Wintertime CO2 exchange in a boreal agricultural peat soil

We measured the carbon dioxide (CO2) exchange with the eddy covariance (EC) method through three winters above a cultivated peat soil. During the first winter, the soil was ploughed, while for the next two winters it was grass-covered. On a weekly timescale, the emission was controlled by the soil temperature, whereas the vegetation had no clear impact. The deeper soil temperatures better correlated with the CO2 efflux, especially in frozen soil. The correlation with the air temperature was poor. After a mid-winter snowmelt, decreased CO2 efflux rates were temporarily detected, probably resulting from a lowered diffusion of CO) from the soil air into the atmosphere. Moderate soil-thaw CO pulses were observed in the springs of 2001 and 2003. CO2 emission rates measured with the EC method were found to be significantly lower as compared to those measured with the chamber method. The cumulative CO, emission between December and mid-March ranged from 80 to 178 m(-2) during three winters, correlating positively with air and soil temperatures and the number of snow-free days during that period. The projected increase in the air temperature related to global warming would boost the wintertime CO) efflux at our site by 30-200% (35-114 g m(-2)), depending on the selected emission scenario.