The seismic structure in northeast China is mainly characterized by the basin group and the northeast-trending faults which controlling basin group. The Songliao basin and the northeast-trending fault controlling the basin are the most representative. The basin is also the region with the strongest seismic activity in the region. Most of the Coulomb rupture stresses triggered by major active faults in northeast China after the Tohoku earthquake increased in the northeast direction, from northwest to southeast, and from southwest to northeast, with gradually increasing characteristics; the distribution in the north-west and near east-west direction is mostly decreasing, and the decreasing value from west to east and from northwest to southeast is a gradual increasing process. The Tohoku earthquake in Japan changed the basic movement trend of northeast China and the main island of Japan. The occurrence of the Tohoku earthquake in Japan changed the original squeezed state of northeast China into a stretched state. In the two trends of movement from 2002 to 2018 and from 2006 to 2018, the impact of the Tohoku earthquake on northeast China is still very significant, and the geodynamic environment in northeast China will still be under the influence of the Tohoku earthquake. It will have a certain influence on the rhythmic characteristics of shallow and deep source earthquakes in the relatively regular northeast China. The fifth activity period beginning in 1999 has ended, or entered the intermittent calm period before the sixth activity period, or it will continue for a period of time, and time verification is required.

Combining high-resolution shallow seismic profiles and regional geological data, we analyze and discuss the shallow structural characteristic of Tangxi fault. The results show that the southern part of Tangxi fault is composed of two normal faults with NE strike and SE dip, and the Neogene sediments of down-thrown side of the main fault in Tangyin graben is obviously thickened. Combining with the results of trial trench, the latest active age of south section of Tangxi fault is middle Pleistocene.

In order to analyze the shallow tectonic background, earthquake inoculation mechanism and the relationship between the earthquake and the structure, we use high-resolution gravity anomaly data and gravity 3D inversion algorithm based on block growth mode to analyze the shallow 3D inversion density structure in the earthquake area. We verify the validity and stability of the method based on block growth mode inversion, and retrieve the high-resolution gravity Bouguer anomaly data in the study area. The results show that the Sanhe-Pinggu M8.0 earthquake is located in the low-value part of intersection transition of the local high gravity of Daxing and Sanhe-Mafang and the low gravity of Dachang. The inversion results of the shallow 3D density structure of the study area show that the Sanhe-Pinggu M8.0 earthquake is obviously controlled by the NE-trending Xiadian fault. The density difference on both sides of the fault is obvious and extended down about 10 km. It is estimated that the earthquake occurred at a depth of about 10 km.

Based on mobile gravity observation data from Leiqiong area since 2016, we process the joint observation data in the second half of 2017 by using different adjustment methods. The Zhanjiang and Qiongzhong absolute gravity points are considered as the spatial reference of the Leiqiong survey network, and we study the dynamic changes of gravity field in Leiqiong area since 2016. We focus on the differential and cumulative dynamic changes of regional gravity fields before the 2018-03-20 Yangjiang M4.2, the 2019-03-05 Leizhou M4.1, the 2019-08-20 Sanya M4.2, and the 2019-10-12 Beiliu M5.2 earthquakes. The results show that there are abnormal changes in the gravity field near the epicenter before the earthquakes, except the Leizhou M4.1 earthquake. The results of this study indicate that the regional gravity field may reflect a certain preseismic anomalous of magnitude 4~5 earthquakes with reasonably dense observation networks and proper data processing methods.

Based on the observation data of flowing gravity in the capital area from 2017 to 2019, we analyze the gravity changes at different time scales before the M4.5 earthquake in Tangshan, Hebei province, on December 5, 2019. The results show:1) The image of the variation of the 0.5 a scale gravity field well reflects the evolution process of the earthquake: “in 1 a before the earthquake, the gravity field continued to rise(positive) change-the local reverse(negative) change before the earthquake-the earthquake occurred near the zero contour line of the gravity gradient zone”; 2) The 0.5 a, 1 a and 2 a scale gravity fields before the earthquake, show that the earthquake occurred near the zero contour line of gravity change; 3) The change of gravity observed before the earthquake may be caused by the upwelling process of deep thermal materials.

