Making use of all available geoscience data to maximise exploration success
Energy Resources of Australia (ERA) has been producing uranium in the Northern Territory for more than 35 years from the two Ranger deposits (Pits 1 and 3). Many of these years have been spent exploring for additional uranium deposits with considerable success, with up to nine smaller deposits discovered on the Ranger Project Area (RPA). Most of these discoveries are associated with a radiometric signature, derived from near surface uranium mineralisation. Since the discovery and recognition of the world class Ranger 3 Deeps deposit, the challenge to discover another uranium deposit at depth elsewhere on the RPA has been presented.
Unfortunately, like a lot of Australia, the majority of the RPA is overlaid by a thick layer of cover, masking the basement rocks and limiting the effectiveness of some geophysical methods like radiometrics. To help in exploration, ERA has an impressive array of geophysical surveys, geochemical assays, downhole geophysical measurements, geological knowledge and logging at its disposal. The challenge was to combine all the data into a single useful platform and 3D model to create exploration targets based on well-defined criteria and learnings already established on the RPA. The 3D model was interrogated using numerous geophysical inversions and forward modelling techniques to further refine the exploration targeting.
The geology of the RPA is very well established given the amount of work that has been completed. Away from the immediate mining area, downhole geological logging was used to ‘snap’ surfaces to known stratigraphic boundaries and interpolated through the model area. Ranger stratigraphy is separated into distinct stratigraphy sequences – the hanging wall sequence (HWS), the upper mine sequence (UMS), the lower mine sequence (LMS), the foot wall sequence (FWS), the Kombolgie sandstone (KOM) and surficial (SUR) surfaces. Other lithological horizons such as magnetite-rich schists and amphibolites were also included in the model.
Recent refinements in the detailed proto-stratigraphy of the Ranger 3 Deeps deposit have further divided the UMS into two major subdivisions: a basal chlorite schist dominated member (the ore host stratigraphy) and an upper meta-arenite member. The basal chlorite schist is upward-coarsening and transitions into a sequence of graded meta-arenites via a package of laminated meta-silts (Pevely et al, 2017). This detailed proto stratigraphy is also recognised in exploration holes away from the Ranger 3 Deeps deposit.
ERA has a comprehensive petrophysical database that has been built over numerous years of mining and exploration. Density and magnetic susceptibility measurements are regularly collected in the core shed by the water immersion method and hand-held magnetic susceptibility meter measurements. Select core samples are also sent off periodically and analysed for density, magnetic susceptibility, resistivity, conductivity, porosity, etc. As part of the resource estimation process, density is regularly measured of both ore and gangue rock types.
Every drill hole completed at ERA is measured for natural gamma radiation. As part of this process, a resistivity tool is often ‘stacked’ with the natural gamma probe, as well as magnetic deviation tool for surveying purposes. As such, ERA has an enormous downhole gamma database as well as an extensive downhole resistivity database. These data have proven invaluable for the definition of lithological changes, stratigraphic boundaries and even structural analysis of the orebodies and in exploration.
Downhole density and magnetic susceptibility probes have also been used periodically to obtain density and magnetic data across the RPA. All these petrophysical property data were imported into the model, and used to constrain geophysical inversions.
Analysis between the downhole resistivity measurements and local stratigraphic boundaries revealed some surprises. Decreasing resistivity (and increasing conductivity) could clearly be seen in the downhole geophysics. Pre-feasibility studies during the development of the Ranger 3 Deeps deposit now show this upward fining of the stratigraphy underpins and explains the morphology and structures of the deposit. This provided confidence that inversion of the data (electro-magnetics) would be providing information at the target stratigraphic boundary and targeting the same prospective stratigraphic package.
A key fault controlling mineralisation in the Ranger 3 Deeps system could also be recognised in the downhole resistivity data (Figure 1). Previously this structure had only been defined by geochemical assays and broad cross-sectional interpretation. The structure occurs in the barren HWS, which is most often reverse circulation drilled, so structural analysis is difficult. As it is barren for uranium mineralisation, assay coverage is also poor. As downhole geophysics is collected in every drill hole, the structure could be much better defined. This suggested resistivity could play an important role in exploring for uranium-bearing structures elsewhere on the RPA.
ERA has a high resolution airborne electromagnetic (AEM) survey dataset covering the entire RPA. The survey shows numerous conductive anomalies across the lease. A 3D inversion of the conductivity data was completed and draped over a key exploration stratigraphic boundary (Figure 2). This immediately highlighted numerous areas for further investigation.
