Short communication: Milking order consistency of dairy cows in large Australian herds

Lameness prevalence and management practices on Irish pasture-based dairy farms

Background

Lameness is a painful disease, which negatively impacts dairy cow production and welfare. The aim of this observational study was to determine herd lameness prevalence, describe current lameness management practices and identify the presence of established risk factors for lameness on Irish pasture-based dairy farms. Farms were visited once during grazing (99 farms) and again during housing (85 farms). Lameness scoring was carried out at each visit (AHDB 0–3 scale); cows were classified as lame if they scored two or three. Farm management practices and infrastructure characteristics were evaluated via farmer questionnaires and direct measurements of farm infrastructure.

Results

Median herd-level lameness prevalence was 7.9% (interquartile range = 5.6 – 13.0) during grazing and 9.1% (interquartile range = 4.9 – 12.0) during housing; 10.9% of cows were lame at a single visit and 3.5% were lame at both visits (chronically lame or had a repeat episode of lameness). Fifty-seven percent of farmers were not familiar with lameness scoring and only one farm carried out lameness scoring. Only 22% of farmers kept records of lame cows detected, and 15% had a lameness herd health plan. Twenty-eight percent of farmers waited more than 48 h to treat a lame cow, and 21% waited for more than one cow to be identified as lame before treating. Six percent of farmers carried out routine trimming and 31% regularly footbathed (> 12 times per year). Twelve percent put severely lame cows in a closer paddock and 8% stated that they used pain relief to treat severely lame cows. Over 50% of farms had at least one cow track measurement that was classified as rough or very rough, and cow tracks were commonly narrow for the herd size. On 6% of farms, all cubicle beds were bare concrete (no matting or bedding) and on a further 6% of farms, there was a combination of cubicles with and without matting or bedding. On 56% of farms, all pens contained less than 1.1 cubicles per cow and on 28% of farms, a proportion of pens contained less than 1.1 cubicles per cow.

Conclusions

Overall, this study identified infrastructure and management practices which could be improved upon. The comparatively low lameness prevalence demonstrated, compared to fully housed systems, also highlights the benefits of a pasture-based system for animal welfare; however, there remains scope for improvement.

Livestock Informatics Toolkit: A Case Study in Visually Characterizing Complex Behavioral Patterns across Multiple Sensor Platforms, Using Novel Unsupervised Machine Learning and Information Theoretic Approaches

Large and densely sampled sensor datasets can contain a range of complex stochastic structures that are difficult to accommodate in conventional linear models. This can confound attempts to build a more complete picture of an animal's behavior by aggregating information across multiple asynchronous sensor platforms. The Livestock Informatics Toolkit (LIT) has been developed in R to better facilitate knowledge discovery of complex behavioral patterns across Precision Livestock Farming (PLF) data streams using novel unsupervised machine learning and information theoretic approaches. The utility of this analytical pipeline is demonstrated using data from a 6-month feed trial conducted on a closed herd of 185 mix-parity organic dairy cows. Insights into the tradeoffs between behaviors in time budgets acquired from ear tag accelerometer records were improved by augmenting conventional hierarchical clustering techniques with a novel simulation-based approach designed to mimic the complex error structures of sensor data. These simulations were then repurposed to compress the information in this data stream into robust empirically-determined encodings using a novel pruning algorithm. Nonparametric and semiparametric tests using mutual and pointwise information subsequently revealed complex nonlinear associations between encodings of overall time budgets and the order that cows entered the parlor to be milked.

Do Walking Distance and Time Away from the Paddock Influence Daily Behaviour Patterns and Milk Yield of Grazing Dairy Cows?

Simple Summary

Dairy cows managed on pasture may need to walk several kilometres to reach the milking parlour to be milked and, thus, may spend an extended time away from the paddock without access to pasture. While cows have a diverse behavioural repertoire, at a minimum they must graze, ruminate and lie down; therefore, time spent away from the paddock will affect how these behavioural needs are met. We investigated how walking distance to the milking parlour and total time spent away from the paddock affected daily grazing, ruminating and lying behaviours of dairy cattle managed in three groups. The devices automatically monitored cow behaviour and the time spent walking and waiting outside the paddock. We showed that cows spent more time grazing and less time ruminating on days with longer walking distances, which may, in part, be due to greater energy expenditure resultant of walking. When cows spent more time away from the paddock, they had a reduced lying time, and the cows in one of the groups produced less milk. Thus, limiting the walking distance and time spent away from the paddock are two factors that could provide greater opportunities for cows to engage in daily behavioural patterns that meet individual needs.

