Modeling of lactation curves of Red Sindhi dairy cows

Abstract

The Lactation curve is the relationship between milk yield and lactation length with respect to time. The present study is based on the data on the lactation records of Red Sindhi cows compiled from Central Cattle Breeding Farm (CCBF), Chiplima, Odisha. Five different lactation curve model, viz. Sikka (SIK), Wood (WD), Polynomial (POL), Dijkstra (DIJK), Morant and Gnanasakthy (MG) models were fitted to the weekly test day milk yield (WTDMY) of Red Sindhi dairy cows. The goodness of fit of the models were judged by adjusted coefficient of determination (Adj. R2), root mean square error (RMSE), Akaike’s information criteria (AIC) and Schwartz Bayesian information criterion (SBC). The normality of the residuals was tested by Shapiro –Wilk’s test and autocorrelation was tested by Durbin Watson test. From the analysis of data it is evident that all the models fitted very well the lactation data of Red Sindhi cows with high degree of accuracy (Adj.R2 > 91%) except for the Wood (WD) model. However among all the models, Polynomial model (Dave 1971) was the best fitted model followed by Dijkstra model. Wood model was the least fitted model among all the lactation models. Highest observed peak yield was at sixth week, the same peak period was given by Wood, Polynomial and Dijkstra models. All the estimated peak yields were ranged between 70.14 kg to 75.96 kg.