Yang Chen

Accurate forecasting of geomagnetic perturbations is essential for assessing space weather risk of geomagnetic activities. Thanks to decades of magnetometer observations from world‐wide distributed networks and upstream solar wind observations at L1, data‐driven modeling of global geomagnetic perturbations has become increasingly feasible. Such models are promising for operational use because they typically run at lower computational cost and can offer higher predictive skill than first‐principles, physics‐based simulations when observations available for training are abundant. However, this potential has not been fully realized as sparse observational coverage in the Southern Hemisphere (SH) has long limited model generalizability. In this work, we use a zeroth‐order mirror‐symmetry assumption in ionospheric electrodynamics under the joint sign reversal of magnetic latitude , interplanetary magnetic field and dipole tilt to improve SH predictions using information from the densely observed Northern Hemisphere (NH). Rather than enforcing exact symmetry, we build on our previous GeoDGP model by introducing a symmetry‐aware kernel for a deep Gaussian process that learns the symmetry‐asymmetry balance from data. For comparison, we also experiment with an alternative approach that imposes symmetry as a hard constraint via data augmentation. We evaluate the models on 22 geomagnetic storms and present a case study of the 2024‐05‐10 Gannon extreme storm. The results show that both methods significantly improve the SH predictions, while the kernel‐based approach achieves the best performance and maintains NH accuracy comparable to GeoDGP.

Ground‐based Global Navigation Satellite System (GNSS) receivers provide primary total electron content (TEC) measurements used to monitor ionospheric variability, yet their sparse distribution over oceanic and equatorial regions limits global coverage and model accuracy. A far‐ultraviolet spectrograph aboard the Global‐scale Observations of the Limb and Disk (GOLD) mission provides continuous observations of the OI 135.6 nm emission radiance, which can be used to infer TEC across these undersampled regions. We present a framework for filtering and calibrating TEC from GOLD observations using the retrieval technique developed by Qin et al. (2023, https://doi.org/10.1029/2023ea003222 ) to produce an enhanced, bias‐corrected data set (gTEC) that can be combined with well‐validated ground‐based TEC sources, such as the Madrigal TEC. After filtering GOLD observations to remove solar and auroral contamination, we apply an adaptive quantile‐mapping calibration method that aligns the cumulative distribution functions of GOLD and Madrigal TEC. The resulting gTEC achieves a mean absolute error of ∼1–2 TECU relative to Madrigal, reproducing ∼53% of its natural variability within one standard deviation. Daily averaged spatial coverage improves by ∼11.62 ± 3.65% for frames with corresponding GOLD data, especially over the equatorial Atlantic, where existing databases lack sufficient observations. The integration of gTEC into the Video Imputation with SoftImpute, Temporal smoothing and Auxiliary data (VISTA) ionospheric TEC algorithm yields considerable improvement in the reproduction of the Equatorial Ionization Anomaly (EIA) structure and asymmetry, as well as mesoscale Equatorial Plasma Bubble (EPB) features. These results demonstrate that gTEC can substantially enhance global ionospheric specification and modeling and improve data assimilation capabilities.

Solar energetic particle (SEP) events pose severe threats to spacecraft, astronaut safety, and aviation operations. Accurate SEP forecasting remains a critical challenge in space weather research as a result of their complex origins and highly variable propagation. In this work, we built SEPNET , an innovative multi‐task neural network that jointly predicts future solar eruptive events, including solar flares and coronal mass ejections (CMEs) and SEPs, incorporating long short‐term memory and transformer architectures that capture contextual dependencies. SEPNET is a machine learning framework for SEP prediction that utilizes an extensive set of predictors, including the properties of solar flares, CMEs, and space‐weather HMI active region patches (SHARP) magnetic field parameters. SEPNET is rigorously evaluated on the SEPVAL SEP data set (Whitman, 2025b, https://doi.org/10.5281/zenodo.15555244 ), which is used to evaluate the performance of current SEP prediction models. The performance of SEPNET is compared with classical machine learning methods and current state‐of‐the‐art pre‐eruptive SEP prediction models. The results show that SEPNET , particularly with SHARP parameters, achieves higher detection rates and skill scores while maintaining the suitability for real‐time space weather alert operations. Although class imbalance in the data leads to relatively high false alarm rates, SEPNET consistently outperforms reference methods and provides timely SEP forecasts, highlighting the capability of deep multi‐task learning for next‐generation space weather prediction.

Solar phenomena such as flares, coronal mass ejections (CMEs), and solar energetic particles (SEPs) are actively monitored and assessed for space weather hazards. In recent years, machine learning has demonstrated considerable success in solar flare forecasting. Accurate SEP forecasting remains challenging in space weather monitoring due to the complexity of SEP event origins and propagation. We introduce SEPNET, an innovative multi-task neural network that integrates forecasting of solar flares and CME summary statistics into the SEP prediction model, leveraging their shared dependence on space-weather HMI active region patches (SHARP) magnetic field parameters. SEPNET incorporates long short-term memory and transformer architectures to capture contextual dependencies. The performance of SEPNET is evaluated on the state-of-the-art SEPVAL SEP dataset and compared with classical machine learning methods and current state-of-the-art pre-eruptive SEP prediction models. The results show that SEPNET achieves higher detection rates and skill scores while being suitable for real-time space weather alert operations.

