Prediction of the chemical composition of white clover by near-infrared reflectance spectroscopy

The aim of this study was to assess near-infrared reflectance spectroscopy (NIRS) as a rapid and reliable method for estimating quality parameters in white clover (Trifolium repens). NIRS calibrations were derived for ash, crude protein (CP), crude fibre (CF), neutral detergent fibre (NDF), acid detergent fibre (ADF) acid detergent lignin (ADL). To maximize variability, the set of samples for analysis was selected from field plots, harvested in two contrasting environmental locations over a period of 5 years. The samples were scanned using a NIR Systems model 5000 monochromator. A total of 145 samples was used to calibrate and crass-validate the equations derived. A separate independent set (sixty-five samples) was used for testing the predictions. Standard errors of calibration and cross-validation, expressed as percentages of the mean, were 3.5 and 5.5 for ash, 3.0 and 4.2 for CP, 4.4 and 5.0 for CF, 3.7 and 4.7 for NDF, 3.3 and 4.2 for ADF and 14 and 16 for ADL respectively. The MRS calibrations obtained from this study have the potential to be used in breeding programmes to assess the quality of white clover and for agronomic evaluations.