To improve the positioning accuracy of mobile intelligent terminals, we analyze the observation data of multiple satellite systems collected by Xiaomi Mi8 and Huawei Mate20 smartphones, estimate the parameters of signal-to-noise ratio model based on pseudorange residuals of different satellite systems, and finally carry out the static pseudorange single-point positioning experiment. The results show that the pseudorange accuracy of BDS and GPS observed on mobile phones is close to each other and is higher than that of GLONASS. The correlation between pseudorange residual and SNR is stronger compared to elevation angle. In the parameter estimation of SNR model, the fitting effect of Xiaomi Mi8 is better than that of Huawei Mate20. The fitting effect of BDS is worse than that of other satellite systems, and the horizontal positioning and elevation positioning accuracy of SNR model after parameter fitting are 2.37 meter and 8.38 meter, respectively, which is about 10% higher than that of elevation model.

We propose a gait recognition model that fuses long short-term memory(LSTM) and convolutional neural network(CNN), the model can automatically extract activity features and classify them using a small number of model parameters, and we utilize this model for user identification.  The experimental results show that the recognition accuracy of the model is about 97.68 % and the loss value is about 0.05, which significantly improves the recognition rate compared to other models.

This paper is based on the AR products from CODE, GFZ, CNES, PRIDE Lab, and in-house AR products for the performance assessment of PPP-AR in terms of convergence time, time to first fix and positioning accuracy. The experiment uses seven-day observation data from 40 IGS stations in 2022 and precise products of each AR product. The results show that under the static solution mode with confidence level of 95%, five products have the most obvious improvement in positioning accuracy at the time of 1 h, which are 46.58%(3.4 cm), 41.10%(3.0 cm), 45.21%(3.3 cm), 34.25%(2.5 cm), and 41.10%(3.0 cm), respectively. In the simulation-kinematic solution mode, the positioning accuracy of the five products in the E and N directions can reach the millimeter level, and the accuracy of the U direction is slightly improved than that of the floating point solution, among which the accuracy of GBM product is the least improved, and 72.73%(2.4 cm), 47.37%(0.9 cm) and 5.41%(0.2 cm) were increased in E, N and U directions. WUM and COM products had better improvement effect. They were 81.82%(2.7 cm), 63.16%(1.2 cm), 24.32%(1.1 cm) and 81.82%(2.7 cm), 63.16%(1.2 cm), 15.58%(0.8 cm).

We use observation data from 5 GNSS/MET stations in Shanghai from July 21 to August 8, 2021 to invert atmospheric precipitable water vapor(PWV) by GAMIT software, and study the relationship between PWV and rainfall in Shanghai and the spatial distribution characteristics of PWV during typhoon “In-Fa”. The results show that the correlation between GNSS PWV and sounding PWV is above 0.9. Before the formation of precipitation, PWV undergoes three obvious rise processes and the precipitation generation time is about 60 h ahead of typhoon landfall.  PWV sharp change will cause a sharp rise or fall in precipitation. After the precipitation, PWV gradually drops below 40 mm. Before typhoon landfall, PWV is basically above 75 mm; PWV above 80 mm is mainly distributed in southern and central areas of Shanghai. After typhoon landfall, PWV is above 80 mm, and higher value areas can reach 90 mm. After the typhoon moved away, the PWV value dropped below 85 mm. The expansion and progression of the high-value area to the northwest is basically consistent with the typhoon path.

This paper introduces the research progress, standardized test and experimental observation on seismic station of the JCZ series ultra broadband seismometer JCZ-360. The results show that the JCZ-360 ultra broadband seismometer can effectively suppress the influence of external atmospheric pressure, external magnetic field interference and temperature change on the instrument. Its performance is more stable and noise level is lower at the ultralow frequency terminal. The JCZ-360 ultra broadband seismometer can not only observe full-band seismic waves, but also record information such as solid tides and free oscillation of the Earth. It has the comprehensive observation ability of traditional seismic and traditional geophysical field.

