Optimizing lactation length for dairy cows

Dairy cow health is at risk around calving. By extending the lactation period a dairy cow will experience a lower frequency of calving events thus, reducing the frequency of associated health risks. However, not every cow seems to be suited for an extended lactation, as cows risk fattening, and at longer lactations dairy production can drop below economically sustainable levels. Different cow characteristics were used to predict if an individual is suited for an extended lactation. Some of these characteristics included: parity, milk yield, body condition, and breeding value for persistency. In this project, I will refine the existing decision support models by incorporating characteristics rom milk robots, cow activity, milk content, and cow body condition score, while accounting for farm-specific effects. Initially, various regression models, including nonlinear regression, will be used to predict the milk yield in an extended lactation. These models will serve as a baseline and will be interpreted to understand the most important characteristics of cow performance in extended lactation. Building upon the knowledge gained from these models I will validate several timeseries classification models to better understand the underlying common patterns in this complex problem. Finally, I will adjust the best performing model such that it can be used as a decision support model in helping farmers to select cows suitable for an extended lactation. The model will be validated with the help of the farmers of the ~40 dairy farms in the OptiLac network. This PhD project is part of the PPP OptiLac, which is financed by Lely, ZuivelNL, Melkveefonds and the Ministry of Agriculture, Fisheries, Food Security and Nature, and in collaboration with ForFarmers and CRV.