Estimation Of Fault Geometry By Seismic Activity Data

Seismic fault geometry is a hotspot in seismology and geodesy research,It determines the motion characteristics of faults to a large extent.Studying fault geometry helps us to understand earthquakes and crustal deformation movement mechanisms more clearly.Meanwhile,it is great significance to predict future seismic activity and its development trend,Seismic activity is one of the most natural disasters that damages the sustainable development of human life,property and economy.It brings rare research opportunities and unavoidable challenges to earthquake workers.At the same time,earthquake activities have left a lot of preciousness observation data.A serious challenge to avoid.In this paper,statistical variance analysis is used to study the influence of various parameters in the fault geometry model on the ground displacement.Based on the seismic activity data,the method of calculating the fault parameters under different conditions and different data is given.Then the fault dip and its variation with depth are quantitatively and finely calculated to determine the geometry of the fault,According to the given fault geometry,the displacement of the surface observation point is predicted,and the GPS data of the ground observation point is used to verify th,e correctness of the fault dip angle result,which provides reliable basic research data for future geodetic inversion.For the geometric model of active faults with multiple fault planes,this paper uses the principle that clustered small earthquakes occur near the fault plane.It is assumed that the source points obey the three-dimensional normal distribution around the center of the fault plane,and the fault planes are divided by clustering method.The traditional clustering algorithm is improved,and then the three-dimensional structure of the active fault network is reconstructed by combining the single fault plane geometric parameter determination method.The main work and achievements of the thesis are as follows:(1)The main parameters of the fault were studied by analysis of variance.Under the given geometric parameters of the fault,by changing the length,width and,dip value,the different ground point displacement results obtained by forward modeling are compared with the theoretical values.Using the analysis and calculation of variance anallysis,it is found that the three factors of length,width and dip have a significant effect on the surface displacement whether it is single action or pairwise interaction.By calculating the effect magnitude of each factor,it is found that the dip angle plays a dominant role in the factors affecting the surface displacement.(2)Three methods commonly used to calculate the fault dip angle are studied and compared.In this paper,the Dujiangyan section of the Longmenshan fault zone is used as an example to test the three methods of least square method,focal mechanism solution and maximum likelihood method.It is found that all three methods can give the fault dip angle,but the calculation results have certain differences.Then,after calculating the correlation and residual verification and analysis,it is concluded that the calculation results of the focal mechanism solution and the maximum likelihood method are more accurate and better than the least squares method.(3)In-depth study of the fault structure of the fault.For the case that the actual fault plane is not a geometric plane,this paper innovates the research method.Taking the Dujiangyan fault of the Longmenshan fault as an example,it divides along the strike direction and the downward extension direction to form a fault layered segmentation model to study the detail of the fault,and then the optimal B23 fault model was given based on the measured surface displacement data.(4)Based on the idea of principal component analysis,this paper improves the traditional clustering algorithm,overcomes the defect that it can not automatically give the number of clusters,and gives the principle of using the principal component analysis method to determine the fault parameters.Then,using the improved clustering algorithm and the earthquake aftershock catalogue,a three-dimensional fault network with 10 fault planes in the Landers region of the United States was successfully constructed,which is similar to the previous research results.