Global Mantle Conductivity Imaging By Joint Analysis Of Satellite And Observatory Geomagnetic Data

Mantle electrical conductivity is a key parameter to infer the water content and thermal state of Earth’s interior.Electrical conductivity directly reflects the connectivity of the components of mantle materials such as water content,fluid,partial melt and volatilities,and is a key rheological parameter of mantle materials.Therefore,to infer the internal state parameters of the mantle which are closely related to the internal motion mechanism,such as the composition and rheology of the mantle materials,it is necessary to study the mantle conductivity.Limited by the simplification of the Earth’s external current sources,the defects of the observed data,and the low accuracy and efficiency of inversion methods,the development of mantle conductivity imaging is slow.(1)Current studies rarely use electromagnetic(EM)data induced by ionospheric currents,limiting the imaging resolution of upper mantle conductivity.(2)Current studies mainly use small and unevenly distributed geomagnetic observatory data,limiting the imaging resolution in oceanic regions and the southern hemisphere.(3)None of the existing global EM forward modeling solvers can accurately simulate the real land,ocean,coastline and deep influences,limiting the further development of three-dimensional(3D)inversion of mantle conductivity.To overcome the above issues,this thesis systematically carried out the research on 3D mantle conductivity by integrating the data of geomagnetic satellites and geomagnetic observatories and presented a high-resolution global 3D mantle conductivity model.The main research contents and innovative works are as follows:1.A 3D finite element forward modeling method of global EM induction considering the effects of real land,ocean and coastline is developed.Firstly,a new global-scale multi-resolution tetrahedral mesh discretization method is established by considering the different characteristics of the geomagnetic observatory and satellite data.Then,the boundary value problem of global EM induction applicable to current excitation in both the magnetosphere and the ionosphere is derived by using the background fieldanomaly field decomposition strategy.Then,the boundary value problem is discretized into large sparse linear equations using the vector finite element method.A variety of efficient iterative solving strategies for linear equations are implemented and tested based on the auxiliary-space preconditioning technique.A multi-level parallel acceleration strategy is implemented based on MPI.Finally,several synthetic models are used to verify the correctness and efficiency of the proposed forward algorithm.A realistic 3D conductivity model is used to simulate the ocean effect on the surface and at satellite altitudes.2.A fully satellite and observatory data processing method integrating ionospheric and magnetospheric current information is developed.Firstly,more than 100 years of geomagnetic observatory data and more than 8 years of Swarm satellite data are collected to establish the longest data set.Then,by considering the different characteristics of the observatory and satellite data,a satellite and observatory data processing method integrating ionospheric and magnetospheric current information is developed based on Fourier analysis,spherical harmonic analysis and Robust estimation methods.Finally,data processing software is independently programmed to estimate the frequency-domain response functions from the original time series,and new satellite and observatory response functions are obtained,which provide new and reliable data sources for the subsequent one-dimensional(1D)and 3D inversions.3.A trans-dimensional Bayesian 1D joint inversion method is proposed for satellite and observatory geomagnetic data based on the idea of iterative correction of the ocean effect.Firstly,using the trans-dimensional Markov chain Monte Carlo sampling algorithm,our algorithm overcomes the disadvantage of uncertainty estimation and the need to manually set the model layer number in conventional geomagnetic data inversion,which makes it possible to automatically determine the electrical interface of the mantle with geomagnetic data.Then,to solve the problem that the data of coastal observatories and geomagnetic satellites would be affected by the ocean effect,an iterative ocean effect correction and 1D inversion algorithm is developed.Finally,the possible lithosphere-asthenosphere boundary and transition-zone interface beneath North America are detected by joint inversion of magnetotelluric and geomagnetic depth sounding data.The horizontal conductivity variation of the upper mantle and transition zone in continental China is revealed by the joint inversion of multi-source data from geomagnetic observatories covering China.A new global 1D conductivity model is established by the inversion of satellite geomagnetic data.4.A global 3D mantle conductivity inversion method based on the fusion of geomagnetic satellite and observatory data is developed,and a new 3D mantle conductivity model is presented.Firstly,the objective function of joint inversion combining geomagnetic satellite and observatory data is established based on the regularized inversion theory.Based on the adjoint principle,the fast calculation method of objective function gradient suitable for ground and satellite data is derived in detail.Then,considering the different characteristics of observatory and satellite data,the inversion model parameterization method based on tetrahedral mesh and spherical harmonic expansion is implemented respectively,which improves the flexibility of inversion.Then,the correctness,reliability and efficiency of the inversion algorithm are verified with synthetic data.Finally,a new global 3D mantle conductivity model is presented by integrating geomagnetic satellite and observatory data.