Current ionospheric tomography faces a bottleneck: heavy reliance on empirical background fields leads to poor spatiotemporal accuracy and robustness. To resolve this, we propose an inversion strategy fusing interpretability analysis, Spatio-Temporal Implicit Neural Representation (ST-INR), and the Multiplicative Algebraic Reconstruction Technique (MART). By training a physics-informed ST-INR model on multi-LEO satellite occultation data, we overcome the spectral bias of traditional networks to generate high-fidelity prior fields with fine physical gradients. These priors serve as initializations for MART-based refinement using ground-based GNSS Slant Total Electron Content (STEC), ensuring alignment with real-time observations. Validation during an independent test period featuring a severe geomagnetic storm (Dst approximately -150 nT) reveals significant improvements. Compared to the IRI2020 model, our method reduces the 3D electron density reconstruction RMSE by 71.9% (quiet) and 61.2% (storm), while improving key physical parameters (NmF2 and hmF2) by 45.0% and 40.5%, respectively. It also significantly outperforms the traditional Back-Propagation (BP) neural network. Comparisons with Global Ionospheric Map (GIM) products from the International GNSS Service (IGS) further verify high horizontal consistency, particularly in regions lacking station coverage. This study confirms that the proposed method effectively mitigates reconstruction challenges in data-sparse areas, significantly enhancing physical realism under complex space weather conditions.

The atmospheric gravity waves (AGWs) generated by severe typhoons can facilitate the transfer of energy from the troposphere to the ionosphere, resulting in medium-scale traveling ionospheric disturbances (MSTIDs). However, the complex three-dimensional nature of MSTIDs over oceanic regions presents challenges for detection using ground-based Global Navigation Satellite System (GNSS) networks. This study employs a hybrid approach combining space-based and ground-based techniques to investigate the spatiotemporal characteristics of ionospheric perturbations during Typhoon Doksuri. Plane maps depict significant plasma fluctuations extending outward from the typhoon’s gale wind zone on 24 July, reaching distances of up to 1800 km from the typhoon’s center, while space weather conditions remained relatively calm. These ionospheric perturbations propagated at velocities between 173 m/s and 337 m/s, consistent with AGW features and associated propagation speeds. Vertical mapping reveals that energy originating from Typhoon Doksuri propagated upward through a 500 km layer, resulting in substantial enhancements of plasma density and temperature in the topside ionosphere. Notably, the topside horizontal density gradient was 1.5 to 2 times greater than that observed in the bottom-side ionosphere. Both modeling and observational data convincingly demonstrate that the weak background winds favored the generation of AGWs associated with Typhoon Doksuri, influencing the development of distinct MSTIDs.

The 60°W International Meridian Circle Project (IMCP) sector is a critical region for global ionospheric space weather monitoring, spanning multiple critical physical zones and exhibiting intense spatiotemporal variability. Accordingly, this paper proposes a hybrid CNN-Transformer deep learning framework integrated with a cross-attention mechanism for Total Electron Content (TEC) forecasting at stations within this region. The model combines a multi-scale CNN and a Transformer to extract local morphological features and capture global long-range dependencies, respectively; on this basis, a cross-attention layer is introduced to achieve adaptive bidirectional fusion between the two, incorporating geomagnetic and solar activity indices as external drivers. The results demonstrate that during geomagnetically quiet periods, the model achieves a Coefficient of Determination (R2) of up to 0.99 and a Root Mean Square Error (RMSE) as low as 0.95 TECu. For multi-level geomagnetic storm events, the model's cross-attention mechanism can dynamically allocate weights to external drivers, enabling the delay-free tracking of extreme evolutionary processes. Even in highly dynamic regions such as the Equatorial Ionization Anomaly (EIA), the RMSE remains stable within 2.60 TECu, demonstrating a significant forecasting error reduction of 20%–30% compared to the BiLSTM and standard Transformer models. Furthermore, comprehensive spatiotemporal analysis confirms that the framework maintains exceptionally high accuracy and robustness across all seasons and latitudinal zones. This study validates the efficacy of the hybrid architecture in complex space weather forecasting, providing a robust and physically consistent technical pathway for TEC prediction in the IMCP sector.