Remote Sensing of Environment 84 (2003) 350 – 366

www.elsevier.com/locate/rse

Lithologic mapping in the Mountain Pass, California area using

Advanced Spaceborne Thermal Emission and
Reflection Radiometer (ASTER) data
Lawrence C. Rowan*, John C. Mars
National Center, U.S. Geological Survey, Mail Stop 954, Reston, VA 20192, USA
Received 23 April 2002; received in revised form 12 August 2002; accepted 24 August 2002

Abstract

Evaluation of an Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image of the Mountain Pass, California
area indicates that several important lithologic groups can be mapped in areas with good exposure by using spectral-matching techniques.
The three visible and six near-infrared bands, which have 15-m and 30-m resolution, respectively, were calibrated by using in situ
measurements of spectral reflectance. Calcitic rocks were distinguished from dolomitic rocks by using matched-filter processing in which
image spectra were used as references for selected spectral categories. Skarn deposits and associated bright coarse marble were mapped in
contact metamorphic zones related to intrusion of Mesozoic and Tertiary granodioritic rocks. Fe-muscovite, which is common in these
intrusive rocks, was distinguished from Al-muscovite present in granitic gneisses and Mesozoic granite.
Quartzose rocks were readily discriminated, and carbonate rocks were mapped as a single broad unit through analysis of the 90-m
resolution, five-band surface emissivity data, which is produced as a standard product at the EROS Data Center. Three additional classes
resulting from spectral-angle mapper processing ranged from (1) a broad granitic rock class (2) to predominately granodioritic rocks and (3) a
more mafic class consisting mainly of mafic gneiss, amphibolite and variable mixtures of carbonate rocks and silicate rocks.
D 2002 Elsevier Science Inc. All rights reserved.

Keywords: Advanced Spaceborne Thermal Emission and Reflection Radiometer; ASTER; Remote sensing; Geologic mapping

1. Introduction A very important aspect of the EOS Program is the open

The Advanced Spaceborne Thermal Emission and
Reflection Radiometer (ASTER), which is aboard the Earth
Observing System (EOS) TERRA platform, records solar
radiation in 14 spectral bands (Table 1). ASTER measures
reflected radiation in three bands between 0.52 and 0.86 Am
(VNIR) and in six bands from 1.6 to 2.43 Am (SWIR), with
15- and 30-m resolution, respectively (Fujisada, 1995).
ASTER also has a back-looking VNIR telescope with 15-
m resolution. Thus, stereoscopic VNIR images can be
acquired at 15-m resolution. In addition, emitted radiation
is measured at 90-m resolution in five bands in the 8.125–
11.65-Am wavelength region (TIR). The swath width is 60
km, but ASTER’s pointing capability extends the total
cross-track viewing capability to 232 km (Fujisada, 1995).
* Corresponding author.
E-mail address: lrowan@usgs.gov (L.C. Rowan).
availability of the data from all the instruments, including
on-demand standard products, for research purposes at low
cost. ASTER standard products include VNIR and SWIR
surface radiance and reflectance, brightness temperature at
the sensor, TIR surface radiance and emissivity, surface
kinetic temperature, decorrelation-stretch images and digi-
tal-elevation models (DEM).
The VNIR, SWIR and TIR wavelength regions provide
complementary data for lithologic mapping. The three
VNIR bands are important sources of information about
absorption in transition metals, especially iron and some
rare-earth elements (REE) (Hunt, Salisbury, & Lenhoff,
1972; Rowan, Kingston, & Crowley, 1986), and for
chlorophyll absorption in photosynthesizing vegetation
(Knipling, 1970). In the six SWIR bands carbonate,
hydrate and hydroxide mineral spectra display molecular
absorption features related to overtones and combination
tones (Hunt, 1977). This wavelength region has been

0034-4257/02/$ - see front matter D 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 3 4 - 4 2 5 7 ( 0 2 ) 0 0 1 2 7 - X
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 351

Table 1
Performance parameters for the ASTER radiometer (Fujisada, 1995)
ASTER baseline performance requirements
Subsystem Band number Spectral Radiometric Absolute Spatial Signal
range (Am) resolution accuracy (r) resolution quantization levels
VNIR 1 0.52 – 0.60 NEDq V 0.5% V F 4% 15 m 8 bits
2 0.63 – 0.69
3N 0.78 – 0.86
3B 0.78 – 0.86
4 1.600 – 1.700 NEDq V 0.5%
5 2.145 – 2.185 NEDq V 1.3%
SWIR 6 2.185 – 2.225 NEDq V 1.3% V F 4% 30 m 8 bits
7 2.235 – 2.285 NEDq V 1.3%
8 2.295 – 2.365 NEDq V 1.0%
9 2.360 – 2.430 NEDq V 1.3%
10 8.125 – 8.475
11 8.475 – 8.825 V3 K (200 – 240 K)
TIR 12 8.925 – 9.275 NEDT V 0.3 K V2 K (240 – 270 K) 90 m 12 bits
13 10.25 – 10.95 V1 K (270 – 340 K)
14 10.95 – 11.65 V2 K (340 – 370 K)

Stereo base-to-height ratio 0.6 (along-track)

Swath width 60 km
Total coverage in 232 km
cross-track direction
by pointing
Mission life 5 years
MTF at Nyquist frequency 0.25 (cross-track)
0.20 (along-track)
Band-to-band registration Intra-telescope: 0.2 pixels
Inter-telescope: 0.3 pixels
of coarser band
Peak data rate 89.2 Mbps
Mass 406 kg
Peak power 726 W
Band number 3N refers to the nadir pointing view, whereas 3B designates the backward pointing view.