Using seismic data of Jianghan-Dongting basin and its surrounding areas from historical records to 2021 to analyze the spatiotemporal distribution characteristics of earthquakes, we obtain the relationship between deep tectonics and earthquakes by using the hypoDD algorithm to relocate 788 earthquakes, and analyze the characteristics of tectonic stress field by using the P-wave first motion method to calculate the focal mechanism solutions of M_L≥2.8 earthquakes. The results show that earthquakes are distributed unevenly in space, the seismicity of basin boundary is stronger than that inside basin, and there is an obvious periodic rule of alternating quiet and active in time. The relocated earthquakes are distributed in clusters in NW and NE directions in Ningxiang, Shimen, Yuanan-Dangyang, Nanzhang- Jingmen, and Ezhou-Huangshi, and the seismogenesis of this five regions are closely related to fault tectonic activity. The type of focal mechanism solution is mainly thrust fault, the dominant orientation of P-axis is nearly EW, and its plunge angle is almost horizontal, the orientation of T-axis is mainly SN and NE, and its plunge angle is large. The horizontal collision between the Indian ocean plate and the Eurasian plate result in the formation of a nearly EW tension stress field in the Tibetan plateau, which pushes the basin eastward. In addition, the Philippine sea plate subducts northwestward and the Pacific plate subducts westward, resulting in the formation of a nearly EW compression stress field in the basin, which make the regional fault structure prone to vertical differential movement, and then led to the gestation and occurrence of earthquakes in the Jianghan-Dongting basin and its surrounding areas.

By fusing the observation data with the model，data assimilation provides an effective solution for realizing accurate mapping of physical and mathematical models to solid Earth as much as possible. We review the application and research status of data assimilation method in various branches of solid Earth geophysics, and make a preliminary discussion on the urgent problems of data assimilation method in solid Earth geophysics.

We compare and analyze the scale factors of long baseline calibration and datum solution of 5 CG-6 relative gravimeters, and process the data of gravity network in Guizhou of CMONOC in 2022 and 2023 using different scale factors. The results show that the scale factors of CG-6 relative gravimeters vary over time, and the precision of processing results is better by using the datum solutionscale factor.

We process 69 Sentinel-1A satellite images using SBAS-InSAR from July 2018 to August 2019, and obtain surface deformation in Shuicheng district, Liupanshui city as dynamic evaluation factors to address the lack of dynamic feature data in traditional landslide susceptibility studies. The results show that, by fusing ten static evaluation factors and InSAR deformation feature data as dynamic evaluation factors, in a weighted information model coupled with analytic hierarchy process(AHP) and information volume method, the model performance improves approximately 13.3% compared to using only static feature data. The area under the ROC curves is 0.756 02 and 0.888 68, respectively. To assess zoning accuracy, we overlay historical disaster sites on two types of zoning maps. Compared to scenarios without the inclusion of deformation features, the introduction of deformation features corrected approximately 12.44% of misclassified areas, significantly enhancing zoning reliability.

This paper introduces the methods and steps for estimating multi-GNSS multi-frequency code biases and validates the effectiveness of estimating inter-frequency code biases using a multi-GNSS ionospheric modeling approach. We compute daily solutions for various code biases using MGEX network observations from January 2019 to July 2023. We further analyze the accuracy of various code biases. The results indicate that the multi-GNSS multi-frequency OSBs calculated in this paper exhibit an average weekly stability ranging from 0.07 to 0.26 ns. Among them, the average weekly stability of BDS-2 and BDS-3 multi-frequency OSBs are 0.17 to 0.26 ns and 0.13 to 0.21 ns, respectively. The consistency of various types of DCBs with products from DLR and CAS can be maintained within 0.3 ns. This research is not dependent on external post-processing products and can calculate code biases with both a rapid turnaround time and high precision.

Based on the satellite orbit characteristics of BDS GEO, IGSO, and MEO, we use multi-point hemispherical grid model(MHGM) and sidereal filtering(SF) method to establish a multipath error correction model for a mixed constellation of BDS. After applying the model to correct multipath errors, the post-fit carrier phase residuals for BDS precise point positioning(PPP) are significantly reduced, and the positioning accuracy in E, N, and U directions is improved by 41%, 37%, and 38%, respectively, with an overall increase in convergence speed of 31%.

