Attitudes of western Canadian dairy farmers toward technology.

Dairy farms have become more reliant on technology. The overall aim of this study was to better understand how dairy farmers view technology and its effects on animal care, including their views on the prospect of integrating gene-editing technology in the future. Virtual-semistructured interviews were conducted with dairy farmers (n = 11) from British Columbia and Alberta. To facilitate discussion, the participants were asked to develop and discuss a timeline describing when and why various technologies were adopted on their farm. Although farmers defined technology broadly and affecting multiple aspects of farm management, this paper focuses on their views regarding how technology can affect animal care. Following thematic analysis of the data, the following 3 themes emerged: (1) the changing role of the farmer (including intergenerational considerations and learning new technology), (2) the effect of technology on the cow and her relationship with the farmer and, (3) technology as the future of the farm. The discussions also highlight the concerns that some farmers have regarding challenges associated with reduced human-animal interactions and effective use of the large amounts of data that are collected through technology. We also specifically asked the participants their views about gene editing as a potential future technology. Most of the participants did not specifically address their views on gene editing, but they spoke about the effect on genetic technologies more generally, often making references to genomic testing. However, some questioned how this technology may affect farmers more generally and spoke about how it could affect human-animal relationships. These results illustrate differences among farmers in the way they view technology and how this can affect the dairy cattle they care for.

Cow- and farm-level risk factors for lameness on dairy farms with automated milking systems.

Lameness is a major concern to animal health and welfare within the dairy industry. Our objectives were to describe the prevalence of lameness in high-producing cows on farms with automated milking systems (AMS) and to identify the main risk factors for lameness at the animal and farm level. We visited 36 AMS farms across Canada and Michigan. Farm-level factors related to stall design, bedding use, flooring, and stocking rates were recorded by trained observers. Cows were scored for lameness, leg injuries, body condition (BCS), and body size (hip width and rump height; n=1,378; 25-40 cows/farm). Mean herd prevalence of clinical lameness was 15% (range=2.5-46%). Stall width relative to cow size and parity was found to be the most important factor associated with lameness. Not fitting the average stall width increased the odds of being lame 3.7 times in primiparous cows. A narrow feed alley [<430cm; odds ratio (OR)=1.9], obstructed lunge space (OR=1.7), a low BCS (OR=2.1 for BCS ≤2.25 compared with BCS 2.75-3.0), and presence of hock lesions (OR=1.6) were also identified as important risk factors for lameness. Only 1 of 36 farms had stalls of adequate width and length for the cows on their farm. For lameness prevention, it can be concluded that more emphasis needs be placed on either building stalls of appropriate width or selecting for smaller-framed cows that fit the existing stalls.

Effects of Early Lactation Milking Frequency in an Automated Milking System on Cow Performance.

Automated milking systems (AMS) are increasingly adopted for dairy cow production, promoting individualized cow management dependent on factors like lactation stage, age, and productivity. The study objective was to investigate the effects of early lactation milking frequency on cows milked via AMS. Multiparous Holstein cows blocked by parity and due date were randomly assigned to treatments (n = 8 per treatment): three (3X) or six (6X) milkings per day (MPD). The experimental phase (EXP) was defined as 4 to 29 days in milk (DIM). The AMS settings were programed so 3X cows were limited to three MPD while 6X cows were allowed six MPD. Afterwards was the carry over phase (CO) ranging from 30 to 90 DIM; all cows were allowed up to six MPD. Measurements by the AMS included bodyweight, milk yield (MY), and pellet intake. Weekly composite milk samples were analyzed for macronutrient composition and fatty acid (FA) profile. Coccygeal blood was sampled at 3, 8 ± 1, and 13 ± 1 DIM; concentrations of blood plasma analytes were quantified. Greater MPD was achieved for 6X cows versus 3X cows during EXP, but similar during the CO. Daily MY was non-separable during the EXP while 6X cows in their third or greater lactation group (3 + LG) had greater MY than 3X cows of the same LG during the CO. Milk fat content and 4% fat-corrected MY were both greater for 6X, 3 + LG cows during the EXP compared to 3X, 3 + LG cows. Milk FA methyl esters (FAME) proportions were different between MPD groups, with 6X, 3 + LG cows having the lowest short, even-chain FA from de novo or post-absorptive origin. Differences in analytes indicated that 6X, 3 + LG cows experienced metabolic stress and incorporated greater FA from adipose tissue. Greater early lactation MPD in AMS may shift cow nutrient partitioning to support greater production in 3+ parity cows.

In-Line Detection of Clinical Mastitis by Identifying Clots in Milk Using Images and a Neural Network Approach.

Automated milking systems (AMSs) already incorporate a variety of milk monitoring and sensing equipment, but the sensitivity, specificity, and positive predictive value of clinical mastitis (CM) detection remain low. A typical symptom of CM is the presence of clots in the milk during fore-stripping. The objective of this study was the development and evaluation of a deep learning model with image recognition capabilities, specifically a convolutional neural network (NN), capable of detecting such clots on pictures of the milk filter socks of the milking system, after the phase in which the first streams of milk have been discarded. In total, 696 pictures were taken with clots and 586 pictures without. These were randomly divided into 60/20/20 training, validation, and testing datasets, respectively, for the training and validation of the NN. A convolutional NN with residual connections was trained, and the hyperparameters were optimized based on the validation dataset using a genetic algorithm. The integrated gradients were calculated to explain the interpretation of the NN. The accuracy of the NN on the testing dataset was 100%. The integrated gradients showed that the NN identified the clots. Further field validation through integration into AMS is necessary, but the proposed deep learning method is very promising for the inline detection of CM on AMS farms.