exploited extensively for mineralogic mapping by using
high-spectral-resolution imaging systems, such as the Air-
borne Visible/Infrared Imaging Spectrometer (AVIRIS)
(Boardman & Kruse, 1994; Clark, Swayze, Gallagher,
Gorelick, & Kruse, 1991; Crowley, 1993; Kruse, 1988;
Rowan et al., 1995; Rowan, Clark, & Green, 1996, among
others). Although important rock-forming minerals, includ-
ing quartz and feldspar, do not exhibit absorption features
in the VNIR and SWIR regions, most of them display
fundamental molecular absorption features in the TIR
wavelength region (Hunt & Salisbury, 1974, 1975, 1976;
Lyon, 1964).
This paper describes an evaluation of ASTER data
recorded on June 4, 2000, of the Mountain Pass, California
area, which is located approximately 60 km south – south-
west of Las Vegas, NV. (Fig. 1). The part of the ASTER
scene selected for this evaluation corresponds to three 10-
km-long AVIRIS segments, which had been calibrated and
evaluated in the field previously (Rowan et al., 1996).
AVIRIS provides high spectral resolution data to correlate
and crosscheck ASTER data analysis results. The study area
is transected by US Interstate Highway 15, with excellent
exposures of a wide range of lithologies in the Clark
Mountain Range north of highway I15 and in the Mescal
and Ivanpah Ranges south of the highway (Fig. 1).
2. Geologic setting
In the northern two-thirds of the study area, Proterozoic
gneisses, granitoids and minor amphibolite underlie most of
the eastern part, whereas the central and western parts are
characterized by folded, thrust-faulted Paleozoic carbonate
and quartzose clastic rocks (U.S. Geological Survey, 1991)
(Fig. 2). Jurassic granitic rocks dominate the southern one-
third of the study area. Mesozoic granodioritic and dioritic
intrusives, which are generally less than 1 km wide, are
important in the northern part, because they are responsible
for several skarn deposits (Fig. 2). The most important
mineral deposit within the study area, however, is the
Sulphide Queen REE deposit (Castor, 1991; Olson, Shawe,
Pray, & Sharp, 1954) located at Mountain Pass, California,
which is associated with Proterozoic carbonatite and alkalic
rocks (Dewitt, Kwak, & Zartman, 1987) (Figs. 1 and 2).
Tertiary volcanic rocks of intermediate composition locally
overlie the Aztec Sandstone and gneiss near the middle of
352 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366

Fig. 1. Color – infrared composite subset ASTER image of the Mountain Pass, California study area showing the location of features mentioned in the text. Red
band 3; green band 2; blue band 1. Inset shows the location of the study area within the full ASTER scene.
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 353

Fig. 2. Generalized lithologic map of the study (from U.S. Geological Survey Open-file Report 91-427).

the study area. Undifferentiated Quaternary alluvial and
colluvial deposits cover broad areas, especially on the
western and southeastern slopes (Fig. 2).
Some of these lithologic units can be distinguished in
standard color – infrared images (Fig. 1) and in decorrela-
tion-stretch image standard products because of their
albedo, color and textural characteristics. The availability
of stereoscopic images enhances this photointerpretative
analysis. However, photointerpretative analysis does not
yield direct compositional information. Maximum litho-
logic information is derived from ASTER data when
spectral analysis and photointerpretative analyses are used
together.
3. Spectral reflectance analysis
An important aspect of this evaluation of the nine VNIR
and SWIR ASTER bands is comparison of the image
analysis results with the results of the AVIRIS data analysis
using similar spectral matching techniques. AVIRIS records
reflected solar radiation in 224 narrow spectral channels that
354 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366

span the 0.4 – 2.5-Am wavelength region (Vane et al., 1993).

The swath width is approximately 11 km, and the spatial
resolution of this data set is about 17 m. Thus, the AVIRIS
results are described first, and then, the ASTER VNIR/
SWIR results are presented. Analyses of the nine VNIR-
SWIR bands and the five TIR bands were conducted
separately, and then, the results were evaluated together.

3.1. AVRIS data analysis

Three segments of calibrated AVIRIS data were analyzed

using in situ and laboratory measurements of spectral
reflectance and XRD analysis of selected samples. The
AVIRIS image spectra shown in Fig. 3A, which focus on
the 2.1 – 2.4-Am wavelength region, exhibit well-defined
CO3 absorption features near 2.33 Am in calcite (limestone)
and dolomite, Al –O – H features between 2.17 and 2.22 Am
in Al-muscovite, Fe-muscovite and kaolinite and Mg –O –H
absorption features at 2.26 and 2.35 Am. Note that the
position of the minimum in the dolomite spectrum is near
2.315 Am, whereas the calcite minimum is at about 2.335
Am. The epidote spectrum is very similar to the dolomite
and calcite spectra in this part of the 0.4 – 2.5-Am wave-
length region, but the weak feature at 2.25 Am in the epidote
spectrum is distinctive. The spectrum, which displays an
Mg –O – H doublet absorption feature in the 2.30 –2.40-Am
region, represents amphibolite. The Al-muscovite spectrum
exhibits intense absorption centered at 2.20 Am, whereas the
Fe-muscovite minimum is located at about 2.22 Am. Kao-
linite is spectrally similar to Al-muscovite in this wave-
length region, but it is distinguished by the slight depression
or shoulder at 2.17 Am, which creates a doublet shape.
The carbonatite reflectance spectrum is especially rich
with features related to REE and carbonate (Fig. 3B). The
sharp features located at 0.58, 0.74, 0.80 and 0.87 Am are
due to electronic transitions in Nd3 +, and the somewhat
broader features in the 1.55-Am region are due to Sm3 +; the
feature centered near 2.33 Am is due to CO3 absorption in
bastnaesite and calcite.
The AVIRIS spectra shown in Fig. 3A were used as
reference spectra in matched-filter processing (Harsanyi &
Chang, 1994). The results indicate the dominance of Al-
muscovite in the granitic gneiss and granitoids in the eastern
part of the study area (A, red, Fig. 4), in the quartzose clastic
rocks located in the northern part (B, Fig. 4), in the granitic
intrusive rocks in the Ivanpah Mountains (C, Fig. 4) and in
Fig. 3. (A) AVIRIS image spectra in the 2.1 – 2.42-Am wavelength region of
the Colosseum mine (D, Fig. 4). Limonitic coatings, which the main spectral – lithologic categories; (B) AVIRIS image spectrum of
are present on some of these rock surfaces, do not mask the REE-rich carbonatite in the Sulphide Queen mine area. ASTER band
muscovite spectral response. centers shown at the top (Fujisada, 1995).

