Aiming at the problem that GPS elevation time series is affected by various types of noise, which makes it difficult to extract useful information, we propose a threshold denoising method based on ensemble empirical mode decomposition (EEMD) and multi-scale permutation entropy (MPE). The method uses EEMD as the core algorithm. EEMD can decompose the original signal into a series of intrinsic modal function (IMF), use MPE as an indicator to classify it into noise IMF, hybrid IMF and information IMF, and then use threshold function to process the mixed IMF to achieve secondary noise reduction. Finally, the noise-reduced data and information IMF are reconstructed to obtain noise reduction results. By analyzing the simulation signals and examples, the results show that compared with the correlation coefficient method and the MPE method, the noise reduction evaluation indexes RMSE, SNR, and d_nSNR are all optimal, indicating that the new method has the best noise reduction effect. It further demonstrates that the noise reduction results obtained by the new method can better reflect the nonlinear variation characteristics of the time series itself, and can provide a reliable basis for GPS elevation time series analysis.

In view of the characteristics of GNSS elevation coordinate time series, including non-stationarity and nonlinearity, based on the in-depth analysis of the characteristics of the Prophet model and random forest(RF), we construct the Prophet-RF combination forecasting model. The combination model solves the defect of weak predictive ability of Prophet model to the nonlinear part of time series, and has strong robustness. In this paper, the continuous observation data in the elevation direction of BJFS station is selected for analysis, and a variety of combination schemes are designed to test the applicability and accuracy of the combination model. The experimental results show that the Prophet-RF composite model is better than the single Prophet model in representing change trends of the time series of the elevation coordinates, and in getting more accurate prediction data.

In this paper, the data of 12 individual MGEX stations in the Asia Pacific region are selected to analyze and compare the positioning accuracy and convergence of BDS-2 and BDS-2+3 kinematic PPP. The results show that the convergence speed and positioning accuracy of dynamic PPP is significantly improved after the addition of BDS-3, but the convergence speed is still worse than that of GPS due to the limitation of BDS-3 precision orbit and clock products.

We comprehensively evaluate the RTPPP service performance of three widely used GNSS open source software programs: BNC, RTKLIB and PPPWIZARD. We select nine IGS-MGEX observation stations evenly distributed all over the world to analyze and compare the integrity rates of data streams from real-time GPS, GLONASS and GALILEO as received by the three software programs. The convergence speed, positioning accuracy and data efficiency of the different software programs are evaluated. The results of numerical examples show that: 1) The communication stability of the three software programs is high and the real-time data integrity rate is basically above 96%. In comparison, the real-time data integrity rate of PPPWIZARD is the best, BNC takes the second place, while RTKLIB is relatively poor. 2) Using the data of the triple systems, BNC, RTKLIB and PPPWIZARD software programs converge to three-dimensional point position with precision better than 0.5m at about 14 min, 82 min and 12 min. 3) In terms of the accuracy of real-time positioning, the accuracy of BNC, RTKLIB and PPPWIZARD software in the horizontal and vertical directions are respectively at about 9 cm and 12 cm, 16 cm and 20 cm, 7 cm and 12 cm, in which PPPWIZARD adopts ambiguity fixed solution. 4) The data efficiency index of the three software programs do not share very much in common, among which BNC and PPPWIZARD data efficiency reach about 95%, and RTKLIB data efficiency is lower than 73%. In general, the service performance indicators of BNC and PPPWIZARD's real-time PPP are basically the same, while RTKLIB results are relatively poorer.

The Beidou ionosphere model includes the Beidou system modified Klobuchar model (BDSKlob), the Beidou global broadcast ionospheric delay correction model (BDGIM) and a wide area differential grid ionosphere model. We analyze in detail the delay jump caused by updating parameters of two ionospheric models. Based on code grid ionosphere data, we compare the changes of two ionosphere models with space and time. We analyze the correction and positioning accuracy of different ionosphere models at different latitudes in different periods. We evaluate the service scope and service accuracy of ionosphere information of Beidou grid network in 2018. The results show that, compared with the BDSKlob ionospheric model, BDGIM and CODE GIM are similar in accuracy and distribution; the delay jump of parameter updating of two adjacent models is less than 1 ns. The accuracy of BDGIM model in the northern hemisphere is better than that in the southern hemisphere. The correction rate of BDGIM in China is more than 75%, and the global correction rate is 60%, which is better than the BDSKlob ionospheric model. The positioning accuracy of BDGIM can be improved by up to 45% compared with the BDSKlob model. The effective ionosphere area of Beidou grid basically covers China, the ionosphere delay error is better than 2.08 TECu, and the correction ratio is better than 80%.

