Ore Geology Reviews 54 (2013) 181–196

Contents lists available at SciVerse ScienceDirect

Ore Geology Reviews

journal homepage: www.elsevier.com/locate/oregeorev

Detection of hydrothermal alteration zones in a tropical region using satellite remote

sensing data: Bau goldfield, Sarawak, Malaysia
Amin Beiranvand Pour, Mazlan Hashim ⁎, John van Genderen
Institute of Geospatial Science & Technology (INSTeG), Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor Bahru, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: Remote sensing for geology in tropical environments is very challenging, because of the dense vegetation cover
Received 24 November 2012 and the problem of persistent cloud cover. In this research paper, we have investigated and demonstrated the
Received in revised form 27 March 2013 detection of hydrothermal alteration zones and structural elements associated with intrusion-related gold min-
Accepted 28 March 2013
eralization using various types of remote sensing data in the Bau gold mining district in the State of Sarawak, East
Available online 15 April 2013
Malaysia, on the island of Borneo. The climate of Bau is tropical with persistent cloud cover and very dense veg-
Keywords:
etation ground. Geological analyses coupled with remote sensing data were used to detect hydrothermally al-
ETM+ tered rocks and structural elements associated with gold mineralization in the Bau area. Landsat Enhanced
Hyperion Thematic Mapper+ (ETM+), Hyperion and Phased Array type L-band Synthetic Aperture Radar (PALSAR) data
PALSAR were used to carry out lithological–structural mapping of the mineralized zones in the study area and surround-
Bau gold mining district ing terrain. Hydrothermal alteration zones were detected along the SSW to NNE structural trend of the Tai Parit
Lithological and structural mapping fault that corresponds with the occurrence of other gold mineralization in the Bau Limestone. The results show
that the known gold prospects and potential areas of mineralization are recognizable by the methods used,
despite limited bedrock exposure. The approach used in this study is broadly applicable to the detection of
gold mineralization using ETM+, Hyperion and PALSAR data in tropical/sub-tropical regions.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction
Numerous remote sensing investigations for mineral exploration
and lithological mapping have been conducted in arid and semi-arid
terrains, with large exposures of geologic materials, allowing the
acquisition of spectral information directly from rock–soil assemblages
(Di Tommaso and Rubinstein, 2007; El Desouky et al., 2008; Gabr et al.,
2010; Massironi et al., 2008; Pour and Hashim, 2011, 2012a, 2012b,
2013; Rajendran et al., 2012; Sabins, 1999; Van Ruitenbeek et al.,
2012; Zhang et al., 2007). However, in tropical environments, the appli-
cation of remote sensing data for geological mapping has been much
more limited (Carranza and Hale, 2002; Galva˜o et al., 2005; Vicente
and Filho, 2011), because of the dense, often complete vegetation cover
in tropical regions. Moreover, the persistent cloud cover and limited bed-
rock exposures are other obstacles imposed by tropical environments.
In this study, the identification of hydrothermally altered rocks,
faults and fractures associated with hydrothermal mineralization in a
tropical environment is examined using Landsat Enhanced Thematic
Mapper+ (ETM+), Hyperion and Phased Array type L-band Synthetic
Aperture Radar (PALSAR) remote sensing data sets. The Bau gold mining
district in Sarawak province, eastern Malaysia, on the island of
⁎ Corresponding author. Tel.: +60 7 5530666; fax: +60 7 5531174.
E-mail addresses: beiranvand.amin80@gmail.com (A.B. Pour),
mazlanhashim@utm.my, profmhashim@gmail.com (M. Hashim),
jlvangenderen@hotmail.com (J. van Genderen).
Borneo in Southeast Asia serves as the study area (Fig. 1). It is located
about the intersection of 1° 25′ N longitude and 110° 10′ E latitude,
25 km southwest of Kuching city, Sarawak (Fig. 2). The climate of
Bau is characterized by heavy but seasonal rainfall, uniform temper-
ature, and high humidity. Some 40% of the land is primary rainforest,
mostly of the Dipterocarpaceae type, restricted to infertile limestone
hills and higher mountains (Andriesse, 1972).
Bau is a goldfield similar to the Carlin district of gold deposits
(Schuh, 1993; Sillitoe and Bonham, 1990). Gold mineralization at Bau
is present from intrusive centers outward to the intrusive peripheries,
but principal gold deposits occupy distal to intrusive sites and have
the same geological, mineralogical, and geochemical characteristics as
Carlin (sediment-hosted) gold deposits in the western United States
(Sillitoe and Bonham, 1990). However, the Bau gold mineralization
does not fit into today's state of knowledge about Carlin-type gold
mineralization. It may be classified as intrusion-related or, using USGS ter-
minology, distal disseminated. Carlin-type gold deposits are disseminated
gold deposits hosted by a variety of permeable sedimentary rocks, espe-
cially thinly bedded, silty dolomites or limestones, cut by high-angle faults
(Bagby and Berger, 1985; Percival et al., 1988; Sillitoe and Bonham, 1990).
Nearly all deposits are associated with felsic intrusive rocks, commonly in
the form of dikes or sills. Orebodies may be confined to fault zones or may
be irregularly-shaped replacements in adjoining rocks. Gold-bearing
rocks underwent decalcification, silicification, and argillic alteration, and
are associated closely with structurally-localized replacement of car-
bonate rock by jasperoid (Arehart, 1996; Rockwell and Hofstra, 2008;

0169-1368/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.oregeorev.2013.03.010
182 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196

Fig. 1. Location of the study area in Southeast Asia. The Bau gold mining district is in westernmost Borneo. It is located in the state of Sarawak, which is part of Malaysia.
Modified from Hall (2012).

Sillitoe and Bonham, 1990; Zhu et al., 2011). These geological character-
istics can be detected as indicators in the reconnaissance stages of explo-
ration for intrusion-related gold mineralization using remote sensing
data.
Gold was first discovered in the Bau goldfield in the 1870s, but there
was very little production until 1909, and only about 22,000 oz was
produced through 1964 (Dill and Horn, 1996; Hon, 1981; Schuh, 1993).
In comparison, by 2008, the intrusion-related mineralization trend in
Sarawak had produced over 70 million ounces of gold (Stevens and
Fulton, 2010). The Bau goldfield is located in this mineralization trend
where a variety of mineralization styles ranging from mesothermal
to low-temperature deposits were recognized. Six distinct types of
ore deposits occur in the Bau mining district, including porphyry-copper
deposits (Gunong Ropih, Gunong Juala), contact metasomatic Cu–Au
skarn deposits (Arong Bakit), Cordilleran vein-type mesothermal pre-
cious metal and Pb–Zn–Ag sulfide deposits (Sarabau Mine, Bekajang,
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Fig. 2. Geologic map of the Bau mining district.
Modified from Percival et al. (1990).
Bukit Young, Tai Parit), disseminated and shale-hosted gold deposits
(Jugan), distal and low-temperature Hg–Ba, and As deposits (Tegora,
Gading) (Schuh, 1993). Previous geological studies indicated that gold,
copper and Pb–Zn–Ag sulfide mineralizations are associated with signifi-
cant structural (faults) and stratigraphic (composition/permeability) con-
trols, which are associated with hydrothermal alteration zones, consisting
of decarbonatization, silicification (jasperiod formation) and argillization
(Dill and Horn, 1996; Kim, 1994; Percival et al., 1990; Schuh, 1993;
Sillitoe and Bonham, 1990).
To our knowledge, a remote sensing study has not been carried out
in the Bau mining district to identify hydrothermally altered rocks and
structure elements associated with mineralization. Therefore, this inves-
tigation attempts to acquire comprehensive and accurate information
for exploring potentially interesting areas of gold mineralization by inte-
grating ETM+, Hyperion and PALSAR remote sensing data sets and ancil-
lary data sets. The objectives of this research are: (1) to introduce an
approach for detecting hydrothermally altered rocks using ETM+ and
Hyperion data at regional and district scales in a heavily vegetated trop-
ical rainforest region; (2) to identify structural elements associated with
fault-controlled and stratiform gold orebodies using microwave PALSAR
data; and (3) to discover areas of potential gold mineralization in the
Bau mining district.
2. Geology of the study area
Borneo Island is located in Southeast Asia (Fig. 1). It is the product of
Mesozoic accretion of an ophiolitic, island arc and micro-continental frag-
ments, together with marginal basin fill, onto a Paleozoic continental core
183
in the southwest of the island (Hall, 2012; Hutchison, 1989; Metcalfe,
1996; Meyerhoff, 1995; Wilson, 2002). The oldest rocks in our test region
are the Late Triassic Serian Volcanics. These are I-type, magnesian, tholei-
itic to boninitic andesites derived from hot, anhydrous mantle-derived
magmas. The Serian Volcanics were emplaced in a shallow-marine to
subaerial island arc to back arc basin environment. The 190 m.a. Jagoi
intrusions are I-type, calc-alkaline, hornblende-bearing, and silicic
granodiorites and were intruded during the Early Jurassic orogeny (Kirk,
1968; Schuh, 1993).
Crustal thickening and uplift during the Middle Jurassic resulted in
erosion and uplift of the older rocks. A slope-trench system devel-
oped. The Late Jurassic to Early Late Cretaceous Pedawan turbidites
were eroded from the andesitic–basaltic Serian Volcanics and were
rapidly buried as slope deposits, remaining only a short time in the
weathering cycle (Wilson, 2002). Simultaneously, gradual subsidence
on the unstable shelf edge led to the development of massive rudist
patch reefs of the Late Jurassic to Middle Cretaceous Bau Limestone
Formation. The generalized stratigraphy of the Bau district is shown
in Fig. 3 (Schuh, 1993).
Cessation of subduction during Late Cretaceous was possibly in-
duced by collision with a microcontinent arriving at the trench. Crustal
thickening and corresponding uplift resulted in erosion that produced
extensive terrestrial molasse basins of the Early Tertiary Kayan Forma-
tion (Schuh, 1993; Schuh and Guilbert, 1990). Worldwide plate reorga-
nization in the Mid-Eocene resulted in a change of plate motion in the
direction of Borneo. Subduction ceased altogether in west and central
Borneo, and shifted towards North Borneo, where an Oligocene volcanic
arc fired up. Molasse sedimentation of the Kayan Formation ended in
184 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196

