b'David Annetts best of Exploration GeophysicsFeatureon a fast PC. Any automated inversion process should avoid local minima, be stable in the presence of noise from any source, account for all physically detectable effects such as IP or magnetic susceptibility, and be tolerant of imperfect system description. A significant research challenge remains to speed up and stabilise this process.AcknowledgmentsWe acknowledge the support of the sponsors of AMIRA project P378 (Pasminco, Aberfoyle, CRA, BHP) and AMIRA project P407 (Pasminco, MIM, Billiton. CRA. BHP. North. World Geoscience), who together with CRCAMET supported the development of software based on the ideas expressed in this paper. The paper benefited immensely from very thoughtful and constructive reviews by Peter Fullagar and Figure 6.An approximate fit of the resistive limit predicted from field dataPerry Eaton. This paper is published with permission of the to a wide dipping target. A single, wide body was not able to fit the peak of theCRC for Australian Mineral Exploration Technologies (CRC field data response. Tbe geology consists of a series of steeply dipping shales inAMET), established and supported under the Australian crystalline host. Governments Cooperative Research Centres Programme.preset iteration limit which can be easily extended by manualReferencesintervention. No simple block model was found that could fit the peak of the resistive limit, where visual inspection showsDyck, A. V, M. Bloore, and M. A. Vallee. 1980. User manual for most difference between the data and the model. The data wereprograms Plate and Sphere: Research in Applied Geophysics 14. obtained over a steeply dipping band of shales hosted in moreGeophysics Laboratory. University of Toronto.resistive crystalline rocks. Eaton, P. A., and G. W. Hohniann. 1989. A rapid inversion With many such anomalies routinely detected in AEM surveys,technique for transient electromagnetic soundings: Phys. faster models or computers are essential if each of the (say)Earth Planet. Int 53: 384404.500 anomalies detected in the course of a day of flying needsFullagar, P. . 1989. Generation of conductivity depth pseudo-to be modelled. If this step could be made more robust andsections from coincident loop and in-loop TEM data. accurate than at present, the task of interpreting local anomaliesExploration Geophysics 20: 4345.contained within the data would be greatly simplified. This mayFullagar, P. K., and J.E., Reid. 1992. Conductivity-depth be achieved in future with better automatic model selectiontransformations of fixed loop TEM data. Exploration and better starting models for inversion. Geophysics 23: 515520.Grant, F. S., and G. F., West. 1965. Interpretation Theory in Applied Geophysics. McGraw Hill.Discussion and conclusions King, A. 1998. Inversion of localised electromagnetic anomalies. The great redundancy in AEM data acquired as equi-spacedPh D. thesis. Macquarie University.samples favours data compression as a first step in processing.Lamontagne, Y,. 1975. Application of wideband, time domain This may involve both time windowing and spatial averaging.measurements in mineral exploration. Ph. D., thesis. As part of this process, it is advantageous to transform data,University of Toronto.whether in time or frequency domain, to time constant (tau)Lamontagne, Y., J. C. Macnae, and B. Poller. 1988. Multiple space in order to remove the waveform dependence of theconductor modelling using program MuliLoop: 58th Annual observed AEM response. Internal. Mtg., Soc. Expl. Geophys., Expanded Abstracts. Session, EM2.5.With many local anomalies and a variable background, the nextLiu, G., and M. Asten. 1993. Fast approximate solutions of logical stage of processing is transformation of the responsetransient EM rsponse of a target buried beneath a conductive to a conductivity-depth image (CDI) to facilitate geologicaloverburden. Geophysics 58: 810817.interpretation of the background and to guide subsequentMacnae, J., and Y. Lamontagne. 1987. Imaging quasi-layered modelling. Use of all the available recorded data, particularlyconductive structures by simple processing of transient the inclusion of on-time data, appears to improve the stability ofelectromagnetic data. Geophysics 52: 545554.the CDI process. Macnae, J. C., R. Smith, B. D. Polzer, Y. Lamontagne, and P. S. Klinkcrt. 1991. Conductivity-depth imaging of airborne The objective of subsequent interpretation in mineralelectromagnetic step-response data. Geophysics 56: exploration is to assess the likelihood that each local anomaly102114.detected represents economic mineralisation. This step involvesMacnae, J. C., A. King, E. Stolz, and P. Klinkert. 1998. 3D EM the extraction of geometrical and conductivity informationinversion to the limit. In Three-Dimensional Electromagnetics, from the AEM data. The only feasible route at the present timeed. M. Oristaglio, and B. Spies. Soc. Expl. Geophys.is to parameterise both the data, to say the inductive and theMencke, W. 1984. Geophysical Data Analysis: Discrete Inverse resistive limit, as well as the model to allow rapid inversion ofTheory. San Diego: Academic Press.the local anomalies. A fit to one or two plate-like conductorsNekut, A. G. 1987. Direct inversion of time-domain can be achieved in seconds; fits to a wide body take minuteselectromagnetic data. Geophysics 52: 14311435.49 PREVIEW OCTOBER 2020'