Plate modelling was used to test the plausibility of a geological model of a strong conductor at the top of a stratigraphic unit and to characterise the conductivity of this unit. Five tie lines from the AEM survey (the tie lines orientated east-west in the preferentially orientated direction for a gently east dipping formation) were selected. Initial results suggested plate modelling is a good reflection of the actual geological conditions – an isolated conductor surrounded by resistive material. The conductor in this case is a graphitic breccia present at the boundary between the overlying schists and the underlying carbonates.
Further inversions were run to explore variations in the thicknesses of the modelled units. Two initial inversions were run: a geometry inversion to adjust thicknesses of the conductive layer, and a conductivity inversion to adjust its conductivity. The geometry inversion increased the thickness of the starting unit significantly, suggesting the modelled stratigraphic unit varies in thickness along the boundary. This has been noted already in the sparsely populated exploration drill holes away from the mine area but not extrapolated or properly constrained in 3D. These thicker zones may present as exploration targets. The final model also showed conductivity anomalies present in the eastern portion of the model, which were unexplained by the geological model. These also presented as exploration targets. A final geological model of a strong conductor at the stratigraphic boundary, representing the graphitic breccia, and a horizon above this unit of moderate conductivity achieved a very good fit after inversion. This unit is representative of the lower portions of the UMS schists, which have lower resistivities than the overlying UMS meta-arenites, as evident in the downhole resistivity data shown in Figure 1.
The total magnetic intensity (TMI) image over the RPA is dominated by strong linear anomalies that follow the regional strike of the geology. These magnetic anomalies are due to magnetite in the micaceous HWS schist. There are numerous ‘breaks’ in the anomalies that are of exploration interest. These ‘breaks’ are almost certainly faults, or are pathways of magnetite-destructive fluids. These pathways may have also been conduits for uranium-bearing fluids at some stage in the past.
The magnetic horizons were interpreted and constructed in the model based on the aeromagnetic data, downhole geological logging and downhole geophysics if available. The horizons were then forward modelled and adjusted if necessary. The model was updated and forward modelling process repeated until a geologically reasonable fit was achieved. A geometry inversion of the geological domains was then completed and the modelled horizons finalised. Finally, any breaks or magnetite depletion zones present in the horizons were identified and highlighted as possible structures. The interpreted faults were then ranked and targeted in conjunction with other factors in the model.
Forward modelling – gravity
One weapon missing from ERA’s exploration arsenal is complete gravity coverage. A ground gravity survey had been conducted in the 1980s, but coverage was patchy and the interpretation limited to broad stratigraphic boundaries. A more complete ground gravity or airborne gravity gradiometry survey could prove helpful to exploration.
A key geological phenomenon associated with the Ranger 1 and Ranger 3 deposits (and to a lesser extent the Ranger 3 Deeps) is slumping associated with hydrothermal dissolution of the underlying LMS carbonate. The absence of carbonate is a key trap site for uranium mineralisation in Ranger Pit 3, with some slumping and faulting of the LMS carbonate controlling mineralisation in the Ranger 3 Deeps. A density contrast of 0.15 g/cc exists between the overlying UMS schists (2.67 g/cc) and LMS carbonate (2.82 g/cc).
A block of lower density representing an absence of dense carbonate was created based on the ‘trough’ present in the LMS stratigraphic surface over the Ranger 3 pit and Ranger 3 Deeps areas. A dip of 25 degrees to the east was applied to simulate the regional dip of the geology. The block was placed at 0, 100, 200 and 400 metre depths and forward modelled to predict the response normally seen in a ground gravity or gravity gradiometry survey.
Amphibolite sills in the HWS also provide a density contrast with the surrounding host muscovite schist (~0.3 g/cc). Large faults have been mapped throughout the RPA with displacement as large as 60 metres of the stratigraphy. A 60-metre offset of an amphibolite sill was forward modelled at 0, 100, 200 and 400 metres to predict the response of both a ground gravity and airborne gravity survey. Various offset thicknesses were also modelled. These predicted gravity results were then analysed to assess how useful a gravity survey could be for use by the exploration team.
This article was originally written in late 2017.
Thank you to Tim Chalke and the Mira geoscience team for their assistance during this project. Thank you also to the ERA exploration team.
Pevely, S, Hinman, M, McLellan, A 2017, Ranger 3 Deeps uranium deposit, AusIMM Monograph 32, 2017 pp. 461–464