Convergent diversity and trait composition in temporary streams and ponds

Background/Question/Methods

Although ponds and streams are fundamentally different, the strong abiotic filter imposed by intermittent hydrologic conditions may cause them to be more similar to each other than they are to their own perennial counterparts. There is a parallel consensus in flowing- and still-water science that flow regime and hydroperiod are the main drivers of ecological communities in freshwater ecosystems. However, there is a lack of comparisons that test whether similar patterns result from similar hydrologic processes across aquatic systems. Convergence between stream and pond communities may be expected if an environmental gradient is driving processes equally in both ecosystems. We took a functional trait perspective to compare invertebrate assemblages from woodland ponds in SE Ontario, Canada to arid-land streams in SE Arizona, U.S.A. from a common hydroperiod gradient. We tested the relationship between functional richness and taxonomic richness with linear regression and used redundancy analysis to analyze trait relationships in response to ecosystem and hydroperiod. We calculated the multivariate dispersion as an index of beta diversity for each hydrologic group. 

Results/Conclusions

Our findings suggest that similar hydrologic processes lead to similarity in invertebrate beta diversity and trait compositions, while biogeographic factors generate taxonomic diversity patterns. A redundancy analysis showed that short-hydroperiod streams and ponds converged in trait composition, whereas long-hydroperiod streams and ponds were divergent in trait composition. Beta diversity of invertebrate assemblages varied significantly between ponds and streams (ANOVA, P = 0.04) with greater beta diversity in intermittent sites. Congruence in beta diversity and trait composition of pond and stream invertebrate communities was highest in the habitats with the harshest hydrologic filter (short hydroperiods; P = 0.38). Intermittent ponds vs. streams and perennial ponds vs. streams were significantly different from one another in trait composition (P <0.001, P = 0.04, respectively). Streams consistently had higher functional richness for any incremental amount of taxonomic richness (ANCOVA: P <0.0001). In both ponds and streams a lengthening in hydroperiod reflected an increase in occupied niche space (functional richness; R² = 0.36, P = 0.008; R² = 0.527, P < 0.0001, respectively). Our results suggest that short hydroperiods imposes similar functional constraints on pond and stream ecosystems, which may allow us to utilize common management strategies for these disparate ecosystems.