Atmospheric delay analysis from GPS meteorology and InSAR APS

Abstract

Radar atmospheric decorrelation due to inhomogeneity of atmospheric refractivity is a critical limitation of satellite SAR interferometry (InSAR) in the high accuracy retrieving of geophysical parameters. With mm precision, a water vapor tracing technique based on GPS meteorology was widely employed to mitigate InSAR atmospheric errors. However, a reliable comparison of atmospheric delay between GPS and InSAR is rarely touched, mainly due to the scarcity of stable and accurate InSAR atmospheric phases. In the paper we propose a comparison methodology between GPS Zenith Tropospheric Delay (ZTD) and SAR Atmospheric Phase Screen (APS) in both differential and pseudo-absolute modes. In the experiment, ENVISAT ASAR APS maps and synchronous GPS campaign measurements in Como, Italy were collected for consistency analysis. Furthermore, the stratification effect of atmospheric delay, in a form of delay-to-elevation ratios, was particularly analyzed for the purpose of separating different components within APSs. Finally, with the above stratification analysis, terms of stratification and assumed turbulence from SAR APS and GPS were compared in differential mode. Presented results show that the stratified ratios from GPS delays and SAR APS maps are in agreement with a std of 7.7mm/km and a bias of 3.4mm/km. Correlation coefficients of stratified ratios are higher than 0.7 in ascending case. In differential mode, the atmospheric total delays coincide with STandard Deviations (STDs) smaller than 4mm (∼0.65mm PWV) and with correlation coefficients higher than 0.6. The comparison of total delays in ‘pseudo-absolute’ mode is provided as an alternative vision of the agreement between GPS and InSAR. The agreement in this mode was slightly worse than that in differential mode. STDs of the difference are smaller than 6mm (∼1mm PWV), and the correlation coefficients are about 0.5 for different implementation approaches. Above comparison results in the work provide a quantitative extent to which atmospheric measurements from GPS and SAR APS are comparable. Another significant finding is that in most cases the STD of difference (between GPS and SAR APS) is slightly smaller than STD of SAR APS itself in both comparison modes. It implies the potentiality to correct atmospheric errors in SAR interferometry with high-precision GPS meteorological products, i.e. tropospheric delay or water vapor.