Persistence and growth of a metapopulation are affected by factors beyond the within-patch dynamics of its constituent subpopulations, including by the topology of the network of dispersal corridors connecting its subpopulations. Metapopulations can vary in their total number of dispersal corridors and how those corridors are distributed spatially – the same number of dispersal corridors could be distributed very evenly such that all subpopulations are equally connected to their neighbors, or they could be clumped, leaving most subpopulations connected to only one or two neighbors while a few hub subpopulations are more highly connected. Because highly connected subpopulations are more likely to be rescued from local extinction by migration from neighbors, a metapopulation containing highly connected subpopulations is predicted to have a higher probability of recovery following disturbance than a metapopulation with fewer highly connected subpopulations.
I tested this prediction by comparing replicate metapopulations of the bacterium Pseudomonas syringae cultured in Metapopulation Microcosm Plates, devices I design and build which resemble 96-well microtiter plates with corridors connecting the wells to form metapopulations with 96 subpopulations. I compared metapopulation growth and subpopulation size following disturbances between treatments that varied in the spatial clustering of dispersal corridors and rate at which subpopulation re-colonization could occur.
Results/Conclusions
P. syringae metapopulations with highly connected subpopulations returned to their pre-disturbance abundances faster than those in evenly-connected networks and had higher variance in subpopulation size. Variance in degree of subpopulation connectivity was positively correlated with metapopulation persistence, ability to recover from disturbance, and heterogeneity of subpopulation persistence. In other words, evenly connected metapopulations recovered more slowly from disturbance, but their constituent subpopulations all shared similar probabilities of remaining occupied, whereas those with highly connected subpopulations were better able to recover from disturbance overall, but their more highly connected subpopulations were more likely to be occupied than less connected subpopulations.