b'David Annetts best of Exploration GeophysicsFeaturewhere conductivity is derived from a predicted depth to anWhen all-time data is processed and plotted at the same scale image of the source. This method was described by Macnae andas Figure 2, we note as expected a significant improvement in Lamontagnc (1987) and by Nekut (1987). Eaton and Hohniannthe coherence of the processed CDI section (Figure 3). Note (1989), Fullagar (1989), Fullagar and Reid (1992) and Smith,that each of the 1600+ individual stations has been processed Edwards, and Buselli (1994) developed approximate imagingwithout regard to its neighbours. All vertical striping has schemes based on the actual depth of maximum current or thebeen eliminated from the section, and the predicted depth maximum sensitivity to a layer in a half-space. to the top of the shallowest conductor is generally more Given data in the tau domain, it is possible to predict the stepconsistent with the surface (0 m), although this is not obvious response and derive a conductivity-depth-image from AEM dataon the plots. Although drilling and physical property logging by the method described in Macnae et al. (1991). In practice, it isare not available to confirm that this CDI section is more not necessary to predict the step response because processinggeologically correct than that shown in Figure 2, the result can be carried out directly in the tau domain. One of the reasonsis consistent with known geology, although less interesting for using fast CDI transformations based on current diffusion isfrom a target generation point of view. Generally, shallow that CDI sections are equivalent to, or even better than, stitcheddips would be expected at the base of the sediments or at the ID inversions over 2D or 3D conductive structures (Nekut 1987.top of unweathered bedrock as seen in this all-time CDI, and Eaton and Hohniann 1989; Stolz and Macnae 1998). geological units in the local Proterozoic basement are expected to be very resistive. Use of the on-time data has thus allowed The use of on-time data production of a more geologically believable image, and which theoretically should be more sensitive to slow decays indicative Recent work by Smith et al. (1996) and by Stolz and Macnaeof conductive targets at depth below the conductive cover than (1998) suggest that on-time AEM data should extend systemwas the off-time data alone. This conclusion can not however be sensitivity to both very rapid and very slow decays. Figure 1verified from this data set.clearly demonstrates this on-time sensitivity in a plot of the secondary response throughout a typical AEM waveform. WhenParameterising local EM responsesboth on- and off-time data are used to derive the AEM response in the tau domain, we would expect that the resolutionGrant and West (1965) described the inductive and resistive should be more stable over a wider range of tau values, i.e., anlimits of an EM response in frequency domain and suggested increased bandwidth, provided that the primary field is wellthat they were generally diagnostic of the main features of known. The following example demonstrates this be the case. an AEM response from a local conductor in free space. In Figure 2 shows a 20-km line of 37.5-Hz Questem data of 1992essence, the inductive limit represents the response in free vintage collected over an area with conductive surface cover,space when current is confined to the surface of a conductor. and a CDI derived from off-time data using the MaxwellMathematically the inductive limit condition is expressed as receding image technique using program EMFlow developedl21, whereis frequency,conductivity,magnetic within AMIRA project P407. The geology consists of transportedpermeability and l a characteristic dimension. For a given sediments overlying a Proterozoic basement, which is generallyconductor, the inductive limit is a function of geometry only. resistive except where weathered at its present or palaeo- The time-domain step response inductive limit is identical to surface. This horizontal (x) component data was collected inthe frequency-domain inductive limit if scaled by the total summer and is quite noisy (50 ppm) at late delay times. Underprimary field (Macnae et al. 1998)the ID assumptions used in CDI processing, a response of smallThe resistive limit in frequency domain consists of the time amplitude will be interpreted as coming from great depth.derivative with respect toof the secondary EM response Thus responses (of signal, and by inference noise) of small toat the low frequency limit (l2 1). In time domain, the moderate amplitude will mostly show up in the CDI at depth.resistive limit is the area under the step response decay curve The effect of the noise thus appears to cause a number of(Appendix A) and is easy to calculate. At the resistive limit, the narrow responses (vertical striping) to appear at depth in theresponse is a linear function of conductivity and the response CDI, which responses do not correlate between lines. Geologicalof bodies that not in galvanic contact do not interact and are reasoning would indicate that most valid basement conductorslinearly additive (Lamontagnc 1975).would show some strike (across-line) consistency, which is not seen in these CDI images. Figure 4 shows an example of the predicted resistive limit estimated from a segment of AEM field data. Because the response is decomposed into a sum of exponential decays, it is possible to calculate the contribution of each of these decays to the resistive limit. Since a good conductor will have a slow decay, we would expect that the long taus would contribute more to the resistive limit than would be the case for a poorer conductor. The total resistive limit profile shown in Figure 4 on the upper plot characterises the spatial amplitude variation of any local anomaly, as indeed does the raw data plotted on the bottom axis. Variations in the contribution to the total resistive limit from early, intermediate and late taus can be used to discriminate adjacent conductors, as is done using time-constant analysis. Figure 1.Exponential decays convolved with a 1992 Questem waveformThe advantage of the resistive limit is that, unlike raw data, its (received from an approximate half-sine transmitter current waveform): theamplitude is directly proportional to the conductivity of the time constants range from 0.01 ms (short) to 10 ms (long). associated conductivity structure.OCTOBER 2020 PREVIEW 46'