Service: Constrained inversion honors additional data
WHEN explorers carry out above-ground gravity surveys, ideally they'd like a large density contrast between barren host rock and the mineralization they're targeting. But if the contrast is so minimal as to make mineralization essentially undetectable this way, all is not lost. Constrained gravity inversion may help in cases where below-ground measurements are available, such as drillcore or borehole density measurements.
The commonly used technique of inverting airborne or ground-based gravity data — to produce a density model of the underlying geology — is an example of an unconstrained gravity inversion.
But let's say you also have a reliable, alternative set of gravity data points on hand, such as below-ground measurements obtained from a drilling campaign. You know your inverted model must conform to these ground-truthed measurements.
By inverting your airborne data and while honoring the alternative data, your model becomes an example of a constrained gravity inversion.
In this way, combining the information in two datasets into one constrained inversion can outperform the use of two individually-processed datasets.
The general method is:
- Construct a density reference model from below-ground density measurements. This involves using an interpolation method such as kriging to fill gaps between data points.
- Create a constrained inversion by inverting the surface gravity data while honoring the below-ground data. This may include using a weighting process. For instance, in a location close to a drillhole, the inversion results will honor the data there exactly, whereas the iterative inversion-finding process is permitted increasing flexibility with increasing distance from drillhole data.
- Calculate a residual density model. This is done by subtracting the density reference model from the results of the constrained inversion.
- Analyze the residual density model. Wherever the residual model shows high anomalies is where there seems to be a dense body that remains untested by drilling. Anomalies identified this way may represent a good shortlist of brownfields exploration targets. Fathom Geophysics recently assisted Vancouver-based First Quantum Minerals with constrained gravity inversion work on the company's 100%-held Kevitsa nickel-copper deposit in northern Finland, where mine development has been progressing.
First Quantum Minerals, a copper and gold producer at its African operations, has been working towards diversifying into nickel production. Part of those nickel production aims (as understood at the time of writing) involve an anticipated start to commercial production from an open pit at Kevitsa in mid-2012.
Ahead of this milestone, the company has been working on locating further mineralization at this low-grade, high-tonnage deposit.* This brownfields exploration program targets sulfides that to date escaped detection in stand-alone resource density modeling.
As is the case for many intrusion-related nickel sulfide deposits, Kevitsa's surrounding lithologies are dominated by mafics and ultramafics, which are relatively dense. They provide little density contrast when conducting surface-gravity surveys in search of targeted sulfides, which are also dense. Kevitsa's geological model includes largely horizontally-trending rock units of gabbro, dunite (which is mostly made up of the mineral olivine), olivine pyroxenite and websterite (which is mainly made up the minerals orthopyroxene and clinopyroxene) (see Figure 1).
FIGURE 1: Geological model at First Quantum Minerals' Kevitsa nickel-copper deposit in Finland. The lithologies there, which are dominated by mafics and ultramafics, largely trend horizontally.
Above-ground data available over the planned open pit at Kevitsa included a detailed ground-based gravity survey done using gravity station intervals of 10 metres and with a line spacing of 50 metres (see Figure 2).
FIGURE 2: Above-ground gravity data available in the Kevitsa nickel-copper deposit area. The outline shows the trace of the planned open pit.
Below-ground data available included 84 holes of in-situ density measurements and specific-gravity measurements of drillcore material (see Figure 3).
FIGURE 3: Below-ground data available at the Kevitsa nickel-copper deposit. Colors along the drillhole locations represent density log data. The planned open-pit shell is also shown.
Fathom Geophysics built a density reference model using a radial basis function interpolation of below-ground data (see Figure 4). Rather than being isotropic (where treatment is equal in all directions), the model incorporates a horizontal bias, in keeping with the geological model understood at Kevitsa.
FIGURE 4: Density reference model built using radial basis function interpolation of below-ground data. The model includes a horizontal bias to reflect lithological trends appearing in the Kevitsa geological model.
Fathom Geophysics carried out a constrained inversion of the ground gravity data, honoring the below-ground data, and then produced a residual density model, also called a difference model. The model results were supplied to First Quantum Minerals in voxel (volumetric pixel) data format.
From there, First Quantum Minerals undertook analysis of the differenced results. Imaging of the difference model shows density differences greater than 0.15 grams per cubic centimetre as orange- and red-colored pods (see Figure 5). These high-density bodies could represent sulfides not yet intersected via drilling — potential targets, in other words. The greenish point cloud shows the density reference model's 3.15 g/cc shell.
FIGURE 5: Residual density model (also called the difference model) calculated by subtracting the density reference model from the results of the constrained inversion. Density differences greater than 0.15 grams per cubic centimetre are shown as orange and red pods. The green point cloud shows the location of the 3.15g/cc density reference model shell.
Extended analysis of results by First Quantum Minerals, which incorporated other information such as electromagnetic data on hand, revealed that some differenced density anomalies and electromagnetic plates were quasi-coincident (see Figures 6 and 7).
FIGURE 6: Wide-view image showing extended analysis of the difference model at Kevitsa. Some differenced density anomalies and electromagnetic plates were quasi-coincident. The planned open-pit shell is also shown.
FIGURE 7: Image showing a more detailed view in the extended analysis of the difference model at Kevitsa.
* As at 31 December 2009, Kevitsa was estimated to host NI-43-101-compliant proven and probable reserves of 107.5 million tonnes grading 0.296% nickel, 0.272% nickel sulfides and 0.418% copper (given a 0.13% nickel cut-off grade). The mineral reserves statement also included gold, paladium, platinum and cobalt contents.