Open File Report 67
Spectral properties of soil overlying the sites of the Bounty
and North Bounty gold mines, Forrestania Region, Western Australia
Cudahy, T.J., Lintern, M.J. and Gabell, A.R.
Gold exploration in the Yilgarn Craton using spectral remote sensing
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 spectral properties of
these weathered surface materials, of very variable crystallinity,
overlying gold mineralisation and background areas. To address these
problems, some of the studies in the P243 project have concentrated
on the measurement and analysis of the 0.4 to 2.5 mm spectra of
a large range of regolith materials overlying gold deposits (reports
160R, 169R and 234R) and background areas (reports 235R and 240R).
This study of the spectral properties of weathered materials overlying
gold mineralisation at the Bounty and North Bounty deposits, Forrestania
region, was conducted in association with P241 mineralogical and
geochemical studies on the same samples which meant that spectral
information could be directly compared with other physicochemical
The exposed regolith units of the study area comprise saprolite,
lateritic duricrust and alluvium-colluvium which have developed
over mafic, ultramafic and sedimentary parent rocks. Primary gold
mineralisation is hosted by shales and cherts. The collection of
samples preceded major disturbance of the surface materials by mining.
Approximately 100 samples were measured, including 57 near-surface
samples from three east-west profiles spanning primary gold mineralisation
and background areas, and 25 samples from drill core and costeans,
which provided information from deeper down the regolith profile.
Only 6 samples were taken from areas overlying primary gold mineralisation
and approximately 12 from areas associated with secondary gold.
The spectral results show no evidence for alteration minerals associated
with primary gold mineralisation. Only one spectrum from weathered
surface materials showed spectral properties related to primary
minerals, including absorptions at 1.1 (ferrous iron), 2.306 and
2.384 µm (Mg-trioctahedral silicate). All the other surface
samples were characterised by spectral properties related to the
products of weathering, namely the ferric oxides (hematite and goethite)
and clays (kaolinite and smectite). The spectral variations associated
with these weathering products show poor relationships with primary
gold mineralisation, though there is information related to the
character of the regolith, and possibly secondary gold.
The associated studies found secondary gold, is in places, associated
with pedogenic carbonate in the top metre of the soil profile. However,
the spectra of even the most carbonate-rich samples showed no evidence
for the CO32- absorptions in the 2.3 µm wavelength region.
This lack of carbonate absorption was found to be caused by insufficient
abundance of carbonate, as a laboratory experiment showed 40% by
weight of carbonate is required in a given mineral mixture (less
than 75 µm particle size) before the absorption at 2.33 µm
is recognisable. The soils from the study area contained less than
30% carbonate by weight
The reflectance data provide other evidence, albeit indirectly,
for the presence of pedogenic carbonate. This includes increased
albedo, particularly in the visible wavelengths. Also, an intimate
relationship between carbonate and smectite was discovered. The
smectite is recognised by characteristic adsorptions at 1.4, 1.46,
1.9 and 2.2 µm. The associated XRD analyses confirmed the
presence of smectite only after XRD was performed on clay-rich concentrates.
Kaolinite is evident in the spectra of carbonate-poor samples and
is recognised by absorption doublets at 1.4 and 2.2 µm.
The spectral data show information related to the hematite-goethite
ratio. This information is indicated by shifts in the wavelength
of the ferric iron charge transfer shoulder, near 0.6 µm,
and the crystal field absorption, near 0.9 µm. As with the
smectites, XRD did not provide much information on the iron oxide
mineralogy. The reflectance data show the carbonate-poor soils are
related to hematite whereas the carbonate-rich soils are related
to goethite. The reasons for these changes in the iron oxide mineralogy
are not clear but may be related to the effects of Ca and Mg on
The depth of the 0.9 µm absorption was found to correlate
with the Fe203 content. A few soils from primary gold mineralisation
show relatively deep 0.9 µm absorption, though further work
is required to establish the wider significance of this result as
regolith-related variations appear to be more important.
The depth of the 2.2 µm absorption was examined because it
can provide information on the relative clay abundances important
for discriminating regolith units and possibly gold mineralisation.
The 2.2 µm absorption depth was found to correlate with the
Al2O3 content but a correction factor had to be applied before the
2.2 µm absorption depth could be related to the total clay
abundance. This correction is related to the different proportions
of A13+ in the smectite and kaolinite structures. The corrected
results show the "clay abundances" in the soils over exposed
saprolite and lateritic duricrust are similar and that there are
no variations spatially related to primary or secondary gold.
The reflectance data provide information consistent with a model
for regolith development (report 243R). Important for the interpretation
of the Mt Hope regolith is the recognition that pedogenic carbonate
(associated with smectite and hematite) overprinted the earlier-formed
laterite mineralogy. According to the regolith model, saprolite
is characterised by well-crystalline kaolinite and goethite. However,
this mineralogy was overprinted by the enrichment of Ca and Mg and
the development of smectite, hematite and carbonate. Regional geochemical
sampling and analysis could be assisted using spectral sensing (remote
or proximal) by mapping both the surface distribution of pedogenic
carbonate and laterite units. Further work is required to establish
the significance of more subtle spectral information to gold mineralisation.
This subtle information may have to be interpreted in the context
of a quantitative regolith model. If for example, gold mineralisation
was found to be related to a small increase in the iron oxide content,
then larger iron oxide variations, related to regolith development,
would first have to be removed (by normalisation with a predicted
regolith model) so that more subtle information can be enhanced.
Last updated: Thursday, January 06, 2000 11:34 AM