Major advances in exploration geochemistry, 1998–2007

D.R. Cohen1,*, D.L. Kelley2, R. Anand3 & W.B. Coker4

1
School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW, 2052, Australia
2
Zinifex Ltd, 14809 Foothill Rd, Golden, CO, 80401, USA
3
Cooperative Research Centre for Landscape Environments and Mineral Exploration, CSIRO Exploration and Mining,
26 Dick Perry Avenue, Kensington, WA, 6151, Australia
4
BHP Billiton World Exploration Inc., Suite 800, Four Bentall Centre, 1055 Dunsmuir St., Vancouver,
British Columbia, V7X 1L2, Canada
*Corresponding author (e-mail: [email protected])

ABSTRACT: Exploration geochemistry continues to be a major contributor to

mineral exploration programs at all scales. The decade has spanned a complete
economic cycle in the minerals industry, during which there has been a change in
exploration emphasis from the dominance of Au and diamonds to the addition of a
range of base metals and the recent revival of U exploration. Increasing metal prices
and the development of new geochemical techniques have allowed exploration
activity to expand in previously under-explored terrains, including those character-
ized by thick, transported regolith cover. There has been significant evolution and
refinement of geochemical dispersion models for transported regolith, especially
processes associated with oxidizing sulphide mineralization. These models have
prompted further developments in sampling methods, analytical techniques and
approaches to anomaly detection. The routine measurement of fundamental soil and
groundwater properties, such as pH, has been demonstrated to assist in detecting
geochemical and mineralogical effects of buried mineralization on regolith. New
lithogeochemical techniques have been developed for terrane selection and early
discrimination of potentially mineralized units. The production of regolith-landform
maps, including 3-D models, and more detailed techniques for differentiating the
origin and nature of regolith materials now underpins the design of exploration
programs in a number of terrains. Various exploration techniques, such as
biogeochemistry and soil gas, have been revitalized by the application of new
technologies. Ongoing improvements to detection limits and the reliability of
multi-element analytical methods have permitted development and routine applica-
tion of a wide range of selective geochemical extraction methods aimed at detecting
subtle geochemical haloes developed in transported regolith cover. The results from
a plethora of selective extraction orientation studies have, however, indicated the
reliability of such methods to be uncertain, especially in arid terrains. Low-density
geochemical mapping of large regions has been undertaken by government agencies
for a variety of resource, environmental and health purposes, including attracting
mineral exploration. Such maps typically demonstrate the dominance of underlying
geology on regional geochemical patterns in regolith, even in industrialized regions.
Philosophical and numerical approaches to defining and detecting geochemical
anomalies continue to be reviewed. There is a need to supplement classical
parametric outlier-detection approaches, routinely applied to geochemical data, with
statistical techniques that reflect a recognition that subtle, multivariate patterns
related to the geochemical effects of mineralization may emanate from within the
main data cloud and not outlying or distinct clusters.

KEYWORDS: Geochemical exploration, minerals, metals, review, advances

Exploration geochemistry continues to be a major contributor conference in 1997 the minerals industry has completed a full
to mineral exploration programs at scales ranging from regional economic cycle and is benefiting from a surge in the price of
reconnaissance down to mining leases. Since the last decennial most mineral commodities. The depressed markets of the late
Geochemistry: Exploration, Environment Analysis, Vol. 10 2010, pp. 3–16 1467-7873/10/$15.00  2010 AAG/Geological Society of London
DOI 10.1144/1467-7873/09-215
4 D. R. Cohen et al.

1990s resulted in companies focusing exploration on locating 2. Evolution and geochemical characteristics of regolith and
more resources in the vicinity of their existing mining opera- the development of secondary dispersion models upon
tions, including the re-evaluation of old datasets. The current which mineral deposit targeting methods can be developed
upswing has resulted in a return to greenfields exploration, or refined. An example of this work is drawn from the arid
including terranes with high mineral potential but which have terrains in Australia.
been previously neglected or under-explored.
The decade has provided various challenges for exploration
geochemical research and applications. A technical challenge Lithogeochemical discrimination of barren versus fertile
has been to develop methods capable of reliably detecting intrusions
mineralization covered by deep transported regolith, using Porphyry Cu deposits can form from melts that undergo
the geochemical and mineralogical characteristics of surface repeated cycles of replenishment by primitive magma and
samples. New models for geochemical dispersion from deeply partial crystallization (Rohrlach & Loucks 2005). With each
buried mineralization have been proposed and new sampling cycle, incompatible components, such as H2O, SO3 and Cl2,
and analytical techniques tested in a large number of orientation become progressively enriched in the residual melt, resulting in
surveys, but with varying degrees of success. Another challenge conditions that are optimal for Cu mineralization. In the case of
has been the adjustment to changing commodity emphases, the Tampakan porphyry Cu–Au deposit in the Philippines,
with the decade spanning the tail-end of the great Au epoch, cogenetic silicic andesite has been shown to have increased
the diamond rush and the base metal revival to a reawakening water content in the melt (3–7.5 wt % H2O) over a 5-Ma
of U exploration, at a time when substantial exploration period. This increased water content suppressed plagioclase
expertise has been lost from the industry (few practicing crystallization relative to hornblende, resulting in a significant
geochemists have had much experience with U exploration). increase in the ratio of the (plagioclase-hosted) Sr to the
There has been some convergence between the disciplines (hornblende-hosted) Y in fertile melts. Systematic whole-rock
of exploration geochemistry and environmental geochemistry, analysis of igneous rocks can, therefore, potentially discriminate
with the principal result being a number of multi-purpose fertile hydrous magmatic events. Application of this method to
geochemical baseline mapping programs, at regional to conti- volcanic rocks derived from hydrous melts permits the regional
nental scales, being undertaken mainly by government agencies. assessment of porphyry copper prospectivity in arc terranes.
One fundamental geochemical question has survived six Extensive reference suites of whole rock data have been
decades of debate: What constitutes a geochemical anomaly and compiled to compare ‘barren’ terranes with ‘productive’ ter-
how can this be enhanced (by suitable combinations of ranes using Sr/Y versus SiO2 (Fig. 1). Although there is some
sampling processing and analytical methods) and detected (by overlap between the barren and productive suite, the produc-
use of various univariate and multivariate mathematical tech- tive suite tends to have elevated Sr/Y over a wide range of rock
niques)? This question has been revived again by the arrival of compositions.
a variety of selective geochemical extraction methods linked to
significant reductions in analytical detection limits. These
methods have further revealed the complexity within geochemi- TerraneChron
cal landscapes, and the challenge of separating subtle geochemi- Zircons have long been used as a geochronometer, but their
cal signals related to the effects of mineralization from the complex crystal structure also accommodates process-sensitive
‘noise’ generated by the many processes that contribute to trace elements useful for understanding crustal development
the overall geochemical and mineralogical characteristics of (Harley & Kelly 2007). Because they are physically and chemi-
regolith. cally stable over long periods of time, they are preserved in
This paper is intended to highlight some of the advances in regolith. The TerraneChron method integrates U–Pb dating,
exploration geochemistry over the last decade and point to Hf isotopic analysis and trace element analysis of zircons
developments for the next, rather than providing a comprehen- obtained from drainage samples to characterize their source
sive overview. A few of the key advances are expanded upon in rocks (O’Reilly et al. 2004). Where applied over a large region,
other papers in this issue. As indicated by Closs (1997) in the method can provide an overview of the geologic history of
the previous decade, the ultimate objective of geochemical a region and crustal evolution. The method is ideally suited for
exploration is to assist the efficient discovery of ore deposits. regional work where the geology is poorly known or correlating
terranes separated by later tectonic events.
Isolation of different populations of zircons, based on size,
colour and chemical composition, allows different events to be
TERRANE SELECTION AND TARGETING characterized, including the age of formation, modelled ages of
Geochemical terrane modelling has proceeded along two lines: source rocks, the composition of parent rock types and the
source region for these rocks. The compilation of this infor-
1. Lithogeochemical characterization of environments condu- mation for a large number of zircon populations collected over
cive to the formation of mineral deposits and application of a broad region is presented on an ‘event signature diagram’,
these methods on a regional scale in the early stages of which allows for interpretation of the crustal evolution for that
exploration programs, to predict and subsequently select region (Fig. 2). The diagram shows the distribution of mag-
regions with potential for discoveries. Terrane targeting matic crystallization ages for the zircon populations versus the
methods are not primarily used to ‘detect’ specific mineral mean crustal residence age of the source. When pre-existing
deposits, but to measure geochemical features consistent crust is being reworked (i.e. no new material is added), the
with processes that can lead to the formation of ore results follow the evolutionary paths for crust at a given age
deposits. In many cases, the geology of these terranes may (toward the lower left in Fig. 2). When juvenile material is
be poorly understood and reliable geological maps absent. added, as might occur during significant compressional or
Two examples discussed here are whole-rock lithogeochemi- extensional events, the results trend to the upper left, normal to
cal discrimination in volcanic arc terranes and the detrital the evolutionary paths. Hafnium isotopic results and rock
zircon TerraneChron method; and compositions deduced from trace element data provide further
 Major advances in exploration geochemistry 5

Characterizing well-endowed terranes with the TerraneChron

method may allow prediction of new terranes with potential for
undiscovered resources.

