Abstract Nitrate (NO 3 ) leaching and nitrous oxide (N 2 O) emissions are known to be high from urine and dung patches in grazed pasture. However, the extent to which estimated total net greenhouse gas (GHG) emissions from grazed pasture are affected by the inclusion of excreta-generated spatial heterogeneity in models is not known. The net GHG emissions include changes in soil carbon (C) storage, direct N 2 O emissions and indirect N 2 O emissions from ammonia (NH 3 ) volatilisation and NO 3 leaching. The FASSET whole-farm model was used to simulate the effect of including heterogeneity on direct and indirect GHG emissions during 2 years of grazed grassland followed by 2 years of spring barley. Simulations were performed with nitrogen (N) fertiliser rates to grassland varying from 0 to 300 kg N ha −1 year −1 , with or without heterogeneity in the simulation model. For N fertiliser rates of 50–150 kg N ha −1 year −1 , the excreta inputs were about 50 kg N ha −1 year −1 greater in simulations without than with heterogeneity. This resulted from a higher plant uptake in simulations without heterogeneity. The simulated N efficiency (plant N off-take in relation to external N input) was about 5% higher for simulations without than with heterogeneity. N 2 O emissions from the grazed grassland were strongly affected by N rate and heterogeneity, whereas there was little effect of these factors on N 2 O emissions from the following spring barley. There was a highly non-linear response of N 2 O emissions to N fertiliser for simulations without heterogeneity and an almost linear response with heterogeneity. This meant that N 2 O emissions were higher for simulations with heterogeneity than without at fertiliser N rates below 100 kg N ha −1 year −1 , whereas the situation was reversed at N rates above 150 kg N ha −1 year −1 . The NO 3 leaching was higher after ploughing of the grassland than during the grazing period and inclusion of heterogeneity in the simulations increased NO 3 leaching in both the grassland and the spring barley. The estimated net GHG emissions increased with increasing N rate in both the grazed pasture and in the following spring barley. The effects were largest in the grazed grassland, and for N rates above 150 kg N ha −1 year −1 simulated GHG emissions were considerably higher without than with heterogeneity. However, for N rates below 150 kg N ha −1 year −1 there was little effect of including heterogeneity in the simulations on net GHG emissions from either grassland or spring barley.