Behavioral responses during and after a postpartum human-animal interaction in rangeland breeding ewes

Abstract

Lamb welfare, performance and survival on rangeland is partially determined by the quality of maternal care received (i.e., high durations of sniffing/grooming, allowing udder access, and staying close to the lamb). Previous studies have explored maternal behavior scores (MBSs), as a tool to select ewes based on their proximity to the lamb and the shepherd, finding variable evidence towards sensitivity and robustness in scores in ability to gauge lamb outcome (birth weight and weaning weight), survival, and maternal care throughout the lambing season. Though this scoring system is convenient, there may be another behavioral attribute related to fear of human or relationship to the human handler may be elicited during lamb collection and processing. This longitudinal study investigated the presence of a trait in response to the human handler employing a Lamb Handling (during processing) and a Lamb Tie Down (after processing) test with N= 42 total ewes and n=26 ewes consistently present across both years of the study. The current study also set out to investigate the relationship between proximity to the shepherd and lamb outcome. Attempts to perform principal component analysis revealed that variables from the Lamb Handling and Lamb Tie down test were not suitable for factor analysis. In the Lamb Handling test, proportion of ‘pacing’ was a nearly significant, negative predictor (P=0.07) term in the binary ‘investigating human’ model, improved model fit and explained more variance compared to the null model [(cumulative weight: 90 vs 10 %); (R²_adjusted: 30.8 vs 15.2 %; deviance= 31 vs 18.5 %)]. A random term for individual also improved fit in this model, indicating a potential behavioral trait driven by the ewe related to fear or perception of the human. The occurrence of grazing (0 or 1) was a nearly significant, positive predictor (P=0.06), improved model fit and explained more variance in the ‘investigating human’ model compared to the null model [(cumulative weight: 76 vs 24 %); (R²_adjusted: 21 vs 15.2 %; deviance= 26.8 vs 18.5 %)]. Individual did not improve model fit in the relationship between ‘investigation human’ and ‘grazing’. Weaning weights were a nearly significant positive predictor in the model for ‘close proximity’ at lamb collection (P=0.07) and a significant positive predictor for ‘close proximity’ during processing (P<0.001), however, did not improve model fit over the null. Birth weights improved model fit for behaviours including ‘sniffing/grooming’ and ‘pawing’ in the Lamb Tie Down test and these behaviours may have been moderated by maternal investment in heavier lambs. Behavioural responses, expanding beyond proximity measures, can be useful for gauging maternal attachment and investment in the lamb during and after processing. Fear or perception of humans should also be considered as contributors to the response to the shepherd during human-animal interactions in the lambing season.