Rapid mineralogical and geochemical characterisation of the Fisher East nickel sulphide prospects, Western Australia, using hyperspectral and pXRF data

Accepted Manuscript Rapid mineralogical and geochemical characterisation of the Fisher East nickel sulphide prospects, Western Australia, using hyperspectral and pXRF data Lauren L. Burley, Stephen J. Barnes, Carsten Laukamp, David R. Mole, Margaux Le Vaillant, Marco L. Fiorentini PII: DOI: Reference:

S0169-1368(16)30784-3 http://dx.doi.org/10.1016/j.oregeorev.2017.04.032 OREGEO 2200

To appear in:

Ore Geology Reviews

Received Date: Revised Date: Accepted Date:

2 December 2016 24 April 2017 27 April 2017

Please cite this article as: L.L. Burley, S.J. Barnes, C. Laukamp, D.R. Mole, M. Le Vaillant, M.L. Fiorentini, Rapid mineralogical and geochemical characterisation of the Fisher East nickel sulphide prospects, Western Australia, using hyperspectral and pXRF data, Ore Geology Reviews (2017), doi: http://dx.doi.org/10.1016/j.oregeorev. 2017.04.032

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Rapid mineralogical and geochemical characterisation of the Fisher East nickel sulphide prospects, Western Australia, using hyperspectral and pXRF data Lauren L. Burleya, c, Stephen J. Barnesb, Carsten Laukampb , David R. Moleb, Margaux Le Vaillantb, Marco L. Fiorentinic

a Geological Survey of Western Australia, 100 Plain Street, East Perth, 6004

b Mineral Resources, Commonwealth Scientific and Industrial Research Organisation (CSIRO), Australian Resources Research Centre (ARRC), 26 Dick Perry Avenue, Kensington WA 6151

c Centre for Exploration Targeting, ARC Centre of Excellence for Core to Crust Fluid Systems, The University of Western Australia, 35 Stirling Highway, Crawley WA 6009

*Corresponding author: Lauren Burley Email: [email protected] Current address: Geological Survey of Western Australia, 100 Plain Street, East Perth, 6004

Abstract

The use of Shortwave Infrared (SWIR) and Thermal Infrared (TIR) hyperspectral data in mineral exploration has been well documented in many mineralisation types, but is limited in komatiite-hosted nickel sulphide deposits. This project combines hyperspectral, Portable X-Ray Fluorescence (pXRF) and whole-rock geochemical data to assess different analytical techniques in the exploration of these deposits. We use the Fisher East nickel sulphide prospects, Western Australia for our case study. The Fisher East prospects lie in an area of the eastern goldfields that has historically been underexplored and understudied. The rocks have undergone intense deformation with primary igneous textures being destroyed, along with strong alteration to talc carbonate

assemblages.

Combining

different

analytical

tools

allowed

for

differentiation of A and B-zones of original komatiite flows, and the reconstruction of original volcanological facies in a setting where whole rock chemistry as well as igneous textures have been substantially modified by metamorphism. By using different lithogeochemical techniques including pXRF, this study shows the Fisher East prospects are hosted within channelised komatiite flows, and have similar characteristics to Kambalda style deposits.

Keywords: Fisher East, komatiite-hosted nickel, hyperspectral, pXRF

1. Introduction

Komatiite-hosted Ni-Cu-PGE systems are found in many greenstone terranes worldwide, including in Canada (e.g. the deposits of the Raglan, Abitibi and Thompson belts; (Barnes et al., 1982; Lesher, 2007; Houle et al., 2008; LaytonMatthews et al., 2011), Zimbabwe (e.g., Hunters Road; Prendergast, 2001) and Finland (e.g., Vaara; Konnunaho et al., 2013). The Yilgarn Craton in Western Australia contains the most voluminous komatiite event preserved on Earth, and the largest associated Ni-Cu-PGE deposits (Barnes and Fiorentini, 2012). The Kalgoorlie Terrane in the Eastern Goldfields Superterrane (EGST), contains more than half of the global magmatic Ni endowment (Barnes and Fiorentini, 2012). Compared to other komatiite-hosted nickel sulphide terranes, the presence of favourable lithospheric architecture (Begg et al., 2010; Mole et al., 2013; Barnes and Van Kranendonk, 2014;

Mole et al., 2014), large amounts of very magnesium-rich adcumulates, and high crustal contamination levels are considered important criteria for the extremely high Ni endowment of the Kalgoorlie Terrane (Barnes and Fiorentini, 2012).

Although the Kalgoorlie Terrane hosts the majority of known komatiite-hosted nickel sulphide deposits in the EGST, recent discoveries of similar deposits in the adjacent Kurnalpi Terrane have recently broadened the exploration scope into less studied areas. This study examines the Fisher East komatiite-hosted nickel sulphide prospects (Figure 1), located in the Kurnalpi Terrane in the northeast Yilgarn Craton, straddling the boundary between this terrane and the Burtville Terrane (Belbin et al., in prep; Burley et al., 2016)Other known deposits located along this boundary include the AK47, Rosie, Mulga Tank and Mt Windarra deposits (Figure 1). Although the first three occurrences are unmined resources at present, the potential exists for larger deposits along this terrane boundary.

Lithogeochemistry has been widely applied in exploration for komatiite-hosted nickel sulphide systems. This includes tracking high flux magma environments through the use of whole rock major elements and Ni/Cr and Ni/Ti ratios (Barnes et al., 2004), assessment of crustal contamination levels using incompatible trace elements such as La, Sm, Ti, Zr, Nb, Th and Yb (Barnes et al., 2004; Le Vaillant et al., 2014; Le Vaillant et al., 2016) and defining komatiitic source environments by evaluating Al contents, and Al2O3/TiO2 and Gd/Yb values (Nesbitt et al., 1979). To date, hyperspectral reflectance data have not been widely used in komatiite-hosted nickel systems, although Laukamp et al. (2012) distinguished mafic and ultramafic lithologies at the Plutonic deposit in Western Australia, on the basis of their relative amphibole and talc abundances and changes in the Mg# of amphibole, interpreted from shortwave infrared (SWIR) HyLogger3TM data. Murdie et al. (2010) discussed the potential for mapping serpentinised and carbonatised ultramafic rocks in the St Ives area by using a combination of airborne hyperspectral, magnetic and radiometric data. SWIR reflectance spectra have been extensively examined in hydrothermal systems (for example, Herrmann et al., 2001; Sonntag et al., 2012; Harraden et al., 2013), but hyperspectral thermal infrared (TIR) reflectance spectroscopy has only been used for analysis of mineral systems in a few cases, such as orogenic gold (e.g. Cudahy et al., 2009) and VMS-style mineralisation (Duuring et al., 2016).

This contribution describes the geology, volcanology, primary geochemical characteristics of komatiites and Ni-Cu-PGE mineralisation at the previously unstudied Mount Fisher greenstone belt, and assesses the potential for mineralisation in the region. We look at trends in SWIR and TIR data for komatiite host rocks, and relate these mineral trends to corresponding geochemical data collected from wholerock and portable X-ray Fluorescence (pXRF) analysis. The hyperspectral data were examined with the aim of detecting subtle changes in mineral abundance and mineral chemistry, considering that primary textures in komatiites at the Fisher East prospects have mostly been destroyed during greenschist-facies metamorphism, deformation, and talc-carbonate metasomatism. The combined techniques allow interpretation of primary magmatic zonation within komatiite flows and ultimately the development of a flow facies model.

2. Regional Geological Setting

The Eastern Goldfields Superterrane comprises four smaller terranes: Kalgoorlie, Kurnalpi, Burtville and Yamarna (Pawley et al., 2012; Huston et al., 2012). The EGST is separated from the Youanmi Terrane by the Ida Fault (Cassidy et al., 2006), and includes volcanic sequences dated between c. 2940 and 2660 Ma (Czarnota et al., 2010; Huston et al., 2012). The Fisher East prospects are hosted in the Mount Fisher greenstone belt, along the boundary between the Kurnalpi and Burtville Terranes in the north-eastern Yilgarn Craton. Several other Ni deposits have recently been discovered along this same boundary (Figure 1).

Most greenstone belts in the EGST formed between c. 2720 and 2655 Ma (Czarnota et al., 2008), with a general younging trend towards the west (Czarnota et al., 2008; Said et al., 2012). A similar age range of 2710 to 2700 Ma was proposed by Swager (1997) for greenstone belts in the southern area of the EGST. In contrast, greenstone belts in the Burtville Terrane have considerably older ages of c. 2805Ma (Kositcin et al., 2008), 2770 Ma and 2960 Ma (Geological Survey of Western Australia, 2009), demonstrating closer temporal relationships with the spatially more distant greenstone sequences to the west in the Youanmi Terrane. Most greenstone sequences include

calc-alkaline felsic, komatiitic and tholeiitic sequences, as well as sedimentary units (Czarnota et al., 2008; Said et al., 2012).

The Mount Fisher greenstone belt is located in the Laverton Domain of the Kurnalpi Terrane. The stratigraphy of the Laverton Domain is poorly constrained, but includes banded iron formations (BIFs), mafic-ultramafic volcanic sequences, and fine-grained volcanogenic sedimentary rocks (Cassidy et al., 2006). Kositcin et al., (2008) propose that this domain could be older than c. 2720 Ma, likewise Cassidy et al., (2006) suggests an age of c. 2800 Ma. The Duketon greenstone belt (Figure 1), located southeast of Mount Fisher, contains metamorphosed ultramafic, mafic, felsic volcanic and volcanoclastic rocks, in addition to sedimentary packages including sandstone, shale, conglomerates and chert (Wyche et al., 1997). The Dingo Range greenstone belt (Figure 1), south-west of the Mount Fisher greenstone belt, also includes BIFs, komatiites and basalts, that are reportedly younger than c. 2870 Ma (Cassidy et al., 2006). The southern portion of this belt is metamorphosed to upper amphibolite facies (Wyche et al., 1997).

