Alternative approaches of transforming bimodal into unimodal mutualistic networks. The usefulness of preserving weighted information

Abstract

Pollination and seed dispersal networks are by definition bimodal, linking two sets of species, plants and animals. Bimodal networks are often analysed after being transformed into unimodal ones, since most attributes in network theory are defined for the latter. Such unimodal projections (e.g. of plants sharing flower visitors or seed dispersers) map potential inter-specific competition or facilitation, and can thus be useful for instance when identifying native species potentially sensitive to aliens in the communities. In this work, we introduce procedures to project unweighted and weighted bimodal networks into unimodals, for animals or plants, and calculate two centrality measures that inform us about the species' role in the communities. By using 20 empirical weighted networks worldwide, we obtained 160 unimodal networks via four projection methods and evaluated correlations among centrality parameters across the different methodologies to assess how consistent the results are when including different link weights between species. Degree centralities obtained by projecting unweighted and weighted bimodal networks were not significantly correlated, suggesting that the role of the species differs when considering link weights in the original bimodal networks. By contrast, betweenness centralities were highly correlated, indicating the consistent importance of the species as connectors regardless of the projection method used. We conclude that preserving the weighted information when transforming bimodal into unimodal networks may allow us to make more realistic predictions on the potential competitive or facilitative interactions among species of one set (e.g. plants) that share species of the other (e.g. flower visitors or dispersers). © 2011 Gesellschaft für ökologie.