Obtaining the Depth of Potential Field Sources

Linear least-squares inversion is a powerful tool which, when applied to potential field data, can yield important insights into the nature of the subsurface geology. Unfortunately, most geophysical problems are non-linear, and the use of linear inversion has several consequences. At best, it now becomes necessary to iterate the inverse process to reach the optimum solution. At worst, the inverse process can diverge to infinity or become stuck in a local minimum of the misfit surface. Noise problems usually prevent second derivative (Hessian) or higher terms being used in the inversion.

This abstract shows how to solve a particular non-linear inverse problem, the depth of one or more potential field sources, in the frequency domain without the need to iterate. The method is applicable to magnetic or gravity data, and can be used with any forward model type (dyke, contact, sphere, arbitrary collection of two or three dimensional bodies, etc).