Mob-based walk-over weights: similar to the average of individual static weights?

Abstract

Merino ewe liveweight represents an objective measure shown to have a profound effect on production outcomes and therefore research into technology that provides accurate and timely feedback of ewe liveweight change is warranted. Most sheep in Australia are not individually identified and therefore management of sheep is on a flock or ‘mob’ basis. Mob-based walk-over weighing (MBWOW) is a remote weighing concept for sheep flocks whereby animal weights are collected as they pass freely over a strategically placed weighing platform. The weights are then collected by the livestock manager, analysed and interpreted to aid nutritional decision making on a whole-flock basis. The hypothesis tested in this study was that data from MBWOW systems is comparable to data from static weighing sessions. At three sites, MBWOW data was collected simultaneously with monthly static weighing sessions. Raw data from MBWOW from each site was manipulated through a series of methodologies that were compared according to their relationship with the static weight data. All forms of MBWOW data showed a significant relationship with static weighing data (P < 0.05). Using a 25% filter (data within 25% of a predetermined central weight kept) and grouping data into 5-day groups strengthens the relationship between MBWOW data and static weighing data. In 1-day groupings, MBWOW data with a 25% filter and subjectively chosen central weight showed the strongest relationship (R2 = 0.89) with static weighing data. In 5-day groupings, MBWOW data with a 25% filter and reference weight from a previous weighing event as a central weight showed the strongest relationship (R2 = 0.88) to static weighing data. The former MBWOW data manipulation methodology had the least mean numerical difference (±s.d.) between MBWOW data and static weighing data (1.86 ± 0.85 kg), yet the latter had the least mean numerical difference in change-in MBWOW data and change-in static weighing data (1.51 ± 0.39 kg), and as change-in liveweight has the most application to industry, it is recommend as the preferred data manipulation technique. These findings suggest that although MBWOW is not fully congruent with static weighing, a strong relationship (R2 > 0.8) between the two and low mean numerical difference in change-in liveweight indicates that MBWOW has potential to be used to established liveweight profiles for Merino ewes that aid nutritional management.