Methods For Monitoring Displacement Field Of Earthquake With High-precision And Constructing3D Coseismic Maps Using InSAR

InSAR (Interferometric Synthetic Aperture Radar) has the potential toproduce millimeter-level displacement maps for coseismic, postseismic, andinterseismic fault slip. However, accumulative errors in InSAR processinghave an impact on the accuracy of measurement. In this paper, we firstlypresent the brief flow of InSAR processing, then systematically analyze theerror-contributions of key procedures of InSAR, and discuss the methods ofeliminating errors. On the basis of above-mentioned discussions, we developsome methods to correcting height-dependent atmospheric delay phase, andapply the developed algorithms to monitor coseismic and postseismicdisplacements of2008Damxung earthquake. The laws of systematic andrandom errors are deduced in resolving3D displacement field, then theoptimal3D displacement fields are estimating in some earthquakes. The mainresults of the work are listed as below:1. Systematically summarize the main error-contributions of InSAR:image registration error, phase noise, residual phase due to DEM error,residual phase due to satellite orbit inaccuracies, and phase unwrapping error.We fully analyze the influence characteristics of above-mentioned InSARphase errors using the methods of combining theoretical inference, numericalsimulation, and real SAR data analysis, and discuss the methods ofeliminating different errors. Moreover, according to the law of errorpropagation, the influence of InSAR atmospheric delay is discussed inmonitoring displacement. From the above-mentioned discussions, we canconcluded that residual phase due to satellite orbit inaccuracies andatmospheric delay are the primary factors in affecting the accuracy of InSARprocessing.2. Based on the spatial delay characteristics, the InSAR atmosphericdelay can be re-parameterized as: the spatial mean, the turbulence and theheight-dependent (vertical stratification) components. Semi-variogram modelis introduced to analyze the characteristics of spacial variability of there-parameterized components. And we also deduce the semi-variogramstructure of InSAR atmospheric delay. According to the characteristics of theexperimental variogram, we propose the flow of correcting height-dependentatmospheric delay phase with DEM. Three algorithms are proposed to determine the optimal function of phase and topography.3. The characteristics of InSAR atmospheric delay in the area of2008Damxung earthquake are analyzed using Precipitable Water Vapor (PWV)derived by MERIS, which indicate that MERIS PWV does not capturesmaller-scale variation of PWV, and there is slight temporal-spatial variationof PWV in study area. Therefore, we chose the TCAD algorithm to correctInSAR images. According to the corrected InSAR coseismic images, wediscover that: except for structure displacement, the Damxung earthquakeresult in much non-structural earthquake damage related to collapse, landslide,and fissure. SBAS-TCAD InSAR time serial analysis algorithm is proposed inthis paper, which greatly removes the residual topography atmospheric delayphase. The algorithm is also used to analyze the evolution of postseismicdisplacement of seismogenic fault: in the first10months, the displacement ofupperwall presents logarithmic decrement in time; while the footwall presentsexponential decay in the first6months, the in4months it presents logarithmicdecrement in time. On the basis of analyzing coseismic and postseismiccharacteristics of displacement, slip distribution of seismogenic fault isinverted with Okada model.4. According to the analysis of influence factors of Offset-Tracking andMAI techniques, the process flows of Offset-Tracking and MAI are proposedto measure azimuth displacement. We summarize the main models of3Ddisplacement, and evaluate the robustness of these models. Based on the lawof error propagation, the laws of systematic and random errors are deduced inresolving3D displacement field. Then we take Bam earthquake, Hector mineearthquake, and Gaize earthquake for examples, the optimal3D displacementfields resolved models are proposed in these earthquakes.