Evaluation of Columbia, USMARC Composite, Suffolk, and Texel rams as terminal sires in an extensive rangeland production system: VII. Accuracy of ultrasound predictors and their association with carcass weight, yield, and value1,2

Use of lamb BW or chilled carcass weights (CCW), live-animal ultrasound or direct carcass measurements of backfat thickness (BF; mm) and LM area (LMA; cm2), and carcass body wall thickness (BWall; mm) to predict carcass yield and value was evaluated using 512 crossbred lambs produced over 3 yr by mating Columbia, U.S. Meat Animal Research Center Composite, Suffolk, and Texel rams to adult Rambouillet ewes. Lambs were harvested at 3 BW endpoints within each year. The predictive value of 3 to 5 additional linear measurements of live-animal or carcass size and shape was also evaluated. Residual correlations (adjusted for effects of year, breed, and harvest group) between ultrasound and direct measurements were 0.69 for BF and 0.65 for LMA. Increasing ultrasound or carcass LMA had positive effects (P < 0.001) on yield of chilled carcass (i.e., on dressing percentage) and, at comparable CCW, on weight of high-value cuts (rack, loin, leg, and sirloin) before trimming (HVW), weight of trimmed high-value cuts (trimmed rack and loin and trimmed boneless leg and sirloin; TrHVW), and carcass value before (CVal) and after (TrCVal) trimming of high-value cuts. By contrast, ultrasound and direct measures of BF had positive effects on yields of CCW and on HVW and CVal but large negative effects on TrHVW and TrCVal. After adjusting for BW at scanning, increases of 1 mm in ultrasound BF or 1 cm2 in ultrasound LMA were associated with changes of US$–0.32 (P < 0.10) and $1.62 (P < 0.001), respectively, in TrCVal. Carcass BWall was generally superior to carcass BF as a predictor of TrHVW and TrCVal. Carcass LMA was superior to ultrasound LMA but carcass BF was inferior to ultrasound BF for prediction of carcass yield and value. Increasing LMA thus would be expected to improve carcass yield and value. Addition of linear measurements of live-animal or carcass size and shape to the prediction model reduced residual SD (RSD) for TrHVW and TrCVal by 0.4 to 2.2%, but subsequent removal of ultrasound or direct measures of BF and LMA from the prediction model increased RSD by 7.4 to 12.2%. Measurements of CCW, LMA, BF, and BWall would thus be appropriate to support programs for value-based marketing of lamb carcasses and are superior to systems based only on measurements of size and shape in unribbed carcasses.