The relative importance of local, regional and historical factors determining the distribution of plants in fragmented riverine forests:: an emergent group approach

Aim To investigate how local, regional and historical factors shape the herbaceous plant communities in fragmented riverine forests, and how the community composition and species richness of these fragments is related to the interplay between the environmental factors and specific plant life-trait combinations. Location Riverine forest fragments in the Grand-duche de Luxembourg. Methods Forest fragments were surveyed for their abundance in herbaceous plant species. All plant species where clustered into Emergent Groups (EG) by means of a formal classification based on 14 life-history traits. Within each EG, the local, regional and historical factors were related to the community composition using partial Canonical Correspondence Analyses (pCCA) and to the species richness using Generalized Linear Models (GLMs). The EG colonization ability was characterized by means of logistic regressions. Results We defined and characterized seven EGs, among which three consisted of forest specialist species (barochorous perennials, short geophytes and zoochorous perennials), which exhibited specific life-trait combinations: large and short-lived seeds and/or vernal phenology. Differences in EG composition between forest fragments were mainly explained by local environmental factors such as soil productivity and pH. The richness of barochorous perennials and short geophytes was well predicted by the historical and regional factors. The colonization ability appeared very low for barochorous perennials and short geophytes. Main conclusions Local environmental conditions appear to drive the differentiation of the riverine forest plant communities owing to the specific habitat requirements of many forest species. Spatial and temporal forest discontinuities affect the richness of forest specialist species, due to dispersal and/or recruitment limitations. The emergent group approach enhances the understanding of the relative influence of local, regional and historical factors by distinguishing between forest specialists from generalists or 'matrix' species, which have a masking effect.