规模化奶牛场泌乳牛粪便氮磷含量预测模型研究  

Prediction of Fecal Nitrogen and Fecal Phosphorus Content for Lactating Dairy Cows in Large-scale Dairy Farms

作　　者：渠清博 杨鹏[1] 翟中葳[1] 李爱秀[1] 张克强[1] QU Qing-bo YANG Peng ZHAt Zhong-wei LI Ai-xiu ZHANG Ke-qiang(Agro-environmental Protection Institute, Ministry of Agriculture, Tianjin 300191, China)

机构地区：[1]农业部环境保护科研监测所,天津300191

出　　处：《农业资源与环境学报》2017年第3期234-241,共8页Journal of Agricultural Resources and Environment

基　　金：天津市科技支撑计划(重点)项目(12ZCZDNC09600);中国农业科学院科技创新工程项目;农业科研杰出人才及其创新团队项目

摘　　要：为了探讨基于泌乳牛日粮营养成分及其基本生产状况预测粪便氮磷含量的可行性,建立粪便总氮(fecal nitrogen,FN)、粪便总磷(fecal phosphorus,FP)含量的预测模型。本试验以中国荷斯坦奶牛为研究对象,选取天津市具有代表性的7家规模化奶牛场为采样点,利用问卷调查收集了20组泌乳期奶牛日粮营养成分及基本生产状况等基础数据,并采集了60份粪便样品,测定其氮磷含量。选取其中14组泌乳期奶牛日粮营养成分及基本生产状况等基础数据和48份泌乳牛粪便的氮磷含量,利用SAS统计分析软件,对其进行相关性分析和回归分析,建立预测模型。结果表明:日粮有机质摄入量(organic matter intake,OMI)和粗脂肪摄入量(crude fat intake,CFi)与泌乳牛粪便总氮含量有显著的正相关性,相关系数分别为0.836和0.705。泌乳牛体重(body weight,BW)与粪便总磷含量呈负相关性,相关系数为-0.525。利用多元线性回归分析建立的预测模型的决定系数R2显著高于一元线性回归方程。其中基于泌乳牛产奶量(milk yield,MY),产奶天数(days in milk,DIM),日粮有机质摄入量OMI和氮摄入量(nitrogen intake,NI)建立的粪便总氮含量的预测模型的决定系数R2可达0.96(P<0.001),预测方程为:y=0.43+0.29×MY+0.02×DIM+0.92×OMI-13.01×NI。粪便总磷含量的预测模型的决定系数R2相对低于总氮含量的预测模型,最高为0.62(P<0.10),预测方程为:y=22.97-0.026×BW-4.02×NI+14.63×PI(phosphorus intake,PI)。最后利用6组泌乳牛日粮营养成分及其基本生产状况的基础数据和对应的18份粪便样品的氮磷含量对最优预测模型进行了外部验证。结果表明,粪便总氮含量和总磷含量的预测值与测定值间的相对误差分别为1.62%和3.81%,预测标准差分别为0.70 mg·g-1和0.68 mg·g-1,模型取得较理想的预测结果。To facilitate efficient and sustainable manure management and reduce potential pollution, it＇s necessary for precise prediction of fecal nutrient content. The aim of this study is to build prediction models of fecal nitrogen and phosphorus content by the factors of dietary nutrient composition, days in milk, milk yield and body weight of Chinese Holstein lactating dairy cows. 20 kinds of dietary nutrient composition and 60 feces samples were collected from lactating daily cows from 7 large-scale dairy farms in Tianjin City; The fecal nitrogen and phosphorus content were analyzed. The whole data set was divided into training data set and testing data set. The training data set, including 14 kinds of dietary nutrient composition and 48 feces samples, was used to develop prediction models. The relationship between fecal nitrogen or phosphorus content and dietary nutrient composition was illustrated by means of correlation and regression analysis using SAS software. The results showed that fecal nitrogen （FN） content was highly positively correlated with organic matter intake （OMI） and crude fat intake （CFi）, and correlation coefficients were 0. 836 and 0.705, respectively. Negative correlation coefficient was found between fecal phosphorus （FP） content and body weight （BW）, and the correlation coefficient was -0.525. Among different approaches to develop prediction models, the results indicated that determination coefficients of multiple linear regression equations were higher than those of simple linear regression equations. Specially, fecal nitrogen content was excellently predicted by milk yield （ MY ）, days in milk （ DIM ）, organic matter intake （OMI） and nitrogen intake （NI）, and the model was as follows ： y=0.43＋0.29×MY＋0.02×DIM＋0.92×OMI- 13.01×NI （R2=0.96）. Accordingly, the highest determination coefficient of prediction equation of FP content was 0.62, when body weight （BW）, phosphorus intake（ PI ） and nitrogen intake （NI） were combined as pre

关 键 词：泌乳牛 粪氮 粪磷 日粮营养成分 预测模型 

分 类 号：X713[环境科学与工程—环境工程]