IEEE Geoscience and Remote Sensing Magazine - June 2019 - 62

formed under different conditions than the assemblages
found in the altered granitic wall rock. The two disparate
formations may represent different alteration zones that
resulted from two distinct hydrothermal alteration episodes, where the less-ordered Al-poor illite-muscovite and
kaolinite formed first by pervasive alteration of the wall
rock and, in a later episode under higher-temperature conditions, the quartz veins formed. Alternatively, the two
different assemblages may have resulted from disparate
physicochemical conditions between the fluid conduits
and the less permeable wall rock during the same hydrothermal event.
The strength of the Wavelength Mapper, as shown in
this study, is the provision of a synoptic view of the compositional information together with the microstructural
information in one color image. The linkage enables the
interpretation of the compositional variation in relation to
the spatial patterns present in the fabric of the rock. The
method's capability of visualizing small shifts in wavelength positions related to changes in mineral chemistry is
another useful feature of the technique.
A limitation of the Wavelength Mapper as used here
is that it shows only the position and depth of single features and does not take into account the shape of the feature and the occurrence of other absorption features at
other wavelength positions. This makes it impossible to
differentiate among minerals with overlapping deepest
features, such as illite-muscovite, kaolinite, and tourmaline, from the wavelength map alone. To discriminate among these minerals, the wavelength positions
of the second-deepest features could be used because
they differ for these three minerals (Table 3). Minerals with shallow absorption features are not visible in
the wavelength map because of the default stretching
interval of the depth image of the absorption feature.
This results in low intensity levels of pixels that contain

shallow features in the wavelength map. A different
choice of the stretching interval for the creation of the
wavelength map, where the shallow features display in
higher brightness levels, will solve this problem.
We conclude that the Wavelength Mapper combines key
compositional and microstructural rock information in a
single color image, which enables a first assessment of the
mineral diversity and rock-forming processes that created
the different minerals and assemblages in a specimen. The
construction and interpretation of the wavelength map is a
first step in the interpretation workflow of a hyperspectral image. The results can be further used as the selection
of image endmember spectra for input into a classification
procedure and to obtain a mineral map.
APPLICATION TO AIRBORNE HYPERSPECTRAL DATA
In this case study, we apply IDL DISPEC and the Wavelength Mapper to highlight mineral alteration patterns in
airborne AVIRIS-NG data over Cuprite, Nevada. Cuprite
has a long tradition of spectral investigations (e.g., [53]-
[57]). Cuprite is particularly useful for testing hyperspectral airborne sensors and algorithms because of the area's
diverse alteration mineralogy, limited vegetation cover,
and excellent outcrop situation. From the literature (e.g.,
[57]), it is known that the area contains various iron-,
Al-OH-, and iron/Mg-OH-bearing minerals that show different spectral positions in their respective absorption feature wavelengths, depending on mineralogy and mineral
chemistry, respectively. We will use IDL DISPEC to check
at which wavelength we can expect the spectral features
and will employ the Wavelength Mapper to highlight the
mineralogy and mineral chemistry spatial patterns in the
airborne data.

Relative Frequency (per Bin)

STUDY AREA
Cuprite is in south-central Nevada, roughly 200 km northwest of Las Vegas. The Cuprite study
area contains advanced argillic alteration in a western and an east0.06
ern center (Figure 11), which are
separated by U.S. Highway 95 and
a 2-km-wide pediment of Quater0.04
nary age. The exposed lithologies in
the western center consist mainly of
0.02
Cambrian metasediments and Tertiary conglomerates, while the east0
ern center is composed mostly of
0
5
10
15
20
25
Tertiary ash-flow and air-fall tuffs.
Crystallinity Value
A late Miocene hydrothermal altera(b)
(a)
tion caused pervasive alteration on
both sides of the highway to mineral
zones of advanced argillic hydrated
FIGURE 10. (a) A crystallinity map (i.e., Depth2,200/Depth1,900) showing the ordering of
silica and alunite, as well as to argilillite-muscovite minerals in gray scale and linear stretch. Low values are in dark tones, and
lic kaolinite and white mica. Addihigh values bright. The image was calculated by dividing the depth of the Al-OH feature near
tionally, there is propylitic chlorite
2,200 nm by the depth of the water feature near 1,900 nm. (b) A histogram of the crystallinin the western center and a relict of
ity values in (a).
62

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE

JUNE 2019

IEEE Geoscience and Remote Sensing Magazine - June 2019

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - June 2019