REGOLITH CHARACTERIZATION AND

EVOLUTION AND REGOLITH MAPPING
Regolith-dominated terrains pose considerable difficulties for
exploration. Over the last decade such terrains have received
considerable exploration attention as the potential for finding
sizeable new mineral deposits becomes greater than in the more
explored outcropping areas. This is largely due to a better
understanding of all aspects of the regolith (its distribution,
characteristics, processes and evolution) that are essential for
effective geochemical and geophysical exploration program
design. Much of this work is summarized by Butt et al. (2000),
Anand & Paine (2002) and Anand & de Broekert (2005).

Regolith-landform mapping
The importance of understanding the position of regolith
materials within regolith stratigraphy and the landscape, for
geochemical exploration, has become increasingly apparent in
recent years. Regolith-landform mapping has gained greater
importance and wider usage within the mineral industry over
the last decade. Regolith features are an integrated expression
of geology, climate, landforms, geomorphic processes and
landscape evolution, and have close generic relationships with
landforms. Thus, landforms become a surrogate for regolith
and the basis of mapping. A crucial component of this mapping
work has been reconciliation of observed regolith-landform
relationships with carefully selected field traverses in each
Fig. 1. Discriminant plot of Sr/Y versus SiO2 showing rock suites mapping unit (Pain et al. 2001; Anand & Paine 2002).
associated with Cu–Au ore deposits (symbols) compared with a Regolith-landform maps are not intended as end-products in
barren reference rock suite (shaded area) defined by hundreds of themselves, but to provide two key pieces of information for
samples from NW Pacific volcanic arcs, including intraoceanic and mineral explorers: whether near-surface material is residual or
continental margin arcs. All analyses are from fresh or weakly altered transported and the expected thickness of transported material.
(weakly propylitic at most) whole-rock samples. Modified from
Rohrlach & Loucks (2005). Both factual and interpretative or derivative regolith-landform
maps have been produced for a number of terrains. Derivative
maps are assisting in identifying probable dispersion models
and subsequent design of geochemical surveys, including media
selection, sampling procedures and intervals. Appropriate
thresholds can then be set, anticipating and allowing for
variations in geochemical responses due to weathering or
masking by transported cover. The likely effects of chemical
and physical dispersion, resulting in displaced or transported
anomalies, can be incorporated into data interpretation.

The third dimension

Three-dimensional regolith models are being built around large
drill-hole datasets. For example, the Yandal project in the
Yilgarn Craton, included a 3-D representation of residual and
transported regolith over a major gold system in an Archaean
greenstone belt using a database of over 70 000 drill-holes and
high-resolution aeromagnetics (Anand 2000) and spectacularly
Fig. 2. Example of an event signature diagram for detrital zircons identified the maghemite-defined palaeodrainages. These 3-D
from the Mt Isa (Australia) and southern Norway terranes (from models can define the base of the transported regolith, weath-
O’Reilly et al. 2004). ering fronts, internal regolith structure and have genuine
predictive value. Although difficult to construct, due to incon-
insight into the tectonic events (e.g. rifting versus subduction) sistencies in regolith logging and problems defining regolith
that have occurred. In the example, results for the Mt. Isa unit interfaces, there have been advances in geophysical
Eastern Succession indicate repeated extension and magmatism methods for regolith mapping methods that require minimal
superimposed on an ancient crust. In contrast, results for drill-hole calibration, including the low-frequency, broad-
southern Norway indicate continental build-up of new crust at bandwidth, time-domain TEMPEST airborne EM system
a continental margin with no indication of ancient crust. (Lane et al. 2000). With the advent of higher resolution
6 D. R. Cohen et al.

aeromagnetic data and the capacity to differentiate regolith

features (e.g. Fe accumulation in palaeochannels) from bedrock
magnetic features, the interaction of drainage and geology
becomes clearer. A significant observation from these trans-
ported regolith models is that, despite the subdued landscapes
across much of Australia, the palaeotopography below the
interface is quite rugged. It is this elusive form of surface that
will continue to host the most reliable direct geochemical
indicators of buried mineral systems.

Landscape evolution
An understanding of the distribution and origin of regolith
materials and palaeotopography have led to the refinement and
development of new landscape evolution models in Australia
and Africa (Bolster 1999; Bamba et al. 2002; Anand & Paine
2002; Anand & de Broekert 2005). These models have affected
the interpretation of geochemical dispersion data and
anomalies. Weathering profiles are recognized to have formed
in both residual materials and also commonly in transported
overburden of various ages (Fig. 3). There have been several
distinct phases of ferruginization, deferruginization, kaoliniza-
tion, silicification and calcification. The polyphase evolution of
regolith has long been known but the advent of various
geochronological techniques now permits quantification. The
basic geochronological framework of the Australian regolith
can be determined by palaeomagnetic dating using secondary
hematite formed in saprolite by chemical weathering. Palaeo-
magnetic dates for regolith from 30 sites throughout Australia
indicate two principal weathering events at c. 10 Ma and
50–60 Ma, with some evidence of Mesozoic or even earlier
weathering in the eastern Yilgarn and Cobar areas (Pillans
2005). It is likely that each weathering event generates a new
metal dispersion event, hence dating such events may contrib-
ute to selection of more favourable geochemical sampling
media, either bulk samples or mineral phases.
It is now clear that ferruginous pisoliths and duricrust have
formed in a variety of materials including residuum, colluvium
and alluvium. In places, ferruginous duricrusts occupy topo-
graphically higher areas due to relief inversion. Thus, a distinc-
tion between residual and transported types has been shown to
be essential for geochemical sampling. Lateritic residuum forms
largely by residual processes, hence elevated concentrations of
Au and other elements in this material are generally the result
of underlying deposits. Ferricretes are Fe-cemented sediments Fig. 3. Generalized regolith profile for the Yilgarn Craton, showing
and have been found to be generally unsuitable as sample a three-layered strato-regolith architecture, each with different
media on a local scale but may contain indications of distal degrees of ferruginization, probably relating to episodic weathering
mineralization, either within their detrital component or in the events.
hydromorphically-derived goethite cement. Elements may be
concentrated in ferricrete following lateral transport from lacustrine sediments. Some of these regions have suffered
sources that may be hundreds of metres to several kilometres under-exploration relative to their indicative mineral potential.
away. Criteria have, therefore, been established to distinguish
lateritic residuum from ferricrete (Anand & Paine 2002).
Dispersion models
Numerous case studies have documented the development of
EXPLORATION UNDER DEEP COVER secondary geochemical dispersion haloes and mineralogical
changes within transported regolith above deeply buried min-
Regions with thick, transported regolith or rock cover present eral deposits in various regolith setting (Mann et al. 1998;
special problems for exploration geochemistry approaches that Lawrence 1999; Wang et al. 1999; Smee 1998; Kelley et al. 2003;
use surface samples. Examples of such terrains include the Cameron et al. 2004; Butt et al. 2005). There is much debate on
northern parts of the Tasmanides in eastern Australia which are the mechanisms by which such dispersion proceeds; these have
variably covered by thick alluvium and on-lapping Mesozoic included combinations of advection, groundwater flow, capil-
to Tertiary sedimentary sequences, the Yilgarn Craton of lary action, diffusion of volatile compounds and biological
Western Australia, the southwestern USA and the Andes where action. Two mechanisms that have received significant atten-
thick gravel deposits fill the basins, and the recently glaciated tion over the last decade are electrochemical dispersion and
terrains of the northern hemisphere with tills and glacio- cyclical dilatancy pumping.
 Major advances in exploration geochemistry 7

Fig. 4. Redox model from Hamilton

(1998, 2000), modified after Cameron
et al. (2004) showing development of a
reduced column over sulphides in
transported regolith. Metals mobile
under reducing conditions migrate
vertically along redox gradients. Above
the water table metals are oxidized,
creating a zone of low pH and
carbonate loss. Carbonate precipitates at
the edge of the reduced column. Metals
such as Zn and Cd accumulate near
surface.

