•Respiration of grazing dairy cows was derived from eddy covariance measurements.•Annual NEE of a pasture was determined including and excluding cow contributions.•Footprint weight of grazing cows was determined with animal GPS sensors. Eddy covariance (EC) is the standard approach for monitoring the CO2 exchange of ecosystems and plays an important role in the assessment of their carbon and greenhouse gas budgets. For pastures the application of EC measurement is challenging due to the uneven spatial and temporal distribution of pasturing animals contributing to the net ecosystem exchange of CO2 (NEE). In the present study, the quantitative contribution of the animals to the measured NEE and its gap filling was investigated for a pasture system with 20 dairy cows during one year. GPS trackers recorded individual animal positions to separate EC data into fluxes with and without influence of animal respiration. Based on animal position data, emissions per animal were calculated using a footprint model. Annual NEE calculated from data including cow respiration was −68gCm−2yr−1 (negative value indicating uptake) whereas annual NEE calculated from data without cow respiration contribution was −248gCm−2yr−1. The difference between the two calculation methods allowed the quantification of the cow respiration. The average emission per animals derived by footprint correction (4.6kgChead−1d−1) was in the upper range of values found in literature for similar dairy cows. The applied standard gap filling and flux partitioning algorithm worked well for annual values, but results for shorter-term (daily to monthly) periods showed considerable uncertainties that can be attributed to the strong variability of cow presence in the flux footprint. However, the rotational grazing system in combination with the flux footprint distribution in the present study led to a reduction of the uncertainty for the annual scale.