Fig. 3. Generalized stratigraphy of the Bau district.

Modified from Schuh (1993).

the Mid-Eocene, and the Kayan sandstones were folded into broad
synclines. The 48 m.a. Sirenggok intrusion in Bau was emplaced. Re-
gional uplift, crustal consolidation, and erosion ensued. Early Miocene
I-type, calc-alkaline granodiorites were embedded between slope sedi-
ments and Early Tertiary fore-arc basin sediments in Central Kalimantan
(Moss, 1998; Tate, 1991).
In West Sarawak, regional extension in Mid-Miocene produced a
200-km-NNE-trending strike-slip fault system with a wrench com-
ponent. The Bau Trend became the site of emplacement for reduced,
calc-alkaline, hydrous, porphyritic, magnetite-, hornblende-bearing
granodiorites. Several intrusions display porphyry-copper style min-
eralization, skarn, limestone polymetallic replacement, epithermal
precious metal, disseminated gold, and Ba–Hg deposits (Dill and Horn,
1996; Kim, 1994; Percival et al., 1990; Schuh, 1993). This Mid-Miocene
mineralization event represented a second enrichment of metals,
already pre-concentrated during the Late Cretaceous accretion.
Lack of regional metamorphism and intensive structural deforma-
tion resulted in excellent permeability, both in the high porosity ma-
trix of turbidite sand beds and along decollement-related structures.
Extension-induced retrograde melting along deep through-going re-
gional fault zone tapped deep-seated reservoirs, with considerable
involvement of the Pedawan Formation. A batholithic intrusion, the
Bau Anomaly, developed along a re-activated ENE structure, and
was located 1.2 km below the present surface. The largest deposit
at Bau, the Tai Parit orebody with over 1 million ounces of gold re-
covered, associated with porphyry microgranodiorite developed at
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
the intersection between the ENE D1 uniform Bau limestone and the
NNE D4 Bau Trend (Schuh, 1993).
There is a strong structural control on essentially all ore deposit
types in the Bau district (Dill and Horn, 1996; Percival et al., 1990;
Sillitoe and Bonham, 1990). Four deformation events (D1 to D4) can
be identified in the structures of the Bau district. The resulting struc-
tures are complex. Structures of the D1 event can be recognized in
several places where they are cut by the regional ENE Bau trend.
Two major anticlines, the Bau Anticline to the north, and the Ropih
Anticline to the south are the most prominent NE-trending features
of the NW–SE compressive D3 event. Intersections of anticlines with
anticlines of the D2 and D3 are thought to have produced structural
and topographic highs, while the additions of two synclines produced
lows by superposition of orthogonal sets of folds (Marjoribanks, 1989;
Wolfenden, 1965).
The NNE trend that developed during the D4 event is characterized
by two dominant faults, the Tai Parit and Totag faults. Numerous apoph-
yses and dikes occupy fractures extending outwards from the intrusions
towards the NW and SE. The spatial distribution of mineralization, felsic
intrusions and openings created by the D4 transtensional shear
zones all indicated their coeval and co-genetic development (Williams
et al., 1988).
3. Materials
3.1. Remote sensing data
Landsat Enhanced Thematic Mapper+ (ETM+), Hyperion and Phased
Array type L-band Synthetic Aperture Radar (PALSAR) data sets were
used in this investigation. Landsat TM/ETM+ data have been widely
used for geological mapping applications for many years (Abdelsalam
and Stern, 2000; Aydal et al., 2007; Crosta and Moore, 1989; Okada
et al., 1993; Podwysocki et al., 1984; Rajesh, 2008; Ranjbar et al.,
2011; Rowan et al., 1977; Zhang, 2010). Unfortunately, no suitable
ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiom-
eter) data could be found, due to persistent cloud cover in the area. Thus,
only Landsat ETM+ data and Hyperion data were processed for hydro-
thermal alteration mapping in this research.
The Landsat ETM + image (Path/Row 121/59) was obtained from
the U.S. Geological Survey's Earth Resources Observation System
(EROS) Data Center (EDC). It was acquired on August 3, 2000 during
dry season with only 5% cloud cover for the Bau mining district and
surrounding areas. The images were pre-georeferenced to a UTM zone
40 North projection using the WGS-84 datum.
Hyperion is the only available satellite with a hyperspectral sensor,
which has 242 spectral channels from 0.4 to 2.5 μm over a 7.6 km
swath width, and 30 m spatial resolution. The first 70 bands are in the
visible and near infrared (VNIR; 0.4 to 1.0 μm) and the next 172 bands
in the shortwave infrared (SWIR; 0.9 to 2.5 μm) (Barry and Pearlman,
2001; Folkman et al., 2001; Pearlman et al., 2003). The VNIR infrared
bands between 0.4 and 1.3 μm can be used to highlight iron oxide min-
erals (Bishop et al., 2011), whereas the SWIR bands (2.0 to 2.5 μm) can
be used to identify and map hydroxyl-bearing minerals, sulfates and car-
bonates in hydrothermal alteration assemblages (Bishop et al., 2011;
Gersman et al., 2008; Kruse et al., 2003).
A cloud-free level 1B Hyperion image (EO1H1210592004230110PZ)
obtained from the EROS Data Center was acquired on August 28, 2004
during the dry season, and was also pre-georeferenced to the UTM
zone 40 North projection using the WGS-84 datum. The Hyperion data
were processed for hydrothermal alteration mapping at a district scale
in this study.
PALSAR, which is an active microwave sensor onboard the Advanced
Land Observing Satellite (ALOS), is not affected by weather conditions
and operable both daytime and nighttime (Igarashi, 2001; Rosenqvist
et al., 2004). It was launched on January 24, 2006. The PALSAR sensor
is an L-band synthetic aperture radar, with multi mode observation
185
functions (Fine mode, Direct downlink, ScanSar mode, and Polarimetric
mode), including multi polarization (HH, HV, VH, and VV), variable
off-nadir angle (9.9 to 50.8°), and spatial resolution (10 m, 30 m,
100 m for Fine, Polarimetric, and ScanSar modes, respectively) and
swath width observation (30 km, 70 km, and 250–350 km for Polari-
metric, Fine and ScanSar modes, respectively) (Igarashi, 2001).
In the tropics, synthetic aperture radar data are particularly appro-
priate, because microwave signals can penetrate the persistent cloud
cover. Structural geology is enhanced by the side looking aspect of the
active SAR sensor, which produces radar “shadows” highlighting
micro relief.
A PALSAR Fine mode Level 1.5 scene (PASL1501103141522141209
100014) acquired on 14 March 2011 was obtained from the Earth and
Remote Sensing Data Analysis Center (ERSDAC) in Japan. A Level 1.5
product scene includes high resolution mode by single polarization
(HH or VV) and 6.25 m pixel spacing, and has an off-nadir angle of
34.3°. The imagery was geo-referenced and geo-coded.
The Landsat ETM +, Hyperion, and PALSAR data sets of the Bau area
were processed using the ENVI (Environment for Visualizing Images)
version 5.0 software package.
3.2. Field reconnaissance and verification of image processing results
A field reconnaissance was carried out to verify the image process-
ing results during 19 to 22 August 2012. Locations were recorded by a
Garmin® eTrex Legend®H GPS. Ground photos of the geomorpholo-
gy, rock units, open-pit quarry, tunnels, and hydrothermally altered
rocks in the Bau and surrounding areas were taken. Samples for labo-
ratory studies were collected through a systematic collection of fresh
and weathered surfaces of representative hydrothermally altered rocks.
The XRD analyses were implemented on bulk powder using an X-ray dif-
fractometer D8ADVANCE model at the Material laboratory in Universiti
Teknologi Malaysia (UTM). Spectral reflectance measurements were
made using an Analytical Spectral Devices (ASD) Spectroradiometer
Fieldspac® model, which records a reflectance spectrum across an over-
all spectral range of 325–2500 nm (nanometer) with a 10 nm individual
band width. The spectral measurements were performed in the Remote
Sensing laboratory at the Universiti Teknologi Malaysia (UTM) using
a contact probe and a built-in illumination source. In addition, the
image processing results were compared with the geological map
(1:25,000 scale), a Quickbird image map of the gold prospects locations
(ore outline), and a geophysical map of 450 Hz resistivity (1:25,000
scale) of the Bau area. These additional data sources were provided by
the Olympus Pacific Minerals Inc. Company in 2012 (North Borneo
Gold SDN BHD).
4. Methods
The main purpose of the methodology adopted was to apply image
processing techniques that could reduce or remove the spectral effects
of vegetation in lithologic and hydrothermal alteration mapping in a
the tropical terrain, and the detection of structure elements associated
with fault-controlled and stratiform gold orebodies.
4.1. Preprocessing of remote sensing data
Atmospheric correction was applied to the ETM + and Hyperion
