Abstract The polarity inversion line (PIL) in active regions (ARs) is considered to be closely associated with solar flare eruptions. In this study, we rigorously constructed standardized data sets based on time series of different lengths using Space‐weather HMI Active Region Patches (SHARP) parameters calculated along the PIL. We compared the performance of traditional non‐sequential models and a time‐series model in solar flare prediction tasks, as well as the predictive performance of time‐series models with different input lengths within the CNN–BiLSTM–AT framework. The main findings of this study are summarized as follows: (a) SHARP parameters computed along the PIL consistently yield better prediction performance than those calculated over entire active regions. (b) In realistic and highly imbalanced prediction scenarios, the time‐series model outperforms non‐sequential models, achieving an F1 score of 0.59 for strong‐flare prediction. (c) Robustness tests and sliding‐window probability forecasts further demonstrate the practical feasibility of the proposed approach. These results provide useful guidance for data representation and model selection in solar flare forecasting.

China has established a ground-based network system, that is the Chinese Meridian Project (CMP), to continuously monitor the geomagnetic field. The superstorm in May 2024 was analyzed using the CMP data. The negative peak of the horizontal geomagnetic field (B h) during the storm main phase at different CMP stations varied between -449 nT and -720 nT while the absolute peak value of the corresponding time derivative (dB h/dt) varied from 32 nT/min to 60 nT/min. We adopted a geomagnetically induced current (GIC) model to investigate the impact of the storm on the Eastern Inner Mongolia Power Grid. The comparison between modeled and measured geomagnetically induced currents (GICs) at two substations showed a good agreement in timing and amplitude. The simulated GICs across the Eastern Inner Mongolia Power Grid service area identified the Tianjin South substation at 1,000 kV level as the most vulnerable part of the grid, with maximum GIC of 168.5 A. The GIC level, as an indicator, supports the conclusion that the effect of the May 2024 storm exceeds that of the November 2004 storm and is on a par with the March 1989 geomagnetic storm. Together with real-time GIC monitoring, our approach has paved the way for future geomagnetic storm risk assessment and extreme space weather safeguard.

Solar active regions (ARs) are the source of solar eruptive events such as flares. The morphology of sunspots within ARs is closely related to solar eruptive activity. The complexity of an AR serves as a critical reference for forecasting various solar eruptive events. The high temporal resolution of current solar observations has led to the rapid accumulation of solar activity data, making accurate and objective automatic identification and classification of sunspot groups in ARs highly important. This paper combines a vision transformer (ViT) with a convolutional neural network (CNN) and introduces the magnetic physical parameter R-value to construct a recognition model named ViT-CNN-R for Mount Wilson magnetic classification. Test results show that the model's classification accuracy for Alpha, Beta, and Beta-x type ARs are 0.9282, 0.8479, and 0.9162, respectively, with true skill statistic scores of 0.8464, 0.6490, and 0.7996, respectively. Comparisons with models from other studies and testing the model's generalization performance using Advanced Space-based Solar Observatory data reveal that the ViT-CNN-R model exhibits high classification performance for complex ARs of type Beta-x. This model can provide accurate magnetic classification information for ARs in subsequent forecasting research.

Coronal mass ejections (CMEs) are the key drivers of both nonrecurrent geomagnetic storms and gradual solar energetic particle events. A near real-time capability of automatic CME detection from coronagraph images is crucial for operational space weather forecasts, particularly solar proton events (SPEs). Utilizing the You Only Look Once (YOLO) algorithm against the C3 coronagraph images taken by the Large Angle and Spectrometric Coronagraph Experiment (LASCO) on board the Solar and Heliospheric Observatory, we have developed a machine learning framework to classify and segment CMEs that tend to cause SPEs. Two different types of YOLO models are implemented. One is used to detect and classify CMEs and the other is for the segmentation purpose. Subsequently, CME characteristic parameters can be computed based on the segmentation results. Statistical comparison to the extensively used CDAW CME catalog proves that the model performs better than both CACTus and SEEDS in determining CME properties. Our approach is novel for the following reasons: (1) conceptually, an evolving CME is regarded as different classes at different times, which improves the classification accuracy; (2) the model directly adopts a single pseudocolor image rather than the running-difference image as the input; and (3) this work is based on a large number (53,082) of LASCO C3 images spanning 1997-2007, 2010-2013, and 2015-2022.

