b'FeatureDiscovery and geophysics of the Khamsin depositthe plate. It is safe to conclude that Plate 5 exhibits GCG effectsData CDIs and depth slices at mid times and LMI effects at late times. Both phenomenonThe Khamsin survey data were converted to apparent conductivity occur at the plate but the CDI algorithm places the LMI apparentand apparent depth for all stations and all times using the method conductivity high at great depth because it is not dominantdescribed for Figure 8.2 (but using a depth adjustment factor until after the dissipation of the host currents at late times (earlyof 2.3 to distribute the error between shallow and deeper workers would have recognised the Plate 5 anomaly in Channelsfeatures). Interrogating the resulting metafile then permitted 25 to 33 in Figure 8.1 as an LMI anomaly to be stripped of anythe generation of the depth slices shown in Figures10.1 to remaining background host response and interpreted using a10.4 and CDIs for all east-west lines, of which one, L6547800N plate-in-free-space code). through the centre of the deposit, is shown in Figure11. Profile Early TEM tests for IOCG mineralisation amplitudes in Figure11 use a sinh scale suggested by James C. Macnae (c.1987, pers comm.), which transitions smoothly A single line of 200 m coincident loop SIROTEM data (5 Hz basefrom linear at low values to logarithmic at high values. Of the frequency) was collected in the early days at Olympic Dam,CDI schemes reported above, the one described here is not but it was concluded that the last channel (50 ms) was notnecessarily the best; rather, it is the only one available to us. late enough to see the basement. Esdale et al. (1987, 2003), however, assure us on the basis of IP/Restivity surveying, thatPlate-in-host model the ore exhibits higher electrical conductivity and polarizabilityThe modelling code that produced Figures 8.1 and 8.2 was than the country rock. Hart and Freeman (2003) come to thethen used in an attempt to simulate the profiles and a CDI for same conclusion for Prominent Hill based on down-hole andLine 6547800N using the same specifications that were set for the surface surveying. Finally, a single (unpublished) line of in-loopsurvey equipment. The result is shown in Figure 12.1. The model TEM (1.6 Hz base frequency; last channel at 148 ms) was readcomprises flat lying (non-interacting) plates in a host environment by TeckCominco at Carrapateena, for which a CDI showed rocksjudged to be similar to that at Khamsin. Early models yielded with elevated conductivity starting near the known depth ofCDIs, which, upon comparison with the data CDI(Figure 11), gave mineralisation. Consequently, a similar prognosis would seemvisual indications of how the model had to be modified and/or reasonable for the mineralisation at Khamsin. expanded to better simulate the data. The computed response Khamsin TEM data of the thin plates shown in Figure 12.1 is the best model found so far, and while it remains an imperfect simulation of the data Survey designprofiles and the data CDI, there are enough similarities to make it worthy of further consideration. As stated above, a CDI is not a In 2018, subsequent to the Khamsin discovery, OZ Minerals Ltd.model; the model, now offered as an interpretation of the data, carried out TEM surveying to determine if IOCG mineralisationis the 0.025 S/m halfspace and the nine plates shown in the generates recognisable responses in CDIs for the area.lower half of Figure12.1 and described in Table2. Theoretical responses were used to assess depth penetrationInterpretation of the CDIsand target sensitivity for the, by then, known conditions at nearby Carrapateena. Based on the results, the selected fieldForward modelling is an arduous, computationally intensive system was a Geonics Ltd transmitter to generate fields withtask, but if it can be agreed that the above model for L6547800N an exponential-on ramp-off bipolar current waveform, a 1 Hzis reasonable, then with little further effort, the depth slices base frequency and a ramp-off duration of 750 microseconds.and the CDIs generated from the full dataset can be taken as a A SMARTem receiver collected in-loop B-field measurementsqualitative, or at best, a semi-quantitative interpretation of local binned into 36 time channels ranging from 0.094 to 225.8 ms.electrical structures (with due regard for the misrepresentations Transmitter loops of 200x200m size were deployed with 100mof any LMI anomalies). stations along 200 m spaced lines. Figure 10.1 shows a near ubiquitous conductive layer at about 100 m depth. At 450 m depth, Figures 10.2 and 11 suggest that the part of the ovoid not covered by the conductive layer exhibits an apparent conductivity high at about the unconformity depth, but a similar claim for the 600 and 900 m depth slices (Figure 10.3 and 10.4) is less convincing.The interpreted palaeochannel, which was so disruptive to the gravity, seems also to disrupt the mid-level depth slices (Figures10.2, 10.3 and possibly 10.4 in the far north) and appears as a linear conductivity low following the same dotted line shown in Figure 3.2. While interpreting this feature as a palaeochannel might seem reasonable as long as only the geophysical data was available, subsequent drilling suggests a normal fault with a slight downthrown eastern side. Porous Whyalla sandstone units are in contact on either side of the Ch 110.87 ms Ch 3053.0 ms interpreted fault, and, to the east, at least, these rocks contain hypersaline water. For the linear feature to appear resistive Figure 9. Profiles of H zdue to currents on 500x500m red plate at 550mwould seem to require low salinity water sealed off from the secdepthwith vector display of Hsec . Even numbered channels are shown in yellow.saline waters in the shallower (?) sandstone, and, if the 900 m Black profiles show Channels 11 and 30. Scale bar ranges from 0to1pT/A. deep conductivity low seen in Figure10.4 is real, then a fault 48 PREVIEWFEBRUARY 2024'