Analysis and prediction of natural pasture CNCPS (cornell net carbohy-drate and protein system) components by near infrared reflectance spec-troscopy (NIRS)
Sixty-six samples collected from alpine grassland of Kobresia hastily in Henan county, Qinghai province were used to investigate the feasibility of predicting the CNCPS ( cornell net carbohydrate and protein system) composition of natural pasture by near infrared reflectance spectroscopy (NIRS). Using modified partial least squares (MPLS) regression method, the models of CP(crude protein), SP(soluble protein), PA(non-protein nitrogen), PB1(rapidly degradable crude protein), PB2(intermediately degradable crude protein), PB3(slowly degradable crude protein), PC(bound crude protein), and models of NDF( neutral detergent fiber) , ADF( acid detergent fiber) , ADL( acid detergent lignin) , CHO( total carbohy-drate ) , CNSC ( non-structural carbohydrates ) , CA (sugars), CB1(starch and pectin), CB2(available fiber), CC(not available fiber) were built. The results showed 1-VR ( cross validation determination coefficient) for CP, PC, NDF, ADF, CHO, CNSC, and CA were 0. 989, 0. 870, 0. 975, 0. 932, 0. 964, 0. 966 and 0. 846, respectively, and RPDCV ( ratios of standard deviation of reference analysis data to SECV) were 9. 336, 2. 913, 6. 353, 3. 758, 5. 306, 5. 521, and 2. 603, respectively. Models with 1-VR less than 0. 9 and RPDCV less than 2. 5 were not ideal. The results indicated that CNCPS components of natural pasture could be fastly and accurately predicted by NIRS, and the models established were applicable for the predictions of CP, PC, NDF, ADF, CHO, CNSC and CA. Further studies should be focusing on improving the precision of the models for PA, PB1 , PB2 , PB3 , CB1 , CB2 , and CC.
near infrared reflectance spectroscopy(NIRS)natural pasturecornell net carbohydrate and protein system(CNCPS)nutritional value
NETL
NSTL
万方数据
维普
