ON THE IMAGING OF RADIOFREQUENCY ELECTROMAGNETIC DATA FOR CROSS-BOREHOLE MINERAL EXPLORATION

Radio-frequency (typically from 0.1 to 20 MHz) electromagnetic methods are powerful tools for locating conductive mineralization in ore expl oration and mine development. Yet data interpretation is complicated b y the non-linear relationship between the observed electric and magnet ic fields and the electrical parameters of the Earth. The principal me ans of quantifying inversion capabilities is to compute synthetic data sets using accurate numerical models and to perform the inversion und er controlled conditions. Our specific interest is in locating 3-D bod ies that are highly conductive relative to the host rock. An excellent approximation for this class of targets, at least at radio frequencie s, is to assume that the bodies are infinitely conductive. The numeric al advantage of this assumption is that inhomogeneities can be represe nted simply as internal boundaries where the total electric and magnet ic fields are identically equal to zero. Ensuring numerical stability thus does not require excessive discretization in conductive regions s ince the maximum grid cell size is determined only by the electrical p arameters of the host material, We use a finite-difference time-domain approach to compute the total electric and magnetic fields everywhere within the background medium and validate the code by comparisons wit h two analytical solutions. One common means of interpreting radio-fre quency electromagnetic data is to assume linearity between the model p arameters and physical response and to apply tomographic image reconst ruction methods. While relatively simple and inexpensive, the limitati ons and applicability of tomographic imaging methods to non-linear ele ctromagnetic data acquired in complicated, 3-D mineral exploration env ironments are not well understood. Our initial study involves applying the simultaneous iterative reconstruction technique to recover images of the electrical properties of a conductive inclusion. Several examp les show that the structural geometry of bodies between boreholes can be reliably imaged using both frequency-domain and time-domain data. P hase data seem more amenable to recovering geometry information from t omographic reconstruction methods than amplitude data. However, attenu ation data provide better constraints on the electrical properties of the geological media and thus form an essential complement to primaril y geometrical information obtained from phase tomography. Non-linear i nversion methods will probably be required to incorporate the amplitud e data for accurate reconstructions of the subsurface.