How does network structure change robustness to co-extinctions in mutualisms?

Background/Question/Methods

Mutualistic interactions are currently threatened by species extinctions. These losses can disrupt multispecies communities which provide critical ecosystem services such as seed dispersal and animal pollination. Mutualisms largely occur in complex networks of interactions with characteristic network structures. These structures are thought to provide redundancy and stability if partners are lost, yet the mechanisms by which whole-network structures lead to network stability are unclear. One such network structure that appears to be important is nestedness. Here we used a simulation modeling framework to isolate two intertwining ways networks can be nested: in their binary structure and in their quantitative structure. Using a fully factorial design we created networks that had high, low, or random amounts of nestedness in terms of the presence-absence of links as well as had high, low, or random amount of nestedness in terms of the interaction strengths of links. We simulated sets of networks with biologically realistic number of species and density of links (connectances). We assessed these networks’ robustness to coextinction via a stochastic simulation model.

Results/Conclusions

Our stochastic model finds that more densely connected networks are less robust to species co-extinctions than networks that are more sparsely connected. This contrasts previous findings with purely deterministic robustness simulations which find greater robustness with increased network connectance. Overall the stochastic model introduces more extinctions into the networks due to cascading co-extinctions leading to lower overall robustness in the stochastic model compared to the deterministic simulations across all connectances and network sizes. Though we continue to collect and analyze data, preliminary results indicate that robustness to co-extinction is sensitive to binary and quantitative nestedness combinations. This indicates that more empirical and theoretical work should be conducted to understand the mechanism driving the ubiquity of this common mutualistic network structure.