Evolution of spatiotemporal ground deformation over 30 years in Xi’an, China, with multi-sensor SAR interferometry

Abstract

Since the 1960s, Xi'an, China, has suffered from severe ground subsidence due to groundwater extraction and continuous urban construction. During the urbanization process, ground subsidence has threatened and will continue to threaten the stability of the city's infrastructure. Additionally, parts of Xi'an have recently experienced a period of rebound. In this context, we must monitor the spatial and long-term evolution of deformation in Xi'an city. First, archived SAR datasets from 1992 to the present, including ERS1/2, Envisat/ASAR, TerraSAR-X(TSX), ALSO/PALSAR-2(ALOS-2), and Sentinel-1(S1), provided independent historical deformation data in Xi’an city; different SAR interferometry methods can obtain the deformation rates and deformation time-series of the region. Second, to fuse the deformation time-series acquired by different sensors, the Quantile-Quantile Adjustment (QQA) algorithm was used to generate the long-term deformation time-series from October 2003 to May 2021, which has a higher precision than when using the Singular Value Decomposition (SVD) method. Quantitatively, the standard deviation (STD) of the QQA method was only 22 mm at maximum, considerably smaller than the 54 mm obtained with the SVD method. Third, ERS1/2 was linked by the segmental interpolation method to obtain the 30-year deformation time-series of Xi'an, which showed that ground deformation in Xi'an from 1992 to 2021 has experienced a “deceleration-acceleration-deceleration-rebound” process, with a maximum cumulative subsidence of over 2 m. Finally, land use type data and hydraulic head change data were also collected to analyze the deformation information. At the same time, the elastic and inelastic water storage coefficients of the aquifer were calculated by combining the long-term deformation time-series and groundwater head data; the deformation mechanism of the aquifer was analyzed accordingly.