Prediction of Botanical Composition Using NIRS Calibrations Developed from Botanically Pure Samples
1990
Coleman, S. W. | Christiansen, S. | Shenk, J. S.
Point quadrat, hand separation, and histological methods for determining botanical composition are tedious, time consuming, and expensive. Near-infrared reflectance spectroscopy (NIRS) was used as a rapid method to predict botanical composition of grass-legume mixtures clipped from field plots. Our objective was to determine if a group of samples, each consisting of one of three botanical components, could be used for calibration in lieu of samples containing mixtures of the components. Calibration equations were developed for NIRS prediction of a number of legumes, Caucasian bluestem [(Bothriochloa caucasica (Trin.) C.E. Hubb.], and cheatgrass (Bromus tectorum L.). Hand separations of clipped mixtures were used to validate the calibration equations. All forage samples were grown in silty soils of the Dale series. Some evidence existed for an interaction between light scatter and mixing of two pure species, but it was eliminated effectively by second derivative transformation of the spectra. Biases existed for all components (36-61 g/kg) but were smaller than standard errors of performance (52-104 g/kg), except for legumes. Linearity was excellent for prediction for percentage legume, and validation was more precise for legume than for Caucasian bluestem or cheatgrass. This method of calibration for a closed population is recommended. Care must be taken to validate equations using mixed samples to insure linearity of unknowns between the extremes. Further, unknowns should be scanned along with calibration samples so that the same day to day instrument variation is associated with both sample sets. Contribution of the USDA-ARS in cooperation with The Pennsylvania State Univ., University Park, PA 16802.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par National Agricultural Library
Découvrez la collection de ce fournisseur de données dans AGRIS