3. Sampling and methods

Eleven drill holes from four different prospects in the Fisher East area; Camelwood, Cannonball, Musket and Sabre (Figure 2) were selected for logging. Ten of these were sampled for geochemical analysis and petrography. Drillholes were chosen from mineralised and unmineralised locations along strike and within the komatiite package. 3.1Whole-rock geochemical analysis Whole-rock geochemistry was performed on unweathered samples from the 10 selected drillholes. Samples were collected down hole to test for likely variations in primary stratigraphy and lithology. Least altered rocks without veins were targeted for analysis. Analyses were completed by Genalysis Intertek Laboratories in Perth, Western Australia for major oxide, trace, and Rare Earth Elements (REEs) using a combination of X-ray Fluoresce (XRF) Spectrometry, Inductively Coupled PlasmaMass Spectrometry (ICP-MS), Inductively Coupled Optical (atomic) Emission

Spectrometry and Infrared Spectrometry. Specific techniques used in preparation for analysis included a combination of 4 acid digest, fire assay and lithium metaborate/tetraborate fusion. Rock samples were cleaned and crushed with a steel jaw crusher, and reduced to powder using a low Cr mill to reduce contamination in komatiitic lithologies. All data obtained and used in interpretations have been corrected for volatile and sulphide content (expect Ni and Cu, which still contains the sulphide component). Results are presented in Appendix 1. 3.2 Portable X-Ray Fluorescence studies Portable X-Ray Fluorescence (pXRF) analyses were performed on drill core using an Olympus-InnovX Handheld XRF analyser, which contains a 4W, 40KV tantalum Xray tube and silicon drift detector (Olympus Corporation, 2016). Clean sample sites were selected every 1m throughout ultramafic intersections, and approximately every 5 m in other rock types. Measurements were taken on surfaces of quarter, half, and full core samples (on flat cut surface where available). ‘Soil’ mode was utilised during this study, as it is designed to measure trace element contents in ppm, and for element concentrations below 1 wt% (Le Vaillant et al., 2014). In this mode, three-beam analysis (50, 35 and 15kv) (Le Vaillant et al., 2014) were set for 20s each. Elements analysed include S, K, Ca, Ti, Cr, Mn, Fe, Cu, Zn, As, Rb, Sr, Zr and Ni.

For calibration purposes, seven known matrix-matched standards were analysed at the beginning and end of each analysis session (standards provided and previously used by Le Vaillant et al. (2014)). These standards, which comprise pulps previously analysed by whole-rock lithogeochemistry (XRF and ICP-MS), were measured through thin plastic bags. The attenuation on the concentration of various elements caused by the presence of the plastic bag was evaluated, and corrections applied for each element before calculating calibration curves. In order to guarantee the accuracy of the data, calibration curves were only applied after all analyses had been collected. To further ensure the quality of the data and monitor possible drifts of the instrument over time, a monitor (one of the standards) was analysed in between 15 unknown measurements, along with instrument calibration.

Data correction, sample precision and instrument precision was undertaken, following procedures outlined in Le Vaillant et al. (2014). Sample precision was evaluated for each lithology encountered in the drillholes (to account for different matrix effects and grainsize variations), while instrument precision was monitored regularly. Supplementary material for pXRF analysis is presented in Appendix 2. 3.3 SWIR and TIR data Ten drillholes were scanned using a HyLogger3 TM at the Geological Survey of Western Australia in Perth. The HyLogger3 TM enables the simultaneous collection of hyperspectral reflectance spectra in the visible near infrared (VNIR; 350 to 1000nm), SWIR (1000 to 2500nm) and TIR (6000 to 14500nm) wavelength ranges (Hancock and Huntington, 2010; Hancock et al., 2013; Schodlok et al., 2016). The three wavelength ranges collected using the HyLogger3TM, allow rapid mapping of all major rock forming minerals, based by means of identifying particular absorption features specific to certain ions or molecular bonds (Hancock and Huntington, 2010). The VNIR can be used to characterise, for example, iron oxides and REE, whereas the SWIR enables the characterisation of hydroxyl-bearing minerals including amphiboles, talc and chlorite. The TIR is used to characterise nominally anhydrous minerals (e.g. quartz, feldspar, pyroxene) and phosphates. Data were processed using The Spectral Geologist (TSGTM) software (www.thespectralgeologist.com).

Three common, but different ways of processing mineral assemblages from the hyperspectral drill core data in TSGTM were chosen for this study: firstly, The Spectral Assistant (TSA) within TSGTM was used to map changes in the mineral assemblages down hole. TSA uses an unmixing algorithm based on a spectral library for modelling the collected spectra (Berman et al., 1999). Secondly, a multiple feature extraction method (Laukamp et al., 2010) was used to identify minerals based on their diagnostic absorption features which are related to the physicochemistry of the respective mineral. For this paper, the main focus was on reliably identifying the relative abundance of 1) talc, based on the hydroxyl-related combination band located at around 2080nm (Laukamp et al., 2012), and 2) chlorite, based on the hydroxyl-related combination band located at around 2250nm (Doublier et al., 2012). Thirdly, a Partial Least Squares method (PLS; Haaland & Thomas, 1988a, b) was chosen to model the

geochemical indices (i.e. Ni/Cr and Ni/Ti ratios) from the SWIR hyperspectral signatures using the PLS-module built in TSGTM. For this, the Ni/Ti and Ni/Cr ratios derived from whole rock geochemistry were imported into TSGTM and calibration samples selected from value ranges between zero and 16 and zero and 4, respectively. After selection of the calibration samples, the following calibration stages were undertaken: 1) specification of input data (i.e. whole SWIR wavelength range, hull removed) and 2) the cross-validation, which is based on a leave-one-out method and comprises the evaluation of the predicted residual error sum of squares (PRESS) and generation of the final regression coefficients (FRC) for each input (i.e. spectral band). During the subsequent prediction stage, an FRC-based scalar was created and applied to the whole spectral data set. The values predicted from the PLS modelling scalar were then compared with the original reported data. For log Ni/Ti and log Ni/Cr a correlation of r=0.755 (using 25 input bands) and r=0.59 (using 26 input bands) between the measured and predicted values was achieved, respectively.

4. Results

4.1 Geology, stratigraphy and mineralisation of the Fisher East prospects

The overturned Fisher East stratigraphy (Figure 3, here, shown stratigraphically right way up), is comprised of felsic metasedimentary rocks, BIFs, shales, sulphidic cherts, quartz porphyries, basalts and komatiites, most of which have undergone deformation, and have been cross-cut by multiple intrusions including mafic dykes, now amphibolites. Komatiites at the Fisher East prospects have been dated between c. 2940 and 2840Ma (Mole et al., 2016), and are pervasively altered to talc-carbonate assemblages. Primary spinifex (A zone) and cumulus (B zone) textures are destroyed in komatiites, but despite this, two predominant secondary mineral assemblages are observed; talc-rich and chlorite-rich. Talc-rich komatiites contain carbonate porphyroblasts (Figure 4) varying in size and colour, with talc and chlorite abundance varying not only in different drillholes, but throughout the same hole. Chlorite-rich komatiites are dominated by chlorite, with local presence of cross cutting carbonatequartz veins (Figure 5). Petrography of talc-rich samples shows large carbonate porphyroblasts with finer grained talc comprising the matrix of the sample (Figure 4b). In places, carbonate looks to be completely recrystallised, however in a thin

section sample from MFED043, carbonate looks to be replacing primary igneous minerals (Figure 6). Carbonate grains often contain planar arrays of sulphide and/or magnetite inclusions. Chlorite is commonly found wrapping around the carbonate porphyroblasts.

A strong pervasive foliation is developed in both talc-rich and

chlorite-rich lithologies.

Basal contacts to komatiite intervals (in the original stratigraphic sense, Figure 3) at the Fisher East prospects are mainly defined by talc-rich komatiites and underlying felsic metasedimentary rocks. The nature of these basal contacts varies throughout the prospects. In drillhole MFED069 (Cannonball prospect, see Figures 2 and 3) the basal contact shows mingling between the two rock units, possibly due to injection of komatiites into the underlying sedimentary rocks (Figure 7). Massive sulphides, bleaching of footwall sediment and rounded quartz clasts are also observed in komatiites along the basal contact. In addition, massive sulphides within komatiite units contain clasts of metasedimentary rocks (Figure 8). These features are the result of primary magmatic processes, evidenced by irregular contacts and an absence of shear zones, fault zones or fracturing. In contrast to basal contacts, upper contacts between komatiites and other lithologies are abrupt, and show little evidence of interaction or alteration. Consequently these features are interpreted to be concordant primary stratigraphic/sedimentary contacts or faulted equivalents. An example of a small alteration zone at one of these upper contacts is shown in Figure 9.

Mineralisation in the Fisher East prospects shows characteristics of type 1 magmatic nickel sulphide deposits (Lesher and Keays, 2002), occurring along the basal contact between originally olivine-rich komatiites, represented by the talc-rich lithology, and felsic metasedimentary rocks. Three of the investigated prospects contain massive sulphides at the basal contact (Figure 10b-c), while the fourth prospect, Cannonball, displays sulphide veins (Figure 10a). Massive and vein sulphides include pentlandite, violarite, pyrite ± pyrrhotite ± chalcopyrite. Evidence is lacking for extensive tectonic mobilisation of sulphide off the primary contact.

