Factor analysis and factor score imaging of transient electromagnetic model responses over simple geometric bodies

Abstract

A factor analysis-based method has been developed for preliminary interpretation of ground-based, time-domain electromagnetic data. We refer to this method as factor score imaging. It provides data volume and dimensionality reduction, allows efficient approximation of the shape of buried conductive bodies and aids selection of initial model geometry for full inversion. In this paper, the method is applied to modelled amplitude responses of simple geometrical bodies, such as a buried conductive sphere and a dipping plate. It was found that five or fewer factors typically describe the 30-dimensional data. Factor score imaging was developed to visualise the results. In the first factor score images, the footprint of the modelled bodies are recovered with remarkable accuracy. The simple examples described in this paper provide insight into the developed methodology.