Electrochemical dispersion highs may be a previously-undocumented form of electro-

Self-potential currents, associated with oxidizing sulphide osmosis, a process where fluid flow is induced by an electrical
deposits, are the basis for one class of dispersion model in field (Hamilton et al. 2003). In addition, selective extractions
which the upward movement of electrons in an electrically and Pb isotope studies carried out in the VMS example (Hall
conductive sulphide body results from the electrochemical et al. 2004) suggest that Pb was transported into the exotic
gradients formed between (underlying) reducing and (over- overburden from the sulphides below.
lying) oxidizing environments (Sato & Mooney 1960; Govett
1973; Bølviken & Logn 1975; Smee 1983). In all cases, the mass
transfer of ionic species facilitates electrical current and the Cyclical dilatancy pumping
maintenance of charge balance, creating various types of Movement of metal-bearing groundwater surrounding mineral
geochemical anomalies in the overburden (Antropova et al. deposits is capable of redistributing metals both laterally and
1992; Goldberg 1998). A revised model (Fig. 4) provides for vertically. In areas with moderate to high relief and a shallow
upward propagation of reduced species to the water table, groundwater table, lateral, gravity-induced groundwater flow
forming a reduced column over the mineralized zone and the causes seeps to form at the base of slope where the ground-
development of geochemical anomalies in the overlying surfi- water table intersects the surface. Seeps produce hydromorphic
cial overburden (Hamilton 1998, 2000). The oxidation of Fe2+ accumulations of metals (mainly hydroxides), reflecting trans-
and its precipitation in ferric oxyhydroxides causes acid pro- port in reduced groundwater and subsequent atmospheric
duction and subsequent dissolution of carbonate which precipi- oxidation at the seep. Vertical migration of groundwater is
tates at the edge of the reduced column where the pH is higher. more difficult and requires an external energy source to propel
Theoretical ion migration rates in electrochemical fields, linked water through the vadose zone. The expulsion of basinal fluids
to redox gradient effects, are much faster than diffusion rates from sedimentary sequences undergoing active tectonic com-
and may permit the formation of geochemical anomalies at pression is well documented (Unruh et al. 1992). Saline peren-
surface in young (c. 8 000 years), thick, glacial sediments over- nial springs in the eastern margin of the northern Coast ranges,
lying mineral deposits (Hamilton 1998; Hall et al. 2004). Given California, are caused by fluid overpressure zones in an actively
that H+ is the most mobile and one of the most reactive deforming anticline in Upper Cretaceous marine shale and
aqueous species, sustaining its presence over the reduced sandstone (Davisson et al. 1994). The discharge is a local
column in a high-rainfall, high-carbonate environment requires phenomenon, with fluid migrating along fluid-pressure gradi-
a continuous process (Hamilton et al. 2004b). ents and expelling along fractures and faults. Geochemical and
Several detailed investigations of surficial geochemical geothermometry evidence indicates that the water expelled at
responses over deeply buried sulphide-hosted gold and VMS an elevation of 230 to 380 m above sea level is dilute seawater
deposits in northern Ontario have been carried out by Cameron that originated from depths of up to 4 km. The surface
et al. (2004), Hall et al. (2004) and Hamilton et al. (2004a, b). expulsion of groundwater from seismic activity is also well
Each of these studies characterized the pH, oxidation–reduction documented (Tchalenko 1973; Nur 1974; Sibson 1981; Muir-
potential (ORP) and self potential (SP) in overburden and Wood 1994). In a process termed cyclical dilatancy pumping
confirmed the presence of a reduced column above zones of (Sibson 2001), groundwater held in fractures is expelled during
buried mineralization, as well as local zones of acidic ground- and after earthquake events. In extensional regimes, ground-
water, metal deposition, CO2 enrichment, O2 depletion, car- water is held in open structures before the earthquake and is
bonate remobilization, groundwater temperature anomalies and expelled during post-seismic compressional elastic rebound. In
piezometric highs at the edge of the reduced column (Hamilton contrast, compressional regimes cause structures that are closed
et al. 2003, 2004a, b). These features all support the redox- during pre-seismic time to open after the earthquake, increasing
gradient model of Hamilton (1998, 2000). The piezometric porosity and allowing water recharge. Hence, deep groundwater
8 D. R. Cohen et al.

Fig. 5. Conceptual model showing the

generation of soil anomalies over the
buried Spence porphyry Cu deposit,
northern Chile. Vertical arrows show
the flow path of deep saline waters
during earthquake events along a
fracture zone. Horizontal arrows show
groundwater flow direction. Anomalous
Cu in soil is interpreted to reflect
injection of Cu-bearing saline
groundwater into overburden following
seismic events. Modified from Cameron
et al. (2004).

surrounding mineral deposits has the potential to disperse Alternate sampling media
metals away from the deposit during seismic events. In areas of transported regolith, a number of alternate sampling
Cameron et al. (2002) proposed that cyclical dilatancy media to regolith are being evaluated. These include ground-
pumping could explain the development of surface geochemical water (Pirlo & Giblin 2004; Giblin & Mazzuchelli 1997; Kirste
anomalies in areas with a thick vadose zone. Results of soil and et al. 2004), soil gas (Highsmith 2004; Polito et al. 2002;
groundwater analyses from an integrated study at the Spence Rutherford et al. 2005) and vegetation (Cohen et al. 1998;
porphyry copper deposit in northern Chile, which is buried Lintern 2005; Dunn 2007; Anand et al. 2007). As with many
beneath 50 to 100 m of Miocene gravels, are consistent with the of the new techniques being evaluated, interpretation of geo-
vertical movement of saline metal-rich groundwater along chemical patterns in such media has tended towards the
fractures during earthquakes in this seismically active area empirical rather than the mechanistic.
(Fig. 5; Cameron et al. 2002; Cameron & Leybourne 2005).
Copper in groundwater is restricted to the mineralized area due
to the tendency of Cu2+ released by oxidation of sulphides to REGIONAL GEOCHEMICAL MAPPING
adsorb to Fe hydroxide colloidal particles and coatings, whereas
elements that dissolve as anions (e.g. As, Mo, Se and Re) are Regional and global baselines
dispersed widely. Multi-element geochemical surveys, principally involving
Field measurement of the conductivity of soil-water slurries stream sediments or soils, continue to be a cornerstone of
showed two zones of salt (NaCl) enrichment, one directly over regional mineral exploration. The development of large-scale,
the deposit and the other 1 km away. Trenching of the soils in regional reconnaissance geochemical exploration surveys
these zones revealed vertical fractures in the gravels, whereas in the west originated with the NURE program in the USA
trenching in a background area showed no fractures. The (Gustavsson et al. 2001), the NGR program in Canada (Coker
fracture zones appear to have formed by reactivation of et al. 1987) and the Wolfson Geochemical Atlas in the UK
basement faults. Elements present in the soils above the (McGrath & Loveland 1992) and has continued into recent
fracture zones are the same as those enriched in groundwater times with the geochemical mapping of China (Xie & Cheng
near the deposit and indicate redistribution of elements by 1997; Li et al. 1999; Xujing et al. 2004) and other countries or
groundwater movement to the surface during earthquakes, provinces. These surveys were mainly designed to support
followed by evaporation and further redistribution by rain. mineral exploration.
Both fracture zones contain NaCl in addition to anomalous In post-mining and industrialized regions, such as Europe,
anions, whereas Cu is restricted to the fracture zone above the regional geochemical surveys are now being undertaken for
deposit. Additional evidence for the Cu anomalies in soil being environmental monitoring, general resource planning and
related to cyclical dilatancy pumping comes from the presence detection of sources of contamination (Volden et al. 1997; Plant
of atacamite, a secondary mineral in the oxide zone of many et al. 2000, 2003). Following IGCP projects 259 and 360
deposits in northern Chile (Cameron & Leybourne 2006). (Global Geochemical Baseline; Darnley et al. 1995), a number
Atacamite is soluble in low-salinity water and would have been of low density geochemical surveys have been completed in
removed from the supergene profile had it formed as a primary Europe. Part of this was prompted by a lack of information
oxidation product during supergene weathering. The compos- upon which to evaluate the impact of fallout from the 1986
ition of saline groundwater at Spence falls within the stability Chernobyl nuclear disaster in Ukraine. The most recent release
fields of the atacamite–brochantite assemblage and suggests has been the FOREGS Geochemical Atlas of Europe
that atacamite is the replacement of original copper oxide min- (Salminen et al. 2005) in which c. 800 sites were sampled for
erals by saline basinal fluids being pumped vertically through soils, floodplain sediments, stream sediments and waters, yield-
the deposit. ing a density of c. 1 site per 5 000 km2. Despite 250 years of
 Major advances in exploration geochemistry 9

