Abstract
The hydrogeochemistry of the Yilgarn Craton and its margins has been extensively investigated, with particular emphasis on the chemistry of Au. Four groundwater regions have been delineated based on variations in salinity, acidity and oxidation potential: (1) Northern (N Yilgarn and margins) – Fresh and neutral, trending more saline in the valley axes; (2) Central – Neutral and brackish (commonly <1% TDS) to saline (about 3% TDS), trending to hypersaline (10–30% TDS) at the salt lakes, with common increases in salinity with depth; (3) Kalgoorlie – Commonly acid (pH 3–5), except where buffered by extremely alkaline materials (e.g. ultramafic rocks), and saline within the top part of the groundwater mass, trending more neutral (pH 5–7) and hypersaline at depth and within a few kilometres of salt lakes; and (4) Eastern (E Yilgarn and Officer Basin) – Saline to hypersaline, neutral to acid and reducing. Dissolved concentrations of many ions are low, due to the presence of lignites in the channel sediments.
These regional variations have major effects on the concentrations of many elements. Aluminium, Li, Y, REE and U are dominantly controlled by pH and thus have higher concentrations in acid groundwaters, such as those in the Kalgoorlie region. Dissolved concentrations of Mn, Co, Ni, Cu and Zn are less closely correlated with acidity, and show scope for lithological discrimination, but there is no apparent relationship with Au mineralization. Dissolved Cr shows an absolute correlation with ultramafic rocks, apparently irrespective of pH, possibly due to its presence as chromate (i.e. Cr6+ as CrO42−). Concentrations of As, Sb Mo, W and Bi are low in acid groundwaters, but are higher above pH 6.5, particularly in the Central region. Therefore, acid groundwaters (particularly in the Kalgoorlie district) will be poor media for the use of these elements as exploration pathfinders. Molybdenum differs from the other elements in this group in having significant concentrations in acid groundwaters, although lower than in neutral and alkaline groundwaters.
Dissolved Au is commonly the best pathfinder for Au mineralization. It occurs dominantly as halides (chloride and/or iodide) and has enhanced concentrations (to >1 ppb) under the acid/saline/oxidizing conditions common in the Kalgoorlie region, whereas concentrations in the northern Yilgarn are two orders of magnitude less. This implies that supergene Au remobilization should be considerably less in the northern Yilgarn than in the Kalgoorlie region. Additionally, the threshold dissolved Au concentration as used for Au exploration differs significantly between regions.
INTRODUCTION
Hydrogeochemistry may be useful for exploration for Au and other metals, and may also provide information on how various materials are weathering. This enhances understanding of active dispersion processes and assists in the development of weathering and geochemical models, which are essential for effective exploration in regolith-dominated terrain. The aims of the hydrogeochemical studies detailed here were to: (a) provide information on whether groundwater can be used successfully as an exploration medium in the Yilgarn Craton and adjoining belts; (b) yield data on geochemical dispersion processes; (c) create a groundwater database of groundwaters at various mineralized sites; (d) enhance understanding of groundwater processes in mineralized zones; and (e) develop techniques for interpretation of groundwater data from mineralized areas.
This paper develops a regional model of groundwater characteristics and discusses important observations from these investigations.
STUDY SITES
Groundwaters within the Yilgarn Craton directly investigated by CSIRO/CRCLEME (⇓Fig. 1) involved approximately 450 groundwater samples. These are mainly from areas of extensive, deep (20–100 m) regolith within unconfined aquifers, with the water-table commonly 10–60 m below surface. Thus, these groundwaters are mostly in contact with weathered Archaean rocks. Some groundwaters have been sampled from palaeochannels, in which the aquifer is commonly confined within permeable sands and gravels overlain by clays. On the basis of salinity, pH and oxidation potential of the waters, the groundwaters are regionally grouped (⇑Fig. 1), as follows:
Location map of Yilgarn Craton groundwater sites, split into four regions as indicated by the dashed lines, with rainfall (mm) isohyets (Australian Bureau of Meteorology).
(1) Northern groundwaters (N Yilgarn and margins) – Baxter/Harmony (⇓Gray 1995), Lawlers (⇓Gray 1994), Windara (⇓Gray & Longman 1997) and Mt Joel (⇓Porto et al. 1999).
Groundwaters in these areas are fresh and neutral, trending more saline in the valley axes.
(2) Central groundwaters (close to and north of the Menzies line, an E–W transitional zone with marked changes in soils, vegetation and groundwaters; ⇓Butt et al. 1977) – Granny Smith (⇓Gray 1993a), Golden Delicious (⇓Bristow et al. 1996), Mt Gibson (⇓Gray 1991) and Boags (⇓Gray 1992a).
Groundwaters are neutral and brackish (commonly <1% TDS) to saline (about 3% TDS), trending to hypersaline (10–30% TDS) at the salt lakes, with common increases in salinity with depth.
