Three-dimensional Numerical Simulation Of Magnetotelluric Sounding For Northern Hainan Volcanic Area

Firstly, some basic preparations are made for the study of this thesis. They include picking up the time series from the instrument V5-2000, noise processing and calculation of fundamental parameters. There is a data processing soft attached the instrument V5-2000, but it does not provide the software for picking up the time series. Only getting the apparent resistivity, we can distinguish whether the original data is good or not. So I analyze the data thoroughly, and use the visual basic language to compile a program for picking up the time series. It can correct parameter errors set in field operation and help to obtain good processing result through truncation of the time sections. But, it is not able to eliminate noise from signals. In the view that the magnetotelluric data acquired in the Hainan area was disturbed by thunders, this work attempts to apply the wavelet analysis to remove peak disturbance and explain the reason of its existence.Then numerical modeling is performed in this work. Since the electric structure of the Earth is inherently heterogeneous, three-dimensional forward and inverse algorithms are required for interpretation of magnetotelluric data. Besides, local electromagnetic fields are prone to effects of the regional features, which are far from the survey area, the applied modeling algorithms must be able to incorporate these features. In this work the three dimensional finite difference program for magnetotelluric sounding compiled by Randy Mackie is employed. Its theory, including establishment of difference equation sets, boundary conditions, conjugate relaxation solutions to difference equation sets, and calculation of three-dimensional impedance, is described in detail in the chapter three of this thesis. The algorithms are difference equation algorithms that are based on the integral forms of Maxwell's Equations rather than the differential forms. This formulation does not lead to the problems of approximating derivatives of Earth properties or derivatives of electromagnetic fields. Finally, this modeling method is applied to interpret magnetotelluric data acquired from the Northern Hainan volcanic area.The Leiqiong region has a largest volcanic area of Cenozoic time in the South China. The volcanic activity began in Eocene Epoch and lasted till Holocene Epoch. The main body of the Qiongbei volcano is trending north west. The magnetotelluric sounding was carried out to explore the structure of this area. Four survey lines, HN1, HN2, HNS, and HN4, were deplored, among which HN1 is normal to the strike of the structure and other lines are parallel to the structure. For the HN1 line, the two-dimensional inversion (conjugate gradient method) is used, and one-dimensionalinversion is for the other three lines.The results of one-dimensional and two-dimensional inversion of this work show that the electric structure beneath the northern Hainan volcanic Area has a good continuity, through the Qiongzhou Strait exists between the survey line HN2 and HN3. Along the line HNl there is a remarkable aggregation of low resistivity in the shallow crust. About 20km east to the Ma'anling volcano, a body of low resistivity, implying a possible magma chamber, is found at the depth of some 20km, which has been confirmed by the survey line HN4 trending NS through the body. The 1-D inversion result along HN4 indicates that a body of high resistivity is presented above this low resistivity, in good agreement with the inversion of HNl. As there is no available method for determining the motion state of this magma chamber, 3-D numerical modeling is conducted to verify its existence.A three-dimensional model is designed based on the two-dimensional inversionresult along the survey line HNl. Lines HN2 and HN4 are used to control the boundary condition of the model. A vertical profile and several horizontal slices are taken to display the forward modeling results. The numbers of grid elements are Nx = 171,Nv = 66, and Nz = 54 in directions x, y, and z, respectively. Using the 3-D techniqueto model the apparent and...