Representative estimates of soil and ecosystem respiration in an old beech forest

Respiration has been proposed to be the main determinant of the carbon balance in European forests and is thus essential for our understanding of the carbon cycle. However, the choice of experimental design strongly affects estimates of annual respiration and of the contribution of soil respiration to total ecosystem respiration. In a detailed study of ecosystem and soil respiration fluxes in an old unmanaged deciduous forest in Central Germany over 3 years (2000-2002), we combined soil chamber and eddy covariance measurements to obtain a comprehensive picture of respiration in this forest. The closed portable chambers offered to investigate spatial variability of soil respiration and its controls while the eddy covariance system offered continuous measurements of ecosystem respiration. Over the year, both fluxes were mainly correlated with temperature. However, when soil moisture sank below 23 vol.% in the upper 6 cm, water limitations also became apparent. The temporal resolution of the eddy covariance system revealed that relatively high respiration rates occurred during budbreak due to increased metabolic activity and after leaf fall because of increased decomposition. Spatial variability in soil respiration rates was large and correlated with fine root biomass (r(2)=0.56) resulting in estimates of annual efflux varying across plots from 730 to 1,258 (mean 898) g C m(-2) year(-1). Power function calculations showed that achieving a precision in the soil respiration estimate of 20% of the full population mean at a confidence level of 95%, requires about eight sampling locations. Our results can be used as guidelines to improve the representativeness of soil respiration measurements by nested sampling designs, being applied in long-term and large-scale carbon sequestration projects such as FLUXNET and CarboEurope.