Biodiversity loss due to human-induced habitat fragmentation is one of the biggest concerns in ecology. Biological populations are rarely isolated in space and usually interact with others via migration in what are known as metapopulations. On a bigger scale, spatial dynamics play an important role in food web structure and maintaining diversity, however, there is not much information on the interplay of these factors and the topology of the metapopulation (i.e., network pattern described by the local populations when they are connected). Network connectivity patterns can have critical effects on network robustness, as some topologies can promote resilience after perturbations. However, at present, experimental evidence of how these patterns affect population persistence in a metapopulation framework is lacking. Such experimental approaches could be useful to identify vulnerable local populations that need protection to maintain regional persistence and stability. In this study, we used aquatic protists to determine how network topology influences the regional persistence of the ciliate Paramecium tetraurelia. We created 18 metapopulations using two attached 24-well plates, where local populations (single wells) were connected using capillary tubes. We compared metapopulations arranged as random networks to power law networks by evaluating local population persistence and abundance throughout ∼30 protist generations.
Results/Conclusions
Metapopulations of P. tetraurelia reached higher densities and higher occupancy (proportion of occupied patches) in randomly metapopulations compared to power law systems. The two types of networks also showed opposite patterns of temporal occupancy or "incidence": in landscapes connected in a random network pattern, incidence decreased with increasing patch degree (number of connections per patch), whereas in landscapes following a power law, incidence augmented with increasing degree. These contrasting results highlight the fact that node degree alone is not a good predictor of occupancy if topology is not accounted for. Differences in patch occupancy in a network have been attributed to large "regional” variations within the network (i.e. effects of large distances on dispersal processes). However, in this study both networks had a similar an average path length (average graph-distance between all pairs of nodes) and thus should have the same “regional” effect. Therefore, here the observed pattern can be attributed network topology, reinforcing the idea that topology itself does affect spatial patterns in populations. In the future, studying the conjugated effects of topology, habitat complexity and population−and community−dynamics, could take us one step closer to successfully conserving biodiversity.