On August 23, 2023, the seismic instrument system of the Hongyanhe Nuclear Power Plant successfully recorded a M4.6 earthquake in Pulandian, Dalian. This is the first complete strong earthquake record around the nuclear power plant area in China. We analyze the seismic data recorded by the system, and the data playback shows that the seismic recorded data was clear and complete, and the peak value of the free field monitoring data is consistent with the calculation result of the seismic attenuation empirical formula. Comparing the peak acceleration values of seven monitoring points data, the result shows that the nuclear power plant building has an amplification effect on seismic acceleration signals, and the amplification coefficient is positively correlated with building elevation. The conclusion of spectral analysis of peak acceleration data shows that predominant frequency points of seismic response data in nuclear power plant buildings are mainly concentrated in the range of 10 to 20 Hz, and seismic acceleration signals in the range of 10 to 20 Hz also have certain destructive effects on the building. This conclusion is inconsistent with the seismic triggering filtering range (1 to 10 Hz) specified in NB/T 20076-2012 seismic instrument criteria, due to the fact that the seismic trigger recognition algorithm of the seismic instrument system is set in a filtering range of 1 to 10 Hz. This defect will reduce the performance of the seismic instrument system’s depth safety defense.

We conduct PPP experiments using 80 global MGEX stations in September 2021, and compare the results with ZTD product released by the international GNSS service (IGS) as a reference. The results reveal significant advantages in ZTD accuracy of multi-system joint estimation. The average RMSE accuracy of GPS+BDS dual-system is approximately 0.6 mm higher than that of single GPS system, while the GPS+BDS+GLONASS+Galileo four-system exhibits a further improvement of about 0.9 mm than dual-system. The ZTD estimation accuracy of GPS under single-system condition is higher than that of BDS. In terms of spatial distribution, the improvement of ZTD estimation accuracy is more significant with increasing latitude. The four-system PPP ZTD accuracy is better than 5 mm when latitude exceeds 50°. Moreover, elevation increase of observation stations can improve ZTD estimation accuracy while latitude remains unchanged. The ZTD estimation accuracy is significantly improved under ambiguity resolution. The average absolute error (MAE) and root mean square error (RMSE) of ZTD estimated by single GPS system are 7.6 mm and 8.4 mm, respectively, which are about 11% and 12% higher than floating-point solutions. The average convergence time is reduced by 20 minutes.

To obtain high-quality original seismic data of the Wuwei basin of the lower Yangtze block, we conduct experimental research based on the reflection seismic data from 2019 to 2020, such as high density wide line acquisition technology, excitation technology, and tomographic static correction technology. We summarize a set of collection and construction parameters suitable for shale gas seismic exploration in the Wuwei basin, and apply these parameters to the newly deployed seismic exploration work in the research area. We obtained seismic data with high signal-to-noise ratio. By performing Kirchhoff pre-stack time migration on the data, we obtain a high-resolution and high-precision time-domain reflection profile. The target layer(Permian) on the profile has strong reflection wave energy and continuous in-phase axes, which can basically clarify the geological structure framework of the area. This can provide reliable seismic imaging data for the drilling and deployment of shale gas in the Wuwei basin.

We use remote sensing data from the C-band Sentinel-1 and L-band ALOS-2 satellites, which cover the Menyuan area, and apply differential interferometry(D-InSAR), pixel offset tracking(POT), and multi-aperture InSAR(MAI) techniques to obtain the coseismic surface deformation fields of the 2022 Menyuan, Qinghai M_W6.6 earthquake in line-of-sight and azimuth directions. By combining the surface stress-strain model with the variance component estimation(SM-VCE) method, we deduce the three-dimensional coseismic displacement. Furthermore, we use both nonlinear and linear approaches to invert the geometric parameters, such as dip angle and strike, as well as the distribution characteristics of coseismic slip. The results reveal that the earthquake was a left-lateral strike-slip event, with the fault extending northwest-southeast from the epicenter. The maximum horizontal displacement reached 1.58 m, and the predominant sliding occurred within a depth range of 0 to 8 km underground, with a maximum slip of 3.54 m. Based on the inversion results, the moment magnitude of earthquake is estimated to be M_W6.67.

