Leveraging the true value of historic AEM data sets with quantitative inversion

Large holdings of airborne electromagnetic data exist in public archives of government agencies that have never been fully utilised. Many of these surveys would have only been interpreted as profile data to identify discrete conductor anomalies, which may in turn have been subjected to parametric plate-modelling techniques. Most datasets from prior to the turn of the century will not have been interpreted with the stitched 1D inversion algorithms that are now used as a routine practice. Since historic datasets were not routinely quantitatively inverted at the time, there was no driver for survey clients to demand accurate system specifications or calibrations. Added to this is the difficulty of dealing with data that are presented in units of parts-per-million, as well as the lack of measurement of the transmitter-receiver separations and orientations. Despite these impediments, we have been motivated to ascertain if further value can be extracted from historic data sets through quantitative inversion techniques. We have investigated various techniques in an attempt to overcome these impediments. This includes using inversion parameter-sweeps on selected subsets of a data set to derive calibration parameters that best represent the known unknowns in the data. Examples of which are the waveform pulse-width and the high-altitude reference geometries by which data are provided in the case of parts-per-million normalised systems. A new bunch-by-bunch inversion algorithm, which allows along-line system geometry constraints to be applied, is also used in the procedure. The methodology shows promise of remedying the issues for some data sets, but good results are not always achievable.