The aim of this study was to develop near infrared spectroscopy(NIRS)calibrations to predict quality parameters,dry matter(DM,g kg1)and crude protein(CP,g kg1 DM),in fresh un-dried grass.Knowledge of these parameters ...The aim of this study was to develop near infrared spectroscopy(NIRS)calibrations to predict quality parameters,dry matter(DM,g kg1)and crude protein(CP,g kg1 DM),in fresh un-dried grass.Knowledge of these parameters would enable more precise allocation of quality herbage to grazing livestock.Perennial ryegrass samples(n=1615)were collected over the 2017 and 2018 grazing seasons at Teagasc Moorepark to develop a NIRS calibration dataset.Additional samples were collected for an independent validation dataset(n=197)during the 2019 grazing season.Samples were scanned using a FOSS 6500 spectrometer at 2 nm intervals in the range of 1100~2500 nm and absorption was recorded as log 1/Reflectance.Reference wet chemistry analysis was carried out for both parameters and the resultant data were calibrated against spectral data by means of modified partial least squares regression.A range of mathematical spectral treatments were examined for each calibration,which were ranked in order of standard error of prediction(SEP)and ratio of percent deviation(RPD).Best performing calibrations achieved high predictive precision for DM(R2=0.86 SEP=9.46 g kg1,RPD=2.60)and moderate precision for CP(R2=0.84 SEP=20.38 g kg1 DM,RPD=2.37).These calibrations will aid the optimisation of grassland management and the development of precision agricultural technologies.展开更多
文摘The aim of this study was to develop near infrared spectroscopy(NIRS)calibrations to predict quality parameters,dry matter(DM,g kg1)and crude protein(CP,g kg1 DM),in fresh un-dried grass.Knowledge of these parameters would enable more precise allocation of quality herbage to grazing livestock.Perennial ryegrass samples(n=1615)were collected over the 2017 and 2018 grazing seasons at Teagasc Moorepark to develop a NIRS calibration dataset.Additional samples were collected for an independent validation dataset(n=197)during the 2019 grazing season.Samples were scanned using a FOSS 6500 spectrometer at 2 nm intervals in the range of 1100~2500 nm and absorption was recorded as log 1/Reflectance.Reference wet chemistry analysis was carried out for both parameters and the resultant data were calibrated against spectral data by means of modified partial least squares regression.A range of mathematical spectral treatments were examined for each calibration,which were ranked in order of standard error of prediction(SEP)and ratio of percent deviation(RPD).Best performing calibrations achieved high predictive precision for DM(R2=0.86 SEP=9.46 g kg1,RPD=2.60)and moderate precision for CP(R2=0.84 SEP=20.38 g kg1 DM,RPD=2.37).These calibrations will aid the optimisation of grassland management and the development of precision agricultural technologies.