Whether the initial search space is set reasonably or not directly affects the reliability and search efficiency of ambiguity resolution. In view of the constellation characteristics of BDS system, it is necessary to analyze the correlation between ellipsoid volume and the number of candidate points, and the effect of data redundancy for BDS2 and BDS3 under different satellite numbers and frequencies. For the drawback that objective function method is prone to produce too many candidate vectors when determining the ambiguity search space, we propose an optimized search space determination method, and verify the effectiveness of improved method through experiments. The results show that the improved search space determination method significantly reduces the number of actual ambiguity candidate vectors, with average improvement up to 35.02%, and ensures that the number of candidate vectors contained in 94% space satisfies the requirements. The improved search space method improves search efficiency by nearly 33% on average after applying it to different search methods, which is conducive to separating the redundant candidate vectors in initial search space, thus improving the ambiguity search efficiency.

Based on the mobile geomagnetic observation data of four consecutive periods from 2020 to 2023 in the border area of Hebei and Shandong, this paper analyzes the temporal and spatial variation characteristics of the regional lithospheric magnetic field and its relationship with the Pingyuan M_S5.5
 earthquake on August 6, 2023. Combined with the laboratory results of rock magnetism and the theory of fault meta-instability, it discusses the meta-instability characteristics before the plain earthquake and the seismogenic mechanism.The results show that:1) The spatial and temporal variation characteristics of the lithospheric magnetic field in the study area have obvious anomalies before the Pingyuan M_S5.5 earthquake. There is always a weakening area of the horizontal vector near the epicentre from 2020 to 2023, and with the passage of time, the weakening area with horizontal vector gradually approaches the epicentre; there is always a 0 value line of the total geomagnetic intensity, magnetic declination and vertical component near the epicentre. 2)The time series variation of the observation points near the epicentre in 2020-2021 is highly consistent, which better reflects the synergistic activity state of the faults near the observation points, and may be used as one of the characteristics to identify the sub-instability stage of the faults.3) By analyzing the relationship between stress and magnetic field intensity, it can be seen that the total intensity of geomagnetic field in the middle and eastern segment of Linnan fault and the north-eastern segment of Liaocheng-Lankao fault increases, and the crustal stress should be in a release state; the total intensity of geomagnetic field in the western segment of Linnan fault and Tangyi fault decreases, and the crustal stress should be in an accumulation state. The stress release of the former can promote the stress accumulation of the latter, which may lead to the instantaneous increase of the shear stress of the latter and instability when it exceeds its strength, thus triggering the plain earthquake.

Based on lithologic characteristics, samples age, sporo-pollen and sedimentary cycle analysis of the controlled drill in northern Taiyuan basin, as well as the known data of sporo-pollen assemblage and strata lithology of late Cenozoic in Taiyuan basin, we divide the Quaternary strata in Taiyuan basin. The Quaternary strata from top to bottom in Taiyuan basin is divided into: from 0 to 31.6 meter is the Holocene; from 31.6 to 95.5 meter is the late Pleistocene, which is further divided into the early(31.6 to 47.8 meter), the middle(47.8 to 78.2 meter) and the late(78.2 to 95.5 meter) periods; from 95.5 to 186.16 meter is the middle Pleistocene; from 186.16 to 250.42 meter is the early Pleistocene. On this basis, we analyze the evolution characteristics of the Quaternary in Taiyuan basin based on the color and particle size features of sediment and sporo-pollen analysis results.

The traditional Kalman filter(KF) of GNSS/SINS integrated navigation system is optimal under the minimum mean square error(MMSE) criterion and under the Gaussian hypothesis. However, when the measurement noise is disturbed by heavy tail pulse noise, the filtering performance of KF is seriously degraded. To solve this problem, we propose a maximum entropy Kalman filter(MCKF) for integrated navigation system. First, we establish the state equation and measurement equation of MCKF. Then, using the principle of maximum entropy, we establish a Kalman filter based on the maximum entropy criterion, and design its filter iteration flow. Finally, we simulate the GNSS/SINS integrated navigation system in the environment of mixed Gaussian noise and heavy tail pulse noise, respectively. The simulation results show that KF performs better than MCKF under mixed Gaussian noise interference, MCKF has better filtering performance than KF under the interference of heavy tail pulse noise, and MCKF is equivalent to KF when the kernel bandwidth tends to infinity.