(3) Kalgoorlie groundwaters – Twin Peaks/Monty Dam (⇓Sergeev & Gray 1999), Carosue Dam (⇓Gray et al. 2000), Mulgarrie (⇓Gray 1992c), Panglo (⇓Gray 1990a), Baseline mine, Steinway palaeochannel (⇓Lintern & Gray 1995a), Golden Hope mine, (⇓Gray 1993b), Wollubar palaeochannel (⇓Gray 1993b) and Lake Cowan (⇓Gray 1992b).
These groundwaters are commonly acid (pH 3–5), except where buffered by extremely alkaline materials (e.g. ultramafic rocks), and saline within the top part of the groundwater mass, trending to more neutral (pH 5–7) and hypersaline at depth and when within a few kilometres of salt lakes. Fresher conditions are only observed close to major drainage divides.
(4) Eastern groundwaters (E Yilgarn and Officer Basin) – Argo palaeochannel (⇓Lintern & Gray 1995b) and Mulga Rock palaeodrainage system (⇓Douglas et al. 1993).
Groundwaters are saline to hypersaline, neutral to acid and reducing. The major ion chemistry is similar to that of the Kalgoorlie region, but the dissolved concentration of many other ions is low, due to the presence of lignites in the channel sediments.
Detailed site information can be obtained from the site reports listed above. Additional groundwater data were obtained from various sources, including the Water and Rivers Commission of Western Australia, open file mineral exploration reports from Department of Minerals and Energy, W.A.; CSIRO research (⇓Giblin 1990; ⇓Turner et al. 1993; ⇓Giblin & Mazzucchelli 1997), and from WMC Resources Ltd., and Centaur Mining Ltd. These additional data were commonly limited to a smaller analytical suite than used for the site studies, but were far more extensive in number and distribution across the Yilgarn Craton.
METHODS
Groundwaters were commonly directly collected within drill holes, preferably 5 m or more below the water-table. Waters were analysed for pH, temperature, conductivity and oxidation potential (Eh) when sampled, and a sample collected, with overfilling to remove all air, for later HCO3 analysis by alkalinity titration in the laboratory. About 1.5 l of water was filtered (0.2 μm) in the field: 100 ml of the filtered solution was acidified and analysed in the laboratory by ICP-OES and ICP-MS for a number of elements, generally including Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hg, Ho, I, K, La, Li, Mg, Mn, Mo, Na, Nd, Ni, P, Pb, Pr, Rb, Sb, Sc, Sm, Sn, SO4, Si, Sr, Tb, Th, Ti, Tl, Tm, U, V, Y, Yb, Zr and Zn; 50 ml of filtered water were separately analysed for Cl and Br; a one litre sub-sample of the filtered water was shaken with a one gram sachet of activated carbon in a saline/acid medium and the carbon analysed for Au (quantitative to <0.005 μg l−1) and other elements (qualitative for As, Mo, W, U, La, Ce and Sm). More saline groundwaters required specific analytical routines, leading to higher detection limits, particularly for As and REE.
Additional understanding of groundwater processes was obtained by computing the solution species of many of the major and trace elements and degree of mineral saturation from solution compositions. For most elements, this was done using the program PHREEQE (⇓Parkhurst et al. 1980) with the supplied thermodynamic database enhanced in-house to include additional chalcophile elements of interest (e.g. Sb, Bi) and REE. To obtain highly accurate speciation data on a limited suite of the major ions (Na, K, Mg, Ca, Cl, HCO3, SO4, Sr and Ba) for highly saline solutions, the specific ion interaction model, the Pitzer equations, was applied, using the program PHRQPITZ (⇓Plummer & Parkhurst 1990). These programs calculate the solubility indices (SI) for each water sample for various minerals. If the SI for a mineral equals zero (empirically from −0.2 to 0.2 for the major element minerals, and −1 to 1 for the minor element minerals), the water is in equilibrium with that mineral, under the conditions specified. Where the SI is less than zero, the solution is under-saturated with respect to that mineral, so that, if present, the phase may dissolve. If the SI is greater than zero, the solution is over-saturated with respect to this mineral, which can potentially precipitate from solution. Note that this analysis only specifies possible reactions, as kinetic constraints may rule out reactions that are thermodynamically allowed. Thus, for example, waters are commonly in equilibrium with calcite, but may become over-saturated with respect to dolomite, due to the slow rate of solution equilibration and precipitation of this mineral (⇓Drever 1982).
COMPILATION OF RESULTS
Median results for groundwaters of the four regions are shown in ⇓Table 1. Better coverage in the Kalgoorlie area allowed further division into four sub-sets (⇓Table 2): (1) groundwaters sampled below c. 50 m depth or sampled from saline playas – neutral and hypersaline; (2) ‘normal’ groundwaters contacting regolith overlying a variety of rock types (except ultramafic rocks), away from the major drainage divides – acid and saline; (3) groundwaters sampled from weathered ultramafic rocks – neutral and saline; and (4) groundwaters close to major drainage divides – neutral and fresh to brackish.
