Abstract
The idea that ecological networks are built in a sequence of colonization events is not new but has been applied mostly to competitive interactions. Similar processes act in trophic networks, i.e. food webs: a regional pool of species act as the source from which species colonize local areas, called the metaweb. Local food webs are realizations of metawebs that result from assembly processes influenced by migration, habitat filtering, stochastic factors, and dynamical constraints imposed by food-web structure. We analyse how the structure of a metaweb influence local food webs with different spatial scales, using an assembly model, a random model and properties at three levels: emergent global properties that take into account the whole network e.g. modularity, sub-structural properties that consider several nodes e.g. motifs, and properties related to one node e.g. topological roles. Three independent data-sets were included: the marine Antarctic metaweb (34.8 million Km2), the Weddell Sea (3.5 million Km2) and Potter Cove (6.8 Km2) food webs. Looking at the global properties, the metaweb presents a structure very different from the random model, while the local food webs follow the same pattern and are very similar to the assembly model. The assembly model only takes into account migration, local extinction and secondary extinctions. For sub-structural properties the metaweb and the local food webs also showed the same pattern against the random model, but we found differences compared to the assembly model that did not increase the local stability of food webs. Topological roles also showed differences between the metaweb and local food webs that were explained by the assembly model. We found that a great portion of the structure of the food webs is determined by evolutionary processes that act on large temporal and spatial scales. On the contrary, dynamical processes that favour stability have a small influence, but habitat filtering or dispersal limitations seem to be important factors that determine food web structure.