In this paper, observation methods applying the GNSS-R soil moisture detection technology and the applied principles are systematically reviewed. Current research achievements are described, indicating shortcomings in the current research status and future directions of GNSS-R development.

In this paper, for highly undulating terrain in western China, we analyze the relationships between T_m and height, T_m and surface temperature in western China using radiosonde data from 2014 to 2016. We establish a new T_m model related to the surface temperature, height and season variations based on Bevis formula. Then, we assess the new T_m model by comparing with the widely used Bevis formula and GPT2w model using 2017 radiosonde data as reference values. The results show that the annual bias and RMS error of the new T_m model are -0.08 K and 3.89 K respectively, and the RMS of the new T_m model has decreased by approximately 14.3%, 20.6% and 9.3% against Bevis model, GPT2w-5 and GPT2w-1, respectively. In addition, the new T_m model has RMS_PWV and RMS_PWV/PWV values of 0.22 mm and 1.43% when used to estimate GNSS-PWV. Therefore, the new model has critical applications in GNSS-PWV remote sensing in western China.

Considering the high false forecast rate and the low accuracy of the existing method, we introduce a variety of meteorological parameters(temperature, relative humidity, dew point temperature, surface pressure) and time-related parameters (day of year, hour of day) correlated to the rainfall process. We use meteorological parameters and rainfall data of two GNSS meteorological stations in Singapore over the period of 2010-2012 to study the time-varying characteristics between meteorological parameters and rainfall data. We find that the meteorological parameters show abnormal variation trends before rainfall, and all kinds of meteorological parameters show weak correlation with rainfall. Based on the above discovery, we firstly forecast short-term rainfall using the least squares support vector machine (LS-SVM). This method takes the meteorology and time parameters as the model input, rainfall data as the model output, and uses the true forecast rate (TFR) and false forecast rate (FFR) to evaluate the accuracy of LS-SVM model. The results indicate that the proposed algorithm can forecast 99% rainfall events, and the FFR is 40%. Compared with the existed least square rainfall forecast model, the FFR of this paper is reduced by nearly 20%, and TFR is increased by nearly 10%.

Based on the general formula of generalized inverse matrix and Kelly inverse four block matrix, we analyze the existence conditions of a special kind of Kelly inverse four block matrix with a rank defective sub-block matrix in the upper left corner and a zero sub-block matrix in the lower right corner. The explicit expression formula of a Kelly inverse matrix of this kind is derived using the generalized inverting matrix method and the matrix transformation method, and is applied to the parameter estimation of rank-defective indirect adjustment model with constraints. The results show that when existence conditions are satisfied, the solution of rank-defective indirect adjustment model with constraints by inverting block matrix is consistent with that of the indirect adjustment model. This proves that the derived explicit expression formula is feasible.

In this paper, we introduce the deep fully connected neural network into the field of dam deformation prediction. Based on the training samples of dam multi-source monitoring data, we establish a dam deformation prediction model based on deep fully connected neural network. We use several common deep optimization learning algorithms to optimize the model training, select the optimal learning algorithm by comparing the change curves of each loss function, and further construct a deep fully connected neural network dam deformation prediction model based on the optimal learning algorithm. Finally, we test and analyze the model in combination with the test samples of the dam multi-source monitoring data, and compare the prediction results with those of the traditional BP neural network. The research shows that the deep fully connected neural network in this paper has high prediction accuracy and strong practicability. It can provide a certain reference value for dam safety monitoring.

