Do ecological specialization and functional traits explain the abundance–frequency relationship? Arable weeds as a case study

International audience

English. Aim: The abundance–frequency relationship (AFR) is among the most-investigatedpatterns in biogeography, yet the relative contributions of niche-based processesrelated to ecological strategies, and of neutral processes related to spatial colonization–extinction dynamics, remains uncertain. Here, we tested the influences of ecologicalspecialization and functional traits on local abundance and regional frequency,to determine the contribution of niche-based processes.Location: France and the UK.Taxon: Vascular plants.Methods: We used two arable weed surveys covering 1,544 fields in Western Europe(France, UK), along with functional traits related to resource acquisition, resource requirements,flowering phenology and dispersal. We quantified specialization both toarable habitat and to individual crop types, and performed phylogenetic path analysesto test competing models accounting for direct and indirect relationships betweentraits, specialization, abundance and frequency. We performed the analysesfor all species in each country, as well as for a subset of the most abundant species.Results: Local abundance of weeds increased with their regional frequency, butthe relationship became negative or null when considering only the most abundantweeds. Specialization to arable habitat and to individual crop type either had a similaror opposite effect on regional frequency and local abundance explaining thesepositive and negative relationships, respectively. Regional frequency was not directlyexplained by any trait but indirectly by resource requirement traits conferring specializationto the arable habitat. Conversely, high local abundance was directly relatedto low seed mass, high SLA, early and short flowering.Main conclusions: Direct/indirect effects of functional traits on local abundance/regionalfrequency, respectively, supports a significant role of niche-based processes inAFR. Neutral spillover dynamics could further explain a direct linkage of abundanceand frequency. Similar causal paths and consistent influences of traits on specializationand abundance in the two studied regions suggest genericity of these findings.