Monitoring, analyzing and predicting urban surface subsidence: A case study of Wuhan City, China

Ding Q., Shao Z., Huang X., Altan O., Zhuang Q., Hu B.

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, cilt.102, 2021 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 102
  • Basım Tarihi: 2021
  • Doi Numarası: 10.1016/j.jag.2021.102422
  • Dergi Adı: INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Aerospace Database, Aquatic Science & Fisheries Abstracts (ASFA), Communication Abstracts, Environment Index, Geobase, Metadex, Civil Engineering Abstracts
  • Anahtar Kelimeler: Urban surface subsidence, Synthetic aperture radar interferometry, Geo-detector, Long short-term memory network, LAND SUBSIDENCE, RADAR INTERFEROMETRY, TIME-SERIES, SAR
  • İstanbul Teknik Üniversitesi Adresli: Evet

Özet

Wuhan, one of China's megacities with rapid development, is facing serious surface subsidence. In this study, we combined MT-InSAR, geo-detector, and LSTM (Long Short-Term Memory) to achieve the monitoring, analysis, and prediction of surface subsidence in the main urban districts of Wuhan. The effectiveness of MT-InSAR in monitoring surface subsidence was validated against leveling results. During the monitoring period, the maximum subsidence velocity and uplift velocity were -53.3 mm/year and 18.0 mm/year, respectively. We identified six subsidence regions and explored their deformation characteristics. Further, we analyzed the relationship between the surface subsidence and influencing factors using the geo-detector in a quantitative manner. Our study revealed that the distance to soft soils had the greatest explanatory power on the subsidence. However, we also confirmed that subsidence was affected via coupling effects from multiple factors, suggesting a complex reinforcing relationship among influencing factors. The interaction between the distance to soft soils and the distance to karst collapse prone areas had the largest joint explanatory power on subsidence. Further, we constructed a data-driven LSTM model to predict and analyze the subsidence. The results showed that the LSTM model achieved great performance and presented strong universality, suggesting that it can be used for subsidence prediction in large geographic areas.