Metapopulation persistence in a complex network results from the interplay between movement pathways and individual-specific movement behavior. In dendritic networks such as streams, some animals are restricted to travel along the network (i.e., within the stream), while others may be able to make out-of-network movements (i.e., over land between streams). The fractal nature of dendritic networks, with many opportunities for movement through the network, may provide substantial benefit to population persistence. However, many real networks (such as urban streams) do not have this regular structure, and should therefore have increased metapopulation extinction rates. I investigated the persistence of metapopulations in networks with two different topologies: (1) a “full” dendritic network with regular bifurcations and (2) a “pruned” (irregular) network with a reduced number of bifurcations. Within these networks I investigated the relationship between movement probabilities and metapopulation persistence. I then compared persistence times of salamander populations in full and pruned networks using empirical data from movement in real stream networks.
Results/Conclusions

 In the theoretical model, I found that out of network connectivity has a large effect on the time to extinction in both network configurations.  This effect was most prominent with high levels of within-network movement, suggesting that out of network movement is not a primary driver of extinction risk.  In the Pruned network with a high extinction probability, out of network dispersal did not ameliorate the extinction risk. Using empirical movement data on a stream salamander, I found that extinction risk was similar in full and pruned networks when both the extinction rate and the out of network dispersal were low.  The full dendritic network had a greater potential for population persistence when out of network movements were proportional to or greater than other modes of dispersal, even at high extinction rates.  These results suggest that understanding species-specific movement probabilities (and the propensity for out of network movements) are important for assessing metapopulation extinction risk. While out of network connectivity generally increases time to metapopulation extinction, the magnitude of the effect is correlated with the within network movement probabilities.  Realistic networks in nature may fall between the dendritic network topologies considered here, and thus understanding how network complexity interacts with population extinction risk is important for managing these habitats.