| The CO2storage flux (Fs) in forest canopy is an important component of net CO2exchange (NEE) between the forest ecosystem and the atmosphere. Based on the measurements of CO2concentration from an8-level vertical profile in a temperate deciduous broadleaved forest at the Maoershan Forest Ecosystem Research Station in Northeast China, this study analyzed the nature of the Fs and compared various methods for calculating of the Fs. The objectives were to (1) analyze the diurnal variation of effective change in storage (Fs_E) and its contribution to the NEE,(2) quantify the errors of Fs calculated with three scalar concentration variables (i.e., CO2density, molar fraction, and mass maxing ratio),(3) compare the Fs calculated by different methods, i.e., one point at the eddy covariance measurement height (one point method) and profile methods; and (4) quantify the effect of resolution and configuration of the profile system on the accuracy of Fs estimation.The results showed that (1) The Fs_E had significantly diurnal variation, with the maximum variation during the transition periods in growing seasons. Ignoring the Fs_E would underestimate the NEE14.8%and11.7%, respectively in the nighttime and daytime on an-half hour scales.(2) The total mass of dry air in the eddy covariance (EC) measurement control volume (storage in dry air) was not constant. Therefore, the storage in dry air fluctuation caused CO2molecules diffusion between the forest ecosystem and the atmosphere, which is called the dry air storage adjustment term (Fsd, an abiotic source/sink of CO2). The relative magnitude of Fsd to the eddy flux (Fc) was large during the day-night transition periods and nighttime, ignoring the Fsd could introduce errors in estimating NEE. Three major error sources may be caused by the atmospheric processes in the Fs calculation:air temperature fluctuation was the most important error source; followed by the effect of water vapor in the warm and moist summer, but negligible in the cold and dry winter; and the influence of atmospheric pressure was generally low throughout the year. The Fs calculated from pc, cc, and χc were overestimated by8.5%,0.6%, and0.1%of the magnitude of the effective change in CO2storage, respectively. To minimize potential errors in Fs calculation, we recommend adopting the CO2mixing ratio method.(3) On half-hour scales, the Fs calculated by one point method underestimated the NEE by10.5%and18.5%in daytime and nighttime, respectively.(4) The effectiveness of the profile system in estimating Fs is determined by the number of sampling levels and their vertical distribution. Satisfactory estimates of Fs could be expected when the profile system was configured as≥4sampling levels with relatively even vertical distribution. Nevertheless, the layers below the canopy contributed to the total Fs more than those above the canopy, because time derivative of χc generally showed decreasing magnitude of mean and standard deviation with increasing heights at night and in the morning and evening transition periods. The lower canopy needs more sampling levels when configuring a profile system. To optimize a profile system, one needs considering both sampling layers and their thickness within a forest ecosystem. |
PDF Full Text Request