Median groundwater data for the Yilgarn regions
Median groundwater data for different sub-sets of the Kalgoorlie region
SALINITY VARIATION ACROSS THE YILGARN CRATON
Various researchers (e.g. ⇓Bestow 1992) have observed regional variations in groundwater salinities across the Yilgarn Craton. ⇓Commander (1989) modelled groundwater salinities (⇓Fig. 2), showing the highest salinities along the major drainages, which generally match palaeodrainage channels (⇓Fig. 3). In addition, there are major regional differences. The SE Yilgarn, roughly centered on the Kalgoorlie mining district, has the most saline conditions. More recent salinity modelling (⇓Kern 1995a, ⇓b, ⇓1996a, ⇓b) shows pervasive salinities >30% within the Kalgoorlie region and eastwards. North of the Menzies Line, a narrow east–west transitional zone across which there are marked changes in soil types and vegetation (⇓Butt et al. 1977), groundwaters are significantly less salty and higher salinities are confined to the immediate vicinity of the major drainage channels. Compiled results from >2900 salinity measurements (D. Gray, unpublished data) support these observations.
Modelled groundwater salinities in the Yilgarn Craton (⇓Commander 1989).
Elevation of the Yilgarn Craton, showing principal sediment-filled channels, N–S drainage divide and Menzies line.
The major origin of the salt appears to be sea water, presumably as aerosols (⇓McArthur et al. 1989), with concentration due to evaporation. ⇓Bestow (1992) calculated that salinities >1.4% (observed throughout much of the southern Yilgarn) would require loss of >99.9% of the rainfall through evaporation. One commonly suggested reason for the salinity difference from north to south are climatic changes from primarily winter rainfall in the south, to irregular summer cyclonic rainfall, with high run-off, in the north. However, differences in elevation may also be important. Elevations along the drainage divide are at least 100 m higher in the northern Yilgarn (leading to much greater piezometric head differences), relative to the south (⇑Fig. 3). Indeed, the extremely flat nature of the SE Yilgarn is demonstrated by the change in elevation of only 18 m over more than 300 km along the Johnston–Lefroy drainage system (⇓Fig. 4). Low piezometric head differences in the southwest will cause very slow groundwater flows and high salinities due to evaporation and recharge of saline waters.
The four groundwater regions are reflected in a plot of pH v. TDS (⇓Fig. 5): Northern groundwaters are generally neutral and fresh; Central groundwaters range from neutral and fresh close to groundwater divides to highly saline adjacent to salt lakes; Kalgoorlie groundwaters vary from saline and acid close to the surface to hypersaline and neutral at depth and close to salt lakes; Eastern groundwaters have characteristics similar to those in the Central and Kalgoorlie districts, although commonly with lower Eh. The few Kalgoorlie groundwaters that are neutral and saline (rather than hypersaline) are shallow groundwaters in contact with weathered ultramafic rocks. In all but the Northern region, salinities show major increases with depth, probably due to back-flow from salt lakes (⇓Fig. 6). These sub-regional effects are specifically deliniated for the Kalgoorlie groundwaters (⇑Table 2, ⇑Fig. 5).
pH v. TDS for groundwaters in Western Australian study areas.
Diagrammatic representation of groundwater evaporation and flow at salt lakes: (1) Evaporation, leading to saline and dense groundwater; (2) Downward flow of dense groundwater; (3) Back-flow of saline waters, leading to higher salinity at depth.
VARIATION IN ACIDITY AND OXIDATION POTENTIAL
Although the major influence on salinity appears to be introduced sea water, there are strong local controls on acidity (⇓McArthur et al. 1989) and, presumably, oxidation potential. Deep groundwaters in contact with mineralization commonly have high concentrations of dissolved Fe and other chalcophile elements, probably derived from the first stage of the oxidation of pyrite and other sulphides:
At depth, acid production is buffered by minerals such as carbonates or feldspar. These groundwaters (solid symbols in ⇓Fig. 7) contain significant dissolved Fe (>0.1 mg l−1), are neutral (pH 6–8) and reducing, commonly having Eh values of 200 mV or less.
Eh v. pH in groundwaters from the Yilgarn Craton: solid symbols indicate samples with [Fe] >0.1 mg l−1 (ppm); grey-filled symbols have [Fe] <0.1 mg l−1 and [Mn] >0.1 mg l−1; open symbols have [Fe] and [Mn] <0.1 mg l−1. The Mn line is derived using data from ⇓Moussard et al. (1974), assuming [Mn]=10−4 M (5.5 mg l−1). The striped area is zone where [Au] >2 μ l−1 (ppb) in 1 M (5.7%) NaCl solution, with stippled area showing the increased Eh field in which 2 μ l−1 Au will dissolve in the presence of 10−5 M (1.3 ppm) I. The dashed line is lower Eh limit for [Au]=0.2 μ l−1 in 10−5 M I/1 M NaCl solution (derived from thermodynamic data compiled in ⇓Gray 1988).
Closer to the surface, conditions are more oxidizing, and soluble Fe2+ oxidizes to Fe3+, which then precipitates as an oxide/hydroxide, generating acidity (Eqn 2). As this occurs higher in the profile, and (unlike the initial phase of sulphide weathering; Eqn 1) buffering minerals are generally absent, highly acid conditions can ensue. This critical groundwater Eh/pH control is known as the ‘ferrolysis’ reaction (⇓Brinkman 1977).