Solar flares are explosive releases of magnetic energy in localized regions of the solar atmosphere that threaten the safety of solar-terrestrial technological systems, making their forecasting critical for space-weather-forecasting support services. However, transitioning flare-forecasting models from research to operational use faces two primary challenges. First, those models relying on active region (AR) magnetograms or AR parameters are subject to projection effects, limiting their reliability to ARs near the solar disk center rather than full-disk ARs. Second, most models utilize only a single data type (images or parameters), failing to leverage complementary information, which restricts their predictive performance. To address these issues, we develop an operational ensemble forecasting model for full-disk ARs that integrates AR magnetograms with traditional solar activity parameters less affected by projection effects. The model comprises three modules: a parameter feature extractor, an image feature extractor, and a feature fusion module. Trained on multimodal data, the model predicts the probability of M-class or above flares occurring within the next 48 hr. On the test set, it achieves an F1 score of 0.4347 and a ROCA of 0.8074. Comparative evaluations against models using only parameters or images demonstrate that the ensemble approach achieves improved and more balanced performance within the same dataset. This ensemble model provides a practical reference for transitioning to operational full-disk AR forecasting using near-real-time, easily accessible solar observations.

China's Space Station (CSS) is equipped with a high-performance BDS/GPS receiver that enables centimeter-level Precise Orbit Determination (POD) and sub-nanosecond time synchronization to support scientific research tasks. This study investigates the scientific and technological applications of the orbital and empirical parameters estimated from the reduced-dynamic Precise Orbit Determination (RPOD) process. Analysis of in-flight data demonstrates the following practical applications: 1) The dynamic and empirical acceleration parameters allow for an independent centimeter-level determination of the CSS's in-flight center-of-mass (CoM), while accurately quantifying the impacts of key events such as spacecraft docking/separation and extravehicular activities. 2) The aerodynamic force modeled using a detailed geometric macro-model can assess the operational status of the onboard external atmospheric detector. 3) The estimated scale factor of the modeled aerodynamic drag effectively reflects the CSS's frontal cross-sectional area, exhibiting a strong positive correlation with the daily reference area provided by the flight control team. 4) The GNSS receiver can run with external time source, thereby, GNSS-based timing with 0.1 ns precision enables independent evaluation of the external clock's frequency stability. 5) Leveraging rapid data downlink via China's Tianlian geosynchronous relay satellite system and near-real-time (NRT) GNSS products, NRT-POD for CSS is achieved with an accuracy of 15 cm and the subsequent orbit prediction errors are 37m@3h and 565m@12h during high solar activity period. Overall, this study demonstrates the feasibility of using spaceborne GNSS receiver as a multi-functional 'sensor' to monitor and predict the status of large space stations. These capabilities can further support space scientific activities, collision avoidance maneuvers and flight plan coordination. (c) 2026 The Author(s). Published by Elsevier B.V. on behalf of COSPAR. This is an open access article under the CC BY license (http:// creativecommons.org/licenses/by/4.0/).

This study examines the spatiotemporal evolution of midlatitude ionospheric disturbances during the intense geomagnetic storm on 10–11 October 2024, focusing on the North American and European sectors. It utilizes multi-instrument datasets from ground-based observations, including Global Navigation Satellite System (GNSS) receivers and ionosondes, supplemented by the measurements from the Swarm, DMSP and GUVI/TIMED satellites. The results reveal significant longitudinal and latitudinal variations in regional ionospheric responses, specifically related to Storm Enhanced Density (SED) and the midlatitude trough. Key findings include: (a) During the main phase of the storm, the North American midlatitude ionosphere exhibited a pronounced longitudinal contrast: a positive SED-driven phase in the west versus a negative trough-dominated phase in the east. In the early recovery phase, the western sector transitioned to a trough-induced negative phase, while the eastern sector showed a positive phase related to auroral particle precipitation during substorms. (b) The North American SED featured a strong northwest-extending plume with a westward shift velocity of 200–300 m/s at 45°N, and a sharp density gradient of 60–65 TECU on its northeastern side, in contrast to the trough. (c) The European sector displayed a “sandwich-like” latitudinal pattern, with “positive–negative–positive” variations during the storm. (d) The European sector’s storm-time trough expanded rapidly equatorward, reaching a minimum of ~35° magnetic latitude (MLAT), while broadening latitudinally to a width of 18–20°. These density gradient structures, along with the longitudinal/latitudinal differences, highlight the dynamic processes occurring in the magnetosphere–ionosphere–thermosphere system during intense storms and contribute to the understanding of storm-response mechanisms across different sectors.