data sets to eliminate the spectral effects of water vapor and aerosols.
During the atmospheric correction, raw radiance data from the imaging
spectrometer was re-scaled to reflectance data. As a result, all spectra
are shifted to nearly the same albedo. The resultant spectra can be di-
rectly compared with the reflectance spectra of the laboratory or field
spectra (San and Suzen, 2010). The ETM+ data were converted to reflec-
tance using the Fast Line-of-sight Atmospheric Analysis of the Spectral
Hypercubes (FLAASH) algorithm. This algorithm contains the standard
MODTRAN model atmospheres, which use the standard column water
186 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
vapor amounts for each model atmosphere. Moreover, scene and sensor
information (solar zenith angle, satellite view angle, and relative azimuth
angle between the satellite and sun), and aerosol model are considered
when running this algorithm (Thome et al., 1998).
Spectral bands covering VNIR and SWIR ranges of Hyperion were
selected for use in this study. Preprocessing included the removal of
overlapping and inactive bands (1–7, 58–81, 164–185 and 225–242),
destriping and atmospheric correction (Gersman et al., 2008). After ra-
diometric corrections, there is still a pronounced vertical striping pattern
in the Hyperion data, such striping is often seen in data acquired using
pushbroom technology (Kruse et al., 2003). Destriping of the Hyperion
Level 1B data was accomplished before atmospheric correction. To
correct the atmospheric effects, Atmospheric CORrection Now (ACORN)
software was used to retrieve the surface reflectance (ACORN TM 5.0,
2004; Kruse et al., 2003).
A property of SAR imagery is the presence of speckle. The presence
of speckle in an image reduces the detectability of ground targets,
obscures the spatial patterns of surface features, and decreases the
accuracy of automated image classification. Therefore, for the SAR data
speckle suppression filter was applied that retains edges and preserves
features (linear features and point features). The adaptive Lee filter was
applied to the PALSAR image to achieve this (Lee, 1980). It is based on
the multiplicative model of speckle noise, which is at first approximated
by a linear model. Then the minimum mean square error criterion was
applied to this linear model (Lee, 1980; Sheng and Xia, 1996; Shi and
Fung, 1994; Sveinsson and Benediktsson, 1996).
4.2. Image processing methods
Spectral signatures observed in tropical remote sensing imagery
are related almost entirely to vegetation. Therefore, the separation
and suppression of vegetation effects are significant to test the appli-
cation of optical remote sensing data for detecting mineralogy of soils
and rocks in areas of the tropical/sub-tropical environment. To accom-
plish this an image enhancement based on a directed principal compo-
nents (DPC) analysis of four band ratio images (Fraser and Green,
1987) was implemented using the following indices: vegetation index
(band ratio of 4/3), clay minerals index (band ratio of 5/7), ferric iron
oxide index (band ratio of 3/1), and ferrous iron oxide index (band
ratio of 5/4) of ETM+ data. The input band ratio images are selected on
the basis that one band ratio contains information related to the compo-
nent of interest (e.g. hydrothermal alteration minerals), the spectral
response of which suffers interference from the spectral response of
another component (i.e. vegetation). Band ratio 4/3 serves as an excel-
lent vegetation indicator while band ratios of 3/1, 5/4 and 5/7 are com-
monly used to enhance iron oxide and clay minerals, respectively.
The DPC, which has loadings of similar signs on both input band
ratio images, explains the variance due to similarities in the spectral
responses of the interfering component and the component of inter-
est. The other DPC, whose loadings are of different signs on either of
the input band ratio images, highlights the contributions unique to
each of the components. The sign of the loadings dictates whether
the component of interest is represented as bright or dark pixels in
the DPC image; a positive loading implies bright pixels while a nega-
tive loading implies dark pixels (Fraser and Green, 1987).
The image eigenvectors and eigenvalues were obtained from the
DPC analysis using a covariance matrix of the indices. The outputs
of DPC analysis, presented as tables of statistic factors, and selected
DPC images from these transformations have been reproduced in
figures to support the discussion. The statistical results are presented
in Table 1.
Linear Spectral Unmixing (LSU) was applied to a selected spatial
subset scene of Hyperion data covering the Bau gold mining district,
for mineral mapping purposes at the district scale. The LSU was ap-
plied on the VNIR and SWIR bands of Hyperion for mapping iron
oxide/hydroxide minerals and clay mineral assemblages associated
Table 1
Principal component analysis on band ratio indices of ETM+, selected spatial subset
scene covering the Bau gold mining district.
Input band Band ratio Band ratio Band ratio Band ratio Eigenvalues
ratio indices 4/3 5/7 3/1 5/4 (%)
PC1 0.677 0.434 0.413 0.427 81.23
PC2 0.730 –0.419 –0.267 –0.470 15.67
PC3 –0.097 0.870 –0.329 –0.352 2.14
PC4 0.027 –0.326 0.576 0.686 0.96
with gold mineralization. The LSU method was used to determine
the relative abundance of materials depicted in the multispectral or
hyperspectral imagery based on the materials' spectral characteristics
(Shimabukuro and Smith, 1991). The reflectance at each pixel of the
image is assumed to be a linear combination of the reflectance of
each material (or end-member) present within the pixel. This tech-
nique, also known as sub-pixel sampling, or spectral mixture analysis,
is a widely used procedure to determine the proportion of constituent
materials within a pixel based on the materials' spectral characteris-
tics (Boardman, 1989, 1992).
Two spectral subsets of Hyperion data were analyzed separately to
detect iron oxide/hydroxide minerals and hydroxyl-bearing (clay)
alteration mineral assemblages. The first subset (VNIR) covering 90
bands (excluded overlapping and inactive bands) between 0.4 and
1.3 μm was used for highlighting iron oxide/hydroxide minerals,
and the second subset (SWIR) of 40 bands between 2.00 and 2.40 μm
(185 to 225) was processed for detecting hydroxyl-bearing (clay) alter-
ation minerals. Analytical Imaging and Geophysics (AIG)-developed
hyperspectral analysis processing methods (Kruse and Boardman,
2000; Kruse et al., 2003) were used to extract end-member spectra
from the Hyperion subsets. The vegetation end-member spectrum has
been removed from the output, in order to suppress the erroneous
effects of vegetation in hydrothermal alteration mapping. In fact, the re-
sultant spectral unmixing bands with high erroneous effects of vegeta-
tion were not used for the subsequent alteration mapping analysis, in
order to visualize the altered areas in the final image map.
The second moment co-occurrence texture filter was applied to
the Lee-filter resultant PALSAR image for extracting the structural
elements (anticline, faults and lineaments associated with gold min-
eralization) at both regional and district scales. The method is based
on the co-occurrence matrix, including mean, variance, homogeneity,
contrast, dissimilarity, entropy, second moment and correlation. The
co-occurrence texture filter uses a gray-tone spatial dependence matrix
to calculate the texture values. This is a matrix of relative frequencies by
which pixel values occur in two neighboring processing windows, sep-
arated by a specified distance and direction. It shows the number of oc-
currences of the relationship between a pixel and its specified neighbor
(Anys et al., 1994; Park and Chen, 2001). We used the pixels in a 3 ∗ 3
base window and also the pixels in a 3 ∗ 3 window that was shifted
by 1 pixel for implementing the co-occurrence matrix. A high gray
scale 64 setting was used to reduce the shade and highlight the dark
pixels of fracture zones.
5. Results and discussion
From the results of directed principal component analysis trans-
formation for DPC analysis of specific spectral indices of ETM + data
(considering magnitude and sign of the eigenvector loadings and per-
centage of eigenvalues for selected spatial subset scene covering the
Bau gold mining district), it is realized that the first principal component
(DPC1) accounts for 81.23% of the total eigenvalues (Table 1). DPC1 con-
tains most of the spectral information in the scene (Jensen, 2005). All of
the eigenvector loadings for the DPC1 are positive (Table 1), thus the
differentiation between materials (vegetation and alteration minerals)
using the specific spectral indices in the DPC1 image is not possible.
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196 187

Fig. 4. A panoramic view of vegetation cover in the Bau area.