At present, the seismic stations need to use two sets of acquisition systems to monitor and record weak signals and large seismic signals. To address this problem, this paper proposes a data acquisition method that automatically switches acquisition modes. When monitoring weak signals under normal conditions, the acquisition system uses small-signal channel ADCs for graded sampling, and when large earthquakes and giant earthquakes occur, the system automatically switches to use large-signal channel ADCs for sampling. Experiments show that the dynamic range of the acquisition system reaches 128 dB for single AD data acquisition at 100 Hz sampling, and 158 dB for double AD data acquisition at 100 Hz sampling.The improved data acquisition method effectively improves the dynamic range of seismic signal acquisition, and improves the performance of the data acquisition system to a certain extent without increasing the hardware cost.

We use the catalog of modern earthquakes in the Beijing-Tianjin-Hebei region from 1980 to 2022 as a statistical sample. To grid the study area, we use the incompletely centralized Voronoi segmentation method based on spatial segmentation. Based on the Poisson model, the probability model of earthquake risk is established, and we conduct the probability prediction of small and medium earthquakes at the medium-term scale of the study area. The results show that the calculated probability of relatively high value region has a certain correlation with the occurrence of earthquakes with M_S≥3.0, which can provide a reference for medium and long term earthquake prediction.

The traditional method of geoid-quasigeoid separation is difficult to meet the accuracy requirements in mountainous regions with complex terrain. We use a strict formula to calculate the geoid-quasigeoid separation in the mount Qomolangma area and analyze spatial changes. The results show that the total correction using the strict formula varies from －3.270 m to －0.119 m, the potential correction reaches 1.272 m, and the gravity gradient correction reaches －0.138 m. The magnitude of these corrections is large enough and needs to be taken into account in mountainous areas. The strict method and the approximate method both have a certain correlation with height, and the former produces smoother results than the latter. Under the influence of Bouguer gravity anomaly and terrain changes, some local features of strict method are negatively correlated with height.

We select 38 views of Sentinel-1A SAR images, and obtain the surface subsidence results within 5 km along the Qufu-Heze section of the Lunan high speed railway using SBAS-InSAR technology for the period from February 2019 to November 2022. We analyze the distribution characteristics and patterns, and use the PSO-BP model to predict the subsidence of some feature points. The results show that the annual average deformation rate is between －20 and 15 mm/a, the maximum subsidence velocity is 25.46 mm/a, and the maximum lifting velocity is 17.43 mm/a within 0.1 km along the high speed railway; the RMSE of the subsidence prediction value obtained by the PSO-BP model ranges from 5.8 to 12.4 mm, which can predict the surface subsidence well.

Based on NGA database and using genetic programming techniques, we give a set of prediction equations for PGA, PGV and PGD. On the basis of the explicit prediction formula, the reliability test and model comparison are carried out by establishing the correlation between ground motion parameters and key seismological parameters such as magnitude, fault distance, fault mechanism, and site shear wave velocity within 30 meters. The results show that: 1) Compared with the traditional decay relation-like prediction equations based on nonlinear regression techniques, genetic programming techniques do not need to specify the equation form of the decay relation, can model the complex behaviors of PGA, PGV and PGD, and give explicit formulas to meet engineering needs. 2) Compared with the Campbell-Bozorgnia attenuation relationship, the prediction effect of PGA and PGV based on the genetic programming techniques is slightly better; the RMSE and MAE of the PGD prediction model are 5.47 and 1.64, respectively, which are significantly smaller than those of the Campbell-Bozorgnia model, which are 45.98 and 4.61. 3) The obtained ground motion prediction equations are characterized by magnitude effect, site amplification effect, and saturation effect of near-field large earthquakes, but fail to reflect the soft-soil damping effect, and the maximum site amplification coefficients of PGA, PGV and PGD are about 1.42, 2.53 and 2.64.

