Taking the 2022 Luding M_W6.7 earthquake as an example, we improve the signal-to-noise ratio of postseismic deformation by calculating the relative displacement time series between two GNSS continuous stations and simulate the viscoelastic relaxation effect using a layered medium model. The results show when the steady-state viscosity coefficient of lower crust is about 10¹⁹ Pa ·s, which can effectively explain the observed postseismic deformation and is consistent with the viscosity coefficient of lower crust beneath Tibetan plateau obtained from previous research, but slightly larger than that beneath western Sichuan plateau calculated by the postseismic deformation of 2008 Wenchuan earthquake.

We use SBAS-InSAR technology to obtain the surface deformation field of eastern Himalaya syntaxis, propose a method that use external GACOS data to eliminate atmospheric effects related to terrain based on interferometric phase, and develop a residual orbital trend phase correction method considering block motion. In addition, we use the elastic rebound theory and negative dislocation model to invert the fault locking depth and creep rate in this area. The InSAR results show that the deformation rate in LOS direction of eastern Himalaya syntaxis is about -10-15 mm/a. There is a significant differential movement on both sides of the southwest section of Dongjiu-Milin fault, which gradually weakens and tends to stabilize along the fault. There is no obvious surface deformation on both sides of Jiali Fault. The inversion results of fault movement model show that the locking degree of the southwest section of Dongjiu-Milin fault is relatively low, with a creep rate of 7-16 mm/a. Near the top of east Himalaya syntaxis, the fault locking degree gradually increases, and the creep rate gradually decreases to 2 mm/a. The locking depth of Jiali fault is 4-25 km, with a creep rate of 1-4 mm/a. The Jiali fault is in a deep locking state.

In order to investigate the applicability of InSAR technology in CO₂ geological storage safety monitoring, we use 43 Envisat images from October 2003 to September 2010, and obtain the deformation rate field around CO₂ injection well during the implementation of In Salah CCS project based on StaMPS-InSAR and SBAS-InSAR technologies. Furthermore, the underground CO₂ reservoir parameters of KB-501 injection well are inverted using the geodetic Bayesian method. The results show that: 1) With the injection of CO₂, deformation of 2-4 mm/a occurred around three injection wells, and the maximum cumulative deformation during monitoring period reached 32 mm. 2) Through the joint analysis of surface displacement time series and CO₂ injection data, we find that CO₂ injection rate significantly affects surface uplift rate. 3) Based on the horizontal Sill model, we use the Bayesian method to invert the depth of underground CO₂ reservoir in KB-501 injection well, and the best fitting value is 1 897 meters. The strike angle of horizontal edge is 37°, and the tensor of dislocation plane is 0.04 meters, which proves the applicability of horizontal Sill model and Bayesian method in geophysical processes analysis related to CO₂ geological storage.

Using fault gas radon geochemical detection technology, combined with geophysical detection, trench exploration, geological exploration and joint drilling data, we conduct a comprehensive study on the hidden section of Heyuan fault. The results show that: 1) Soil gas radon measurement reveals significant anomalies in the main peak value of radon gas at the ground projection position corresponding to the fault under cover layer. 2) The tendency of Heyuan fault determined by fault gas radon pulse curve has been verified by joint drilling and geophysical exploration. The fault occurrence is N30°-56°E/SE∠45°-55°. 3) The relative activity of Heyuan hidden fault is identified through soil gas radon measurement method, and the activity of northern section of Heyuan fault(Shuangtang-Sanchashui) is stronger than that of southern section(Beijiao village-Shuangtang).

We detect the Fendong fault in Linfen basin using shallow artificial seismic exploration technology, obtain two high-resolution seismic exploration profiles, and analyze the basic characteristics of shallow strata of Ganting section of Fendong fault. The results indicate that the shallow strata of Ganting section shows a west-dipping morphology, with the overall thickness decreasing from west to east. The Ganting section of Fendong fault is characterized by a north-east strike, a north-west dip, a visual dip angle of 67°-72°, and a fault drop of 26-182 meters. The shallowest upper breakpoint that can be identified is buried at a depth of 85 meters, indicating a Quaternary active fault. By comparing the profile characteristics of Ganting section in the north and Linfen section in the south of Fendong fault, we infer that the activity intensity of Ganting section during the Cenozoic is higher than that of Linfen section, and the activity intensity of Fendong fault gradually decreases from north to south.