Based on the relative gravimetric data measured by LCR-G and CG-5 relative gravimeter and hydrological data in Changyi-Xinhe area from 1991 to 2019, we find that: 1) The increasing relative gravimetric data from 2009 to 2018 is mainly caused by the decline of groundwater level in Changyi city, and the maximum effect can reach to 132 μGal. 2) The relative gravity decreased by 72 μGal cumulatively from May 2014 to March 2016, and it recovered rapidly from March 2016 to May 2018. It is not clear whether this phenomenon is related to the dynamics.

Based on 8 glacial isostatic adjustment(GIA) models commonly used in the world, we calculate the global mass change trends from 2005 to 2014 using GRACE data corrected by GIA. When correcting in Antarctic region, the correction values of the three models: ICE-6G_C, ICE-6G_D and IJ05, are similar. When correcting the global ocean mass changes, the mean value of the impact of different correction models on global ocean mass changes is below -2.0 mm/a. We contrast the contribution of land water and ice sheet to the trend of ocean mass changes, and combining satellite altimetry and Argo data to further verify that the Paulson07, Geruo13 and ICE-6G_D models are relatively suitable for global ocean mass correction. Integrating the entire land and ocean, ICE-6G_D model is more suitable for adjusting the global mass change.

The high-precision and high-resolution earth gravity field can be used to unify the global height datum and determine the geoid. So, it is very important to estimate the accuracy of the gravity field model and various gravity quantities (such as height anomaly). Based on VS2015 + Intel XE 2016 platform and OpenMP parallel technologies, we develop EGMTools software for gravity field model evaluation and gravity quantities calculation. The main functions of EGMTools include: 1) estimating the accuracy of the gravity field model truncated to any degree/order by using measured data such as gravity anomaly, height anomaly, deflection of vertical data and so on; 2) calculating gravity anomaly, height anomaly, deflection of vertical data and other gravity quantities of points or grid (SRTM topographic surface) using the gravity field model; 3) comparing the gravity quantities (such as height anomaly) represented by different gravity field models according to a certain grid interval on a global scale; 4) converting gravity field model format, analyzing degree variance and data preprocessing for RTM gravity quantities calculation.

In order to improve the timeliness and accuracy of PM_2.5 concentration prediction, this paper integrates observation factors such as atmospheric pollutants, GNSS PWV and wind speed, and uses the methods of FFT and LSTM neural network to build the PM_2.5 concentration prediction model to predict PM_2.5 concentration in the next 24 hours. Firstly, fast Fourier transform is applied to the observation elements such as air pollutants, GNSS PWV and wind speed, and the common change period of various elements is extracted to obtain the optimal common period of 216 hours. Then, various elements of the optimal common period length are selected as the model input, and the PM_2.5 concentration of 24-hour sequence are taken as the model output. The RBF neural network of PM_2.5 single elements and the LSTM neural network integrating atmospheric pollutants, wind speed and GNSS PWV are respectively used to construct the PM_2.5 concentration prediction model. Finally, the measured PM_2.5 concentration sequence is used to test the external reliability of the two models,RMSE and IA are used as evaluation indexes to evaluate the model accuracy. The results show that the RMSE and IA tested by the PM_2.5 concentration prediction model based on FFT-LSTM are 16.22 μg /m³ and 84.36%, respectively. The prediction accuracy of the model could effectively predict the PM_2.5 concentration in the next 24 hours. The model can be used as a reference of air quality prediction for air pollution prevention department.

Multipath error is the main factor affecting the measurement accuracy of GNSS and the difficulty of error correction. Combining the advantages of empirical wavelet transform(EWT) algorithm in the decomposition of complex signals with the daily repetition characteristics of GNSS multipath error, we propose a new method of multipath error extraction and correction based on EWT. The screening coefficient is defined by combining the energy entropy with the correlation coefficient, and the points with significant changes are taken as the dividing points of the noisy components and the information components. A new denoising method based on EWT is established by denoising the noisy components and reconstructing the information components. Through the analysis of simulation and experiment data, it is shown that EWT can extract and correct the GNSS multipath error better than EMD and wavelet method.

