To enhance the computational efficiency of filter-based joint reduced-dynamic precise orbit determination (POD) for LEO satellites and GNSS, and to strengthen its feasibility for realtime applications, this study performs a simplified analysis and validation of the dynamical parameters to be estimated for LEO satellites. Based on science orbit data from seven LEO satellites during 2019 to 2024, it is found that Swarm-A/C and GRACE-C/D during periods of high solar activity exhibit relatively large estimated accelerations (90 to 140 nm/s²), while the remaining satellites show smaller estimated accelerations (≤30 nm/s²). Validation through four designed POD schemes demonstrates that when the estimated acceleration is small, the best orbit accuracy can be achieved without introducing any dynamical parameters. If the estimated acceleration is large, only adding an atmospheric drag scaling parameter yields optimal accuracy. Simulation experiments further show that in a joint POD scenario involving 120 LEO satellites and BDS-3 satellites, the proposed simplified LEO dynamical parameter model improves computational efficiency by about 30% compared with traditional models, thereby enhancing the realizability of realtime joint orbit determination.

To address the issues of limited nonlinear modeling capability and low prediction accuracy in single-model satellite clock bias prediction, a combined prediction model integrating long short-term memory (LSTM) neural network and random forest (RF) is proposed. This model fully leverages LSTM's ability to deeply mine the temporal characteristics of clock biases, while utilizing the ensemble learning mechanism of Random Forest to non-linearly compensate for the prediction residuals of LSTM, thereby achieving a significant improvement in prediction accuracy. Experiments demonstrate that, compared to the single LSTM model, the LSTM-RF combined model enhances the prediction accuracy for GPS, BDS, Galileo, and GLONASS satellites by 74.4% to 86.8%, 27.5% to 32.0%, 80.5% to 82.7%, and 0 to 18.2%, respectively, in short-term, medium-term, and long-term predictions. This verifies the feasibility and universality of the LSTM-RF combined model for predicting nonlinear clock bias sequences.

In the maximum correntropy Kalman filter based on equal weighting (MCKF), since the error samples with different qualities are weighted equally, it affects the filtering accuracy of the GNSS/SINS integrated navigation system based on MCKF. To solve this problem, a non-equal-weighted maximum correntropy Kalman filter (NWMCKF) is proposed. Firstly, based on the equal weighting maximum correntropy criterion (MCC), a non-equal-weighted maximum correntropy criterion (NWMCC) is proposed to reduce the adverse effects caused by MCC applying the same weight to all error samples. Then, based on NWMCC, the algorithm equation of NWMCKF is derived, and the fixed-point iterative recursive algorithm flow of NWMCKF is given. Finally, experiments on the GNSS/SINS integrated navigation system based on KF, MCKF and NWMCKF are carried out for verification. The experimental results show that when the measurement noise is Gaussian noise, the filtering performance of NWMCKF is slightly better than that of MCKF; when the measurement noise is non-Gaussian noise, the filtering performance of NWMCKF is significantly better than that of MCKF.

Based on PWV data from 240 GNSS stations of the Crustal Movement Observation Network of China from 2017 to 2019, combined with PWV data retrieved from two AI-based meteorological models, Fengwu and Pangu (hereafter referred to as FW PWV and PG PWV), this study investigates bias correction methods for AI-derived PWV. First, using GNSS PWV as a reference, the accuracy of FW PWV and PG PWV were validated. Subsequently, to improve the accuracy of the two AI-based models, a bias correction method was constructed using a long short-term memory (LSTM) network. The input features included longitude, latitude, elevation, day of year, and AI-derived PWV from 2017 to 2018 at GNSS stations, while the output feature was the difference between FW PWV or PG PWV and GNSS PWV. The method was then applied to correct the AI-derived PWV for 2019. The results show that the mean bias and RMSE of FW PWV and PG PWV are -0.01 mm, 2.88 mm and 0.28 mm, 3.04 mm, respectively. After correction, the RMSE of FW PWV_L and PG PWV_L decreased to 2.29 mm and 2.38 mm, with improvement rates of 20.48% and 21.71%, respectively. The reliability of the corrected results was further validated using radiosonde PWV data from 2019. The bias correction method effectively reduced RMSE across all months, with monthly RMSE values of FW PWV_L and PG PWV_L concentrated between 1.4 mm and 3.4 mm, and an average RMSE improvement rate of 22.30% in summer. For PG PWV_L, the absolute bias in the elevation range of 0 to 1.5 km decreased by 0.45 mm, 0.46 mm, and 0.32 mm, respectively, while the RMSE improvement rate reached 23.55% in the 1.5 to 2.0 km range. FW PWV_L performed best in regions above 2 km elevation, with an absolute bias reduction of 0.23 mm and an RMSE improvement rate of 29.50%.

