Journal of Rangeland

There are various methods to measure forage quality but these methods require labor and are expensive and time consuming. Near infrared reflectance spectroscopy (NIRS) has been used to predict forage quality since 1970. This method is a reliable and fast method for measuring forage quality but it requires appropriate calibration instruments with adequate samples. The aim of this study was to develop optimal NIRS calibrations for predicting forage quality of forb species. In order to predict N, CP, ADF, DMD and ME content of forb species through NIRS, a total of 641 samples from three phenological stages (vegetative, flowering and seeding) were scanned by DA7200 Perten instrument. Spectral data was exported into the Unscrambler software to conduct multivariate analysis. The spectral data was smoothed and normalized by S.Golay and SNV methods respectively. Calibration models were developed through PLS1 method and then evaluated based on R², RMSECV and RPD values. According to the results, the nitrogen and CP contents of forb species can be measured more appropriately than other forage quality indicators. This study indicated that the forage quality of forbs species can be determined by NIRS method, however the accuracy values (PRD values of validation) obtained from this technique in some cases were less than the optimum thresholds.