Winter CO2 flux from soil and snow surfaces in a cool-temperate deciduous forest, Japan

We measured diurnal and wintertime changes in CO2 fluxes from soil and snow surfaces in a Japanese cool-temperate Quercus/Betula forest between December 1994 and May 1995. To evaluate the relationship between these winter fluxes and temperature, flux measurements were made with the open-flow infrared gas analyzer (IRGA) method rather than with the more commonly used closed chamber method or the snow CO2 profile method. The open-flow IRGA method proved to be more successful in measurements of winter CO2 fluxes than the two standard methods. Despite colder air temperatures, soil temperature profiles were greater than 0 degreesC because of the thermal insulation effect of deep snowpack. This reveals that soil temperature is satisfactory for microbial respiration throughout the winter. Unfrozen soils under the snowpack showed neither diurnal nor wintertime trends in CO2 fluxes or in soil surface temperature, although there was a daily snow surface CO2 flux of 0.18-0.32 g m(-2). By combining this with other reference data, Japanese cool-temperate forest soils in snowy regions can be estimated to emit < 100 g m(-2) carbon over an entire winter, and this value accounts for < 15% of the annual emission. In the present study, when data for all winter fluxes were taken together, fluxes were most highly correlated with deep soil temperatures rather than the soil surface temperature. Such a high correlation can be attributed to the relatively increased respiration of the deep soil where the temperature was higher than the soil surface temperature. Thus, deeper soil temperature is a better predictor of winter CO2 fluxes in cold and snowy ecosystems.