Multi-Temporal InSAR Technology For Potential Landslide Identification And Deformation Monitoring In The Longyangxia-Liujiaxia Section

The country’s vast land area of 9.6 million square kilometres,its diverse topography and complex geological formations have led to the occurrence of geological disasters,of which earthquakes are one of the most far-reaching.Due to the large number of mountainous areas in China,landslide disasters have become one of the most widespread geological disasters,causing extremely serious economic losses and casualties to the country and its people.Located at the junction of the Loess Plateau and the Qinghai-Tibet Plateau,the middle and upper reaches of the Yellow River are rich in water resources and have a large number of hydropower stations.However,there are numerous landslide developments within the Longyangxia-Liujiaxia section,which not only seriously threatens the lives and property of residents on both sides of the Yellow River,but also affects the safe operation of hydropower stations and the exploitation of water resources.Therefore,the identification and monitoring of potential landslides in this area is crucial to the prevention and management of regional geological hazards.Interferometric Synthetic Aperture Radar(InSAR)is an emerging measurement technology in the field of geological hazard identification and monitoring,which has the advantages of round-the-clock and all-weather that traditional ground monitoring such as GNSS technology and level measurement do not have,and has been widely used in landslide identification and monitoring.However,the practical application of InSAR technology should take into account the effects of various errors,such as out-of-coherence errors and atmospheric errors,and the resulting errors should be corrected or attenuated in the data processing to obtain more accurate monitoring results.In this paper,the Long Yangxia-Liujiaxia section is used as the study area,and the potential landslides in the study area are first identified by combining lift-track Sentinel-1A SAR images to reduce the effect of SAR image geometric distortion.Then IPTA and Stacking InSAR techniques were combined to identify unstable slopes in the key reservoir area-Longyangxia reservoir area in the study area,and the advanced time-series InSAR technique and 2D deformation decomposition technique were used to monitor the Chana landslide with high accuracy.Finally,the deformation characteristics and time series features of the Gob landslide were obtained and analysed.The specific research contents and results of this paper are as follows:(1)The multi-temporal InSAR technique for potential landslide identification and monitoring in the study area was systematically summarised,and various errors in InSAR data processing were analysed,focusing on the atmospheric distribution characteristics of the study area,and three atmospheric error correction methods,namely GAGOS,linear fitting model and quadratic tree image segmentation,were compared and analysed.The results show that the quadratic tree image segmentation algorithm is more effective than the other two methods in correcting the atmospheric errors in the study area.(2)Due to the numerous high and steep slopes in the study area,geometric distortions are highly likely to occur when SAR images are used for monitoring.Therefore,the geometric parameters and topographic factors of the Sentinel-1A SAR images were combined to analyse the visibility of the lift-track Sentinel-1A SAR images covering the study area,and to derive the perspective shrinkage,shadowing and overlay mask areas of the lift-track SAR images in the study area.Finally,the elevation track SAR images were used to identify and catalogue potential landslides in the study area.(3)The Long Yangxia reservoir area is the largest reservoir area in the Yellow River basin in terms of water resources storage,and its near-dam reservoir banks are distributed with numerous dangerous high and steep slopes,which greatly threaten the safe operation of the Long Yangxia hydropower plant.Therefore,joint elevated-track Sentinel-1A SAR imagery and combined improved IPTA algorithm with the traditional Stacking InSAR technique are used to identify unstable slopes in the Longyangxia reservoir area.(4)Taking the Chana landslide in the Longyangxia reservoir area as the research object,the DS-InSAR technique was used to obtain the annual average deformation rate of its elevation track,and the results were compared with those of IPTA to obtain a significant increase in the point density obtained by the DS-InSAR technique compared with those of IPTA,which restored the points that were not selected due to spatial and temporal decoherence in the deformation area and obtained a complete deformation field of the Chana landslide.The 2D deformation decomposition of the obtained lift-track deformation was then carried out to obtain the 2D deformation rate and deformation time series of the Chana landslide,which were analysed to obtain the 2D deformation characteristics of the landslide and the landslide development characteristics.(5)The annual average deformation rate and deformation time series of the Guobu landslide were obtained using the DS-InSAR technique and SBAS-InSAR technique,and their comparative analysis showed that the deformation characteristics obtained by the two monitoring techniques were basically the same.However,the DS-InSAR technique obtained more monitoring points compared with the SBAS-InSAR technique.The deformation results were also verified by using GNSS data to verify the outer conformal accuracy.(6)Wavelet analysis and singular spectrum decomposition(SSA)techniques were used to analyse the time series of landslide deformation obtained,and the correlation and lag between rainfall and reservoir water level on landslide deformation were obtained.