From soaking wet to bone dry: predicting plant community composition along a steep hydrological gradient

QUESTION: Why do plants grow where they grow? Prediction of species' occurrence and abundance in relation to the environment is a core aim of ecology, as is understanding the link between environmental stressors and adaptive traits. Community assembly may be viewed as a sequence of filters, sorting species according to their functional traits. We ask if the strength of filtering changes along a strong hydrological gradient? Can we predict species′ relative abundance using few, but biologically relevant traits? And are strongly filtered traits better predictors of community assembly? LOCATION: A hydrological gradient from pond to dry limestone pavements on the Island of Öland, South Sweden. METHODS: Plant community composition and six morpho‐physiogical plant traits were measured along a pronounced gradient in water supply and soil depth. The strength of filtering was quantified using a trait dispersion index, while the prediction of species′ relative abundance and importance of individual traits was assessed with the community assembly by trait selection (CATS) model. RESULTS: We show that species are filtered by the hydrological environment through the traits root porosity, specific leaf area and resistance to water loss on drying. For individual traits, the strength of filtering waxes and wanes along the gradient. This strongly suggests that the mechanism, through which species are filtered into communities, acts through different traits as environmental conditions change along the gradient. The CATS model predicted 66% of the variation in species' relative abundances using six traits. In general, the traits subject to filtering also were most important in predicting species abundance. CONCLUSIONS: Few plant traits are exposed to environmental filtering across the entire hydrological gradient, and most traits are strongly filtered only in parts of the gradient (e.g. root porosity in wet soils and water loss on drying on thin dry soils). Evidence for congruence between trait dispersion indices and the CATS model was established, underpinning the importance to plant community assembly of environmental filtering of species through their traits. New functional traits relevant to a specific environmental gradient – and not just some standard traits from a public database – can contribute significantly to resolve how plant communities are assembled.