While enteric methane (CH₄) is the main source of greenhouse gas (GHG) emissions from dairy farms, the potential for CH₄ and N₂O emissions from manure is also significant. Dairy manure is typically managed in liquid form, but management practices range from year-round daily spreading to mandatory storage to recycle manure nutrients for crop production. In the former case, CH₄ emissions will be small while direct and indirect N₂O emissions are high. In the latter case CH₄ emissions during storage dominate the GHG balance of manure management, and manure nitrogen (N) may substitute fertilizer N. The GHG mitigation options available will depend on the baseline situation, and therefore prediction and mitigation at farm level requires a dynamic description of site conditions, management and associated emissions. Currently GHG inventories are based on guidelines from the Intergovernmental Panel on Climate Change (IPCC). With this methodology GHG emissions from manure management are calculated using statistical data and annual emission factors for a limited number of housing systems and climates. Each category must represent a range of site-specific conditions and practices, and the uncertainty of emission factors is therefore high. Country-specific emission factors may be adopted, but verification is costly and inflexible, and effects of management changes are not easily documented. Mechanistic models have been proposed to predict CH₄ and N₂O emissions during manure storage, but data requirements prevent routine use. An empirical model of CH₄ emissions during storage is currently explored which is based on daily time steps and few input parameters. The model allows for experimental verification and uncertainty estimation, and is suitable for investigating effects of management on GHG emissions. Water balance is key to control of N₂O emissions during storage, where populations of nitrifiers and denitrifiers can develop in natural crusts on stored manure. After field application soil water content is also critical by determining the redistribution of slurry constituents and subsequent turnover. Models to predict N₂O emissions must account for the distribution of reactive C and N in manure, and the balance between oxygen supply and demand.