The vegetation index 4/3 band ratio is very sensitive to vegetation
amount and relatively insensitive to variations in lithology. The clay
mineral index 5/7 band ratio generally varies with the abundance of
hydroxyl-bearing minerals. However, the clay mineral index can also
be affected by the amount of vegetation (Crippen and Blom, 2001).
Eigenvector loadings for the DPC2 indicate that DPC2 image has
the highest value of vegetation variations, which is statistically dom-
inant in the image due to the high positive contribution from the
vegetation index (0.730) and negative contribution from the specific
hydrothermal mineral indices (Table 1). Fig. 4 shows a panoramic

Fig. 5. A regional view of unveiling the lithology in the background of DPC3 image.
188 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
view of the vegetation cover in the Bau area. The existence of vegeta-
tion within a pixel can reduce the 2.20 μm absorption depth related
to AlOH/MgOH-content (Bierwirth et al., 2002; Galva˜o et al., 2005;
Rodger and Cudahy, 2009).
The DPC3 image shows the desired information relating to Al
(OH)-bearing (clay) minerals with very limited disturbance of vege-
tation, due to the very low negative (contribution) eigenvector load-
ings for the vegetation index (− 0.097) but with very strong positive
eigenvector loadings for the clay mineral index (0.870). The eigen-
vector loadings for the ferric iron oxide index (3/1) and ferrous iron
oxide index (5/4) are (− 0.329) and (− 0.352), respectively. Hence,
the contribution of iron oxide minerals in the extracted spectral infor-
mation in the DPC3 image is very low. Fig. 5 shows a regional view of
enhancing the lithology in the background of the DPC3 image of the
study area. Geological structures (Bau anticline), bedrock lithology,
hydrothermally rocks (as bright pixels), alluvial patterns and rivers
can be easily distinguished in the DPC3 image (Fig. 5).
Fig. 6 is a spatial subset covering the Bau gold mining district and
surrounding terrain. Hydrothermally altered rocks are depicted by
red pixels. Most of the detected altered zones are well-known pros-
pects indicated by numbers in Fig. 6 (Stevens and Fulton, 2010).
These are the Jugan (110° 11′ 37″ E, 1° 26′ 10″ N), Bukit Sarin (110°
10′ 12″ E, 1° 26′ 09″ N), Sirenggok (110° 09′ 23″ E, 1° 25′ 35″ N),
Jumbusan (110° 11′ 48″ E, 1° 24′ 16″ N) and Kapor (110° 06′ 58″ E,
1° 19′ 21″ N) prospects. Precise GPS measurements taken over and
around known prospects emphasized the accuracy of the identified
Fig. 6. Selected spatial subset scene covering the Bau gold mining district and surrounding terrains. Hydrothermally altered rocks (clay minerals) are indicated as red color.
prospect locations in Fig. 6. Some GPS sample points were also col-
lected over areas of unaltered rocks that do not carry mineralization.
Some identified potentially interesting areas for gold mineralization
warrant further investigation. However, these potentially interesting
areas have only very small rock exposures relative to the scene size.
Several other new areas with small rock exposures can be seen in the
north/northeast and scattered in the background of the selected spatial
subset scene. Two new potentially interesting areas with large rock ex-
posures have been detected in the southeastern part of the scene (Fig. 6).
It is clear that iron oxide minerals can be distinguished in the DPC4
image. Vegetation has spectral characteristics similar to iron oxide min-
erals in the VNIR from 0.7 to 1.2 μm (Jensen, 2005; Ruiz-Armenta and
Prol-Ledesma, 1998). Vegetation shows absorption features from 0.45
to 0.68 μm and high reflectance in the near infrared. It is observed
that iron oxide/hydroxide minerals have high reflectance in the range
of 0.63 to 0.69 μm, while the range of 0.76 to 0.90 μm covers higher
range of the vegetation red-edge reflectance features in the near infra-
red. This characteristic can be used to differentiate iron oxide/hydroxide
minerals from vegetation (Crosta and Moore, 1989; Ruiz-Armenta and
Prol-Ledesma, 1998).
Iron oxides could be mapped in the DPC4 image because of the
strong positive contribution from ferric iron oxide index (0.576) and
ferrous iron oxide index (0.686) and negative weighting of clay mineral
index (−0.326) for the selected spatial subset scene covering the Bau
gold mining district and surrounding terrain (Table 1, Fig. 7). Iron-rich
areas are shown as purple pixels, and the distribution of moderate
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
iron oxide soils is depicted as light green pixels. Although positive con-
tributions from the vegetation index (0.027) can also be seen in the
DPC4 image, this is evidence for squat disturbance of vegetation effects
in the image.
Iron oxide/hydroxide minerals produced during supergene alter-
ation are among the important mineral groups associated with hydro-
thermally altered rocks over porphyry copper–gold deposits (Sabins,
1999). Interaction with air and surface water oxidizes sulphides, yield-
ing a Fe-rich crust termed gossan (Abdelsalam and Stern, 2000; Sato,
1974). In tropical regions, laterites with ubiquitous bright red soil are fa-
miliar elements of the landscape. Primary minerals are transformed into
characteristic assemblages of secondary minerals, including kaolinite,
hematite, goethite and residual quartz. The outcome of lateritization is
a build-up of iron and aluminum oxides at the top of the paleosol
profile, resulting in the formation of ferricretes (iron crust or duricrust)
in tropical and subtropical terrains (Butt and Zeegers, 1992). Therefore,
iron oxides in tropical soils may have large specific surface areas (Osei and
Singh, 1999), and hence their presence, even in small amounts, could
have a great influence on the remote sensing results (Deller, 2006).
Fig. 7 shows the selected spatial subset scene covering the Bau
gold mining district and surrounding terrain. Iron oxides are illustrated
as iron rich zones or gossan in purple and moderate iron oxides soils or
laterites in light green. Most of the detected iron oxides zones in Fig. 7
are moderate iron oxide soils or laterites especially in the center parts
of the image, where there is no reported prospect or mineralized area.
Fig. 7. Selected spatial subset scene covering the Bau gold mining district and surrounding terrains. Iron oxide minerals are indicated as purple (iron-rich) and green (iron-fair)
colors.
189
Several well-known prospects such as Jugan, Bukit Sarin, Sirenggok,
Jumbusan and Kapor are associated with iron-rich rocks or gossan,
which are indicated by numbers in Fig. 7. Potentially interesting areas
have also been detected in the northeastern and southeastern parts of
the scene. Prior to the opening of the pit of the gold deposits in the
Bau area, the mining areas were covered by iron-rich rocks or gossan,
which presented the surficial expression (Schuh, 1993). Fig. 8 shows
the association of gossan with the surficial exposure of the Jugan pros-
pects. Bau gold deposits occur near or adjacent to areas of silicification
and argillically altered zones in the Bau Limestone and the Pedawan
Formation. The alteration mineralogy includes kaolinite, dickite, illite,
minor sericite, calcite, quartz, marcasite, and pyrite. Iron oxides are
prevalent in the weathered altered rocks (Percival et al., 1990).
This approach detected hydrothermal alteration zones associated
with hydrothermal activity of igneous intrusions, which are potential
areas for gold mineralization in the Bau mining district. It provides
image maps of alteration minerals and highlights the lithology of
the background with minimizing the spectral effects of vegetation.
The linear spectral unmixing technique was also applied to the se-
lected spatial subset scene of the Hyperion data set covering the Bau
gold mining district for detailed hydrothermal alteration mapping at
the district scale. The end-member spectra were extracted using the
AIG-developed hyperspectral analysis processing methods, and
then compared with the USGS spectral library as reference spectra.
Figs. 9 and 10 show the extracted end-member spectra for two
190 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Fig. 8. Association of iron-oxide-rich rocks (gossan) with surficial exposure of the Jugan prospect.
spectral subsets, excluding the vegetation end-member spectra.
Considering the shape and position of the absorption feature, the
minerals are characterized as hematite–limonite–goethite for the
first subset (Fig. 9). The extracted signatures for the second subset
suggest the existence of hydroxyl minerals such as kaolinite–sericite
(Fig. 10).
Goethite, hematite and limonite have strong Fe3+ absorption features
at 0.48 μm and 0.83–0.97 μm (Hunt et al., 1971; Fig. 9). Sericitically-
altered rocks typically contain sericite, a fine-grained form of muscovite
that has a distinct Al\OH absorption feature at 2.2 μm and a less intense
absorption feature at 2.35 μm (Abrams and Brown, 1984; Spatz and
Wilson, 1995; Fig. 10). Kaolinite is a typical mineral of advanced argillic
alteration that exhibits Al\OH 2.165 μm and 2.2 μm absorption features
(Hunt, 1977; Hunt and Ashley, 1979; Rowan et al., 2003; Fig. 10).
Fig. 11 shows an image map of the selected spatial subset scene for
first subset (VNIR), showing spectrally predominant iron oxide minerals
as colored pixels that are overlaid on the gray-scale image background of
Hyperion band 36. The abundance of iron oxide minerals is represented
as green (high abundance) and magenta pixels (low abundance) in the
Bau subset scene (Fig. 11). Iron oxide concentrations are obviously asso-
ciated with lineament structures as well as known prospects in the Bau
area. Most of the gold prospects such as Boring (110° 07′ 11″ E, 1° 22′
Fig. 9. Extracted mineral signatures (goethite–hematite–limonite) from VNIR bands of
Hyperion data.
39″ N), Kapor, Umbut (110° 07′ 33″ E, 1° 23′ 59″ N), Sirenggok, Jugan
and Bukit Sarin are areas associated with high abundance of iron oxides,
although some of the prospects are associated with low abundance
of iron oxides (Taiton (110° 81′ 22″ E, 1° 23′ 21″ N), Bukit Young
(110° 09′ 28″ E, 1° 24′ 38″ N), Tai Parit (110° 09′ 18″ E, 1° 24′ 52″ N),
Bekajang (110° 10′ 13″ E, 1° 24′ 13″ N) Pejiru (110° 05′ 48″ E, 1° 22′
05″ N)). Some potentially interesting areas with high abundance of
iron oxide minerals are detected in the central and western parts of
the scene (Fig. 11). As mentioned before, iron-oxide-rich rocks or gos-
san are associated with known mines before the opening of the pit of
the gold deposits in the Bau area, hence, new indentified gossan zones
can be considered for future exploration. The XRD analysis indicated