4.2 Komatiite geochemistry

Whole-rock and pXRF geochemical data derived from the Fisher East prospects are interpreted to determine flow sequences within the komatiite intervals. Analysis of komatiites displaying the two styles of secondary mineralogy and textures at Fisher East has revealed that both types have distinct geochemical signatures. Talc-rich komatiites have higher MgO and Ni contents than chlorite-rich komatiites; the latter are more rich in Al2O3, TiO2 and Zr. These geochemical trends can all be explained from secondary mineralogy. Talc is only found in large amounts when the precursor rock was rich in forsteritic olivine cumulates (Barnes, 2006), explaining the high MgO and Ni content in talc-rich komatiites at Fisher East. Relationships between mineralogy and geochemistry of komatiite flows are consistent across all 10 drillholes sampled in this study, with an example of this demonstrated in Figure 11 from drillhole MFED029 (Camelwood prospect, see Figures 2 and 3). Drill logs, mineralogy and geochemical summaries for all drillholes are presented in Appendix 3. A summary of bulk whole-rock geochemical differences between the two styles of altered komatiites at the Fisher East prospects is presented in Table 1.

Al2O3/TiO2, MgO/Al2O3 MgO/SiO2, Ni/Cr vs. Ni/Ti and incompatible trace element ratios have all been assessed to understand petrogenetic processes occurring in the Fisher East area. The Al2O3/TiO2 ratios are used to distinguish Al-undepleted (Munrotype) from Al-depleted (Barberton-type) type komatiites (Arndt et al., 1977; Nesbitt et al., 1979). These elements are typically immobile during alteration (Hill et al., 1988; Barnes et al., 2004). Munro komatiites share a similar Al2O3/TiO2 ratio to the chondritic mantle, having values of ~20, while Barberton komatiites have Al2O3/TiO2 ratios of ~11 (Nesbitt et al., 1979). Al2O3/TiO2 ratios for komatiites at the Fisher East prospects range from ~10-37, with an average value of 25. Ratios plot just below the chondritic mantle line, with limited scatter (Figure 12). Kalgoorlie Terrane komatiites are typically Al-undepleted (Barnes, 2006). Plots of MgO vs Al2O3 and MgO vs SiO2 (Figure 13), show very weakly defined linear trends, in contrast to the strong negative correlations found in most komatiite suites, implying that there has been considerable scatter of MgO and probably SiO2 during alteration.

Incompatible trace elements, such as La, Sm, Ti, Zr, Nb, Th and Yb can be used to assess crustal contamination (e.g., assimilation with sediment, granite, intermediatesilicic volcanic rocks) of komatiites upon emplacement into the crust (Barnes et al.,

2004; Barnes, 2006). The Fisher East komatiites show deviation towards bulk continental crust in all incompatible trace element ratios presented (Figure 14). Measured Th values are mostly below the analytical detection limit for whole-rock geochemical analysis, particularly in olivine-rich cumulates (characteristic of BZones), and consequently Th/Nb and Th/Yb plots have limited use. Plots of La/Sm ratio show that approximately half of the samples exhibit signs of crustal contamination, while the remaining demonstrates sub-chrondritic La/Sm values. This latter feature is common in uncontaminated Munro-type komatiites and may be attributed to the derivation of komatiites from previously depleted mantle sources (Barnes, 2006).

Plots of Ni/Cr vs. Ni/Ti for the Fisher East prospects are shown in Figure 15. This plot was originally proposed to discriminate between different volcanic facies in a komatiite system (Barnes et al., 2004), and is of use as a discriminant where intense talc-carbonate alteration and/or weathering have disrupted the major element chemical composition of the rock, and also destroyed primary textures (Le Vaillant et al., 2016). CSF (channelised sheet flows) and DC (dunitic conduit) fields represent high-flow magma pathways with high olivine abundance (Barnes et al., 2004). Ponded lavas represented by LLLS (layered lava lakes and sills), characterised by more chromite and hence lower Ni/Cr, have low Ni sulphide potential, as does the TDF (thin differentiated flows) facies marked by multiple thin (few meters) flow lobes with high proportions of spinifex textured A-zones to cumulate textured Bzones (Barnes et al., 2004). The Fisher East data array corresponds closely to the field for Kambalda-style channelised sheet flows (CSF). When comparing whole-rock (black squares) and pXRF geochemistry (red circles) in Figure 15, both provide a similar result and there is a good agreement between the two analytical techniques. Ni/Cr vs. Ni/Ti ratio plots have also been contoured to discriminate rock type (Le Vaillant et al., 2016) where MgO is unreliable, as is the case here. Figure 15 indicates a range of lithologies at Fisher East, ranging from typical liquid-rich flow top (spinifex) compositions to transitional meso-adcumulates, with a predominance of ortho-mesocumulates. Two samples plot only in the sulphide-bearing cumulate field, reflecting a general lack of disseminated sulphides more than a few m away from the mineralised zones. This plot also shows a lack of komatiitic basalt compositions, chromite-enriched cumulates or true dunites at the Fisher East prospects.

4.3 Hyperspectral data Hyperspectral techniques (HyLogger3TM; Schodlok et al., 2016) were employed to help differentiate komatiite units and flows based on mineral abundance and mineral chemistry. The most dominant trends in HyLogger3 TM data obtained from the Fisher East prospects were observed for relative talc and chlorite contents using SWIR and TIR detectors (e.g., Figure 11). Examples of typical reflectance spectra for chloriterich and talc-rich komatiite zones are shown in Figure 16. Talc-rich komatiites positioned lower in the stratigraphic sequence displays less chlorite than similar zones located higher in the sequence. The detected carbonate mineral occurrence varies throughout scanned drill core but is more common in talc-rich komatiites. Chlorite was confirmed as the dominant constituent of chlorite-rich komatiites, with minor amounts of amphibole interpreted from TIR spectra.

Modelling of hyperspectral data shows a direct link between spectral and geochemical data. Figure 17 shows plots for the 2250nm (Doublier et al., 2012) vs. 2080nm (Laukamp et al., 2012) spectral indices, with colours representing different proportions of Al2O3 (wt%), TiO2 (wt%) and Mg# in samples, derived from wholerock geochemical data. In the investigated data set, the 2250nm SWIR absorption feature is diagnostic for chlorite, while the 2080nm SWIR absorption feature is a key indicator for the presence of talc. It should be noted that epidote and biotite can be present in the investigated data set, which both are also characterised by a major absorption feature at around 2250nm. However, based on field observations their presence is only minor, so that the 2250nm feature can be used as an indicator for the presence of chlorite. Figure 17 shows that in rocks where Al2O3 and TiO2 content are highest (samples in red) chlorite is the dominant mineral in that particular sample ( increasing along the 2250nm axis). The opposite relationships are demonstrated in the Mg# plot, where the highest Mg# occurs in talc-rich samples (i.e. samples with increasing values along the 2080nm axis). Additional PLS-based modelling of geochemical indices from hyperspectral data in TSGTM shows the density of talc-rich vs. chlorite-rich samples in Ni/Cr vs. Ni/Ti plots (Figure 18). A majority of chloriterich samples plot in the spinifex/orthocumulate field, while the talc-rich samples fall in the mesocumulate compositional field. This opens up the potential for classifying

komatiites on a much larger sample data base provided by the hyperspectral data when compared to the whole-rock geochemical data.

By using the PLS modelling method, hyperspectral data can be used to predict geochemical values such as Ni/Cr and Ni/Ti (explained above). Figures 19 and 20 compare downhole Ni/Ti and Ni/Cr whole rock and pXRF geochemistry, with PLS modelled predicted values, as well as changing abundance of talc and chlorite in drillhole MFED029. Different coloured points in this diagram represent different lithologies. These plots show that pXRF and PLS modelled Ni/Ti and Ni/Cr data are remarkably similar. Whole rock geochemistry shows the same trend when compared to PLS modelled values, but is harder to depict due to a lower sample density. Higher Ni/Ti values, and to a lesser extent, Ni/Cr, have a direct correlation with talc abundance. Chlorite abundance in higher Ni/Ti and Ni/Cr zones tends to be low. As there are a large amount of hyperspectral data points compared to geochemical data, PLS modelling can indicate how Ni/Ti and Ni/Cr could potentially vary in different units. Talc-rich komatiite zones show limited spread compared to other lithologies, particularly sediments.

These downhole figures (Figures 19 and 20) also show more clearly how mineralogical data obtained from the HyLogger3TM can be useful in these systems. For example, in drillhole MFED029, drillcore logging suggests that there is a talc poor, siliceous patch of komatiites at ~410m, and a talc increase at ~415m. This can be seen in both talc and chlorite abundance plots, as well as being marked by a decrease in the case of ~410m, and increase at ~415m, in PLS modelled Ni/Cr and Ni/Ti data. These trends are not clear in whole rock and pXRF data.

All of these methods, especially the PLS method for predicting geochemical values from hyperspectral data (Figure 19 and 20), demonstrate the potential for classifying komatiites on a much larger sample data base when compared to the whole-rock and pXRF geochemical data.