Fig. 6. Comparison of aqua regia

extractable Ni in topsoil (0–25 cm)
with subsoil (50–75 cm) from the
Geochemical Atlas of Europe
(Salminen et al. 2005).

industrialization (including smelting operations) and high varia- such as Australia, which will form the basis for more
bility in the nature of the regolith (from residual regolith to rigorous sampling design and data interpretation in regional
glacial sediments), the dominant control on most elements in reconnaissance surveys (Brown & Hill 2004).
the topsoil and subsoils is the underlying geology, with the
example of Ni shown in Figure 6. This must have an impact on
ADVANCES IN ANALYTICAL TECHNIQUES
the definition of ‘contamination’ and in the setting of soil
Multi-element analysis and detection limits
clean-up targets, given some regulatory authorities have set
environmental action levels for trace elements in soils at The wave of developments in selective geochemical extractions
industrial sites that are lower than natural concentrations and the implementation of large-scale geochemical surveys
(Salminen et al. 1998). using these methods has largely ridden on the back of devel-
Apart from high quality data, such surveys can provide a opments in analytical methods, including the progressive low-
valuable sample archive for future reference. The value of this ering of routinely achievable detection limits, reduction in real
is demonstrated by the recent production of a the National costs and improvements in the reliability of analytical devices
Geochemical Survey of the USA based on the multi-element such as ICP-MS (inductively coupled plasma-mass spec-
reanalysis of c. 35 000 stream sediment samples collected in the trometry). Some new exploration techniques, such as the use of
1960s under the NURE program (USGS 2004). ultra-trace stable isotopes, have also been made possible by the
availability of high resolution analytical devices such as multi-
Role of governments collector HR (high resolution)-ICP-MS. For many elements,
routine detection limits for geochemical samples have
The economic advantages gained by government investment in improved by half to one order of magnitude (Fig. 7), with more
provision of pre-competitive exploration data, including geo- elements being added to routine commercial multi-element
chemical maps, are agreed to by industry and geological surveys packages (Caughlin 2010). Most exploration geochemical data-
(Blight 2002; Macdonald 2002). Whereas the value of such sets now contain determinations on >40 elements, prompting
databases has been demonstrated by the strong response to further consideration of the support to target definition pro-
initiatives such as Discovery 2000 (Geological Survey of New vided by previously difficult-to-determine elements such as Re
South Wales), SAEI (Primary Industries and Resources of and Te.
South Australia) and regolith mapping programs in Western
Australia (Cornelius et al. 2006), similar initiatives commonly
founder when funding requests are considered by state or Selective extractions
federal treasuries (Bowler 2002). Provision of high quality, In response to new models that account for the development of
publicly-funded, geoscience databases reduce both the costs weak dispersion haloes in transported cover and the routine
and the risks involved in greenfields exploration and target availability of ICP-based methods, there have been significant
generation. advances in the use of selective extractions to isolate geochemi-
The value in such databases (though not necessarily the cal signals associated with specific mineral components or
purpose in accessing) is indicated by the NGS survey where, element forms in regolith. These are designed to physically
over the last 30 months, there have been 5 000 browser hits separate the (low concentrations of) trace elements contributed
(on individual web pages in the NGS), c. 50 downloads of to overburden by these secondary mechanisms, and associated
the complete dataset and c. 65 customized datasets for with transient mineral phases in the regolith, from the trace
specific regions or specific sets of samples (J. Grossman, USGS element contents inherited from the transported regolith itself.
pers. comm. 2007). A number of case studies are presented in Hall & Bonham-
An associated task has been government support for the Carter (1998). In comparing the results of studies from a range
development of regolith and landform maps in countries of geochemical terrains (Yeager et al. 1998; Gray et al. 1999;
10 D. R. Cohen et al.

Fig. 7. Typical improvements to

detection limits for ultra trace
ICP-MS/ICP-AES (inductively coupled
plasma–atomic emission spectrometry)
analytical package based on aqua regia or
4-acid digestion of regolith, reported by
commercial analytical companies.

Fig. 8. Relationship between various

selective geochemical extractions and
regolith mineral components attacked.

Cohen et al. 1998) it would appear that selective extraction has has been limited work on the actual mechanisms of the
met with more limited success in deeply weathered and arid extractions or the mineralogical–digestion response relationship
terrains than in terrains such as the recently glaciated areas of (Cave et al. 2004; Dalrymple et al. 2005; MacFarlane et al. 2005).
the northern hemisphere. Various commercial extraction methods have been developed
Although by no means a new concept (the use of dilute HCl with secret ingredients, but the more explicit methods typically
and similar selective extractant on soils and stream sediments contain various combinations of acids, chelating agents, reduct-
dates from the 1950s), the development and application of ants and exchangeable ions.
selective extractions has tended to dominate analytical methods One feature common to selective extraction data, though
development in geochemical exploration over the last decade. A usually more evident in the less vigorous digestions, is the high
general relationship between some common extractions and the degree of spatial variability over small distances (Seneshen
mineral components they theoretically attack is presented in 1997; Gray et al. 1999; Rutherford 2002). While this phenom-
Figure 8. Whereas most studies have tended to be empirical enon was initially attributed to problems in controlling reac-
investigations of the selective extraction response of regolith tion conditions and the effects of small variations in sample
components to deeply buried underlying mineralization, there mineralogy on digestion performance (e.g. carbonate altering
 Major advances in exploration geochemistry 11

Fig. 9. Effects of pH versus

temperature and reagent concentration
versus temperature for the extraction of
Mn from a soil by hydroxylamine. HCl,
showing pH and temperature to be the
most significant factors (from
Dalrymple 2007).

reaction pH or Fe oxides causing reprecipitation of elements in mineral analysis by portable spectral analytical devices which
the digestion), it is more likely that such variation over short can provide a rapid and detailed mineralogical characterization
distances (metre-scale) simply reflect the distribution of the of samples, which may itself provide indicators of mineraliza-
migration paths (fractures and other discontinuities in the tion or alteration, and provide a more solid basis for subsequent
regolith) within which elements may concentrate, producing interpretation of geochemical data (Taylor 2000; Mauger et al.
discrete anomalies within surface transported regolith until 2004).
subsequent processes laterally redistribute those transported
elements. Conversely, one effect of buried mineralization may
be to take a relatively uniform initial element distribution in
transported regolith cover and induce ‘geochemical mottling’ Quality control (QC)
through the various processes described above (i.e. an increase The need for quality control in geochemical exploration, from
in spatial variance within the regolith without altering the bulk sampling to chemical analysis, remains critical in exploration
composition of the block of regolith overlying the mineraliza- geochemistry. Systematic QC protocols are especially critical
tion). Rather than obscuring the geochemical response to for selective extractions as minor changes in extraction con-
mineralization, this high spatial variability may be the signal we ditions can have severe effects on the extraction rates (Fig. 9;
need to detect. Dalrymple 2007). In relation to compilations of existing data-
The development of field-based analytical devices has made set, significant problems exist in stitching together different
some progress in the areas of field-portable XRF (Glanzman & geochemical data-sets, or levelling data derived from different
Closs, pers. comm.) although proper understanding of the regolith materials, especially where the metadata and quality
limitations of such methods is essential. The other area is in control information are scarce (Agnew 2004).
12 D. R. Cohen et al.

