Seasonal and annual variations in soil respiration in a cool-temperate deciduous broad-leaved forest in Japan

We investigated the seasonal and annual variations in soil respiration in a cool-temperate oak–birch forest in central Japan and, in particular, to reveal factors determining seasonal variations in the temperature dependence of respiration based on soil CO2 efflux data for 1999–2002. The daily soil carbon efflux was moderate in late spring (1.8–2.9 g C m−2 day−1 in May), increased sharply to a peak in summer (4.6–6.0 g C m−2 day−1 in August), and decreased in autumn (1.5–2.5 g C m−2 day−1 in November). In winter, the carbon efflux from the snow surface was low (0.29–0.71 g C m−2 day−1). Soil temperature exerted principle control on the seasonal and annual variation of soil respiration. Furthermore, reduced soil water content decreased soil respiration in summer when droughts occurred. The temperature function driven on a seasonal scale showed large variations in seasonal Q10 and R0 (simulated soil respiration at a soil temperature of 0 °C), e.g. a higher Q10 and lower R0 in spring and a lower Q10 and higher R0 in autumn. These large variations in seasonal Q10 and R0 might reflect confounding effects of temperature sensitivity and seasonal changes in physiological activities induced by root phenology, microbial biomass, and other factors. These results suggest that if the objective is to simulate soil respiration in a particular season, a seasonal Q10 function derived from the target period must be used, especially in a forest ecosystem possessing distinct seasonal changes. Conversely, the annual Q10 function derived from all measurements of daily soil temperature at −1 cm across the 4 years (Q10 = 3.8) was adequate for estimating total annual soil respiration because it may integrate all processes that influence soil carbon efflux, despite the fact it may under- or overestimate daily soil carbon efflux on a seasonal scale. Based on this single annual Q10 function, the annual mean soil respiration in 1999–2002 was estimated to be 853.5 ± 42.4 g C m−2 without topography correction, and 725.5 ± 36.0 g C m−2 with topography correction for a 1 ha experimental area. The estimated annual soil respiration (1999–2002) gave coefficient of variation (CV) value of less than 10%, suggesting a small inter-annual variability in annual soil respiration during 1999–2002 in this forest, since the annual mean temperature was similar in the 4 years of the survey. Carbon efflux from the snow surface accounted for 10% of the annual soil respiration, with a value of 84.3 ± 7.0 g C m−2, indicating the importance of quantifying winter CO2 efflux within a forest ecosystem in a cold and snowy region.