5. Discussion

5.1 Analytical techniques

5.1.1 Hyperspectral data

The SWIR and TIR mineralogical data for the Fisher East prospects accurately discriminate subtle changes in talc and chlorite abundances in drillcore, relating to changes in komatiite texture and mineralogy, which ultimately reflects different komatiite flow units. Differentiating different zones within komatiite flow is vital in the exploration of komatiite- hosted Ni-Cu-PGE deposits.

While summaries of mineralogical data (Figure 11, Appendix 3) are useful in depicting the overall different komatiite flow units (based on talc and chlorite abundances) and what minerals occur in different lithologies, downhole plots of changing mineral abundance (Figures 19 and 20) provide a clearer picture of changes within separate flow units. For example, in MFED029, there is a change in komatiite mineralogy (identified by drillcore logging) at ~410m and ~415m. This is displayed nicely in downhole 2080nm plots, and to a lesser extent, in 2250nm plots. In highly altered talc-carbonate rocks such as the ones at Fisher East, these plots could potentially be a very useful tool to supplement drillcore logging, and depict subtle and potentially important changes within individual komatiite flows. The changes in mineralogy stated here can be related back to modelled PLS Ni/Cr and Ni/Ti values. While mineralogical data from the HyLogger3TM is valuable in this case study, certain aspects of this technique must be considered. The investigated data set consists of mineral assemblages that contain many different SWIR-active minerals, which precludes an interpretation solely based on spectral similarity measures (for example by comparison with spectral libraries or data set endmembers). For example, amphiboles and talc can show almost identical SWIR spectral signatures and the identification of minor absorptions such as the talc-related feature at around 2080nm is required to successfully differentiate between talc and amphibole (Laukamp et al., 2012). This could potentially be a problem in talc-carbonate altered komatiite systems. It is therefore advised to manually check some interpretations using particular scalars. In this study, amphiboles, biotite, epidote and zoisite were verified. The HyLogger3TM may also suggest certain minerals that are questionable in specific

circumstances. At the Fisher East prospects, this is the case with white mica and feldspars in komatiite zones. White mica and feldspar are both found in other lithologies at Fisher East, such as in footwall metasedimentary rocks. However, the HyLogger3 TM also depicts white mica in talc-rich komatiites, and a combination of white mica and feldspar in chlorite-rich komatiite samples. While only minor, the presence of those minerals is still questionable within ultramafic rocks, which although they have been highly altered, should not contain enough Na or K to form these minerals. However, the white mica related spectral signatures in some of komatiites may point to narrow intervals of felsic rocks or sediments that were overlooked during visual drill core logging. The sample area of the HyLogger3 TM should also be taken into consideration when assessing data. The sample area of the HyLogger3 TM is only 10mm around the centre of drillcore (Hancock et al., 2013), which could possibly lead to a misrepresentation of the mineral assemblage in coarse grained, inhomogenous sections. Various modelling in TSGTM can directly link hyperspectral data to geochemistry, which is discussed in section 5.1.3 below.

5.1.2 Comparison of pXRF and whole-rock geochemical data

In most cases pXRF and whole-rock geochemical data are comparable, but some discrepancies were observed, notably between Zr, Ni and Cr values. Inconsistencies in Zr values could be the result of poor pXRF instrument precision, or heterogeneous distribution of Zr-bearing phases. The cause of variation between Ni data, after examination of instrument precision and calibration curves for pXRF, is unknown (Appendix 2). When discrepancies occur in Ni/Cr data, the corresponding Ni comparison between pXRF and whole-rock geochemistry are acceptable, and therefore the problem is likely due to Cr data. pXRF instrument precision for Cr is respectable at 2%, and therefore sample imprecision due to a nuggety distribution of chromite may be the cause. While minor discrepancies occur throughout the dataset, they are not significant enough to rule out the use of pXRF in these systems as a rapid exploration tool to vector towards high-flux magma pathways.

5.1.3 Comparison of hyperspectral and geochemical data

Although measured at different scales, the mineral abundances interpreted from hyperspectral data can be broadly correlated to the corresponding geochemical data for the same section of core (Figure 11, Figure 19, Figure 20, Appendix 3). The highest MgO% values derived from whole rock geochemistry correlate with higher abundances of talc as documented in HyLogger3TM data. In lithologies where little or no talc is recognised by the HyLogger3TM, MgO% content decreases significantly, typically to less than 10 wt%. In komatiites, Al2O3 also shows a trend with chlorite, increasing in wt% in chlorite rich sections. Al2O3 is at its highest when plagioclase, chlorite, white mica, and sometimes amphiboles are present in hyperspectral data. Hyperspectral data also allows us to predict and model different element concentrations through entire drillhole sequences more rapidly than other methods currently used. Figure 17 shows how different element concentrations can be directly related to certain minerals, using the 2250nm and 2080nm absorption features. In this case, Al and Ti contents are related to higher chlorite sequences and therefore depict A-zone komatiites, and Mg# to talc, or B-zone komatiites. The biggest advantage of combining geochemical and hyperspectral data together is being able to quickly and accurately model different elements in drillcore where data may be lacking. PLS modelling, demonstrated in Figures 19 and 20 show great similarities to geochemical data, particularly pXRF data. Due to the large sample density of hyperspectral data, we can predict possible variations of Ni/Cr and Ni/Ti within the same lithological unit, which are otherwise overlooked by conventional intervals of geochemical sampling. B zone komatiites show limited spread in these two element ratios compared to other lithologies. This is most likely due to the fact that the other lithologies contain small amounts of these elements, so any difference in concentration, even if minor, is likely to have an effect and hence the large spread. As Ti is relatively immobile, it is most likely a change in Ni or Cr concentration that produces such large spread. Changes in modelled Ni/Cr and Ni/Ti ratios can be directly related to mineral abundance (Figure 19 and 20). As shown, where Ni/Ti and to a lesser extent, Ni/Cr increase, talc is more abundant, delineating talc-rich B-zone komatiites. These changes can also potentially depict differences within individual

flow units where whole rock and pXRF geochemical data does not, as shown at ~410m and ~415m in MFED029 (Figures 19 and 20).

While pXRF data in itself is relatively quick to gather, it can be time consuming and requires specific calibration methods and conditions throughout data collection to insure accurate results. The collection of hyperspectral data can also vary on certain conditions. However, hyperspectral data collected through the HyLogger3TM are collected in controlled conditions, therefore minimising errors. PLS modelling of particular geochemical indices such as Ni/Cr and Ni/Ti is a new technique that produces a larger dataset to work with, and can fill in the geochemical gaps when lacking, rapidly and accurately. Combining all hyperspectral plots with whole rock and pXRF data provides a complete and robust set of data to help differentiate A and B zones in komatiites flow, and ultimately aid in the exploration of komatiite-hosted Ni-Cu-PGE in highly altered talc-carbonate assemblages.

5.2 Fisher East mineralisation

5.2.1 Volcanic architecture using komatiite zonation and geochemistry

Komatiite flow units are commonly differentiated into A and B zones (summarised in Table 2; Barnes, 2006), where A-zones refer to upper zones in a komatiite flow, typically containing spinifex textures, while B-zones refer to cumulate rich lower flow zones (Hill et al., 1988; Hill, 2001). The A and B zones can be distinguished based on incompatible element contents, such as Al, Ti and Zr. Cumulates (B-zone), with a low proportion of trapped liquid, have low incompatible elements contents, increasing slightly from adcumulate to mesocumulate to orthocumulate. A-zones represent komatiite flow tops, and hence contain quenched komatiite as well as variably fractionated spinifex zones; these have much higher incompatible element contents. Olivine spinifex-textured rocks from other localities typically contain between 20 and 32 wt% MgO (Barnes, 2006). Based on the data presented above, it is clear that the talc-rich lithologies were originally B-zone, or cumulate textured komatiites, while chlorite-rich komatiites were most likely originally A-zone or spinifex textured.

The difference in chemistry and mineralogy between the two zones reflects different proportions of original cumulus olivine before talc-carbonate alteration. Lower MgO komatiite zones at Fisher East, interpreted as A-zones, typically contain on average 29% MgO, which is similar to A-zones in other areas. These values, however, may have been modified during alteration. Specific cumulate textures in B zones at Fisher East are not recognisable texturally owing to pervasive alteration; however the presence of meso and orthocumulates is indicated by the whole-rock MgO, Al and Ti contents. This is also seen in Ni/Cr vs. Ni/Ti ratios (Figure 15), whereby A-zone samples that plot in the ortho- and mesocumulate fields contain higher Mg, and lower Al2O3 and TiO2 than other A-zone samples that plot in the spinifex field. While these samples have the mineralogy and texture often seen in other A-zone samples, there geochemistry suggest that they should be classified as B-zone komatiites. Olivine adcumulates, characterised by very low Al and Ti contents and MgO contents >50 wt% were not observed at Fisher East.

Based on stratigraphy and geochemical differentiations of A and B zones, an interpretive long section (Figure 21) has been created for the Fisher East prospects. The B zone komatiites are commonly much thicker compared to A zones. This is important as the presence of a high proportion of B zone within a flow profile can represent high-flux magma pathways (Hill, 2001). High-flux komatiitic pathways can be important for mineralisation, as they imply low rates of heat loss compared to flow rate, facilitating thermomechanical erosion of the underlying rocks (Lesher et al., 1984; Lesher, 1989; Hill, 2001; Barnes, 2006). Basal contacts between komatiites and felsic metasedimentary rocks across the section show evidence of thermomechanical erosion, for example, rip-up clasts of substrate material (Figure 8) and evidence for injection of sulphide liquid into the footwall.

