Scholarlink

<p>Calving prediction is critical for enhancing reproductive management and animal welfare in dairy buffaloes. This study was conducted at the National Dairy Research Institute (NDRI), Karnal, involving detailed behavioural monitoring of 25 Murrah buffaloes from 7 days before to 5 days after calving using continuous CCTV camera surveillance. Day-wise analyses revealed significant behavioural shifts, with tail raising frequency rising sharply from 10.12 (± 0.64) on day − 1 to 24.48 (± 1.13) on the calving day (day 0). Similarly, head turning towards the abdomen increased nearly threefold, from 14.56 (± 0.74) to 45.08 (± 0.99). Six-hour interval data showed a gradual escalation in tail raising from 0.34 movements 120&#xa0;h pre-calving to 2.57 movements near 6&#xa0;h pre-calving, while head turning increased from 0.54 to 5.77 movements in the same timeframe. High-resolution hourly monitoring in the last 24&#xa0;h before calving highlighted an even more pronounced surge in behaviours. Tail raising increased from 0.67 movements at 24&#xa0;h prior to 8.88 movements at 1&#xa0;h before calving, and head turning increased from 0.60 to 12.04 movements, a 1906% rise. Using these behavioural indicators, two predictive models—logistic regression and random forest—were developed to identify calving risk within a practical 3-day window. The logistic regression model highlighted tail raising and head turning as significant predictors, achieving an area under the ROC curve (AUC) of 0.78, accuracy of 86%, specificity of 99%, and moderate sensitivity of 42%. The random forest model exhibited similar performance with an AUC of 0.75, 83% accuracy, 94% specificity, and slightly higher sensitivity of 47%. These results demonstrate the potential of behavioural analytics combined with modern predictive modeling to support timely, automated calving detection systems, facilitating improved reproductive outcomes and animal care.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Tail-raising and head-turning as predictive behavioural indicators of calving in Murrah buffaloes

  • 期刊论文

  • 发表日期：

    2026年2月20日

查看全文
  • Deepak Chopra,
  • Arun Kumar Misra,
  • T. K. Mohanty,
  • Rubina K. Baithalu,
  • Indu Devi,
  • Chinmayee Sahu,
  • Simran Kamboj,
  • Birender Singh,
  • Pratibha Jareda,
  • Prachi Chandrakar

摘要

Calving prediction is critical for enhancing reproductive management and animal welfare in dairy buffaloes. This study was conducted at the National Dairy Research Institute (NDRI), Karnal, involving detailed behavioural monitoring of 25 Murrah buffaloes from 7 days before to 5 days after calving using continuous CCTV camera surveillance. Day-wise analyses revealed significant behavioural shifts, with tail raising frequency rising sharply from 10.12 (± 0.64) on day − 1 to 24.48 (± 1.13) on the calving day (day 0). Similarly, head turning towards the abdomen increased nearly threefold, from 14.56 (± 0.74) to 45.08 (± 0.99). Six-hour interval data showed a gradual escalation in tail raising from 0.34 movements 120 h pre-calving to 2.57 movements near 6 h pre-calving, while head turning increased from 0.54 to 5.77 movements in the same timeframe. High-resolution hourly monitoring in the last 24 h before calving highlighted an even more pronounced surge in behaviours. Tail raising increased from 0.67 movements at 24 h prior to 8.88 movements at 1 h before calving, and head turning increased from 0.60 to 12.04 movements, a 1906% rise. Using these behavioural indicators, two predictive models—logistic regression and random forest—were developed to identify calving risk within a practical 3-day window. The logistic regression model highlighted tail raising and head turning as significant predictors, achieving an area under the ROC curve (AUC) of 0.78, accuracy of 86%, specificity of 99%, and moderate sensitivity of 42%. The random forest model exhibited similar performance with an AUC of 0.75, 83% accuracy, 94% specificity, and slightly higher sensitivity of 47%. These results demonstrate the potential of behavioural analytics combined with modern predictive modeling to support timely, automated calving detection systems, facilitating improved reproductive outcomes and animal care.