Search

Land subsidence can be observed with time-series of Interferometric Synthetic Aperture Radar (InSAR) data. However, existing approaches only reveal subsidence signals that are multi-scale mixed, which is not conducive to the systematic analysis of subsidence of different mechanisms. A deformation signal decomposition (DSD) method based on spectral analysis is used to decompose the deformation extracted by time-series InSAR into three classes of deformation signals. They refer to large-scale deformation related to geological settings, medium-scale deformation caused more by group excavation, and small-scale deformation along linear infrastructures. TerraSAR-X datasets for Shanghai spanning April 2013 to September 2020, and Sentinel-1A datasets spanning January 2016 to September 2020 are used in this study. The results were cross-verified between the TerraSAR-X and Sentinel-1A datasets, and validated against levelling measurements. Subsidence signals caused by different mechanisms were automatically decomposed, which facilitates a systematic analysis for targeted diagnosis of land subsidence signals. A detailed analysis was conducted jointly at three scales of surface displacement, geological conditions, major construction activities, and subsidence mechanisms. It indicated that construction activities were the leading cause of land subsidence, and suggests that local authorities that wish to mitigate surface subsidence may benefit from primarily considering this process. [ABSTRACT FROM AUTHOR]

Shanghai has the largest newly reclaimed area in China's coastal regions, and dredger fills are universally distributed in reclaimed lands there. Multi-scale geotechnical properties of dredger fills in region-scale (>1 km2), site-scale (1 m2–1 km2), mesoscale (1 cm2–1 m2), and microscale (1 µm2–1 cm2) were studied mainly through cone penetration investigation, borehole investigation, and microscopic inspections. Dredger impact craters, soil structures, and soil textures were investigated qualitatively and quantitatively to identify the dredger fills. Ultimate land subsidence since 1 July 2018 and land subsidence velocity characterized by the land subsidence from 1 July 2018 to 1 January 2019 were evaluated in Shanghai's typical reclaimed regions. Afterwards, Bayesian network including sixteen variables was proposed for land subsidence risk assessment. Surface subsidence, land subsidence velocity, and comprehensive land subsidence were the most sensitive to data point location and average initial void ratio of underlying strata. These three risk variables were also highly correlated to the thickness of the dredger fill layer and the reclamation time. This finding indicates that variances of the four independent variables (i.e., data point location, average initial void of underlying strata, land reclamation time, and thickness of dredger fill layer) significantly decide variances of the three risk levels. [ABSTRACT FROM AUTHOR]