Inverse and Forward Modelling Using Random Dipoles — Case Study

Abstract

A recently published method of automatically finding magnetic depths to magnetic layers is demonstrated, finding depths to significant details in the magnetic basement under Melville Island, Northern Territory.

By using a method that simulates a magnetic basement as a series of layers of random dipoles, the depths to basement are satisfactorily obtained. Multiple layers appear in the results. However the inversion method used has coarse horizontal resolution, and the layers may be separated horizontally within the sample.

To resolve the ambiguity a forward model, also composed of layers of dipoles, is built on the information obtained from the inversions. Forward modelling requires Fourier convolution for speed. The cycle of analysis is logically completed by comparing the synthetic depth profile with the depth profile obtained by inverting the survey data.