Assessment of biophysical soil properties through spectral decomposition techniques

A mixture model was utilized to extract soil biophysical properties from fine resolution soil spectra (400-900 nm) measured outdoors with a portable spectroradiometer. The objective of this study was to fully characterize soil spectral signatures in the visible and near-infrared in terms of underlying "basis" curves, key wavelengths, and dimensionality. Through spectral decomposition and mixture modeling, the reflectance response of a wide, genetic range of soil materials were separated into four independent sources of spectral variability (basis curves), which in linear combination were able to reconstitute the experimental data set. Stepwise spectral reconstruction was then utilized to isolate organic carbon and free iron oxide "basis" curves. This enabled a good "global" measure of soil properties irrespective of soil type or brightness. We anticipate the EOS-MODIS and HIRIS sensors to provide the spectral data needed for inversion of satellite data into soil surface properties and processes. © 1991