联合InSAR与PCA的北京平原地面沉降时空分析

Spatio-Temporal Analysis of Land Subsidence in Beijing Plain Based on InSAR and PCA

20世纪70年代以来,由于地下水超采、可压缩层厚度不均等引起的不均匀地面沉降已逐渐发展成为北京平原最严重的地质灾害之一。目前,北京平原最新时段的地面沉降时空分析的研究报道较少。根据主动微波获取的39景Sentinel-1A光谱影像,使用SBAS-InSAR技术,获取了北京平原2017年5月—2020年5月的地面形变数据;利用主成分分析法对北京平原的地面沉降时空特征进行了分析。在SBAS-InSAR技术处理中设置时间基线120天、空间基线的阈值设置为最大临界基线的45%,生成154个干涉对;对所有干涉对进行配准、干涉,对干涉后的结果进行去平、使用Goldstein算法进行滤波、生成相干系数图并使用最小费用流算法进行相位解缠,筛选73个优质的干涉对进行轨道精炼与重去平,以估算和去除残余相位;通过时间高通滤波、空间低通滤波去除大气相位;最后采用最小二乘法和奇异值分解获取北京平原2017年5月—2020年5月的地面形变数据。在监测时间段内平均沉降速率最大为114.9 mm·yr^(-1),最大累积沉降量为345.9 mm·yr^(-1),均位于朝阳金盏。2019年平原中部沉降速率超过50 mm·yr^(-1)的范围相比2018年有所减小;海淀区、昌平区、大兴区沉降速率超过50 mm·yr^(-1)的影响范围在逐渐增加。利用主成分分析法对北京平原地面沉降进行分析,得出其前三个主成分解释了99.11%的数据特征。第一主成分解释了96.48%的特征,反映了因地下水长期开采引起地面的长期沉降且不可恢复的过程,但南水进京对地下水的补给减缓了平原中部地区的沉降速率;第二主成分解释了2.11%的特征,突出了以年际为周期的沉降过程,该特征与可压缩层厚度和土地利用类型等因素有关;第三主成分解释了数据集0.52%的特征,强调了由降雨调控的季节性弹性形变。研究结果在北京平原地面沉降综合治理方面具有一定的科学价值。

Since the 1970 s,uneven ground subsidence caused by groundwater overdraft and uneven thickness of compressible layers has gradually developed into one of the most serious geological hazards in Beijing Plain.There are few research reports on the temporal and spatial analysis of land subsidence in the latest period in Beijing Plain.Therefore,this paper used SBAS-InSAR technology to obtain the ground deformation data of the Beijing Plain from May 2017 to May 2020 based on 39 Sentinel-1 A spectral images acquired by active microwaves.The principal component analysis method was used to analyse land subsidence’s temporal and spatial characteristics in the Beijing Plain.In the SBAS-InSAR technology processing,the time baseline was set to 120 days,and the threshold of the space baseline was set to 45%of the maximum normal baseline,and 154 interference pairs were generated,and then all interference pairs were registered and interfered.After the interference,the results were flattened,Goldstein flattening for filtering,generated coherence coefficient and used minimum cost flow algorithm for phase unwrapping.The 73 high-quality interference pairs were screened for orbit refinement and re-leveling to estimate and remove the residual phase.The atmospheric phase was removed through temporal high-pass filtering and spatial low-pass filtering.Finally,the least-squares method and singular value decomposition were used to obtain land subsidence data of the Beijing Plain from May 2017 to May 2020.During the monitoring period,the maximum average deformation rate was-114.9 mm·yr^(-1),and the maximum cumulative subsidence was 345.9 mm.which was located in Chaoyang Jinzhan.Compared with 2018,the ranges where the settlement rate of the central plain decreased by exceeding 50 mm·yr^(-1) in 2019 in the Haidian District,Changping district,but the settlement rate range of Daxing District was gradually increasing.It analysed the land subsidence of the Beijing Plain by using the principal component analysis method.It was concluded that the first three principal components explained 99.11% of the data characteristics.The first principal component explained 96.48% of the characteristics,reflecting the long-term and unrecoverable process of groundwater subsidence caused by long-term groundwater extraction.However,the replenishment of groundwater by the South Water entering Beijing slowed down the settlement rate in the central plain.The second principal component explained the 2.11% feature,highlighting the interannual settlement process,which was related to factors such as the thickness of the compressible layer and the type of land use;The third principal component explained 0.52% of the characteristics of the data set and emphasized the seasonal elastic deformation regulated by rainfall.The research results of this paper can provide a certain scientific basis for the comprehensive management of land subsidence in the Beijing Plain.