In this study, micro-level data from wood energy producers in Hedmark County were gathered and analysed. The aim was to find how much greenhouse gas (GHG) emissions various kinds of wood energy cause (not only CO2, but also CH4 and N2O), which energy they substitute, their potential to reduce GHG emissions, and the major sources of uncertainty. The method was life cycle assessment. Six types of wood energy were studied: fuel wood, sawdust, pellets, briquettes, demolition wood, and bark. GHG emissions over the life cycle of the wood energy types in this study are 2-19% of the emissions from a comparable source of energy. The lowest figure is for demolition wood substituting oil in large combustion facilities, the highest for fuel wood used in dwellings to substitute electricity produced by coal-based power plants. Avoided GHG emissions per in 3 wood used for energy were from 0.210 to 0.640 tonne CO2-equivalents. Related to GWh energy produced, avoided GHG emissions were from 250 to 360 tonne CO2-equivalents. Avoided GHG emissions per tonne CO2 in the wood are 0.28-0.70 tonne CO2-equivalents. The most important factors were technology used for combustion, which energy that is substituted, densities, and heating values. Inputs concerning harvest, transport, and production of the wood energy are not important. Overall, taking the uncertainties into account there is not much difference in avoided GHG emissions for the different kinds of wood energy. (c) 2006 Elsevier Ltd. All rights reserved.