The algorithm for solving the geostationary orbit(GEO) using 11 parameters published by the International Telecommunication Satellite Organization(Intelsat) has been extended and applied to the rapid fitting and forecasting of the orbits of four GEO satellites. Firstly, the inverse solution of the 11 parameters is derived in detail. Then, the dynamic prediction of GEO satellite orbits and their onboard GNSS data is simulated. Finally, the accuracy of the orbits with varying fitting and prediction arcs is evaluated. The results show that when fitting data for 6 h, the computation time is less than 0.9 s, and the fitting orbit accuracy is better than 100 m. When using the inverse 11 parameters from different fitting data arcs for prediction, the accuracy of the predicted orbits initially increases and then decreases as the fitting time increases. When the fitting time is 2 days, the prediction orbit accuracy is the highest, and the prediction orbit accuracy of 1 day is better than 1 200 m.

This paper analyzes the orbit determination accuracy and orbit extrapolation accuracy under different gravity field models and finds that the accuracy is higher when the gravity field model order is 100 and 150 compared to when it is 50. In the reduced dynamic orbit determination of GRACE-C and GRACE-D, the 3D directional orbit determination accuracy of EGM2008 and Tongji-Grace02k can reach up to 22.6 mm and 24.0 mm, respectively, both of these gravity field models have better orbit determination accuracy than ITG-Grace03.In orbit extrapolation, Tongji-Grace02k has the highest accuracy, capable of reaching the sub-decimeter level, followed by EGM2008, while ITG-Grace03 has the lowest accuracy, both in decimeter level. Optimizing the random impulse parameters can improve the orbit extrapolation accuracy of the three gravity fields, with the optimization effect being more significant when the model order is 100 and 150 compared to when it is 50.

GNSS-R(global navigation satellite system-reflectometry) technology has been widely verified as an effective means of monitoring ocean tide levels, but the commonly used LSP(lomb-scargle periodogram) spectral analysis method still has issues with low precision and time resolution in storm surge monitoring. This paper, based on observational data from the HKQT station in Hong Kong, China, uses a non-linear fitting method to conduct inversion studies on the storm surge event of Typhoon "Mangkhut" in 2018. By combining wind speed change data during the storm surge period, the paper compares the performance of spectral analysis and non-linear fitting methods in tide level inversion. The results show that the measurement accuracy using the non-linear fitting method is better than 8 cm, with a correlation coefficient of 0.99 compared to the observational data from tide gauge stations. The tide level representation is smoother and more continuous than that of the spectral analysis method, effectively improving monitoring accuracy and time resolution, and providing more reliable technical support for storm surge disaster monitoring.

This paper applies generative adversarial networks (GAN) to the denoising of GPS coordinate time series. The generator produces time series data, which is then distinguished from clean time series data by the discriminator. During training, the generator is iteratively optimized until the discriminator can no longer differentiate between the generated data and the real clean data, thereby achieving the denoising effect. Simulation results show that, compared with traditional methods such as WNNM, BiLSTM, and SSA, the GAN method achieves the closest correlation with the original noise in key indicators such as residual amplitude, residual spectral index, and velocity uncertainty, demonstrating more efficient denoising performance. Experimental results using real data further validate the effectiveness of the proposed algorithm in denoising actual GPS coordinate time series.

This paper employs VTEC data from the Jason-1/2/3 satellite series to analyze the accuracy of global ionospheric maps (GIMs) from temporal and spatial perspectives, and establishes corresponding correction equations using solar activity components, diurnal variation components, and seasonal variation components. The results indicate: 1) The residuals(G-J) between CODE GIMs and Jason series satellite data exhibit a quadratic function relationship with F_10.7, showing a 24-hour periodic characteristic in diurnal variation and a semi-annual variation pattern in seasonal changes. The correction equations can effectively amend the GIMs data. 2) The mean value of G-J reaches its maximum in the low-latitude regions of the Pacific and its minimum in southern Australia; the plasma transport within the equatorial anomaly leads to a bimodal feature in the median error of G-J; the relative error of G-J attains its maximum in the southern Atlantic Ocean, particularly in the marine areas off southern Africa.