We introduce the residual-based RAIM model and protection level calculation method based on the single-fault assumption, analyze the relationship between the beam angle and defilade angle of LEO satellites, and propose the improvement value and improvement ratio of protection level as the integrity enhancement index of LEO constellations. Based on the simulation data of BDS and LEO constellation, we analyze the influence of defilade angle and equivalent error level on protection level and RAIM availability. The results show that the lower the defilade angle and equivalent error level, the better the enhancement effect on protection level and RAIM availability.

Based on conventional space radiometer method, we integrate the vertical stratification effect of water vapor in troposphere into the correction model, and propose an improved InSAR atmospheric phase delay correction method. To validate the feasibility of the proposed method, the MERIS near-infrared water vapor product is used to reduce the atmospheric phase delay in ground subsidence InSAR monitoring in Beijing. The deformation monitoring results of GNSS stations of crustal movement observation network of China (CMONOC) are used as benchmarks to verify the monitoring accuracy of the improved atmospheric correction method. The results show that RMSE of the improved atmospheric correction method, space-based radiometer method and uncorrected InSAR monitoring results are 0.388 cm, 0.603 cm and 0.685 cm, respectively, compared with the monitoring results of CMONOC GNSS stations, which shows that the improved atmospheric correction method is with higher monitoring accuracy than uncorrected and conventional method, and can effectively reduce the atmospheric phase delay errors in SAR interferograms.

We use measured surface temperature data to calculate the vertical deformation effect of temperature change on pillar and bedrock of GNSS reference station. Combining with the atmospheric load change, the influence of temperature change on vertical deformation of reference station is comprehensive analyzed. We use Xinjiang GNSS continuously operating reference station data as an example, the results show that the annual impact of temperature change on Xinjiang GNSS continuously operating reference station is 1.4－2.1 mm, and the half-annual amplitude is 0.01－0.33 mm. The annual amplitude accounts for the majority. When the vertical time series change minus the effect of temperature change, the average RMS value is reduced by 0.55 mm, and the accuracy is improved by about 9%. The effect of temperature change and atmospheric load on the vertical deformation is basically same, showing an increase in the first half of the year and a decrease in second half. At the same time, when the two effects are deducted, the accuracy is improved by about 13%. The research in this paper shows that the changes in temperature and atmospheric load are very important factors for the periodic displacement of reference station.

Due to the sensor limits and effect of noise and atmospheric delay, the maximum deformation gradient detected by InSAR technology has a limit. In order to accurately predict the detection ability of InSAR technology in deformation monitoring of mining area, this paper establishes the dynamic detection function model of maximum deformation gradient of InSAR based on the geological mining data of 65 mining areas, combined with Knothe time function and probability integral model, using TerraSAR data of different spatial resolutions. The model can use existing geological mining data to judge the detectable range of InSAR in different mining stages without SAR image in advance. In this paper, taking Daliuta mining area in Shanxi Province as an example, seven TerraSAR data from November 21, 2012 to January 26, 2013 are used for experimental verification. The results show that the model can accurately predict the detectable range and determine the detectable boundary of InSAR technology in different mining stages of mining area,which can provide important reference for the application of InSAR technology in mining area.

Based on data of three GNSS observation stations in Shaanxi from January to June in 2018, the water vapor inversion is carried out by using Beidou satellite navigation system. We use different ephemeris products to calculate the water vapor, and compare them with the results calculated by IGF and radiosonde data. At the same time, we compare the water vapor results obtained by different ephemeris with each other. The results show that the three ephemeris can obtain high-precision and high-reliability water vapor results. The accuracy of water vapor inversion with precision ephemeris and fast ephemeris are equivalent, the deviation of each station is better than 1 mm, the standard deviation and root mean square error are better than 3.5 mm. The ultra-fast ephemeris (forecast part) is used to invert water vapor at each station with a deviation of about 1 mm, the standard deviation and root mean square error are better than 5 mm. Combining the corresponding accuracy of water vapor inversion of three products, it can be obtained that the three ephemeris products can invert precipitable water vapor and the water vapor value calculated by the ultra-fast ephemeris (forecast part) can provide a reference for weather forecast.

