Using animal-mounted sensor technology and machine learning to predict time-to-calving in beef and dairy cows.

Worldwide, there is a trend towards increased herd sizes, and the animal-to-stockman ratio is increasing within the beef and dairy sectors; thus, the time available to monitoring individual animals is reducing. The behaviour of cows is known to change in the hours prior to parturition, for example, less time ruminating and eating and increased activity level and tail-raise events. These behaviours can be monitored non-invasively using animal-mounted sensors. Thus, behavioural traits are ideal variables for the prediction of calving. This study explored the potential of two sensor technologies for their capabilities in predicting when calf expulsion should be expected. Two trials were conducted at separate locations: (i) beef cows (n = 144) and (ii) dairy cows (n = 110). Two sensors were deployed on each cow: (1) Afimilk Silent Herdsman (SHM) collars monitoring time spent ruminating (RUM), eating (EAT) and the relative activity level (ACT) of the cow, and (2) tail-mounted Axivity accelerometers to detect tail-raise events (TAIL). The exact time the calf was expelled from the cow was determined by viewing closed-circuit television camera footage. Machine learning random forest algorithms were developed to predict when calf expulsion should be expected using single-sensor variables and by integrating multiple-sensor data-streams. The performance of the models was tested using the Matthew's correlation coefficient (MCC), the area under the curve, and the sensitivity and specificity of predictions. The TAIL model was slightly better at predicting calving within a 5-h window for beef cows (MCC = 0.31) than for dairy cows (MCC = 0.29). The TAIL + RUM + EAT models were equally as good at predicting calving within a 5-h window for beef and dairy cows (MCC = 0.32 for both models). Combining data-streams from SHM and tail sensors did not substantially improve model performance over tail sensors alone; therefore, hour-by-hour algorithms for the prediction of time of calf expulsion were developed using tail sensor data. Optimal classification occurred at 2 h prior to calving for both beef (MCC = 0.29) and dairy cows (MCC = 0.25). This study showed that tail sensors alone are adequate for the prediction of parturition and that the optimal time for prediction is 2 h before expulsion of the calf.

Use of novel sensors combining local positioning and acceleration to measure feeding behavior differences associated with lameness in dairy cattle.

Time constraints for dairy farmers are an important factor contributing to the under-detection of lameness, resulting in delayed or missed treatment of lame cows within many commercial dairy herds. Hence, a need exists for flexible and affordable cow-based sensor systems capable of monitoring behaviors such as time spent feeding, which may be affected by the onset of lameness. In this study a novel neck-mounted mobile sensor system that combines local positioning and activity (acceleration) was tested and validated on a commercial UK dairy farm. Position and activity data were collected over 5 consecutive days for 19 high-yield dairy cows (10 lame, 9 nonlame) that formed a subset of a larger (120 cow) management group housed in a freestall barn. A decision tree algorithm that included sensor-recorded position and accelerometer data was developed to classify a cow as doing 1 of 3 categories of behavior: (1) feeding, (2) not feeding, and (3) out of pen for milking. For each classified behavior the mean number of bouts, the mean bout duration, and the mean total duration across all bouts was determined on a daily basis, and also separately for the time periods in between milking (morning = 0630-1300 h; afternoon = 1430-2100 h; night = 2230-0500 h). A comparative analysis of the classified cow behaviors was undertaken using a Welch t-test with Benjamini-Hochberg post-hoc correction under the null hypothesis of no differences in the number or duration of behavioral bouts between the 2 test groups of lame and nonlame cows. Analysis showed that mean total daily feeding duration was significantly lower for lame cows compared with non-lame cows. Behavior was also affected by time of day with significantly lower mean total duration of feeding and higher total duration of nonfeeding in the afternoons for lame cows compared with nonlame cows. The results demonstrate how sensors that measure both position and acceleration are capable of detecting differences in feeding behavior that may be associated with lameness. Such behavioral differences could be used in the development of predictive algorithms for the prompt detection of lameness as part of a commercially viable automated behavioral monitoring system.

Associations between locomotion score and kinematic measures in dairy cows with varying hoof lesion types.

During this study we explored the gait attributes commonly used in subjective locomotion scoring systems and use new technology to evaluate these gait attributes objectively on 60 Holstein lactating dairy cattle. Kinematic gait analysis more commonly used in sports and equine science was adapted for use on dairy cattle to assess stride characteristics, joint flexion, and spine posture in dairy cows with different lameness status. Cows that were lame had shorter stride length and had negative tracking distance compared with nonlame cattle. Lame cattle did not show any difference in spine posture when walking. Gait alterations were more evident in cows with sole ulcers, which showed considerable shortening of stride and had more negative tracking compared with cows with no hoof lesions. Cows with sole ulcers also showed significant shortening of the spine when walking than cows with no hoof lesions.