Fig. 10. Comparison of Enzyme Leach

Cu and As, and a seven-point moving
average and moving variance for As,
from a traverse over buried
mineralization at Chimborazo, Chile
(data from Kelley et al. 2003).

DATA ANALYSIS may be that the samples displaying the geochemical or minera-
logical effects of mineralization lie within the main cluster of
Pattern recognition and anomaly detection observations rather than as outliers. The mineralized signal may
A dichotomy continues to exist between stochastic modelling just be just denser sub-clusters of points within a larger and
of geochemical data (exemplified by techniques such as factor more dispersed cluster of observations. Gauss is out and chaos
analysis or fractal analysis) and deterministic modelling of is in.
geochemical data (exemplified by Pearce Element Ratio An example of a new paradigm in geochemical signal
methods; Madeisky & Stanley 1993). detection is drawn from the Enzyme Leach extraction element
The traditional approach to the interpretation of exploration patterns for surface regolith samples collected over the
geochemical data has been to define a ‘background’ population Chimborazo deposit where mineralization is covered by a thick
based on some a posteriori distance metric relating observations gravel unit (Fig. 10; Kelley et al. 2003). Elevated, but highly
to the centre of the dataset, or the a priori characteristics of variable, Cu values form a continuous zone across mineraliz-
samples collected distal to known mineralization. Geochemical ation. Despite the appearance of ‘anomalies’ in the As data, the
boundaries are subsequently set, beyond which observations are application of a seven-point moving average across the profile
considered ‘anomalous’ (either high or low), or data may simply indicates no addition of As to the gravels above mineralization
be clustered with the assumption that the common or more (i.e. all the As was probably inherited with the gravel itself).
highly populated groups are related to regional-scale or Conversely, a seven-point moving variance for the As data
‘environmental’ geological processes and the samples affected indicates significant increase in variance over mineralization.
by mineralization will occur in small clusters or as extreme Hence, the ‘signal’ is a change in spatial variance (in the As
values (Cheng et al. 1996). Element associations are still inherited with the gravels) rather than addition or depletion of
generally determined either by parametric clustering methods or that element.
various forms of principal components or factor analysis Whereas the detection of population outliers or atypical
(Reimann et al. 2002), including spatial factor methods (Bailey values by classical means is still an important step in initial
& Krzanowski 2000). exploratory data analysis (Grunsky 2007), most practitioners
A number of inherent problems have been identified in such still focus on univariate rather than multivariate methods
an approach, especially for large areas where geology, regolith (Filzmoser et al. 2005). This probably reflects a lack of good
and landscape vary substantially (Riemann & Garrett 2005) as grounding in mathematical theory and methods, the limitations
they fail to recognize that background varies on all scales of standard software packages when arrayed against the com-
(Riemann & de Caritat 1998), from continents down to plexity of geochemical data, and the limited time allocated for
individual minerals. company staff to properly analyse their geochemical datasets.
Recent experience with selective geochemical data would Isolating complex patterns of correlation between variables,
suggest that the old notions of ‘background’ and ‘anomalous’ associations among samples and the identification of
be abandoned in favour of recognizing that the multifaceted mineralization-induced effects in data through conventional
geochemical characteristics of a suite of samples, reflecting the multivariate parametric statistical procedures may be biased or
influence of a large number of complex interacting processes obscured by the presence of multivariate outliers and non-
(ancient and modern). Instead, the objective of data processing normal variable distributions (Reimann et al. 2002). Such
should be to define that component of the total geochemical procedures generally require substantial pre-processing of data
process variability that is associated with the influence of prior to modelling and the recognition that geochemical vari-
mineralization, and establish the spatial relationship between ables display distributions that are complex and difficult to
mineralization and samples in which that signal is present. It transform (Reimann & Filzmoser 2002; Pison et al. 2003).
 Major advances in exploration geochemistry 13

Neural networks and the derivative self-organizing maps are

an alternate approach to revealing complex relationships within
multivariate data sets and identifying anomalies, which are less
dependent on the characteristics of the input data than conven-
tional methods (Kohonen 1995; Foody 1997; Lalor & Zhang
2001; Clare & Cohen 2002). Other matrix decomposi-
tion methods have been considered, including semi-discrete
decomposition (Skillicorn & Cohen 2004).
One area of statistical analysis that has received significant
attention has been the use of various forms of fractal and
related analysis to separate univariate or multivariate popula-
tions (Cheng et al. 1994; Cheng & Agterberg 1996; Gonçalves
2001; Shen & Cohen 2005; Li et al. 2004). Incorporation of the
de Wijs model of element redistribution has also been incor-
porated (Xie & Bao 2004; Agterberg 2007). Variation in the
fractal dimension of some geochemical patterns represents
changes in the processes or factors controlling the geochemical
values, for which one change (relating to one dimension) may
be the introduction of the effects of mineralization.
Ultimately the efficiency of a geochemical exploration pro-
gram will be defined by the percentage of errors in the
categorization of samples into the set whose geochemistry (and
mineralogy) reflects the effects of mineralization and the set
that does not. The data analysis component of this task may
require the provision of substantial resources in technical staff Fig. 11. Distribution of ‘exploration geochemistry’ papers published
in Geochemistry: Exploration, Environment, Analysis; Journal of Geochemical
and time. Exploration; and Applied Geochemistry, 1997–2007, allocated to main
subject areas.
Data integration and spatial analysis
Over the last decade the increase in performance of personal
computers and availability of a range of user-friendly software A major discovery might be that the best tool for detecting
packages permit easier handling of large datasets in spatial deeply buried sulphide mineralization is the simple pH meter
frameworks (GIS) and advanced iterative manipulation linked (Smee 2003).
to more flexible data visualization options (Aaron & Brown
2004; Racic 2006). An example is the integration of multivariate
geochemical data, PCA (Principal Components Analysis) and
topography has been shown for Sumatra (Grunsky & Smee Current challenges to mineral exploration: new
2002). commodities, new terrains, no workers
The detection of mineralization in areas dominated by trans-
ported regolith cover remains the principal challenge for
CONCLUSIONS exploration geochemistry, with U now added to the list of
commodities for which immediate future demand seems set to
Key advances in exploration geochemistry over the rise. Continuing interaction between exploration geochemistry
decade: did we progress? and environmental geochemical practitioners is certain: it is
The decade has been marked by a series of small to medium common for both groups to be asking the same fundamental
steps in the development of exploration geochemical theory questions about the factors controlling the behaviour of metals
and practice. While the promise of selective geochemical in the surficial environment (Plumlee & Logsdon 1997). Such
extractions as the ‘silver bullet’ for areas with thick transported convergence has been recognized in the renaming of the
or deeply weathered cover has not materialized, it has Association of Exploration Geochemists into the Association
prompted research into the fundamental processes of geo- of Applied Geochemists in 2004, and moves to provide
chemical dispersion and new approaches to the definition and opportunities (publications and conferences) where both sides
the detection of geochemical patterns that might be related to of the artificial geochemical divide can meet.
the effects of underlying mineralization. With new models there The most serious challenge is not technical. The upturn in
is opportunity to ‘mine’ old datasets (Stegman 2002; Agnew commodities has been met by increasing severity in the
2004). The generation of a series of high quality, regional to shortage of suitably qualified and experienced scientists,
sub-continental scale geochemical and regolith maps has rekin- especially exploration geochemists and geophysicists, coupled
dled general interest in landscape geochemistry and demon- to a progressive contraction of geology in the university sector
strated that geology is usually the main determinant of regolith (Cohen 2004). This will have further critical implications for
geochemistry. the future research support provided to industry (Hogan 2004).
These features can be seen in the distribution of papers in As governments now view education as another commodity,
the three principal journals in which exploration geochemical the laws of supply and demand need to be carefully considered
research tends to be published (Fig. 11). Whereas data process- by industry and ways in which the decline in the number of
ing is the surprise winner, selective extractions is probably the universities able to deliver credible geoscience programs capa-
underlying area of research focus if the papers designated as ble of providing the initial training and career gateway into the
being mainly analytical, exploration in covered terrains and exploration and mining industry can be reversed. It is a cold
partial extractions are examined. fact: no staff, no graduates and no graduates, no exploration.
14 D. R. Cohen et al.