Aiming at the shortcomings of the seafloor topography inversion based on gravity data in irregular shape sea area (e.g. the sea-land boundary area), such as the implementation difficulty, we propose the improved gravity-geologic method (iGGM) and regression analysis method (iGA) based on the approximate linear relationship between seafloor topography data and gravity data, combined with depth soundings and gravity anomaly data, to predict seafloor topography in the area. We select the range of 2°×2°(6°N~8°N, 115°E~117°E) in the south China sea as the study area. Using iGGM and iRA, and compared with the SCS_Grid model (gridded acoustic bathymetric data in the study area) and topo_23.1 model, etc., we construct the SCS_iGGM and SCS_iRA seafloor topography models. The experimental results show that the overall statistical results of the SCS_iGGM and SCS_iRA are similar to the SCS-DTU18 model, SCS-ETOPO1 model and SCS_topo_23.1 model in the sea-land boundary sea area, and the correlation coefficient is higher than 0.99. The external checking results in the study area shows that the accuracy of the SCS_iGGM and SCS_iRA is comparable. Compared with the SCS-DTU18 model, SCS-ETOPO1 model and SCS_topo_23.1 model, the accuracy of the SCS_iGGM and SCS_iRA models are improved by about 76%, 70% and 53% respectively. Combining the sea gravity anomaly and ship depth soundings to establish seafloor topography model, either in terms of the seafloor topography data recovery capacity, or the quality of model construction, is better than the seafloor topography model that relies only on ship depth soundings data.

Acoustic positioning technology can transfer the position reference frame maintained by GNSS into the ocean. Domestic GNSS-acoustic seabed positioning experiments typically involve the cooperation of multiple hardware devices. However, the observation data outputted by different devices may suffer from time stamp jumps. While time stamp jumps can be resolved through redundant observations, there is currently no effective solution for cases without redundant observations. Therefore, we propose a time stamp jump correction method based on parameter search. By constructing a search space for time stamp deviations and using the unit weight variance of acoustic positioning results as a criterion, the time stamp jump in observation data is searched numerically. The proposed method is validated using measured sea trial data. The results show that the parameter search method successfully detects a time stamp jump of approximately 1 second in acoustic device data and improves the accuracy of acoustic positioning from decimeter-level to millimeter-level by correcting the data time stamps.

In this paper, the interior layer of Dushan substation in Guizhou province is taken as the research object. We generate a composite dispersion curve by superimposing the dispersion energy of multichannel analysis of active surface waves(MASW) and spatial auto correlation(SPAC) of passive surface waves. The curve has a wider frequency band, which can not only ensure the resolution of the surface layer and avoid the formation of a large exploration blind area, but also effectively improve the inversion depth and obtain a more refined shear wave velocity model. The model can be given geological significance in combination with the drilling data. The results show that the combination of MASW and SPAC is beneficial to the fine imaging of the strata from the surface to the deep layer, and to understand the lithological structure characteristics of the strata. Compared with drilling, this method has low field construction cost and high efficiency, and it can provide an important reference for the detection of unfavorable geological bodies in complex urban areas.

Using high resolution remote sensing images, we accurately extract the water body boundary of the head area of the Three Gorges reservoir. A precise discretized numerical model is constructed for the water body to simulate and calculate the gravity effect of the head area of the Three Gorges reservoir at different storage and release water levels. We finely preprocess the continuous observation data of the gPhone 101 gravity gravimeter at Zigui station in 2017 and the gPhone 113 gravity gravimeter at Maoping station in 2019 to 2021. The results show during the water storage process from 145 m to 175 m, the simulation values of gravity change on the bank at different storage levels (distance is 5 m) are 0 to 28 μGal, 0 to 47 μGal, 0 to 60 μGal, 0 to 97 μGal, 0 to 170 μGal, and 0 to 210 μGal, respectively. There is an effect of 10 to 20 μGal within 2 km from the bank at the time when the storage level is elevated by 5 m. When the water storage level increases 30 m, there is an effect of 30 to 40 μGal within 5 km from the bank. The gravity residuals mainly reflect the hydrological gravity effect information, the gravity residuals at Zigui station recorded about 40 μGal of gravity change, and the gravity residuals at Maoping station recorded about 50 μGal of gravity change. The gravity change trend and the water level change show very good consistency, verifying the accuracy of the model simulation results, and provides a reference for the stability monitoring of the head area of the Three Gorges reservoir and its banks.

