EU en RR e QU UH A eh ud Bie
GEOBOTANICAL TECHNIQUES FOR DISCRIMINATING SERPENTINE ROCK TYPES
IN WESTERN UNITED STATES
D. MOUAT, NASA-Ames Research Center, Moffett Field, CA, U.S.A.
C. ELVIDGE, R. Lyon, Stanford University, Stanford, CA, U.S.A.
Serpentine-derived soils have a significant affect on species composition,
vegetation density, and vegetation spectral response as a result of several
factors including low calcium/magnesium ratios and high concentrations of
chromium, cobalt, and nickel.
Remote sensing techniques involving airborne scanner imagery and several
statistical and image processing techniques were used in three diverse test
sites in western United States to discriminate vegetation parameters
associated with serpentine rock types.
An important area of ultramafic lithology containing economic deposits of
chromium and nickel occurs within the Siskiyou Mountains of southwest Oregon
and northwest California. This region of high annual precipitation is
characterized by dense conifer vegetation on non-ultramafic volcanics and
sedimentary rocks. The ultramafic rock types, however, are characterized
by unique vegetation assemblages often consisting of grassland or scrubland
types which have only a sparse conifer cover. These types are easily
discriminated with the airborne scanner imagery employing relatively simple
analytical techniques. More sophisticated analytical techniques, however,
were neded to differentiate rock units (serpentine and non-serpentinized
peridotites) from within the ultramafics based upon their vegetation char-
acteristics. The differentiation of these rock units is important in the
exploration of economic deposits of chromium and nickel.
The other serpentine-bearing regions were examined in central California. One,
situated near the coast, is located near Stanford University's Jaspar Ridge
Biological Preserve. The other is located approximately 250 kilometers to the
east in the foothills of the Sierra Nevada. In both sites, multidate airborne
scanner imagery was useful in discriminating the vegetation associated with
the serpentine rock types on the basis of plant phenology as well as anomalous
vegetation occurrences. Step-wise discriminant analysis was used to select
the best multispectral channels for discriminating the serpentine areas. On
account of phenological changes, the best wavelengths for separating the
rock types varied.
499
= ER ERE cR lett
EPA E E