Surface Deformation, Co-seismic And Post-seismic Activity Constrained By Advanced InSAR Time Series Analysis

Differential interferometry techniques (D-InSAR) are developed based on InSARtechnology, which are used to measure small deformation of earth surface. They arewidely applied to earthquake and volcanic activity, glacier monitoring, urban settlementand landslides, and other field. They can also supply the constraints for the earthdynamics research and physical model inversion. With the widely application ofD-InSAR techniques, a few limitations have significantly affected the analysis inpractical research fileds, which include the orbit parameters error, DEM data error,atmospheric delay error, phase unwrapping error associated with the temporal andspatialsignaldecorrelation, and system noise, etc. These factors have been the mainrestrictions for InSAR high-precision deformation detection applications.InSARtime-series analysis method is the one of the best ways to solve the conventionalD-InSAR restrict accuracy factors. This paper carries out the key technologyinvestigation of InSAR time-series analysisand innovation points selection are asfollows:(1)This paper proposes a new method of fullrank coherent points based on thetraditional coherent points extracting method. Full rank coherent points methodcombines redundant network nodes and edges in the graph together to build a full rankmatrix and accordingly as a coherent point extraction indicator. The new method canimprove the density of space distribution of coherent effectively and preserve optimalmeasurement precision.(2) This paper uses a network method to correct three types of atmospheric delay phases,long wavelength atmospheric phase, short wavelength atmospheric phase andTopography Corrected Atmospheric Delay (TCAD) respectively. The network methodcan estimate the atmospheric delay error of eachepoch, and then reconstructatmospheric delay error for each interferogram. The proposed method can accuratelysimulate the atmospheric delay phase in each interfereometric image.(3) The unwrapped phase error of the discrete coherent points is one of the main errorsources in the process of time series analysis, directly affects the precision of thesubsequent time-series analysis results.This paper proposes a method of using closedloop network to detecte and correct the discrete phase unwrapped errors of coherentpoints. The method could efficiently detecte the main phase jump errors of thediscretecoherent point unwrapped phase and restore the new corrected solution for thesephase jumps. (4) This paper also proposed a new InSAR time series inversion strategybased on fullrank matrix conditions. The strategy applied the relationship between the nodes andedges in theredundancy network and the least squares method to calculate deformationcharacteristics of each time epoch to solve the traditional InSAR temporal rank deficientproblem in the process of inversion. The new method is applied to the postseismicdeformation analysis associated with the2003Mw6.5Bam earthquake, the landsubsidence of Yambajan plant geothermal exploitation caused by groundwaterextraction and land subsidence monitoringof Beijing city, which validate the efficiencyand accuracy of the new method.The coseismic and post-seismic deformation of the2008Mw6.3Damxungearthquake on Tibetan Plateau is investigated using the new InSAR time series. Thispaper determines the coseismic distributed slip model and its potential potentialpostseismic physical mechanisms. Extracting fault geometry and slipmodelusingcoseismic deformation field from the InSAR and as a prior knowledge, combinedwith post-seismic deformation, to analysis the Damxung earthquake fault ruptureprocess using slip model andlower crust or upper mantle viscosityand other majorgeophysical parameters in plateau central using viscoelastic relaxation model.