Hongwei Gong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hongwei Gong has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Chao Xiao
College of Advanced Interdisciplinary Studies National University of Defense Technology Changsha China
2 shared publications - Chenghao Li
College of Meteorology and Oceanography National University of Defense Technology Changsha China
2 shared publications - Chuyue Zhou
College of Meteorology and Oceanography National University of Defense Technology Changsha China
2 shared publications - Die Duan
College of Advanced Interdisciplinary Studies National University of Defense Technology Changsha China
2 shared publications - Ganming Ren
College of Advanced Interdisciplinary Studies National University of Defense Technology Changsha China
2 shared publications - Han Wu
College of Advanced Interdisciplinary Studies National University of Defense Technology Changsha China
2 shared publications
Latest Publications
- 8 September 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Space Physics
- Vol. 131 (9)
Physics‐Motivated Middle‐ to Low‐Latitude NmF2 Reconstruction Using Thermospheric ΣO/N 2 and Solar EUV Radiation
Traditional deep learning models for the ionosphere often rely on space weather indices as inputs, offering predictive capability but limited physical interpretability. We present SRON2NN, a physics‐motivated artificial neural network (ANN) that incorporates two geophysically motivated features: effective solar radiation (ESR) and the thermospheric column density ratio of atomic oxygen to molecular nitrogen (ΣO/N 2 ). Using a comprehensive data set comprising COSMIC‐1 radio occultation data, TIMED/Global Ultraviolet Imager ΣO/N 2 observations, and solar EUV measurements from 2008 to 2019, we find that SRON2NN consistently outperforms an index‐based ANN using only Dst, F10.7, time, and location across all testing conditions. The improvement is most evident during geomagnetic storms in mid‐latitude regions, especially in the Southern Hemisphere, where the contribution of ΣO/N 2 leads to an 8.20% reduction in root mean square error. Compared with adding vertical total electron content and IRI‐NmF2, our physics‐motivated features yield a more pronounced improvement in reconstructing NmF2. Cross‐model experiments suggest that incorporating these two features can consistently boost performance not only in ANN frameworks but also in tree‐based models, including random forests and XGBoost. Further analysis of four geomagnetic storm events shows that SRON2NN is generally consistent with storm‐time ΣO/N 2 depletion and the associated decrease in NmF2. Overall, these results suggest that incorporating ESR and ΣO/N 2 as physically motivated inputs can improve NmF2 reconstruction while enhancing interpretability.
- 8 September 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Space Physics
- Vol. 131 (9)
Physics‐Motivated Middle‐ to Low‐Latitude NmF2 Reconstruction Using Thermospheric ΣO/N 2 and Solar EUV Radiation
Traditional deep learning models for the ionosphere often rely on space weather indices as inputs, offering predictive capability but limited physical interpretability. We present SRON2NN, a physics‐motivated artificial neural network (ANN) that incorporates two geophysically motivated features: effective solar radiation (ESR) and the thermospheric column density ratio of atomic oxygen to molecular nitrogen (ΣO/N 2 ). Using a comprehensive data set comprising COSMIC‐1 radio occultation data, TIMED/Global Ultraviolet Imager ΣO/N 2 observations, and solar EUV measurements from 2008 to 2019, we find that SRON2NN consistently outperforms an index‐based ANN using only Dst, F10.7, time, and location across all testing conditions. The improvement is most evident during geomagnetic storms in mid‐latitude regions, especially in the Southern Hemisphere, where the contribution of ΣO/N 2 leads to an 8.20% reduction in root mean square error. Compared with adding vertical total electron content and IRI‐NmF2, our physics‐motivated features yield a more pronounced improvement in reconstructing NmF2. Cross‐model experiments suggest that incorporating these two features can consistently boost performance not only in ANN frameworks but also in tree‐based models, including random forests and XGBoost. Further analysis of four geomagnetic storm events shows that SRON2NN is generally consistent with storm‐time ΣO/N 2 depletion and the associated decrease in NmF2. Overall, these results suggest that incorporating ESR and ΣO/N 2 as physically motivated inputs can improve NmF2 reconstruction while enhancing interpretability.
- Open Access
- 22 July 2026
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 53 (14)
Physics‐Embedded Deep Learning Reveals the Memory Effect of Ionospheric Electron Density in Response to Solar EUV Forcing
Precise characterization of the electron density profile is crucial for optimizing high frequency communications. Reconstruction models often assume an instantaneous ionospheric response to solar forcing or using time delays, neglecting the time‐integrated cumulative “memory effect” of plasma dynamics. Here, we develop the Electron Density Profile Reconstruction Neural Network (EDPRNN), a physics‐embedded framework that parameterizes and dynamically learns this memory representation from data. Using observations from three mid‐ and low‐latitude digisondes, we reveal that the learned memory exhibits distinct altitude and latitude dependence. Below the F1 layer, this data‐driven memory strongly correlates ( r ≈ 0.97) with theoretical photochemical time constants at all stations. EDPRNN also performs well on an independent solar maximum test set, with coefficient of determination ( R 2 ) values of 0.939, 0.905, and 0.841 at Mohe, Zuoling, and Fuke, respectively. These results indicate that EDPRNN provides an effective ionospheric nowcasting method and a novel data‐driven framework for identifying a physically consistent memory effect.
- 22 July 2026
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 53 (14)
Open AccessPhysics‐Embedded Deep Learning Reveals the Memory Effect of Ionospheric Electron Density in Response to Solar EUV Forcing
Precise characterization of the electron density profile is crucial for optimizing high frequency communications. Reconstruction models often assume an instantaneous ionospheric response to solar forcing or using time delays, neglecting the time‐integrated cumulative “memory effect” of plasma dynamics. Here, we develop the Electron Density Profile Reconstruction Neural Network (EDPRNN), a physics‐embedded framework that parameterizes and dynamically learns this memory representation from data. Using observations from three mid‐ and low‐latitude digisondes, we reveal that the learned memory exhibits distinct altitude and latitude dependence. Below the F1 layer, this data‐driven memory strongly correlates ( r ≈ 0.97) with theoretical photochemical time constants at all stations. EDPRNN also performs well on an independent solar maximum test set, with coefficient of determination ( R 2 ) values of 0.939, 0.905, and 0.841 at Mohe, Zuoling, and Fuke, respectively. These results indicate that EDPRNN provides an effective ionospheric nowcasting method and a novel data‐driven framework for identifying a physically consistent memory effect.