Fig. 4. Results of matched-filter processing of AVIRIS image spectra shown in Fig. 3A and B. Color code of spectral – lithologic categories: red—Al-
muscovite; blue—dolomite; yellow—limestone; cyan—Fe-muscovite; orange—epidote-bearing rocks; green—Fe, Mg – O – H + Al – O – H minerals;
magenta—REE-rich rocks. Areas discussed in the text: (A) dominant Al-muscovite in granitic gneiss and granitoids; (B) Al-muscovite in quartzose clastic
rocks; (C) Al-muscovite in granitic intrusive rocks in the Ivanpah Mountains; (D) Al-muscovite in hydrothermally altered rocks in the Colosseum mine; (E), (F)
and (G) dolomitic areas with low density of mapped dolomite pixels; (EP1) mafic epidote-bearing gneiss; (EP2) skarn; (H) small concentrations of epidote-
bearing skarn; (J) dolomitic area which is mapped as epidote-bearing in ASTER image (Fig. 9); (K) epidote-bearing gneiss which is mapped as Mg – O –
H + Al – O – H in the ASTER image (Fig. 9); (L) siltstone and shale; (GD) granitic to granodioritic intrusives.
L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 355
356 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
Calcite and dolomite are indicated in carbonate rocks by
dense, coherent yellow and dark blue patterns, respectively,
in the central and western parts of the study area (Fig. 4).
Two types of epidote-bearing rocks are present, although
they are not distinguished in this image (orange, Fig. 4): (1)
mafic gneiss located along the southeastern and central
margins and west of the REE mine (EP1, Fig. 4) and (2)
skarns related to granodioritic intrusive rocks (EP2, Fig. 4).
Some of the intrusive rocks contain Fe-muscovite (GD,
cyan, Fig. 4), which is also common in clastic rocks present
in the northwestern part of the area. Epidote, as well as
chlorite, is present in these clastic rocks.
The green pattern in the AVIRIS image map represents
mainly amphibole, biotite and chlorite, as indicated by the
doublet absorption feature at 2.32 and 2.38 Am, and a minor
feature located near 2.25 Am (amphibole, Fig. 3A). Lith-
ologies are mainly amphibolite and biotite schist within the
gneiss complex along the eastern margin of the study area
(A, Fig. 4), and in the northernmost labeled granodioritic
intrusive (GD, Fig. 4).
Carbonatite is restricted to the REE mine area, represent-
ing stock-piled processed ore, the open-pit mine and some
of the dump area (magenta, Fig. 4). The kaolinite and
alunite categories are very sparsely represented and are
not shown in Fig. 4.
3.2. ASTER VNIR-SWIR data analysis
The following steps constituted the data processing and
analysis of the spectral radiance data contained in the nine
VNIR-SWIR ASTER bands: (1) spatial registration of the
30-m SWIR data to the 15-m VNIR data; (2) calibration of
the nine-band spectral radiance data to spectral reflectance;
(3) evaluation of selected ratio images; (4) selection of
spectral lithologic categories; (5) mapping the selected
spectral lithologic categories.
3.3. Calibration
The ASTER scene subset was calibrated to surface
reflectance by reference to in situ spectral reflectance
measurements of the Ivanpah playa surface (Fig. 5). The
average in situ spectrum was resampled to the nine ASTER
VNIR and SWIR bandpasses, and then, the quotient of this
spectrum and the average ASTER image spectrum repre-
senting the playa sample area was used to rescale the initial
radiance values. Comparison of this surface reflectance
image and the standard surface reflectance product
(AST_07) obtained from the EROS Data Center (EDC)
indicates some significant differences. The largest difference
is in band 9, which is 10 –20% higher in the EDC standard
product. In addition, band 3 is consistently higher in EDC
image. Therefore, the in situ calibrated ASTER VNIR and
SWIR data were evaluated and are discussed here. Although
the in situ playa spectra were recorded on Nov. 9, 2000 and
the elevation of this surface is approximately 800 m lower
than the study area average, comparisons of spectra from the
in situ calibrated ASTER image with AVIRIS spectra and in
situ spectra measured within the study area indicate that any
changes in the playa surface and the elevation differences
did not cause significant spectral reflectance changes in the
ASTER bands.
The in situ calibrated ASTER image may be affected by
the ‘crosstalk’ instrument problem, which is caused by light
reflected from the band 4 optical components leaking into
the other SWIR band detectors (Iwasaki, 2001, personal
communication), particularly band 9. However, the magni-
tude of any anomalous reflectance appears to be small
compared to the spectral reflectance differences among the
main spectral – lithologic categories.
3.4. Spectral –lithologic category selection
3.4.1. ASTER laboratory reflectance spectra
Comparison of the high-resolution laboratory reflectance
spectra of selected minerals and the corresponding
resampled ASTER spectra (Fig. 6) shows that these ASTER
mineral spectra are distinguishable in spite of their much
lower spectral resolution. The spectral reflectance differ-
ences between the carbonate minerals (calcite and dolomite,
Fig. 6) and the muscovite minerals (Al-muscovite and Fe-
muscovite, Fig. 6) are conspicuous in the ASTER spectra, as
well as in the laboratory spectra. However, the important but
subtle differences between calcite and dolomite, between
Al-muscovite and Fe-muscovite and between epidote and
calcite are much clearer in the high-resolution laboratory
spectra. At ASTER spectral resolution, the CO3 absorption
feature in the dolomite spectrum is distinguished from the
calcite CO3 feature and the epidote Mg – O –H absorption
feature by the slight depression at 2.26 Am in the dolomite
spectrum (Fig. 6). Fe-muscovite, which exhibits an Al –O –
H absorption feature centered near 2.22 Am (Rowan et al.,
1996) (Fig. 6), displays a broad, step-type 2.20– 2.26-Am
absorption feature at ASTER spectral resolution (Fig. 6).
Epidote and calcite are similar in the 2.1 –2.4-Am region,
but their slopes between 2.17 and 2.33 Am are different (Fig.
6).
Reflectance spectra extracted from the calibrated nine-
band ASTER image subset display CO3, Al – O – H and
Mg – O –H absorption features. The overall shapes of the
ASTER image spectra, however, are different from the
ASTER laboratory spectra due to the effects of several
factors which influence the image spectra but are obviated
in the laboratory, including the presence of variable min-
eral –mineral mixtures, mineral –vegetation mixtures, grain-
size variations and residual atmospheric absorption features
(Fig. 7). These spectra, which were selected by using the
pixel –purity index (PPI) procedure (Boardman, Kruse, &
Green, 1995), were used for mapping the lithologic catego-
ries, because they produced better lithologic separation in
the ASTER images than the laboratory spectra (Fig. 6). The
pixel –purity index algorithm selects pixels whose spectra
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 357