Because the reaction is governed by both Eh and pH, the ferrolysis control is an angled line on an Eh/pH diagram (denoted by the Fe line in ⇑Fig. 7). Groundwaters with significant Fe concentrations (>0.1 mg l−1; solid symbols in ⇑Fig. 7) congregate around this line.
Under most weathering conditions, there is an absolute groundwater pH limit of 3, due to buffering by dissolution of aluminosilicates such as kaolinite:
This maintains the solution pH at about 3.2. The control is Eh independent (the Al line in ⇑Fig. 7). In the more acidic groundwaters, dissolved Si concentrations are high, and reach saturation with amorphous silica. Under these acid conditions, Al is reprecipitated as alunite [KAl3(SO4)2(OH)6], as evidenced by the major K and moderate SO4 depletion (⇓McArthur et al. 1989). This effect is particularly marked in the Kalgoorlie area (⇓Fig. 8) where groundwaters are acidic.
Dissolved K v. salinity for Yilgarn Craton groundwaters.
Many acidic groundwaters have significant Mn contents (⇑Table 2), resulting in very high but stable Eh values (up to 850 mV; shown as the Mn line in ⇑Fig. 7) in the Mn analogy to ferrolysis.
The final type of groundwater observed in the study areas is that containing neither significant Fe nor Mn (open symbols in ⇑Fig. 7). These waters are almost exclusively neutral (pH 6–8.5), and include virtually all of the Northern, some Central, but few of the Kalgoorlie groundwaters (⇑Figs 7 & ⇓9). Although there are redox couples within the range measured for these waters (the most probable being H2O2/O2; ⇓Sato 1960), these reactions are slow and the solution Eh will be weakly controlled. This group is the least chemically active of the waters, being neither strongly reducing, strongly acidic or strongly oxidizing. Iron-rich groundwaters have been observed in the northern Yilgarn Craton, but only where they are in contact with weathered massive sulphide orebodies, such as the Harmony Ni deposit near Leinster (⇓Gray et al. 1999).
Groundwater pH distribution in the Yilgarn Craton.
The four Eh/pH groups [Fe2+/Fe(OH)3, Mn2+/MnXOY, Al3+/kaolin and H2O2/O2)] discussed above appear adequate to model the Eh/pH data from all Yilgarn Craton sites investigated to date. The Eh/pH characteristics vary from neutral to moderately acid and reducing at the extreme southeast edge of the Yilgarn Craton (where there are extensive lignite within the palaeochannels), neutral to highly acid and oxidizing in Kalgoorlie, and neutral in the Central and Northern Regions (⇓Fig. 10).
Groundwater Eh distribution in the Yilgarn Craton.
MAJOR ELEMENT HYDROGEOCHEMISTRY
The major elements and/or ions can be classified as those primarily controlled by salinity effects (i.e. Na, K, Mg, Ca, Cl, SO4, Br) as well as HCO3, Si, Al and Fe. The concentrations of the salinity-controlled ions in Central, Kalgoorlie and Eastern groundwaters suggest a sea water source, as discussed above. The most critical modifications are reductions in SO4 and K, particularly in the Kalgoorlie region (⇑Fig. 8), due to alunite precipitation (Eqn 3) and significant depletion of Br in the Kalgoorlie and some of the Central region, probably due to oxidation of Br− to Br2 in these more acidic groundwaters. However, for the low salinity Northern groundwaters, the host rocks appear to be a major control (⇓Gray 1994).
In some circumstances, the concentrations of most of the major elements in groundwater are controlled by equilibria with secondary minerals, as demonstrated using speciation analysis. Depending on conditions, the important phases are halite and gypsum in saline environments, calcite, magnesite and sepiolite [Mg2Si3O7.5(OH) · 3H2O] in neutral/saline conditions, and alunite and amorphous silica in acidic environments.
Ion/TDS ratios that differ from a regional norm may relate to mineralization or other sources that differ from the surrounding country rock. For example, localized enrichments in SO4, as observed for the Golden Hope (SO4/TDS=0.13) and Boags (SO4/TDS=0.17) mines in the Kalgoorlie region compared to sea water (SO4/TDS=0.08), and in groundwaters from overburden above the Four Corners mineralization at Lawlers in the Northern region (SO4/TDS>0.23; ⇓Fig. 11), may reflect oxidation of major sulphide bodies.
SO4/TDS distribution in groundwaters at Four Corners, Lawlers study area, Northern Yilgarn Craton.
MINOR ELEMENT HYDROGEOCHEMISTRY
The median concentrations of a range of minor elements in Yilgarn Craton groundwaters are listed in ⇑Tables 1 and ⇑2. Many elements, including Al, Li, Y, REE and U, have higher concentrations in acid groundwaters, such as those in the Kalgoorlie region. In particular, light REE concentrations (e.g. La and Ce) in the Kalgoorlie region are commonly greater than those for most of the base metals. A compendium of recently published data on dissolved REE concentrations (⇓Gray 1996) shows the average REE content of the Wollubar groundwaters (e.g. 0.47 mg l−1 La) to be about five times that of the highest recorded REE concentrations outside the Yilgarn Craton (Carnmenellis metasediment, England; e.g. 0.095 mg l−1 La; ⇓Smedley 1991), and the highest individual sample is about 20 times greater. Calculations indicate that approximately 4000 kg per annum of total REE are extracted from the Wollubar palaeochannel bore field (⇓Gray 1996). The high REE concentrations are primarily due to the reaction of highly acid and saline groundwaters with regolith.