that hematite, arsenopyrite, limonite and goethite are major minerals
in samples of iron-oxide-rich rocks or gossan collected from the field.
Fig. 12 shows image map of the selected spatial subset scene for
the second subset (SWIR). Kaolinite–sericite abundance is illustrated
as red (high abundance) and yellow (low abundance) pixels that are
overlaid on the gray-scale image background of Hyperion band 205.
Detected pixels are associated with fault and fracture structures in
the Bau scene. Most of the gold prospects, including Kapor, Taiton,
Umbut, Bukit Young, Tai Parit, Bukit Sarin, Jugan, Bekajang, Jambusan
are located in areas with high abundance clay shown by numbers in
Fig. 10. Extracted mineral signatures (kaolinite–sericite) from SWIR bands of Hyperion
data.
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Fig. 11. Image map of VNIR bands of Hyperion data showing the abundance of iron oxide minerals in the Bau gold mining district. Gold prospects and potentially interesting areas
are marked by numbers.
the scene (Fig. 12). Potentially interesting areas are located in the
center, western and northern parts of the scene (Fig. 12), which are
scattered along the SSW to NNE structural elements. Field and geological
investigations indicate that they are associated with advanced argillic al-
teration, decalcification, brecciation, and silicification of the sedimentary
host rocks (Percival et al., 1990; Schuh, 1993; Stevens and Fulton, 2010).
Fig. 13 shows the association of advanced argillic alteration with
ore mineralization (gold-bearing arsenopyrite) in the Jugan prospect.
According to XRD analysis of collected rock samples from advanced
argillic alteration zones, the minerals detected in these altered rocks
include kaolinite, dickite, illite, limonite, sericite, calcite and quartz.
The average spectra measurements of collected rock and soil samples
from hydrothermally altered rocks associated with the gold deposits and
surface geologic materials from non-mineralized zones in the Bau area are
shown in Fig. 14. The reflectance spectra of the collected samples from
iron-oxide-rich areas (gossan) exhibit strong absorption features in the
VNIR, centered at 450 nm and 850 to 950 nm (Fig. 14). Reflectance curves
exhibit similarities to spectra of hematite, goethite and limonite. Strength
191
of the 0.9 μm absorption and amount of drop off toward 0.35 μm are re-
lated to the total concentration of iron oxide/hydroxide minerals present.
When both hematite and goethite are present in significant proportions,
the combination of these two features produced absorption with its min-
imum between 0.85 μm and 0.93 μm (Hunt and Ashley, 1979).
Spectra from moderate iron oxides and clay soils (laterites) show
distinctive absorption features in the VNIR, centered at 480, 700, 850
to 950 nm, as well as less pronounced ones in the SWIR, centered at
2160 and 2200 nm (Fig. 14). Results derived from the samples indicate
the presence of hematite, goethite and kaolinite. The mere appearance
of the 2.2 μm absorption feature shows the presence of clay minerals.
Clear double sharp features near 2.16 and 2.2 μm emphasize the existence
of kaolinite, which are attributed to vibrational processes associated with
Al\OH bonds (Clark et al., 1990; Hunt, 1977; Hunt and Ashley, 1979).
Samples from argillically-altered areas show strong absorption
features in the SWIR, centered at 2170, 2200 and 2350 nm (Fig. 14),
typical of clay minerals (kaolinite, dickite, illite, and sericite). All the
absorption features that appear at wavelengths longer than 1.3 μm are
192 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Fig. 12. Image map of SWIR bands of Hyperion data showing the abundance of clay minerals in the Bau gold mining district. Gold prospects and potentially interesting areas are
marked by numbers.
caused by vibrational processes that occur in the hydroxyl-bearing min-
erals that the rock contains, and quite typically the features of just one
of these minerals will dominate the spectrum (Hunt, 1977; Hunt and
Ashley, 1979). Two diagnostic absorption features appear near 1400
and 1900 nm in the average spectra of all the samples measured.
Whenever molecular water is present in minerals and rocks, two fea-
tures always appear: one near 1.4 μm and the other near 1.9 μm. They
may be broad, indicating that the molecules occupy unordered, or
several non equivalent sites, but the presence of these two features to-
gether is completely diagnostic of the presence of water molecules
(Hunt, 1977).
Accordingly, the LSU method could detect the alteration minerals
using the extracted end-member minerals spectra directly from the
Hyperion image at a district scale in the tropical regions. The results
derived from VNIR and SWIR bands of Hyperion compare quite well
with each another, and most of the detected hydrothermal alteration
zones are spatially associated with fault and fracture structures in the
Bau gold mining district. In this regard, the co-occurrence texture fil-
ter was used to identify the edges of surrounding pixels in the PALSAR
image to highlight the fault and lineament structures. Hence, geolog-
ical structures were detected in the study area at both regional and
district scales using the PALSAR image. Fig. 15 shows the geological
structures, including Bau anticline, Bau anticline axis, Tai Parit fault
and several linear and curvilinear features in the Bau and surrounding
areas at a regional scale. The Bau anticline (NE–SW trending) and its
axis are located in the south and southwestern parts of the PALSAR
image (Fig. 15). The Tai Parit fault (NNE-trending) and several linear
and curvilinear features (WNW-to NW-trending) dissected the Bau
limestone, which are identifiable in the north western part of the scene
(Fig. 15).
A major regional lineament, the 200-km-long NNE-striking Bau
Trend, was the locus of Mid-Miocene intrusions. All major ore de-
posits of the Bau district have been found along this trend. Explora-
tion away from the trend was generally less successful (Dill and
Horn, 1996; Schuh, 1993; Sillitoe and Bonham, 1990). The intrusions,
associated dikes, and ore deposits are spatially associated with sites
of pronounced structural weakness along the NNE Bau Trend. These
sites are: (i) intersection of Dl with D4 lineaments; (ii) intersections of
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Fig. 13. Association of advanced argillic alteration with ore mineralization (gold-bearing arsenopyrite) in the Jugan prospect.
lineaments with folds; (iii) intersections of folds; and (iv) a combination
of lineament intersection with two folds. The district-wide asymmetric
distribution of ores with numerous deposits on the western side, but
few deposits on the eastern side of the district, is attributed to a more
dilational stress field west of the Bau Trend. Evidence for open fracture
zones are numerous NW-trending dikes linking the intrusions with the
NNE fault zones (Schuh, 1993).
Several prominent faults are exposed for 3 to 4 km within and
parallel to the trend of the Bau gold deposits (e.g. Tai Parit Fault,
Fig. 2) and represent exposures of the deep-seated structural zone.
Most gold deposits in the Bau district lie within the area where this
northeast-trending structural zone intersects the axial zone of the
east-to northeast-trending Bau anticline (Fig. 2) (Percival et al.,
1990). Fig. 16 illustrates the selected spatial subset scene of PALSAR
image, showing the fault, lineament structures and the Bau minerali-
zation trend at a district scale.
The regional and local structures influenced the location of the
mineral deposit mineralization because they acted as major conduits
for hydrothermal fluids. Therefore, faults and fractures served as a
primary structural control on ore emplacement in the Bau area. Previous
geological studies verified that the mineralized areas are cut by a net-
work of NW and NE trending faults and joints, along which mineralizing
Fig. 14. Laboratory reflectance spectra of altered rock samples. Arrows point a maxi-
mum absorption features.
193
fluids have passed upward into the calcareous host rocks (Dill and Horn,
1996; Percival et al., 1990; Schuh, 1993).
Intersections of ENE and NNE structures were favorable sites for
intrusions and mineralization. Differential movement along several
fault strands of the D4 Bau trend with an overall right-lateral sense
produced locally NE–SW-directed extension. The resulting pathways
were occupied by hydrothermal mineralization and by numerous
dikes (Schuh, 1993). The entire Bau district is anomalous in gold, arsenic,
and antimony, while anomalous concentrations of other elements are re-
stricted to particular types of ore deposits. Porphyry-copper, skarn, and
Cordilleran vein-type base metal deposits occur in a narrow, NNE zone,
following essentially the Bau Trend of Miocene intrusions. Antimony
and gold-rich epithermal ores occur in three ENE belts across the district.
Arsenic-rich epithermal ores occur in the western areas (Dill and Horn,
1996; Kim, 1994; Schuh, 1993).
Gold mineralization at Bau is structurally controlled, for this reason a
detailed detection of geological structures is very useful in identifying
potentially interesting areas in this region and surrounding terrain.
Fig. 16 shows lineament map of the Bau mining district. It is evident
that hydrothermal fluids have passed upward into the structure ele-
ments and altered the carbonate host rocks. Therefore, the potentially
mineralized areas intersect ENE and NNE structural elements and are
the recommended areas for detailed ground investigations for gold
exploration.
6. Conclusions
This study demonstrates the importance and advantages of the
combined use of the ETM +, Hyperion and PALSAR remote sensing
data sets in detecting hydrothermal alteration zones and structural
elements associated with gold mineralization in a tropical/sub-tropical
region, at Bau goldfield Sarawak, Malaysia. Intrusion-related gold min-
eralization indicators, including iron oxides, clay minerals and structur-
al elements in the Bau gold mining district have been detected using the
remote sensing satellite data sets and the approach used in this investi-
gation. Results of the DPC analysis of four appropriate ETM+ band ratios
produced DPC images, allowing the removal of the spectral effects of
vegetation from ETM+ data and the detection of separate mineral
images. DPC3 image revealed the lithology of the background and
detected clay-rich zones. DPC4 image identified moderate iron
oxide soils and iron-oxide-rich rocks or gossan. The results of the
DPC analysis of four band ratio images indicate that this algorithm
194 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196