Future research direction: grasping the nettle B, T.C. & K, W.J. 2000. extensions to spatial factor methods
with an illustration in geochemistry. Mathematical Geology, 32, 657–682.
Our geochemical crystal ball suggests a number of future B, O., P, J.C., G, G. & B, A. 2002. Ferricrete
research directions that will be supported by further develop- genesis and supergene behaviour in Burkina Faso, West Africa.
ments and improvements in analytical methods and the assem- Geochemistry: Exploration, Environment, Analysis, 2, 15–25.
bly of multi-disciplinary research teams. A number of these are B, D.F. 2002. Observations on the funding of geological survey
further considered by Coker (2010). The mechanisms of organizations. Proceedings of the 2nd World Mining Ministers Forum. Toronto,
element dispersion must be extended by more detailed investi- 13–15 March.
B, J.S.M. 1999. Regolith mapping. is it necessary? Exploration Geo-
gation of the role of biological systems in element mobilization chemistry for the New Millennium. Australian Institute of Geoscientists Bulletin,
and precipitation (the presence of specific gene sequences may 30, 125–135.
become another variable to measure in regolith samples). Bø, B. & L, O. 1975. An electrochemical model for element
Broad-based isotopic analysis, from the light to the heavy distribution around sulphide bodies. In: E, I. & F, K. (eds)
elements, will become more routine. This will be supported by Geochemical Exploration 1974. Elsevier, Amsterdam, 631–648.
increased research efforts to establish the regolith isotopic B, J. 2002. Ministerial inquiry into greenfields exploration in Western Australia
signatures characteristic of different mineral deposits under and Government response to the inquiry. Western Australian Government, Perth.
B, A.D. & H, S.M. 2004. Regolith-landform maps are an essential
various regolith setting and geological terrains.
tool for interpreting regolith geochemistry – the White Dam, SA, experi-
We will abandon the concept of fixed geochemical ‘base- ence. In: R, I.C. (ed.) Proceedings of the CRC LEME Regional Regolith
lines’, ‘backgrounds’ and ‘anomalies’ in favour of (1) recogniz- Symnposia 2004. CRC LEME, Perth, 37–41.
ing geochemical signals relating to the effects of mineralization, B, C.R.M., L, M.J. & A, R.R. 2000. Evolution of regolith and
especially in multivariate data, and (2) focusing on suitable landscapes in deeply weathered terrain-implications for geochemical explo-
methods to isolate those processes that have affected the ration. Ore Geology Reviews, 16, 167–183.
geochemistry of our samples and which are of interest to B, C.R.M., R, I.D.M., S, K.M. & C, M. (eds) 2005.
Regolith expression of Australian ore systems. CRC LEME, Perth.
explorers. Anomaly detection will migrate from the algorithmic
C, E.M. & L, M.I. 2005. Relationship between ground
to the heuristic, from the parametric to more robust and water chemistry and soil geochemical anomalies at the Spence copper
knowledge-based approaches. porphyry deposit, Chile. Geochemistry: Exploration, Environment, Analysis, 5,
There will be continued integration between the exploration 135–145.
and environmental geochemical disciplines in recognition that C, E.M. & L, M.I. 2006. Atacamite in the oxide zone of
deposits cannot be mined until (1) they are discovered and (2) copper deposits in northern Chile – Involvement of deep formation
proper environmental controls on mining can be maintained. waters. Mineralium Deposita, 42, 205–218.
C, E.M., H, S.M., L, M.I., H, G.E.M. &
We are grateful for assistance in the preparation of this paper by S.G. MC, M.B. 2004. Finding deeply buried deposits using
Gatehouse and G.R. Taylor, UNSW (Sydney), N.F. Rutherford, geochemistry. Geochemistry: Exploration, Environment, Analysis, 4, 7–32.
RMRC (Sydney), B. Cauglin, ALS-Chemex (Vancouver), J. C, E.M., L, M.I. & K, D.L. 2002. Exploring for
Grossman, USGS (Denver) and I.J. Dalrymple, Actlabs (Perth). deeply-covered mineral deposits. Formation of geochemical anomalies in
Valuable criticism was provided by L.G. Closs, CSM (Golden) and northern Chile by earthquake-induced surface flooding of mineralized
K. Witherly, Condor Consulting (Lakewood). A brief overview such ground waters. Geology, 30, 1007–1010.
as this will inevitably fail to acknowledge many of the significant C, B. 2010. Developments in analytical technology. Geochemistry:
steps that have been taken by geochemists in the development of Exploration, Environment, Analysis, 10.
geochemical exploration theory and practice, and we apologize to C, M.R., M, A.E. & F, E.N. 2004. Evaluation of a method
those individuals in advance. With the passing of Professor J. Webb, for identification of host physico-chemical phases for trace metals and
the first generation of exploration geochemical gurus (the primary measurement of their solid-phase partitioning in soil samples by nitric acid
anomalies) is now almost at an end and the subsequent generation extraction and chemometric mixture resolution. Geochemistry: Exploration,
(the secondary anomalies) acknowledge the strong foundations upon Environment, Analysis, 4, 71–86.
which the discipline continues to grow. C, Q. & A, F.P. 1996. Multifractal modelling and spatial
statistics. Mathematical Geology, 28, 1–16.
REFERENCES C, Q., A, F.P. & B, S.B. 1994. Separation of
geochemical anomaly from background by fractal methods. Journal of
A, D. & B, D.A. 2004. Production of geochemical atlas pages Geochemical Exploration, 51, 109–130.
using Datadesk and ArcGIS Toronto. In: R, I.C. (ed.) Regolith 2004. C, Q., A, F.P. & B-C, G.F. 1996. A spatial
CRC LEME, Perth, 32–36. analysis method for geochemical anomaly separation. Journal of Geochemical
A, P.D. 2004. Applications of geochemistry in targeting with emphasis
Exploration, 56, 183–195.
on large stream and lake sediment data compilations. In: M, J.,
C, A.P. & C, D.R. 2002. A comparison of unsupervised neural
G, R.J. et al. (eds) Predictive Mineral Discovery Under Cover. Society of
networks and k-means clustering in the analysis of multi-element stream
Economic Geology Symposium, 27 Sept–1 Oct, Society of Economic
sediment data. Geochemistry: Exploration, Environment, Analysis, 1, 119–134.
Geologists, Perth, 232.
A, F. 2007. New applications of the models of de Wijs in regional C, L.G. 1997. Exploration Geochemistry. Expanding Contributions to
geochemistry. Mathematical Geology, 39, 1–25. Mineral Resource Development. In: G, A.G. (ed.) Proceedings of
A, R.R. 2000. Regolith and geochemical synthesis of the Yandal Exploration 97, Fourth Decennial International Conference on Mineral Exploration.
greenstone belt. In: P, G.N. & A, R.R. (eds) Yandal Greenstone Paper 1, 3–8. Prospectors and Developers Association, Toronto.
Belt, Regolith, Geology and Mineralization. Australian Institute of Geoscientists, C, D.R. 2004. Who will deliver the last classical exploration
Bulletin, 31, 79–112, Perth. geochemistry course. Explore, 123, 1.
A, R.R. &  B, P. (eds) 2005. Regolith-landscape Evolution across C, D.R., S, X.C., D, A.C. & R, N.F. 1998. A
Australia. CRC LEME, Perth. comparison of selective extraction soil geochemistry and biogeochemistry
A, R.R. & P, M. 2002. Regolith geology of the Yilgarn Craton, in the Cobar Area, NSW. Journal of Geochemical Exploration, 61, 173–190.
Western Australia – Implications for exploration. Australian Journal of Earth C, W.B. 2010. Future research directions in exploration geochemistry.
Sciences, 49, 4–162. Geochemistry: Exploration, Environment, Analysis, 10, 75–80.
A, R.R., C, M. & P, C. 2007. Use of vegetation and soil C, W.B., H, E.H.J. & C, E.M. 1987. Lake sediment
in mineral exploration in areas of transported overburden, Yilgarn Craton, geochemistry applied to mineral exploration. In: H, P.H. (ed.) Geophysics
Western Australia – A contribution towards understanding metal transpor- and Geochemistry in the Search for Metallic Ores. Proceedings of Exploration, 77,
tation processes. Geochemistry: Exploration, Environment, Analysis, 7, 267–288. Ottawa, Canada, 435–477.
A, L.V., G, I.S., V, N.A. & R, J.S. 1992, New C, M., M, P.A. & C, A.J. 2006. Laterite geochemical
methods of exploration for blind mineralization in the USSR. Journal of database for the southwest Yilgarn Craton, Western Australia. CRC LEME Open
Geochemical Exploration, 43, 157–166. File Report 201 and CSIRO Report P2006/75.
 Major advances in exploration geochemistry 15

