Research On Tropospheric Delay Correction Of SAR Interferometry And Its Application In Land Subsidence Monitoring

Over the past 30 years,Interferometric Synthetic Aperture Radar(InSAR)has been widely used as a geodetic technique to measure the Earth's surface deformation.The time-series InSAR method provides a possibility to overcome the phase decorrelation and tropospheric delay error,but there are still some difficulties and problems,such as the stability and density of coherent targets identification,the accuracy and reliability of tropospheric delay estimation,etc.Among them,the tropospheric delay is still one of the biggest challenges for InSAR geodetic applications.Without proper delay estimation and correction,it is difficult to acquire an acceptable level of accuracy(millimeter level)for small amplitude surface displacement.Therefore,developing robust and reliable algorithms to estimate InSAR tropospheric delay remains one of the key challenges.This thesis aims at studying and exploring the key theory and method to overcome the influence of atmospheric effects,improving the accuracy and reliability of the InSAR technique for ground deformation monitoring.This thesis studies the following questions:(1)In-depth analysis and discussion on the effect of InSAR troposphere delay.Using meteorological data from radiosonde to analyze the distribution characteristics of the tropospheric delay.Our experiments show that InSAR tropospheric delay has convergence properties with the increase of height,verifying that the dominant delay contribution come from the wet delay,which is the main factor to improve the accuracy of tropospheric delay correction.Generally,vertical stratified component dominates the tropospheric delay and is related to the topography,and the turbulent mixing effects are random in interferograms.(2)Application research of time-series InSAR technique integrating with atmospheric correction techniques.The conventional spatio-temporal filtering method has some problems for estimating the troposphere delay,such as estimation bias due to the seasonality of the atmospheric delay.We proposed a method that incorporated the ERA-Interim model into time-series InSAR techniques to correct the tropospheric delay,and applied this method to monitoring the land subsidence in Changzhou by ENVISAT ASAR data and TerraSAR-X data.The accuracy is evaluated by inside and outside.The advantage of high resolution SAR is discussed.The characteristics and the reason of land subsidence are discussed.The extent of land subsidence acquired by three methods is different:the inversion results for SBAS method and PSM method have higher accuracy and consistency.The results obtained by TerraSAR-X show highest accuracy that the RMSE is 2.1 mm/y.From July 2004 to May 2013,serious land subsidence happened in Changzhou Wujin which is mainly characterized by two subsidence center and a subsidence zone.The geological conditions in Changzhou are the material basis for the land subsidence,which is closely related to ground water levels and building load.(3)Current phase-estimated correction techniques depended on the SAR scene and may fail to capture the spatial variability of the tropospheric properties.This thesis developed a robust and multi-weighted approach(RMW)and proposed the solution for time-series InSAR tropospheric delay correction.The basic idea of RMW method is:(1)using an expanded robust estimation algorithm to calculate the phase-height scale factor;(2)using a multiple weighted method to overcome dependence on SAR scenes of empirical model.(3)using a spatial band-filtering to reduce the effects from other signals(such as deformation,orbital error,etc.).This thesis used the PSI method to extract the time-series subsidence and deformation velocity based on ENVISAT ASAR images over Southern California to verify the correction capability of the RMW method.The experimental results show that RMW method mitigated the atmospheric noise better than the conventional phase-based method.The corrected ground surface deformation agreed better with those measured from GPS.(4)ERA-I reanalysis products are plagued by low spatio-temporal resolution,and need temporal,lateral,and vertical interpolations for InSAR observation correction,which could introduce uncertainty into the final result.This thesis proposed a method to correct InSAR tropospheric delay using a power law model based on ERA-I reanalysis(PLE),assuming that the interferometric tropospheric delay trend(power exponent)and convergence properties(constrained height)are sensitively simulated from the ERA-I reanalysis integrated with empirical model.Similar with the RMW method,the scaled factor is effectively and robustly estimated based on the relationship between band-filtered phase and reference topography.Through the experiment,this thesis compared and analyzed InSAR tropospheric delay estimated by PLE method to the linear method and the ERA-I method.The results show that the PLE method performed better than the other two and agreed well with MERIS measurements.