We use 58 days of teleseismic waveform data collected from 225 short-period dense seismic arrays arranged across Jiangxi, Guangdong, and Fujian provinces, and use P-wave receiver function H-κ scanning stack and common conversion point (CCP) stack methods to analyze the crustal thickness, Poisson's ratio distribution, and receiver function imaging beneath seismic stations in the study area. The results show that: 1) The crustal thickness and Poisson's ratio in the study area exhibit clear block feature, closely related to the distribution of faults. The overall crustal thickness varies from thin to thick and gradually thins from northwest to southeast, with a range of 27.1 to 33.9 km and an average of 29.9 km. The variation of crustal thickness and surface fluctuation are consistent with the crustal isostasy theory. 2) The Poisson's ratio of the study area ranges from 0.21 to 0.33, with an average of 0.26. 70.83% of the whole survey line has a Poisson's ratio greater than 0.25, which belongs to the high Poisson's ratio area. It may be related to fault fracture zones in the study area. The deep melting of subducted plate materials leads to the upwelling of mantle materials along faults, which reduces the quartz content in the crust. 3) In the depth range of 10-15 km from the surface, there exists a relatively continuous negative polarity interface in the study area. Combining with previous studies, we interpret this as the top interface of a widespread low-velocity layer in eastern Cathaysia block. Within the depth range of 15-20 km from the surface, a low-velocity body is present and tends towards the inland direction.

Taking Beijing-Tianjin-Hebei region as the research area, we use the stratified PWV calculated by ERA5 reanalysis data to replace the total PWV. Combined with the FFT ConvLSTM model, we comprehensively consider the spatiotemporal correlation to predict PM_2.5 concentration. The model can effectively capture the spatiotemporal patterns of atmospheric pollutants and meteorological elements, and extract the optimal common change period through FFT technology and partition modeling to weaken the influence of spatiotemporal heterogeneity, realizing high-precision prediction of PM_2.5 concentration in the next 24 hours. The results show that in the plains, mountains and plateaus of the study area, the best prediction performance is achieved when 1-4 layers of ERA5 PWV is involved, reducing the RMSEs by 2.862 μg/m³, 5.384 μg/m³, and 2.283 μg/m³ respectively compared to that of total ERA5 PWV.

Based on the meteorological elements of precipitable water vapor(PWV), temperature, air pressure, and wind speed at GNSS station, and air quality data at neighboring environmental monitoring station in Suzhou city from 2021 to 2022, we analyze the temporal variation characteristics and their correlations between GNSS meteorological elements and PM_2.5 under different seasons and precipitation conditions. The results show that GNSS meteorological elements and PM_2.5 display seasonal variations. The PWV, ZTD, temperature, and wind speed are negatively correlated with PM_2.5, their correlation coefficients are -0.389, -0.389, -0.289, and -0.318. The air pressure and relative humidity are positively correlated with PM_2.5, their correlation coefficients are 0.250 and 0.017. This suggests that PWV, ZTD, temperature, air pressure, and wind speed are major factors influencing PM_2.5 concentration. During haze, the PWV and relative humidity are high, leading to the convergence of PM_2.5 particles. The correlation coefficients between PWV and relative humidity and PM_2.5 increase by 29.14% and 96.99%, respectively. Rainfall can significantly reduce PM_2.5 concentration, and the correlation coefficient between relative humidity and PM_2.5 increases by 90.39%. The correlation between GNSS meteorological elements and PM_2.5 is significantly improved after wavelet decomposition reconstruction, which can more clearly reflect the overall trend of GNSS meteorological elements and PM_2.5 changes.

In order to evaluate the long-term performance of GPS Ⅲ satellite clocks, we analyze the data characteristics and time-frequency characteristics of satellite clocks. The results show that GPS Ⅲ satellite clocks have a relatively fixed frequency drift rate, with a mean of (4.7±0.1)×10^-14/d. The fitting noise remains stable for a long time, with a mean of (2.0±0.007)×10^-10 s, but there is a cycle fluctuation of about 200 days. The frequency accuracy has relatively large fluctuations at the early operation stage but then goes into a relatively stable state, with a mean of (7.3±0.5)×10^-12. The mean value of frequency stability at average intervals of 1 000 s, 10 000 s and 86 400 s are (2.5±0.02)×10^-14, (5.9±0.04)×10^-15 and (4.2±0.04)×10^-15, respectively. There are significant periodic terms in satellite clock offset that are approximately one-half, equal or double times the satellite orbital period. The overall performance of GPS Ⅲ satellite clock has improved and enhanced compared with that of GPS Ⅱ F, but in addition to its accuracy index, the overall performance is not as good as that of BDS-3 and Galileo.

