| Xiaojiang River Basin is a typical development area of landslide and debris flow disasters in China.Landslide disaster not only destroys the integrity of local geomorphic units,but also leads to the fragmentation of surface landscape ecological pattern,causes a series of ecological and environmental problems,and restricts the sustainable development of local economy and society.Early identification,hidden danger investigation and risk analysis of potential landslide areas are the key links of prevention and control of landslide and debris flow disaster in local areas,which are of great significance to protect people’s lives and property and ecological environment safety.Surface deformation is the most intuitive performance of monitoring landslide occurrence and one of the preferred indicators to identify potential landslide disaster.It is well-known that Time Series In SAR(TS-In SAR)is an advanced active technology for surface deformation monitoring,with the advantages of wide coverage,high spatial resolution,as well as working in all day and all weather.TS-In SAR is able to obtain the millimeter-scale deformation information of ground targets,and retrieve the slow evolution of surface physical phenomenons,which make it a powerful means for monitoring wide-area surface deformation.However,TS-In SAR technology can only directly detect the surface displacement signal in Line of Sight(LOS),which makes it difficult to interpret accurately the actual evolution process of surface physical phenomena in mountainous areas with complex terrain.How to improve the efficiency and accuracy of RS-In SAR to obtain surface deformation characteristics in complex terrain areas,and establish an early identification and analysis system of slow-growing landslides is still at the exploratory stage.Aim at the above problems,this paper took the Xiaojiang River Basin as the research area.The surface deformation detection,potential landslides’early identification,stability analysis and susceptibility evaluation carried out by using TS-In SAR method.The main research results obtained in this paper are as follows:(1)The surface’s visibility of Xiaojiang River Basin in sentinel-1 satellite images was be assessed,and the feasibility of using multi-orbit SAR data to detect the surface deformation of research area was be proved.The sentinel-1 Satellite SAR images covering the Xiaojiang River Basin were collected,including 100 abcending data and 92 descending data.Based on the satellite flight parameters and terrain parameters,surface terrain visibility of the research area under different flight modes were obtained by mountain shadow calculation model and R-index model.The results showed that ascending data have better visibility on west-facing slope,accounting for 48.37%,while descending data have better visibility on east-facing slope,accounting for 36.31%because of the smaller incident angle.Shadow distortion occurred on some areas during imaging,respectively accounting for 3.86%in ascending data and 8.72%in descending data.In order to maximize the good visibility area,based on the spatial distribution complementarity of good visibility area in the two data sets,the research mode of detecting surface deformation combined with near-synchronous ascending and descending orbit data was proposed in the paper.It showed that,with the combination operation,good visible surface’area has been greatly increased,while the shadow area reduced to only 0.55%,and proved that surface deformation dectection in Xiaojiang River Basin with multi-orbit Sentinel-1 SAR data was feasible.(2)PS feature points were integrated into the framework of SBAS-In SAR technology,and the surface deformation characteristics of research area were detected and analyzed.By analyzing the problems faced by the application of traditional TS-In SAR in Xiaojiang basin,an improved method of introducing PS feature points into SBAS-In SAR workflow as stable ground feature points(PS-GCP)was proposed.The PS-GCP obtained based on the coherence threshold and ADI threshold could effectively reduce or avoid the human error introduced by the traditional manual point selection method,ensure the high coherence and signal-to-noise ratio of GCP points,and take into account the phase stability.By using the improved method,the surface deformation of the study area was detected,and the prominent deformation in Taipingcun gully,Jiangjia gully,Shawandagou gully,Jianshan gully and other areas are identified.Because the deformation monitoring results in LOS direction could not be directly used to describe the actual displacement of the surface,in the paper,the projection components of the surface deformation in the vertical and east-west directions were recalculated by establishing a two-dimensional deformation projection model,and the vertical deformation was used to describe the degree of surface deformation.It could be seen from the experimental results that many prominent subsidence centers have been formed in the study area.The slope deformation degree of typical debris flow gullies was higher,such as Jiangjia gully,Dabaini gully,Xiaobaini gully and Shawandagou gully,and the average annual subsidence velocity of Duozhao gully’south bank was the fastest,reaching-106.94 mm/yr.Based on the multi-dimensional deformation monitoring results,the temporal and spatial movement characteristics of the accumulation migration of slopes in Xiaobaini gully were deeply analyzed,and the migration earthwork was estimated.Meanwhile,the significance of multi-dimensional monitoring based on time series In SAR Technology in surface deformation and landslide disaster identification was emphasized.