Evaluating Equations for Predicting Enteric Methane Emissions in Dairy Cattle.

Abstract

Several prediction equations have been created, based on various dietary composition variables, to predict dairy cattle enteric methane emissions (EMEs). Inconsistencies in measuring EMEs have created difficulties comparing dairy cattle emissions between farms and inhibits certain in efforts to reduce emissions and work towards Net Zero. The aims of the current study were to gather existing EME prediction equations and evaluate the variability in their prediction results. An additional aim was to create a combined prediction equation, based on the dietary components with the highest predictive ability, representing the average prediction across existing equations, which accounted for the variation amongst existing equations. The 32 equations produced large variation in the prediction of EMEs for each of the 15 example diets, ranging from 12.49 to 34.27 g CH₄/kg DM. To create a combined EME prediction equation, twelve combinations of dietary variables were evaluated using a mixed-effects model. An equation based on metabolised energy (ME) and neutral detergent fibre (NDF) was chosen (methane (CH₄) = 0.33 × ME + 0.31 × NDF + 3.47), due to the significance of the predictor variables and low prediction error (RMSE = 1.47 g CH₄/kg DM), with a random-effects residual variance of 2.32. The combined equation may act as a suitable compromise to compare emissions between studies accounting for unexplained variation.