Estimation Volcano Deformation With PSI And Pixel Offsettracking For Source Parmeters Inversion

Volcano hazard, one of the worst natural disasters, can trigger a series of surface deformation. Therefore, understanding subsurface processes such as the deep of magma chamber and magma volume changing are essential to any attempt to assess the risk of potential damage and reduce the loss caused by it. Because these processes are hidden beneath the surface, geophysical measurements may represent the best tool to study them.It is practical to use space-born interferometric synthetic aperture radar (InSAR) measurements to detect deformation of volcano, and other methods such as Pixel Offset-Tracking also can be used and combined with InSAR. But those measurements still have some defects, such as loss of coherence caused by bare area and large deformation gradient in volcanic area. What more, the negative effects of image registration and deformation detection model etc. also make low deformation detection accuracy.This thesis, based on method of Pixel Offset-Tracking and PSI, presents a series of studies (e.g. Combinatorial Models and Interferometric Combination Mode) carried out at Sierra Negra volcano to improve the accuracy and spatial coverage of measurements with SAR images, and provides reliable deformation of volcano to invert the process beneath the surface. There are four main contents of work, listed below, to ensure the accuracy and spatial coverage of deformation detection:(1) Because the precision of image registration, affected by the length of space-baseline between two SAR image, plays an essential role in Pixel Offset-Tracking method. Therefore, least-Squares registration method has been used for that to optimize image registration with long space-baseline in this paper, and sequentially provides highly precision registration data sources for detecting horizontal deformation with Pixel Offset-Tracking.(2) Combinatorial models have been proposed as the constraint for deformation detection, which take the characteristics of volcanic deformation into consideration. Real GPS data was used for models validation, and show that these models can provide high-precision deformation field. More specifically, the deformation precision of new models on azimuth and range direction were up to 0.070 m and 0.160 m, respectively, and more precision than the traditional model which comes from SBAS.(3) Interferometric processing of time series of SAR images is one of important parts in PSI technique for monitoring ground deformation of volcano. The mode of interferometric combination directly affects the quality of interferometric phases. In order to detect the deformation of crater, an improved interferometric combination mode based on traditional ways is proposed and validated by GPS data. The accuracy can reach centimeter.(4) Jointing the deformation, detected by PSI and Pixel Offset-Tracking, the deep of magma chamber and magma changing are inverted by Mogi model and conclusion are carried out. The results show that, the velocity of volume changing before volcanic eruption arrived at 1.59×107 m3/years, while the dramatic change in volume during eruption was about 1.02×108 m3. Moreover, the trend of magma variation is relatively smooth when compared with traditional model, which because the combinatorial models are less effected by noise. And final inversion results indicate that jointing the combinatorial models and deformation in different directions can provide more effective information to monitor volcanic subsurface processes.Previous studies have shown that the proposed method can improve the accuracy and spatial coverage of deformation in volcano area, and provide great practical information to monitor volcanic activity (e.g. volcanic parameters inversion and hazard assessment).