Abstract. Intensification of pastoral agriculture is occurring rapidly across New Zealand, including increasing use of irrigation and fertiliser application in some regions. While this enables greater gross primary production (GPP) and livestock grazing intensity, the consequences for the net ecosystem carbon budget (NECB) of the pastures are poorly known. Here, we determined the NECB over one year for an irrigated, fertilised and rotationally grazed dairy pasture and a neighbouring unirrigated, unfertilised, winter-grazed pasture. Primary terms in the NECB calculation were: net ecosystem production (NEP), biomass carbon removed by grazing cows and carbon (C) input from their excreta. Annual NEP was measured using the eddy-covariance method. Carbon removal was estimated with plate-meter measurements calibrated against biomass collections, pre- and post-grazing. Excreta deposition was calculated from animal feed intake. The intensively managed pasture gained C (NECB = 103 ± 42 g C m−2 yr−1) but would have been subject to a non-significant C loss if cattle excreta had not been returned to the pasture. The unirrigated pasture was C-neutral (NECB = −13 ± 23 g C m−2 yr−1). While annual GPP of the former was almost twice that of the latter (2679 vs. 1372 g C m−2 yr−1), ecosystem respiration differed by only 68 % between the two pastures (2271 vs. 1352 g C m−2 yr−1). The ratio of GPP to the total annual water input of the irrigated pasture was 37 % greater than that of the unirrigated pasture, i.e. the former used the water input more efficiently than the latter to produce biomass. The NECB results agree qualitatively with those from many other eddy-covariance studies of grazed grasslands, but they seem to be at odds with long-term carbon-stock studies of other New Zealand pastures.