Significant CO2 fluxes from snow-covered soils occur in cold biomes. However, little is known about winter soil respiration on the eastern Tibetan Plateau of China. We therefore measured winter soil CO2 fluxes and estimated annual soil respiration in two contrasting coniferous forest ecosystems （a Picea asperata plantation and a natural forest）. Mean winter soil CO2 effluxes were 1.08 μmol m-2 s-1 in the plantation and 1.16 μmol m-2 s-1 in the natural forest. These values are higher than most reported winter soil CO2 efflux values for temperate or boreal forest ecosystems. Winter soil respiration rates were similar for our two forest ecosystems but mean soil CO2 efflux over the growing sea- son was higher in the natural forest than in the plantation. The estimated winter and annual soil effluxes for the natural forest were 176.3 and 1070.3 g m-2, respectively, based on the relationship between soil respiration and soil temperature, which were 17.2 and 9.7 % greater than their counterparts in the plantation. The contributions of winter soil respiration toannual soil efflux were 15.4 % tor the plantation and 16.5R for the natural forest and were statistically similar. Our results indicate that winter soil CO2 efflux from frozen soils in the alpine coniferous forest ecosystems of the eastern Tibetan Plateau was considerable and was an important component of annual soil respiration. Moreover, reforesta- tion （natural coniferous forests were deforested and refor- ested with P. asperata plantation） may reduce soil respiration by reducing soil carbon substrate availability and input.
