Open File Report 69
Spectral properties of soil and lag overlying the site of the
Beasley Creek Gold Mine, Laverton Region, Western Australia
Cudahy, T.J., Robertson, I.D.M. and Gabell, A.R.
Gold exploration, using spectral remote sensing in the Yilgarn
Craton, is difficult for at least two reasons. Firstly, very deep
weathering has changed the complex fresh rock mineralogy to a surface
mantle, largely consisting of ferric oxides, clays and quartz. Secondly,
there is little detailed knowledge of the reflectance properties
of these weathered surface materials, of very variable crystallinity,
overlying gold mineralisation and in background areas. To address
these problems, some of the studies in the P243 project have concentrated
on the measurement and analysis of 0.4 to 2.5 µm reflectance
spectra of a large range of regolith materials overlying gold deposits
at Beasley Creek (this study), Bounty (in the Forrestania region,
report 169R) and Panglo (in the Ora Banda region, report 234R).
Background areas have included studies at Laverton (report 235R),
Lawlers (report 240R) and Ora Banda.
The Beasley Creek gold mine is located in the Laverton area, north-eastern
Yilgarn Craton. Before disturbance of the ground by open-pit mining,
the Beasley Creek gold deposit was situated under the crest of a
small, 3.5 m high rise. Gold mineralisation was related to deeply
weathered carbonaceous shale and chert rocks, within an Archaean
greenstone sequence. Laboratory reflectance spectra were measured
of surface samples collected from two east-west traverses across
gold mineralisation and extending into background areas. The samples
comprised soil and ferruginous lag overlying subcropping lateritic
duricrust, exposed saprolite and mottled zone.
The spectral results show no evidence for sericite or other primary
minerals related to the original fresh rock. Instead, the spectra
show information related to the products of weathering, namely,
ferric oxides (hematite and goethite) and kaolinite.
The spectra of the coarser, ferruginous lag (10-50 mm diameter)
show variations in the wavelengths of the ligand-metal charge transfer
shoulder, near 0.6 µm, and the iron crystal field absorption
minima, near 0.9 µm, indicating hematite or goethite. The
relative changes in the wavelength of these parameters are invariant
over broad zones (>500 m wide), consistent with hematite-goethite
relationships measured by XRD and spatially related to particular,
exposed, lateritic units. The goethitic lag is located over saprolite
and mottled zone and is interpreted to be the residual product of
deflation of the upper, lateritic horizons. The hematitic lag is
located over lateritic duricrust.
The soil shows a relatively consistent spectral mineralogy comprising
hematite and poorly crystalline kaolinite. The poor kaolinite crystallinity
is indicated by the weakly developed absorption doublets at 1.4
and 2.2 µm. Within a 100 m wide zone, over gold mineralisation,
there is an 8 nm shift to shorter wavelengths of the charge transfer
shoulder, indicating slightly more goethite-rich soil. This shift
in wavelength is much less than that shown by the coarse lag.
The soil and lag data indicate a weak relationship between gold
and the wavelength of the charge transfer shoulder. This relationship
shows that the weathered materials, with more gold, are relatively
goethite-rich though, there are too few data to consider this result
The associated P240 and P241 investigations noted an increase in
the overall abundance of iron-rich lag in the vicinity of gold mineralisation.
The increase in the total ferric iron content of the soil and lag
was examined, using the depth of the crystal field absorption near
0.9 µm. This spectral parameter showed weak correlation with
the Fe2O3 content but no relationship to areas of gold mineralisation.
However, this does not preclude gold being associated with higher
ferric iron content at the surface. An increased amount of ferruginous
lag at the surface will increase the total iron content, especially
in the context of remote sensing applications.
According to a regolith model (report 243R), passage from saprolite
and mottled zone to lateritic duricrust is hypothetically associated
with a decrease in the abundance of clays. This relationship was
tested for the soil developed over these particular units, using
the depth of the AlOH-related 2.2 µm absorption. The depth
of the 2.2 µm absorption was found to correlate with the A12O3
content of the soil but there was poor correlation with the position
in the regolith. The results showed soil mantling saprolite and
mottled zone has relatively shallow 2.2 µm adsorptions (lower
clay abundances) compared to the soil covering lateritic duricrust.
It is suggested that this anomaly is caused by increased winnowing
of the soils, overlying saprolite and mottled zone on the western-side
of the low rise, by the prevailing winds.
The associated P240 and P241 studies found that powdery carbonates
occur as patches within the soil. However, spectral examination
of carbonate-rich soil showed no evidence of the diagnostic carbonate
absorption at 2.33 µm, even though scanning electron microscope
examinations showed calcium-rich particles were well exposed at
the surface of the soil minerals. A related P243 study (report 169R)
discovered more than 40% by weight of sub-75 µm carbonate
powder was required in a soil before carbonates were spectrally
recognisable. This is much greater than the maximum carbonate content
of the Beasley Creek soil. The only indication of carbonate in the
spectra was an increase in the albedo, particularly the visible
Spectral analysis of the depth, width and wavelength of the 1.9
µm, water-related absorption, can provide information about
the content and site (free, adsorbed, trapped) of water molecules.
The soil shows two distinct populations of 1.9 µm adsorptions,
one associated with free water (driven off after heating to 100°C),
and the other associated with water trapped in quartz as fluid inclusions
(unrelated to gold mineralisation).
The coarse, ferruginous lag commonly shows an upward ramp in reflectance
from 0.9 and 1.3 µm which can dominate spectral properties
in this region and will influence the geometry of the ferric iron
absorption at 0.9 µm. Another P243 study (report 244R) found
the intensity of this feature is correlated with the depth and width
of the 1.9 µm absorption and suggested this water is either
adsorbed on the surfaces of the iron oxide crystals, by hydrogen
bonding, or is intimately associated with silica and iron oxide
and formed after lateritisation (report 243R). This spectral property
may be associated with the affects of desert varnish, though the
underlying rock mineralogy is clearly evident in the spectra (for
example, the hematite-goethite ratio).
The spectral results show that the soil and ferruginous lag overlying
the Beasley Creek gold mineralisation are not mineralogically distinct
from the background materials. An increased development of goethite
(shorter wavelength charge transfer shoulder) appears to be associated
with gold at Beasley Creek but goethite is pervasively developed
in varying abundances throughout the regolith. Therefore, the significance
of ferric oxide development to gold mineralisation needs to be further
evaluated, especially in the context of the regolith. The spectral
results appear to be useful for the regolith characterisation of
the soil and lag and so could be used to help classify regolith
materials to assist geochemical gold exploration.
Last updated: Thursday, January 06, 2000 11:42 AM