The functional trait distinctiveness of plant species is scale dependent

Beyond the local abundance of species, their functional trait distinctiveness is now recognized as a key driver of community dynamics and ecosystem functioning. Yet, since the functional distinctiveness of a species is always relative to a given species pool, a species distinct at regional scale might not necessarily be distinct at local or community scale, and reciprocally. To assess the importance of scale (i.e. the definition of a species pool) when quantifying the functional distinctiveness of species, and how it might distort the ecological conclusions derived from it, we quantified trait distinctiveness of 1350 plant species at regional, local and community scales over ca 88 000 grassland plots in France. We measured differences in functional distinctiveness of species between regional (mainland France), local (10 x 10 km cell) and community (10 x 10 m plot) scale, and tested the influence of environmental predictors (climate and nitrogen input) and contexts (environmental distinctiveness, frequency and heterogeneity) on these variations. In line with theoretical expectations, we found large variation in functional distinctiveness (in particular between regional and community scales) for many species, with a general tendency of lower distinctiveness at smaller scales. We also showed that nitrogen input - a key aspect of high land use intensity - and environmental frequency partly explained the differences between local and regional scale only. These results suggest the role played by environmental filtering on species distinctiveness at local scale, but the determinant of distinctiveness variations at community scale still need to be elucidated. Our study provides robust empirical evidence that measures of ecological originality are strongly scale-dependent. We urge ecologists to carefully consider the scale at which they measure distinctiveness, as ignoring scale dependencies could lead to biased (or even entirely wrong) conclusions when not considered at the scale of interest for the respective research question.