PUB - Publikationen an der Universität Bielefeld

Gradual changes in the environment could cause dynamical ecological networks to suddenly shift from one state to an alternative state. When this happens ecosystem functions and services provided by ecological networks get disrupted. We, however, know very little about how the topology of such interaction networks can play a role in the transition of ecological networks when spatial interactions come into play. In the event of such unwanted transitions, little is known about how statistical metrics used to inform such impending transitions, measured at the species-level or at the community-level could relate to network architecture and the size of the metacommunity. Here, using hundred and one empirical plant-pollinator networks in a spatial setting, I evaluated the impact of network topology and spatial scale of species interactions on transitions, and on statistical metrics used as predictors to forecast such transitions. Using generalized Lotka-Volterra equations in a meta-network framework, I show that species dispersal rate and the size of the metacommunity can impact when a transition can occur. In addition, forecasting such unwanted transitions of meta-networks using statistical metrics of instability was also consequently dependent on the topology of the network, species dispersal rate, and the size of the metacommunity. The results indicated that the plant-pollinator meta-networks that could exhibit stronger statistical signals before collapse than others were dependent on their network architecture and on the spatial scale of species interactions.