We present a Solar Situation Awareness and Flare Forecasting Dataset designed to support research and operational applications in space weather events monitoring and prediction. The core dataset spans a 14-year period from May 2010 to May 2024 and integrates multi-source solar observations from the Helioseismic and Magnetic Imager and Atmospheric Imaging Assembly onboard the Solar Dynamics Observatory together with ground-based Hα observations. During this period, the dataset provides a complete set of observations including line-of-sight magnetograms, continuum images, multi-band EUV images, and Hα observations. The dataset includes manually annotated segmentation masks for major solar structures, including active regions, coronal holes, and filaments. To extend the historical coverage of solar magnetic and coronal observations, earlier data from the Solar and Heliospheric Observatory are also incorporated, including MDI magnetograms from 1996 to 2010 and EIT extreme ultraviolet images at 195 and 304 Å. All observations are spatially standardized to a unified solar-disk coordinate system extending to 1.1 solar radii and resampled to a consistent spatial scale. The dataset further includes flare labels within a 24-hour prediction window, operational magnetic parameters derived from magnetograms, and a manual mapping between SHARP patches and NOAA active region numbers. This structured multimodal dataset provides a standardized resource for solar structure analysis, machine-learning research, and operational flare forecasting.

The Global Positioning System (GPS) satellite constellation has provided long-term observations of energetic electrons in Earth’s radiation belts, but inconsistencies among different satellites limit the direct use of their combined measurements. A comparative analysis of electron flux data from year 2000 to 2020 reveals significant deviations for several satellites, particularly NS41 and NS48 in low-flux regions, as well as scattered observations from satellites such as NS74 and NS69. These discrepancies highlight the necessity of cross-calibration to ensure data consistency. To address this, we conducted the first systematic cross-calibration of energetic electron fluxes from 25 GPS satellites. Taking the 2.0 MeV average unidirectional differential electron fluxes and the ≥ 2.0 MeV omnidirectional integral electron fluxes as examples, we adopted the conjunction method in magnetic coordinates (Lm, B/B_0 ) (Lm: McIlwain L-shell parameter, B/B_0 : magnetic field ratio) and applied cubic polynomial fitting to achieve unified calibration using NS59 as the reference. For the 2.0 MeV differential electron fluxes, the Root-Mean-Square Deviation (RMSD) before calibration is on average 3.08 times larger than that after calibration, while the Correlation Coefficient (CC) increases by a factor of 1.14 on average. For the ≥ 2.0 MeV integral electron fluxes, the corresponding values are 1.68 and 1.01, respectively. This method can be extended to other energy channels and satellites. The calibrated dataset facilitates quantitative analysis and modeling of variations in the high-energy electron distribution in medium Earth orbits.

Abstract We investigate the response of thermospheric neutral density and associated Joule heating (JH) with different interplanetary structures during 73 intense geomagnetic storms. Neutral density is derived from accelerometers onboard CHAMP, GRACE‐A, SWARM‐C, and GRACE‐FO satellites during 2001–2024. We estimate the JH variation using the data from the DMSP spacecraft and the Weimer electric potential model. Four typical interplanetary structures identified as the primary drivers of intense storms (Dst ≤ −100 nT) are analyzed in terms of their influences on neutral density and JH: isolated interplanetary coronal mass ejections (ICMEs) (29% storms), successive ICMEs (44% storms), interacting ICME‐HSS structures (26% storms), and high‐speed streams (HSSs) alone (1% storm). The middle two categories represent the complex interplanetary structures. The results show that neutral density and associated JH were significantly enhanced during the May 2024 superstorm driven by successive ICMEs. The complex interplanetary structures, including successive ICMEs and interacting ICME‐HSS structures, lead to significant and prolonged neutral density enhancements during intense storms, in contrast to isolated ICMEs. During successive ICME‐driven storms, enhanced interplanetary magnetic field and solar wind (SW) are associated with increased JH, which has a strong impact on storm‐time neutral density. The interaction between ICMEs and HSSs causes the same distribution in density intensity as that during isolated ICME‐induced intense storms but substantially extends the duration of density enhancement, plausibly resulting from enhanced SW due to interacting ICME‐HSS structures.

