USGS logo : science for a changing world

Descriptions, Spectral Plots, and Digital Reflectance Spectra of Samples Applied to Spectral Analysis of Imaging Spectroscopy Data:

Utah (East Tintic Mountains, Oquirrh Mountains, Wasatch Mountains, and Tushar Mountains), Nevada (Goldfield Hills), and New Mexico (Jemez Mountains), USA, 1999-2002

by Barnaby W. Rockwell1

Open-File Report 02-407

Version 1.2

2002  





This report is preliminary and has not been reviewed for conformity with U.S. Geological Survey editorial standards or with the North American Stratigraphic Code. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

U.S. DEPARTMENT OF THE INTERIOR
U.S. GEOLOGICAL SURVEY

1barnabyr@usgs.gov, Denver, Colorado



This report documents rock samples that were analyzed in the laboratory as a part of several projects, including the USGS/USEPA Utah Abandoned Mine Lands Imaging Spectroscopy Project. These samples are intimate mixtures of various minerals associated with hydrothermal alteration and supergene weathering.  Reflectance spectra of these samples were used as standard reference spectra for the spectral analysis of AVIRIS data covering parts of Utah, Nevada (Goldfield Hills, Esmeralda County), and New Mexico (Jemez Mountains near Los Alamos). Sample compositions were determined using bulk X-ray Diffraction (XRD) analysis, the results of which are listed in the description pages for each sample.

The samples where spectrally characterized in the laboratory using an ASD FieldSpec Pro FR spectrometer over the 0.35 to 2.50 µm range. Forty spectral measurements were integrated to generate each saved spectrum, and from 10 to 15 saved spectra were averaged to generate the spectra published here. The ASD FR spectrometer has three separate detector arrays that acquire data from different wavelength regions of the electromagnetic spectrum (array 1: VNIR, 0.35-1.05 µm; array 2: SWIR1, 1.00-1.80 µm; and array 3: SWIR2, 1.80-2.50 µm). Offsets in reflectance level between data acquired by each of the three detector arrays were computed and removed by extrapolating the linear trends from several channels at each end of the more stable SWIR1 array, calculating the offsets between the extrapolated trends and the first channels of the adjacent arrays, and adding the offsets to adjust all data from the adjacent arrays. The resultant spectra were converted to absolute reflectance (ABS REF) by multiplying the reflectance spectrum of the Spectralon white reference standard by each measured average sample spectrum.  The ABS REF spectra (717 channels) were convolved to 2151 channels to increase the trapezoidal integration for the final convolutions to the wavelength and bandpass of the AVIRIS instrument. All spectral convolutions were performed using cubic spline algorithms (Carnahan and Wilkes, 1973).

The spectra published herein were incorporated into the Tetracorder expert system (Clark and others, 2002), which was used for materials identification from laboratory and imaging spectroscopy data. One or more diagnostic spectral absorption features were identified in each spectrum for analysis by the Tetracorder system. The wavelength intervals of these absorption features are indicated on the description page of each sample, and are defined by two endpoints, each of which is represented by a short wavelength interval (usually 0.03 microns) over which spectral data are averaged to reduce effects of noise. The wavelength intervals defining an absorption feature represent the spectral regions over which a modified least-squares curve-fitting algorithm compared the reference spectra to AVIRIS image spectra. After the absorption features are isolated from broader, more general curve shapes using continuum removal (Clark, 1999), each reference spectrum was scaled to match the feature depth of the AVIRIS spectra. Curve-fitting was then performed between reference and AVIRIS spectra, generating a value of “fit” for each absorption feature. The degree of curve "fit" for each diagnostic feature was weighted based on the relative area of that feature beneath the continuum, with weights of all diagnostic features summing to unity. Therefore, the “fit” values of deep and broad features with large areas had more influence on material identification than the “fits” of features with smaller areas. The weighted curve "fits" for each diagnostic feature were then summed to derive a single composite "fit" value for each reference spectrum which was compared to the composite "fit" values of other materials characterized by vibrational absorptions in the 1.4 – 2.5 µm spectral region. The material with the largest composite "fit" for this group of materials was selected as the identified material.

Some diagnostic absorption features are shallow in depth, narrow in width, and/or occur in wavelength regions subject to low signal-to-noise in data from a particular sensor, and are, thus, not ideal for curve fitting. These features were identified as "weak" and were utilized by the Tetracorder expert system in a manner different from that used for the principal diagnostic features. In the analysis of AVIRIS spectra, the “weak” features had to exist above a globally-defined “fit” threshold for the composite “fit” of a material to be included in the selective competition between materials leading to an identification, but the “fit” value for the “weak” feature was not included in the composite “fit” for that material. Features designated as "weak" are indicated with a "W" on the description pages. An example of a "weak" feature is the jarosite feature at 0.43 µm evident in sample BRCM-1.

NOTE: All latitude/longitude coordinates for sample collection locations are relative to the NAD 27 horizontal datum.



Title

Description

Spectral Plot
(0.35 - 2.50 µm)

Ascii Spectrum
for Download

Natrojarosite + Gypsum + Illite BRCM-1

Description

Small
Large

brcm1.txt

K Alunite + Muscovite + Pyrophyllite + Jarosite GF00-10

Description

Small
Large

gf00-10.txt

Pyrophyllite + Muscovite JH_PYRM1

Description

Small
Large

jh_pyrm1.txt

Na Alunite + Kaolinite MV00-11a

Description

Small
Large

mv00-11a.txt

K Alunite + Kaolinite MV2-AR3

Description

Small
Large

mv2-ar3.txt

Na Alunite + Dickite MV99-6-26b

Description

Small
Large

mv99-6-26b.txt

Talc + Calcite PC99-1G

Description

Small
Large

pc99-1g_talc.txt



References:

Carnahan, B., and Wilkes, J.O., 1973, Digital Computing and Numerical Methods: John Wiley & Sons, New York, p. 307.

Clark, R.N., 1999, Spectroscopy of rocks and minerals, and principles of spectroscopy:  in Rencz, A.N., ed., Remote Sensing for the Earth Sciences,  Manual of Remote Sensing, Ryerson, R.A., ed., Volume 3, John Wiley & Sons, Inc., New York, pp. 3-58, see http://speclab.cr.usgs.gov/PAPERS.refl-mrs/refl4.html.

Clark, R. N., Swayze, G.A., Livo, K.E., Kokaly, R.F., Sutley, S.J., Dalton, J.B., McDougal, R.R., and Gent, C.A., 2002, Imaging Spectroscopy: Earth and Planetary Remote Sensing with the USGS Tetracorder and Expert Systems: Journal of Geophysical Research, in press.


USA.gov logo