Simple and robust algorithms to estimate liveweight in African smallholder cattle

Abstract

Measurement of liveweight of stock is one of the most important production tools available to farmers – playing a role in nutrition, fertility management, health and marketing. Yet most farmers in sub-Saharan Africa do not have access to scales on which to weigh cattle. Heart girth measurements (and accompanying algorithms) have been used as a convenient and cost-effective alternative to scales, however despite a plethora of studies in the extant literature, the accuracy and sensitivity of such measures are not well described. Using three datasets from phenotypically and geographically diverse cattle populations, we developed and validated new algorithms with similar R2 to extant studies but lower errors of prediction over a full range of observed weights, than simple linear regression, that was valid for measurements in an unassociated animal population in sub-Saharan Africa. Our results further show that heart girth measurements are not sufficiently sensitive to accurately assess seasonal liveweight fluctuations in cattle and thus should not be relied on in situations where high precision is a critical consideration.