Dissolved concentrations of the base metals (Mn, Co, Ni, Cu and Zn) and Ga are less closely correlated with acidity than the REE. Thus, these elements have potential for lithological discrimination (as discussed later in this paper), but there is no apparent relationship with Au mineralization. On the basis of its normal aqueous chemistry, Cr shows a surprising lack of correlation with acidity: dissolved Cr concentrations can be very high, with no pH relationship, and are above detection only where groundwaters are in contact with ultramafic rocks. Moreover, Cr concentrations appear to be strongly over-saturated with respect to secondary Cr oxides above pH 6. Spectrophotometric analyses indicate that this Cr is present as Cr6+ (i.e. CrO42−), which has a much higher mobility than Cr3+. A high oxidation state of Cr is also suggested by its highly anti-pathetic relationship with Fe (⇓Gray 1996), possibly due to the capacity of dissolved Fe2+ to reduce CrO42− to the less soluble Cr3+ ion.
Other elements, including As, Sb Mo, W and Bi, have low concentrations in acid groundwaters, but have higher concentrations in waters with pH>6.5, particularly in the Central and Northern region (⇓Fig. 12). These elements commonly occur as oxy-anions (e.g. H2AsO4−), which are better adsorbed by Fe oxide (or other) surfaces at low pH. This is because surfaces are positively charged when the pH is below the point of zero charge (PZC) of the mineral phase, and are therefore more effective at adsorbing negative ions. This implies that the acid (particularly Kalgoorlie) groundwaters will be poor media for the use of these elements as exploration pathfinders. Molybdenum differs from the other elements in this group in having significant, though lower, concentrations at low pH. The reason for this difference is unclear, but the potential of Mo as a pathfinder element in groundwater should be investigated further.
As v. pH for Yilgarn Craton groundwaters.
The concentration and speciation of dissolved I may affect the solubility of Au. Iodine may be present in a number of forms with contrasting oxidation states, including I− (as free, weakly or strongly bound iodide), I2, IO− and IO3−. The iodide ion may be an important ligand for Au (as discussed below), whereas the other, oxidized, forms could be important in maintaining oxidizing conditions for Au dissolution. Groundwater I contents (commonly >1 ppm; ⇑Tables 1 and ⇑2) are much greater than for sea water. Iodine (⇓Gray 1996; references therein) may behave as a chalcophile element, with major enrichments associated with sulphides.
The Northern groundwaters, in contrast to other regions, have low concentrations of most elements, with a few exceptions:
P and V concentrations are higher than in other groundwaters, probably due to the low concentration of metals that cause precipitation of these elements, particularly Ca (which precipitates P as apatite), and Fe (P as strengite or adsorbed on Fe oxides, V as Fe vanadate);
Si concentrations match those at the other regions;
Sc concentrations are only slightly lower on average than the other regions;
Cr concentrations can be high in the Northern region, with a similar amount of enhancement in ultramafic groundwaters as elsewhere in the Yilgarn Craton;
although dissolved concentrations of the oxy-anions are lower in the Northern groundwaters than in the Central groundwaters, As, Mo, Sb and W still have significant concentrations;
Rb concentrations are much higher than expected, on the basis of the normal sea water Rb/TDS and/or Rb/K ratios. Rubidium contents may reflect the presence of muscovite, which commonly occurs as an alteration mineral in Archaean Au deposits (⇓Eilu & Groves 2001), hence Rb in non-saline groundwater could be useful as a Au pathfinder.
GOLD HYDROGEOCHEMISTRY
Important ligands
Gold may be dissolved by a variety of ligands, including halides, S compounds having intermediate oxidation states, and a range of organic ligands (⇓Gray et al. 1992). However, the poor solubility of Au in groundwaters in most environments, except in highly saline conditions and close to weathering sulphides, indicates that only the first two ligands are important in most groundwater situations.
Thiosulphate dissolution
Initial weathering of the sulphides accompanying Au mineralization might generally be expected to create acid conditions and generate sulphate. However, if sufficient carbonates are present, neutral to alkaline conditions will prevail, and significant concentrations of the intermediate sulphur compounds thiosulphate (S2O32−) and sulphite (SO32−) are observed (⇓Granger & Warren 1969; ⇓Goldhaber 1983; ⇓Mann & Webster 1990):
Under such conditions, Au will dissolve as the thiosulphate complex [Au(S2O23)3−]. However, in acid environments, thiosulphate is unstable and Au thiosulphate is only expected to be important for the mobilizing of Au at the active weathering front.
