use hourly forecasts from the GDAS-driven Pangu Weather Model during January and July 2024 as inputs. Over 20 medium-range forecast tests, our approach (Pangu-ML) yields a day-ahead (24-h lead) RMSE of 62.5 (January) /95.4 (July) W m-2 and a 10-day lead RMSE of 92.3 (January) /110.1 (July) W m-2. For comparison, we conduct parallel simulations using the GFS-driven WRF v4.6 model at 9-km resolution over eastern China. The WRF-based irradiance forecasts produce day-ahead and 10-day RMSEs of 78.4 (January) /107.6 (July) W m-2 and 109.8 (January) /130.3 (July) W m-2 across the 15 sites, demonstrating that Pangu-ML achieves comparable or even superior accuracy.
In summary, our approach takes advantage of the computational efficiency of AI large meteorological models. It enables rapid generation of solar irradiance forecasts with minimal computational cost, thereby offering a practical pathway for subsequent operational ensemble irradiance forecasting.
How to cite: Yan, M., Zhang, M., Yang, K., Shu, Z., and Shao, C.: Bridging AI Large Meteorological Models and Solar Irradiance Forecasting Through Machine Learning Approaches, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10225, https://doi.org/10.5194/egusphere-egu26-10225, 2026.
Traditional operational weather prediction systems are driven by physics-based numerical simulations, which demand substantial computational resources. With the advancement of Artificial Intelligence (AI), modern transformer architectures have emerged as powerful alternatives, delivering high accuracy in data-driven weather forecasting. Despite this progress, transformers inherently operate on discrete representations and do not follow the underlying physical laws, thereby limiting their effectiveness in modelling the continuous spatio-temporal evolution of atmospheric processes. To mitigate this issue and inject physical structure, we introduce continuous-depth dynamics within the encoder and attention mechanism of a transformer. We propose the dual attention mechanism that jointly captures spatial and temporal dependencies. The spatial mode is modelled as a simple multi-head attention which is fused with the temporal component. The temporal attention operates on finite-difference derivatives of token embeddings across successive time steps, allowing the network to infer local temporal gradients and represent continuous evolution in feature space. Furthermore, we introduce continuous-depth Neural ODE layers in transformer encoder which models smooth transitions replacing the discrete residual updates. Finally, we propose a customized physics-informed loss function which is applied during training as a soft-constraint. This loss penalizes deviations from established thermodynamic and kinetic energy relationships governing temperature and wind evolution. By constraining the learned dynamics to respect these physical laws, the model produces forecasts that are not only data-accurate but also energetically consistent with the underlying principles of the atmospheric system.
How to cite: Saleem, H., Salim, F., and Purcell, C.: PINN-Cast: Exploring the Role of Continuous-Depth NODE in Transformers and Physics Informed Loss as Soft Physical Constraints in Weather Forecasting, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-78, https://doi.org/10.5194/egusphere-egu26-78, 2026.
Seasonal climate forecasts are critical for disaster management across the fragile Himalayan ecosystem, particularly during winter. However, these forecasts often exhibit strong spatial and temporal biases that reduce their reliability for predicting extremes at longer lead times. Traditional postprocessing methods such as quantile mapping and linear scaling assume stationarity and have limited ability to capture complex spatiotemporal error structures. To address these limitations, this study introduces Hima-Net (Himalayan-Net), a hybrid deep learning model that combines U-Net and Conv-LSTM architectures. Hima-Net is designed to improve the skill of sub-seasonal-to-seasonal (S2S) daily precipitation forecasts from the ECMWF S2S system by learning season-specific spatial and temporal patterns in forecast errors. The model is trained with a loss function that jointly emphasizes magnitude and correlation, enhancing its ability to represent the distribution and evolution of precipitation across lead times. Evaluation using metrics such as root mean square error (RMSE) and anomaly correlation coefficient (ACC) shows that Hima-Net consistently outperforms the raw forecasts across lead times over the Himalayan region. These findings demonstrate the potential of deep learning–based postprocessing to better capture and enhance spatial and temporal forecast patterns, offering a promising pathway for more accurate wintertime precipitation forecasts over the complex terrain of the Himalayas.
How to cite: Dar, J. and Ghosh, S.: Hima-Net: Deep Learning Enhancement of ECMWF S2S Winter Precipitation Forecasts over Northern India, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-526, https://doi.org/10.5194/egusphere-egu26-526, 2026.