Fig. 5. Flow diagram showing the steps in calibrating the ASTER data by using an average in situ reflectance spectrum of the Ivanpah playa surface.
358 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366

continuum increases the intensity of the absorption feature

(Crowley et al., 1989):
RBD ¼ ðband 1 þ band 2Þ=band 3

Comparison of the thresholded ASTER RBD images with

the generalized lithologic map and AVIRIS results (Figs. 2
and 4, respectively) shows that the pattern of the pixels
representing Ca – CO3 absorption ((band 7 + band 9)/band 8,
Fig. 8A) corresponds well with the limestone distribution in
the lithologic map and AVIRIS results, and the Ca,Mg –CO3
absorption RBD image ((band 6 + band 8)/band 7, Fig. 8B)
exhibits more limited correspondence with the dolomite
distribution (Figs. 2 and 4). Mg –O –H absorption may also
contribute to the CO3 ASTER RBD image, but thresholding
at high digital number (DN) levels has eliminated the weaker
Mg – O – H absorption in Fig. 8A and B. The Al – O – H
absorption RBD image ((band 5 + band 7)/band 6, Fig. 8C)
corresponds to the distribution of granitoids and gneisses
along the eastern margin of the study area, granitic and
granodioritic rocks in the southern part and quartzose rocks
in the western and northwestern parts (Figs. 2 and 4).
Fig. 8D is a band 2/band 1 ASTER ratio image which
shows the distribution of pixels with intense Fe3 + absorp-
tion, mainly in the granitoids and gneisses, where Al – O –H
absorption is also intense, and locally within the granitic –
granodioritic rocks in the Ivanpah Mountains (Fig. 2). This
ratio image has been filtered using a 3  3 median filter and
then thresholded.

3.4.3. Matched-filter mapping of spectral reflectance litho-

logic categories
Fig. 6. Five pairs of laboratory reflectance spectra of rock-forming minerals Seven lithologic units were mapped in the study area by
which are important in the study area. The upper spectrum of each spectral using the ASTER image spectra (Fig. 7) as reference in a
pair is a full resolution spectrum, and the lower spectrum of each spectral
pair is a spectrum resampled to ASTER bandpasses 1 – 9. ASTER band
centers shown at the top (Fujisada, 1995).

are extreme or spectrally pure and therefore represent the

most suitable spectral categories.

3.4.2. Mapping spectral lithologic units

3.4.2.1. ASTER ratio images. Ratio images designed to

display the spectral contrast of specific absorption features
have been used extensively in geologic remote sensing
(Rowan, Goetz, & Ashley, 1977; Rowan, Wetlaufer, Goetz,
Billlingsley, & Stewart, 1974). Relative absorption-band
depth (RBD, Crowley, Brickey, & Rowan, 1989) images
are an especially useful three-point ratio formulation for
displaying Al – O –H, Mg – O –H and CO3 absorption inten-
sities prior to conducting more detailed, time-consuming
spectral analysis. For each absorption feature, the numerator
is the sum of the bands representing the shoulders (bands 1 Fig. 7. Reflectance spectra of lithologic categories from the in situ
and 2), and the denominator is the band located nearest the calibrated ASTER image of the study area. ASTER band centers shown at
absorption feature minimum (band 3); removal of the the top (Fujisada, 1995).
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 359

Fig. 8. Relative band-depth images of the study area for major absorption features: (A) RBD=(band 7 + band 9)/band 8, Ca – CO3 (limestone); (B) RBD=(band
6 + band 8)/band 7, Ca,Mg – CO3 (dolomite); (C) RBD=(band 5 + band 7)/band 6, Al – O – H (muscovite); (D) band 2/band 1, Fe3 + (hematite, goethite).