Risk factors for increased rates of sole ulcers, white line disease, and digital dermatitis in dairy cattle from twenty-seven farms in England and Wales.

Claw lesion treatment records were recorded by farmers on 27 dairy farms (3,074 cows, 36,432 records) in England and Wales between February 2003 and February 2004. These were combined with farm environment and management data collected using a combination of direct observations, interviews with farmers, and milk recording data. Multilevel models were constructed for the 3 most frequently reported lesions related to lameness, namely, sole ulcers, white line disease, and digital dermatitis. Risks associated with an increased incidence of sole ulcers were parity 4 or greater, the use of roads or concrete cow tracks between the parlor and grazing, the use of lime on free stalls, and housing in free stalls with sparse bedding for 4 mo or more. The risks for white line disease were increasing parity and increasing herd size, cows at pasture by day and housed at night, and solid grooved concrete floors in yards or alleys. Solid grooved flooring was also associated with an increased risk of digital dermatitis, and cows 6 or more months after calving had a decreased risk of a first case of digital dermatitis. These results improve our understanding of the specific risks for 3 important lesions associated with bovine lameness and could be used as interventions in future clinical studies targeted at the reduction of specific lesions.

Associations between sole ulcer, white line disease and digital dermatitis and the milk yield of 1824 dairy cows on 30 dairy cow farms in England and Wales from February 2003-November 2004.

The milk yields of 1824 cows were used to investigate the effect of lesion-specific causes of lameness, based on farmer treatment and diagnosis of lame cows, on milk yield. A three-level hierarchical model of repeated test day yields within cows within herds was used to investigate the impact of lesion-specific causes of lameness (sole ulcer, white line disease, digital dermatitis and other causes) on milk yield before and after treatment compared with unaffected cows. Cattle which developed sole ulcer (SU) and white line disease (WLD) were higher yielding cattle before they were diagnosed. Their milk production fell to below that of the mean of unaffected cows before diagnosis and remained low after diagnosis. In cattle which developed digital dermatitis (DD) there was no significant difference in milk yield before treatment and a slightly raised milk yield immediately after treatment. The estimated milk loss attributable to SU and WLD was approximately 570 and 370 kg, respectively. These results highlight that specific types of lameness vary by herds and within herds they are associated with higher yielding cattle. Consequently lesion-specific lameness reduction programmes targeting the cow and farm specific causes of lameness might be more effective than generic recommendations. They also highlight the importance of milk loss when estimating the economic impact of SU and WLD on the farms profitability.

Management factors associated with impaired locomotion in dairy cows in England and Wales.

Forty-nine farms in England and Wales were visited on 4 occasions between February 2003 and March 2004. A total of 21,693 scores of locomotion were assigned to 7,722 cattle. Locomotion was assessed on a 3-point scale by observing the posture of a cow's back while standing and walking (1 = sound, 2 = not sound, 3 = lame). Data on measurable factors potentially associated with locomotion were collected from all farms using direct observations of the farm environment and a comprehensive farmer interview. The mean herd locomotion score was 1.77 +/- 0.02. There was no significant difference in mean herd locomotion scores between 5 herds housed in straw yards (1.72 +/- 0.02) and 44 herds housed in free stalls (1.78 +/- 0.02), possibly because of lack of power. A GLM was produced using data from the 44 herds housed in free stalls, with the mean farm locomotion score of all cows examined on all 4 visits as the outcome variable. Factors associated with an elevated locomotion score were dry cows kept in straw yards compared with free stalls (increase in locomotion score = 0.06 +/- 0.03), pregnant heifers kept with milking cows in winter compared with being kept with dry cows (increase in locomotion score = 0.09 +/- 0.03), aisle widths of < 3 m compared with widths of > or = 3 m (increase in locomotion score = 0.06 +/- 0.02), a curb height of < or = 15 cm compared with a height of > 15 cm (increase in locomotion score = 0.07 +/- 0.03), routine trimming of hooves of all cows by a hoof trimmer or by the farmer compared with no routine hoof trimming (increase in locomotion score = 0.18 +/- 0.04 and 0.13 +/- 0.03 respectively), feeding corn silage to milking cows compared with feeding other forage types (increase in locomotion score = 0.10 +/- 0.03), and the use of automatic scrapers in the free-stall barn compared with tractor scrapers (increase in locomotion score = 0.10 +/- 0.03). These variables were correlated with many other management variables. The use of automatic scrapers was correlated with the use of sawdust on rubber mats in free stalls. Curb height was negatively correlated with increasing distance of the neck rail from the front (head end) of the free stall. These putative risk factors support the hypothesis that locomotion score is linked to management factors; in particular, the combination of sawdust on rubber mats with automatic scrapers was associated with elevated locomotion scores.

