Evaluation of a neck mounted 2-hourly activity meter system for detecting cows about to ovulate in two paddock-based Australian dairy herds

Two studies were conducted to assess the performance of a commercially available neck-mounted activity meter to detect cows about to ovulate in two paddock-based Holstein-Friesian dairy herds. The activity monitoring system recorded cow activity count in 2-hourly periods. Study I investigated the ability of the system to detect cow ovulatory periods in dairy herds managed in two different Australian environments and breeding systems using five activity alert algorithms. Herd 1 consisted of approximately 130 milking cows calving year-round in a sub-tropical environment and kept in a single dry lot paddock. Herd 2 consisted of approximately 400 milking cows calving seasonally in a temperate climate and fed pasture by rotation through multiple grazing paddocks. Ovulatory periods and non-ovulatory days were identified using milk progesterone monitoring alone or in combination with ovarian ultrasonography; using these 'gold standards' 141 and 135 ovulatory periods were identified in 64 and 135 cows in Herds 1 and 2 respectively. Sensitivity of the activity monitoring system for detecting cow ovulatory periods ranged from 79.4% to 94.1%, specificity from 90.0% to 98.2% and positive predictive value from 35.8% to 75.8%. Study II investigated the ability of the activity meter system to predict the timing of ovulations in paddock-based pasture-fed dairy cattle (Herd 2). The time of ovulation was estimated by repeat trans-rectal ovarian ultrasonography at approximately 0, 12, 24 and 36 h after artificial insemination (AI). The mean times (±SD) from onset and end of increased activity to ovulation were 33.4 ± 12.4 and 17.3 ± 12.8 h respectively (n = 94). Fifty per cent of cows (n = 47) ovulated within the 8-h period between 30 to 38 hs after the onset of increased activity, 76.6% (n = 72) within the 16 h between 24 to 40 h, 85.1% (n = 80) within the 24 h between 18 and 42 h and 90.4% (n = 85) within the 32 h from 19 to 51 h after the onset of increased activity. Results from these studies show that in paddock-based dairy cows in two diverse management systems, this neck-mounted activity meter system detects high proportions of cows that are about to ovulate and provides a useful indication of when ovulation is likely to occur. However, the specificities and positive predictive values using the algorithms assessed may be lower than desirable. © 2009 Blackwell Verlag GmbH