Correcting Atmospheric Delay for InSAR Images using Regional GPS Network Atmosphere Model and Independent Component Analysis: A Case Study in Los Angles

Abstract:

Interferometric synthetic aperture radar (InSAR) has been widely used to monitor urban surface subsidence, groundwater extraction, and many geological disasters. However, tropospheric effect brings InSAR a big challenge for obtaining high-precise surface deformation, especially significant in the coastal area. Surface deformation in Los Angles mapped by Sentinel-1 SAR imagery (2015-2018) reveals a serious tropospheric delay, especially for C-band data. Here, we use the GPS-derived zenith tropospheric delay (ZTD) estimation to mitigate the InSAR tropospheric delay in this area. An independent component analysis method is also used to further remove this delay in time domain. Our corrected results are finally compared with the corresponding GPS vertical coordinate timeseries. The results show that, comparing with the weather reanalysis model correction, GPS ZTD works better by at least 15%. And the ICA method could further improve the correction results by at least 10%.