Advancing space weather forecasting for human space exploration requires not only advanced scientific models, but also demonstration of their operational readiness, validation in realistic environments, and sustained feedback between research and operations (R2O2R). The CLEAR Space Weather Center of Excellence (CLEAR center) focuses on developing and transitioning advanced solar energetic particle (SEP) forecasting capabilities into operationally viable, real-time systems to support future missions.In this presentation, we will describe the research to operation activities conducted within the CLEAR center, with an emphasis on the past testbed-based exercise, operational co-development, and real-time implementation. The CLEAR center actively participated in the 2025 Space Weather Prediction Testbed Exercise in support of Human Space Exploration and the Artemis-II Mission, which provided a realistic operational context to assess model performance under constraints relevant to flight decision support, including latency, robustness, automation, interpretability, and uncertainty communication.We will also report on the deployment of the CLEAR center’s physics-based, empirical, and machine-learning SEP models into an automated, near-real-time forecasting framework, designed to operate continuously during mission-critical periods. Particular attention is given to operational architecture, including data acquisition, computational optimization, automation and fail-safe design, enabling timely delivery of prediction products for Artemis launch windows. Feedback from operators and forecasters has directly informed pipeline design, product placement, delivery timing, and visualization - closing the O2R loop.This work demonstrates how sustained engagement with operational partners accelerates the transition of SEP research into actionable forecasting capabilities. The CLEAR experience provides a concrete example of effective R2O2R pathways for next-generation space weather modeling in support of Moon and Mars exploration.

Climate warming is altering the timing of seasonal growth in cold-temperate and boreal ecosystems, with potential consequences for forests on multiple continents. It has been hypothesized that in a warming climate, earlier growth in spring might lead to a longer growing season and thus higher annual total growth. However, whether the period of active growth is longer in a warming climate is unclear, as is whether changes in growth rate might reinforce or offset the effect of changing growing season length in modulating annual total growth. Understanding responses in growth to warming is crucial to projecting forest dynamics under future climate change. To address this, we experimentally warmed (by +1.62 and 3.26 °C on average) open-air southern boreal forest plots for 14 years (2010 to 2023) and measured juvenile height growth from spring until fall for 20 tree and shrub species, with a total of 11,565 individual-year growth trajectories. Across all species and in various canopy (open, closed) and rainfall (ambient, reduced) contexts, experimental warming consistently advanced the timing of height growth (4.32 days per 3 °C warming). However, species differed in their annual total growth responses. Evergreen gymnosperms such as spruce, fir, and pine tended to exhibit suppressed annual total growth under warming (−15.8% per 3 °C warming), driven by a shorter duration of active growth, despite an earlier start, with little change in relative growth rate. In contrast, deciduous angiosperms and non-native shrubs often show enhanced annual total growth under warming (19.3 and 37.4% per 3 °C warming), driven by faster rate, longer duration, or a combination of both. More broadly, species with a more conservative lifestyle responded less positively in the duration and rate of growth, thus placing them at a disadvantage under warming. These contrasting responses challenge assumptions that an earlier start of the growing season universally enhances growth and highlight how warming may shift both the timing and pace of growth and thus alter community composition within mixed-species forests.

Geomagnetically induced currents (GICs) are a major space weather hazard during geomagnetic storms and pose ongoing threats to technologies using long conductor systems. Accurate forecasting of geomagnetic perturbations is essential for estimating GIC risk. Thanks to decades of magnetometer and solar wind observations from the SuperMAG collaboration and upstream space instruments at L1, data-driven modeling of global geomagnetic perturbations has become increasingly feasible. Such models are promising for operational use because they typically run at lower computational cost and can offer higher predictive skill than first-principles, physics-based simulations when observations are abundant. However, this potential has not been fully realized as sparse observational coverage in the Southern Hemisphere (SH) has long limited the generalizability of data-driven models. In this work, we leverage mirror symmetry in ionospheric electrodynamics to improve SH predictions using information from the densely observed Northern Hemisphere (NH). Building on our previous GeoDGP model, we introduce a novel kernel that encodes mirror symmetry into the prior of the Bayesian model, enabling data-driven learning of covariance between interhemispheric conjugate locations. We also experiment with an alternative approach that imposes symmetry as a hard constraint via data augmentation. We evaluate the models on 22 geomagnetic storms and present a case study of the 2024-05-10 Gannon extreme storm. The results show that both methods significantly improve SH predictions, while the kernel-based approach achieves the best performance and maintains NH accuracy comparable to GeoDGP.