Approximately 90% of the total precipitation in Senegal is produced by convective storms. The most intense rainfall events are associated with mesoscale convective systems (MCSs), frequently producing high-intensity rainfall that triggers pluvial flooding. Flood vulnerability is particularly high in the Greater Dakar area due to surface sealing and high population exposure. Timely and reliable short-term precipitation forecasts are therefore essential for effective early warning systems and flood risk reduction.
Precipitation nowcasting aims to describe the current atmospheric state and predict weather evolution at short lead times using real-time observations. The quality and availability of input data are key factors determining the nowcasting performance. In this study, three main data sources are employed: (i) in-situ observations from the High-resolution weather observations East of Dakar (DakE) station network, (ii) satellite-based products such as cloud-top temperature (CTT) from EUMETSAT and precipitation estimates from the Integrated Multi-satellitE Retrievals for GPM (IMERG) algorithm provided by NASA, and (iii) modeled data from the Weather Research and Forecasting (WRF) model.
The objective of this project is to identify a suitable nowcasting approach while weighing the strengths and limitations of the available data sources. Extrapolation-based methods, such as optical-flow techniques implemented in the pySTEPS library, estimate future precipitation by extrapolating observed patterns under the assumption of steady system evolution. These approaches perform well for large, long-lived convective systems, but they are unable to predict convective initiation, decay, and growth. Their applicability is further limited by the temporal resolution and detection uncertainties of the available satellite-based precipitation products identified in comparisons with station observations.
To address these limitations, a machine-learning-based nowcasting framework is developed, primarily relying on the high-temporal-resolution DakE station data to accurately capture atmospheric boundary conditions. Given the limited time span of data collection and the high predictor dimensionality, a Random Forest model was chosen as a robust approach. To mitigate challenges like zero inflation and the underestimation of extreme events, a two-step model architecture is developed: in a first step, a classification forest (I) is used to determine precipitation occurrence and the duration of the predicted event in the lead time horizon. If precipitation is expected, the model is coupled to a regression forest (II) that returns the rainfall intensity of the detected event. Future work will assess potential performance improvements from incorporating CTT-satellite and WRF-modeled data using feature importance analysis, which can also inform the placement of hypothetical new automatic weather stations.
How to cite: Berghoefer, M.-B., Haerter, J. O., and Monroy, D. L.: Random forest based precipitation nowcasting for Dakar , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18808, https://doi.org/10.5194/egusphere-egu26-18808, 2026.
Accurate short-term precipitation nowcasting is crucial for disaster risk reduction, flash-flood early warning, and water resource management. Conventional nowcasting approaches, such as extrapolation-based radar methods or numerical weather prediction models, often struggle to capture the nonlinear evolution of convective systems and are computationally demanding for rapid updates at high spatial and temporal resolution. The ability to provide reliable high-resolution forecasts at lead times of minutes to hours is particularly important for mitigating the societal and economic impacts of intense rainfall events. Recent developments in deep learning (DL), in combination with high-resolution radar observations, represent a compelling alternative for improving short-term precipitation forecasting. Radar-based precipitation data are particularly well suited for nowcasting applications due to their fine spatio-temporal resolution and ability to capture the dynamic structure and movement of precipitation systems. In this study, we develop and evaluate an operationally oriented DL framework for precipitation nowcasting that integrates multi-source data including high-resolution radar and satellite observations and automatic weather station measurements via the qPrec system over Slovakia. By incorporating satellite-derived forcing, the framework accounts for convection initiation and cloud development stage, providing a physical advantage over both classical extrapolation and radar-only deep learning methods. The framework leverages modern DL architectures, including convolutional encoder-decoder models such as U-Net and spatio-temporal transformer-based models (e.g., Earthformer), to learn the temporal evolution of precipitation fields inputs. The use of transformer-based models allows the network to capture long-range spatial dependencies and complex motion patterns that traditional CNNs may miss.
The proposed models generate precipitation forecasts at a spatial resolution of 1 km and a temporal resolution of 5 minutes, with lead times of up to 60 minutes. In addition to instantaneous precipitation estimates, the framework produces 15-minute accumulated precipitation for horizons up to 120 minutes. Unlike traditional methods where predictability skill remains static across resolutions, our DL approach leverages varied spatial representations to enhance predictability at these coarser temporal scales, optimizing the forecast for different hydrological requirements. These accumulated fields can be directly applied to flash-flood hazard assessment, enabling estimation of flood likelihood as a function of rainfall intensity and duration. Model performance is evaluated using standard verification metrics such as the Fractions Skill Score, and continuous ranked probability score (reducing to MAE on deterministic outputs), showing improvement over conventional radar extrapolation methods. This study demonstrates that modern DL approaches, particularly when combined with high-resolution radar observations, offer a promising path toward next-generation operational nowcasting.
