Context. Knowledge of genetic and phenotypic variation and the accuracy of different measurement techniques is needed to successfully reduce livestock methane (CH4) emissions.
Aims. To estimate repeatabilities, heritabilities and genetic correlations of respiration-chamber (RC) and portable accumulation-chamber (PAC) measurements using two different protocols but the same management and feeding conditions.
Methods. Australian Information Nucleus Flock ewes were measured in seven test-batches. The 510 ewes were removed from pasture and habituated to chaffed alfalfa and cereal hay at 1.5-1.6 times maintenance. Methane was measured in RC for two 22-h periods approximately 14 days apart, and 40 min in PAC, either immediately after removal from individual pens (with feed as described above, PAC0), or 1-h after withdrawing feed (PAC1). There were up to 48 PAC0 tests per day (at 0930 hours, 1100 hours, 1230 hours, 1400 hours in 12 PAC) and 24 PAC1 tests per day (at 1100 hours and 1300 hours). Test methods (RC, PAC0, PAC1) were analysed as different traits in a multi-trait repeated-measures model.
Key results. Before adjustment for liveweight (Lwt) or feed intake (FI), CH4 was highly repeatable (RC 78%, PAC0 83%, PAC1 82%), with heritabilities of 39-55%, permanent environmental (PE) animal variances 23-43% of phenotypic variances (Vp), high genetic correlations between methods (98-100%), and lower PE correlations (44-58%). A second PAC test on the same day decreased CH4 by 8-12% compared with the ewe's first test that day. Heritabilities of FI from 0800 hours until the test was complete (FIOD) were 16-17% (PAC) and 25% (RC) before adjusting for Lwt, with high PE variances (PAC 67-73%, RC 41% of Vp). FI in the previous 24 h was highly heritable and much less variable than was FIOD in the RC, suggesting that testing introduced additional variation by disrupting feeding patterns. After adjusting CH4 for Lwt, FIOD and FI in the previous 24 h and Lwt, some additive genetic variation remained, averaging 17% of Vp. Multivariate models of CH4 and FI, fitting a single animal term (representing genetic+PE variation) showed high animal correlations between FI and CH4, namely, 90-95% before, and 86-95% after adjusting for Lwt.
Conclusions. PAC measurements are heritable and highly correlated with RC measurements under similar management conditions. The high genetic and animal correlations of PAC CH4 and FI imply that CH4 is a useful proxy for FI of grazing animals.