The metacommunity concept provides a flexible framework to ecologists considering intercommunity dynamics and local assembly mechanisms; however, disentangling the multiple sources of variation affecting community composition remains a challenge. The functional traits of species can potentially play important roles in determining their response to environmental change and biotic interactions. For instance, if dispersal limitation is trait-based, then the spatial configuration of suitable habitats may impose non-random selection.
Here, we integrate zooplankton species records from 1,234 lakes and ponds along the North American Cordillera – from Yukon Territory to California – to evaluate the factors driving community structure and generate hypotheses about how communities respond to environmental change. A variation partitioning approach was used to identify total and unique contributions to community composition from environmental predictors (climate, catchment/lake features, and fish stocking history), spatial structure (distance-based Moran’s eigenvector maps), and the geographic positions of waterbodies spanning more than 29° latitude, 25° longitude, and 3,700 meters elevation. Spatial patterns were further investigated using multispecies patch connectivity estimates, based on distances between occupied habitats, to infer community dispersal limitation. Analyses were conducted for entire communities, as well as functional groups defined by reproductive mode (asexual versus obligate sexual) and body size.
Results/Conclusions
We discovered that spatial structure accounts for the most variation in community composition across functional groups, while geographic positions explain the least. A closer examination of environmental predictors indicated that the combined contribution of catchment/lake features (morphometry, land cover, and lithology) explains more variation than climate (air temperature, precipitation, and solar radiation) or predation by introduced fish, highlighting the importance of terrestrial-aquatic habitat coupling. However, individual climate variables generally explain more variation than any single catchment/lake features, and the relative importance of variables differs between functional groups. For instance, fish stocking exerts a greater impact on slower moving asexual species and climate conditions have more influence on larger species, indicating potential environmental sensitivities based on species’ traits.
The patch connectivity model explains less variation than the full spatial structure or environmental models; however, connectivity is most important to obligate sexual species, suggesting that Allee effects constrain their colonization success. Although unknown sources of spatial autocorrelation explain more variation than our environmental predictors, a comparatively small portion of spatial structure is attributed to dispersal. These findings suggest that while environmental filtering may exceed dispersal limitation in mountain lake communities, both processes are deterministic and mediated by species’ functional traits.