Thick B-zones, asymmetric flow features, evidence of multiple flows and the lack of olivine adcumulates implies a similarity in facies architecture to the type example of channelised sheets flows (CSF) at Kambalda (Lesher et al., 1984). This particular volcanic facies, along with dunitic conduits, is known to be a favourable setting to host nickel sulphide mineralisation.

5.2.2 Primary melt characteristics

The differences in geochemistry between the main geochemical types of komatiite are a reflection of mantle melting conditions. Al-depletion in Barberton-type komatiites is due to retention of garnet in the mantle source during high pressure partial melting (Nesbitt et al., 1979). The Al-undepleted Munro-type komatiites, which have Al2O3/TiO2 ratios similar to that of the chondritic mantle, only leave behind olivine when the melt is separating from its source. Recent studies by Sossi et al., (2016) however suggest that both komatiite types can come from the same mantle plume, but form in different positions in the plume under different conditions. It is stated that at a given pressure, Al-undepleted komatiites will form over Al-depleted komatiites at higher rates of partial melting (Sossi et al., 2016). Al-undepleted komatiites form along the main plume axis, while Al-depleted komatiites are formed at outer edges of the plume at cooler temperatures (Sossi et al., 2016).

Compared with the typical Munro-type komatiites of the Kalgoorlie Terrane (Barnes, 2006), the Fisher East komatiites fall slightly below the chondritic mantle line (Figure 12) and have slightly higher Al2O3/TiO2 ratios than typical Munro-type. This difference could potentially be due to crustal contamination, post-depositional alteration, or a slight difference in komatiite source composition. The Fisher East data points show a high degree of Al-Ti correlation with little scatter, making it more likely this trend is inherited from source melting. Unusually high Al/Ti is observed in a rare type of Al-enriched komatiite found in the Commondale greenstone belt in South Africa (Wilson, 2003) and is attributed to derivation from mantle that had undergone very strong depletion due to previous melting; a similar explanation may apply to the Fisher East komatiites, although the degree of Al enrichment is far less: mantle normalised Al/Ti of 1.1 to 1.4 as opposed to around 3 at Commondale. Nonetheless, the Fisher East komatiites are somewhat anomalous in this respect relative to typical komatiites in the Eastern Goldfields Superterrane, and this characteristic may enable correlative units to be identified elsewhere. An attempt was made to assess primary melt compositions at Fisher East using FeO-MgO relationships, but deemed unreliable as the system has almost certainly undergone metasomatic change in Fe and Mg.

5.2.3 Mineralisation and contamination

Type 1 and 2 komatiite-hosted nickel sulphide deposits (Lesher and Keays, 2002) are the most common types found in the Yilgarn Craton (Barnes, 2006). Mineralisation at Fisher East shows type 1 mineralisation characteristics, with nickel sulphides occurring at the basal contact between felsic metasedimentary rocks and B-zone (talcrich) komatiites. Sulphides are typically massive, but are sometimes observed as veins.

The stratigraphic long-section presented for Fisher East (Figure 21) shows that mineralised prospects are associated with greater thicknesses of B-zones relative to Azones compared with the non-mineralised areas. This is interpreted as the signal of ore formation beneath high-flux flow channels, as in the accepted model for the Kambalda area (Lesher and Keays, 2002), and implies that there has been relatively little tectonic mobilisation of the sulphides from their original position.

As shown in incompatible trace element data, crustal contamination is evident in Fisher East komatiites. Along with geochemical evidence for thermo-mechanical erosion/crustal contamination, physical evidence revealed by detailed core logging includes sulphides invading the footwall, and rip-up clasts of felsic metasedimentary rocks in massive sulphide zones.

Mineralisation at the Fisher East prospects mostly likely occurred via thermal erosion of underlying sedimentary rocks. The likely source of S is from sulphide bearing BIF, fragments of which are preserved in drillcore adjacent to identified channels. Textures indicating potential BIF presence before komatiite emplacement are demonstrated in Figure 22. This example may represent a peperitic section of BIF due to interaction between BIF and komatiite on a channel flank, or due to invasive flow action. In the northern and southern-most holes in this study, MFED016 and MFED071, basal komatiites look as though they were previously in contact with BIFs, suggesting the mineralised area represents a ‘BIF gap’, as observed at Mt Windarra (Marston, 1984), indicating a significant amount of thermal erosion has occurred. Rox Resources (2015) state that BIF units rich in sulphides proximal to ultramafic flows (as seen in MFED071) tend to represent the edge of a flow. BIFs are absent in mineralised holes, suggesting that complete assimilation of these units occurred in these areas, and that

the mutually exclusive relationship between basal contact ores and substrate BIF is analogous to that with contact sediment at Kambalda (Lesher et al., 1984; Lesher, 1989). Ore zones are broadly unaffected by major structural mobilisation, with basal sulphides texturally associated with footwall and hangingwall lithologies. However, the degree of shearing and deformation observed in the rocks indicate that unseen structural remobilisation of some parts of the orebody cannot be ruled out.

5.2.4 Prospectivity of Fisher East

The Fisher East region has close affinities with the Type 1 style mineralisation seen at Kambalda, and particularly at Mt Windarra, with which it shares the characteristic association with sulphidic BIF. It can be concluded that the Fisher East prospects have the potential to host larger-scale komatiite-hosted nickel sulphides; evidence for which includes crustal contamination (physical and geochemical), occurrence of a Srich sedimentary substrate (S-rich BIF), the presence of a high-flux magma environment, and large amounts of cumulus material indicating channelization and high magma volumes. Barnes and Fiorentini (2012) document the same characteristics in highly-endowed komatiites in numerous localities of the Kalgoorlie Terrane. However, the key difference between the two areas is that the Kalgoorlie Terrane, particularly in its northern part, appears to have a large proportion of magnesium-rich olivine adcumulates, such as those in the Agnew-Wiluna belt; a feature not currently observed at Fisher East. However, this does not discount the potential of this region to form a world-class nickel camp via the definition of multiple type-1 deposits, similar to that found in Kambalda. The potential for these camps along the Kurnalpi-Burtville Terrane margin, highlighted in the location of other small discoveries, is presently untested.

6. Conclusions

This study shows that the use of hyperspectral and pXRF data in talc-carbonate altered komatiites can be useful in confirming mineralogy and assessing differences between A and B-zones when primary textures such as spinifex or cumulates are not preserved. Combining all hyperspectral data outputs with pXRF data provides a complete and very robust dataset to help differentiate A and B zones within a

komatiite flow, and provides much more geochemical data, more quickly and for a fraction of the cost of high-precision lithogeochemistry, which can now be restricted to specific intersections of interest and calibration of other methods. The reconstruction of volcanological flow fields and identification of channelised flows with thick olivine cumulate sections is vital in defining zones that are prospective for Ni-Cu-PGE mineralisation. This study also recognises that the Fisher East prospects have the key features required to form a large komatiite-hosted nickel sulphide deposit. Large camp equivalents to those at Kambalda and the Agnew-Wiluna belt are currently absent in the Kurnalpi Terrane; however, this may be due to low levels of exploration activity in the eastern terranes of the EGST compared to the Kalgoorlie Terrane. A significant factor may be the absence of magnetic anomalies over the komatiites related to the talc-carbonate alteration, which can be readily detected by using hyperspectral reflectance data collected from drill core and chips. Further exploration and follow up studies are warranted to unlock the full potential of this terrane. Detailed isotopic mapping to pin down the location of the Burtville-Kurnalpi terrane boundary would be of great value.

7. Acknowledgments

Part of this research was conducted as part of a Master of Ore Deposit Geology degree through the University of Western Australia. This work was financially supported by the Geological Survey of Western Australia. Lauren Burley publishes with permission of the Director of the Geological Survey of Western Australia. The authors wish to thank Rox Resources Limited for its financial and logistical support and permission to publish this paper. We thank Paul Duuring for comments and consultation during the draft stages of the manuscript.

References

Arndt, NT, Naldrett, AJ and Pyke, DR 1977, Komatiitic and iron-rich tholeiitic lavas of Munro Township, northeast Ontario: Journal of Petrology, v. 18, p. 319–369. Barnes, S-J, Coats, CJA and Naldrett, AJ 1982, Petrogenesis of a Proterozoic nickelsulfide-komatiite association ; the Katiniq Sill, Ungava, Quebec: Economic Geology, v. 77, p. 413–429. Barnes, SJ 2006, Nickel deposits of the Yilgarn Craton: geology, geochemistry,and geophysics applied to exploration (1 edition): Society of Economic Geologists, 210p. Barnes, SJ and Fiorentini, ML 2012, Komatiite magmas and nickel sulphide deposits: a comparision of variably endowed Archean terranes: Economic Geology, v. 107, p. 755–780. Barnes, SJ, Hill, RET, Perring, CS and Dowling, SE 2004, Lithogeochemical exploration of komatiite-assocaited Ni-sulphide deposits: strategies and limitations: Mineralogy and Petrology, v. 82, p. 259–293. Barnes, SJ and Van Kranendonk, MJ 2014, Archean andesites in the east Yilgarn craton, Australia: Products of plume-crust interaction?: Lithosphere, v. 6, p. 80–92. Begg, GC, Hronsky, JAM, Arndt, NT, Griffin, WL, O'Reilly, SY and Hayward, N 2010, Lithospheric, cratonic, and geodynamic setting of Ni-Cu-PGE sulfide deposits: Economic Geology, v. 105, p. 1057–1070. Berman, M, Bischof, L and Huntington, J 1999, Algorithms and software for the automated identification of minerals using field spectra or hyperspectral imagery, in Proceedings of the 13th International Conference on Applied Geologic Remote Sensing: Vancouver, British Columbia, Canada, 1 March 1999, p. 222–232. Burley, LL, Barnes, SJ, Fiorentini, ML, Mole, DR and Belbin, W 2016, The geology of the Fisher East komatiite-hosted nickel sulfide prospects, in GSWA 2016 extended abstracts promoting the prospectivity of Western Australia: Geological Survey of Western Australia, Record 2016/2, p. 8–11.

