Abstract. Transpiration is a key process in the hydrological cycle,
and a sound understanding and quantification of transpiration and its
spatial variability is essential for management decisions as well as for
improving the parameterisation and evaluation of hydrological and
soil–vegetation–atmosphere transfer models. For individual trees,
transpiration is commonly estimated by measuring sap flow. Besides
evaporative demand and water availability, tree-specific characteristics
such as species, size or social status control sap flow amounts of
individual trees. Within forest stands, properties such as species
composition, basal area or stand density additionally affect sap flow, for
example via competition mechanisms. Finally, sap flow patterns might also be
influenced by landscape-scale characteristics such as geology and soils,
slope position or aspect because they affect water and energy availability;
however, little is known about the dynamic interplay of these controls. We studied the relative importance of various tree-, stand- and
site-specific characteristics with multiple linear regression models to
explain the variability of sap velocity measurements in 61 beech and oak
trees, located at 24 sites across a 290 km2 catchment in
Luxembourg. For each of 132 consecutive days of the growing season of 2014
we modelled the daily sap velocity and derived sap flow patterns of these 61
trees, and we determined the importance of the different controls. Results indicate that a combination of mainly tree- and site-specific
factors controls sap velocity patterns in the landscape, namely tree
species, tree diameter, geology and aspect. For sap flow we included only
the stand- and site-specific predictors in the models to ensure variable
independence. Of those, geology and aspect were most important. Compared to
these predictors, spatial variability of atmospheric demand and soil
moisture explains only a small fraction of the variability in the daily
datasets. However, the temporal dynamics of the explanatory power of the
tree-specific characteristics, especially species, are correlated to the
temporal dynamics of potential evaporation. We conclude that transpiration
estimates on the landscape scale would benefit from not only consideration of
hydro-meteorological drivers, but also tree, stand and site characteristics
in order to improve the spatial and temporal representation of transpiration
for hydrological and soil–vegetation–atmosphere transfer models.
Tree-, stand- and site-specific controls on landscape-scale patterns of transpiration