b'FeatureDiscovery and geophysics of the Khamsin depositconductances because they cannot gather more current than is flowing in the host and do not highlight high conductance features as well as LMI effects. One further complication is that LMI currents in a confined target decay exponentially, which is faster than the t3/2 power-law decay of the fields in the semi-infinite host. Consequently, GCG effects dominate again at very late times, although this may be later than the last channel.The complex 3-dimensional interplay of the electric and magnetic effects in a semi-infinite medium with inhomogeneities does not make for easy conceptualisation. While geophysics students are usually taught about magnetic induction, it seems that the GCG phenomenon is often ignored so that graduates may be prone to entering the exploration workforce ill-equipped to understand, and maybe disinclined to believe, that important information can reside in the early and mid-times, and not just in the late times, of TEM datasets. Figure 8.1. CDI derived from profiles of model response at 36 measurement Conductivity-depth images times. Apparent conductivity highs and lows relate to bumps and troughs in During or shortly after the development of comprehensivetheimages. modelling codes, workers like Macnae and Lamontagne (1987), Fullagar (1989), Smith et al. (1994) and others, began creating conductivity-depth images that transformed inscrutable profiles of TEM channel amplitudes into intuitively credible Conductivity-Depth Images (CDIs) for the earth under a given survey line. Such images are not models, but they provide a more effective way of viewing TEM data because of the knowledge imparted about the existence and approximate locations of conductive or resistive features that could be of interest. By way of a numerical experiment, the upper half of Figure 8.1 shows profiles of the modelled responses of five flat-lying plates in a 0.022 S/m halfspace for the in-loop prospecting systemFigure 8.2. Using the integrated conductance over all shallower depths to discussed below. The lower half shows a CDI derived from theestimate a mean conductivity encountered by the descending smoke-ring profiles, which uses Nabighians theoretical descent rate ofplaces the maximum apparent conductivity high at the plateat least for the the smoke-ring maximum to assign an apparent depth to theshallowest plate at a given location. Profiles are the same as in Figure 8.1.apparent conductivity for each reading (while holding in reserve one final depth adjustment factor). Algorithms by Hanneson and West (1984) and Holladay (1981) were used to generate theInterestingly, for a given point on the image, integrating all profiles shown in Figure 8.1. shallower conductivities seems to give a better estimate for In consideration of the overall process, there must be time fortheeffective conductivity of the path followed by the smoke-the field disturbance caused by transmitter shut-off to diffusering, so that when the new descent rate and the time are used to target depth, excite scatter currents on the target, and for theto assign a depth, the location of the maximum conductivity anomalous scatter currents to radiate their fields back to thehigh is much closer to the known depth of a plate in the model, surface (similar to the notion of a two-way travel time in seismicbut unfortunately the one-to-one correspondence between processing). In consideration of this, the CDI in Figure8.1,highs in the image and highs in the profiles is no longer results from a final adjustment factor of 2.0, and while it placesevident.See Figure8.2 (for which the profiles are the same as the apparent conductivity highs at about the right lateralinFigure 8.1). location, the highs are slightly deeper than the known depthsPlate 5 has a high enough conductance (1000 S) to generate an of the plates. Nevertheless, this image is reassuring becauseLMI anomaly that persists at least until Channel 33. It exhibits bumps (and troughs) in the channel amplitudes corresponda deep (1200 to 2000 m) conductive lobe in Figure 8.1, but is to relative conductivity highs (and lows) in the image. Everygrossly misrepresented in Figure 8.2. fifth channel is annotated to make clear how each point on a profile relates to its corresponding point on the CDI. Plate 1 atFigure 9 simulates a fixed loop survey for a 200 x 200 m loop 90 m depth exhibits a bump in Channel 1 (0.10ms) because theabove the left edge of the plate and helps explain what is measurement post-dates the arrival of the smoke-ring at thehappening on the plate at station 731400E in Figure 8.1. It target (plus the return of the scattered fields). Plate 2 is resistiveshows the secondary Hz profiles (all channels: yellow) with (negative conductivity contrast) and exhibits a conductivityHsec vectors near the red horizontal Plate 5. At 0.87 ms, the low. Plates 3, 4 and 5 are deep enough that the smoke-ringHsec vectors circle the entire plate from which can be inferred travel time (multiplied by 2) is later than Channel 1 so that thea bundle of gathered, unidirectional (GCG) current filaments shallowest conductivities approximate the intrinsic conductivityflowing into the page. At 53.0ms, the Hsec vectors have a of the host. This last effect is called early-time blanking and hasdipolar form centred on the plate and are caused by closed been mentioned by Rai (1982) and others.horizontal loops (or vortex) of magnetically induced current on FEBRUARY 2024PREVIEW 47'