Fig. 15. Regional view of the geological structures in the Bau and surrounding area.

Fig. 16. Selected spatial subset scene of PALSAR image showing local structure elements, including fault and lineament structures and the Bau gold mineralization trend at district
scale.
 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
could achieve better results than the Crosta (Crosta and Moore, 1989)
and software defoliant (Fraser and green, 1987) techniques for detecting
alteration associated with hydrothermal activity in tropical/subtropical
regions.
The LSU technique can be used to produce image maps of hydroxyl-
bearing minerals using Hyperion data at a district scale in tropical re-
gions. Results derived from VNIR and SWIR bands of Hyperion repre-
sented iron oxide/hydroxide and clay-rich zones associated with the
known gold prospects in the Bau district. Field reconnaissance and
laboratory studies, including accurate GPS surveying, XRD analysis
and spectral reflectance measurements of collected rock samples
verified the optical remote sensing results. Many of the mapped
iron-rich (gossan) and clay mineral zones coincide spatially with the
known prospects and mines, and few of them coincide with probable
non-vegetated ground.
PALSAR data can be very useful to highlight general and detailed
geological structures, including anticline, faults, joints, and fractures
associated with structurally controlled gold mineralization in tropical/
subtropical regions. The potential mineralized areas associated with
the intersection of ENE and NNE structure elements are recommended
for detailed ground investigations for gold exploration in the Bau area.
The results of this study indicate that hydrothermal alteration zones
and structural elements associated with hydrothermal gold mineral-
ization can be detected in tropical/sub-tropical regions using ETM +,
Hyperion and PALSAR remote sensing data sets and the data integra-
tion approach used in this research.
Acknowledgments
This study is conducted as a part of post-doctoral fellowship scheme
granted by Universiti Teknologi Malaysia (UTM). We acknowledge the
assistance of the Olympus Pacific Minerals Inc. Company (North Borneo
Gold SDN BHD) for their logistic support during the field investigations
and ground truth data collection, as well as appreciate their assistance
in various other ways during this research. The authors would like to
express their sincere thanks to Murray Stevens (Geology manager of
North Borneo Gold) for helpful discussions and his continuous support
for this study. Sincere thanks to Rubi Dihom (Community liaison officer
of North Borneo Gold), David G. Gotera (Geologist of North Borneo
Gold), Muhammad Izuan Bin Nadzri and a Sarawakian guide for very
helpful support during the geological field work. We also would like
to express our great appreciation to Prof. Dr. John Carranza and the
anonymous reviewers for their very useful and constructive comments
and suggestions for improvement of this manuscript.
References
Abdelsalam, M., Stern, R., 2000. Mapping gossans in arid regions with Landsat TM and
SIR-C images, the Beddaho Alteration Zone in northern Eritrea. J. Afr. Earth Sci. 30
(4), 903–916.
Abrams, M.J., Brown, D., 1984. Silver Bell, Arizona, porphyry copper test site report:
Tulsa, Oklahoma, The American Association of Petroleum Geologists. The Joint
NASA–Geosat Test Case Project, Final Report, Chapter 4. 4–73.
ACORN TM 5.0, 2004. Tutorial, ImSpec LLC, Advanced Imaging and Spectroscopy.
ImSpec LLC.
Andriesse, J.P., 1972. The soils of West-Sarawak East-Malaysia. East Malaysia Memoir I,
vol. 1. Department of Agriculture, Sarawak.
Anys, H., Bannari, A., He, D.C., Morin, D., 1994. Texture analysis for the mapping of
urban areas using airborne MEIS-II images. Proceedings of the First International
Airborne Remote Sensing Conference and Exhibition, Strasbourg, France, vol. 3,
pp. 231–245.
Arehart, G.B., 1996. Characteristics and origin of sediment-hosted disseminated gold
deposits: a review. Ore Geol. Rev. 11, 383–403.
Aydal, D., Ardal, E., Dumanlilar, O., 2007. Application of the Crosta technique for alter-
ation mapping of granitoidic rocks using ETM + data: case study from eastern
Tauride belt (SE Turkey). Int. J. Remote Sens. 28 (17), 3895–3913.
Bagby, W.C., Berger, B.R., 1985. Geologic characteristics of sediment-hosted, disseminated
precious-metal deposits in the western United States. In: Berger, B.R., Bethke, P.M.
(Eds.), Geology and Geochemistry of Epithermal Systems: Society of Economic
Geologists: Reviews in Economic Geology, vol. 2, pp. 169–202.
195
Barry, P.S., Pearlman, J., 2001. The EO-1 Mission: Hyperion Data. National Aeronautics
and Space Administration (NASA) 35–40.
Bierwirth, P., Huston, D., Blewett, R., 2002. Hyperspectral mapping of mineral assemblages
associated with gold mineralization in the Central Pilbara, Western Australia. Econ.
Geol. 97 (4), 819–826.
Bishop, C.A., Liu, J.G., Mason, P.J., 2011. Hyperspectral remote sensing for mineral explora-
tion in Pulang, Yunnan Province, China. Int. J. Remote Sens. 32 (9), 2409–2426.
Boardman, J.W., 1989. Inversion of imaging spectrometry data using singular value decom-
position. IGARSS'89, 12th Canadian Symposium on Remote Sensing, pp. 2069–2072.
Boardman, J.W., 1992. Sedimentary Facies Analysis Using Imaging Spectrometry: A
Geophysical Inverse Problem. (Unpublished Ph.D. thesis) University of Colorado.
Butt, C.R.M., Zeegers, H., 1992. Regolith Exploration Geochemistry in Tropical and Sub-
tropical Terrains. Elsevier, Amsterdam and London.
Carranza, E.J.M., Hale, M., 2002. Mineral mapping with Landsat Thematic Mapper data
for hydrothermal alteration mapping in heavily vegetated terrain. Int. J. Remote
Sens. 23 (22), 4827–4852.
Clark, R.N., King, T.V.V., Klejwa, M., Swayze, G.A., 1990. High spectral resolution reflec-
tance spectroscopy of minerals. J. Geophys. Res. 95, 12653–12680.
Crippen, R.E., Blom, R.G., 2001. Unveiling the lithology of vegetated terrains in remotely
sensed imagery. Photogramm. Eng. Remote. Sens. 67, 935–943.
Crosta, A., Moore, J.M.M., 1989. Enhancement of Landsat Thematic Mapper imagery for
residual soil mapping in SW Minais Gerais State, Brazil: a prospecting case history
in Greenstone belt terrain. Proceedings of the 7th ERIM Thematic Conference:
Remote Sensing for Exploration Geology, pp. 1173–1187.
