Soil greenhouse gas emissions from cattle grazed and un-grazed temperate upper salt marsh were measured using dark static chambers, monthly for one year. Below-ground gas sampling tubes were also used to measure soil methane (CH 4) concentrations. CH 4 efflux from grazed and un-grazed salt marsh did not differ significantly although grazing did lead to 'hotspots' of underground CH 4 (up to 6% of total air volume) and CH 4 efflux (peak of 9 mg m -2 h -1) significantly linked to high soil moisture content, low soil temperatures and the presence of Juncus gerardii. Carbon dioxide (CO 2) efflux was greater from the un-grazed marsh (mean of 420 mg m -2 h -1) than the grazed marsh (mean of 333 mg m -2 h -1) throughout most of the year and was positively correlated with the deeper water table and greater soil temperatures. Grazing was not a significant predictor of nitrous oxide (N 2O) soil emissions. Global Warming Potential (GWP; over 100 years), calculated from mean yearly chamber fluxes for CH 4 and CO 2, did not differ significantly with grazing treatment. Seasonal variation in the key drivers of soil greenhouse gas efflux; soil temperature, moisture and water table, plus the presence or absence of aerenchymatous plants such as J. gerardii were more important to the magnitude of greenhouse gas emissions than grazing management per se.