Impact of alien plant invaders on the pollination networks of oceanic and continental islands

Background/Question/Methods

A growing number of studies is showing that positive interactions among species (mutualisms), specifically those established between plants and animals, promote the integration of invasive species in native communities. On the other hand, once integrated in the receptive community, the invasive species can dramatically alter the mutualistic interactions present in it, which, in turn, can influence the dynamics of the invasion. To date, such studies have adopted a ‘single species’ approach, focusing on one or a few plant or animal species. This approach is fundamental, but an important further step is to broaden the focus to the level of entire communities. The study of these mutualisms at community level is as novel as necessary, particularly on island ecosystems, since the current situation of disturbances is alarming. The main goal of this study is to test the effect of alien species on different network parameters that describe the network topology. For such purpose, we compare pollination networks in sites invaded by the alien cactus Opuntia spp. with sites where these alien species are absent. We build both 2-mode networks (mapping the interactions between plants and pollinators in the community) and 1-mode plants networks (mapping the facilitation-competition processes within plant species in the community). The study is performed both in the oceanic island of Tenerife (Canary Islands) and in Menorca (Balearic Islands, W. Mediterranean)

Results/Conclusions

Due to the effect of insularity and to the time (late spring) that Opuntia spp flower, pollination networks are much smaller than those reported from continental areas. Consistently in the two islands, the presence of Opuntia appears to decrease the linkage level for both plants and pollinators besides increasing the level of nestedness. By contrast, centrality and centralization measures in the 1-mode plant networks did not differ significantly between invaded and non-invaded sites, and had also similar path length and clustering coefficient. Moreover, such 1-mode networks behave as dissortative networks, in which highly connected plants (hubs) are connected to plants with low connectivity. This network topology appears to provide robustness against different types of disturbances. Our results point out to the usefulness of the network framework in mutualistic interactions to help understanding the ecology of the invasions, both their integration as their impact at the community level.