Belbin, W, Mulholland, I, Burley, L, Mole, D and Barnes, S in prep, The Fisher East nickel sulphide deposits. Cassidy, KF, Champion, DC, Krapez, B, Barley, ME, Brown, SJA, Blewett, RS, Groenewald, PB and Tyler, IM 2006, A revised geological framework for the Yilgarn Craton, Western Australia: Geological Survey of Western Australia, Record 2006/8, 8p. Cudahy, T, Hewson, R, Caccetta, M, Roache, A, Whitbourn, L, Connor, D, Mason, P, Yang, K, Huntington, J and Quigley, M 2009, Drill core logging of plagioclase feldspar composition and other minerals associated with Archean gold mineralization at Kambalda, Western Australia, using a bidirectional thermal infrared reflectance system: Reviews in Economic Geology, v. 16, p. 223–235. Czarnota, K, Blewett, RS and Goscombe, B 2008, Structural and metamorphic controls on gold through time and space in the central Eastern Goldfields Superterrane - a field guide: Geological Survey of Western Australia, Report 2008/9, 66p. Czarnota, K, Champion, DC, Goscombe, B, Blewett, RS, Cassidy, KF, Henson, PA and Groenewald, PB 2010, Geodynamics of the eastern Yilgarn Craton: Precambrian Research, v. 183, p. 175–202. Doublier, MP, Roache, T, Potel, S and Laukamp, C 2012, Short wavelength infrared spectroscopy of chlorite for determination of low metamorphic grades: European Journal of Mineralogy, v. 24, p. 891–902. Duuring, P, Hassan, L, Zelic, M and Gessner, K 2016, Geochemical and spectral footprint of metamorphosed and deformed VMS-style mineralization in the Quinns district, Yilgarn Craton, Western Australia: Economic Geology, v. 111 (1411), p. 1438. Geological Survey of Western Australia 2009, Compilation of Geochronology Data, 2009 Update: Geological Survey of Western Australia. Geological Survey of Western Australia 2015, 1:500 000 tectonic units of Western Australia: Geological Survey of Western Australia.

Haaland, DM and Thomas, EV 1988a, Partial least-squares method for spectral analyses. 1. Relation to other quantitative calibration methods and the extraction of qualitative information: Analytical Chemistry, v. 60, p. 1193–1202. Haaland, DM and Thomas, EV 1988b, Partial least-squares method for spectral analyses. 2. Applications to simulated and glass spectral data: Analytical Chemistry, v. 60, p. 1202–1208. Hancock, EA, Green, AA, Huntington, JF, Schodlok, MC and Whitbourn, LB 2013, Hylogger-3: implications of adding thermal-infrared sensing: Geological Survey of Western Australia, Record 2013/3, 24p. Hancock, EA and Huntington, JF 2010, The GSWA NVCL HyLogger: Rapid mineralogical analysis for characterizing mineral and petroleum core: Geological Survey of Western Australia, Record 2010/17, 21p. Harraden, CL, Mcnulty, BA, Gergory, MJ and Lang, JR 2013, Shortwave Infrared spectral analysis of hydrothermal alteration associated with the Pebble porphyry copper-gold-molybdenm, deposits, Ilimna, Alaska: Economic Geology, v. 108, p. 483–494. Herrmann, W, Blake, M, Doyle, M, Huston, D, Kamprad, J, Merry, N and Pontual, S 2001, Short wavelength infrared (SWIR) spectral analysis of hydrothermal alteration zones associated with base metal sulfide deposits at Rosebery and Western Tharsis, Tasmania, and Highway-Reward, Queensland: Economic Geology, v. 96, p. 939–955. Hill, RET 2001, Komatiite volcanology, volcanological setting and primary geochemical properties of komatiite-assocaited nickel deposits: Geochemistry: Exploration, Environment, Analysis, v. 1, p. 365–381. Hill, RET, Gole, MJ and Barnes, SJ 1988, Physical volcanology of komatiites: Geological Society of Australia, Excursion Guide Book No. 1, 74p. Houle, MG, Gibson, HL, Lesher, CM, Davis, PC, Cas, RAF, Beresford, SW and Arndt, NT 2008, Komatiitic Sills and Multigenerational Peperite at Dundonald

Beach, Abitibi Greenstone Belt, Ontario: Volcanic architecture and nickel sulfide distribution: Economic Geology, v. 103, p. 1269–1284. Huston, DL, Blewett, RS and Champion, DC 2012, Australia through time: a summary of its tectonic and metallogenic evolution: Episodes, v. 35, p. 23–43. Konnunaho, JP, Hanski, EJ, Bekker, A, Halkoaho, TAA, Hiebert, RS and Wing, BA 2013, The Archean komatiite-hosted, PGE-bearing Ni-Cu sulfide deposit at Vaara, eastern Finland: evidence for assimilation of external sulfur and post-depostional desulfurization: Mineralium Deposita, v. 48, p. 967–989. Kositcin, N, Brown, SJA, Barley, ME, Krapez, B, Cassidy, KF and Champion, DC 2008,

SHRIMP

U-Pb

zircon

age

constraints

on

the

Late

Archaean

tectonostratigraphic architecture of the Eastern Goldfields Superterrane, Yilgarn Craton, Western Australia: Precambrian Research, v. 161, p. 5–33. Laukamp, C, Caccetta, M, Chia, J, Cudahy, T, Gessner, K, Haest, M, Liu, YC, Ong, C and Rodger, A 2010, The uses, abuses and opportunites for hyperspectral technologies and derived geoscience information, in Geo-computing 2010 Conference: Brisbane, September 2010; AIG Bulletin 51, p. 73–76. Laukamp, C, Termin, KA, Pejcic, B, Haest, M and Cudahy, T 2012, Vibrational spectroscopy of calcic amphiboles - application for exploration and minng: European Journal of Mineralogy, v. 24, p. 863–878. Layton-Matthews, D, Lesher, CM, Liwanag, J, Halden, N, Burnham, OM, Hulbert, L, Peck, DC and Keays, RR 2011, Geochemistry and genesis of komatiite-associated Ni-Cu-PGE mineralization in the Thompson Nickel Belt, Manitoba, in Magmatic Ni-Cu and PGE deposits: geology, geochemistry and genesis (17 edition) edited by C Li and EM Ripley: Society of Economic Geologists, Reviews in Economic Geology. Le Vaillant, M, Barnes, SJ, Fisher, L, Fiorentini, ML and Caruso, S 2014, Use and calibration

of

portable

X-Ray

fluorescence

analysers:

application

to

lithogeochemical exploration for komatiite-hosted nickel sulphide deposits: Geochemistry: Exploration, Environment, Analysis, v. 14, p. 199–209.

Le Vaillant, M, Fiorentini, ML and Barnes, SJ 2016, Review of lithogeochemical exploration tools for komatiite-hosted Ni-Cu-(PGE) deposits: Journal of Geochemical Exploration. Lesher, CM 1989, Komatiite-assocaited nickel sulphide deposits, in: Society of Economic Geologists, Special Publication 4, p. 44–101. Lesher, CM 2007, Ni-Cu-(PGE) deposits in the Raglan area, Cape Smith Belt, New Quebec, in Mineral deposits of Canada: A synthesis of major deposit-types, district metallogeny, the evolution of geological provinces and exploration methods edited by WD Goodfellow: Geological Association of Canada, Mineral deposits division, St. John's (NL), p. 351–386. Lesher, CM, Arndt, NT and Groves, DI 1984, Genesis of komatiite-associated nickel sulphide deposits at Kambalda, Western Australia: a distal volcanic model, in Sulphide deposits in mafic and ultramafic rocks edited by DL Buchanan and MJ Jones: The Institution of Mining and Metallurgy, London, p. 70–80. Lesher, CM and Keays, RR 2002, Komatiite-associated Ni-Cu-(PGE) deposits: mineralogy, geochemistry, and genesis, in The geology, geochemistry, mineralogy, and mineral benefication of the Platinum-Group Elements edited by LJ Cabri: Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Special Volume 54, p. 579–617. Marston, RJ 1984, Nickel mineralization in Western Australia: Geological Survey of Western Australia, Mineral Resources Bulletin 14, 271p. Mole, DR, Burley, L and Barnes, SJ 2016, A new komatiite-hosted Ni-Cu-PGE event in the Eastern Goldfields Superterrane, in 13th international Ni-Cu-PGE symposium, Fremantle, Australia: Abstracts edited by B Godel, S Barnes, I Gonzales-Alvarez, M Fiorentini and M Le Vaillant: Geological Survey of Western Australia, Perth, Record 2016/13, p. 57. Mole, DR, Fiorentini, ML, Cassidy, KF, Kirkland, CL, Thebaud, N, McCuaig, TC, Doublier, MP, Duuring, P, Romano, SS, Maas, R, Belousova, EA, Barnes, SJ and

