Remotely sensing large- and small-scale ground subsidence: A case study of the Guangdong–Hong Kong–Macao Greater Bay Area of China

Abstract

As the world's largest city cluster, the Guangdong–Hong Kong–Macao Greater Bay Area (GBA) is vulnerable to significant subsidence. The widely distributed sediments and rapid urbanization in the GBA result in the concurrence of large- and small-scale subsidence. Mono-sensor synthetic aperture radar (SAR) images usually suffer from the limitation of either low resolution or small coverage, and thus, are not applicable to completely monitoring the GBA. In this study, we used Sentinel-1 (S1), COSMO-SkyMed (CSK) and TerraSAR-X (TSX) images jointly to reveal multi-scale subsidence of the GBA. The overall subsidence ranged from 0 to 112.3 mm/yr derived from the 2015–2017 S1 images. Three regional subsidence bowls (Zhuhai–Zhongshan, Jiangmen, and Guangzhou–Zhongshan) formed in the western alluvial plain. The high correlation between regional subsidence and Quaternary sediments confirms that sediment consolidation is the main cause of subsidence. The land use and numerical modeling results suggest that groundwater extraction and artificial loading are the triggering factors. Two representative local subsidence cases were analyzed using high-resolution images: reclamation settlement at the Hong Kong International Airport (HKIA) and sinkhole subsidence due to the excavation of the Shenzhen Wenbo skyscraper foundation pit. The validation at the HKIA showed that the measurements from the CSK and S1 data both agreed well with the leveling data. However, CSK outperformed S1 in the sense that it increased the point density by a factor of 5, improved the height precision by a factor of 4, and showed fewer false alarms. CSK is therefore more applicable to monitoring the local subsidence of key infrastructures. The cross-heading tracks of TSX and CSK images detected precursory subsidence before the sinkhole collapse from two sides, indicating that the cross-heading tracks benefit the comprehensive monitoring of local subsidence in high-rise and high-density built environments. In summary, the synergistic use of multi-sensor SAR images demonstrates the practicability of the operational surveillance of multi-scale subsidence from regional surveying to the fine monitoring of local areas in the GBA.