Monitoring Co-seismic, Inter-seismic Deformation And Ground Subsidence By InSAR Observation

By subtract topographic phase from InSAR image by differential processingtwo InSAR images that one contains crustal deformation phase while the other not,Massonnet first derived spatial continuous co-seismic deformation filed first inhistory, which applied to Landers Earthquake. We called this technique as D-InSAR,which contributes greatly to deriving co-seismic deformation field and inversingfocal mechanism in the following decades. Despite its high precise and spatialcontinuous deformation field, D-InSAR is limited by coherence which declinesquickly while spatial perpendicular baseline and temporal baseline increases. Toovercome such limitation, PS-InSAR, by applying a stack of SAR images acquire bythe same antenna in the same mode, same track and same frame, estimating SNRand/or phase stability pointwise. Those pixels with high SNR and show phasestability are man-made construction or bedrock which are regard as persistentscatterers. Even if temporal baseline last for years, or perpendicular baseline variedgreatly, persistent scatterers can show high coherence and a stability in phase.Therefore by only analyzing phase variation of persistent scatterers, PS-InSAR canderived mean velocity and time series of crustal deformation even if perpendicularbaseline varied which may leads to failure of conventional D-InSAR. HencePS-InSAR shows greatly potential in analyzing urban subsidence, volcanicdeformation and inter-seismic deformation.LOS deformation derived from PS-InSAR can be caused by hydrologicalloading, loss of support beneath the ground, glacial isostatic adjustment, glacialmelting, and inter-seismic deformation. To study inter-seismic deformation, it isnecessary to analyzing and subtract those nuisance deformation terms detected byPS-InSAR. For seasonal deformation is so common that most roving GPS survey aretaken in the same season to avoid such influence. Vertical deformation series ofPermanent GPS stations also show a seasonal change which correlated strongly withhydrological loading. Global climate change leads to glacial melting in Tibet Plateau,which introduces that lakes which mainly supplied by glacial melting get large. Thisphenomenon leads to a long term subsidence around the lake and compromisingroving GPS survey. We derived LOS deformation field around Bamucuo Lake inTibet which shows a tunnel-shaped subsidence around the lake. We also derived areachange of this lake by using SAR amplitude images after geocoding and coherence maps of InSAR images to assist.Since its limited precision of satellite borne DOIRS or GPS receiver, SARsatellite orbit determination accuracy limits to20cm. Discrepancy between real orbitand precise orbit released by different agency such as ESA, JAXA would result in abias which called orbital ramp, similar to flat-earth effect, which can be eliminatedbased on the assumption that the displacement field in the area of interest does notcontain any linear spatial trend (in a full SAR scene) when there are no GPSdeformation field to constrain. For PS distribution is highly correlated with thecoherence map, if decorated area divided the area into different isolated patch, theywould share no same unwrapping datum, which would leads to error in orbital rampestimation. We proposed an algorithm to estimate orbital ramp based on DelaunayTriangulation and Tarjan’s depth first search algorithm. We seek to find the largeststrongly connected component of PS point cloud to estimate the orbital ramp insteadof all the PS pixel. We applied this algorithm to an example in south Tianshan, and itis effective.Collision of India and Eurasia Plate not only cause significant uplift of TibetPlateau but also wide spread and large scale deformation within the Eurasia Plate.Pamir is the most far place India Plate indenting into Eurasia Plate. In the lastcentury, many earthquake above M6.5struck this area, which indicate strongtectonic. We acquired SAR image spread for seven years after Bachu-Jiashiearthquake of M6.8occurred on Feb.24,2003and before orbit of Envisat get low inOct,2010, while no significant earthquake struck this area. We used both PS-InSARand SBAS methods to derived LOS deformation field in these area, and by applyinghorizontal GPS deformation field we derived vertical deformation field, whichindicate strong deformation and collision at about1cm/yr。 We found thatTuotegongbaizi-Arpaleike fault where1902Atushi8occurred show no significantinter-seismic deformation while buried fault north to Kashi city and S-shaped faulton the west edge of Tarim basin show the attractive uplift in these area. Kashidepression shows an obvious subsidence with respect to Kashi Anticline. We alsoused Savage’s negative dislocation model and Okada dislocation model to simulatethe deformation field, which indicate the locking depth, dip angle and slip rate ofthese buried faults.