International Journal of Biometeorology, 2010, Vol.54(1), pp.23-36
We investigated the water balances of two beech stands ( Fagus sylvatica L.) on opposite slopes (NE, SW) of a narrow valley near Tuttlingen in the southern Swabian Jura, a low mountain range in Southwest Germany. Our analysis combines results from continuous measurements of forest meteorological variables significant to the forest water balance, stand transpiration (ST) estimates from sap flow measurements, and model simulations of microclimate and water fluxes. Two different forest hydrological models (DNDC and BROOK90) were tested for their suitability to represent the particular sites. The investigation covers the years 2001–2007. Central aims were (1) to evaluate meteorological simulations of variables below the forest canopy, (2) to evaluate ST, (3) to quantify annual water fluxes for both beech stands using the evaluated hydrological models, and (4) to analyse the model simulations with regard to assumptions inherent in the respective model. Overall, both models were very well able to reproduce the observed dynamics of the soil water content in the uppermost 30 cm. However, the degree of fit depended on the year and season. The comparison of experimentally determined ST within the beech stand on the NE-slope during the growing season of 2007 with simulated transpiration did not yield a reliable statistical relationship. The simulation of water fluxes for the beech stand on the NE- and SW-slopes showed similar results for vegetation-related fluxes with both models, but different with respect to runoff and percolation flows. Overall, the higher evaporation demand on the warmer SW-slope did not lead to a significantly increased drought stress for the vegetation but was reflected mainly in decreased water loss from the system. This finding is discussed with regard to potential climate change and its impact on beech growth.
Water fluxes ; Beech stand ; Swabian Jura/SW Germany ; Forest meteorological measurements ; Sap flow measurements ; DNDC ; BROOK90 ; Modelling
View full text in Springer (Subscribers only)