The primary aim of this study was to identify and compare the most significant sources of nitrous oxide (N2O) emissions from soils within a typical mixed livestock farm in Scotland. The farm area can be considered as representative of agricultural soils in this region where outdoor grazing forms an important part of the animal husbandry. A high temporal resolution dynamic chamber method was used to measure N2O fluxes from the featureless, general areas of the arable and pasture fields (general) and from those areas where large nitrogen additions are highly likely, such as animal feeding areas, manure heaps, animal barns (features). Individual N2O flux measurements varied by four orders of magnitude, with values ranging from −5.5 to 80,000 μg N2O-N m−2 h−1. The log-normal distribution of the fluxes required the use of more complex statistics to quantify uncertainty, including a Bayesian approach which provided a robust and transparent method for “upscaling” i.e. translating small-scale observations to larger scales, with appropriate propagation of uncertainty. Mean N2O fluxes associated with the features were typically one to four orders of magnitude larger than those measured on the general areas of the arable and pasture fields. During warmer months, when widespread grazing takes place across the farm, the smaller N2O fluxes of the largest area source – the general field (99.7% of total area) – dominated the overall N2O emissions. The contribution from the features should still be considered important, given that up to 91% of the fluxes may come from only 0.3% of the area under certain conditions, especially in the colder winter months when manure heaps and animal barns continue to produce emissions while soils reach temperatures unfavourable for microbial activity (<5 °C).