Landscape-scale measurements of greenhouse gas exchange from whole farms can bracket the magnitude of fluxes, identify sources and sinks, and provide data for model development and validation. Greenhouse gas fluxes were measured for an annual cycle on a 123-ha beef-cattle farm in southwestern Manitoba, Canada. Carbon dioxide and methane fluxes were measured from a central eddy covariance flux tower. These measurements were supplemented by a secondary flux tower, and by static chambers that measured carbon dioxide, methane and nitrous oxide fluxes. The farm had 104 backgrounding steers and an additional 160 cow-calf pairs occupied the farm for part of the period. Cattle occupied a winter bale-grazing field, a confined feeding paddock, a summer pasture, and a cereal-crop swath-grazing field in sequence. Cattle within the flux footprint exhibited large respiration and methane fluxes, and the cattle respiration was separated from the land CO2 exchange. The largest fluxes of nitrous oxide from cattle excreta (urine, manure) were emitted from the confined feeding paddock and from the location of the winter-feed bales, but only the confined feeding paddock had a net emission of methane from excreta. Despite challenges with cattle movements and scaling in space and time, a greenhouse gas budget was estimated: over the annual cycle, cattle respiration dominated the budget with an emission of 20t CO2 equivalent ha−1y−1; CO2 from the plant/soil system was a net emission of 10t CO2 equivalent ha−1 y−1; enteric methane was 11t CO2 equivalent ha−1y−1; methane from soil/excreta was 0.06t CO2 equivalent ha−1y−1; and nitrous oxide from soil/excreta/fertilizer was 4t CO2 equivalent ha−1y−1. The farm was a net greenhouse gas source of 46t CO2 equivalent ha−1 y−1 and the plant/soil system was a contributing source in this year, partly because of respiration of imported feed.