D, I.J. 2007. An approach to the optimisation of partial extractions for H, S.L. & K, N.M. 2007. Zircon tiny but timely. Elements, 3, 3–18.
geochemical exploration. PhD thesis, University of New South Wales, Sydney. H, P. 2004. Overview of soilgas theory. Explore, 122, 1.
D, I.J., C, D.R. & G, S.G. 2005. Optimization of H, L. 2004. Research and development in exploration and mining – Implications
partial extraction chemistry for buffered acetate and hydroxylamine for Australia’s gold industry. Australian Bureau for Agricultural and Resource
leaches. Geochemistry: Exploration, Environment, Analysis, 5, 279–285. Economics, e-Report 043.
D, A.G., B̈, A. et al. 1995. A global geochemical database for K, D.L., H, G.E.M., C, L.G., H, I.C. & ME, R.M.
environmental and resource management. Recommendations for international geochemical 2003. The use of partial extraction geochemistry for copper explora-
mapping. Final Report of IGCP Project 259. UNESCO Publishing, Paris. tion in northern Chile. Geochemistry: Exploration, Environment, Analysis, 3,
D, M.L., P, T.S. & C, R.E. 1994, Geochemistry of tectoni- 85–104.
cally expelled fluids from the northern Coast ranges, Rumsey Hills, K, D.,  C, P. & D, R. 2004. The application of the stable
California, USA. Geochimica et Cosmochimica Acta, 58, 1687–1699. isotopes of sulfur and oxygen in groundwater sulfate to mineral explora-
D, C.E. 2007. Biogeochemistry in Mineral Exploration. Handbook of tion in the Broken Hill region of Australia. Journal of Geochemical Exploration,
Exploration and Environmental Geochemistry, Volume 9. Elsevier. 78/79, 81–84.
F, P., G, R.G. & R, C. 2005. Multivariate outlier K, T. 1995. Self-Organizing Maps. Springer-Verlag, Berlin.
detection in exploration geochemistry. Computers and Geosciences, 31, L, G.C. & Z, C. 2001. Multivariate outlier detection and remedia-
579–587. tion in geochemical databases. The Science of the Total Environment, 281,
F, G.M. 1997. Fully fuzzy supervised classification of land cover from 99–109.
remotely sensed imagery with an artificial neural network. Neural Computing L, R., G, A. et al. 2000. An example of 3D conductivity mapping
and Applications, 5, 238–247. using the TEMPEST airborne electromagnetic system. Exploration
G, A. & M, R. 1997. Groundwater geochemistry in explo- Geophysics, 31, 162–172.
ration. an investigation in the Black Flag district, Western Australia. L, L.M. 1999. Multi-element dispersion in Mesozoic basin sediment
Australian Journal of Earth Sciences, 44, 433–443. over the Osborne deposit, northern Queensland – implications for regional
G, R.K. & C, L.G. 2010. Field portable XRF geochemical geochemical exploration in buried terrain. In: Australian Institute of
analysis – Its contribution to in situ, real-time project evaluation. Geoscientists (eds) Exploration Under Cover. Australian Institute of Geosci-
Geochemistry: Exploration, Environment, Analysis, 10, xx–yy. entists, Bulletin, 28, 73–81.
G, I.S. 1998. Vertical migration of elements from mineral deposits. L, C., M, T. & C, J. 2004. A fractal interpolatory approach to
Journal of Geochemical Exploration, 61, 191–202. geochemical exploration data processing. Mathematical Geology, 36, 593–606.
Gç, M.A. 2001. Characterization of geochemical distributions using L, J., W, G.(Chief compilers) et al. 1999. Atlas of the Ecological Environmental
multifractal models. Mathematical Geology, 33, 41–61. Geochemistry of China. Geological Publishing House, Bejing.
G, G.J.S. 1973. Differential secondary dispersion in transported soils L, M.J. 2005. Biogeochemical anomalies at barns gold prospect (Eyre
and post-mineralization rocks. An electrochemical interpretation. In: J, Peninsula, South Australia). In: R, I.C. (ed.) Regolith 2005 – Ten Years
M.J. (ed.) Geochemical Exploration 1972. Institute of Mining and Metallurgy, of CRC LEME. CRC LEME, Perth, 195–196.
London, 81–91. M, J. 2002. Obtaining leverage through geoscience. Options for
G, D.J., W, J.E. & L, G.D. 1999. Selective and partial governments and mining companies. Proceedings of the 2nd World Mining
extraction analyses of transported overburden for gold exploration in the Ministers Forum, 13–15 March, Toronto.
Yilgarn Craton, Western Australia. Journal of Geochemical Exploration, 67, MF, W.R., K, T.K., C, D., B, D. & O, C.
51–66. 2005. Continuous leach inductively coupled plasma mass spectrometry.
G, E.C. 2007. Evaluating and Interpreting Regional Geochemical applications for exploration and environmental geochemistry. Geochemistry:
Survey Data. In: M, B. (ed.) Proceedings of Exploration 07: Fifth Exploration, Environment, Analysis, 5, 123–134.
Decennial International Conference on Mineral Exploration. Toronto, Sept 9–12, M, H.E. & S, C.R. 1993. Lithogeochemical exploration for
Prospectors and Developers Association, Toronto. metasomatic zones associated with volcanic-hosted massive sulphide
G, E.C. & S, B.W. 2002. The differentiation of soil types and deposits using Pearce Element Ratio analysis. International Geological Reviews,
mineralization from multi-element geochemistry using multivariate 35, 1121–1148.
methods and digital topography. Journal of Geochemical Exploration, 67, M, A.W., B, R.D., M, A.T., H, D.B. & P, J.L.
287–299. 1998. Application of the mobile metal ion technique in geochemical
G, N., Bø, B., S, D.B. & S, R.C. 2001. Geo- exploration. Journal of Geochemical Exploration, 61, 87–102.
chemical landscapes of the conterminous United States—new map presentations for 22 M, A.J., K, J.L. & H, J.F. 2004. Bring Remote
elements. U.S. Geological Survey Professional Paper, 1648. Sensing Down To Earth. CSIRO Hylogger as applied to the Tarcoola
H, G.E.M. & B-C, G.F. 1998. Selective extractions. Journal of Goldfield, South Australia. In: Proceedings of the 12th Australasian Remote
Geochemical Exploration, 61. Sensing and Photogrammetry Conference. Freemantle. 18–22 October, 65.
H, G.E.M., H, S.M., MC, M.B. & C, E.M. MG, S.P. & L, P.J. 1992. The Soil Geochemical Atlas of England
2004. Secondary geochemical signatures in glaciated terrain. In: M, and Wales. Blackie Academic and Professional, London.
J., G, R.J. et al. (eds) Predictive Mineral Discovery Under Cover, Perth, M-W, R. 1994, Earthquakes, strain-cycling and the mobilization of
27 Sept–1 Oct. Society of Economic Geology symposium. SEG, Perth, fluids. In: P, J. (ed.) Geofluids: origin, migration and evolution of fluids in
79–84. sedimentary basins. Geological Society, London, Special Publications, 78,
H, S.M. 1998. Electrochemical mass-transport in overburden. a new 85–98.
model to account for the formation of selective-leach geochemical N, A. 1974, Matsushiro, Japan, earthquake swarm: confirmation of
anomalies in glacial terrain. Journal of Geochemical Exploration, 63, 155–172. dilatancy-fluid diffusion model. Geology, 2, 217–221.
H, S.M. 2000. Spontaneous potentials and electrochemical cells. In: O’R, S.Y., G, W.L. & B, E.A. 2004. TerraneChronTM –
H, M. & G, G.J.S. (eds) Geochemical remote sensing of the sub-surface. delivering a competitive edge in exploration. In: M, J., G,
Handbook of Exploration Geochemistry, volume 7. Elsevier, Amsterdam, R.J. et al. (eds) Predictive Mineral Discovery Under Cover. Society of Economic
81–119. Geology Symposium, 27 Sept–1 Oct, Society of Economic Geologists,
H, S.M., H, G.E.M., MC, M.B. & C, E.M. Perth, 156.
2003. Rapid mass-transport of elements through clay and the possible role P, C.F., C, R.A., C, M.A., G, D.L., K, P. & W,
of redox gradients. Proceedings of the 6th International Symposium on Environ- J.R. 2001. RTMAP Database Field Book and Users Guide. CRC LEME,
mental Geochemistry, September 8, Society for Environmental Geochemistry Report 13, Perth.
and Health, Edinburgh. P, B. 2005. Geochronology of the Australian regolith. In: A, R.R.
H, S.M., C, E.M., MC, M.B. & H, G.E.M. &  B, P. (eds) Regolith landscape Evolution across Australia. CRC
2004a. Redox, pH and SP variation over mineralization in thick glacial LEME, Perth, 41–52.
overburden, Part I. Methodologies and field investigation at the Marsh P, M.C. & G, A.M. 2004. Application of groundwater-mineral
zone gold property. Geochemistry: Exploration, Environment, Analysis, 4, 33–44. equilibrium calculations to geochemical exploration for sediment-hosted
H, S.M., C, E.M., MC, M.B. & H, G.E.M. uranium. observations from the Frome Embayment, South Australia.
2004b. Redox, pH and SP variation over mineralization in thick glacial Geochemistry: Exploration, Environment, Analysis, 4, 113–127.
overburden, Part II. field investigation at Cross lake VMS property. P, G., R, P.J., F, P. & C, C. 2003. Robust factor
Geochemistry: Exploration, Environment, Analysis, 4, 45–58. analysis. Journal of Multivariate Analysis, 84, 145–172.
16 D. R. Cohen et al.