Continuous forecasting of the SYM-H index from real-time solar wind is a practical problem in space weather monitoring, especially during intense geomagnetic storms when the target evolves rapidly on sub-hourly time scales. We present a state–driver conditional augmented neural ordinary differential equation (ANODE) for continuous SYM-H forecasting at 5-minute cadence with horizons up to 3 hours. The model separates recent SYM-H history and solar wind drivers into a latent system state and a driver context, and couples them through conditional continuous-time latent dynamics. With about 0.1M trainable parameters, the proposed model provides a compact and structured forecasting framework. On four storm intervals with SYM-H below −200 nT, it improves the 3-hour average RMSE from 23.52 to 20.70 nT and the average PCC from 0.926 to 0.938 relative to an LSTM baseline. For the October 2024 superstorm, the 3-hour peak intensity error is reduced from 96.63 to 36.59 nT. These results show that a lightweight continuous-time model can provide effective high-resolution SYM-H forecasting from real-time solar wind under strong-storm conditions.

This study investigates midlatitude ionospheric and thermospheric variations during the 16 April 2025 geomagnetic storm through multi-instrument analysis and numerical simulation. The analysis uses a comprehensive data set comprising ground-based Global Navigation Satellite System Total Electron Content (GNSS TEC) and ionosonde measurements, along with space-based observations from the Swarm, Defense Meteorological Satellite Program (DMSP), Global-scale Observations of the Limb and Disk (GOLD), and TIMED Global Ultraviolet Imager (GUVI), as well as the Thermosphere-Ionosphere-Electrodynamics General Circulation Model (TIEGCM). We found distinct midlatitude density gradient structures in both the ionosphere and thermosphere during the storm's main phase as follows: (a) An unusual storm-enhanced density (SED) was observed originating in the dawn sector, which persisted over North American longitudes for 8-9 hr. In contrast, the local afternoon and dusk sectors over Europe showed no SED signature but instead exhibited a negative ionospheric storm phase. (b) Coincident with the morning SED, we observed a distinct tongue-like structure in neutral composition with O/ enhancement of 20%-40% in the North American sector, which was bounded by areas of O/ depletion. This formed a coherent thermosphere-ionosphere structure with concurrent enhancement and depletion. (c) The dawnside SED can be largely attributed to two synergistic mechanisms: a chemically driven increase in plasma density from an Oxygen-rich neutral tongue, coupled with an electrodynamically driven zonal electric field increased by the sunrise enhancement of the penetration electric field. These results underscore the critical role of neutral composition and electrodynamic forcing in the formation of storm-time midlatitude ionospheric gradient structures, offering new insights into the dynamics of the coupled ionosphere-thermosphere system.

Abstract Coronal mass ejections (CMEs) are among the key solar eruptive activities, triggering space weather disturbances. Thus, forecasting their geoeffectiveness has become a research focus. This study constructs a model to recommend similar events for forecasting the geoeffectiveness of CMEs. The input parameters are optimized via feature dimensionality reduction, while the cosine similarity algorithm, combined with logistic regression, is employed to perform four tasks: binary classification of whether a CME will reach Earth, prediction of travel time, assessment of the intensity of geomagnetic disturbances, and matching of historical similar events. Experimental results show that the model achieves an F1‐score of 0.43 for binary classification, with a mean absolute error (MAE) of 13.75 hr for travel time prediction and a MAE of 1.68 for the maximum Kp value (Kp max) on the test set. The findings indicate that the model not only achieves staged multi‐task forecasting for the geoeffectiveness of CMEs but also enhances the interpretability of forecasting results through similar event recommendation, providing pragmatic decision support for space weather warnings.