matched-filtering procedure (Harsanyi & Chang, 1994) (Fig.
9). Comparisons of image mapping results based on
matched filtering using all nine bands, bands 4 –9, bands
5 –9 and bands 6– 9, showed that the five bands from band 5
to band 9 generally produced the best separation of the
spectral –lithologic categories and less scatter, except for
REE and ferric iron deposits. REE-bearing materials
exposed in the Sulphide Queen mine area were mapped
by using nine bands to detect the strong Nd3 + absorption
band at 0.80 Am along with the CO3 band at 2.33 Am (Figs.
3B and 7). Note that the level of Nd3 + absorption present in
these deposits is extraordinarily intense and accounts for
low reflectance in ASTER bands 1 –3 (Fig. 7).
Comparison of the ASTER results with the AVIRIS
matched-filter mapping results (Figs. 9 and 4, respectively)
shows that the patterns of the image map units are similar,
but some substantial differences are present in a few areas.
Al-muscovite (red, Figs. 4 and 9) dominates the areas
underlain by gneissic rocks along the eastern margin of
the study area, quartzose clastic rocks located in the north –
central part, granitic rocks in parts of the Ivanpah Mountains
and in the Colosseum mine (A, B, C, and D, respectively,
Figs. 4 and 9).
Broad dark blue areas representing dolomitic rocks
correspond to the thrust-faulted Paleozoic carbonate units
(blue, Figs. 4 and 9). In the AVIRIS image map, dolomite
is shown in more dense, coherent patterns than in the
ASTER image map, except in areas E, F and G (Figs. 4
and 9). Limestone, which is denoted by the yellow pattern
within the same tectonic setting, compares favorably in
Figs. 4 and 9. The main difference is exhibited in area H
where more skarn –marble assemblage is indicated in the
ASTER image than in the AVIRIS image map (H, Figs. 4
and 9). Approximately 4 km east of area H, the skarn –
marble assemblage is more widespread in the ASTER
image map than dolomite which is represented in the
AVIRIS image by a dense, coherent dark blue pattern (J,
Figs. 4 and 9).
The green pattern in the ASTER image map represents
rocks containing Fe,Mg – O – H minerals, including horn-
blende, chlorite, epidote and biotite, as well as the Al– O –
H mineral illite/smectite; the former group of minerals
causes the 2.33-Am absorption features, whereas the latter
mineral causes the 2.20-Am absorption features displayed
in the Mg –O –H + Al– O – H spectrum in Fig. 7. Reliable
identification and mapping of these Fe,Mg – O – H minerals
360 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 361

individually has proven difficult using ASTER data of the

study area due to the overlap of the molecular absorption
features in the 2.30 – 2.40-Am region (Rowan et al., 1996).
Lithologies represented by the green pattern include epi-
dotized amphibolite and biotite schist along the south-
eastern margin (K, Fig. 9), granodioritic intrusives (GD,
Fig. 9) and siltstone and shale in the northwestern part of
the area (L, Fig. 9). Fe-muscovite was also mapped in
these areas, but it is obscured in Fig. 9 by the overlap of
the green and cyan patterns. The largest difference between
the ASTER and AVIRIS image maps is at location K
where epidote was mapped in the AVIRIS image (K, Figs.
4 and 9).
The orange color in the ASTER image map (Fig. 9)
represents white, coarse-grained marble and, adjacent to
the intrusive contact, rocks containing calcite, epidote,
chrysotile and more locally, garnet and diopside (orange,
Fig. 9). These areas are generally quite bright in the
color – infrared composite image (Fig. 1). As noted above,
the orange pattern in the ASTER image map is more
extensive in areas H and J than in the AVIRIS image (Figs.
4 and 9). Epidote-bearing amphibolite and schist which
were mapped in the epidote category in the AVIRIS image
(EP1, orange, Fig. 4) were included in the Mg – O –
H + Al – O –H class in the ASTER image map (EP1 and
K, green, Fig. 9).
REE-rich deposits were mapped in the Sulphide Queen
mine area based on the combination of intense CO3 absorp-
tion in ASTER band 8, and Nd3 + absorption in bands 1 –3
(Fig. 7; magenta, Figs. 4 and 9). However, the mapped
pixels occupy only small areas, mainly mine workings, and
are not readily apparent in Figs. 4 and 9.

4. ASTER spectral emittance analysis
4.1. ASTER laboratory emissivity spectra
Laboratory emissivity spectra of important rock-forming
minerals retain distinct spectral signatures after resampling
to the ASTER TIR bandpasses (Fig. 10). Calcite, as well as
dolomite, laboratory spectra are distinctive because of the
presence of a weak absorption feature centered near 11.2 Am
and otherwise generally flat spectra. However, this weak
absorption feature is commonly not evident in ASTER TIR
spectra (Fig. 10) due to the breadth of these bandpasses.
This spectral shape is similar to that of vegetation in this
Fig. 10. Seven pairs of laboratory emissivity spectra of several important
rock-forming minerals. The upper spectrum of each spectral pair is a full-
resolution spectrum, and the lower spectrum of each spectral pair is a
spectrum resampled to ASTER bandpasses 10 – 14. ASTER band centers
shown at the top (Fujisada, 1995).
wavelength region (Salisbury & Milton, 1987), which can
lead to ambiguities where VNIR data are not available. Note
that several of these minerals exhibit important features near
10.0 Am in the laboratory spectra, but ASTER does not have
a band in this region due to atmospheric absorption. Never-
theless, spectra of mafic minerals, such as biotite, epidote
and hornblende, display minima in ASTER band 13,

Fig. 9. Results of matched-filter mapping lithologic categories using ASTER spectra shown in Fig. 7 as reference. Color code: red—Al-muscovite; blue—
dolomite; yellow—limestone; cyan—Fe-muscovite; orange—skarn – marble assemblage, and epidote-bearing amphibolite and schist; green—Fe,Mg – O –
H + Al – O – H minerals; magenta—REE-rich rocks. Areas discussed in text: (A) Al-muscovite in granitic gneiss and granitoids; (B) Al-muscovite in quartzose
clastic rocks; (C) Al-muscovite in granitic intrusive rocks; (D) Al-muscovite in hydrothermally altered rocks in the Colosseum mine; (E), (F) and (G) areas
where dolomite class is denser in the ASTER image map than in the AVIRIS image map (Fig. 4); (H) area where skarn – marble assemblage is more extensive
than epidote category in the AVIRIS image map (Fig. 4); (J) area where skarn – marble assemblage was mapped instead of dolomite shown in the AVIRIS
image map (Fig. 4); (K) area where Fe,Mg – O – H + Al – O – H minerals (green) are indicated in ASTER image map, but epidote-bearing rocks are shown in the
AVIRIS image map (Fig. 4); (L) siltstone and shale; (EP1) epidote-bearing gneiss; (GD) granitic to granodioritic intrusive rocks.
362 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366 363

whereas quartz, microcline and muscovite minima are

centered in shorter wavelength bands (bands 11 and 12,
Fig. 10), which causes the shift of the restrahlen band
towards longer wavelengths with increasing mafic content,
as noted initially by Lyon (1964).

