Analysis of Forest Foliage Using a Multivariate Mixture Model

Abstract

Data with wet chemical measurements and near infrared spectra of ground leaf samples were analyzed to test a multivariate regression technique for estimating component spectra which is based on a linear mixture model for absorbance. The resulting unmixed spectra for carbohydrates, lignin, and protein resemble the spectra of extracted plant starches, cellulose, lignin, and protein. The unmixed protein spectrum has prominent absorption spectra at wavelengths which have been associated with nitrogen bonds.