Proposed methods for estimating loss of saleable milk in a cow-calf contact system with automatic milking

Mikhail Churakov,Hanna K Eriksson,Sigrid Agenäs,Sabine Ferneborg

DOI: https://doi.org/10.3168/jds.2022-23099

Abstract:Cow-calf contact (CCC) systems, although beneficial in many respects, introduce additional challenges to collect reliable data on milk production, which is important to assess individual cow efficiency and dairy farm profitability. Apart from weighing calves before and after each feeding, the amount of saleable milk lost due to calf suckling is practically impossible to measure. Here, we assess 2 indirect methods for estimating loss of saleable milk when housing cows and calves together in a robotic milking unit. In our study, treatment (CCC) cows and calves were kept together full time until the calves were 127 ± 6.6 d old (mean ± SD). Control cows were separated from their calves within 12 h of birth and then kept in the same unit as the treatment cows but with no access to either their own or treatment calves. Milk yield recording of both groups was performed from calving until pasture release at 233 ± 20 d in milk. The first estimation method relied on observed postseparation milk yield data, which were fed into a modified Wilmink regression model to determine the best-fitting lactation curve for the preseparation period. The second method was based on the cows' daily energy intake postseparation, calculated by measuring the daily feed intake and analyzing the energy content of the ration. The calculated energy intake was used to determine the average ratio between energy intake and the observed milk yield the following day for each individual cow, assuming constant rates of mobilization and deposition of body fat. The obtained ratio was then used to calculate the expected daily milk yield based on daily energy intake data during the preseparation period. In this paper, we analyzed data from 17 CCC cows kept together with their calves and 16 control cows; both groups calved from September to October 2020 and were followed up until release to pasture in May 2021. Saleable milk yield was lower in CCC cows than in control cows, both before and after separation. The 2 methods were used on data for control cows and showed milk yield loss using the lactation curve method (average of -3.4 ± 2.8 kg/d) and almost no loss using energy intake data (average of -1.4 ± 2.7 kg/d). Milk yield loss for CCC cows was estimated at average 11.3 ± 4.8 and 7.3 ± 6.6 kg milk/d, respectively. The proposed lactation curve estimation method tends to overestimate milk yield loss, whereas the method based on energy intake is more accurate. However, collecting detailed energy intake data per individual cow requires additional effort and equipment, which is not always feasible on commercial farms. Further research is needed to improve milk loss estimation and to better understand trade-offs in CCC systems.