Environment and dispersal paths override life strategies and residence time in determining regional patterns of invasion by alien plants
MetadataShow full item record
We describe a novel dissimilarity framework to analyze spatial patterns of species diversity and illustrate it with alien plant invasions in Northern Portugal. We used this framework to test the hypothesis that patterns of alien invasive plant species richness and composition are differently affected by differences in climate, land use and landscape connectivity (i.e. Geographic distance as a proxy and vectorial objects that facilitate dispersal such as roads and rivers) between pairs of localities at the regional scale. We further evaluated possible effects of plant life strategies (Grime's C-S-R) and residence time. Each locality consisted of a 1 km(2) landscape mosaic in which all alien invasive species were recorded by visiting all habitat types.Multi-model inference revealed that dissimilarity in species richness is more influenced by environmental distance (particularly climate), whereas geographic distance (proxies for dispersal limitations) is more important to explain dissimilarity in species composition, with a prevailing role for ecotones and roads. However, only minor differences were found in the responses of the three C-S-R strategies. Some effect of residence time was found, but only for dissimilarity in species richness. Our results also indicated that environmental conditions (e.g. climate conditions) limit the number of alien species invading a given site, but that the presence of dispersal corridors determines the paths of invasion and therefore the pool of species reaching each site. As geographic distances (e.g. ecotones and roads) tend to explain invasion at our regional scale highlights the need to consider the management of alien invasions in the context of integrated landscape planning. Alien species management should include (but not be limited to) the mitigation of dispersal pathways along linear infrastructures. Our results therefore highlight potentially useful applications of the novel multimodel framework to the anticipation and management of plant invasions. (C) 2013 Elsevier GmbH. All rights reserved.