As robotic milking systems become increasingly popular on dairy farms and producers seek to improve animal welfare and efficiency, research being conducted at the University of Guelph in Ontario may provide real-world applications for the industry that contributes to healthier and more efficient dairy herds.
Meagan King, a doctoral candidate in the University of Guelph department of animal biosciences, is working with adviser Trevor DeVries to harness data collected by robotic milking systems and use those data to develop an early indicator for illness in the herd.
“This has obvious health and welfare implications for the cows,” she said. Ultimately, if illness can be detected earlier, it can also be treated earlier and restore cows to full health faster.
Working with 10 farms in Ontario, King will collect milk yield and rumination data from the robotic milking systems for approximately six months. These data will be supplemented by her own tests for lameness, subclinical ketosis and endometritis and farmer and veterinary records for other illnesses. Combining the production, behavior and health data, she will work to identify combinations of parameters that can provide earlier indicators of disease.
“These data could lead to the development of an illness-predicting model, which can then be used to create an alert sent out by the robotic milking system,” King explained. “This can help producers maintain healthy herds and improve cows’ welfare.”
This research builds upon two previously completed studies in which King collected data from robotic milking systems as well as farmer and veterinarian records to assess animal health.
From the first study, which included 41 farms across Ontario and Alberta, she concluded that cows identified as being clinically or severely lame had lower milk yield and visited the robotic milking system less often compared to the healthy cows in the herd. In addition, over-conditioned (overweight) cows produced less milk.
Results of the second study, conducted at the University of Guelph's Kemptville Campus, support the idea that data from the robotic milking systems could be used as early indicators of illness. For example, King found that rumination behavior and milk yield started to decrease significantly four to eight days before the identification of pneumonia, twisted stomach and subclinical ketosis in the cows.
This provided the foundation for a further understanding of the relationships between health and production, which King will continue to explore in her upcoming third study.