The ASTER dataset used to cover the central and southern parts of
the Basin and Range province in the United States consists of 247
ASTER_Level 1B scenes. The ASTER_Level 1B radiance data consist of
all VNIR, SWIR and TIR bands, including the backward-looking VNIR
band. The VNIR and SWIR radiance data were calibrated to the
reflectance data using Moderate Resolution Imaging Spectrometer
(MODIS) water-vapor data, radiometric "crosstalk" correction
software, radiance "gain" correction coefficients, and Atmospheric
Correction Now (ACORN) software. The TIR radiance data were
calibrated to emissivity using atmospheric removal and emissivity
normalization algorithms in Environment for Visualizing Images
(ENVI), an image calibration and processing software package. Each
ASTER scene was georectified to a Landsat TM 30 m orthorectified
image with a root mean square error of less than 60 m.
Rocks containing hydrous quartz, chalcedony, opal, and amorphous
silica (hydrothermalsilicarich rocks), calcite-dolomite and
epidote-chlorite (propylitic), alunite-pyrophyllite-kaolinite
(argillic), and sericite-muscovite (phyllic) were mapped using
Interactive Data Language (IDL) logical operators (Details in
report table 1). The IDL logical operators consist of band
thresholds and band ratios strung togetherto map spectral
absorption features of minerals. All of the logical operator
algorithms mask green vegetation using a ratio of VNIR band 3/band
2, which detectsthe chlorophyll absorption feature at 0.67 µm. In
addition, a band 4 threshold is used to mask low signals, and thus
noisy spectra.
Hydrothermally altered phyllic and argillic rocks were mapped using
ASTER VNIR and SWIR data at 30-m spatial resolution. SWIR band
ratios were used in IDL logical operators to map Al-O-H spectral
absorption features associated with alunite, kaolinite, and
sericite-muscovite. The ratio of SWIR band 4/band 5 maps the 2.
165-µm spectral absorption feature associated with alunite,
pyrophyllite, and kaolinite, and the ratios of SWIR band 4/band 6
and band 7/band 6 map the 2.2-µm spectral absorption feature
exhibited by alunite, kaolinite, and sericite.
Hydrothermal silica-rich rocks were mapped using ASTER SWIR and TIR
band ratios at 90-m resolution. The ratio of SWIR band 4/band 7 is
typically higher for hydrothermal silica-rich rocks, which have
lower overall SWIR reflectance in the 2.0- to-2.4-µm region than
nonhydrothermal silica-rich rocks because of residual molecular
water or an O-H absorption feature spanning 2.26 to 2.4 µm. The
ratio of TIR band 13/band 12 maps the 9.09-µm quartz reststrahlen
absorption feature. Thus, silica-rich rocks were mapped using the
TIR emissivity data and hydrothermal silica-rich rocks were
discriminated from the non-hydrothermal silica-rich rocks using the
corresponding SWIR reflectance data for each pixel.
Hydrothermally altered propylitic rocks were mapped using ASTER
SWIR and TIR band ratios at 90-m spatial resolution. Calcite,
dolomite, epidote, and chlorite typically exhibit overlapping CO3
and Fe,Mg-O-H spectral absorption features at 2.31 to 2.33 µm and
have been difficult to map separately using SWIR data in previous
studies. In the TIR region, however, rocks containing calcite and
dolomite exhibit an 11.2-µm spectral absorption feature, whereas
epidote- and chlorite-rich rocks have TIR spectral absorption
features centered at approximately 10.2 µm. Thus, TIR calcite-
dolomite spectra exhibit higher band-13 emissivity and lower band-
14 emissivity, whereas epidote-chlorite TIR spectra exhibit lower
band-13 emissivity and higher band-14 emissivity. The calcite-
dolomite and epidote-chlorite logical operators use the ratio of
SWIR band 6/band 8 to map the 2.31- to 2.33-µm absorption feature
of both groups of minerals and the ratio of TIR band 13/band 14 set
to greater than 1.005 to separate and map calcite-dolomite-rich
rocks and less than 0.999 to separate and map epidote-chlorite-rich
rocks.
The argillic and phyllic units are mapped at 30-m spatial
resolution and the hydrothermal silicarich rocks and propylitic
units are mapped at 90-m spatial resolution. Each logical operator
algorithm produces an image with pixel values of one or zero which
is converted to a polygon-based shapefile for import into ArcGIS.
The logical operators correctly mapped (with ±20 percent error)
hydrothermal alteration at three calibration-validation test sites
on the basis of field data and mineral maps compiled from previous
studies.
