Improving the accuracy of beef cattle methane inventories in Latin America and Caribbean countries

Abstract

On-farm methane (CH₄) emissions need to be estimated accurately so that the mitigation effect of recommended practices can be accounted for. In the present study prediction equations for enteric CH₄ have been developed in lieu of expensive animal measurement approaches. Our objectives were to: (1) compile a dataset from individual beef cattle data for the Latin America and Caribbean (LAC) region; (2) determine main predictors of CH₄ emission variables; (3) develop and cross-validate prediction models according to dietary forage content (DFC); and (4) compare the predictive ability of these newly-developed models with extant equations reported in literature, including those currently used for CH₄ inventories in LAC countries. After outlier's screening, 1100 beef cattle observations from 55 studies were kept in the final dataset (∼ 50 % of the original dataset). Mixed-effects models were fitted with a random effect of study. The whole dataset was split according to DFC into a subset for all-forage (DFC = 100 %), high-forage (94 % ≥ DFC ≥ 54 %), and low-forage (50 % ≥ DFC) diets. Feed intake and average daily gain (ADG) were the main predictors of CH₄ emission (g d^-1), whereas this was feeding level [dry matter intake (DMI) as % of body weight] for CH₄ yield (g kg^-1 DMI). The newly-developed models were more accurate than IPCC Tier 2 equations for all subsets. Simple and multiple regression models including ADG were accurate and a feasible option to predict CH₄ emission when data on feed intake are not available. Methane yield was not well predicted by any extant equation in contrast to the newly-developed models. The present study delivered new models that may be alternatives for the IPCC Tier 2 equations to improve CH₄ prediction for beef cattle in inventories of LAC countries based either on more or less readily available data.

Keywords: Dietary nutrients; Greenhouse gas; Linear regression; Livestock; Methane conversion factor; Model cross-validation.