Using gCAP method based on waveform data of digital seismic network, we study the focal mechanism solution of 2020-01-23 Shiqu M_S4.3 earthquake in Sichuan province. The result shows that strike/dip/slip of nodal plane Ⅰ are 134°/82°/11° and those of nodal plane Ⅱ are 42°/79°/171°, the centroid depth is 9 km and the moment magnitude is M_W4.53. The crustal velocity structure, positioning error and data quality are analyzed to test the stability and reliability of the focal mechanism solution. The results show that these factors have little influence on the inversion results. The aftershock sequences are extended along a NW direction, located in the northeast area of main shock, dipping to northeast. According to geological structure characteristics of the source region, the focal mechanism solution of main shock and space distribution of aftershock sequences, we conclude that nodal plane Ⅰ is inferred to the fault plane and the seismogenic fault is Changshagongma fault. The Shiqu M_S4.3 earthquake is a left-lateral strike-slip earthquake event.

We take the data of local earthquakes in Jiangsu-Anhui section and adjacent area of Tan-Lu fault zone in recent 50 years as data samples. Based on the quantitative analysis of the minimum complete magnitude(M_C) of this area, and the systematically derived principle of the maximum likelihood method for the calculation of b-value, this paper focuses on the study of the G-R relationship in the area, especially the evolution characteristics of the temporal and spatial distribution of the b-value. According to the indication of the b-value, the correlation between the b-value and earthquakes with M_L≥5 in the area, and the distribution of the crustal tectonic stress indicated by the b-value, are analyzed. According to the characteristics of the temporal distribution of b-value, the period of occurrence of moderate and strong earthquakes obviously coincides with the period of medium and short-term decline of b-value, and the correlation rate is over 80%. The spatial distribution of b-value indicates that the crustal stress of the whole north Jiangsu basin, the south of the basin and the southern margin of Hefei basin, is relatively high; it is consistent with the distribution of UHP metamorphic belt in the region. Combined with relevant geological data, we infer that the region is in the stage of earthquake preparation.

We use the shear-wave splitting (SAM) method to calculate the fast shear-waves polarization direction and slow shear-waves time delay of 8 stations in southern Liaoning, based on near-field small seismic waveform data of seismic network from 2008-01 to 2018-12 in Liaoning Region. We mainly analyze the shear-wave splitting parameters of Yingkou (YKO) station and Gaixian (GAX) station. The results show that: 1) The dominant polarization directions of the fast shear-waves at the 8 stations are mainly NE-NEE and NW. Among them, the fast wave polarization directions of 5 stations are NE and NEE, which are basically consistent with the regional geological structure and the direction of the maximum principal stress field in northern North China. 2) The detailed analysis of the shear-wave splitting results of YKO and GAX stations show that the dominant polarization directions of the shear-waves of YKO station have basically been consistent since 2012, indicating that the tectonic stress field near YKO station is relatively stable or in the process of uniform stress loading; the evolution of shear-wave splitting parameters of GAX station over time is more complicated, and its fast-wave polarization direction has changed significantly before and after the small earthquake intensification in 2013. The time delay is larger than other stations, which may reflect that the crustal anisotropy below GAX station is stronger than other regions. That can provide a theoretical basis for further understanding of regional stress field characteristics and tectonic environment.

We use wavelet multi-scale analysis method to separate the gravity field in Jiangsu coastal area, and obtain the details of multi-order gravity wavelet. We obtain the spatial distribution of depressions and uplifts in different depths by analyzing the details of different order wavelets, and then provide constraints on the orientation and tendency of seven deep faults in the study area. We find that the orientation and tendency of most of the faults in the north of Xinghua-Dafeng line tend to be NNE and NW, respectively, with steep slopes; while in the south of the Xinghua-Dafeng line, the orientation and tendency of most of the faults are NEE and SE, respectively, and with relatively small slopes. We suggest that the difference patterns between these faults may be a reflection of the ramp structure of the lower Yangtze. After a combination with seismic observations, we find that the distribution of earthquakes is closely related with faults with NE-NEE, NW and nearly EW orientations. Therefore, the faults in the northwest of Donghai-Ganyu and Dafeng harbor exhibit strong activities and unstable stress field, and we speculate there is a higher risk of earthquakes in this area.

Based on the computer algebraic system, we develop the normal gravity formula in series. Then, the function formula of the geocentric latitude at the average value of the normal gravity with respect to the first eccentricity of the ellipsoid is deduced. In addition, it is considered that the formula can be extended to the planets in the solar system; when the first eccentricity of the ellipsoid is assumed to be a variable, we deduce the function formula of the geocentric latitude at the average value of the normal gravity with respect to the first eccentricity of the ellipsoid.

