Three-dimensional inversion of SQUID TEM data at Lalor Lake VMS deposit

Abstract

The combination of high fidelity and late time channels make SQUID magnetometers an attractive sensor for deep and highly conducting targets. Here we focus upon a data set acquired over the Lalor Lake VMS deposit. The deposit consists of a zinc zone between 700m and 1000m depth and a deeper gold/copper zone. The SQUID data set had previously been interpreted using the plate modeling to yield two conductors respectively corresponding to the zinc zone and the gold-copper zone. In this paper we invert these data with our 3D TEM inversion algorithm to produce a 3D voxel conductivity volume. Our 3D model recovers a shallow dipping conductor, which coincides well with the plate model of the zinc zone, as well as a large conductor below 1000m depth. Our inversion is done without incorporating prior knowledge and there is always the potential that large features at depth can be an artifact of the inversion algorithm. To investigate this, and to have confidence, or not, in the existence of the deep body, we carry out a hypothesis analysis where we attempt to find a model that does not have a high conductivity at depth. Forward modeling and subsequent inversion confirms that there must be another conductor below the zinc zone. Much of the concentration of conductivity lies near the region indicated by the initial blind inversion, but the amplitudes and distribution are different. Nevertheless, despite lack of knowledge about geometric details, there is some highly conductive material at depth and it would warrant a drill hole. In a final analysis we look at the relative merits of using B or dB/dt for the particular geometry of this survey. We generate synthetic B and dB/dt data based on our inversion model of Lalor Lake deposit. While the B-field data inversion recovers the correct locations and geometries of the two compact conductors, the dB/dt inversion shows the shallow conductor in a distorted geometry and the deep conductor as a blurred conductive region with conductivity much smaller than the true model. This demonstrates that B-field data can be are superior to dB/dt data for this survey in which we have surface transmitters and receivers.

3D conductivity models of Lalor Lake VMS deposit from ground loop and airborne EM data sets

“…We are also interested in the credibility of the deep extension of the conductor as revealed by the ground data inversion. Our previous work using inversion with hypothesis testing (Yang and Oldenburg, 2012) indicated that some conductive material at depth is required to fit the ground-based data.…”