Deller, M.E.A., 2006. Facies discrimination in laterites using Landsat Thematic Mapper,
ASTER and ALI data—example from Eritrea and Arabia. Int. J. Remote Sens. 27 (12),
2389–2409.
Di Tommaso, I., Rubinstein, N., 2007. Hydrothermal alteration mapping using ASTER
data in the Infiernillo porphyry deposit, Argentina. Ore Geol. Rev. 32, 275–290.
Dill, H.G., Horn, E.E., 1996. The origin of a hypogene sarabauite–calcite mineralization
at the Lucky Hill Au–Sb mine Sarawak, Malaysia. J. SE Asian Earth Sci. 14 (l/2),
29–35.
El Desouky, H.A., Muchez, P., Dewaele, S., Boutwood, A., Tyler, R., 2008. Postorogenic or-
igin of the stratiform Cu mineralization at Lufukwe, Lufilian Foreland, Democratic Re-
public of Congo. Econ. Geol. 103, 555–582.
Folkman, M., Pearlman, J., Liao, L., Jarecke, P., 2001. EO-1/Hyperion hyperspectral imag-
er design, development, characterization, and calibration. Hyperspectral Remote
Sensing of the Land and Atmosphere. Proc. SPIE 4151, 40–51.
Fraser, S.J., Green, A.A., 1987. A software defoliant for geological analysis of band ratios.
Int. J. Remote Sens. 8, 525–532.
Gabr, S., Ghulam, A., Kusky, T., 2010. Detecting areas of high-potential gold mineralization
using ASTER data. Ore Geol. Rev. 38, 59–69.
Galva˜o, L.S., Almeida-Filho, R., Vitorello, I., 2005. Spectral discrimination of hydrother-
mally altered materials using ASTER short-wave infrared bands: evaluation in a
tropical savannah environment. Int. J. Appl. Earth Obs. Geoinformation 7, 107–114.
Gersman, R., Ben-Dor, E., Beyth, M., Avigad, D., Abraha, M., Kibreba, A., 2008. Mapping
of hydrothermally altered rocks by the EO-1 Hyperion sensor, northern Danakil,
Eritrea. Int. J. Remote Sens. 29 (13), 3911–3936.
Hall, R., 2012. Late Jurassic–Cenozoic reconstructions of the Indonesian region and the
Indian Ocean. Tectonophysics 570–571, 1–41.
Hon, V., 1981. Physical controls of mineralization in the Bau town area, west Sarawak,
Malaysia. Sarawak Min. Bull., 1, pp. 43–54.
Hunt, G., 1977. Spectral signatures of particulate minerals in the visible and near infrared.
Geophysics 42, 501–513.
Hunt, G.R., Ashley, P., 1979. Spectra of altered rocks in the visible and near infrared.
Econ. Geol. 74, 1613–1629.
Hunt, G.R., Salisbury, J.W., Lenhoff, C.J., 1971. Visible and near-infrared spectra of min-
erals and rocks: III. Oxides and hydroxides. Mod. Geol. 2, 195–205.
Hutchison, C.S., 1989. Geological Evolution of South–East Asia. Clarendon Press, Oxford
(368 pp.).
Igarashi, T., 2001. ALOS mission requirement and sensor specifications. Adv. Space Res.
28 (1), 127–131.
Jensen, J.R., 2005. Introductory Digital Image Processing. Pearson Prentice Hall, Upper
Saddle River.
Kim, Y.U., 1994. Repeated mineralization ages and remobilization of elements in gold ore
deposits from the Chonsan, Rumoh, and Chokei mines. Resour. Geol. 44 (5), 339–352.
Kirk, H.J.C., 1968. The igneous rocks of Sarawak and Sabah. Geol. Survey of Malaysia
Borneo region, Bull., 5, p. 210.
Kruse, F.A., Boardman, J.W., 2000. Characterization and mapping of kimberlites and re-
lated diatremes using hyperspectral remote sensing. IEEE Trans. Geosci. Remote.
Sens. 0-7803-5846-5.
Kruse, F.A., Boardman, J.W., Huntington, J.F., 2003. Comparison of airborne hyperspectral
data and EO-1 Hyperion for mineral mapping. IEEE Trans. Geosci. Remote. Sens. 41
(6), 1388–1400.
Lee, J.S., 1980. Digital image enhancement and noise filtering by use of local statistics.
IEEE Transaction on Pattern Analysis and Machine Intelligence. PAMI-2, 2,
pp. 165–168.
Marjoribanks, R.W., 1989. A structural interpretation of the West Sarawak Region, Ma-
laysia. RGC Research Group Unpublished report, pp: 23.
Massironi, M.L., Bertoldi, P., Calafa, D., Visona, A., Bistacchi, C., Giardino, A., Schiavo, B.,
2008. Interpretation and processing of ASTER data for geological mapping and
granitoids detection in the Saghro massif (eastern Anti-Atlas, Morocco). Geosphere
4 (4), 736–759.
Metcalfe, I., 1996. Pre-Cretaceous evolution of SE Asian terranes. In: Hall, R., Blundell,
D.J. (Eds.), Tectonic Evolution of South-east Asia: Geological Society of London,
Special Publication, vol. 106, pp. 97–122.
196 A.B. Pour et al. / Ore Geology Reviews 54 (2013) 181–196
Meyerhoff, A.A., 1995. Surge-tectonic evolution of southeastern Asia: a geohydrodynamics
approach. J. SE Asian Earth Sci. 12 (3/4), 145–247.
Moss, S.J., 1998. Embaluh group turbidites in Kalimantan: evolution of a remnant oceanic
basin in Borneo during the Late Cretaceous to Palaeogene. J. Geol. Soc. Lond. 155,
509–524.
Okada, K., Segawa, K., Hayashi, I., 1993. Removal of the vegetation effect from LANDSAT
TM and GER imaging spectroradiometer data. ISPRS J. Photogramm. Remote. Sens.
48 (6), 16–27.
Osei, B.A., Singh, B., 1999. Electrophoretic mobility of some tropical soil clays: effect of
iron oxides and organic matter. Geoderma 93, 325–334.
Park, B., Chen, Y.R., 2001. Co-occurrence matrix texture features of multi-spectral im-
ages on poultry carcasses. J. Agric. Eng. Res. 78 (2), 127–139.
Pearlman, J.S., Barry, P.S., Segal, C.C., Shepanski, J., Beiso, D., Carman, S.L., 2003. Hyperion, a
space-based imaging spectrometer. IEEE Trans. Geosci. Remote Sens. 41 (6), 1160–1173.
Percival, T.J., Bagby, W.C., Radtke, A.S., 1988. Physical and chemical features of
precious-metal deposits hosted by sedimentary rocks in the western United
States. In: Schafer, R.W., Cooper, J.J., Vikre, P.G. (Eds.), Bulk Mineable Precious
Metal Deposits of the Western United States: Reno. Geological Society of Nevada,
pp. 11–34.
Percival, T.J., Radtke, A.S., Bagby, W.C., 1990. Relationships among Carbonate-replacement
gold deposits, gold skarns, and intrusive rocks, Bau Mining District, Sarawak, Malaysia.
Min. Geol. 40 (1), 1–16.
Podwysocki, M.H., Mimms, D.L., Salisbury, J.W., Bender, L.V., Jones, O.D., 1984. Analysis
of Landsat-4 TM data for lithologic and image mapping purpose. Proceedings of
Landsat-4 Science Investigations Summary, Greenbelt, Maryland, vol. 2, pp. 35–39.
Pour, B.A., Hashim, M., 2011. Identification of hydrothermal alteration minerals for
exploring of porphyry copper deposit using ASTER data, SE Iran. J. Asian Eart Sci.
42, 1309–1323.
Pour, B.A., Hashim, M., 2012a. The application of ASTER remote sensing data to porphy-
ry copper and epithermal gold deposits. Ore Geol. Rev. 44, 1–9.