Based on the data of 56 sounding stations in China from 2009 to 2019, we analyze the correlation between precipitable water vapor(PWV) and zenith tropospheric delay(ZTD), water vapor pressure (E_s), atmospheric pressure (P_s), surface temperature (T_s), and weighted mean temperature (T_m), and then construct the China PWV models using multiple linear regression (MLR), random forest (RF), multilayer perceptron (MLP), and MLP-RF integration algorithm. The accuracy of these models are evaluated. The results show that the correlation coefficients between PWV and T_s, P_s, E_s, T_m and ZTD are 0.78, -0.71, 0.77, 0.74 and 0.99, respectively. The average RMSE of yearly PWV models based on MLR, RF, MLP and MLP-RF are 1.25 mm, 0.44 mm, 0.28 mm, and 0.21 mm, and the average MAE are 0.98 mm, 0.31 mm, 0.18 mm, and 0.14 mm, respectively. The average RMSE of seasonal MLR, RF, MLP, and MLP-RF PWV models are 0.60 mm, 0.45 mm, 0.29 mm, and 0.21 mm, and the MAE are 0.49 mm, 0.33 mm, 0.19 mm, and 0.14 mm, respectively. The yearly MLP-RF PWV model is with the highest accuracy, which is suitable for real time PWV acquisition and research in China region.

High-precision and high-resolution tidal models play a crucial role in the inversion of time-varying gravitational fields, but the published tidal models generally only provide the main tidal components. The fitting coefficient matrix for the secondary tidal components needs to be published by other institutions or calculated independently. This paper first derives the calculation formula for the linear fitting of secondary tidal components in tidal models based on the theory of admittance and tidal potential data. Then, taking the EOT11a, EOT20, and FES2014b tidal models as examples, we calculate the fitting coefficient matrices for the secondary tidal components of each tidal model, as well as the corresponding ocean tide perturbation forces for CHAMP, GRACE, and GOCE satellites for a period of 1 day. Finally, we select the FES2014b tidal model to invert the 120th-order time-varying gravitational field of GRACE-FO for January 2021. The precision and impact of the secondary tidal linear fitting method in this paper are analyzed from different perspectives. The results show that the differences between the fitting coefficient matrices of the three tidal models and the corresponding ocean tide perturbation forces for the three satellites calculated in this paper and the results from the ITSG can be neglected. Considering the secondary tides can improve the precision of gravity potential coefficients above the 40th order and reduce the north-south strip errors in the time-varying gravitational field. The method presented in this paper can provide a reference for independently calculating the impact of secondary tidal components.

Anisotropy exists in the Earth's inner circle, and its genetic mechanism is different.Shear wave splitting can reveal the characteristics of anisotropic media at various depths and structures, providing crucial information for the study of deep Earth dynamics and earthquake prediction. This paper introduces the mechanism of anisotropy in the Earth's interior, summarizes the development of shear wave splitting methods, and the achievements and progress of study the anisotropy of the crust-mantle coupling model by shear wave splitting method.

To address the challenges of traditional hard-threshold methods in dealing with ambiguous boundaries when processing abnormal stations, this paper proposes a new method that treats station anomalies caused by tectonic movements and observational errors as gross errors and uses a membership function to construct an IGGIII equivalent weight function to mitigate the impact of abnormal stations. The new method incorporates fuzzy set theory and is an improved IGGIII robust estimation method based on the membership function of the fuzzy subset to which standardized residuals belong to gross errors(membership function IGGIII, MF-IGGIII). Firstly, the membership function of the fuzzy subset most affected by gross errors in standardized residuals is determined through fuzzy statistics; then, an equivalent weight function is constructed based on these membership function values, and a weight iteration process is performed. Simulation experiments and the fitting analysis of the GNSS velocity field in the Bohai rim area show that this method outperforms hypothesis testing, mean shift methods, and the IGGIII method in abnormal station detection, demonstrating strong capability in identifying abnormal stations.

This paper uses the time series of the geocentric motion provided by the Center for Space Research at the University of Texas at Austin as experimental data. First, empirical mode decomposition (EMD) is applied to the data for noise reduction. Then, fast Fourier transform (FFT) and continuous wavelet transform (CWT) are utilized to perform frequency domain transformation, power spectrum analysis, and extraction of periodic terms. Finally, an autoregressive integrated moving average (ARIMA) model and exponential smoothing method are employed to forecast the geocentric motion for the next 20 months. The results indicate that the amplitude and phase of the annual term extracted using FFT are quite close to previous studies on the geocentric motion. The ARIMA model provides better predictions for the Y-direction motion of the geocentric motion over the next 20 months, while exponential smoothing method offers more accurate predictions for the X and Z directions.