We estimated the terrestrial water storage(TWS) variation of Yangtze from CSR and JPL Mascon solutions, CSR and JPL spherical harmonic (SH) solutions and level-3 products and then calculated the uncertainties of these solutions by the generalized three-cornered hat method. Then evaluated the precipitation variation based on global precipitation measurement (GPM) product. This study also introduced water storage deficit index (WSDI) to analyse the spatial and temporal feature of drought events of Yangtze, compared with scPDSI (self-calibrated Palmer drought severity index). The results showed that: Mascon solution’s uncertainty is less than SH and the uncertainty of JPL_M, CSR_M, CSR_L3, JPL_L3, CSR_SH and JPL_SH is 3.51 mm, 3.78 mm, 5.45 mm, 9.87 mm, 9.12 mm, 10.71 mm, respectively; Precipitation, which shows significant seasonal variation as well as TWS, with correlation coefficients 0.67, is the important trigger of TWS change in Yangtze; We detected six drought events in Yangtze based on WSDI from 2003-01 to 2014-12, with the longest duration of 24 month starting from 2003-02. 2006 summer and 2011 spring experienced the severest droughts with intensity of 2.15 and 1.97, respectively . Spatial distribution of mean WSDI in Yangtze indicated that the drought, starting from 2006-07 to 2007-06, almost hit all the region of Yangtze basin while the drought that began in 2011-03 to 2011-11 appeared in downstream, midstream and northwest upstream of Yangtze basin.

Based on GRACE gravity satellite data, in this paper we adopt the improved combination Gaussian filtering method and scale factor method to improve data accuracy. The water storage changes of Guizhou from August 2003 to July 2013 are retrieved using the theory of terrestrial storage changes; the relative water storage index is calculated as the drought index. Meanwhile, drought duration and intensity are identified by Run theory, so as to monitor and analyze the drought characteristics of Guizhou. We found that the relative water storage index obtained from GRACE satellite data is basically consistent with the trend of SPI3. Besides, GRACE shows that drought events occurred in Guizhou every year from August 2003 to July 2013, with the shortest lasting for a month, and nine drought events lasting more than or equal to 3 months. Years with high drought intensity occurred in 2003-2006, 2009-2011. The loss of water storage in 2006 and 2011 were especially more serious, which is consistent with the actual drought situation in Guizhou for many years.

This paper uses various methods such as seismicity, surface seismic geological survey, shallow seismic exploration, high-density electrical method exploration and trench excavation to investigate the Zhazi’ao fault. Through investigation, the Zhazi’ao fault is developed in the Paleozoic Permian sandstone and Mesozoic Triassic limestone and sandstone formations. It is a reverse fault. The fault is generally NNE-oriented and the length is about 15 km. The width of the fracture zone is generally less than 20 m, and the fault scale is small. Combined with shallow geophysical exploration and trenching profiles, it is preliminarily presumed that the Zhazi’ao fault does present late Pliocene fault activity, and is the seismogenic structure of the Junlian M4.9 earthquake on January 28, 2017.

The mechanical principle of volume borehole strain gauge is simple and reliable data can be obtained easily, but it is limited by its size. The diameter of existing TJ-2 body strain installation hole must be greater than 130mm, and the drilling cost is high. In order to reduce drilling expenses, we research the miniaturization of volume borehole strain gauge and calculate the sensitivity coefficient and apparent compression modulus of volume borehole strain gauge. According to the calculation results and the material technology, we develop a strain gauge which is suitable for drilling holes with a diameter of 110 mm.

Taking high-speed railway exploration in Harbin as an example, we use the spatial autocorrelation method to extract microtremor signal dispersion characteristics, then obtain shear wave velocity structure of the underground 150 m medium and its layering features. The results show that the microtremor method can penetrate thicker overburden to detect the bedrock surface, and the formation interface is consistent with the borehole data. This method has convenient construction and strong applicability. Combining it with drilling data, we can accurately describe the underground geological horizon distribution and reduce the drilling workload. It saves economic costs and has high application value and scientific significance.