Influence of rearing conditions and respiratory disease on haptoglobin levels in the pig at slaughter.

Associations between serum concentrations of haptoglobin, pathological lung lesions indicative of Mycoplasma hyopneumoniae (EP) or Actinobacillus pleuropneumoniae (PL) infection at slaughter and previous rearing environment were investigated in 510 pigs (90-100 kg live weight) from 17 farms in England. Haptoglobin concentrations were significantly higher in pigs showing pathological signs of EP infection compared to those without signs of this disease (EP positive median 0.43 mg ml(-1) vs. EP negative median 0.26 mg ml(-1), p<0.01). However, there were no significant associations between serum haptoglobin concentrations and pathological signs of PL. The presence of solid partitions compared with barred or similar open partitions was associated with a decrease of 0.44 mg ml(-1) farm mean haptoglobin concentration, whilst an increase in pen size of 10 m(2) was associated with a decrease of 0.08 mg ml(-1) farm mean haptoglobin concentration. The findings indicate that pathological signs of EP were associated with increased serum haptoglobin at slaughter, which in turn was influenced by components of the farm environment.

Risk factors for reduced locomotion in dairy cattle on nineteen farms in The Netherlands.

A 3-point locomotion scoring system was used that incorporated the position of the back of cows while standing and when walking to investigate risk factors for elevated locomotion scores of 1,450 dairy cows on 19 farms in The Netherlands. Each of the farms was visited twice in an 18-mo period from February 2003 to July 2004. At each visit, all milking and dry cows were scored for locomotion by a single observer. Two multivariable regression models were constructed to identify factors associated with elevated mean locomotion score (increased abnormality) and the percentage of cows with the highest score (score 3). Risk factors for increased locomotion score were having a hoof-trimming stall with foot-lifting apparatus compared with not having such apparatus (increase in locomotion score = 0.15), presence of a footbath at the parlor exit or other site compared with not having a footbath on the farm (increase in locomotion score = 0.17 and 0.19, respectively), not providing supplemental vitamins and minerals to lactating cows compared with supplementing animals (increase in locomotion score = 0.17) and feeding corn silage to heifers compared with not doing so (increase in locomotion score = 0.10). The results provide a framework for hypotheses for future investigations of risk factors for high locomotion scores.

Factors in the housing environment of finisher pigs associated with the development of gastric ulcers.

The prevalence and severity of gastric ulcers were investigated in a sample of 50 slaughter pigs from each of 16 commercial farms in the uk. The mean prevalence of ulcers was 19.1 per cent and the mean (se) severity score on the 16 farms was 2.2 (0.15) on a scale from 0 (normal) to 6 (severe). Multivariate regression analysis showed that the type of lying surface provided for the pigs during the finishing phase was the most important factor associated with the occurrence of gastric ulcers in the farms studied. Finisher pigs housed on slatted floors had a significantly higher mean ulcer severity score (P<0.001) than pigs housed on solid concrete floors, which had a significantly higher score (P<0.01) than pigs housed on straw bedding. Feeding a pelleted finisher ration was also associated with significant increase in ulcer severity (P<0.01) compared with feeding an unpelleted ration, but was of less importance than the type of floor. The detrimental influence of feeding pellets was ameliorated by the provision of straw bedding.

Behavioural and physiological responses of sheep to 16 h transport and a novel environment post-transport.

The effect of a novel lairage environment on the ability of sheep to recover from 16 h of transport was investigated. Sheep were transported from grass paddocks to either novel outside paddocks or inside pens, and housed groups were transported to either familiar or novel inside pens. During transport, sheep from outside paddocks lay down less than those from inside pens. In sheep transported to inside pens, those from outside paddocks spent more time lying and spent less time eating; hay and water intakes during the first 12 h post-transport were lower than those previously kept inside. There was no obvious effect of a novel environment post-transport on blood biochemistry, suggesting that the lower post-transport feed and water intakes in a novel environment did not have a significant effect on the ability of the sheep to recover from the feed and water deprivation associated with transport.