CO2 fluxes over an old, temperate mixed forest in northeastern China

Eddy-covariance measurements of carbon dioxide exchange were made above a temperate mixed forest for 16 months in northeastern China. The stand is a natural grown, unmanaged mixture of deciduous broad-leaved and coniferous forest and the trees cover a wide range of age classes with a maximum about 450 years. The objective of this study was to test the traditional hypothesis that forests at a late stage of development are insignificant carbon sinks. CO2 exchanges of the ecosystem were studied under conditions of adequate soil water supply. The main results are that:(1)Ecosystem respiration and its components, including soil, stem and leaf respiration, all fit exponential correlations with soil or air temperature, but the ratios of these components to total ecosystem respiration varied during the year. When temperature rose, the ratio of soil respiration to total respiration was relatively stable, while that of leaf and stem respiration increased and decreased, respectively. Soil respiration was the biggest component (48% of the whole ecosystem respiration on average, without including the respiration of fallen trees and coarse woody debris). Leaf respiration was the second largest (22–40%) and stem respiration was the least (13–29%).(2)The ecosystem has very large assimilation potential as environmental factors are adequate in growing season, but the actual gross ecosystem exchange (Gee) was only about 45% of the potential due to the influence of water vapor pressure deficit (VPD). High VPD reduced the photosynthetic capacity and the highest relative reductions occurred in afternoon when VPD was higher, with relative reductions between 0.3 and 0.6 during this time. Bigger reduction occurred in May and July, corresponding the 2 months when VPD are higher.(3)Respiration (Re) and net ecosystem exchange (Nee) were about 74% and 26% of actual Gee during growing season. Net assimilation began in the middle of April and ended in early September. The largest uptake occurred in June, then July and August. Large net CO2 losses occurred in late September and October because of relatively high temperature and leaf fall from most trees in the forest.(4)This forest was a carbon sink of 169–187 g C m−2 yr−1 over a 12-month-period commencing August 2002. Application of the u* correction decreased the annual uptake of carbon by 24.6 g C m−2 on average, corresponding to 13.5% of the corrected annual Nee.There are many uncertainties for determination of absolute values of long-term net carbon exchange even in sites with ideal topography using Eddy-covariance. More detailed experiments and related theoretical studies are needed in the future.