4.2. Mapping spectral emittance lithologic categories

The ASTER spectral emittance data evaluated in this

study is the standard product (AST_05), which was pro-
duced at the EROS Data Center using the Temperature-
Emissivity Separation (TES) algorithm developed by Gil-
lespie, Matsunaga, Rokugawa, and Hook (1998); the
AST_05 product was not fully validated when this study
was conducted. The spectral emissivity data were co-regis-
tered to the VNIR and SWIR data by using a 3  pixel
duplication of the TIR data. Analysis of the TIR data was
hampered by the presence of prominent banding in all five
bands. The banding is partially subdued in the decorrela-
tion-stretch image (Gillespie, Kahle, & Walker, 1986)
shown in Fig. 11 by superposition of a mask. The mask
was constructed by thresholding the first eigen channel of
minimum-noise-fraction (MNF) transformation (Green, Ber-
man, Switzer, & Craig, 1988) of the emissivity data at a
digital number (DN) of 8.52 (dark blue, Fig. 11). This mask
is confined to areas underlain by carbonate rocks, which is
where the banding is most conspicuous. In addition, areas
with dense vegetation are represented in Fig. 11 by yellow
pixels, which resulted from match-filtering using ASTER
bands 5– 9.
Red in this image corresponds to quartzose rocks (red,
Figs. 2 and 11), mainly quartzite, sandstone and siltstone,
and silicified rocks in the Colosseum mine located in the
northeastern corner (Q and A1, respectively, Fig. 11), and in
the Morning Star mine situated in the southern part (B1, Fig.
11). Broad areas of colluvial and alluvial deposits appear to
be quartz-rich in the southern part of the study area. Pink
pixels correspond to colluvial –alluvial mixtures of carbo-
nate rocks and silicate rocks (MQ, pink, Fig. 11). Dark red
corresponds mainly to granitic gneiss along the eastern
margin of the study area (Fig. 2; GN, Fig. 11). Blue and
cyan generally indicate areas underlain by rocks of inter-
mediate to mafic composition, including biotite schist and
amphibolite in the granitoid – gneiss unit situated west of the
dark red band of granitic gneiss (GN, Fig. 11), dacitic
volcanic rocks overlying a narrow exposure of Aztec Sand-
stone (D1 and SS, respectively, Fig. 11; Fig. 2) and a small
area of alkalic intrusive rocks in the Mountain Pass REE
mine area (M, Fig. 11).
Fig. 12. Emissivity spectra of lithologic categories from the ASTER
standard product emissivity image (AST_05) of the study area. ASTER
band centers shown at the top (Fujisada, 1995).
4.3. Spectral Angle Mapper classification of spectral
emittance categories
Additional lithological information was derived from
analysis of spectral emittance categories identified through
pixel – purity processing of all five TIR bands. The six
image spectra shown in Fig. 12 exhibit less spectral contrast
than the ASTER-convolved laboratory spectra of selected
minerals (Fig. 10), because the image spectra represent
mixtures of minerals. An absorption feature is exhibited in
band 12 in all six image spectra (Fig. 12) due to the
widespread distribution of quartz, even in carbonate terrain
as a result of weathering and soil formation. Although the
general shape of these spectra is similar, the following
differences are potentially useful for mapping these litho-
logic categories: (1) the spectrum representing carbonate
displays the smallest variation, (2) the intermediate + mafic
rocks spectrum exhibits low slope between band 12 and
band 14 and (3) the depth of the band 12 feature is variable.
These six image spectra were used as reference spectra
for Spectral Angle Mapper (SAM) processing (Kruse et al.,
1993). SAM compares the angle between the reference
spectrum and each pixel vector in n-dimensional space,
and smaller angles represent closer matches to the reference
spectrum (Kruse et al., 1993). The SAM classification
result, on which both the MNF eigen channel mask and

