基于Sentinel-1 SAR数据的黑河上游冻土形变时序InSAR监测

多年冻土活动层变化导致冻土区大范围地面变形,严重破坏区域内基础设施和水文地质条件,亟需加强活动层季节冻融过程的观测研究.本文提出一种基于分布式目标的小基线集时序InSAR(DSs-SBAS)的冻土形变监测方法.该方法采用分布式目标提取和特征值分解算法,并结合基于地温-形变约束关系的参考点选取新策略,提高了冻土形变监测结果的时空分辨率和可靠性.以祁连山黑河西支源头的野牛沟为研究区域,通过对27景Sentinel-1SAR影像进行时序InSAR分析,获取了2014-2016年该区多年冻土的形变时间序列和年均形变速率,并利用Stefan模型联合地温数据估算其季节性形变幅度.实地踏勘和结果分析表明:(1)研究区大部分多年冻土处于稳定状态(-1.0^+1.0cm·a^-1),在地形陡峭的南坡边缘及含冰量丰富的野牛沟河上游两侧沟底部分区域存在较大形变;(2)区域内冻土形变时间序列呈现年周期变化,冻土冻融形变存在季节性周期形变和季节性波动下沉两种形变特征,形变幅度和速率最大可达6.0cm和-3.0cm·a^-1;(3)不同区域的活动层冻结/融化始日和冻土形变存在明显差异,主要和冻土地貌、土壤类型以及活动层厚度有关.本文提出的方法在青藏高原多年冻土区大范围冻融监测和活动层厚度反演研究方面具有很大的应用潜力.

Monitoring permafrost deformation in the upstream Heihe River,Qilian Mountain by using multi-temporal Sentinel-1 InSAR dataset

Changes in active layer thickness (ALT) over permafrost regions lead to a large-scale ground deformation, affecting infrastructure stability and hydrogeological conditions. Therefore, accurate measurements of such deformation are urgently needed for characterizing the seasonal freeze-thaw process. In this paper, we present a time-series InSAR method for monitoring permafrost deformation, named "DSs-SBAS", which combines both cut-edge algorithms for identification of radar distributed scatters (DSs) and eigenvalue-decomposition-based optimization of DS interferometric phases to improve the spatial and temporal resolutions of deformation measurements. Taking into account difficulty in selecting reference points for the InSAR analysis in the permafrost region, we introduce a new strategy based on the relationship between ground temperature and deformation to improve the reliability of the selected reference point. This DSs-SBAS InSAR method is applied to investigate ground deformation over permafrost regions at the upstream west branch of the Heihe River, Qilian Mountain. A total of 27 Sentinel-1 SAR images are employed to derive deformation time series and annual deformation rates between 2014 and 2016. In addition, the Stefan model constrained by the ground temperatures is adopted to calculate the seasonal deformation amplitude. The results indicate that (1) most of the permafrost in the study area was stable (-1.0~+1.0 cm·a^-1), and large deformations occurred in the southern slope and the bottom part of the upper reaches of the Yeniugou River, (2) there are two main temporal processes of deformation with annual cycles, namely seasonal cycle deformation and seasonal subsidence. The largest deformation rate and maximum deformation amplitude were up to 6.0 cm and -3.0 cm·a^-1, respectively, (3) a significant heterogeneous pattern in freeze-thaw cycles and permafrost deformation is found, which is mainly controlled by permafrost landscapes, soil types and ALT. This proposed DSs-SBAS InSAR method has a great potential for monitoring large-scale deformation and retrieving ALT variations over permafrost regions in the Tibetan Plateau.