Study On The Spatiotemporal Evolution Of Multi-factors Of Landslides Monitored By GNSS/InSAR Fusio

Geological disasters have brought great threats to the natural environment and human society.Compared with traditional monitoring technology,small baseline integrated aperture radar interferometry(SBAS-In SAR)plays an important role in preventing geological disasters and reducing disaster losses.The xining city is located in the Huangshui valley along the Qinghai-Tibet Plateau and Loess Plateau,with high southeast and low northwest,surrounded by mountains on all sides and striped from east to west,with broken topography,steep slopes and deep ditches,and strong topographic cuts.The deformation rate field and time-series displacement data of Xining City from November 2018 to June2022 were obtained using SBAS-In SAR technology.The ASTGTMV003 digital elevation(30 m)data were used to extract the elevation,slope direction,and slope information in the study area to obtain the slope-oriented deformation rate field in Xining City.The CLCD land use dataset(500 m)and 30 m resolution Landsat 8 OLI remote sensing images were used to extract vegetation cover data and precipitation data from the national surface weather stations to study the spatial and temporal variation characteristics of landslides in areas with frequent surface activity and to investigate the coupling relationship between the spatial and temporal distribution of landslides and environmental factors.Using GNSS data and In SAR data,the tangent angle model is used to analyze and determine the development stage and trend of landslides.Using the existence of a spatio-temporal correlation between GNSS data and In SAR data,the errors in In SAR observations are corrected by GNSS data.Then equations are established through the spatiotemporal correlation between GNSS and In SAR data,and the three-dimensional surface deformation information is obtained by fusing GNSS and In SAR observations based on the Kalman filtering model.The results of the study are as follows.(1)The surface deformation results of Xining City were obtained by using the 103-view Sentinel-1A satellite downlink data released by ESA(ESA)from November 15,2018,to June 3,2022,and processed by SBAS-In SAR technique.The accuracy of In SAR monitoring results was evaluated by GNSS observation data.Among them,the mean absolute error of In SAR monitoring results does not exceed 11 mm and the mean square error does not exceed 13 mm,indicating that the SBAS-In SAR monitoring results meet the accuracy requirements of surface deformation.The overall deformation rate of Xining City is small,concentrated in-14～12mm/a,but the spatial variation is large,and the surface settlement areas are mainly distributed in the ring of mountains such as Huohuegou,Haizigou,the slope belt of North Mountain,and the slope belt of South Mountain,and the surface of the main urban area is more stable.(2)Subject to the error of monitoring equipment,the deformation rate and cumulative displacement curves obtained from GNSS data will have obvious up-and-down oscillations,and the obtained tangent angle is oscillating up and down substantially.After the comparison between the Savitzky-Golay method and the five-point triple smoothing method,it is found that the five-point triple smoothing method shows a better filtering effect than the Savitzky-Golay method in the process of landslide deformation monitoring data filtering,and the corresponding tangent angle value fluctuation also decreases significantly,which achieves more effective data smoothing and makes the landslide deformation monitoring data based on tangent angle warning less likely to(3)Due to the presence of radar in SAR satellites,the tangent angle value of the tangent angle is significantly reduced.(3)Due to the uncertainties of SAR satellites such as radar parameters,phase observations,and terrain data,their errors will be propagated to the deformation results.If directly fused with GNSS data,the results also contain these errors.However,the systematic errors of In SAR have a high spatial correlation with GNSS,which can be weakened by GNSS data.The high-precision GNSS observations are constrained to the In SAR observations,and the difference between GNSS and In SAR observations at the same point is obtained,and the difference is fitted out by the fitting rule,the difference is corrected,and finally,the In SAR observations with higher accuracy are obtained.The Kalman filter is then used to fuse and solve the system-corrected In SAR observations and GNSS observations,and the 3D surface deformation results with higher accuracy are obtained.The experimental results show that the overall mean error of the In SAR observations corrected by the fitted extrapolation method is improved by about 53% and the root mean square error is improved by about 44% compared with the uncorrected In SAR observations.The results of the fusion of the corrected In SAR observations with GNSS observations by the Kalman filter are 45% more accurate than the fusion of the uncorrected In SAR observations with GNSS observations,and 57% more accurate than the In SAR observations.(4)Using SBAS-In SAR technology,the surface slope deformation information of Xining City,Qinghai Province from November 15,2018,to June 3,2022,was extracted and combined with satellite,vegetation,and precipitation data to reveal the distribution pattern of surface subsidence in Xining City and its connection with climate and environmental factors.The results show that the spatial variation of slope subsidence in the study area is large,and it is mainly concentrated on the sunny slopes with slopes ranging from 5° to 35°,among which the average slopes of southwest and south slopes are large and less than-50mm/a are mainly concentrated on the southwest slope.The lower the vegetation cover factor is,the higher the slope deformation rate is,which is not conducive to surface stability;precipitation in Xining is mainly in a decreasing trend from north to south,i.e.,precipitation comes from the direction of the negative slope.The north slope is the slope direction with the largest proportion of precipitation area except for the southwest slope and the south slope.Precipitation and morphological variables of feature points are closely related,and precipitation may cause a short-term increase in morphological variables of feature points,but the overall situation is still decreasing.The threat of features mainly comes from the sunny slope,and the higher slope area is often the place with the largest rate of the landslide area.