Abstract

In pasture-based systems, cows may spend several hours away from the paddock and may also walk several kilometres to meet daily milking requirements; this could lead cows to experience time constraints for grazing, ruminating and lying time in the paddock. This study investigated how walking distance and time spent away from the paddock affected daily behavioural patterns (i.e., grazing, ruminating and lying time) and milk yield. Dairy cows were managed in three rotationally grazed groups (n = 29 cows each) on the same farm and milked twice daily. A triaxial ear tag accelerometer on each cow recorded daily duration of grazing and ruminating, and a leg-based accelerometer recorded the daily lying time, for 13 days. GPS collars on four cows per group recorded the daily walking distance and total time away from the paddock for the group. A mixed repeated measures model tested how time off-paddock and walking distance affected the daily behavioural patterns; age, breed, milk yield and maximum ambient temperature were used as covariates with group as the observational unit. A second similar model tested how these factors affected milk yield. Walking distance and time spent away from the paddock were not correlated. When daily walking distance increased (to a maximum of 4 km/d), cows spent more time grazing and less time ruminating, but lying time was not affected. This result may, in part, be related to the greater energy expenditure demands for walking longer distances and milk production. When time away from the paddock increased (to a maximum of 4 h/d), cows spent less time lying, but grazing and ruminating times were not affected. Milk yield was not affected by walking distance, but one of the groups experienced a lower milk yield when time away from the paddock was increased. This result suggests that, for some cows, lying times may be shorter when experiencing a longer time away from the paddock, which may also affect milk yield. Overall, this study indicates that paddock behaviours are associated with walking distance to the milking parlour and time spent away from the paddock. Efforts to reduce walking distance and time spent away from the paddock are likely to provide cows with greater opportunity to engage in daily behaviours in the paddock that meet their needs and maintain their milk yield.

Spatial and temporal variation in proximity networks of commercial dairy cattle in Great Britain

The nature of contacts between hosts can be important in facilitating or impeding the spread of pathogens within a population. Networks constructed from contacts between hosts allow examination of how individual variation might influence the spread of infections. Studying the contact networks of livestock species managed under different conditions can additionally provide insight into their influence on these contact structures. We collected high-resolution proximity and GPS location data from nine groups of domestic cattle (mean group size = 85) in seven dairy herds employing a range of grazing and housing regimes. Networks were constructed from cattle contacts (defined by proximity) aggregated by different temporal windows (2 h, 24 h, and approximately 1 week) and by location within the farm. Networks of contacts aggregated over the whole study were highly saturated but dividing contacts by space and time revealed substantial variation in cattle interactions. Cows showed statistically significant variation in the frequency of their contacts and in the number of cows with which they were in contact. When cows were in buildings, compared to being on pasture, contact durations were longer and cows contacted more other cows. A small number of cows showed evidence of consistent relationships but the majority of cattle did not. In one group where management allowed free access to all farm areas, cows showed asynchronous space use and, while at pasture, contacted fewer other cows and showed substantially greater between-individual variation in contacts than other groups. We highlight the degree to which variations in management (e.g. grazing access, milking routine) substantially alter cattle contact patterns, with potentially major implications for infection transmission and social interactions. In particular, where individual cows have free choice of their environment, the resulting contact networks may have a less-risky structure that could reduce the likelihood of direct transmission of infections.

Cow milking order and its influence on milk production in a pasture-based automatic milking system

Context While several studies have documented the consistency of milking order and its association with milk yield in herds with conventional milking systems, there is limited data available on herds in the automatic milking systems (AMS) where cows move to the dairy voluntarily to be milked. Aims The present study was conducted to examine the consistency of milking order in AMS and its association with milk yield and cow characteristics. Methods The study was performed at The University of Melbourne Dookie Dairy in northern Victoria, Australia. The milking herd had up to 158 Holstein–Friesian cows in a pasture-based AMS with a three-way grazing system. The study utilised the individual-cow milking times, parity number, days in milk and data on daily production (milk yield in kilograms, fat and protein percentages and liveweight) from August 2017 till April 2018. Monthly milking order was determined for each milking session by ranking individual cows on the basis of their recorded time of milking. Key results A consistent milking order was observed with the order of cows at the beginning (percentile rank position 0–33) and end of the milking order (percentile rank position 68–100) being less variable than cows in the middle positions. Milking orders from any two consecutive months were highly correlated (P < 0.01). Energy-corrected milk yield was negatively associated with the milking position (5 of 9 months) and days in milk (8 of 9 months), but positively associated with parity number and liveweight (5 of 9 months). The cow factors such as energy-corrected milk yield, liveweight, parity and days in milk were poor predictors of milking order. This suggests that other factors such as health and social dominance might be of importance. Conclusions This observational study indicated that cows at the beginning of the milking order have a higher milk yield than do cows at the end of the milking order in pasture-based automatic milking systems. Implications Grazing-management strategies that allow cows at the end of the milking order to access fresh pasture are worthy of further investigation.

