Retrieval of ground subsidence by integrating GPS observations and persistent scatterer InSAR and using a small number of ASAR images

Persistent scatterer (PS) InSAR technique relies on a significant number of SAR images (typically suggested 25 or more) to achieve reliable ground deformation measurement. This often limits its practical applications for a limited number of existing SAR images for a given area. We present the method of full interferometric combination from a small number of SAR images to detect ground deformation considering the fact that each interferogram has its coherence that cannot be considered strictly as a linear combination of other ones. The integration method between PS InSAR technique and GPS observations for improving InSAR results is also proposed in this paper. Twenty-eight radar interferograms from eight Envisat ASAR acquisitions over Hong Kong are generated in our study. The tropospheirc wet-component phase delays derived from 12 permanent GPS stations operating in the area are used to correct radar interferograms. A joint deformation monitoring network comprising GPS stations and identified PS is then formed. The weighted least square adjustment algorithm is employed to estimate the linear deformation of PS targets based on the formed network. The validation results show that the derived subsidence velocity of PS targets using the proposed method is well consistent with GPS derived values with the assessed accuracy of about 2.0 mm/y.