The value of carcass weight, fat depth measures and eye muscle area for predicting the percentage yield of saleable meat in Australian grass-fed beef carcasses for Japan

Abstract

The percentage saleable meat yield was determined for 42 carcasses from grass-fed steers representing a range of breed types purchased in Australia for the Japanese market. Their mean (s.d.) carcass weight and P8 fat depth were 329 (28.1) kg and 17.3 (4.3) mm, respectively. All measures of subcutaneous fat depth were significantly ( P < 0.05) and moderately correlated with percentage saleable meat yield, with measures in the loin region showing a stronger association than those taken in the rump region. The association between P8 fat depth and the loin measures of subcutaneous fat were low and there was a significant ( P < 0.05) association only between P8 and fat depth at the 10th rib ( FD 10). The use of fat depth measurements from the loin region reduced the error associated with predicting saleable meat yield more than those from the rump region and significantly increased the amount of variation ( R 2) in saleable meat yield that was explained. Hot carcass weight (HCW) did not significantly ( P > 0.05) improve prediction models when combined with subcutaneous fat depth measurements and overall, the R 2 values were low ranging from 0.19 to 0.42. The models indicated that fat depth measures and carcass weight are poor predictors of percentage meat yield in heavy-weight carcasses from mixed breed cattle as produced and processed in Australia. The prediction of percentage yield was in general significantly ( P < 0.05) improved when measures of M. longissimus thoracis et lumborum (LD) area were added as independent variables to models based on hot carcass weight and subcutaneous fat depth measurements. With LD area added the amount of variation in yield that could be explained by the models increased by as much as 28%. Overall, the best model was based on fat depth at the 12th rib and LD area measured at the 5th rib for which the R 2 was 0.58 and the residual standard deviation was 1.63%. The next most accurate prediction of yield was provided by a model which included the independent variables used in the Australian Chiller Assessment Scheme namely HCW, FD10 and LD area at the 10th rib upon which breed type had no significant ( P > 0.05) effect.