Designing Statistical Models for Holstein Rearing Heifers' Weight Estimation from Birth to 15 Months Old Using Body Measurements.

Abstract

Simple SummaryThe growth monitoring process represents an important part of rearing heifers. The use of a scale is not feasible in some breeding conditions; it may be interesting to investigate the possibility of evaluating body weight (BW) with body measurements. The aim of this study was to estimate heifers’ weight based on their body dimension characteristics. A total of 25 Holstein rearing heifers were monitored after birth, weekly until 2 months of life and monthly until 15 months of age. Animals were weighed, and their wither height (WH), shin circumference (SC), heart girth circumference (HG), body length (BL), hip width (HW) and body condition score (BCS) were measured using tape measure. Equations were built with a stepwise regression to estimate the BW at each time using body measures for the study group. Equations were able to estimate the BW of heifers under a 0.800 kg as an average weight gain target using different variables, representing an alternative method of BW evaluation without a scale. Three variables or fewer were needed for BW estimation at crucial growing times, making these models feasible for use in the field. Different growing rate target may be studied in order to evaluate possible modifications to our equations.Body measurements could be used to estimate body weight (BW) with no need for a scale. The aim was to estimate heifers weight based on their body dimension characteristics. Twenty-five Holstein heifers represent the study group (SG); another 13 animals were evaluated as a validation group (VG). All the heifers were weighed (BW) and their wither height (WH), shin circumference (SC), heart girth circumference (HG), body length (BL), hip width (HW) and body condition score (BCS) were measured immediately after birth, and then weekly until 2 months and monthly until 15 months old. Equations were built with a stepwise regression in order to estimate the BW at each time using body measures for the SG. A linear regression was applied to evaluate the relationship between the estimated BW and the real BW. Equations found were to be statistically significant (r2 = 0.688 to 0.894; p < 0.0001). Three variables or fewer were needed for BW estimation a total of 11/23 times. Regression analysis indicated that the use of HG was promising in all the equations built for BW estimation. These models were feasible in the field; further studies will evaluate possible modifications to our equations based on different growing rate targets.