大地电磁测深和重力数据贝叶斯联合反演

Bayesian Joint Inversion of Magnetotelluric and Gravity Data

贝叶斯反演是地球物理联合反演领域的研究热点。目前,虽已有其他方法之间的贝叶斯联合反演文献发表,但鲜有文献涉及大地电磁测深(MT)和重力数据贝叶斯联合反演。现基于非常快速的模拟退火(VFSA)算法实现了MT和重力数据贝叶斯同步联合反演。将MT和重力联合反演的解表达为电阻率和密度的后验概率密度分布,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,参考多种解的表达有助于获得更加接近实际的解,认识解的多解性和不确定性。

Bayesian inversion is the hotspot of geophysical inversion research. At present,lots of papers about Bayesian joint inversions of other geophysical methods are published,but less work in the joint inversion for data of magnetotelluric（ MT） and gravity. The synchronous Bayesian joint inversion of MT and seismic were realized based on very fast simulated annealing（ VFSA） method. The solution of joint inversion for magnetotelluric and gravity data was expressed as posterior probability density distribution for resistivity and density,and extract the maximum posterior solution,the mean solution and final iterative solution of resistivity and density. Testing on synthetic model shows that,reference on multiple solutions can help us to obtain a solution which is close to the fact and understand the non-uniqueness and uncertainty of solutions.