Habitat availability mediating tadpole richness in streamside ponds of Central Amazonia

Background/Question/Methods

Studies that identify spatial and environmental factors driving species richness in Amazon forest are important because they can clarify the process of structuring communities in a mega-diverse region. In this study, we addressed two questions: (1) Are environmental factors (biotic and abiotic variables) more important than spatial ones in tadpole richness distribution?; (2) Are habitat structure variables more important than biotic variables in tadpole richness distribution? Tadpoles and predators were sampled at 20 riparian plots over three sampling periods between February and June 2010 in a terra firme forest in Central Amazonia. In each plot we measured the eight environmental parameters at each plot: the width of the stream, width of the valley, canopy openness, number of ponds, distance from each pond to the stream, area and depth of each pond, and the depth of the leaf litter layer at the bottom of each pond. The spatial structure of tadpole assemblages was modeled using Principal Components of Neighbor Matrices analysis. To assess the influence of environmental and spatial variables on species richness of tadpoles in plots, we implemented a partial regression analysis. We used generalized linear models to assess the effect of environmental and biotic variables on tadpole species richness.

Results/Conclusions

Spatial and environmental components jointly explained 60.4% and 54.6% of the variation in observed and estimated tadpole richness, respectively. The pure environmental component accounts for the greatest fraction of variation in observed species richness (37.8%), while the spatial component explains the greatest fraction of the variation (37.9%) in the estimated species richness. Among all fitted models with the variables pond morphology, environmental descriptors and predators, three models are equally parsimonious. The best models are: i) with only the number of ponds, ii) with only valley width, iii) with number of ponds and valley width combined. Our results showed that tadpole species richness in streamside ponds in terra firme forests of Central Amazonia is influenced by niche-based processes and can be explained by local factors related to habitat structure (e.g., number of ponds and valley width) in the plots. However, neutral processes are important because spatial variation can explain the spatial distribution of species richness, probably as a result of dispersal limitation that reflects the biology and dispersal abilities of the species.