Changing assembly processes during a primary succession of plant communities on Mediterranean roadcuts

Aims Studying plant ecological succession provides insights into the spatiotemporal processes underlying community assembly and is of primary importance for restoration ecology. We investigate here colonization events and local community assembly over an original primary succession occurring on roadcuts after roadwork. For this, we addressed both the changes in species presence-absence (incidence data) to highlight pre-establishment filters and in species relative abundances to further assess the influence of local biotic processes.Methods We studied 43 limestone roadcuts in Mediterranean France, covering five age classes up to an age of 80 years, along with 13 natural cliffs as a reference, and we counted 14322 plant individuals on these sites. We applied a constrained nonsymmetric correspondence analysis of both the incidence (presence-absence) and abundance data to assess the variation of these data along the chronosequence.Important findings Along the first 30 years, the initially abundant short-lived species declined both in terms of incidence and abundance and were replaced by longer lived herbaceous and woody species. This first phase was characterized by species that are widespread in the surrounding scrublands and was comparable to an early secondary succession there. After 30 years, there were continuing changes in incidence data with age, but no more significant change in species' abundances. This second phase was marked by the late colonization of specialists that did not become dominant. Although colonization and establishment limitation was thereby apparent for specialist species, a slow convergence of community composition toward the situation of natural cliffs could be detected in the older stages of the chronosequence. These findings convey insights into the natural dynamics of man-made outcrop plant communities and may be useful for the ecological management and restoration of such contexts. It also illustrates the interest of comparing incidence and abundance data to investigate the relative influence of ecological determinants on the assembly of plant communities.