Abstract :
[en] This thesis is dedicated to the study of climatic, physiological and phenological causes of the interannual variability (IAV) of the net CO2 exchange (or net ecosystem productivity (NEP)) of a temperate forest.
Until today, temperate forests act as a carbon sink. They absorb more carbon than they emit at annual scale without perturbations. However, the intensity of the sink is highly variable from one year to another and the causes of this variability remain to be explored. Furthermore, the fate of this sink itself is uncertain under the current climate change. This is why it is essential to study NEP variability in these ecosystems.
To do so, a measurement station was set up in 1996 at the heart of a beech-dominated forest in east Belgium. The forest also contains numerous coniferous species (Douglas fir, spruce, silver fir,...).
The mean NEP over the studied period (1997-2016) was around 411 gC/(m².an). The forest acted as a CO2 sink every single year. The range of the IAV was around 348 gC/(m².an). No trend was observed for the annual NEP. For the monthly NEP, an increasing trend was observed in April.
At the opposite, trends were observed for climatic (drier, sunnier and hotter spring and autumn, more precipitation in summer), phenological (shortening of the leafed period and of the carbon uptake period) and physiological (reduction of the quantum efficiency) variables. In particular, it was shown the climate evolution impacted negatively beech phenology (earlier senescence without a significant advance of leaf out) and positively coniferous phenology (earlier carbon uptake start). Combined with the lack of interannual trend for the photosynthetic capacity, these results suggest the lack of trend for annual NEP is essentially due to forest heterogeneity.
The analysis of annual NEP and its interactions with climatic, phenological and physiological variables has brought to light the role of the previous year climate conditions. More specifically, it was shown that a hot and dry summer the previous year diminish the growing season length, the photosynthetic capacity, the quantum efficiency and the annual NEP of the forest for the current year. These processes may likely be explained by the sensitivity to previous year climatic conditions of carbon allocation (directly or through carbon storage) to the fruits, the leaves, the wood or the roots. The main recommendation of the thesis for the future of the research at the Vielsalm station was to deepen the understanding of the carbon allocation processes that are likely controlling the forest annual NEP variability.
Before studying the IAV of the NEP, the validity and representativeness of the data were checked. In this context, the canopy proximity variability (due to tree growth and measurement height changes) impact on eddy covariance measurements was explored. No significant impact was observed for CO2 turbulent transport at the Vielsalm station. However, the CO2 flux was impacted suggesting an evolution of the composition of the area contributing to the measured fluxes. As a second step, the heterogeneity of the forest had, therefore, to be considered. A robust methodology was developed in order to allow the comparison of the measured CO2 flux for the whole period despite canopy proximity variability.
The thesis underlines the necessity, in the context of the climate change, to continue to measure the productivity, the phenology and the physiology at the Vielsalm station and to start collecting auxiliary measurements in order to confirm the obtained results and the formulated hypothesis.
