Abstract. Winter CO₂ flux is an important element to assess when estimating the annual carbon budget on regional and global scales. However, winter observation frequency is limited due to the extreme cold weather in sub-Arctic and Arctic ecosystems. In this study, the continuous monitoring of winter CO₂ flux in black spruce forest soil of interior Alaska was performed using NDIR CO₂ sensors at 10, 20, and 30 cm above the soil surface during the snow-covered period (DOY 357 to 466) of 2006/2007. The atmospheric pressure was divided into four phases: >1000 hPa (HP: high pressure); 985<P<1000 (IP: intermediate pressure); <986 hPa (LP: low pressure); and a snow-melting period (MP); for the quantification of the effect of the environmental factors determining winter CO₂ flux. The winter CO₂ fluxes were 0.22 ± 0.02, 0.23 ± 0.02, 0.25 ± 0.03, and 0.17 ± 0.02 gCO₂-C/m² d⁻¹ for the HP, IP, LP, and MP phases, respectively. Wintertime CO₂ emission represents 20 % of the annual CO₂ emissions in this boreal black spruce forest soil. Atmospheric temperature, pressure, and soil temperature correlate at levels of 56, 25, and 31 % to winter CO₂ flux, respectively, during the snow-covered period of 2006/2007, when snow depth experienced one of its lowest totals of the past 80 years. Atmospheric temperature and soil temperature at 5 cm depth, modulated by atmospheric pressure, were found to be significant factors in determining winter CO₂ emission and fluctuation in snowpack. Regional/global process-based carbon cycle models should be reassessed to account for the effect of winter CO₂ emissions, regulated by temperature and soil latent-heat flux, in the snow-covered soils of Arctic and sub-Arctic terrestrial ecosystems of the Northern Hemisphere.