Understanding patterns of invasions on Mediterranean coastal dunes : from environmental filtering to biotic interactions across scales

Abstract

As a consequence of the human mediated blurring of biogeographical barriers to species dispersal, many plant communities are becoming increasingly dominated by introduced species (“aliens”). Mediterranean sandy shores are threatened conservation targets at the European level, for which plant invasions are a recognized prime agent of alteration. This thesis takes advantage of the value of coastal dunes as model systems to understand features that make plant communities more invasible and to determine the implications of the spatial resolution used for the analysis. I examine alien distribution patterns on holocenic dunes of several central Italian regions both within large grid cells (ca. 35 km2) and within small random plots (4 m2 and 64 m2). At the coarse resolution I analyze broad latitudinal gradients building a data set of environmental drivers (climate and coastline movements) and of human drivers (land cover, urbanization and tourism). For the small resolution I built a dataset of environmental field measurements along the sea-inland gradient: soil features related to level of environmental stress (e.g. content in organic matter, soil moisture and salinity, etc.) and wind factors related to natural disturbance (sand burial, wind abrasion, etc.). I found that presence and abundance of alien species on Mediterranean coastal dunes is driven in different measure by environmental filtering, biotic interactions and human impact depending on the spatial scale and on the sampling resolution examined. With respect to extrinsic abiotic factors there are several main differences between regional and local scales. Along a regional latitudinal gradient a rich alien plant species pool is promoted mostly by mild climatic conditions and broad heterogeneous dune systems. Instead, large scale modifications brought about by humans impact Mediterranean littorals mostly directly by reducing native biodiversity, rather than indirectly by favoring alien species abundance. The situation is turned around when focusing on fine scale patterns. Propagule pressure mediated by human recreational structures explains most of the variability in the fine resolution alien distribution data. Nonetheless, clear directional gradients of environmental stress and natural disturbance can be quantified along the dune profile. This small scale spatial variation in abiotic conditions also explains differences in invasibility. Specifically, intermediate levels of environmental stress offer the best conditions for spread of alien species locally. Turning to intrinsic biotic factors, the sampling resolution influences two main aspects. First, the relationship between native diversity and invasion success turns from positive at large sampling resolutions to negative at fine resolutions supporting a recently proposed conceptual model. The positive relationship can be explained with comparable responses to broad scale extrinsic factors and to varying heterogeneity while the small scale negative relationship can be ascribed to competitive interactions. Second, by addressing a long lasting debate in invasion ecology (Darwin’s naturalization conundrum), this thesis demonstrates that the measure of the phylogenetic distance from successful invaders to the recipient community is also sensitive to sampling resolution. At the smallest spatial scale existing members of invaded plant communities are less related to most invaders than expected by chance (phylogenetic overdispersion). Yet with increasing scale, since species can cooccur while avoiding direct competition, native assemblages become more invasible for closely related species because of habitat filtering (phylogenetic clustering). In conclusion these findings suggest that many apparently contrasting patterns in the study of invasibility of communities can actually be reconciled by recognizing the implications of different sampling resolutions.