To explore high-accuracy autonomous polar motion (PM) prediction using purely data-driven methods, we propose a hybrid model named PSO-FFRLS+MAR (forgetting factor recursive least squares optimized by particle swarm optimization and multivariate autoregressive). This model dynamically weights historical data to track slow system variations and jointly models the co-oscillations of PMX and PMY residuals. Experiments show that during the 1 to 90 days prediction horizon, the proposed model outperforms mainstream models in both accuracy and stability, with particularly significant improvements for PMY (compared with the classical LS+AR model, a 35.4% reduction in MAE). Day-by-day validation confirms the statistical significance of these improvements. The greater enhancement in PMY accuracy may be physically attributed to the model's enhanced ability to capture the strong seasonal atmospheric excitation that dominates this component. This study provides a high-performance data-driven method for medium- to short-term autonomous prediction of Earth orientation parameters and, through error analysis, links the model's physical mechanisms with geophysical angular momentum excitation functions.

To address the challenges of strong dependence on initial models, low convergence efficiency, and susceptibility to local optima in traditional Rayleigh wave dispersion curve inversion for subsurface shear-wave velocity structure, this study proposes two enhanced algorithms based on the whale migration algorithm (WMA). By introducing Lévy flight and spiral flight strategies, we develop the Lévy flight whale migration algorithm (LF-WMA) and the spiral flight whale migration algorithm (SF-WMA) to strengthen global search capability and local exploitation ability. The performance of five algorithms-damped least squares (Levenberg-Marquardt method, LM), whale optimization algorithm (WOA), WMA, LF-WMA, and SF-WMA-was systematically evaluated using Rastrigin function tests, theoretical geological models, random noise tests, and field data applications. The results demonstrate that SF-WMA outperforms the other methods in inversion accuracy, stability, convergence speed, and noise robustness. It achieves the optimal solution of 4.74×10^-10 in the Rastrigin test, exhibits the smallest error and fastest convergence in noisy models, and, when applied to real seismic data, clearly reveals stratigraphic structures within the 0 to 30 m depth range. The inversion results are in good agreement with

Based on the earthquake catalog data from 1970 to 2025 for the Wuhan-Ezhou-Huanggang-Huangshi metropolitan area and its adjacent regions, the b-value was calculated using the maximum likelihood method. Leveraging the negative correlation between the b-value and differential stress, the stress state and seismic hazard of fault zones were analyzed in the context of the tectonic setting. The results indicate that: 1) The temporal evolution of b-values in the study area reveals a seismogenic process of "balanced accumulation-localized concentration-energy release-re-accumulation". Specifically, the overall low b-values from 1996 to 2010 corresponded to a strong stress-locking state prior to the Jiujiang earthquake. Since 2011, although b-values have rebounded, persistently low b-value anomalies on the northern side of the Xiangfan-Guangji fault suggest that this region is currently under a state of high-stress accumulation, posing a heightened risk of strong earthquakes. 2) Analysis of the Jiujiang earthquake comfirms that the b-value is sensitive to changes in tectonic stress. The low b-value anomaly in the source area before the earthquake reflected a high-stress locking state in the crust. The significant increase in b-values after the earthquake indicates stress release and crustal fragmentation caused by the mainshock rupture. Notably, analysis of b-value differences (Δb) reveals post-seismic stress migration and reloading toward the northwest of the epicenter, uncovering a complex process of post-seismic stress adjustment. This study enhances the understanding of the seismogenic dynamics of the Wuhan-Ezhou-Huanggang-Huangshi metropolitan area from the perspective of b-value spatiotemporal evolution and can provide a key basis for regional seismic hazard assessment and disaster prevention and mitigation.