How to cite: Almeida, R., Göttlich, J., Otero, N., Jurasek, M., Méri, L., Shenga, Z. D., Atencia, A., and Ma, J.: Deep Learning-Based Precipitation Nowcasting for Operational and Flash-Flood Applications, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8109, https://doi.org/10.5194/egusphere-egu26-8109, 2026.
Weather Generator (WGs) are tools, which produce synthetic weather series which are statistically similar to the weather series used to calibrate the WG. Though the underlying models of the WGs (frequently based on Markov chains and autoregressive models) include a prognostic component, so that the WGs could be hypothetically used to make a weather forecast, the precision of such forecast quickly converge (with increasing lead time) to zero. In our contribution, we do not use our generator for weather forecasting, but we use it to produce an ensemble of synthetic weather series which fit an available weather forecast.
One of the hot challenges in agrometeorology is a seasonal crop yield forecasting, which is a critical aspect of food production planning. The seasonal crop yield forecasting may be based on crop growth models run with daily time step. In this approach, the meteorological data fed into these models typically consist of observational weather data up to the forecast date, followed by weather forecast data (WF), mean climatic data, or weather generators (WGs).
In our contribution, we propose an improvement of the WG-based methodology. In contrast to approaches described in the literature, where WGs synthesize data independently of any WF, we are developing a methodology in which our single-site parametric M&Rfi WG (run with daily step) synthesizes multiple realisations of weather series which fit available WFs. Two approaches are proposed: (A) For use in operational crop yield forecasting, WG produces synthetic weather series starting with D0 day (which comes after the last day with weather observations and for which WF is available), so that the synthetic series smoothly follows available observations. In our experiments, (a) WF is defined for the upcoming days/weeks/months either in terms of the absolute values of individual weather variables or deviations from their climatological normals, (b) WF may optionally include information on its accuracy (e.g. in terms of standard errors or min-max intervals), (c) Precipitation forecast is assumed to be given in terms of amount and probability of precipitation occurrence, (d) WF may be defined separately for a set of time intervals (e.g. for next three days, next week, next months, etc.). The procedure for linking the generation process with WF is based on a continuous adjusting the stochastically generated series in a way resulting in a series that fits the WF while the internal structure (e.g. relations between variables) of the series remains realistic. (B) the “Research” approach: Unlike A approach, the B approach aims to answer the question: How the use of WF of given accuracy may contribute to the accuracy of seasonal forecast of the crop yields? The process of adjusting the stochastically generated series is similar to A method, but now, we care only about the dispersion of individual realisations, so that the magnitude of the dispersion corresponds to the known accuracy of the weather forecast.
Acknowledgements: The experiments were made within the frame of projects PERUN (supported by TACR, no SS0203004000), OP JAK (supported by MSMT, no. CZ.02.01.01/00/22_008/0004605) and AdAgriF (supported by MSMT, no. CZ.02.01.01/00/22_008/0004635).
How to cite: Dubrovsky, M., Trnka, M., Bartosova, L., Stepanek, P., Pohankova, E., and Balek, J.: Linking the Weather Generator with Weather Forecasts for Use in Forecasting Weather-Dependent Processes , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15394, https://doi.org/10.5194/egusphere-egu26-15394, 2026.
