Abstract. Methane (CH₄) is considered to be the second main contributor to the global greenhouse gas effect, with major CH₄ emissions originating from livestock. Accurate measurements from ruminating herds are required to improve emission coefficients used in national emission inventories, and to evaluate mitigation strategies. Previous measurements of enteric methane emissions from domestic animals have been carried out in artificial conditions such as laboratory chambers, or by fitting individual animals with capillary tubes and using SF₆ as a tracer. Here we evaluated the reliability of eddy covariance technique (EC), already used for CO₂ fluxes, for continuous CH₄ measurements over a grazed field plot. Analyzer accuracy and reliability of eddy covariance technique were tested against field scale measurements with the SF₆ tracer technique, Gaussian plume model and emission factors (i.e. IPCC). Results indicate a better agreement between EC and SF₆ method when grazing heifers were parked close to the EC setup. However, a systematic underestimation of EC data appeared and even more when the distance between the source (ruminating heifers) and EC setup (mast) was increased. A two-dimensional footprint density function allowed to correct for the dilution effect on measured CH₄ and led to a good agreement with results based on the SF₆ technique (on average 231 and 252 g CH₄ ha⁻¹ over the grazing experiment, respectively). Estimations of the CH₄ budgets for the whole grazing season were in line with estimates (i.e. emission factor coefficients) based on feed intake and animal live weight as well as SF₆ technique. IPCC method Tier 2, however, led to an overestimation of CH₄ fluxes on our site.