(3)The early identification and stability dynamic analysis of potential landslide were carried out by integrating the deformation monitoring results of sentinel-1abcending and descending.Based on the analysis of the topographic characteristics of the study area by ALOS DEM,the average annual sliding velocity of the slope was calculated through a down-slope deformation projection model,and a slope classification method was established.According to the terrain and geomorphic features,743 potential landslides were identified in the study area,covering an area of 433.51 km~2.Among them,L-III landslide covered an area of 22.50 km~2,accounting for 5.19%,and the fastest sliding velocity reached-391.40 mm/yr.However,L-III landslide mainly distributed in the typical debris flow gullies or landslide areas,such as Jiangjia gully,Nuole landslide,Maojiadagou gully,etc.Compared with landslide inventory,59 records were identified again,and 684potential landslides were updated.According to the sliding velocity of potential landslides in different periods,the dynamic characteristics of landslide stability in the study area were analyzed.The results showed that the potential landslide activity increasing,stability decreasing and instability deformation mainly occurred in the annual rainfall concentration period(from April to October),and the fastest stage average sliding rate could reach-878.23 mm/yr(Nuole landslide).When lack of rainfall(from October to April of the next year),at least 60%of the landslides were restored to be stable,and only some areas,such as nuole landslide and Jiangjiagou gully,still maintained a high-intensity active state.(4)A coupling model of TS-In SAR and LR model was established to evaluate and analyze the landslide susceptibility in Xiaojiang River Basin.In order to avoid the classification accuracy of noise factor reduction model for susceptibility evaluation,in the paper,CF model was firstly used to analyze the influence weight and contribution rate of disaster-causing factors,and verify the independence,excluding slope direction,land cover type,road distance,curvature and profile curvature.Based on the LR model,an In SAR-LR coupling model was established,and the slope sliding velocity obtained by TS-In SAR was introduced as the error correction parameter to carry out the landslide susceptibility evaluation of Xiaojiang basin.The susceptibility evaluation accuracy of the two models was tested by means of accuracy statistics,frequency ratio and ROC curve.The results showed that the frequency ratio of In SAR-LR coupling model in very low prone area and low prone area were 0.65 and 0.83,which slightly higher than 0.55 and 0.82 of LR model.In the extremely high prone area,the landslide rate and frequency ratio of In SAR-LR coupling model reached 21.70 and 4.20respectively,which significantly higher than 17.63 and 3.41 of LR model.According to the accuracy test results of ROC curve,the AUC indexes of In SAR-LR coupling model in sample training and evaluation analysis were 0.827 and 0.819 respectively,which higher than 0.783 and 0.769 of LR model.The research showed that the In SAR-LR coupling model could more accurately fit the nonlinear response relationship of landslide susceptibility to its influencing factors.Especially in the extremely high landslide prone area,it showed better prediction ability and classification accuracy,effectively reduces the misclassification caused by false positive and false negative errors.At the same time,the significance of TS-In SAR technology in landslide susceptibility evaluation was highlighted.(5)Combined with the multi-dimensional deformation monitoring results of TS-In SAR,the temporal and spatial characteristics of surface deformation in typical areas of Xiaojiang River Basin were deeply analyzed.According to the multi-dimensional monitoring results of TS-In SAR,the temporal and spatial evolution characteristics of surface deformation in Jiangjiagou gully,Shawandagou gully and Dongchuan urban area were deeply analyzed.The results show that:a)The landslides in Jiangjiagou guuly mainly developped on Xiaowanzigou ditch,Mayiping ditch and Wangjiayakou ditch in the resource area,as well as Menqiangou ditch,Duozhaogou ditch and Daaozi ditch in the circulation area.There was a widest instability range on the northern slope of Menqian ditch,which had the hightest sliding strength and fastest sliding velocity.The fastest sliding velocity reached-347.52 mm/yr.b)NL2 sliding zone of Nuole landslide was the most prominent deformation zone in Shawandagou gully.From the rear edge to the main gully,two obvious uplift areas and three subsidence areas have been developed,with the fastest subsidence velocity of-347.54 mm/yr.The deformation intensity gradually decreased with the decreasing of elevation along the main sliding direction of NL2 landslide.c)The surface of Dongchuan City was relatively stable,and the deformation area was mainly distributed along the east bank of Xiaojiang River and’the five gullies’.The deformation intensity varied alternately in dry and wet seasons,indicating that rainfall and fluvial abrasion were the main causes of deformation in these areas. |
PDF Full Text Request