The global transition toward low-carbon energy systems has increased the reliance on renewable energy sources and driven solar power to become a key component of sustainable electricity generation, thereby increasing the importance of accurate irradiance forecasting. As solar penetration grows, power system operations increasingly depend on reliable short-term forecasts to support grid balancing, reserve allocation, and real-time decision-making. Global Horizontal Irradiance (GHI) represents the integrated influence of atmospheric conditions and cloud processes on surface solar radiation and governs short-term variability in photovoltaic power output. However, rapid cloud evolution introduces strong spatiotemporal variability in GHI, making accurate prediction at sub-hourly lead times a persistent challenge for short-term solar forecasting. In this study, we develop a real-time nowcasting system to predict GHI over the western region of India at 15-minute resolution with effective lead times of up to 2 hours. The system is based on a convolutional long short-term memory (ConvLSTM) model that learns spatiotemporal cloud–radiation relationships from high-frequency geostationary satellite observations. We utilize INSAT-3DR and INSAT-3DS products obtained from the MOSDAC archive, which provide continuous monitoring of cloud evolution over the region. The nowcasting framework is implemented using routinely available satellite observations and is evaluated over a large spatial domain covering western India, a region characterized by strong seasonal variability and diverse cloud regimes associated with pre-monsoon, monsoon, and post-monsoon periods. The results demonstrate consistent performance across seasons and show that the system captures the mean diurnal evolution of GHI with stable skill during daytime solar-active periods. Evaluation results indicate mean absolute errors of approximately 60 W m-2 for 1–2 hour lead times and 72 W m-2 for 2–3 hour lead times, corresponding to about 7–12 % of typical daytime GHI under moderate to high irradiance conditions. Overall, this work demonstrates the feasibility of satellite-driven deep learning systems for real-time GHI nowcasting and highlights the potential of integrating geostationary satellite observations and spatiotemporal learning models to support renewable energy forecasting and real-time grid decision-making in regions with high and growing solar power penetration.
How to cite: Garg, S., Ghosh, S., Murtugudde, R., and Banerjee, B.: Real-Time Solar Irradiance Nowcasting for Renewable Energy Forecasting over Western India, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18585, https://doi.org/10.5194/egusphere-egu26-18585, 2026.
Wind speed is a critical factor influencing the transport and dispersion of atmospheric pollutants in air quality models. However, numerical weather prediction (NWP) models, such as the weather research and forecasting (WRF) model, typically overestimate surface wind speeds, leading to inaccuracies in air quality predictions. To address this limitation, we developed an Artificial Intelligence (AI)-based Wind Field Correction (WFC) model aimed at improving PM2.5 forecasts over East Asia. The WFC model was constructed using the eXtreme Gradient Boosting (XGBoost) algorithm and trained on eight years of data, incorporating WRF-simulated meteorological variables as input features and in situ, ship-based, buoy, and radiosonde observations as targets. The WFC model effectively reduced the positive bias in WRF-simulated wind speeds, achieving a 90.15% reduction at the surface level and a 94.6% reduction from the surface to 850 hPa. The bias-corrected wind fields, when incorporated into the GIST Multiscale Air Quality model (GMAQ v1.0) developed by the Gwangju Institute of Science and Technology (GIST), resulted in substantial improvements in PM2.5 predictablity. In Central Eastern China (CEC), the wind field correction mitigated the underestimation of PM2.5 by suppressing excessive plume dilution in the model. In South Korea (SK), the correction slowed down accelerated plume advection, leading to a closer agreement between the simulated and observed PM2.5 plume locations. In addition, the correction enhanced the representation of daily PM2.5 variability and improved statistical metrics over the capital cities of Seoul and Beijing.
How to cite: Kim, J., Yu, J., Kim, H. S., Park, S., Woo, J.-H., and Song, C. H.: Bias-correction of wind speeds to improve PM2.5 predictability in chemical transport model, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2677, https://doi.org/10.5194/egusphere-egu26-2677, 2026.
Artificial intelligence (AI) models have demonstrated advancements in computational efficiency and forecast accuracy relative to the Numerical Weather Prediction (NWP), but they are unable to fully represent high-dimensional atmospheric dynamics. Thus, some AI-NWP coupled frameworks have been proposed, such as integrating AI-driven boundary conditions with numerical models to leverage the strengths of both approaches. However, in this coupled framework, ensemble forecasts and associated error propagation and energy dynamics remain under-explored. In this study, an AI-NWP coupled system that also uses the stochastic kinetic energy backscatter scheme (SKEBS) to generate ensemble forecasts is established. Ensemble simulations of Typhoon Yutu (2018) are carried out with the Weather Research and Forecasting (WRF) model employing Pangu-Weather and FuXi forecast data as boundary forcing. The results show that the ensemble WRF_Pangu (WRF_FuXi) improved Yutu’s track forecast by 67% (50%) compared to the traditional physics-based WRF_GFS (Global Forecast System), and reduced its intensity underestimation by about 67% relative to their AI global counterparts. Nonetheless, WRF_FuXi and WRF_Pangu exhibited limited ensemble spread and linear error growth, reflecting deterministic tendencies. Comparison of global and regional experiments show that Pangu-Weather is more physically constrained and thus better aligned with the WRF model for regional applications, while the adaptation of FuXi to the regional model is less robust. Spectral analysis revealed that AI-derived boundaries introduced excessive small-scale energy and underestimated larger-scale energy. The regional model WRF acted as a “conveyor belt”, propagating additive small-scale energy upscale, ultimately overwhelming the stochastic perturbations for ensemble generation. These findings underscore the need to incorporate more physical features into the AI-derived boundary conditions for ensemble forecasting.
