Estimation of sustained peak yield interval of dairy cattle lactation curves using a broken-line regression approach
A broken-line regression model with three straight lines and two breakpoints was used to estimate the sustained peak yield of the lactation curve. A sample of 1548 lactation records of 425 Holstein-Friesian cows was provided by the Research and Application Farm of the Agriculture Faculty of Çukurova University in Adana, Turkey. A total of 13463 test-day milk yields (kg/day), recorded once a month with electronic identification and automatic milking recording systems, between November 1994 and January 2006, were used. The data was classified as first, second, third, fourth, fifth, sixth and seventh parity, and included 4105, 3238, 2495, 1800, 1113, 641 and 71 test-day records, respectively. Times (days) at the beginning and end of sustained peak yield per parity were 73.9 - 160.2, 70.6 - 131.9, 70.8 - 130.7, 71.0 - 130.3, 71.3 - 130.4, 71.2 - 128.2 and 71.1 – 129.8, respectively. The peak yields of lactation curves were 19.1 kg/d, 21.8 kg/d, 23.8 kg/d, 24.2 kg/d, 24.2 kg/d 24.4 kg/d and 21.3 kg/d, and days in milk (DIM) at peak yields of lactation curves per parity were 126, 103, 102, 104, 106, 103 and 104, respectively. Persistency values (days) and total lactation milk yields (kg/lactation) per parity were 86.4, 61.3, 59.9, 59.3, 59.1, 57.0, 58.7 and 4852.2, 5105.7, 5503.5, 5503.7, 5449.3, 5416.3, 4802.4, respectively. Results showed that the first parity had a lactation curve with the lowest milk yield at peak that reached the peak point at the latest time (DIM) after parturition, but the largest interval between the beginning and end of the sustained peak yield among all parities. This means that the cows in the first lactation were more persistent than those in the later lactations.