Fig. 11. Decorrelation-stretch ASTER image of the study area with vegetation overlay shown in yellow color. Dark blue, MNF band 1 mask. Red band 13;
green band 12; blue band 10. Areas discussed in the text: (A1) hydrothermally altered rocks at the Colosseum mine; (B1) Morning Star mine; (D1) volcanic
rock; (Q) quartzite and siltstone; (MQ) mixed carbonate and silicate colluvial and alluvial deposits; (M) Sulphide Queen mine area; (GD) granitic to
granodioritic intrusive rocks (Fig. 2); (GN) granitic gneiss; (MG) mafic gneiss; (SS) Aztec Sandstone.
364 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
the lavender pixels representing vegetation are superposed,
comprises five classes (Fig. 13). Comparison of these
classes with the geologic map indicates that the extensive
yellow class includes felsic gneissic rocks, granitic rocks
and associated colluvial and alluvial deposits. The red
category, which is the combination of the sandstone and
quartzite classes (Fig. 13), is similar to the red distribution in
the decorrelation-stretch image (Q, A1, and B1, Figs. 11 and
13), which corresponds to quartzose deposits.
The green class, referred to as the granodioritic rock
class, is particularly interesting because it provides some
lithological information not readily evident in the decorre-
lation-stretch image (Fig. 11). The intrusive rocks on the
western slope of the Clark Mountain Range are included in
this class primarily, which is consistent with their grano-
dioritic composition (GD, Fig. 13; Fig. 2), whereas they are
not readily distinguished from the granitoid –granitic gneiss
category in the decorrelation-stretch image (GN, dark red,
Fig. 11). This class also includes areas underlain by argilla-
ceous rocks in the northwestern part of the study area (SH1,
green, Fig. 13), and a broad area of colluvial and alluvial
material in the southeastern part, which was derived from
granitic –granodioritic rocks in the Ivanpah Mountains (GQ,
green, Fig. 13; Fig. 2). The cyan class consists mainly of
gneissic rocks of intermediate composition and mixtures of
carbonate and silicate rock colluvium and alluvium (cyan,
GI and MQ, respectively, Fig. 13). The fifth class comprises
carbonate rocks, which is masked in Fig. 13 (dark blue).
5. Summary of spectral reflectance and spectral
emittance results
Analysis of ASTER spectral reflectance data provides a
basis for mapping limestone and dolomite exposures, dis-
tinguishing Al-muscovite from Fe-muscovite, delineating
rocks containing Fe,Mg –O –H minerals, such as chlorite,
hornblende, epidote and biotite, and Al– O – H minerals and
skarn deposits with associated marble. Delineation of the
REE-rich deposit at the Sulphide Queen mine was limited by
the narrowness of the diagnostic Nd3 + absorption bands and
the breadth of the ASTER VNIR bandpasses. Lower REE
concentrations might not be detectable in ASTER images.
Spectral emittance data analysis complements the spec-
tral reflectance results by mapping five lithologic classes,
each of which typically include a range of lithologies.
Quartzose rocks, including quartzite, sandstone and silici-
fied rocks, are especially well defined (red, Fig. 13), but the
other classes appear to encompass wider lithologic ranges.
Fig. 13. Spectral angle mapper classification of the study area based on the six ASTER image endmember emissivity spectra shown in Fig. 12. Lavender
represents dense vegetation, and dark blue represents MNF band 1 mask. Areas discussed in text: (A1) hydrothermally altered rocks in the Colosseum mine;
(B1) Morning Star mine; (Q) quartzite and siltstone; (GD) granitic to granodioritic intrusive rocks; (GQ) granitic and granodioritic alluvial and colluvial
deposits; (SH1) argillaceous rocks; (MQ) mixtures of carbonate rock and silicate rock colluvial and alluvial materials; (GI) intermediate composition gneissic
rocks.
365
In general, the classes coded yellow (granitic), green (gran-
odioritic) and cyan (intermediate and mafic) represent pro-
gressively more mafic rocks. However, small exposures of
mafic rocks are not well mapped probably because of the
90-m resolution pixels.
Acknowledgements
We thank James Crowley and Bernard Hubbard for
suggestions which improved this manuscript substantially.
This evaluation of ASTER data was supported by National
Aeronautics and Space Administration Contract Number
S46458-E.
References
Boardman, J. W., & Kruse, F. A. (1994). Automated spectral analysis: a
geologic example using AVIRIS data, north Grapevine Mountains, Ne-
vada. Proceedings of the Tenth Thematic Conference on Geologic Re-
mote Sensing ( pp. I407 – I418). Ann Arbor: Environ. Res. Inst. Of
Michigan.
Boardman, J. W., Kruse, F. A., & Green, R. O. (1995). Mapping target
signatures via partial unmixing of AVIRIS data. Proceedings of the Fifth
JPL Airborne Earth Science Workshop. JPL Publication, vol. 95-01
(pp. 23 – 26). Pasadena, CA: Jet Propul. Lab.
Castor, S. B. (1991). Rare earth deposits in the southern great basin. Geol-
ogy and Ore Deposits of the Great Basin ( pp. 523 – 528). Reno: Geol.
Soc. of Nevada.
Clark, R. N., Swayze, G. A., Gallagher, A. J., Gorelick, N., & Kruse, F. A.
(1991). Mapping with imaging spectrometer data using the complete
band shape least-squares fit to multiple spectral features from multiple
materials. Proceedings of the Third AVIRIS Workshop. JPL Publication,
vol. 91-28 (pp. 2 – 3). Pasadena, CA: Jet Propul. Lab.
Crowley, J. K. (1993). Mapping playa evaporite minerals with AVIRIS
data: a first report from Death Valley, CA. Remote Sensing of Environ-
ment, 44, 337 – 356.
Crowley, J. K., Brickey, D. W., & Rowan, L. C. (1989). Airborne imaging
spectrometer data of the Ruby Mountains, Montana: mineral discrimi-