How to cite: Ge, Y.: Ensemble Experiments in an AI-NWP Coupled Framework: A Typhoon Case, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10396, https://doi.org/10.5194/egusphere-egu26-10396, 2026.
While the effects of subgrid orographic drag on large-scale circulation have been extensively studied, its influence on typhoon precipitation remains less understood. Using the Weather and Research Forecasting model, this study investigates impacts of subgrid orographic drag components (gravity wave drag (GWD), flow-blocking drag (FBD), and turbulent orographic form drag (TOFD)) on landfalling typhoon precipitation and explores their resolution sensitivity through two representative cases: Super Typhoon Lekima (2019) and Severe Typhoon In-Fa (2021). Results reveal distinct distributions of GWD and TOFD over southeastern coastal China, which significantly modulate precipitation during strong landfalls like Lekima: GWD enhances precipitation in southern land areas affected by the typhoon while suppressing it in northern regions, whereas TOFD exerts precisely opposing effects. This is mainly due to enhanced (weakened) lower-tropospheric wind speed and water vapor transport caused by GWD (TOFD). GWD is highly sensitive to horizontal resolution, exhibiting more pronounced effects on the wind, moisture, and precipitation at coarser resolutions, while TOFD remains relatively invariant to horizontal resolution changes. Resolution of subgrid orography dataset driving these parameterizations is essential for accurately simulating drag distributions and impacts. Finally, typhoon intensity modulates these effects: stronger background circulation exacerbates the precipitation impacts of both GWD and TOFD.
How to cite: Zhang, M., Yan, M., Ma, Y., Yang, K., and Shu, Z.: Impacts of subgrid-scale orographic drag on landfalling typhoon precipitation, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11583, https://doi.org/10.5194/egusphere-egu26-11583, 2026.
Global Navigation Satellite System Radio Occultation (GNSS RO) observations are increasingly important for improving atmospheric profiling and numerical weather prediction (NWP), especially in cloudy, moisture-rich tropical environments where other satellite observations are often degraded. This study presents two complementary advances: (1) an improved regional quality-control strategy for preserving COSMIC-2 bending-angle data in cloudy regions, and (2) an assessment of the impact of assimilating Tianmu-1 RO observations from a newly deployed 23-satellite commercial constellation on the prediction of Typhoon Gaemi (2024).
First, we show that the widely used latitude-based quality control of COSMIC-2 bending-angle data leads to excessive removal of observations between 6–8 km near the Solomon Islands, where persistent summertime altostratus frequently reach above 6 km. Despite the long-wavelength nature of RO measurements—which makes them less sensitive to clouds—these regions were incorrectly flagged as outliers. By implementing a 2.5° × 2.5° local quality-control approach, the number of discarded observations in cloudy areas is substantially reduced, yielding a more spatially uniform deviation structure relative to the local mean. This regionally adaptive method better preserves high-quality RO data in both mid-tropospheric altostratus and lower-tropospheric Intertropical Convergence Zone environments.
Second, we evaluate the impact of assimilating over 30,000 daily RO profiles from the Tianmu-1 constellation using the GSI–WRF system. Assimilating Tianmu-1 data alone—without other satellite observations—reduces 120-hour track errors of Typhoon Gaemi by 20–40%, with the largest improvements beyond 48 hours. Diagnostics show that enhanced prediction skill arises mainly from improved inner-core temperature structure and better representation of the large-scale steering flow. Remarkably, the track forecasts with Tianmu-1 assimilation are even slightly better than the operational forecasts from the NCEP Global Forecast System (GFS).
Overall, these results highlight the increasing importance of high-density GNSS RO constellations in forecasting tropical cyclone intensity and track, and emphasize the value of cloud-aware, adaptive regional quality-control techniques in preserving cloud-affected observations. Future work will extend these adaptive quality-control strategies globally and examine synergistic assimilation of COSMIC-2, Tianmu-1, and other commercial RO datasets.
How to cite: Yang, S. and Zou, X.: Positive Impacts of Tianmu-1 RO Data Assimilation on Tropical Cyclone Forecasts and the Non-negligible Influence of Altostratus Clouds on RO Data Quality, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1396, https://doi.org/10.5194/egusphere-egu26-1396, 2026.
