Estimation of the carbon sequestration by a heterogeneous forest: night flux corrections, heterogeneity of the site and inter-annual variability

Continuous measurements of the net CO2 flux exchanged in a mixed forest with the atmosphere were performed over 5 years at the Vielsalm experimental site. The carbon sequestration at the site was deduced by a summation of the measurements. Problems associated with this summation procedure were discussed. The carbon sequestration in the ecosystem was presented and its interannual variability was discussed. An estimation of the night flux correction was given. The correction was applied by replacing measurements made during quiet nights by a parameterization. The impact of the correction was shown to vary between 10 and 20% of the uncorrected flux, according to the year. The need to include the storage flux during turbulent periods was emphasized: its neglect leads to an error which will be greater than the one it tries to correct. It was also shown that the heterogeneity of the site made it necessary to split the data into separate series corresponding to the different vegetation patches and to fill the data gaps by using an algorithm that takes account of the weather conditions. Two series were defined, one corresponding to a beech subplot, the other to a conifer subplot. The uncertainty owing to the data split and the data gap-filling was about 15-20% annually. The carbon sequestration was then analysed in both the subplots. The length of the growing season was about 210 days in the beech and 240 days in the conifer. The carbon sequestration over 5 years was 2.28 kg C m(2-2) in the beech and 3.58 kg C m(2-2) in the conifer. The main difference between the species appeared in spring, between March and May, when the beeches were leafless. Significant interannual variations were observed in both the subplots. They appeared mainly in summer and were primarily because of the variations in the radiation and air humidity regimes. In addition, an impact of the interannual variation of the vegetation area index (VAI) and of the leaf initiation date was observed in the beech. Finally, a decline of the carbon sequestration efficiency of the ecosystem during the season was observed in both the subplots. It was because of neither the variation in any climatic variables nor VAI variation.