Identifying areas of animal welfare concern in different production stages in Danish pig herds using the Danish Animal Welfare Index (DAWIN).

Animal welfare is of increasing public interest, and the pig industry in particular is subject to much attention. The aim of this study was to identify and compare areas of animal welfare concern for commercial pigs in four different production stages: (1) gestating sows and gilts; (2) lactating sows; (3) piglets; and (4) weaner-to-finisher pigs. One welfare assessment protocol was developed for each stage, comprising of between 20 and 29 animal welfare measures including resource-, management- and animal-based ones. Twenty-one Danish farms were visited once between January 2015 and February 2016 in a cross-sectional design. Experts (n = 26; advisors, scientists and animal welfare controllers) assessed the severity of the outcome measures. This was combined with the on-farm prevalence of each measure and the outcome was used to calculate areas of concern, defined as measures where the median of all farms fell below the value defined as 'acceptable welfare.' Between five and seven areas of concern were identified for each production stage. With the exception of carpal lesions in piglets, all areas of concern were resource- and management-based and mainly related to housing, with inadequate available space and the floor type in the resting area being overall concerns across all production stages. This means that animal-based measures were largely unaffected by perceived deficits in resource-based measures. Great variation existed for the majority of measures identified as areas of concern, demonstrating that achieving a high welfare score is possible in the Danish system.

Erratum: Identifying areas of animal welfare concern in different production stages in Danish pig herds using the Danish Animal Welfare Index (DAWIN) - CORRIGENDUM.

[This corrects the article DOI: 10.1017/awf.2023.37.].

Automatic monitoring and detection of tail-biting behavior in groups of pigs using video-based deep learning methods.

Automated monitoring of pigs for timely detection of changes in behavior and the onset of tail biting might enable farmers to take immediate management actions, and thus decrease health and welfare issues on-farm. Our goal was to develop computer vision-based methods to detect tail biting in pigs using a convolutional neural network (CNN) to extract spatial information, combined with secondary networks accounting for temporal information. Two secondary frameworks were utilized, being a long short-term memory (LSTM) network applied to sequences of image features (CNN-LSTM), and a CNN applied to image representations of sequences (CNN-CNN). To achieve our goal, this study aimed to answer the following questions: (a) Can the methods detect tail biting from video recordings of entire pens? (b) Can we utilize principal component analyses (PCA) to reduce the dimensionality of the feature vector and only use relevant principal components (PC)? (c) Is there potential to increase performance in optimizing the threshold for class separation of the predicted probabilities of the outcome? (d) What is the performance of the methods with respect to each other? The study utilized one-hour video recordings of 10 pens with pigs prior to weaning, containing a total of 208 tail-biting events of varying lengths. The pre-trained VGG-16 was used to extract spatial features from the data, which were subsequently pre-processed and divided into train/test sets before input to the LSTM/CNN. The performance of the methods regarding data pre-processing and model building was systematically compared using cross-validation. Final models were run with optimal settings and evaluated on an independent test-set. The proposed methods detected tail biting with a major-mean accuracy (MMA) of 71.3 and 64.7% for the CNN-LSTM and the CNN-CNN network, respectively. Applying PCA and using a limited number of PCs significantly increased the performance of both methods, while optimizing the threshold for class separation did result in a consistent but not significant increase of the performance. Both methods can detect tail biting from video data, but the CNN-LSTM was superior in generalizing when evaluated on new data, i.e., data not used for training the models, compared to the CNN-CNN method.

Risk factors associated with tail damage in conventional non-docked pigs throughout the lactation and rearing period.

Tail biting is an abnormal behaviour in pigs, and remains an economic and welfare problem in modern pig production. Reasons for performing tail-biting behaviour are of multifactorial origin, and are often related to the current environment or internal characteristics of pigs. The objective of the present study was to identify early life risk factors connected to tail damage in non-docked pigs in a commercial Danish piggery, and to further compare the effects of cumulative cross-life experience throughout the early rearing. In an observational study, 741 piglets from 51 sows born in six batches were individually marked at birth and followed until nine weeks of age. Litter related variables and individual piglet characteristics were collected during lactation. The pigs' performance parameters were recorded from birth to nine weeks of age. The association between putative risk factors and tail damage assessed at different stages during lactation and rearing was analysed using multinomial mixed regression models. Prior to weaning, the odds of having tail damage were higher for piglets originating from litters with a high birth weight variation (P = 0.012) and for piglets that were heavier at weaning (P < 0.001). Piglets born to an aggressive sow had 2.7-fold increased odds of having tail damage (P = 0.003), while piglets of sows treated after farrowing had a lower odds (P = 0.015). Post-weaning, the most significant risk factor(s) associated with tail damage was the previous tail status of the pigs. Pigs with bite marks/ scratches in previous assessments had an on average 4-fold and pigs with a tail wound 11-fold increased odds of having tail damage during subsequent assessments. Similarly, pigs with a tail wound pre-weaning had 7-times higher odds of having tail damage at the end of rearing (P = 0.033). Pigs in groups with a higher weight variation (P = 0.030) and pigs with a greater weight gain (P < 0.001) had higher odds of having tail damage at the end of rearing. There was an increased chance of having tail damage post-weaning for piglets that were cross-fostered (P = 0.032) or that had a clinical impairment (P = 0.047) during lactation. Females generally had a lower chance of having tail damage compared to castrated males. Early life risk factors were especially associated with tail damage in pigs pre-weaning. However, the results of this study suggest that early life risk factors are secondary to the previous tail status of pigs as risk factors for later tail damage.

Should the Contribution of One Additional Lame Cow Depend on How Many Other Cows on the Farm Are Lame?

Welfare Quality® proposes a system for aggregation according to which the total welfare score for a group of animals is a non-linear effect of the prevalence of welfare scores across the individuals within the group. Three assumptions serve to justify this: (1) experts do not follow a linear reasoning when they assess a welfare problem; (2) it serves to prevent compensation (severe welfare problems hidden by scoring well on other aspects of welfare); (3) experts agree on the weight of different welfare measures. We use two sources of data to examine these assumptions: animal welfare data from 44 Danish dairy farms with Danish Holstein Friesian cows, and data from a questionnaire study with a convenience sample of 307 experts in animal welfare, of which we received responses from over 50%. Our main results were: (1) the total group-level welfare score as assigned by experts is a non-linear function of the individual animal welfare states within the group; (2) the WQ system does not prevent what experts perceive as unacceptable compensation; (3) the level of agreement among experts appears to vary across measures. Our findings give rise to concerns about the proposed aggregation system offered by WQ.