Abstract
Understanding how water use and drought stress in woody plants change in relation to compositional, structural and environmental variability of mixed forests is key to understand their functioning and dynamics. Observational and experimental studies have so far shown a complex array of water use and drought stress responses to species mixing, but progress is hampered by the costs of replicating measurements. A complementary approach consists in using in silico experiments with trait-based forest ecosystem models, which have the advantage of allowing the interpretation of the net mixing effect as the result of specific combinations of trait differences. We explore the potential of such an approach using a novel trait-based forest ecosystem model with a strong focus on plant hydraulics and data from 186 mixed forest inventory plots including holm oak (Quercus ilex L.) and eight co-occurring species. Sensitivity analyses focusing on the effect of differences in individual plant traits indicate that water use and summer drought stress of holm oak trees respond primarily to the variation in competitor's height, root distribution and xylem hydraulic efficiency and safety. Simulations of pure and mixed stands across different combinations of climate aridity and stand leaf area index indicate that differences in traits may compensate for one another, so that the influence of a given trait (e.g. tree height) on water use or drought stress can be decreased or offset by the influence of another one (e.g. hydraulic efficiency). Importantly, we show that species mixing does not always have positive effects at the stand level. Overall, our simulation study shows that the complexity of species- and stand-level mixing effects on water use and drought stress arises primarily as the result of differences in key functional traits of the competitor, although stand structure and climate aridity may modulate mixing effects.