Evaluation of ultrasound scanning to predict carcass composition of Austrian meat sheep

Aims of the study were to evaluate the routine ultrasound scanning plus subjective muscle scoring system for meat sheep in Austria in terms of (1) their ability to predict carcass quality and composition, (2) the repeatability of ultrasound scanning and (3) a comparison of three anatomical scanning sites. Lambs of six breeds (n=189; mean bodyweight 39kg) were scored for the muscling of their shoulder, back and hindquarter and were scanned with an ultrasound device for back fat and longissimus dorsi muscle depth lateral of the spine at 10th/11th (US1) and 13th (US2) thoracic vertebrae as well as at 3rd/4th (US3) lumbar vertebrae. Each ultrasound picture was taken twice within a few minutes to check on within operator repeatability. After slaughter the carcasses were classified according to the EUROP system and back fat and muscle depth were measured on a carcass cross section, and 36 carcasses were dissected to lean meat, fat and bone to evaluate carcass composition. Relationships between carcass and dissection traits and routine performance testing traits (live weight, fat and muscle depth at US3) were evaluated based on partial regression coefficients additionally considering breed, sex (carcass traits only) and birth type as fixed effects. Further, fat and muscle depth at scan sites US1 and US2 were fitted alternatively and Pearson's correlation coefficients were calculated. Correlations between ultrasonic and carcass measures ranged from r=0.60 (muscle depth at US1 and EUROP conformation class) to r=0.84 (muscle depth at US1 and muscle depth at carcass). Repeatabilities for muscle and backfat thickness ranged from 0.90 to 0.95. The results support the usefulness of the currently routine ultrasound scans as relatively easy method to predict carcass composition in live lambs of different breeds. Muscle scans are valuable to estimate amount of carcass lean and EUROP conformation class, but fat scans have greater power to predict the fattiness of the carcass as well as lean percentage. Subjective muscle scoring of live animals seems to be mainly influenced by the fattiness of the animal. The comparison of three anatomical scanning sites did not give definite results. US1 seems to be favourable for estimating muscle depth, for the prediction of lean and in terms of repeatability whereas US2 and US3 had small advantages in scanning fat depth and in the prediction of EUROP classification and carcass fat.