Dispersal limitation may result in the unimodal productivity‐diversity relationship: a new explanation for a general pattern

Abstract

Summary Variation in diversity with habitat productivity has long been a central ecological topic. Plant diversity is mostly highest at intermediate productivity, exhibiting the unimodal (so‐called ‘hump‐back’) relationship. This relationship has been explained by both evolutionary and ecological processes, but the potential role of dispersal limitation has not been considered. We used European flora data to show that dispersal limitation may contribute to the unimodal productivity‐diversity relationship. Species were characterized by their habitat productivity preference and dispersal probability (determined by the number of seeds and the presence of a dispersal syndrome). We calculated average relative dispersal probabilities for species assemblages occurring preferentially in different habitat productivity levels. At low productivity levels, species without dispersal syndromes predominate (R2 = 0.89), but at high productivity levels, species with a low number of seeds are common (R2 = 0.89). The total relative dispersal probability, combining both the dispersal probabilities attributable to the number of seeds and to the presence of dispersal syndrome, had a unimodal relationship with habitat productivity (R2 = 0.86). Thus, the unimodal productivity‐diversity relationship may arise due to the minimal dispersal limitation of local richness in conditions of moderately low productivity. At very low productivity, the lack of dispersal syndromes may limit dispersal. At high productivity, the low number of seeds may limit dispersal. Consequently, in conditions where species pool size and biotic interactions do not vary along productivity gradients, the variation in dispersal probabilities with productivity alone can produce unimodal relationships between diversity and productivity. Thus, dispersal limitation may contribute to the observed diversity pattern and ecosystem functioning to a greater extent than usually assumed.