Spectral Unmixing Applied to Desert Soils for the - Naval ...

VII. DISCUSSION AND CONCLUSIONS
A. SPECTRAL MEASUREMENTS AND SOIL DISTURBANCE
Analysis of the collected spectral libraries established that there were measurable
differences between spectra of disturbed versus undisturbed areas of the same
composition materials. Endmembers collected in the field were all categorized by a
material first undisturbed and then disturbed from the same location. The disturbances
were all caused by impact-related use such as walking, driving over, or dropping
something onto the soil, OHV and equestrian use (horses). Firerelated disturbances were
also used in order to have information regarding camp fires to help associate the source
as anthropogenic. For the initial clay collection the samples were all taken from the same
area and spanned a section of roughly 1.62 Km 2 . The continuum removed spectral
differences occur in the wavelength regions of approximately 500 nm, 1200 nm, and
2200 nm. The changes observed at these wavelengths involve differences in apparent
reflectance received at the sensor and calculated differences in band depth and observable
width differences of absorption features in the same material. Similarities in the amount
of difference in depths between ASD collected endmembers and imagery derived
endmembers suggested that these differences made the spectra unique enough to be
successfully used to detect features such as trails based on differences in spectral
characteristics.
Clearly properties of soils are altered once they have been exposed to some
surface disturbance, particularly those caused by impacts. It is the chemical and
structural make-up of soil that causes it to exhibit its characteristic absorption feature.
Previous studies have found similar changes in reflectance and spectra shape, depth, and
width of vegetation exposed to gasses injected into soils. These changes were the result
of a stress response by vegetation to the gasses as a function of distance from the source.
It makes sense then that if reflectance changes in vegetation are a response to an
ecosystem disturbance, then the reflectance changes in this study can be correlated to
anthropogenically caused surface disturbances as well. The data shown in this report
support this line of thinking because the changes in reflectance are similar enough to the
75