Dual Communities Characterize Structural Patterns and Robustness in Leaf Venation Networks

The structural organization of supply networks plays an important role in their functioning. In this work, we discuss how standard community detection analysis can be complemented by the definition of communities in the networks’ plane dual. Such communities are characterized not by weak but by strong connectivity between the different components of the network. We extract dual communities in leaf venation networks, finding that they can reveal structural features not visible to traditional community detection methods. Furthermore, we show that the presence of dual community structures suppresses failure spreading and that dual communities can be used to classify different leaf types.