nation using relative absorption band-depth images. Remote Sensing of
Environment, 29, 121 – 134.
Dewitt, E., Kwak, L., & Zartman, R. E. (1987). U – Th – Pb and Ar/Ar
dating of the Mountain Pass carbonatite and alkalic rocks, S.E. CA.
Geological Society of America, Abstracts with Programs, 19(7), 642.
Fujisada, H. (1995). Design and performance of ASTER instrument. Pro-
ceedings of SPIE, the International Society for Optical Engineering,
2583, 16 – 25.
Gillespie, A. R., Kahle, A. B., & Walker, R. E. (1986). Color enhancement
of highly correlated images: I. Decorrelation and HSI contrast straeches.
Remote Sensing of Environment, 20, 209 – 235.
Gillespie, A. R., Matsunaga, T., Rokugawa, S., & Hook, S. J. (1998).
Temperature and emissivity from Advanced Spaceborne Thermal Emis-
sion and Reflection Radiometer (ASTER) images. IEEE Transactions
on Geoscience and Remote Sensing, 36, 1113 – 1126.
366 L.C. Rowan, J.C. Mars / Remote Sensing of Environment 84 (2003) 350–366
Green, A. A., Berman, M., Switzer, B., & Craig, M. D. (1988). A trans-
formation for ordering multispectral data in terms of image quality with
implications for noise removal. IEEE Transactions on Geoscience and
Remote Sensing, 26(1), 65 – 74.
Harsanyi, J. C., & Chang, C. (1994). Hyperspectral image classification
and dimensionality reduction: an orthogonal subspace projection ap-
proach. IEEE Transactions on Geoscience and Remote Sensing, 32,
770 – 785.
Hunt, G. R. (1977). Spectral signatures of particulate minerals in the visible
and near infrared. Geophysics, 42(3), 501 – 513.
Hunt, G. R., & Salisbury, J. W. (1974). Mid-infrared spectral behavior of
igneous rocks. Technical Report AFRCL-TR-75-0356, US Air Force
Cambridge Research Laboratory, Cambridge, MA.
Hunt, G. R., & Salisbury, J. W. (1975). Mid-infrared spectral behavior of
sedimentary rocks. Technical Report AFRCL-TR-75-0356, US Air
Force Cambridge Research Laboratory, Cambridge, MA.
Hunt, G. R., & Salisbury, J. W. (1976). Mid-infrared spectral behavior of
metamorphic rocks, Technical Report AFRCL-TR-76-0003, US Air
Force Cambridge Research Laboratory, Cambridge, MA.
Hunt, G. R., Salisbury, J. W., & Lenhoff, C. R. (1972). Visible and near-
infrared spectra of minerals and rocks: V. Halides, phosphates, arsen-
ates, vanadates, and borates. Modern Geology, 3, 121 – 132.
Knipling, E. B. (1970). Physical and physiological basis for the reflectance
of visible and near-infrared radiation from vegetation. Remote Sensing
of Environment, 1, 155 – 159.
Kruse, F. A. (1988). Use of airborne imaging spectrometer data to map
minerals associated with hydrothermally altered rocks in the northern
Grapevine Mountains, Nevada – California. Remote Sensing of Environ-
ment, 24, 31 – 52.
Kruse, F. A., Lefkoff, A. B., Boardman, J. B., Heidebreicht, K. B., Shapiro,
A. T., Barloon, P. J., & Goetz, A. F. H. (1993). The Spectral Image
Processing System (SIPS)—interactive visualization and analsis of
imaging spectrometer data. Remote Sensing of Environment, 44,
145 – 163.
Lyon, R. J. P. (1964). Evaluation of infrared spectrophotometery for com-
positional analysis of lunar and planetary soils, II. NASA Contract.
Rep. NASA CR-100.
Olson, J. C., Shawe, D. R., Pray, L. C., & Sharp, W. (1954). Rare-earth
mineral deposits of the Mountain Pass, District, San Bernardino County,
CA. U.S. Geological Survey Professional Paper, 261, 1 – 75.
Rowan, L. C., Bowers, T. L., Crowley, J. K., Anton-Pacheco, C., Gumiel,
P., & Kingston, M. J. (1995). Analysis of Airborne Visible-Infrared
Imaging Spectrometer (AVIRIS) data of the Iron Hill, CO, carbona-
tite – alkalic igneous rock complex. Economic Geology, 1966 – 1982.
Rowan, L. C., Clark, R. N., & Green, R. O. (1996). Mapping minerals in
the Mountain Pass, California area using Airborne Visible – Infrared
Imaging Spectrometer (AVIRIS) data. Proceedings of the 11th Confer-
ence on Geologic Remote Sensing, vol. 1 (pp. I177 – I178). Ann Arbor,
Michigan: Environmental Institute of Michigan (ERIM).
Rowan, L. C., Goetz, A. F. H., & Ashley, R. P. (1977). Discrimination of
hydrothermaly altered rocks and unaltered rocks in visible and near
infrared multispectral images. Geophysics, 42(3), 522 – 535.
Rowan, L. C., Kingston, M. J., & Crowley, J. K. (1986). Spectral reflec-
tance of carbonatite and related alkalic igneous rocks from four North
American localities. Economic Geology, 81, 857 – 871.
Rowan, L. C., Wetlaufer, P. H., Goetz, A. F. H., Billlingsley, F. C., &
Stewart, J. H. (1974). Discrimination of rock types and detection of
hydrotermally altered areas in south – central Nevada. U.S. Geological
Survey Professional Paper, 883, 1 – 35.
Salisbury, J. W., & Milton, N. M. (1987). Preliminary measurements of the
spectral signatures of tropical and temperate plants in the thermal infra-
red. Proceedings of the 5th Conference on Remote Sensing and Ex-
ploration Geology, vol. I (pp. 131 – 143). Ann Arbor, Michigan:
Environmental Institute of Michigan (ERIM).
Vane, G., Green, R. O., Chrien, T. G., Enmark, H. T., Hansen, E. G., &
Porter, W. M. (1993). The airborne/infrared imaging spectrometer
(AVIRIS). Remote Sensing of Environment, 44, 127 – 143.
Further reading
Green, R.O. (1991). Retrival of reflectance from AVIRIS radiance using a
radiative transfer code. In R.O. Green (Ed.), Proceedings of the Third
AVIRIS Workshop. JPL Publication, vol. 91-28 (pp. 200 – 210). Pasa-
dena, CA: Jet Propul. Lab.
Salisbury, J. W., Walter, L. S., & D’Aria, D. (1991). Mid-infrared (2.5 to
13.5 micrometers) spectra, U.S. Geological Survey Open-file Report 88-
686, 1988. U.S. Geological Survey, Evaluation of metallic mineral
resources and their geologic controls in the Eastern Mojave National
Scenic Area, San Bernardino County, CA, Open-file Report 91-427,
278 pp.