InSAR Tropospheric Delay Correction Using Atmospheric Reanalysis And Water Vapor Mapping

Over the past three decades,repeat-pass spaceborne interferometric synthetic aperture radar(InSAR)has been widely used as a geodetic technique to generate maps of the Earth's topography and to measure the Earth's surface deformation.During the acquisition time of two SAR images,the inconsistence of atmospheric conditions will contribute to propagation delay for radar signals,leading to atmospheric delay errors in interferograms.In the case of InSAR applications for geodesy,to acquire a high accuracy of topography or displacement information,atmospheric delay noise must be eliminated or reduced.Developing robust algorithms to estimate tropospheric delay remains one of the key challenges in the field of InSAR applications for geodesy.Conversely,the "noise" in this field has become a very useful "signal" in atmospheric sciences.The spatial and temporal variations in water vapor content are the major factors causing tropospheric delay in interferograms.Once the tropospheric delay"signal" is accurately extracted from the interfero grams,then it can be converted to the precipitable water vapor(PWV).Regarding to the above problems in InSAR,this thesis studies the following major issues:(1)Recent studies showed that global atmospheric reanalysis has great potential for mitigating large-scale tropospheric delay,motivating us to further explore the effectiveness and robustness of those atmospheric reanalysis products.We compare the performances of two reanalysis products in our study,namely ERA-Interim reanalysis and North American Regional Reanalysis(NARR).We examine their efficiency by processing a set of interferometric SAR images(N=51)acquired by Envisat ASAR over Southern California,USA from May 2005 to September 2010.We validated our approach by comparing these products with atmospheric delay derived from the passive multispectral imager Medium-Resolution Imaging Spectrometer(MERIS),onboard the Envisat satellite.(2)The InSAR time series methods assume that the atmospheric delay is uncorrelated with time,so the spatial-temporal filtering method was applied to separate deformation and atmospheric delay.This assumption,however,is not always valid because the seasonal variation of the atmospheric conditions could remain.We use InSAR time series analysis to investigate land subsidence in the entire region of Taiyuan Basin and demonstrate that the seasonal variation of tropospheric delay in this area.The tropospheric wet delay can be divided into systematic and stochastic components in both space and time,while the former one is the main factor to bias the InSAR deformation results.Instead of applying the conventional spatial-temporal filtering,we propose a new method that employs the stratified delay obtained from ERA-I to eliminate the systematic component and then the residual(stochastic)component can be effectively estimated by the spatial-temporal filtering.The estimated atmospheric delay and surface deformation agree well with those measured from GPS and MERIS.Finally,we analyszed the correlation between gravity change and surface displacement in Taiyuan basin.(3)InSAR can serve as a new approach for meteorological study that we can investigate the tropospheric delay "noise" as a meteorological signal to measure the water vapour content in the atmosphere.Here,we will present a new approach for accurate water vapour estimation with a high spatial resolution by combing InSAR observations,GPS data and ERA-I.Time series maps of differential PWV were obtained by processing a set of Envisat ASAR images covering the area of southern California,USA from 6 October 2007 to 29 November 2008.To get a more accurate PWV,the component of hydrostatic delay was calculated and subtracted by using ERA-I.In addition,ERA-I was used to compute the conversion factor required to convert the zenith wet delay to water vapour.The InSAR-derived differential PWV maps were calibrated by means of the GPS PWV measurements over the study area.We validated our results against the measurements of PWV derived from the MERIS.Our comparative results show strong spatial correlations between the two data sets.The difference maps have Gaussian distributions with mean values close to zero and standard deviations below 2 mm.The advantage of the InSAR technique is that it provides water vapour distribution with a spatial resolution as fine as 20 m and an accuracy of?2mm.Such high spatial-resolution maps of PWV could lead to much greater accuracy in meteorological understanding and quantitative precipitation forecasts.