Building risk monitoring and prediction using integrated multi-temporal InSAR and numerical modeling techniques

Abstract

• Large-scale ground subsidence was measured using an improved MTInSAR method. • Building risk level was assessed at the district scale using InSAR measurements. • The simulated subsidence due to underground construction agreed with InSAR data. • The integrated method achieves building risk monitoring and prediction. Frequent anthropogenic activities associated with rapid urbanization may cause ground subsidence and endanger the adjacent buildings. Interferometric synthetic aperture radar (InSAR) has generated dense measurement points for large-scale deformation monitoring and risk assessment. However, the applications of InSAR for building deformation and risk level prediction are still rare, especially under complex loading scenarios. In this study, we present an integrated method to monitor and predict building deformation and risk levels using multi-temporal InSAR and numerical modeling. Persistent and distributed scatterers were jointly detected to monitor deformation using TerraSAR-X, COSMO-SkyMed, and Sentinel-1 images over the Kowloon district of Hong Kong and the Futian district of Shenzhen. First, the building risk level was assessed at the district scale using ground subsidence and angular distortion to identify potentially damaged buildings. 10 out of 807 buildings in the Kowloon and 2 out of 145 buildings in the Futian were classified as medium risk. Back-analysis was used to retrieve the geological and hydrological conditions of buildings at the medium risk level. Then, time-series subsidence of the Lucky and BMW buildings during 2012 to 2016 and the Jingli building during 2010 to 2014 was simulated by numerical modeling, which agreed well with the InSAR measurements. By defining the groundwater level changes and upper loading, the predicted subsidence of the Lucky, BMW, and Jingli buildings reached 47.9, 62.8, and 39.9 mm by 2021, respectively, with a decelerated trend. Time-series results showed that the Lucky and BMW buildings were still at the negligible-to-medium risk level in 2021, and should be constantly inspected; the risk level of the Jingli building became negligible from 2020. This method can be transferred to other rapidly urbanized areas prone to ground subsidence for geohazard assessment and improve the understanding of the dynamic behaviors of anthropogenic activity-induced ground subsidence.