Modelling the bi-peak lactation curves of summer calvers in New Zealand dairy farm systems

Abstract

New Zealand dairy cows traditionally calve in late winter, so their peak lactation demands are met by the increased growth of pasture in spring, and they are dried off in autumn before the decreased growth in winter. Financial premiums are offered to encourage farmers to produce milk in winter by calving cows in summer or autumn (out-of-season calving). This requires farming systems that can feed lactating cows in winter, which may include extra silages and fertilisers. The associated extra costs must be weighed against the value of the winter premium. This paper assesses the ability of Dexcel’s Whole Farm Model (WFM) to simulate out-of-season calving by comparing the model output with the observed data from an out-of-season calving trial run at Dexcel’s No. 2 dairy in Hamilton from 1998 to 2001. Trial data showed the typical lactation curve for July calvers with 1 peak in early lactation, whereas for January calvers the curve was atypical with a lower peak in early lactation and a second peak in late lactation. July and January calvers were described in the animal metabolic submodel (‘Molly’) at the start of the simulations by using observed liveweights at the beginning of the season and estimated peak daily milk yields. The management submodel used best-practice policies and the pasture submodel was driven by actual climate data. The WFM with an unmodified ‘Molly’ (without the photoperiod effect) did not predict the flatter bi-peak lactation curves of the January calvers. Driving forces responsible for the differences in the shape of lactation curves were identified and a sine function reflecting the photoperiod effect on lactation hormones was incorporated into ‘Molly’. This reduced the mean prediction error of milk yields for January calvers from 46 to 19%, and for July calvers, from 19 to 15%. The modified ‘Molly’ also showed potential to predict the atypical lactation curves of October and April calvers. After evaluating the WFM, farm systems can be simulated by altering factors such as stocking rate, fertiliser quantity and timing, and proportion of the herd calving out of season. Model output can be used for cost–benefit analysis of a wide variety of potential systems.