Uncertainty assessment of the breath methane concentration method to determine methane production of dairy cows

The breath methane concentration method uses the methane concentrations in the cow's breath during feed bin visits as a proxy for the methane production rate. The objective of this study was to assess the uncertainty of a breath methane concentration method in a feeder and its capability to measure and rank cows' methane production. A range of controlled methane fluxes from a so-called artificial reference cow were dosed in a feed bin, and its exhaled air was sampled by a tube inside the feeder and analyzed. The artificial reference cow simulates the lungs, respiratory tract, and rumen of a cow and releases a variable methane flux to generate a concentration pattern in the exhaled breath that closely resembles a real cow's pattern. The strength of the relation between the controlled methane release rates of the artificial reference cow and the measured methane concentrations was analyzed by linear regression, using the coefficient of determination (R²) and the residual standard error as performance indicators. The effect of error sources (source-sampling distance, air turbulence, and cow's head movement) on this relation was experimentally investigated, both under laboratory and barn conditions. From the laboratory to the dairy barn at the 30-cm sampling distance, the R²-value decreased from 0.97 to 0.37 and the residual standard error increased from 75 to 86 ppm as a result of barn air turbulence, the latter increasing to a theoretical 94 ppm if modeled variability due to cow's head movement was accounted for as well. In practice, the effect of these random errors can be compensated by sampling strategies including repeated measurements on each cow over time, thus increasing the distinctive power between cows. However, systematic errors that may disturb the relation between concentration and production rate, such as cow variation in air exhalation rate and air flow patterns around sampling locations that differ between barns, cannot be compensated by repeated measurements. As a result, the methane concentrations of breath air will vary between cows with the same methane production. We conclude that the capability of the breath concentration measurement method to adequately measure and rank methane production rates among cows is highly uncertain and requires further investigation into variation sources with a systematic nature.