Metacommunities in river networks: Elucidating mechanisms that structure communities through surveys and experimentation

Background/Question/Methods

Metacommunity theory explains diversity by simultaneously considering local forces (species interactions and local environment), and regional forces that are the products of dispersal, and all local communities are structured by some combination of those forces.  When a metacommunity is located within a strongly delineated dispersal network, the relative balance of local and regional forces can differ greatly in different sections of the network.  For example, in the drainage network of a river, environmental conditions change markedly from headwaters to mainstems as does the degree of isolation.  We investigated how the relative balance of local and regional forces changes with location in river networks.  To accomplish this goal, we used two complimentary approaches:  1) Using a survey of 13 restoration sites + adjacent unrestored reaches, we decoupled local and regional effects since restoration sites represent a drastic change in local conditions yet share the same regional species pool as the unrestored reaches.  2) We used artificial streams in a 3-way factorial experiment manipulating a local factor (habitat complexity) and two aspects of the regional species pool—rate and source of dispersal—and examined the effects of these manipulations on macroinvertebrate community composition over a 10 week period. 

Results/Conclusions

In our survey, we found significantly lower community similarity in headwater reaches compared with larger (>2nd order) streams in the river network. These results suggest that the relative influence of local and regional forces is skewed towards local forces in in smaller, more isolated sections of river networks.  In our flume experiment, there were significant effects of all three treatments on community structure.  Dispersal and source pool had strong effects on taxon richness, while level of determinism (repeatability among replicates) was strongly affected by a habitat x source interaction.  In both high and low dispersal treatments, there was significant temporal shift in community composition that of in-stream community in the direction of the composition represented by dispersing organisms.  However, that shift was greater in the high dispersal treatment.  One contrast between the two studies was a strong role of local effects in the “mainstem” treatment regardless of dispersal rate.  This result may have been the result of underestimating the necessary dispersal to produce a “high dispersal” treatment.  Taken together, these results clearly indicate that position within a dispersal network affects the relative balance of forces shaping communities, and mechanistically demonstrates how those forces interact.