Integrating SBAS-InSAR and LSTM for subsidence monitoring and prediction at Hong Kong international airport

Abstract

Hong Kong International Airport (HKIA) is one of the busiest airports in the world, and much of its land is reclaimed from the sea, making it prone to uneven subsidence of the ground. Monitoring and predicting the subsidence of its surface are crucial for ensuring the operational safety of the airport. This paper firstly obtained the surface subsidence characteristics of the HKIA through applying the Small Baseline Subset Interferometry Synthetic Aperture Radar (SBAS-InSAR) technology, and then the spatial–temporal evolution was analyzed by using the Standard Deviational Ellipse (SDE) method. Moreover, the Long Short-Term Memory (LSTM) was employed to perform surface trend prediction of HKIA. The results show that the HKIA presents different levels of subsidence and uplift, with a maximum average subsidence rate of −64 mm/year and a maximum cumulative subsidence of −199 mm. The comparison between predicted curves and the actual subsidence revealed by InSAR from 2019 to 2023 is highly consistent, with the average absolute error and root mean square error less than 5 mm, and a coefficient of determination greater than 0.99. The LSTM model utilized in this paper can achieve reliable results in subsidence prediction based on time-series InSAR, and provide alternative means for geohazard prediction.