A few groundwaters from mineralized rock or saprock at Lawlers, Boags and Mt Gibson have anomalous Au (1–41 μg l−1), but only when sampled in mine pits. Thus, these high dissolved Au concentrations may be partially or wholly caused by the release of unnaturally high concentrations of thiosulphate when pit excavation exposed bedrock. In comparison, groundwaters from drill holes in otherwise undisturbed mineralized rocks at Lawlers and Mt Gibson contain much lower dissolved Au concentrations (0.001–0.03 μg l−1; ⇓Gray 1991, ⇓1994). Whatever the cause for the high dissolved Au concentrations in these conditions, it is unlikely that sampling deep groundwaters for Au will be an effective exploration method.
Halide (iodide and chloride) dissolution
In shallow groundwaters, particularly in the Kalgoorlie region, the high acidity destabilizes thiosulphate. However, saline/acidic/oxidizing groundwaters are highly effective in dissolving Au as the chloride or iodide complex (⇑Fig. 7). The conditions under which up to 2 μg l−1 Au will be dissolved as AuCl2− in a 1 M Cl solution (about twice sea water) are indicated in ⇑Figure 7 as the striped grey area. This approximates the mean salinity for Kalgoorlie groundwaters and the upper range for Central groundwaters. A significant proportion of the Kalgoorlie groundwaters are within the Eh range required. However, thermodynamic calculations suggest that AuCl2− may not be the most significant complex, even in these highly saline conditions. Based on free iodide and total iodine determinations, available iodide (free plus loosely complexed iodide) are commonly expected to be greater than 10−5 M (1.3 ppm). Because Au complexes strongly with iodide, this concentration of available iodide considerably extends the theoretical Eh range for the dissolution of at least 2 μg l−1 Au from 690 to 600 mV (⇑Fig. 7). If a lower dissolved Au concentration of 0.2 μg l−1 is used, the required Eh for Au dissolution is lowered to about 550 mV (i.e. to a range which includes many of the Kalgoorlie groundwaters).
The possibility of Au iodide dissolution is also important because sorption studies (⇓Gray 1990b, unpublished data) indicate that Au iodide is more poorly absorbed than Au chloride by most regolith and soil materials, even in neutral conditions. This implies a high mobility, and suggests the possibility that Au could diffuse upwards into the overlying transported overburden or soil to form anomalies above mineralization.
As expected, there is a weak correlation between dissolved Au concentrations and Eh (⇓Fig. 13). However, it is clear that virtually none of the Central groundwaters have Eh values high enough to explain the observed concentrations of dissolved Au. One possibility is that surface or micro-processes generate transient highly oxidizing conditions that can dissolve Au. Once dissolved, the Au-halide complex is meta-stable, and may be only slowly removed from solution, except in Fe-rich solutions.
A number of acid groundwater sites (⇓Gray 1996) show a strong antipathetic relationship between dissolved Au and Fe, supporting the hypothesis that dissolved Fe is important in precipitating Au from halide complexes (Eqn 5). Such a mechanism has major implications for the control of supergene depletion and enrichment zones, and for the interpretation of drilling data.
Dissolved Au v. Eh for Yilgarn Craton Groundwaters.
Regional differences in Au hydrogeochemistry and implications for supergene mobilization
The importance of halides for the dissolution of Au is illustrated by the low dissolved Au concentrations in the Northern region (⇓Fig. 14), which have groundwaters with low Cl, I and Eh. Over 80% of the Northern groundwaters have dissolved Au concentrations ≤0.01 μg l−1, well within background for the other regions. The few Au-bearing groundwaters in the Northern region are from within pits and appear to represent localized thiosulphate dissolution (see above). The Kalgoorlie region has the highest mean dissolved Au content (⇑Table 1), consistent with the observed optimal conditions for dissolution (previous section), with the Central region having moderate levels of dissolved Au.
Distribution of dissolved Au concentrations for the Northern, Central and Kalgoorlie groundwater regions.
Extensive supergene mobilization of Au has occurred, and is occurring, in the Central and Kalgoorlie regions. Oxidizing conditions in the upper parts of the groundwater lead to extensive dissolution of Au, which may diffuse though the water mass. Where dissolved Au–halide interacts with dissolved Fe derived from ongoing weathering, Au is reduced in a redox front reaction (Eqn 5), possibly resulting in horizontal and sub-horizontal supergene anomalies (⇓Gray et al. 1992). The lower dissolved Au contents of Central groundwaters, which are postulated to be due to the less acidic groundwater conditions, suggests this process may take longer in this area, although extensive redistribution would still be expected over the long period of weathering.
IMPLICATIONS FOR EXPLORATION
Lithological discrimination and influence of fault zones
Hydrogeochemistry is potentially useful for lithological discrimination, as illustrated at Panglo (⇓Fig. 15). These groundwaters were characterized into those in contact with shales, mafic and ultramafic lithologies by comparing sample locations with known geology (⇓Gray 1990a). These different groundwater groups are clearly delineated by plotting each water, using an ultramafic (Ni+2Cr) v. a mafic (Mn+13Co+9Zn+36Cu) index. This discrimination is effective even for waters in contact with highly weathered rocks. Hydrogeochemistry may also indicate shear zones; two samples at the top right of the plot appear to be from shear zones which, at Panglo and elsewhere, are sources of dissolved base metals.
Mafic v. ultramafic indices for groundwaters at Panglo, Kalgoorlie region.
