Calibrating NIRS Equations For Predicting Chemical Composition And In Vitro Dry Matter Digestibility Of Siberian Wildrye(Elymus Sibiricus)

Siberian wildrye (Elymus sibiricus Linn.) is one of the predominant pasture species in Qinghai-Tibet plateau of China. It supplies a large number of fodders to domestic animals in spring and winter, and provides a large proportion of the summer and autumn grazing in these alpine regions. Crude protein (CP), acid detergent fiber (ADF), neutral detergent fiber (NDF) and in vitro dry matter digestibility (IVDMD) are the most important aspects of nutritive value. The objective of the present work was to develop NIRS equations for predicting chemical composition (CP, ADF and NDF) and in vitro dry matter digestibility of Siberian wildrye and to evaluate the feasibility of using near infrared spectroscopic technique as a reliable and efficient means to determine these values.Samples of Siberian wildrye covering a wide range, which varied in locations, varieties, organs, maturation stages and farm managements, were collected during2012. FT-NIR spectrometer was used to obtain dried and milled samples spectra. The calibration models for the prediction of nutritive attributes (CP, ADF, NDF and IVDMD) are based on correlation between forage spectra and the results of conventional chemical analysis (i.e. reference data). Different spectral pretreatments and regression methods were used to optimize the calibration. The optimum ranges of wavenumber for developing NIRS equations of CP and NDF are4400-4800cm-1,5400-6600cm-1and7800-10000cm-1.The best spectroscopy for ADF is5000-7144cm-1and7404-10000cm-1. And the best spectroscopy for IVDMD is4000-10000cm-1. The optimum spectral pretreatments for establishing NIRS models of CP, ADF, NDF and IVDMD are smoothing (average3points gap2)+normalization (by closure)+derivatives (first derivative BCAP gap2), standard normal variable+smoothing(average3points)+derivatives(first derivative BCAP gap2), standard normal variable+derivatives (first derivative BCAP gap2) and multiple scatter correction (full)+derivatives (1st BCAP)+normalization(to unit length). The predicted CP, ADF, NDF and IVDMD were accurately calibrated and validated. Coefficient of determination in calibration and standard error of calibration were 0.9945,0.9499,0.9133,0.9822and0.3229%DM,0.7799%DM,1.3430%DM,1.3726%DM, respectively. Coefficient of determination in prediction and standard error of prediction were0.9938,0.9449,0.8907,0.9790and0.3261%DM,0.7878%DM,1.3852%DM,1.4303%DM, respectively. The RPD values for validation are all above3.0.Based on our results, the study has demonstrated that FT-NIR spectroscopic technique associated with chemometrics is sufficiently sensitive to allow the accurate prediction of the concentrations of components (CP, ADF and NDF) and in vitro dry matter digestibility of Siberian wildrye.