Comparison of eddy covariance and chamber-based methods for measuring CO2 flux in a temperate mixed forest

Two methods, eddy covariance and chamber-based measurements, were employed to measure the net ecosystem CO(2) exchange in a mature temperate mixed forest in 2003. The eddy covariance system was used as a reference, which was compared with the chamber-based method. Based on chamber fluxes, the ecosystem had a gross primary production of 1490 g C m(-2) year(-1), 90% of which was released as efflux back into the air via respiration of the entire ecosystem. This was comprised of about 48% from soil surface CO(2) efflux, 31% from leaf respiration and 21% from stem and branch respiration. Net ecosystem exchange (NEE), estimated from the sum of daily component fluxes, was 146 g C m(-2) year(-1). Ecosystem respiration (ER), estimated from the sum of daily ecosystem respiration, was 1240 g C m(-2) year(-1). NEE was 9.8% of actual gross primary production (GPP). The eddy covariance estimates of NEE, ER and GPP were 188, 1030 and 1220 g C m(-2) year(-1), respectively. The eddy covariance estimation of NEE was higher than that of the chamber-based estimation by 22.5%. On a daily basis, NEE of the scaled chamber measurements was in acceptable agreement with eddy covariance measurement data with R(2) values of 0.71. The discrepancy between the measurement of the two methods was greater in the non-growing season primarily due to the lack of spatial variability in the scaled chamber estimates and weak atmosphere turbulence by eddy covariance measurements. There are many uncertainties for determination of absolute values of ecosystem component flux. More detailed experiments and related theoretical studies are needed in the future.