Do metacommunity theories explain spatial variation in fish assemblage structure in a pristine tropical river?

Abstract

Abstract Understanding processes driving patterns of species distribution and diversity is one of the main objectives of community ecology. There has been a growing recognition that local environmental conditions are not the only factor structuring ecological communities, and that large‐scale spatial variation and dispersal also have major influences. The aim of our study was to evaluate spatial variation in fish assemblage structure along the longitudinal fluvial gradient of the Bita River, a nearly pristine tributary of the Orinoco River in the Llanos region of Colombia. Standardised surveys conducted at 34 sites throughout the basin during the low water period in January and March 2016 yielded 25,928 fish specimens representing 201 species. Twenty‐seven environmental variables were recorded at each site, and asymmetric eigenvector maps were used to model spatial variables. To understand spatial variation in local fish assemblages and their relationships with the environmental and spatial variables, two approaches were used. First, we applied the elements of metacommunity structure framework, followed by a variation decomposition analysis that allowed the metacommunity to be classified according to six alternative patterns of species distribution and four alternative metacommunity paradigms. We hypothesised that at a basin scale a major fraction of variation in structure is explained by a pure environmental effect and metacommunity patterns should reveal a Clementsian distribution. At a more regional scale (localities within a river section), assemblages in upstream and downstream regions may reflect different metacommunity processes. Because headwater streams are isolated within the river network, they should receive fewer migrants and may have local assemblages strongly influenced by local environmental conditions and species sorting with one of three possible distributional patterns (Clementsian, Gleasonian, or evenly spaced) would be observed. Conversely, downstream sites closer to the river mouth should be influenced by high dispersal, resulting in a greater importance of spatial factors and the mass effect, with metacommunity patterns nested along the longitudinal gradient. Our results suggest that the fish metacommunity in the Bita River exhibits a Clementsian distribution, implying that species respond to the environmental and fluvial gradient as groups along the longitudinal gradient. These replacements were associated with environmental heterogeneity, especially regarding habitat features. Similarly, the variation partitioning analysis showed that the pure environmental component was higher than the pure spatial component, which is consistent with species sorting. In this paper, we demonstrated that variation partitioning and metacommunity structure analyses provided complementary findings to infer processes structuring fish assemblages in the Bita River. Both approaches identified species sorting as the principal structuring processes in this system. Therefore, strategies to preserve fish diversity in this system must emphasise maintenance of habitat heterogeneity and connectivity.