Estimation of the sample size to determine gas flux in finishing steers under confined conditions using alternative statistical approaches

Abstract In ruminants, gas flux evaluation (i.e., methane [CH4] and carbon dioxide [CO2] production and oxygen [O2] consumption) has primarily been conducted in experimental research conditions. Despite its importance for accurate gas flux estimation, to our knowledge, limited guidance exists on how many animals are required to be sampled to represent a specific population of animals adequately. This study aimed to predict the number of growing steers required to characterize gas flux within contemporary groups during the finishing phase using alternative statistical approaches. The analysis included data from six research studies conducted during the finishing phase using Angus cross-bred steers (n = 572 cattle, 552 ± 60 kg of body weight). In each experiment, gas flux was individually determined using an automated head-chamber system. The required number of steers was estimated using three independent approaches: a conventional statistical sampling method, a bootstrap evaluation, and a Monte Carlo simulation. The number of steers required to determine CH4 production varied between 15 and 55 (i.e., between 15 to 125% of the original population), while to determine CO2 production and O2 consumption varied between 5 and 14 (i.e., between 6 to 43% of the original population) according to each experimental condition. Conventional methods could be applied a priori, whereas computational techniques can only be performed a posteriori. These results illustrate how methodological choices influence the estimated sample size for gas flux evaluation in finishing steers.