Carbon dioxide efflux from the floor of a boreal aspen forest. I. Relationship to environmental variables and estimates of C respired

Soil CO2 efflux (Soil Fc) in a boreal aspen (Populus tremuloides) forest was related to environmental variables to estimate the mass of carbon (C) annually respired. Measurements of soil Fc were made between April and September with a dynamic closed chamber via an infra-red gas analyzer. Over the course of the study day-time soil Fc ranged from 0.027 to 0.411 mg CO2 m(-2) s(-1). Summertime estimates of soil Fc were higher than expected, and this was attributed to underestimation by static chamber methods. Strong relationships were found between seasonal patterns of (1) soil Fc and soil temperature (Ts) at various depths (R-2 = 0.77 to 0.87, Q(10) = 3.9 to 5.1), (2) soil Fc and humus [CO2] and volumetric moisture (theta(v); R-2 = 0.76 to 0.88) and (3) humus [CO2] and humus Ts and theta(v) (R-2 = 0.74). On their own, theta(v), litter and soil organic matter content explained little (<5%) of the variation in soil Fc. Ts was the single most effective variable used to predict soil Fc. Annual masses of soil Fc C were estimated from the relationship between soil Fc and 0.1 m Ts in 1994. Long-term (49-yr) estimates were generated from monthly mean air temperatures (Ta) using the relationship between Ts (0.10 m) and Ta (2 m) at the site throughout 1994-1995. In both years, Ta explained about 70% of the variability in Ts. Estimates of annual soil Fc C were 905 (1994), 870 (1995) and 809 (long-term) g m(-2). Growing season soil Fc approximated 75% of the annual C mass respired. These annual C fluxes far exceed current estimates for boreal and most temperate forest ecosystems. 1994 soil Fc C also exceeded the site estimate of aboveground net primary productivity (361 g C m(-2) yr(-1)). This extreme shortfall, along with the inability of soil organic matter or litter mass to explain soil Fc, suggests that about 60% of annual soil Fc C is attributable to the presence of roots. If soil Fc in vegetated environments is primarily due to the presence of roots then the significance of soil Fc to global C cycling can only be understood with concurrent estimates of net ecosystem C exchange.