基于岩石物性约束的大地电磁与重力贝叶斯联合反演

Bayesian joint inversion of magnetotelluric and gravity based on petrophysical constraints

地球物理反演的多解性是固然存在的,贝叶斯反演理论为表达反演结果的多解性提供了依据,是联合反演领域的研究热点.岩石物性约束是地球物理联合反演中不同物性参数之间相互约束的方法之一,本文将岩石物性关联引入到大地电磁测深(MT)和重力的贝叶斯联合反演中,利用不同物性参数的后验概率密度分布来表达解,并提取了电阻率和密度的最大后验概率解、均值解和最终迭代解.模型试验表明:最大后验概率解和均值解相比传统意义上的最终迭代解能更好地还原真实模型,且岩石物性约束的引入可提高物性参数之间的耦合程度.

Solution of the mult-solution is to be sure. Bayesian inversion theory provides a basis for expressing the non-uniqueness of solutions,and is is an active area of joint inversion. Petrophysical constraints is one of the methods of mutual restraint between different physical parameters in a geophysical joint inversion. In this paper,we introduces petrophysical correlation into a bayesian joint inversion of Magnetotelluric sounding（ MT） and gravity. posterior probability density distribution of different physical parameters is used to express solutions, and extract the maximum posterior solutions,the mean solutions and final iterative solutions. Model test shows, compared to traditional final iterative solutions,maximum posterior solutions and mean solutions can better restore real models. In addition, The introduction of petrophysical constraints can increases a coupling degree between the physical parameters.