天然牧草营养成分的近红外光谱定量分析

Quantitative Analysis of Nutrition Composition of Native Grasses by Near Infrared Reflectance Spectroscopy

为探讨近红外光谱技术(NIRS)定量分析天然牧草营养成分的可行性,采集高寒草甸类高山嵩草草地天然牧草样品210份,用修正偏最小二乘法(MPLS),建立天然牧草中粗蛋白(CP)、中性洗涤纤维(NDF)、粗脂肪(EE)的近红外定量分析模型。结果表明:CP、NDF模型的交叉验证决定系数(1-VR)分别为0.9886、0.9554,交叉验证标准误差(SECV)分别为0.4393、1.7169,交叉验证相对分析误差(RPDCV)分别为9.3355、4.7194;EE模型的SECV、1-VR分别为0.6264、0.668,模型有待于进一步优化。试验初步建立了河南县高山嵩草草地天然牧草CP、NDF、EE定量分析模型,为青海省各地区天然牧草营养价值数据库的建立提供了基础数据。

In order to evaluate the potential of using near infrared reflectance spectroscopy （NIRS） in analyzing the nutrition composition of native grasses quantitatively, content of CP, NDF and EE in 210 native grass samples were tested by modified partial least square （MPLS） regression. The results showed that cross validation determination coefficients （1-VR） of CP and NDF were 0.9886 and 0.9554, standard errors of cross validation （SECV） were 0.4393 and 1.7169, and ratios of standard deviation of reference analysis data to SECV （RPDCV） were 9.3355 and 4.7194, respectively. With low value of 1-VR （0.668） and SECV （0.6264）, the model of EE needed to be further optimized. This study preliminarily established model of CP, NDF and EE in native Kobresia in Henan County, and provided basic data for establishing database of native grasses in different areas in Qinghai Province.