In order to conduct a more detailed analysis of the length-of-day (LOD) interannual variation, we use the normal Morlet wavelet transform method to identify and extract six main interannual signals from the LOD time series, which are 2.3 a (2.4 a), 3.3 a, 3.7 a, 4.8 a, 6.1 a and 8.1 a signals, respectively. Then, based on the time-domain extracted result, we acquire corresponding average amplitudes, which are 0.08 ms, 0.05 ms, 0.05 ms, 0.07 ms, 0.10 ms and 0.07 ms, respectively. Refer to the method of extracting interannual signals of the LOD, we extract the corresponding signals in the atmospheric angular momentum series, and perform correlation analysis between them. The results show that the atmosphere is closely related to four high frequency interannual signals (2.3 a (2.4 a), 3.3 a, 3.7 a and 4.8 a) corresponding to the LOD, and the correlation coefficients are 0.99, 0.93, 0.99, 0.91, respectively. The contribution rates of the atmosphere to the 2.3 a (2.4 a), 3.3 a, 3.7 a and 4.8 a signals of LOD are about 99.7%, 63.1%, 94.7% and 69.3%, which indicates that 2.3 a (2.4 a) and 3.7 a signals of the LOD can be almost completely explained by atmosphere, and the other two signals are also mainly affected by atmosphere. The 6.0 a and 8.5 a atmospheric signals are irrelevant or weakly related with the 6.1 a and 8.1 a signals corresponding to the LOD, and the correlation coefficients are -0.11 and -0.56, respectively.

In order to study the uncertainty of single point positioning deviation caused by the spatiotemporal difference of tropospheric delays, firstly, we use the IGS ZPD product to analyze the correlation between the maximum, minimum, mean, STD and spatial distribution of the station. The results show that the ZPD average is around 2.4 m. There is an overall tendency to decrease with increasing latitude. However, it is not completely symmetrical along with the equator, and the dispersion is slightly larger in the northern hemisphere. Then, based on single point positioning model, the influence formula of tropospheric delay on the calculation of positioning parameters is derived, and its influence on single point positioning is evaluated. The results show that the positioning deviations caused by tropospheric delay are the largest in the U direction, reaching 7 to 15 m; the most insignificant deviations are shown in the E direction, within ±0.2 m; and the deviations in the N direction are centered within ±0.6 m.

We design a circuit for borehole seismometer to deal with azimuth deviation. This device can correct the output of seismometer before recording, so that the data recorded by data collector is the right data of the required azimuth. The results of Simulink simulation tool based on MATLAB show that the scheme can be applied to correct azimuth deviation of borehole seismometers, and is with high feasibility.

Based on water level data from the deep well, the flow data of Longzici spring, and water level data of some long-term observation holes in the structural area of Linfen seismic station, we analyze the overall change of groundwater state at Linfen seismic station and its surrounding areas. We further analyze the short level anomalies at Linfen seismic station in recent years according to the characteristics of regional water level change. The results show that the groundwater level of Linfen short-leveling site and surrounding areas changes significantly from 2011 to 2013, which is consistent with the change of the leveling observation curve. The time of leveling anomalies lags behind the change of water level by 2-3 a. The rising water level in a large range, especially in shallow water wells, which is caused by the decrease of groundwater exploitation not by structure. The ground rebound caused by water level rise accords with the principle of Terzaghi effective stress; that is to say, the water level rise recovers the pore-water pressure, the sand layer rebound and then causes the ground rebound, which further causes the change of the short-leveling height difference measurement value.