Undergraduate student McMaster University Hamilton, Canada
Abstract: Background/Question/Methods
Premises and ample evidence support the view that metapopulation or any of its constituent populations have a lower probability of extinction than nonconnected populations. For example, dispersal may rescue local populations whenever they decline or disappear. This mechanism might also contribute to higher mean metapopulation density by reducing extremes of local variability. The dispersal delays local crashes in high-density populations because the dispersing numbers climb with the size of the population. At the same time, dispersing individuals counter incipient declines in other populations or accelerate their recovery. Thus, we hypothesize that dispersal should add up to a higher metapopulation density by delaying the time to a crash and accelerating recovery. Approach: We tested this expectation in an agent-based metapopulation model (NetLogo). The model used three levels of dispersal, landscapes of different heterogeneity (inter-habitat difference) and habitat suitability. We recorded population changes over 360-time steps and ten replicates for each combination of dispersal (0, 10, and 20% of individuals at each time step) and habitat quality.
Results/Conclusions
Metacommunity densities differed among simulated landscapes in response to their quality and structure. As expected, dispersing metapopulations persisted in the landscape at stable levels. In contrast, in the absence of dispersal, metapopulations showed an asymptotic trend towards 0, although none went extinct in 360 steps. With dispersal, terminal densities of metapopulations were 20 to 100 times higher than those of the isolated populations. We found some surprises, too. The 10% dispersal rate in every simulated scenario accompanied the highest and most stable population density. With equal consistency, a 20% dispersal rate led to lower metapopulation densities. This density drop ranged from 4 to 22% and was significant in every case. Thus, even though dispersal allowed the metacommunity to stabilize and maintain high densities across the landscape, it appears that the effect of dispersal is non-linear. We rule out one possible mechanism: a high rate of dispersal that might cause individuals to die in unsuitable patches. While plausible, the simulation scenario where all patches were the same (no interhabitat differences) produced the same pattern. In conclusion, the emerging possibility of non-linear effects of dispersal on a landscape level density may represent an intriguing flag for conservation strategies.
