%0 Journal Article %A Li, Yongxing %A Smith, Richard S. %T Modelling and straight‐ray tomographic imaging studies of cross‐hole radio‐frequency electromagnetic data for mineral exploration %D 2018 %J Geophysical Prospecting, %V 66 %N 2 %P 282-299 %@ 1365-2478 %R https://doi.org/10.1111/1365-2478.12516 %K Tomography %K Electromagnetic %K Numerical modelling %I European Association of Geoscientists & Engineers, %X ABSTRACT Radio‐frequency electromagnetic tomography (or radio imaging method) employs radio‐frequency (typically 0.1–10 MHz) electromagnetic wave propagation to delineate the distribution of electric properties between two boreholes. Currently, the straight‐ray imaging method is the primary imaging method for the radio imaging method data acquired for mineral exploration. We carried out synthetic studies using three‐dimensional finite‐element modelling implemented in COMSOL Multiphysics to study the electromagnetic field characteristics and to assess the capability of the straight‐ray imaging method using amplitude and phase data separately. We studied four sets of experiments with models of interest in the mining setting. In the first two experiments, we studied models with perfect conductors in homogeneous backgrounds, which show that the characteristics of the electromagnetic fields depend mainly on the wavelength. When the borehole separations are less than one wavelength, induction effects occur; conductors with simple geometries can be recovered acceptably with amplitude data but are incorrectly imaged on the phase tomogram. When the borehole separations are longer than two wavelengths, radiation effects play a major role. In this case, phase tomography provides images with acceptable quality, whereas amplitude tomography does not provide satisfactory results. The third experiment shows that imaging with both original and reciprocal datasets is somewhat helpful in improving the imaging quality by reducing the impact of noise. In the last experiment, we studied models with conductive zones extended into the borehole plane with different lengths, which were not accurately recovered with amplitude tomography. The experiment implies that it is difficult to determine the extent of a mineralised zone that has been intersected by one of the boreholes. Due to the large variation of the wavelength in the radio‐frequency range, we suggest investigating the local electric properties to select an operating frequency prior to a survey. We conclude that straight‐ray tomography with either amplitude or phase data cannot provide high‐quality imaging results. We suggest using more general methods based on full electromagnetic modelling to interpret the data. In circumstances when computational time is critical, we suggest saving time by using either induction methods for borehole separations less than one wavelength or wave‐based methods (only radiation fields are considered) for borehole separation larger than two wavelengths. %U https://www.earthdoc.org/content/journals/10.1111/1365-2478.12516