Prediction of browse nutritive attributes in a Pinus radiata D. Don silvopastoral system based on visible-near infrared spectroscopy

Browse is an important source of food for rustic livestock, particularly when herbaceous forage is scarce. The Atlantic Basque Country (Northern Spain) forest landscape is dominated by Pinus radiata D. Don plantations where Rubus sp. and Ulex gallii are understorey dominant species. Knowledge of the nutritive value of these species is needed in the context of silvopastoralism, primarily because do not always meet livestock requirements. The objective of this study was to evaluate the potential of Visible-Near Infrared Spectroscopy to determine the quality attributes of Rubus sp. and U. gallii, using a Sample Turn Table probe to acquire spectra on non-dried samples. VIS-NIRS calibrations were developed for dry matter (DM), crude protein (CP), crude fibre (CF), neutral detergent fibre (NDF), acid detergent fibre (ADF) and ashes. Spectra were pre-treated and Partial Least Squares Regression models were constructed. Calibration models were accurate for most of the considered variables, both for Rubus sp. (DM: Rc(2)=0.95, RPD=2.53; CP: Rc(2)=0.90, RPD=2.39; CF: Rc(2)=0.86, RPD=2.30; NDF: Rc(2)=0.93, RPD=2.80; ADF: Rc(2)=0.95, RPD=3.12; Ashes: Rc(2)=0.91, RPD=2.15) and for U. gallii (DM: Rc(2)=0.98, RPD=3.67; CP: Rc(2)=0.94, RPD=1.84; CF: Rc(2)=0.98, RPD=4.74; NDF: Rc(2)=0.94, RPD=3.91; ADF: Rc(2)=0.98, RPD=3.62; Ashes: Rc(2)=0.82, RPD=1.65). In general, ADF and DM were the most accurately predictable variables and ash content, the least predictable one. The results showed VIS-NIRS potential for the rapid and accurate prediction of quality attributes in non-dried samples and proved as a useful tool for making decisions in silvopastoral systems.