Based on seismic waveform data, earthquake catalogs, and historical focal mechanism solution data, we systematically analyze the stress evolution characteristics of the seismic source area of the 2021 Yangbi, Yunnan M_S6.4 earthquake from the perspective of three types of stress parameters, namely apparent stress, b-value, and tectonic stress field. The results indicate that: 1) The apparent stress continuously increased before the main shock and decreased significantly after the shock, reflecting the physical process of pre-seismic stress accumulation and co-seismic stress release. 2) The b-value decreased obviously prior to the main shock, indicating a sustained enhancement of regional tectonic stress. 3) Results from the stress field inversion show that the main shock induced perturbations in the stress field of seismic source region, though these were not significant. Comprehensive analysis demonstrates that the variations in apparent stress and b-value effectively reveal the dynamic stress evolution characteristics of the seismic source area during the gestation and occurrence of the Yangbi earthquake sequence, which provides important insights for understanding the mechanisms of fault activity and stress evolution.

Based on three periods of crustal deformation monitoring data from 2018 to 2020 of Wudongde reservoir on the lower reaches of Jinsha River in China, we study the crustal deformation characteristics and mechanisms before and after reservoir impoundment. The results show that: 1) Reservoir impoundment has a significant impact on crustal deformation in the impoundment area, but has a weak impact on crustal deformation in the non-impoundment area. 2) Reservoir impoundment causes a crustal relative uplift in the impoundment area, which generally exceeds 10 mm and the maximum value exceeds 30 mm. 3) Reservoir impoundment induces a tensile strain towards both banks in the impoundment area, which generally exceeds 5 mm and the maximum value exceeds 10 mm. 4) Reservoir impoundment causes a noticeable increase in gravity values in the impoundment area, which generally exceeds 200 μGal and reaches 900 μGal in some local areas. 5) The increase of water mass and the rise of pore water pressure in the rock layers on both banks after reservoir impoundment may be the main reason for the crustal deformation of Wudongde reservoir. 6) The findings of this study differ from the crustal deformation results caused by the impoundment of the Three Gorges reservoir, which has an important reference significance for understanding the characteristics and deformation mechanisms of crustal deformation caused by reservoir impoundment.

Through visco-elastic-plastic-thermodynamic numerical simulation, this study reveals the dynamic processes following magma intrusion into the mid-crust and elucidates the synergistic control mechanism of magma properties and wall-rock rheological characteristics on migration pathways. The results show that: 1) The density difference between magma and wall rock is the key factor controlling whether magma can ascend. When the density difference is low, magma tends to spread laterally in the mid-crust; when the density difference is high, magma has sufficient buoyancy to initiate ascent. 2) Given that the density difference provides the driving force for ascent, viscosity dominates the migration style. Low-viscosity magma can rapidly intrude into tensile fractures at the top to form vertical dikes, whereas high-viscosity magma, constrained by flow resistance, ascends obliquely along shear zones at the margins of sills. 3) The presence of a mid-crustal magma chamber alters the surrounding stress field, prompting magma to ascend asymmetrically along ring faults and form double-wing transport channels. The numerical simulation results reveal the physical mechanism governing the divergence of mid-crustal magma migration pathways, clarify the potential constraints of deep migration processes on shallow volcanic eruption styles, and provide an understanding of the complexity of volcanic conduit systems and the spatial differentiation of eruptive activity.