Low-cost real-time kinematic (RTK) receivers exhibit weaker resistance to multipath effects, leading to a significant decline in single point positioning (SPP) performance in remote ocean environment. We comprehensively analyze the impact of carrier phase smoothing multipath reduction algorithm on the performance of low-cost RTK receivers in remote ocean environment. The results show that the multipath effects on RTK receivers in remote ocean environment are 2~3 times greater than that in land environment, with the strongest resistance observed at E5 frequency. The effectiveness of smoothing multipath reduction is related to pseudorange multipath effects. After using smoothing multipath reduction algorithm, the SPP accuracy of RTK receivers improves by 62% in remote ocean environment and 24% in land environment. The SPP accuracy of GPS/Galileo/BDS-3 multi-systems is superior to 1 meter, meeting the precision requirements for fine ocean operations.

We optimize the back propagation neural network(BPNN) by using particle swarm optimization(PSO), and establish an optimized GPT3 model(MGPT3). We use the zenith tropospheric delay(ZTD) data of 30 IGS stations in Europe for 366 days in 2020 as an example to compare the accuracy of MGPT3, UNB3m and GPT3 models in predicting ZTD. The results show that the RMSE of MGPT3 model is 18.49 mm, which improves the accuracy by 55.0% and 47.7% compared to UNB3m and GPT3 models, respectively.

We proposed an adaptive spatiotemporal filtering method for GNSS coordinate time series, which dynamically subdivides the filtering region into several subregions, facilitating the extraction and elimination of common mode errors under specified thresholds. We conduct a case study on 184 vertical GNSS coordinate time series from CMONOC. The results of three groups of randomized experiments show that the average RMS values of the station sequences after adaptive PCA spatiotemporal filtering are reduced by about 39.7%, 38.4%, and 39.7% on average compared with that before the filtering, and the filtering effect is better than that of the overall PCA spatiotemporal filtering method. Analysis of pre- and post-filtering site noise characteristics reveals that, compared to traditional PCA filtering, our method resulted in an additional attenuation of approximately 17.8% in power-law noise amplitudes within site residual time series.

To study the applicability of the H/V spectral ratio method for determining the predominant period of laterite sites, we conduct systematic microtremor measurement on different classifications of laterite sites and process the microtremor measurement data using different combinations of horizontal components in the H/V spectral ratio method to get the HVSR curves. We analyze the HVSR curves to obtain the predominant period of the sites, and then determine the site classifications and divide the site geotechnical types. Finally, we verify the applicability of the microtremor H/V spectral ratio method for determining the predominant period of laterite site by borehole exploration. Moreover, using the combination of the sum of east-west and north-south vectors as horizontal components in the microtremor H/V spectral ratio method can obtain the predominant period of the site more easily and accurately.

In order to accurately estimate the GNSS velocity field in Jiangxi province and its adjacent area, we select the coordinate time series of 133 GNSS stations in the study area from the crustal movement observation network of China(CMONOC), and deeply study the influence of environmental loading changes on noise models. We calculate the nonlinear change caused by environmental loading, use the maximum likelihood criterion and Bayesian information criterion to determine the optimal noise model of each coordinate time series before and after loading correction, and estimate the velocity uncertainty. The combination of flicker noise and white noise is the main noise model of 133 GNSS stations in Jiangxi and its adjacent area. The noise characteristics of coordinate time series can be directly affected by environmental loading correction. Moreover, the velocity uncertainty of E, N and U directions decrease by 23%, 22% and 21%, respectively, after adding the environmental loading correction. The results show that environmental loading correction is helpful to accurately determine the optimal noise model of GNSS stations in Jiangxi and its adjacent area, and improve the accuracy and reliability of velocity field estimation. It is suggested to consider the influence of environmental loading when calculating GNSS velocity field.

