Deep conductor delineation through improved EMFlow data processing

Abstract

The EMFlow program is an industry standard for the rapid transformation of AEM data to a conductivitydepth section. Written in the 1990?s, it provides a means to deconvolve data from any definable AEM system to the time-constant domain, and thereafter to predict conductivity-depth structures. The deconvolution to time-constant domain requires constraints for optimum performance: these constraints are user-controlled and can include causality, power-law or exponential extrapolation and geometrical amplitude constraints. Geotem Deep and VTEM data collected over the Table Hill prospect in WA showed very small amplitude responses, which when processed with commercial EMFlow indicated an extensive conductor at depth. Initial drilling encountered manganese mineralisation consistent with this deep conductor. However, step-out drilling did not encounter similar mineralisation even though the holes were sited to intersect the imaged good conductor. The data were therefore reprocessed with an improved deconvolution method in an RMIT version of EMFLow. The specific improvement trialled was a novel matrix conditioning method that provided significantly more stable CDI results than simple matrix normalisation used in the commercial EMFLow program. The resulting CDI sections showed considerable detail within the deep conductive layer, where the most conductive region was consistent with the discovery hole, and less conductive regions were coincident with the lack of mineralisation in the step-out drilling. Improvements in data processing therefore were shown to provide very useful if not essential geological detail at depth.