Seismic Tomography Application In The Ningxia Hui Autonomous Region And Adjacent Areas

By analyzing kinematics (travel time, ray path) and dynamics (waveform, amplitude, phase, frequency) data of seismic wave phase observed, Seismic tomography is an important geophysical method which could used to inverse impotant information of earth medium covered with a mass ways,such as structure ,velocity distributing and elasticity parameters The origin of Seismic Tomography method is from the medical CT, with being developed in geoscience ,Seismic Tomography has gotten significant advances ,so it is being accepted by more and more people.In this paper, development history of Seismic Tomography and its principles model constructing, finite difference forward modeling and two methods of inversion.These two inversion methods are applied to Ningxia Hui Autonomous Region and adjacent areas. Useing 6 years'digital seismic first arrival recorded Pg and Pn data .and chosing initial model of damped least square method according to the observed line. and then chosing genetic algorithm initial model according to the result of damped least square method. by inversing 3-D velocity structure(damped least square method), 1-D velocity structure and source parameters(genetic algorithm),we obtained distribution of velocity at different depth,and profile images at different locations.The following conclusions are obtained by comparing two inversion algorithm as well as the results with reference of of geological structure information and previous research results.1. There are academic differences between two diffrent inversion Algorithm :The damped least square method is a linear solution,which can give discernibility matrix and covariance matrix of the solution estmate directly. damped factorλ, both play a role in promoting the balance, but also affects the convergence speed.The genetic algorithm is a global optimization algorithm, which applies the stochastic switching rules to the optimal solution searching and suits in various optimized questions.However it is the only way to obtain the covariance matrix with the linearized means and unable to solve the model parameter information directly such as resolution and variance component,et al.2. Two results of two inversion algorithms are as follows: damped least square method showed that①depth range 0-5km which velocity range is 4.5-5.3km / s between.②depth range 5-25km which velocity range is 5.0-6.1km / s, the average velocity about 5.5km / s.③depth range 25-35km which velocity range is 5.6-6.6km / s, the average velocity about 6.2km / s. This result compared to initial model are similar to but not the same. Genetic algorithm showed that①sediment,1-8km deep,,has an average velocity 5.27km,②upper crust:,9-22km deep, has an average velocity 5.55km/s;③Lower crust 23-48km deep, has an average velocity 6.27km/s,④and upper mantle depth 49-52km, has an average velocity 7.79km/s.According to damped least square method ,there are five low velocity areas, sorting in size are as follows:①the western margin of the Yinchuan graben, Junction region of Yinchuan graben and Alashan Block;②Zhongwei in Ningxia including the shallow low velocity area at Qingshuihe basin in Zhongwei;③the northern margin of the Yinchuan graben, Junction region of Yinchuan graben and Alashan Block;④Tongxin Luoshan in Ningxia⑤deep low velocity area.at easten part of Haiyuan fracture , southern Haiyuan-Guyuan in Ningxia.Higher velocity are mainly distributed in the region: Near qinshuihe fault. which in shallow at the ease of Zhongwei Xianshan only. (105.5°east) , at 20km depth it extend Westward to Alashan zuoqi in Mongolia.Two shallow high velocity region can also be seen :①the Southeast part of Yinchuan graben ,from Zongwei in Ningxia NE extend to Etuokeqi in Mongolia;②in the western part of Haiyuan fault, from Jingtai in Gansu extend NNE to west Hayuan in Ningxia.Based on genetic algorithm, the source parameters on the map gather more or less close to faults. Earthquake occurred in the low velocity or at the edge of low velocity to high velocity zone, we can see the genetic algorithm positioning results is better in depth than network positioning results in line with the the high and low velocity area, but not very obvious advantages. Overall, the genetic algorithm inversion of source parameters location was not significantly improved than the network.