Automated anomaly modelling of AEM data with magnetic dipoles buried inside a layered earth

Abstract

An algorithm is described that determines, models and classifies spatially discrete anomalies in airborne EM (AEM) data sets. The first module scans segments of EM profiles for anomalies wide enough to have their origin in the subsurface and narrow enough to be caused by a discrete conductor. Next, background conductivity models are determined with layered-earth inversions from the spatially smoothed EM data. Finally, the identified EM anomalies are modelled first with rectangular current filaments in free-space and then with magnetic dipoles buried inside a layered-earth. The approach used takes into account the effect of overburden blanking and is applicable for discrete, sheet-like conductors inside a resistive host, i.e. in scenarios where vortex currents dominate. Computing the current-excitation ratio for each solution monitors the validity of this assumption. The model parameters determined from each data segment include the target conductor position, depth, dip, size and conductance. The method is fully automated with the filament start model being determined by curve matching from a digital look-up table. Results from synthetic data indicate the efficiency and reliability of the method. Automated anomaly modelling of TEMPEST data acquired across the Bull Creek prospect, Queensland, provides a sensible description of the main mineralisation and indicates the presence of other minor conductors.