Geometric indices of population persistence in stochastically generated continuous space river networks

Background/Question/Methods

Freshwater scientists are increasingly demonstrating that the branching structure of river networks has substantial ecological consequences. This is especially true when considering population dynamics in river watersheds, where natural and anthropogenic disruptions to connectivity can dramatically reduce long term persistence potential.  We study population persistence using metric graphs which provide a continuous, spatially explicit model of the river network habitat. Unlike standard graphs, metric graphs encode a continuous branching system where edges represent actual habitat rather than just connections among discrete nodes. Within this framework, we stochastically generate hypothetical river networks with a variety of geometric features and explore the effects of network structure and the distribution of habitat within the network on the persistence of a hypothetical, highly mobile population.

Results/Conclusions

We identify an index, CM, related to the distribution of habitable volume in the network as a promising indicator of population persistence potential. The index CM is the distance from the river outflow point at which half of the habitable volume of the river network lies upstream of that distance. This index outperforms other metrics such as the maximum and minimum distance from the river outflow to an upstream boundary and the total habitable volume of a tree.  In doing so, it provides a better generalization of habitat length in the classical linear space models of a river segment than these other metrics, when considered in the context of persistence in river networks.