Modelling of heat stress in a robotic dairy farm. Part 3: rumination and milking performance

Robotic milking systems (RMS) have been demonstrated to reduce on-farm labour requirements and collect significant individual-level data in relation to animal health, welfare and production, but they are still largely underutilised. Studies on the relationship between heat stress, animal behaviour and robotic milking performance in a RMS are still insufficient. To model such a relationship, this study focused on analysing the data collected from a RMS system. Animal response indicators of heat stress assessment were rumination time (RT), milk temperature (MT) and daily milk yield (DMY). In addition, RMS milking behaviour, i.e. time of milking (TM), milking frequency (MF), milking duration (MD), milking speed (MS) and milk yield per milking (MY) were also monitored. A new index of rumination efficiency (REI) was created to evaluate rumination efficiency under heat stress, defined as the ratio between DMY and RT. Using multiple broken-line regression, it was found that a 1 °C rise in daily mean temperature could reduce RT by 5.12 min, decrease REI by 0.07 kg·cow−1·h−1, and increase low efficiency milking by 1%. Moreover, the study also found cows prefer to milk between 7:00–9:00AM, and 86% of milking events happened during this period. No significant correlation was found between heat stress and milking behaviour. However, delaying the first milking event of the day and controlling milking intervals to