Mind the Queue: A Case Study in Visualizing Heterogeneous Behavioral Patterns in Livestock Sensor Data Using Unsupervised Machine Learning Techniques

Sensor technologies allow ethologists to continuously monitor the behaviors of large numbers of animals over extended periods of time. This creates new opportunities to study livestock behavior in commercial settings, but also new methodological challenges. Densely sampled behavioral data from large heterogeneous groups can contain a range of complex patterns and stochastic structures that may be difficult to visualize using conventional exploratory data analysis techniques. The goal of this research was to assess the efficacy of unsupervised machine learning tools in recovering complex behavioral patterns from such datasets to better inform subsequent statistical modeling. This methodological case study was carried out using records on milking order, or the sequence in which cows arrange themselves as they enter the milking parlor. Data was collected over a 6-month period from a closed group of 200 mixed-parity Holstein cattle on an organic dairy. Cows at the front and rear of the queue proved more consistent in their entry position than animals at the center of the queue, a systematic pattern of heterogeneity more clearly visualized using entropy estimates, a scale and distribution-free alternative to variance robust to outliers. Dimension reduction techniques were then used to visualize relationships between cows. No evidence of social cohesion was recovered, but Diffusion Map embeddings proved more adept than PCA at revealing the underlying linear geometry of this data. Median parlor entry positions from the pre- and post-pasture subperiods were highly correlated (R = 0.91), suggesting a surprising degree of temporal stationarity. Data Mechanics visualizations, however, revealed heterogeneous non-stationary among subgroups of animals in the center of the group and herd-level temporal outliers. A repeated measures model recovered inconsistent evidence of a relationships between entry position and cow attributes. Mutual conditional entropy tests, a permutation-based approach to assessing bivariate correlations robust to non-independence, confirmed a significant but non-linear association with peak milk yield, but revealed the age effect to be potentially confounded by health status. Finally, queueing records were related back to behaviors recorded via ear tag accelerometers using linear models and mutual conditional entropy tests. Both approaches recovered consistent evidence of differences in home pen behaviors across subsections of the queue.

Thermodynamics of standing and lying behavior in lactating dairy cows in freestall and parlor holding pens during conditions of heat stress

Heat load is a challenge for high-producing dairy cows, with adverse consequences on production, reproduction, and welfare. The objectives of this observational study in 2 commercial dairy herds were to determine the effects of environmental heat stress on standing and lying behavior, to monitor the changes in core body temperature (CBT) of cows during lying and standing bouts, and to compare changes in CBT during standing bouts in freestall pens versus standing in holding pens. High-producing cows were selected for data collection over a period of 6 d of increasing heat stress during a heat wave to which they were not acclimated. A total of 20 cows were fitted with leg accelerometers that recorded lying behavior and with vaginal temperature loggers that recorded CBT. These data were recorded at 30-s intervals. Time in the parlor holding pen was calculated from automated parlor software that recorded milking unit attachment and removal times. Mean daily temperature-humidity index in the pens increased from 68.5 to 79.0 during the 6-d trial, whereas mean daily lying time decreased from 9.5 to 6.2 h/d. The mean number of lying bouts per day remained similar at 11.1 to 12.2, but duration of lying bouts decreased from a high of 49.7 min on the coolest day to 32.8 min on the hottest day. During lying bouts, CBT increased at a mean rate of 0.50°C/h. In contrast, CBT changed at a mean rate of -0.25°C/h when standing in the freestall pens and only -0.09°C/h when standing in the milking-center holding pens. Explanatory models for the CBT at which cows ended either standing or lying bouts were derived from 6 selected lying bouts and 3 selected standing bouts for each cow on each day. The CBT at which a cow ended a lying bout was positively related to CBT and temperature-humidity index at the start of the bout, bout duration, and rate of CBT change during the bout. The CBT at which a cow ended a standing bout was negatively related to bout duration and positively related to start CBT, start temperature-humidity index, and rate of CBT change. Insights into the thermodynamics of standing and lying behavior in dairy cows during periods of heat stress provided by this study may contribute to the development of more effective strategies to mitigate heat load in dairy cattle.

Milk Yield, Milk Composition, and the Nutritive Value of Feed Accessed Varies with Milking Order for Pasture-Based Dairy Cattle

(1) Background: Pasture varies in its chemical composition from the top of the sward to the base and cattle prefer to eat the leaf fraction. In pasture-based dairy systems, cattle predominantly walk back to pasture voluntarily after each milking, with the first cattle arriving to pasture hours before the last. Here we study the impact of pasture composition according to milking order on milk yield and milk composition for dairy cattle offered grazed ryegrass pasture. (2) Methods: In the first experiment, individual cow milk yield data were recorded on six farms over 8 months. The herd was divided into groups of 50 cows for analysis according to milking order. In the second experiment, the impact of milking order on milk composition and pasture composition accessed was determined in addition to milk yield on three farms. (3) Results: After accounting for age and stage of lactation effects, cattle milked first in experiment 1 produced, on average, 4.5 L/cow/day (+18%; range 14 to 29%) more than cattle milked last. In experiment 2, dairy cattle milked first (first 50 cows) in farm 1 had greater milk, protein, and solids non-fat (SNF) yield; and less lactose content than those milked last (last 50 cows). In farm 2, dairy cattle milked first had greater milk yield, SNF yield, lactose yield, and fat yield; but less protein and SNF content than cattle milked last. In farm 3, cattle milked first produced milk with greater fat and protein content than cattle milked last. In line with these differences in milk yield and composition, the composition of pasture across vertical strata differed, particularly for crude protein (CP) and acid detergent fiber (ADF) content. Conclusion: This work highlights the opportunity to increase herd nutrient use efficiency for improved milk production through strategic pasture allowance and supplementation strategies.