Near-infrared spectroscopy calibration strategies to predict multiple nutritional parameters of pasture species from different functional groups

Abstract

Near-infrared reflectance spectroscopy (NIRS) has been used by the agricultural industry as a high-precision technique to quantify nutritional chemistry in plants both rapidly and inexpensively. The aim of this study was to evaluate the performance of NIRS calibrations in predicting the nutritional composition of ten pasture species that underpin livestock industries in many countries. These species comprised a range of functional diversity (C3 legumes; C3/C4 grasses; annuals/perennials) and origins (tropical/temperate; introduced/native) that grew under varied environmental conditions (control and experimentally induced warming and drought) over a period of more than 2 years (n = 2,622). A maximal calibration set including 391 samples was used to develop and evaluate calibrations for all ten pasture species (global calibrations), as well as for subsets comprised of the plant functional groups. We found that the global calibrations were appropriate to predict the six key nutritional quality parameters studied for our pasture species, with the highest accuracy found for ash (ASH), crude protein (CP), neutral detergent fibre and acid detergent fibre (ADF), and the lowest for ether extract (EE) and acid detergent lignin parameters. The plant functional group calibrations for C3 grasses performed better than the global calibrations for ASH, CP, ADF and EE parameters, whereas for C3 legumes and C4 grasses the functional group calibrations performed less well than the global calibrations for all nutritional parameters of these groups. Additionally, our calibrations were able to capture the range of variation in forage quality caused by future climate scenarios of warming and severe drought.