Differences in carbon uptake and water use between a managed and an unmanaged beech forest in central Germany

• Mean net carbon uptake did not differ between the managed and the unmanaged forest.
• Net carbon exchange showed a larger interannual variability in the managed forest.
• The managed forest had a lower evapotranspiration than the unmanaged forest.

Net atmospheric carbon dioxide exchange, total evapotranspiration and net primary production of two neighbouring beech (Fagus sylvatica L.) forests in central Germany differing in site management were measured using the eddy covariance technique and biometric methods. The unmanaged site was an old-growth forest with admixtures of ash (Fraxinus excelsior L.) and sycamore (Acer pseudoplatanus L.) trees whereas the managed site was a regularly thinned, even-aged (about 130 years old), pure beech stand. Average carbon fluxes measured over seven years did not differ significantly between the two forests. Evapotranspiration was slightly higher (and consequently water use efficiency lower) at the unmanaged site. The maximum rates as well as the interannual variability of both net ecosystem exchange and net primary production were considerably larger at the managed, even-aged stand. The lowest annual carbon sequestration rates were observed in years with high fruit production in beech combined with cold and dry spells during leaf development which affected the carbon balance of the managed forest more than that of the unmanaged forest. In contrast, an extraordinarily dry period in late summer 2003 caused a stronger reduction in net carbon uptake in the old-growth forest which can probably be attributed to the contribution of ash to the ecosystem fluxes. Ash has a different phenology which made it more susceptible to the late summer drought. The relative importance of tree and structural (age, size) diversity, leaf area index and regenerative growth as well as the temporal frame and extent of single weather extremes for the forest–atmosphere exchange is discussed and it is concluded that site management and forest structure needs to be included in soil–vegetation–atmosphere-transfer models as it can often override effects of land use or plant functional type.