With the increasing number of large constellations, it is crucial to accurately predict satellite positions and movements using upper atmosphere models driven by geomagnetic indices. Machine learning (ML) can quickly provide geomagnetic index predictions. However, previous research using ML to forecast Stream Interaction Region-driven (SIR-driven) geomagnetic storms primarily follows two approaches: using real-time observations at the L1 point with historical geomagnetic indices, which lacks solar source information, or using solar images as input. To forecast SIR-driven storms over longer horizons (3 days), a neural network model is constructed by incorporating coronal magnetic field physical parameters and the coronal hole features extracted from solar images (Pch ${P}_{\text{ch}}$). We first select the appropriate combination of coronal magnetic parameters for forecasting geomagnetic indices. Using the Explainable AI technique, we find that the model can learn the propagation characteristics of interactions between high- and low-speed solar wind from historical coronal parameters, which relate to the physical process of SIR formation, providing a physical basis for the approach. We then introduce Pch ${P}_{\text{ch}}$ and the historical Kp index to develop an optimal model. Based on accuracy evaluations across different activity levels and solar cycle phases, the model has proven to provide reasonable forecast results for SIR-driven events over the next 2-3 days. The model outperforms an empirical model and two ML models in terms of event detection ability for horizons longer than one day. This study offers a new approach for 3-day geomagnetic disturbance forecasting using a deep neural network.

Accurate prediction of ionospheric scintillation is essential for ensuring the reliability of spaceborne and ground‐based radio wave technology infrastructures, including but not limited to navigation and communication systems. In this study, we propose a deep learning‐based Ionospheric Scintillation Network (ISNet), which can predict the regional scintillation index S 4 1 hour in advance. The novel ISNet decomposes the S 4 index into background and disturbance fields and treats them separately. The model also employs specialized modules to capture the time‐delay effect between external disturbances and the associated scintillation. A flexible dynamic data reconstruction strategy is adopted, which allows the model to learn directly from high‐fidelity scintillation observations. Our results show that ISNet can provide accurate regional ionospheric scintillation forecasts in the low‐latitude regions of China. The RMSEs for weak, moderate, and strong scintillations in the test set are 0.053, 0.124, and 0.183, respectively. Compared with 12 other methods, ISNet displays a significant advantage in predicting moderate‐to‐strong scintillations. We also quantified the relative importance of input features to model predictions through the interpretable technique, which may provide valuable insights into model interpretation and feature selection.

Using new observations from the Chinese Meridian Project (CMP), this study examines the characteristics of neutral winds in the East Asian sector during the Mother's Day super‐intense storm in May 2024, primarily focusing on its effects on the disturbances over China and adjacent areas. It is the first time that new measurements from three Dual‐Channel Optical Interferometers (DCOIs) are utilized to analyze storm‐time neutral winds at an altitude of 250 km in northern China. By developing all northern‐hemisphere Super Dual Auroral Radar Network radars including six Chinese Dual Auroral Radar Network (CN‐DARN) radars, the newly derived ionospheric convection pattern and its impacts on neutral winds can be well analyzed. The results show that a strong equatorward wind with a maximum amplitude of ∼400 m/s in the meridional component was observed for the first time during the storm main phase. In the East Asian sector, a negative ionospheric storm over China and adjacent areas was accompanied by this enhancement in equatorward wind in the night of May 10. Additionally, ionospheric convection expanded to 43° magnetic latitude (MLAT) with eastward ion velocities exceeding 800 m/s around 50° MLAT. This can strengthen zonal wind in northern China, producing a notable eastward surge of ∼230 m/s measured by new DCOIs in the dawnside sub‐auroral region. Wave‐like oscillations in neutral winds were observed by multi‐DCOI stations, which were associated with the storm‐time Traveling Atmospheric Disturbances (TADs). During the storm recovery phase, a high‐level total electron content cluster shifted from eastern China to the central regions, which may be attributed to enhanced ∑[O]/[N 2 ].