Miller, J 2013, Crustal evolution, intra-cratonic architecture and the metallogeny of an Archaean Craton: Geological Society, London, Special Edition. Mole, DR, Fiorentini, ML, Thebaud, N, Cassidy, KF, McCuaig, TC, Kirkland, CL, Romano, SS, Doublier, MP, Belousova, EA and Barnes, SJ 2014, Archean komatiite volcanism controlled by the evolution of early continents., in Proceedings of the National Academy of Sciences 111: p. 10083–10088. Murdie, R, Cudahy, T and Laukamp, C 2010, Exploration for gold using geophysical, hyperspectral and geological data: 21st International geophysical conference and exhibition, Sydney, Australia, 22 August 2010, 4p. Nesbitt, RW, Sun, SS and Purvis, AC 1979, Komatiites: geochemistry and genesis: Canadian Mineralogist, v. 17, p. 165–186. Olympus Corporation 2016, XRF and XRD Analyzers: viewed 26 August 2016, . Pawley, MJ, Wingate, MTD, Kirkland, CL, Wyche, S, Hall, CE, Romano, SS and Doublier, MP 2012, Adding pieces to the puzzle: episodic crustal growth and a new terrane in the northeast Yilgarn Craton, Western Australia: Australian Journal of Earth Sciences, v. 59, p. 1–20. Prendergast, M 2001, Komatiite-hosted Hunters Road nickel deposit, central Zimbabwe: physical volcanology and sulfide genesis: Australian Jounral of Earth Sciences, v. 48, p. 681–694. Rox Resources 2015, Nickel Sulphide Discovery Confirmed at Sabre Prospect: Report to ASX, 13/05/2015. Said, N, Kerrich, R, Cassidy, K and Champion, DC 2012, Characteristics and geodynamic setting of the 2.7Ga Yilgarn heterogeneous plume and its interaction with continental lithosphere: evidence from komatiitic basalt and basalt geochemistry of the Eastern Goldfields Superterrane: Australian Journal of Earth Sciences, v. 59, p. 737–763.

Schodlok, MC, Whitbourn, LB, Huntingtion, JF, Mason, P, Green, AA, Berman, B, Coward, D, Connor, P, Wright, W, Jolivet, J and Martinez, R 2016, HyLogger-3, a visible to shortwave and thermal infrared reflectance spectrometer system for drill core logging: functional description: Australian Journal of Earth Sciences, v. 63, no. 4, p. 1–12. Sonntag, I, Laukamp, C and Hagemann, SG 2012, Low potassium hydrothermal alterationin low sulfidation epithermal systems as detected by IRS and XRD: An example from the Ca-O miner, Eastern Mindanao, Philippines: Ore Geology Reviews, v. 45, p. 47–60. Sossi, PA, Eggins, SM, Nesbitt, RW, Nebel, O, Hergt, JM, Campbell, IH, O'Neill, HSC, Van Kranendonk, M and Davies, DR 2016, Petrogenesis and geochemistry of Archean komatiites: Journal of Petrology, p. 147–184. Swager, CP 1997, Tectono-stratigraphy of late Archaean greenstone terranes in the southern Eastern Goldfields, Western Australia: Precambrian Research, v. 83, p. 11–42. Wilson, AH 2003, A new class of silica enriched, highly depleted komatiites in the southern Kapvaal Craton, South Africa: Precambrian Research, v. 127, p. 125–141. Wyche, S, Farrell, TR and Liu, SF 1997, Introduction, in Archaean geology and mineralization of the northern part of the eastern goldfield province Yilgarn Craton, Western Australia - a field guide Kalgoorlie '97: Geological Survey of Western Australia, Record 1997/7, p. 1–7.

Figure Captions

Figure 1: Major greenstone belts in the Eastern Goldfields Superterrane, with selected nickel deposits and prospects highlighted. Terrane boundaries from Geological Survey of Western Australia, (2015).

Figure 2: Location of the Fisher East nickel sulphide prospects and drillholes within the Mount Fisher greenstone belt used in this study.

Figure 3: Stratigraphy in the Fisher East area from studied drillholes. The basal contact between komatiites and felsic metasedimentary rocks referenced at 0 level. No horizontal scaling has been applied to this figure.

Figure 4: Talc-rich komatiites: a) MFED069, ~250m; b) Thin section photo from drillhole MFED043-317m. Images taken by L Burley.

Figure 5: Chlorite-rich komatiites. a) MFED029-435m; b) MFED055-303m. Images taken by L Burley.

Figure 6: Thin section photo showing carbonate replacing primary igneous minerals in MFED043. Image taken by L Burley.

Figure 7: Komatiite interaction with metasedimentary rock on basal contact in drillhole MFED069 (Cannonball prospect, see Figures 2 and 3). Komatiite is outlined in red. Image taken by L Burley.

Figure 8: Metasedimentary rock clast within massive sulphides in drillhole MFED043 (Musket prospect, see Figures 2 and 3). Image taken by L Burley.

Figure 9: Upper contact zone between komatiite and BIF in drillhole MFED055 (Musket prospect, see Figures 2 and 3). Image taken by L Burley.

Figure 10: Differing styles of mineralisation at the Fisher East prospects. a) Sulphide veins in MFED069, Cannonball prospect; b) Massive sulphides from drillhole MFED043-305m, Musket prospect; c) Sulphides from drillhole MFED018-416m, Camelwood prospect.

Figure 11: Comparison of SWIR, TIR, whole-rock geochemical and pXRF data in drill hole MFED029. Mineral assemblages were interpreted from the SWIR and TIR hyperspectral drill core data using the TSA algorithm (Berman et al., 1999) built into TSGTM.

Figure 12: Al2O3/TiO2 ratios of the Fisher East prospects. Chondritic mantle line used following Barnes (2006).

Figure 13: Al2O3/MgO and SiO2/MgO plots of samples taken from the Fisher East prospects. Pure olivine lines used following Barnes (2006).

Figure 14: Incompatible trace elements of the Fisher East prospects, obtained from whole-rock geochemical data. Assumed pristine mantle ratio line and crustal contamination fields used following Barnes (2006).

Figure 15: Ni/Cr vs Ni/Ti ratios obtained from whole-rock (black) and pXRF (red) geochemical data to depict volcanic facies and discriminate rock type. Volcanic facies

subdivisions in the diagram enclose 80% of data points for each of the following facies assemblages: thin differentiated flow (TDF), channelised sheet flow (Kambalda style - CSF), dunitic channelised flows (DC) and layered lava lakes or sills (LLLS) (Barnes et al., 2004). Rock type discriminations fields from Le Vaillant et al., (2016).

Figure 16: Example of reflectance spectra for chlorite-rich komatiites (MFED069) and talc-rich komatiites (MFED043).

Figure 17: Comparison of hyperspectral indices derived from HyLogger3 TM data (n=43371) with geochemical indices derived from whole-rock XRF (n=67) and pXRF data (n360). Relative depth of the 2080nm absorption feature indicating talc abundance (x-axis) vs the relative depth of the 2250nm absorption feature indicating chlorite abundance (y-axis), coloured by a) Al2O3 (wholerock XRF), b) TiO2 (wholerock XRF), c) Mg# (wholerock XRF), and d) Ti (ppm; pXRF). Data points in grey have no associated geochemical values. Due to differences in sampling distribution between geochemical and HyLogger3TM data, each geochemical value displayed in these plots have several associated hyperspectral measurements.

Figure 18: a) Ni/Cr vs Ni/Ti plots showing empirically assigned fields of komatiiterelated rock types. b) Ni/Cr (x-axis) and Ni/Ti (y-axis) indices for talc-rich komatiites modelled from hyperspectral drill core data using the Partial Least Squares-based method; c) Ni/Cr (x-axis) and Ni/Ti (y-axis) indices for chlorite-rich komatiites modelled from hyperspectral drill core data using the Partial Least Squares-based method.

Figure 19: Drillhole MFED029; a) Ni/Ti values derived from whole rock geochemical data; b) Ni/Ti values derived from pXRF data; c) expected Ni/Ti values as modelled in TSGTM; d) relative abundance of talc derived from hyperspectral data; e) relative abundance of chlorite (+/- biotite) derived from hyperspectral data. Plots coloured by lithology type.

Figure 20: Drillhole MFED029; a) Ni/Cr values derived from whole rock geochemical data; b) Ni/Cr values derived from pXRF data; c) expected Ni/Cr values as modelled in TSGTM ; d) relative abundance of talc derived from hyperspectral data; e) relative abundance of chlorite (+/- biotite) derived from hyperspectral data. Plots coloured by lithology type.

Figure 21: Schematic flow field model of the prospects at Fisher East. The basal komatiite contact has been used as the horizontal datum, in order to depict variation in thickness in the flows. This model is more constrained in the northern area due to the large abundance of drillholes. The southernmost drillhole at Sabre is ~7km south of the Musket prospect, and therefore has been showed separated from the rest of the model. Note: this reconstruction is not to true horizontal scale, and the vertical scale is greatly exaggerated relative to horizontal.

Figure 22: Plastic deformation of BIF in contact with komatiite in drill hole MFED016. Image taken by L Burley.

Table Captions Table 1: Geochemical characteristics of chlorite-rich (A-zone; n=31) and talc-rich (Bzone; n=36) komatiites, as summarized by whole-rock geochemistry. Talc-rich komatiites contain nickel sulphides, and therefore on average, are expected to contain more Ni (ppm).

Table 2: Characteristics of A and B zone komatiites, as summarised from Barnes, 2006.