Hydrogeochemical discrimination of lithology will presumably be most effective for a consistent sample medium (in terms of acidity and, possibly, salinity), so that different sites may well have to be treated separately. The concentrations of base metals in acidic, shallow Kalgoorlie groundwaters will be significantly higher than those in Central, Northern or deeper Kalgoorlie groundwaters. However, dissolved Cr contents can be used to recognize ultramafic rocks across all regions. Groundwaters in contact with fresh and weathered ultramafic rocks contain consistently high (0.01–0.43 mg l−1) dissolved Cr concentrations, whereas waters in contact with other lithologies have Cr concentrations below detection. This effect is highly robust and offers a straight-forward method for recognizing the presence of ultramafic rocks.
Fault zones might well be expected to influence groundwaters characteristics. Weathering is generally deeper along faults and shears, which may have distinct chemistries and are commonly zones of enhanced water flow. At the Wollubar channel (⇓Gray 1993b), groundwater close to the Boulder–Lefroy shear has a much higher Al concentration than expected at that particular pH, and also has raised concentrations of Si, Fe, Sc, Y, REE, Pb and U. This may reflect the influence of highly active weathering in the shear zone on the chemistry of the groundwater in the palaeochannel. As mentioned previously, Panglo has a similarly strong base metal enrichment of groundwaters (⇑Fig. 15) that may reflect the proximity of a fault zone.
Presence of sulphides
Groundwaters in contact with sulphide-rich rocks buffered by alkaline minerals (e.g. carbonates) may be neutral to alkaline at depth, becoming acidic and more oxidizing closer to the water-table, particularly in the Kalgoorlie region. The deep, neutral waters close to sulphides are characterized by:
near neutral pH and relatively low Eh;
enrichments in Fe, SO4 (possibly from sulphides), HCO3, Mg and, to a lesser extent, Ca and Sr (possibly from carbonates). This association probably reflects oxidation of sulphides at depth, with neutral pH conditions being maintained by carbonate buffering;
enrichment in several anionic chalcophile pathfinder elements, including As, Sb, Mo, W, Bi, Ag and I.
Acid-oxidized waters lack most of these signatures, because most of the anionic chalcophile elements are precipitated in acid conditions (with the possible exception of Mo, as indicated above). In these groundwaters, Fe is precipitated, HCO3 removed, and in saline groundwaters SO4, Mg, Ca and Sr enrichments may be obscured. Thus, with the exception of dissolved Au itself, few elements give useful Au exploration data in acid groundwaters.
Presence of Au mineralization
The clearest indication of Au mineralization in Kalgoorlie and Central groundwaters is given by dissolved Au. It is not clear whether other elements, such as As or Sb, specifically indicate Au mineralization or merely the presence of sulphides. Interpretation of dissolved Au concentrations is complicated by there being two mechanisms for transport of Au in groundwater – thiosulphate and halide complexing. Where Au appears to be dissolving as a thiosulphate complex, as at Boags and the Hornet pit at Mt. Gibson, the distribution of dissolved Au closely matches that of mineralization, but this effect is highly localized, possibly anthropogenic, and would probably be missed in an exploration sampling program.
In oxidizing environments, Au dissolves to form chloride or iodide complexes and, where this mechanism is expected to be active, high concentrations of dissolved Au are observed. However, Au concentration is strongly affected by factors not directly related to mineralization, such as Eh and dissolved Fe content, and the distribution of dissolved Au only approximately matches that of primary mineralization. Nevertheless, this technique could still provide useful extra information. A few other elements, such as As, Sb, Mo, I and various base metals may also have value as pathfinders, although they commonly have low concentrations in acid conditions (e.g. ⇑Fig. 12) and analyses for them in saline groundwaters may be difficult, expensive or of poor sensitivity. In the Kalgoorlie region, a limited suite of parameters, namely salinity, pH, Eh, dissolved Au, Fe, Cr and, possibly, other base metals, could be analysed cheaply (using standard probes, sorption onto carbon for Au, and ICP-AES and/or colorimetric analyses for Fe and Cr) but extending this to the other elements of interest will add considerably to cost, with little added exploration benefit. A threshold dissolved Au concentration of approximately 0.05 μg l−1 would appear to locate most mineralized areas. For the Central areas, a slightly lower dissolved Au threshold (0.02 μg l−1) appears appropriate (⇓Fig. 16), and a number of chalcophile elements (e.g. As, Sb, Mo, W, Tl, Bi) may also give valuable exploration data. However, further work is required to study the relative degrees of groundwater dispersion of these elements.
Dissolved Au distribution at Golden Delicious, Central Yilgarn Craton, superimposed on regolith Au contours (from ⇓Bristow et al. 1996).
In the Northern region, dissolved Au contents are much lower, although higher values may still correlate with mineralization. However, even though Au and indicator elements occur at low concentrations, the low salinity means that multi-element ICP-MS analyses are both cheap and highly sensitive. At the Harmony (Baxter) Au deposit (see ⇓Robertson 2001), a number of indicator elements either correlate closely with the position of buried mineralization (Rb, Sc, W, Mo; ⇓Fig. 17) or have use for lithological discrimination (Cr, Ni, As) (⇓Gray 1995). In addition, the low variation in salinity, Eh and pH with depth indicates that sample depth is less critical than at other sites.
