Robust scanning of AEM data for IP effects

The industry has widely accepted that AEM data are more frequently affected by induced polarisation (IP) effects than previously acknowledged. However, we still lack a clear understanding of how much, where, and when IP is present. Full modelling of airborne IP (AIP) is time and computationally intensive. As an alternative, we derive a novel tool, the “AIP scanner”, based on a combination of extensive data–space and limited model–space analysis. The basic assumption is that failing to model IP, when present, increases AEM inversion misfits. Several data space metrics, on negatives and on decay rates, are correlated to misfit from inversion ignoring IP over a small portion of the dataset. The correlation is used to predict the presence of AIP over the entire dataset. The last step is a recursive comparison between the map of predicted AIP and the results of full AIP modelling over a few selected lines. The resulting “AIP scanner” map indicates areas of definite AIP effects, areas possibly affected, and areas probably unaffected by AIP. Such maps are extremely useful tools for the exploration industry wishing to leverage AEM data information content. A case study from South Australia illustrates the scanner results relative to mapped geology and demonstrates the relationships between chargeability and the often unpredictable consequences for the resistivity inversions.