This study reports the annual carbon balance of a drained riparian fen under two-cut or three-cut managements of festulolium and tall fescue. CO2 fluxes measured with closed chambers were partitioned into gross primary production (GPP) and ecosystem respiration (ER) for modelling according to environmental factors (light and temperature) and canopy reflectance (ratio vegetation index, RVI). Methodological assessments were made of (i) GPP models with or without temperature functions (Ft) to adjust GPP constraints imposed by low temperature (< 10 degrees C) and (ii) ER models with RVI or GPP parameters as biomass proxies. The sensitivity of the models was also tested on partial datasets including only alternate measurement campaigns and on datasets only from the crop growing period. Use of Ft in GPP models effectively corrected GPP overestimation in cold periods, and this approach was used throughout. Annual fluxes obtained with ER models including RVI or GPP parameters were similar, and also annual GPP and ER fluxes obtained with full and partial datasets were similar. Annual CO2 fluxes and biomass yield were not significantly different in the crop/management combinations although the individual collars (n = 12) showed some variations in GPP (-1818 to -2409g CO2-C m(-2)), ER (1071 to 1738g CO2-C m(-2)), net ecosystem exchange (NEE, -669 to -949g CO2-C m(-2)) and biomass yield (556 to 1044g CO2-C m(-2)). Net ecosystem carbon balance (NECB), as the sum of NEE and biomass carbon export, was only slightly negative to positive in all crop/management combinations. NECBs, interpreted as emission factors, tended to favour the least biomass producing systems as the best management options in relation to climate saving carbon balances. Yet, considering the down-stream advantages of biomass for fossil fuel replacement, yield-scaled carbon fluxes are suggested to be given additional considerations for comparison of management options in terms of atmospheric impact.