Dissolved Mo distribution (dots) at the Harmony (Baxter) Au deposit, superimposed on contoured maximum Au content in the regolith.
Groundwaters seem to have a limited ability to ‘see through’ barren transported overburden, as opposed to its success when sampling groundwater even from supposedly depleted in situ material (e.g. Panglo; ⇓Gray 1990a). In the highly active Kalgoorlie groundwaters, those sampled from the Wollubar palaeochannel appear to show significant influences from underlying rocks (⇓Gray 1993b), but results are poor for the least active groundwaters in the Northern region such as at Lawlers (⇓Gray 1994) or Mt Joel (⇓Porto et al. 1999). However, the degree of masking by the overburden will depend on the element used. Chemically conservative species, such as SO4, appear to give good signals (e.g. ⇑Fig. 11), but more chemically active elements, such as Mo or Au, give a poor response in groundwaters from barren overburden. It is recommended that, if possible, groundwaters be sampled below the transported overburden for optimum results.
CONCLUSIONS
Using regional variations of salinity, acidity and oxidation potential, the eastern Yilgarn Craton has been divided into four groundwater regions:
Northern (N Yilgarn and margins) – Fresh and neutral, trending more saline in the valley floors.
Central – Neutral and brackish (commonly <1% TDS) to saline (about 3% TDS), trending to hypersaline (10–30% TDS) at the salt lakes, with common increases in salinity with depth.
Kalgoorlie – Commonly acid (pH 3–5), except where buffered by extremely alkaline materials (e.g. ultramafic rocks), and saline within the top part of the groundwater mass, trending to more neutral (pH 5–7) and hypersaline at depth and within a few kilometres of salt lakes.
Eastern (E Yilgarn and Officer Basin) – Saline to hypersaline, neutral to acid and reducing. Dissolved concentration of many ions is low, due to the presence of lignites in the channel sediments.
Although there may be important local effects, these regional variations are major controls on the concentrations of many elements. Aluminium, Li, Y, REE and U have higher concentrations in acid groundwaters, such as those in the Kalgoorlie region. Dissolved concentrations of the base metals (Mn, Co, Ni, Cu and Zn) are less closely correlated with acidity, and these elements do show scope for lithological discrimination, but there is no apparent relationship with Au mineralization. Dissolved Cr is strongly correlated with ultramafic rocks, and appears to be dominantly present as chromate (i.e. Cr6+ as CrO42−). Concentrations of As, Sb Mo, W and Bi are low in acid groundwaters, but are higher above pH 6.5, particularly in the Central region. Therefore acid groundwaters (particularly in the Kalgoorlie district) will be poor media for the use of these elements as exploration pathfinders. Molybdenum differs from the other elements in this group in having significant concentrations in acid groundwaters, although lower than in neutral and alkaline groundwaters.
Dissolved Au is commonly the best pathfinder for Au mineralization. It occurs dominantly as halides (chloride and/or iodide) and has enhanced concentrations (to >1 ppb) under the acid/saline/oxidizing conditions common in the Kalgoorlie region, whereas concentrations in the northern Yilgarn are two orders of magnitude less. This implies that supergene Au remobilization should be considerably less in the northern Yilgarn than in the Kalgoorlie region. Additionally, the threshold dissolved Au concentration as used for Au exploration will differ significantly in different regions of the Yilgarn Craton.
It is recommended that use of groundwater for exploration in Kalgoorlie and Central regions is best restricted to shallow samples, with depth being less critical in the Northern region. Waters should be in contact with, or within a few metres of, in situ material. In the Kalgoorlie region, it may well be more cost-effective to restrict analyses of the saline waters to a select group of parameters, which should at least include salinity, pH, Eh, Au, Fe and Cr. A more expanded analytical suite, including As, Sb, Mo, W and Bi, may well be useful, though at significant cost, for Central groundwaters. In comparison, for the fresh Northern groundwaters, multi-element analyses are cheaper and several elements have been shown to have exploration potential.
The results and hypotheses presented above may have significance for the use of hydrogeochemistry for exploration in the Yilgarn Craton, and indeed, elsewhere in the world. Orientation studies should always be done, but these general Yilgarn-wide observations may be useful in setting general criteria for initial investigations.
Acknowledgments
Much of this research was the result of collaboration between CRC LEME and the Australian minerals industry through AMIRA, and the companies involved are thanked for their ongoing financial and logistic support and scientific collaboration. CSIRO staff who have assisted in the analysis of waters included G. D. Longman, A. Howe, J. E. Wildman, J. K. Hosking, T. T. Pham and B. van Herk at Floreat Park and L. Dotter at North Ryde, and carbons were analysed by Becquerel Laboratories at Lucas Heights. Finally, C. R. M. Butt has given extensive scientific guidance, and advice in the preparation of this paper. CRC LEME is supported by the Australian Cooperative Research Centres Program.
- © 2001 AAG/The Geological Society of London