Appendices

Appendix 1: Whole-rock geochemical data, recalculated for volatiles and sulphides (except Ni and Cu component). Negative and 0 values are reported as ‘x’

Appendix 2: pXRF analysis

2.1: Whole-rock geochemical data for standards used during pXRF analysis, as used in Le Vaillant et al., 2014.

2.2: Plastic attenuation values, as previously used and documented in Le Vaillant et al., 2014

2.3: Calculated calibration curves for data correction

2.4: Calibrated pXRF data. Negative and 0 values are reported as ‘x’.

2.5: Instrument precision (using calibrated data). Process outlined in Le Vaillant et al., 2014.

2.6: Sample Precision (using calibrated data). Process outlined in Le Vaillant et al., 2014. 2.6.1: Calibration curves for sample precision 2.62: Calculated sample precision

Appendix 3: Lithological, geochemical and hyperspectral plots of all studied DDHs (refer to Figures 2 and 3 for locations of individual drillholes).

AK47

Wiluna

Fisher East

Honeymoon Well

Collurabbie

Mt Fisher greenstone belt

Fig. 2

Mt Keith

Wiluna YILGARN CRATON

Perth

Dingo Rangebelt greenstone

Yakabindie Cosmos

Duketon greenstone belt

Rosie

Leinster

Kalgoorlie

Rockys Reward Perseverance

28°

Weebo Bore

Waterloo

EASTERN GOLDFIELDS SUPERTERRANE Mt Windarra Windarra South

Terrane

i Kurnalp

errane

Ida Fault

Terrane

T Kalgoorlie

Leonora

Laverton

Burtville

Marriotts

Cambridge Dragon

Menzies Mulga Tank Scotia Mt Jewell Silver Swan

30°

Emu Lake

Kalgoorlie–Boulder Duplex Hill

Coolgardie Nepean

Kambalda

Spargoville

Talc Lake Lanfranchi

Miitel

Major lithologies 32°

Norseman 120°

122°

100 km Ni sulfide deposit/Prospect

Ultramafic rock EGST greenstones Granitic rock Granitic gneiss Volcanic/volcaniclastic felsic rock Volcanogenic sedimentary rock Siliciclastic sedimentary rock

Town Figure 1, Burley et al

Kurnalpi Terrane greenstones Ultramafic rocks Banded iron-formation Granite 26°30'

Granitic gneiss Felsic volcanic rock Sedimentary rock Sedimentary rock Sedimentary rock Sedimentary rock

MFED016 Camelwood MFED015 MFED018 MFED001 MFED029 MFED060

Cannonball MFED069

MFED043

MFED064

Musket

MFED055

Enlargement

27°00'

Sabre

MFED071

121°30'

10 km

2 km Figure 2, Burley et al

N

S CAMELWOOD

MFED016

MFED015

MFED018

MFED001

MFED029

CANNONBALL MFED060

MFED069

MFED064

MUSKET MFED043

SABRE MFED055

MFED071

Stratigraphic way up

~20m

Legend: Lithology Amphibolite Chlorite-rich komatiites Talc-rich komatiites Banded Iron Formation Basalt

‘Felsic’ Dyke Metasedimentary rocks ‘Mixing’ between metasedimentary rocks and komatiites

Porphyry Shale Sulphidic Chert Unknown Figure 3, Burley et al

a)

talc rich matrix

carbonate

2 cm

b) carbonate

talc 500 µm

Figure 4, Burley et al

a)

1 cm

b)

1 cm

Figure 5, Burley et al

500 µm Figure 6, Burley et al

Felsic metasedimentary rock

Komatiite

Figure 7, Burley et al

2 cm

Figure 8, Burley et al

BIF

Contact zone

Komatiite

5 cm Figure 9, Burley et al

a)

2 cm

b)

2 cm

c)

2 cm

Figure 10, Burley et al

MgO%

448

Al2O3%

Ni ppm

TiO2%

Zr ppm

Cr ppm

Mg#

Ni/Cr

15

10

5

0

7.5

5

2.5

92 0

52 72

12 32

3000

2000

1000

0

300

200

100

0

1.5

1

0.5

4000 0

3000

1000 2000

16 0

11

6

1

TIR

40

SWIR

20

Stratigraphic Log

0

MFED029 Ni/Ti

444 440 436 432 428 424 420 416 412 408 404 400 396 392 388 384 380

Legend: Lithology Chlorite-rich (A-zone) komatiites Talc-rich (B-zone) komatiites BIF Sediments

Legend: SWIR Shale Sulphidic Chert Unknown

White-mica Chlorite Talc and Brucite Carbonate Invalid

Legend: TIR Silica K-feldspar Plagioclase White-mica Chlorite Talc and Brucite Carbonate

Legend: Geochemical plots Whole-rock pXRF Whole-rock geochemical sample point (approximate)

Burley et al, Figure 11

0.8

TiO2 (%)

Al2O3/TiO2

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Al2O3 (%) 0 0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

11.0

12.0

13.0

14.0

Figure 12, Burley et al

Al2O3 wt%

Pure Olivine

SiO2 wt%

MgO wt%

Pure Olivine

MgO wt%

Figure 13, Burley et al

Incompatible Trace Element Ratios 2.0

0.8

TiO2 0.7 (%)

Sm (ppm)

1.8

Assumed pristine mantle ratio

Assumed pristine mantle ratio

1.6

0.6

1.4

0.5

1.2

Crustal Contamination

0.4

Crustal Contamination

1.0 0.8

0.3

0.6

0.2

0.4 0.1

Zr (ppm) 0

0

3

40

20

60

80

0.2 0

La (ppm) 0

2.5

Nb (ppm) Assumed pristine mantle ratio

2.5

Yb (ppm)

2

2

4

6

Assumed pristine mantle ratio

2 1.5

1.5 1

1

Crustal Contamination

Crustal Contamination

0.5

0.5

Th (ppm) 0

0

0.2

0.4

0.6

0

Th (ppm) 0

0.5

1

Figure 14, Burley et al

2.0

Fisher East pXRF geochemistry Fisher East whole-rock geochemistry

1.5

DC 80

chromite-rich cumulates LLLS 80

1.0 Log (Ni/Ti)

Adcumulates

sulfide-bearing cumulates

0.5 Mesocumulates Orthocumulates

0

Volcanic facies CSF TDF LLLS DC

–0.5 Spinifex zones, flow tops –1.0

Komatiitic basalts TDF

CSF 80

–1.5

–0.8

–0.6

–0.4

–0.2

0

Log (Ni/Cr)

0.2

0.4

0.6

0.8

1.0

Figure 15, Burley et al

Chlorite-rich komatiites

Figure 16, Burley et al

Figure 17, Burley et al

a)

b)

c)

Figure 18, Burley et al

Figure 19, Burley et al

Figure 20, Burley et al

POSSIBLE FLOW CHANNELS Camelwood

Sabre

Musket

Cannonball

N

S

?

?

? ?

?

?

~20 m

?

?

?

?

? ?

Amphibolite Basalt Banded ironformation Felsic dyke

Felsic metasedimentary rock Chlorite-rich (A-zone) komatiite

Porphyry

Talc-rich (B-zone) komatiite

Shale

‘Mixing’ between metasedimentary rocks and komatiites

Sulfidic chert

?

Unknown lithology Figure 21, Burley et al

Figure 22, Burley et al

Table 1 Geochemistry

Chlorite-rich komatiites

Talc-rich komatiites

MgO (%) average

29.15

36.6

MgO (%) lowest value

24.21

25.72

MgO (%) highest value

33.7

48.76

Al2O3 (%) average

6.02

2.34

Al2O3 (%) lowest value

2.14

0.39

Al2O3 (%) highest value

11.96

4.08

Ni (ppm) average

1641

6543

TiO2 (%) average

0.24

0.09

Zr (ppm) average

17

8

Table 2

Primary igneous textures Primary mineralogy Geochemical characteristics Mineralogy in CO2 rich conditions – greenschist facies

A zone komatiites Spinifex textures Olivine and/or pyroxene MgO wt%: 20-32 Higher Al2O3, TiO2 and SiO2 Tremolite-chlorite-dolomite

B-zone komatiites Cumulate textures (ad, meso and ortho cumulates) Olivine MgO wt%: >32 Lower Al2O3, TiO2 and SiO2 Quartz-magnesite-dolomitechlorite or talc-chloritedolomite-magnesite (ortho and mesocumulates) or talcmagnesite (adcumulates)

MgO%

448

Al2O3%

Ni ppm

TiO2%

Zr ppm

Cr ppm

Mg#

444 440 436 432 428 424 420 416 412 408 404 400 396 392 388 384 380

Legend: Lithology Chlorite-rich (A-zone) komatiites Talc-rich (B-zone) komatiites BIF Sediments

Legend: SWIR Shale Sulphidic Chert Unknown

White-mica Chlorite Talc and Brucite Carbonate Invalid

Legend: TIR Silica K-feldspar Plagioclase White-mica Chlorite Talc and Brucite Carbonate

Legend: Geochemical plots Whole-rock pXRF Whole-rock geochemical sample point (approximate)

Ni/Cr

15

10

5

0

7.5

5

2.5

92 0

52 72

12 32

3000

2000

1000

0

300

200

100

0

1.5

1

0.5

4000 0

3000

1000 2000

16 0

11

6

1

TIR

40

SWIR

20

Stratigraphic Log

0

MFED029 Ni/Ti

Highlights: •

Looks at nickel mineralization in a previously unstudied area

•

The Fisher East prospects have characteristics similar to the Kambalda deposits

•

Hyperspectral and pXRF data can aid exploration of komatiite-hosted NiS