In view of the problem that GNSS observation data in complex environments such as landslide is affected by factors such as high occlusion and multipath, it is easy for the success rate of ambiguity resolution to decrease and the positioning reliability to deteriorate. We propose an improved algorithm of GNSS partial ambiguity resolution based on IGGⅢ robust estimation. The robust estimation algorithm is introduced to weaken the influence of gross error on parameter estimation and improve the accuracy of floating-point ambiguity. In addition, to avoid the influence of residual error on the subsequent integer ambiguity resolution, and to further improve the success rate of ambiguity resolution, we adopt a partial ambiguity resolution improvement strategy based on conditional variance matrix method. The experimental results show that compared with the full ambiguity resolution method, the algorithm can significantly improve ambiguity resolution rate, and the results of three stations are increased by 5.9%, 52.7% and 48.0%, respectively. At the same time, compared with the traditional Kalman filter algorithm, the improved robust estimation algorithm significantly improves the situation of incorrect ambiguity resolution and greatly improves the GNSS positioning accuracy. The horizontal RMS is better than 1 cm and the vertical RMS is better than 2 cm.

It is difficult to accurately model atmospheric drag, which is the biggest error source of low-orbit satellite precise orbit determination and orbit prediction. In orbit determination, the time-varying atmospheric drag coefficient (C_D) is considered to absorb errors of atmospheric drag model, obtaining better orbit fitting results. However, the time-varying characteristics of C_D parameter lack accurate modeling methods, resulting in gradually divergent orbit prediction errors. To solve this problem, we propose an orbit forecasting method for predicting C_D parameter based on Bi-LSTM neural network. Firstly, we calculate the C_D parameter of GRACE-C and Sentinel-3A satellites by dynamic orbit determination method. Then, we predict C_D parameter by Bi-LSTM neural network. The results indicate that the mean MAE values of C_D parameters predicted by GRCC and SN3A satellites are 0.030 2 and 0.077 4, and the mean RMSE values are 0.041 6 and 0.101 8. The results of four groups of orbit prediction experiments show that the highest average accuracy of GRCC satellite is 12.28 meters, and the average accuracy improvement rate is above 90%. The highest average accuracy of SN3A satellite is 16.00 meters, and the average accuracy improvement rate is up to 74.82%.

Celestialpole offset is the difference between observed celestial intermediate pole and precession-nutation model, mainly including free core nutation and forced nutation. Using ERA5 atmospheric reanalysis data, we calculate the atmospheric angular momentum function sampled at 1 hour and compare it with the atmospheric angular momentum functions sampled at 3 hours and 6 hours provided by IERS. We analyze the excitation effects on celestial pole offsets by high-frequency atmospheric angular momentum variations under different sampling rates. The results show that excitation effects on forced nutation and free core nutation by atmospheric angular momentum variations with different sampling rates have the same overall trend, but there are obvious differences in some harmonic frequency bands and other details. The atmospheric angular momentum function sampled at 1 hour can more effectively separate the harmonic components corresponding to diurnal frequency band in forced nutation term, and shows a certain advantage in the excitation of free core nutation, indicating the necessity of higher sampling rate atmospheric data in the study of high-frequency Earth rotation changes.

Based on variational mode decomposition (VMD) method, combined with the inherent digital signal characteristics of apparent resistivity observation data, we study the VMD algorithm that can extract long-term and annual changes of apparent resistivity by analyzing the center frequency of IMF component in each layer after VMD. The results show that a reasonable decomposition layer K can be given by determining the second layer IMF component after VMD is annual variation. Adaptive VMD is helpful in apparent resistivity anomaly recognition before an earthquake, especially in determining the turning time of long-term changes of apparent resistivity and identifying the annual changes anomaly.

We analyze the spatio-temporal characteristics of salinity drift using multiple sets of Argo datasets released by different organizations(SIO, JAM and EN4). The results show that since 2016, there is an obvious systematic drift in different salinity datasets, with significant differences in the magnitude of salinity drift at different depth layers. We propose a polynomial correction method based on the AIC criterion to correct salinity drift in specific depth layers. The method can effectively correct salinity deviation of Argo datasets within the depth range of 0-2 000 meters. This reduces the GMSL budget balance bias by approximately 43% from 2005 to 2021. Salinity drift displays intricate spatial correlations. There are significant differences in the spatial distribution of global halosteric sea level change trends, especially in the north Atlantic region, which shows a more consistent spatial distribution after correction.

