Title

Prediction of milk fat, protein and yield from the morning or afternoon milking in grazing holstein cows

Document Type

Article

Publication Date

9-1-2017

Publication Title

Livestock Research for Rural Development

Abstract

Partial milk yield and percentage of milk fat and protein were used to determine percentage and total production per day in grazing Holstein cows. A total of 51369 daily records of milk yield and milk fat and protein (AM/PM) from 2616 cows in 21 dairy herds from north and east Antioquia (Colombia) were used. The milk control was performed with the A4 X2 methodology involving monthly visits and milking control during the morning and afternoon. Milking intervals ranged from 9 to 11.5 h for the PM milking and 12.5 to 15 h for the AM milking. Only data between days 5 to 599 of lactation were considered. The Eureqa program was used to assess linear and nonlinear models for estimating the best prediction of milk yield, fat and protein in 24 hours. The r2 statistic, Akaike´s information coefficient (AIC), bias and its standard deviation, and the concordance correlation coefficient (CCC) were estimated to compare the equations. The best equations selected for yield -for having a low AIC were those that included partial milk yield (AM or PM), days in milk (DEL) and the time between milkings. The best models for protein estimation were those that exclusively considered the percentage of protein (AM or PM). The best models for fat content were those considering only fat percentage. Nevertheless, the models considering days in milk (PM), time of milking, and daily yield (AM) had better fit. The models obtained in this study can be used reliably, with good precision, and at low cost to predict milk, fat and protein production during 24-hours periods.

Volume

29

Issue

9

ISSN

01213784

Identifier

SCOPUS_ID:85028777311

Documento original

Compartir

COinS