P, J.A., R, S. et al. 2003. The distribution of uranium over Europe: S, R.H. 2001. Seismogenic framework for hydrothermal transport and
Geological and environmental significance. Applied Earth Science, 112, ore deposition. In: Structural controls on ore genesis. In: R, J.P. &
221–238. T, R.M. (eds) Reviews in Economic Geology. Society of Economic
P, J.A., S, D., S, B. & W, L. 2000. Environ- Geologists, 14, 25–50.
mental geochemistry at the global scale. Journal of the Geological Society, 157, S, D.B. & C, D.R. 2004. Detecting mineralisation using partial
837–849. element extraction – A Case study. In: 7th workshop on mining scientific and
P, G.S. & L, M.H. (eds) 1997. The environmental geochemistry of engineering datasets, 4th SIAM international conference on data mining, Florida,
mineral deposits. Society of Economic Geologists, Reviews in Economic April 24.
Geology, 6. S, B.W. 1983, Laboratory and field evidence in support of the
P, P.A., C, J.D.A., B, Y. & V, J. 2002. A CO2–O2– electrochemically-enhanced migration of ions through glaciolacustrine
light hydrocarbon-soil-gas anomaly above the Junction orogenic gold sediment. Journal of Geochemical Exploration, 19, 277–304.
deposit: A potential, alternative exploration technique. Geochemistry: S, B.W. 1998. A new theory to explain the formation of soil geochemical
Exploration, Environment, Analysis, 2, 333–344. responses over deeply covered gold mineralization in arid environments.
R, L. 2006. Extending the limits of earth data visualization and exploration. First Journal of Geochemical Exploration, 61, 149–172.
Break, European Association of Geoscientists and Engineers, 24, 61–65. S, B.W. 2003. Theory behind the use of soil pH measurements as an
R, C. &  C, P. 1998. Chemical elements in the environment – inexpensive guide to buried mineralisation, with examples. Explore, 118, 1.
factsheets for the geochemist and environmental scientist. Springer-Verlag, Berlin. S, C. 2002. Near mine exploration in the Cobar Goldfield – Making
R, C. & F, P. 2000. Normal and lognormal data distribution use of our greatest asset. In: T, C. (ed.) Exploration in the Shadow of the
in geochemistry: Death of a myth. Consequences for the statistical Headframe – Getting a New Life on Lease, SMEDG-AIG Symposium, Sydney,
treatment of geochemical and environmental data. Environmental Geology, 39, Oct 11.
1001–1014. T, G.R. 2000. Mineral and lithology mapping of drill core pulps using
R, C. & G, R.G. 2005. Geochemical background-concept and visible and infrared spectrometry. Natural Resources Research, 9, 257–268.
reality. Science of the Total Environment, 350, 12–27. T, J.S. 1973, The Kashmir (Turshiz) 1903 and Torbat-e Heidariyeh
R, C., F, P. & G, R.G. 2002. Factor analysis applied 1923 earthquakes in Central Khorassan (Iran). Annali di Geofisica, 26,
to regional geochemical data: Problems and possibilities. Applied 29–40.
Geochemistry, 17, 185–206.
U, J.R., D, M.L., C, R.E. & M, E.M. 1992, Implica-
R, B.D. & L, R.R. 2005. Multi-million-year cyclic ramp-up of
tions of perennial saline springs for abnormally high fluid pressures and
volatiles in a lower crustal magma reservoir trapped below the Tampakan
active thrusting in western California. Geology, 20, 431–434.
copper-gold deposit by Mio-Pliocene crustal compression in the southern
US G S (USGS). 2004. The national geochemical survey – Database
Philippines. In: P, T.M. (ed.) Super porphyry copper and gold deposits. A
and documentation. US Geological Survey, Open File Report, 2004–1001.
global perspective. PGC Publishing, Perth, 2, 270.
V, T., R, C., P, V.,  C, P. & Ä̈, M. 1997.
R, N.F. 2002. The regolith, background, anomalies and partial leaches.
ALS-Chemex News 2001–1. Overbank sediments from the surroundings of the Russian nickel mining
R, N.F., L, L.M. & S, G. 2005. Osborne copper- and smelting industry on the Kola Peninsula. Environmental Geology, 32,
gold deposit, Cloncurry Mining District, north west Queensland. In: 175–185.
Regolith Expressions of Australian Ore Systems. CRC LEME, Adelaide, 369– W, X., X, X., C, Z. & L, D. 1999. Delineation of regional
407. geochemical anomalies penetrating through thick cover in concealed
S, R. (ed.) & 27 others. 1998. FOREGS Geochemical sampling Field terrains – A case history from the Olympic Dam deposit, Australia. Journal
Manual. Geological Survey of Finland, Guide 47. of Geochemical Exploration, 66, 85–97.
S, R. (ed.) & 30 others. 2005. The Geochemical Atlas of Europe. X, S. & B, Z. 2004. Fractal and multifractal properties of geochemical
Geological Survey of Finland. fields. Mathematical Geology, 36, 847–864.
S, M. & M, H.M. 1960. The electrochemical mechanism of sulphide X, X. & C, H. 1997. The suitability of floodplain sediment as a global
self-potentials. Geophysics, 25, 226–249. sampling medium. Evidence from China. Journal of Geochemical Exploration,
S, D.M. 1997. The applicability of partial extractions to mineral exploration 58, 51–61.
in areas of transported overburden. PhD thesis, Queens University, Kingston. X, X., D, L., Y, X., G, Y. & C, L.
S, W. & C, D.R. 2005. Fractally invariant distributions and an 2004. Geochemical blocks for predicting large ore bodies – Concept and
application in geochemical exploration. Mathematical Geology, 37, 895–913. methodology. Journal of Geochemical Exploration, 82, 77–91.
S, R.H. 1981. Fluid flow accompanying faulting. Field evidence and Y, J.R., C, J.R., M, W. & R, R. 1998. Enzyme
models. In: S, D.W. & R, P.G. (eds) Earthquake prediction: Leach anomalies associated with deep Mississippi Valley-type zinc ore
An internal review.. American Geophysical Union, Maurice Ewing Series 4, bodies at the Elmwood Mine, Tennessee. Journal of Geochemical Exploration,
593–603. 61, 103–112.

Received 6 September 2008; revised typescript accepted 10 October 2008.