Crustal Deformation Observation And Simulation In Middle Balyankalla Block Based On Long-strip InSAR Data

This thesis took central Balyanlkalla block as the study area where the 2001Ms8.1 Kunlun earthquake took place. By means of comprehensively using GAMMA,PI_RATE and other professional software and selecting two tracks(T133 and T176)with the length of more than 500 km, we put forward a enhanced processing method for long strip In SAR data, obtained the crustal deformation rate of study area as well as the accumulative time sequences of ground deformation which cover several main faults, utilized the hierarchical model to simulate the post-seismic deformation field,and acquired viscosity coefficient of the crust by inversion. Then, we took 2015Mw7.8 Nepal earthquake and 2015 Mw8.3 Chile earthquake as examples to explore and investigate both the characteristics and processing techniques of a new generation of radar satellite data, Sentinel-1A, for studying the possibility of its applications in the co-seismic deformation observation and detection. Our findings and main conclusions in this paper can be summarized as the following several aspects:1. For the SLC data in T133 and T176 track, We selected interfere image pairs with perpendicular baseline of less than 200 m, carried on the conventional D-In SAR processing, chose these with better coherence and smaller orbital errors in far field,adopted quadratic polynomial fitting approach to remove phase errors induced by residual orbital error and DEM error. We obtained multiple strip deformation fields covering the main interesting area. We calculated and analyzed cumulative deformation values, ?S, and corresponding time span, ?T, of many deformation fields,as well as resolved the relative deformation rate, V, about 10-15mm/yr. Balyanlkalla block shows the trend of the eastward movement, and the variation of deformation rate across the Kunlun Mountain after the year of 2003 is not obvious.2. We developed a new interface program for interference processing software,GAMMA, and stacking processing software, PI_RATE, which could integrate a large number of unwrapped interferograms and the corresponding SAR intensity images derived by GAMMA into the new defined dual-band files, and reset the relatedprocessing parameters. We import these new dual-band files into PI_RATE, correct orbital errors and atmospheric errors, and calculate the average velocity field as well as the displacement time series of the earth's crust within observation period by setting different correlation coefficient thresholds, forming a new set of improved long strip dataset processing method(G-P method). Results show that the deformation patch is dominantly distributed in the south area with width of about 130-150 km and in the north area with width of about 100 km. The maximum average deformation rate reaches about 10.6mm/yr. Results from time series analysis show that the largest LOS(line of sight) relative movement of fault reaches 100 mm from January 8, 2004 to July 15, 2010.3. Based on the layered crust model and displacement field derived by strip In SAR data, we tried E-M-M and E-M model to simulate the post-seismic LOS deformation, and acquired the viscosity coefficient of the medium by inversion.Through analyzing the results we can draw such conclusions as:(1) E-M-M model cannot resolve the viscosity coefficient of the upper mantle.(2)The results derived by two models present that the optimum viscosity coefficient is 7.00E+18Pa·s.(3)After-slip has been basically terminated three years after the 2001 Ms8.1 Kunlun earthquake, and crust deformation within the year of 2004-2010 is main controlled by lower crust viscoelastic relaxation regime.4. We also explored the characteristics and processing technologies of a new generation of radar satellite data, Sentinel-1A, as well as its effect on detecting and acquiring co-seismic deformation field. We used Sentinel-1A SLC data to acquire widely distributed co-seismic deformation of April 25, 2015 Nepal Mw7.8 earthquake and September 16, 2015 Chile Mw8.3 earthquake, then we inversed the fault slip distribution, and optimum results were obtained. By comparative study and investigation, we found that the far field terrain-related atmospheric error is obvious and notable, and it would produce excess fringes in seismic deformation field by Sentinel-1A In SAR. Terrain-related atmospheric error will bring serious interference to very small magnitude deformation observation. Therefore, it is urgent to develop new method in removing terrain-related atmospheric error in interferogram.