By measuring the water temperature gradient and conducting hydrochemical analysis of Yining 1 and Yining 2 wells in hanging wall and footwall of Kashgar River fault zone, combined with regional geological data and wellbore column, we analyze the water temperature gradient characteristics and influencing factors of the two observation wells. The results show that the water temperature change of Yining 1 well is mainly controlled by rock geothermal temperature, while Yining 2 well is comprehensively affected by water temperature of supply water source and rock geothermal temperature. The water rock reaction degree and circulation depth of the groundwater in Yining 1 well are higher than those in Yining 2 well. Different levels of water rock reaction and circulation depth of supply water source will affect the water temperature gradient characteristics of observation well. The development degree of fractures in the stratum and aquifers distribution are the main factors leading to complex changes in water temperature gradient. The rock mass in the area where Yining 2 well is located is relatively fragmented, with developed fractures and large fluctuations in water temperature gradients. The rock mass in the area of Yining 1 well is relatively intact, and the water temperature gradient is relatively stable.

Based on the continuous second data of five VP vertical pendulum tiltmeters in Inner Mongolia from March 2020 to February 2021, we calculate the acceleration power spectral densities(PSD), and analyze the background noise characteristics at different regions and different frequencies. The results show that the background noise characteristics of VP vertical pendulum basically conform to the frequency features of global seismic background noise model. The background noise level in secondary microseismic frequency band is strong in autumn and winter, and weakest in summer, with significant seasonal variation, displaying no obvious spatial distribution characteristics. The background noise level in main microseismic frequency band has some seasonal variation, which is slightly higher in autumn and winter than that in spring and summer, similar in spring and summer, and significantly varies amongst stations. In the long-period frequency band, the noise levels of four seasons are usually similar, with no obvious seasonal variation. If the instrument is susceptible to natural environment interference, the background noise level will also show seasonal variation, which is higher in autumn and winter than that in spring and summer. In addition, fine cave conditions can effectively reduce the noise level in long-period frequency band.

To calculate the station noise acceleration power spectral density and the corresponding probability density function distribution, we select the three-direction continuous waveform data recorded by 14 seismic stations in Ningxia seismic network from January 1, 2022 to December 31, 2022 and 23 newly built reference stations from January 1, 2023 to March 31, 2023. Then, we analyze the characteristics of noise levels in different regions and different frequency bands. The results show that in the frequency band greater than 1 Hz, the stations with cave installation method have a mean circadian variation of noise level that is about 2 dB lower than that of other stations. In the frequency band greater than 3 Hz, there are obvious circadian variations in noise level, which is consistent with the law of human activities. The mean diurnal variation of noise levels in southern Ningxia and Yanchi region are 6 dB and 4 dB. The average circadian variation of noise level in northern and central Ningxia is 7 dB. The mean noise level in southern Ningxia is lower than that in other regions. In the frequency bands of 0.1-0.3 Hz, 0.3-1 Hz, 1-10 Hz and 10-35 Hz, the average noise levels in southern Ningxia are 3 dB, 5 dB, 8 dB and 3 dB lower than that in northern region. In the frequency bands of 0.1-0.3 Hz, 0.3-1 Hz and 1-3 Hz, the mean noise levels in central Ningxia are 3 dB, 4 dB and 3 dB lower than that in northern region. In the 3-10 Hz and 10-35 Hz frequency bands, the average noise levels in northern Ningxia are 2 dB and 4 dB lower than that in central Ningxia. In the frequency bands of 0.1-0.3 Hz and 0.3-35 Hz, the mean noise levels in Yanchi region are similar to that in northern and central regions.

We use Baker formula, Kinney formula and Broad formula to calculate the energy released by a conventional test of Liujiaxia air gun active source in Gansu province, and use the empirical formula between energy and magnitude to estimate the earthquake magnitude generated by the test. At the same time, we analyze the single shot data recorded by fixed digital seismic station near Liujiaxia active source test site. The results show that the energy released by air gun during the shot is not completely converted into seismic waves. For example, the main air chamber of air gun is not completely insulated during blasting, there is some energy lost along with the heat transfer, so the magnitude calculated by energy conservation method is slightly larger. The conclusion is in good agreement with the result calculated by Kinney formula, which conforms to the actual working conditions of experiment.