Pour, B.A., Hashim, M., 2012b. Identifying areas of high economic-potential copper
mineralization using ASTER data in Urumieh-Dokhtar Volcanic Belt, Iran. Adv. Space
Res. 49, 753–769.
Pour, B.A., Hashim, M., 2013. Fusing ASTER, ALI and Hyperion data for enhanced mineral
mapping. Int. J. Image Data Fusion. http://dx.doi.org/10.1080/19479832.2012.753115.
Rajendran, S., Al-Khirbasha, S., Pracejusa, B., Nasira, S., Al-Abria, A.H., Kusky, T.M.,
Ghulam, A., 2012. ASTER detection of chromite bearing mineralized zones in Semail
Ophiolite Massifs of the northern Oman Mountains: exploration strategy. Ore Geol.
Rev. 44, 121–135.
Rajesh, H.M., 2008. Mapping Proterozoic unconformity-related uranium deposits in the
Rockole area, Northern Territory, Australia using Landsat ETM+. Ore Geol. Rev. 33,
382–396.
Ranjbar, H., Masoumi, F., Carranza, E.J.M., 2011. Evaluation of geophysics and spaceborne
multispectral data for alteration mapping in the Sar Cheshmeh mining area, Iran. Int. J.
Remote Sens. 32 (12), 3309–3327.
Rockwell, B.W., Hofstra, A.H., 2008. Identification of quartz and carbonate minerals
across northern Nevada using ASTER thermal infrared emissivity data Implications
for geologic mapping and mineral resource investigations in well-studied and
frontier areas. Geosphere 4 (1), 218–246.
Rodger, A., Cudahy, T., 2009. Vegetation corrected continuum depths at 2.20 μm: an ap-
proach for hyperspectral sensors. Remote Sens. Environ. 113, 2243–2257.
Rosenqvist, A., Shimada, M., Chapman, B., McDonald, K., Grandi, G., De Jonsson, H.,
Williams, C., Rauste, Y., Nilsson, M., Sango, D., Matsumoto, M., 2004. An overview
of the JERS-1 SAR Global Boreal Forest Mapping (GBFM) project. Geoscience and
Remote Sensing Symposium, IGARRS 0.4, Proceedings. 20–24 September 2004,
vol:2, pp. 1033–1036.
Rowan, L.C., Goetz, A.F.H., Ashley, R.P., 1977. Discrimination of hydrothermally altered
and unaltered rocks in visible and near infrared multispectral images. Geophysics
42 (3), 522–535.
Rowan, L.C., Hook, S.J., Abrams, M.J., Mars, J.C., 2003. Mapping hydrothermally altered
rocks at Cuprite, Nevada, using the Advanced Spaceborne Thermal Emission and
Reflection Radiometer (ASTER), a new satellite-imaging system. Econ. Geol. 98 (5),
1019–1027.
Ruiz-Armenta, J.R., Prol-Ledesma, R.M., 1998. Techniques for enhancing the spectral re-
sponse of hydrothermal alteration minerals in Thematic Mapper images of Central
Mexico. Int. J. Remote Sens. 19, 1981–2000.
Sabins, F.F., 1999. Remote sensing for mineral exploration. Ore Geol. Rev. 14, 157–183.
San, B.T., Suzen, M.L., 2010. Evaluation of different atmospheric correction algorithms
for EO-1 Hyperion imagery. The International Archives of Photogrammetry: Re-
mote Sensing and Spatial Information Sciences. Vol. XXXVII. Part 8. Kyoto Japan
2010, pp. 392–397.
Sato, T., 1974. Distribution and geological setting of the Kuroko deposits. In: Ishihara, S.
(Ed.), Geology of Kuroko deposits: The Society Mining Geologist Japan Special Issue, 6,
pp. 1–9.
Schuh, W.D., 1993. Geology, Geochemistry, and Ore Deposits of the Bau Gold Mining
District, Sarawak, Malaysia. (Unpublished Ph.D thesis) The University of Arizona.
Schuh, W., Guilbert, J.M., 1990. Gold in Borneo. Eine Disseminations vererzung yom Typ
Carlin als distales Produkt einer porphyritischen Intrusion? Nachr. Deut. Geol. Ges.
143, 189–190.
Sheng, Y., Xia, Z.G., 1996. A comprehensive evaluation of filters for radar speckle sup-
pression. Geoscience and Remote Sensing Symposium, IGARSS.96, Remote Sensing
for Sustainable Future, vol. 3, pp. 1559–1561.
Shi, Z., Fung, B., 1994. A comparison of digital speckle filters. Proceedings of IGARSS, In-
ternational Geoscience and Remote Sensing Symposium, 94, pp. 2129–2133.
Shimabukuro, Y.E., Smith, J.A., 1991. The least-squares mixing models to generate frac-
tion images derived from remote sensing multispectral data. IEEE Trans. Geosci.
Remote Sens. 29, 16–20.
Sillitoe, R.H., Bonham, H.F., 1990. Sediment-hosted gold deposits: distal products of
magmatic-hydrothermal systems. Geology 18, 157–161.
Spatz, D.M., Wilson, R.T., 1995. Remote sensing characteristics of porphyry copper sys-
tems, western America Cordillera. In: Pierce, F.W., Bolm, J.G. (Eds.), Arizona Geological
Society Digest, vol. 20, pp. 94–108.
Stevens, M., Fulton, G.W., 2010. Technical Report on Bau Project in Bau, Sarawak. East
Malaysia for Olympus Pacific Minerals Inc. (www.olympuspacific.com/. Access date
2/8/2012).
Sveinsson, J.R., Benediktsson, J.A., 1996. Speckle reduction and enhancement of SAR
image in the wavelet domain. Geoscience and Remote Sensing Symposium, IGARSS.96,
Remote sensing for sustainable future, vol. 1, pp. 63–66.
Tate, R.B., 1991. Cross-border correlation of geological formations in Sarawak and Kali-
mantan. Bull. Geol. Soc. Malaysia 28, 63–96.
Thome, K., Palluconi, F., Takashima, T., Masuda, K., 1998. Atmospheric correction of
ASTER. IEEE Trans. Geosci. Remote Sens. 36 (4), 1119–1211.
Van Ruitenbeek, F.J.A., Cudahy, T.J., Van der Meer, F.D., Hale, M., 2012. Characterization
of the hydrothermal systems associated with Archean VMS — mineralization at Pan-
orama, Western Australia, using hyperspectral, geochemical and geothermometric
data. Ore Geol. Rev. 45, 33–46.
Vicente, L.E., Filho, C.R.S., 2011. Identification of mineral components in tropical soils
using reflectance spectroscopy and advanced spaceborne thermal emission and re-
flection radiometer (ASTER) data. Remote Sens. Environ. 115, 1824–1836.
Williams, P.R., Johnston, C.R., Almond, R.A., Simamora, W.H., 1988. Late Cretaceous to
Early Tertiary structural elements of West Kalimantan. Tectonophysics 148, 279–297.
Wilson, M.E.J., 2002. Cenozoic carbonates in Southeast Asia: implications for equatorial
carbonate development. Sediment. Geol. 147, 295–428.
Wolfenden, E.D., 1965. Bau mining district, west Sarawak, Malaysia, I, Bau. Geol. Survey
Borneo Region, Malaysia, Bull., 7, pp. 1–147.
Zhang, J.X., 2010. Multi-source remote sensing data fusion: status and trends. Int. J.
Image Data Fusion 1.1, 5–24.
Zhang, X., Pazner, M., Duke, N., 2007. Lithologic and mineral information extraction for
gold exploration using ASTER data in the south Chocolate Mountains (California).
J. Photogramm. Remote. Sens. 62, 271–282.
Zhu, Y., An, F., Tan, J., 2011. Geochemistry of hydrothermal gold deposits: a review.
Geosci. Front. 2 (3), 367–374.