Patrones emergentes en redes de polinización en comunidades costeras de Baleares y Canarias

Abstract

Complex networks analysis allows ecologists to study the multitude of interactions that occur in ecosystems. The description of more realistic and complete interaction networks provides information not only about the characteristics of the communities but also of each one of its components. Using the global perspective offered by these plant-pollinator networks, this thesis provides relevant information that can be used in the field of conservation and island ecosystem restoration. To move towards a more functional description of the structure of ecological networks, we need interaction weights to be as accurate as possible. Despite the considerable work carried out to assess the structure of mutualistic networks, little is known about how different ways of measuring the interaction weights can influence network parameters used to describe such structures. In the first chapter of this thesis we evaluated how the use of five different link weights, including progressively more detailed information on the plant-pollinator interactions, influences widely used descriptors of network structure. In the two communities studied, the type of interaction weight used significantly alters the network structure. Thus, carefully considering the most appropriate interaction weight for each ecological network study can minimize potential biases and/or misinterpretation of network structures. Much ecological research has focused on trying to understand why species are generalized or specialized in their interactions and how networks develop in a certain environment. However, less information is available on whether generalized and specialized species differ in particular traits of their interacting partners. In the second chapter, differences in partner’ traits (specifically, their abundances and functional richness) among species, previously categorized according to two indices of specialization, are studied, as well as the temporal variation of these relationships. The indices of specialization used are: the number of partners, also called linkage level (L), and complementary specialization d´. Specialized and selective species, considered the most vulnerable to perturbations, reduce the risk of extinction by interacting with abundant species and/or functionally different "partners", respectively. On the other hand, generalized flower-visitor species are the only ones showing seasonal shifts in their partners’ traits, driven by changes in community species composition. Islands are of particular importance for the conservation of global biodiversity. Their small population sizes and the high level of endemism make them particularly susceptible to anthropogenic change. Two important biological differences between oceanic and continental islands are that the former usually have lower species richness and a higher degree of taxonomic disharmony than the latter. Although there is some evidence of less complex interaction networks on oceanic islands, only a few studies have compared island communities with each other or with mainland communities. In the chapter three, species composition and structural properties of quantitative flower-visitor networks in two communities of each type of island are analyzed: Canary Islands (oceanic origin) and Balearic Islands (continental origin). A greater number of opportunistic species and the presence of supergeneralist species on the oceanic island led to significantly higher values of quantitative connectance and nestedness compared to those on the continental island. Given that these two parameters appear to be associated with network stability, pollination communities on oceanic islands appear to be less vulnerable to perturbations than those of continental islands. Habitat loss or habitat modification represents the main threats to the survival of both plants and their flower visitors. The impact of such disturbances on those species interactions can lead to changes in the entire flower-visitor network structure. In the fourth chapter, the effect of habitat loss (more than 50% of land cover) on interaction patterns between plants and their flower visitors in a dune community, on Mallorca Island, is evaluated. A significant decrease in both species richness and abundance of floral visitors as well as in the number of network interactions was observed after the disturbance. Beetles, flies and ants were the functional groups most affected, whereas bees and wasps were the least affected, probably because of their strong flight capacity. Moreover, lower nestedness but higher selectivity and modularity of the community were detected after habitat loss. Overall, these results indicate that the flower-visitor network studied will be more vulnerable and less resilient to future perturbations. In conclusion, thanks to the complex networks analysis used in this thesis we comprehend that using an appropriate interaction weight when building interaction networks gives us more realistic and complete information about the communities of study. We also know, with more detail, the dynamism and species survival capacity in different communities using diverse strategies. Moreover, we observed the increased vulnerability of a community after the destruction of much of its habitat. The results of this thesis may contribute to protecting, as far as possible, natural ecosystems in the face of the huge kind of disturbances caused by human activities.