This paper investigates the characteristics and variability of equatorial plasma bubbles (EPBs) over the American-Atlantic longitude sector during an intense geomagnetic storm on 10-11 October 2024. The study utilizes multi-instrument data sets from ground-based observations (Global Navigation Satellite System receivers and a Fabry-Perot Interferometer) and satellite measurements (Global-scale Observations of Limb and Disk, Swarm, and DMSP), as well as an equatorial electric field model. The observed EPBs exhibited a significant east-west longitudinal variability with contrasting behaviors in the American and Atlantic sectors, with the main findings summarized as follows: (a) In the American longitudes west of 60 degrees ${}<^>{\circ}$W, super EPBs were observed during the main phase after 23:40 UT on October 10, and EPB-induced plasma depletion extended to the midlatitude ionosphere around +/- 35 degrees $\pm 35{}<^>{\circ}$-40 degrees ${}<^>{\circ}$ geomagnetic latitude. (b) These super EPBs persisted for 12-15 hr throughout the entire local night until sunrise, which exhibited an anomalous westward drift feature due to the penetration of subauroral polarization stream electric field plus induced vertical Hall electric field. (c) In contrast, over the Atlantic longitudes east of 60 degrees ${}<^>{\circ}$W, the EPBs' activity was largely suppressed after 23:00 UT on October 10, exhibiting substantial differences from that of nearby western longitudes and quiet-time conditions. Such significant longitudinal variability of EPBs can be attributed to the electrodynamic driver of storm-time perturbations in the equatorial electric field, which is highly dependent on local-time sectors with sharp contrast in the dusk and premidnight sectors.

Solar activity drives space weather, affecting Earth's magnetosphere and technological infrastructure, which makes accurate solar flare forecasting critical. Current space weather models under-utilize multi-modal solar data, lack iterative enhancement via expert knowledge, and rely heavily on human forecasters under the Observation-Orientation-Decision-Action (OODA) paradigm. Here we present the "Solar Activity AI Forecaster", a scalable dual data-model driven framework built on foundational models, integrating expert knowledge to autonomously replicate human forecasting tasks with quantifiable outputs. It is implemented in the OODA paradigm and comprises three modules: a Situational Perception Module that generates daily solar situation awareness maps by integrating multi-modal observations; In-Depth Analysis Tools that characterize key solar features (active regions, coronal holes, filaments); and a Flare Prediction Module that forecasts strong flares for the full solar disk and active regions. Executed within a few minutes, the model outperforms or matches human forecasters in generalization across multi-source data, forecast accuracy, and operational efficiency. This work establishes a new paradigm for AI-based space weather forecasting, demonstrating AI's potential to enhance forecast accuracy and efficiency, and paving the way for autonomous operational forecasting systems.

The rapid development of low‐Earth orbit (LEO) satellites brings increased attention to spacecraft collisions, space debris, orbital decay, and satellite reentry. Neutral density and associated drag force on the satellite orbits elevate space risks, significantly determined by space weather disturbances, particularly geomagnetic storms. On 3 September 2024, the Australian Binar‐2, 3, and 4 satellites were deployed, but their actual lifetimes were only around 2 months, much shorter than designed, only 20% that of the nearly identical Binar 1 satellite launched in 2022. For the first time, we analyze the premature reentries of the Binar‐2, 3, and 4 satellites to unveil the severe space weather impact on their lifetimes, especially the influence of accuracy of medium‐ and long‐term space weather prediction on satellite lifetime designs. Our findings reveal that the premature reentries of Binar‐2, 3, and 4 satellites were caused by enhanced neutral density due to much higher solar and geomagnetic activities than predicted. The actual satellite lifetimes align well with estimations based on observed space parameters, whereas large deviations occur when using predicted parameters. The widely used predictions for medium‐ and long‐term space weather underestimated the F10.7 index and sunspot numbers during Solar Cycle 25 (SC25), especially during the peak time, leading to discrepancies in the designed satellite lifetimes. Our results illustrate the importance of medium‐to long‐term space weather forecasting for the lifespan of LEO satellites.