AbstractLivestock agriculture is a major source of anthropogenic greenhouse gas (GHG) emissions, with a substantial proportion of emissions derived from manure management. Accurate estimates of emissions related to management practices and climate are needed for identifying the best approaches to minimize, and potentially mitigate, GHG emissions. Current emissions models such as those of the IPCC, however, are based on emissions factors that have not been broadly tested against field‐scale measurements, due to a lack of data. We used a diverse set of measurements over 22 months across a range of substrate conditions on a working dairy to determine patterns and controls on soil‐based GHG fluxes. Although dairy soils and substrates differed by management unit, GHG fluxes were poorly predicted by these or climate variables. The manure pile had the greatest GHG emissions, and though temperature increased and O2 concentration decreased following mixing, we detected almost no change in GHG fluxes due to mixing. Corral fluxes were characterized by hotspots and hot moments driven by patterns in deposition. Annual scraping kept the soil and accumulated manure pack thin, producing drier conditions, particularly in the warm dry season. Summed over area, corral fluxes had the greatest non‐CO2 global warming potential. The field had net CH4 consumption, but CH4 uptake was insufficient to offset N2O emissions on an area basis. All sites emitted N2O with a similar or greater climate impact than CH4. Our results highlight the importance of N2O emissions, a less commonly measured GHG, from manure management and present potential opportunities for GHG emissions reductions.