By analyzing the problems existing in the cycle slip detection method of wide-lane phase minus narrow-lane pseudorange and geometry-free combination, we propose a cycle slip detection method that considers the GNSS satellite elevation factor and is suitable for universal GNSS observation data. After detecting the epoch position where the cycle slip occurs, we use the method based on spatial search and minimum criterion of objective function to repair cycle slip. We then validate the proposed cycle slip processing method by selecting satellite data of different elevations observed from the universal GNSS receiver located at a landslide hazard near the Linxia Highway in Gansu Province. The results show that the proposed method is effective. the cycle slip detection and repair method can realize the accurate detection and repair of cycle slips in carrier phase data with different elevations and can reduce the influence of GNSS data observation error on the cycle slip test. It is feasible for the cycle slip detection and repair in universal GNSS observation data.

Addressing the under-quantified crustal deformation characteristics and the unclear stress-strain coupling relationship of the Ordos block, this study employs a spherical least squares collocation model to calculate the characteristic strain field of the study area's crust and inverts the principal stress distribution using seismic source mechanism data. The results indicate that crustal deformation is primarily concentrated along plate boundaries and within fault basins, with less deformation within the Ordos block itself. Notably, at the junction of the northeastern Tibetan plateau and the Alxa block, the crustal deformation is particularly intense, characterized by strong compression along the plate boundary, with the main deformation mechanism being a coexistence of NW-SE compression and NE-SW extension, with a maximum principal compressive strain rate exceeding 0.05 μstrain. The crust in the study area is under a relatively uniform tensile stress environment, with the direction of the maximum principal stress being nearly horizontal and pointing towards SEE, aligning with the direction of the principal strain in the corresponding area. The minimum principal stress direction is closer to the direction of the principal compressive strain.

In view of the non-linearity and non-stationary characteristics of satellite clock bias(SCB) time series, as well as the interference between trend and noise components that may affect the accuracy of prediction, this paper proposes a SCB prediction model(SSA-ANFIS) based on singular spectrum analysis(SSA) and adaptive neuro-fuzzy inference system(ANFIS). This paper first uses SSA to decompose and reconstruct the first-order difference sequence of clock bias, obtaining the trend component and the residual component. Then, it uses the ANFIS model to predict the reconstructed components, and superimposes and restores the predicted results to obtain the final predicted clock bias value. Finally, through experiments, this paper compares the proposed model with GM, QP, LSTM and ANFIS models. The results show that SSA-ANFIS model can effectively improve the prediction accuracy of the single model. Compared with the LSTM and ANFIS models, its prediction accuracy increased by 25.7%-40.7% and 39.4%-45.7%, respectively.

Based on the mechanism of deformation anomaly, we analyze the correlation of several anomalies before the earthquake. Combined with observation facts and experiments, we explore the possible organic connections between them. The explanations are given uniformly from the mechanism for the anomalies before and after the earthquake, such as the severe drought before the earthquake, the heavy rain after the earthquake, the sand jetting and ground collapse during the earthquake, as well as ground sound and light, the ionosphere anomaly before the earthquake, the atmospheric electrostatic anomaly before the earthquake, and the electromagnetic wave anomaly before and during the earthquake. We consider that these anomalies are all related to the formation, expansion and rapid disappearance of micro-crack expansion bodies near the seismic source before the earthquake.

We use the data of 24 stations in the reservoir head area collected by Three Gorges reservoir induced earthquake monitoring system from 2016 to 2018, and use PALM method to analyze microseismic detection and positioning in Three Gorges reservoir area. A total of 12 814 seismic events are detected, which is about 4 times the number of seismic events in Three Gorges seismic network directory during the same period. The complete magnitude decreased from M_L0.8 to M_L0.3, proving that the PALM method is an effective way to improve the detection ability of seismic network. Based on the construction of a more comprehensive microseismic catalog, we delineate the spatial distribution characteristics of seismic activity in Three Gorges reservoir area. The results show that the focal depth in Badong area is relatively shallow, mostly within 5 km, mainly related to non-structural factors such as karst, coal mine collapse, and shallow unloading. The focal depth in Zigui area is relatively deep, mostly within 5 to 15 km, mainly related to the Jiuwanxi and Xiannüshan faults in the reservoir area.