COS 9-2 - Effects of the interaction between network position and dispersal mode on the structure and stability of metacommunities in a spatially-structured dispersal network

Abstract: While a large number of factors can contribute to the assembly and stability of ecological communities, when communities occur in a metacommunity with a strongly spatially-defined dispersal network, the properties of the network itself can exert a strong influence. Such spatially defined dispersal networks are common across many systems, but are perhaps most pronounced in riverine systems. In such systems, many organisms are confined to aquatic corridors as the only routes for dispersal. However, other organisms, like most benthic macroinvertebrates, have complex life cycles that allow for multiple modes of dispersal. To explore how the interaction between network architecture and dispersal mode can interact to affect community assembly, we conducted a large field experiment using replicated stream mesocosms. These systems were large (7m in length) flow-through systems that provided for near-natural stream conditions while allowing us to control some aspects of the system. The experiment featured two treatments, 1) location of mesocosms within the river network, and 2) a manipulation of dispersal. For the first treatment, we located 18 mesocosm systems throughout the network in sites spanning stream orders 1-3. For the dispersal manipulation, there were 4 treatment levels: 1) reduced in-stream dispersal; 2) reduced aerial dispersal; 3) reduced in-stream and aerial dispersal, and 4) a control with no dispersal manipulation. We ran the experiment for 18 months, with bi-weekly monitoring of the macroinvertebrate communities in the systems. At 25 days in the experiment, we also imposed a raking disturbance that mimicked disturbances created by flooding events in order to look at multiple metrics of community stability. We predicted that communities more central in the river network would recover more quickly from disturbance than more isolated communities, and that higher levels of dispersal would facilitate recovery.

Our results partially supported our predictions. While there were treatment effects on basic metrics like abundance and diversity, the responses did not conform to predictions. However, with stability, we found that sites which were more central in the network recovered more quickly from disturbances, and that treatments with higher dispersal also recovered more quickly. However, those results were most pronounced with regard to resilience, and much weaker for resistance and temporal variability. We also investigated determinism—the repeatability of true replicates—as a potential indicator of mechanism. Pieced together, our evidence suggested that in less-central sites, dispersal was fueling deterministic assembly, while in more well-connected sites, dispersal was leading to more stochastic assembly.