Interacting effects of spatial gradients and fishing gears on characterization of fish assemblages in large reservoirs

Understanding the variation of fish assemblages in reservoirs is crucial for precise interpretations of ecological processes. However, representative samples in such environments are subjected both to spatial zonation (longitudinal and transversal gradients) and limitations of fishing gears. We used a large zoned reservoir (Itaipu, Brazil) to perform a quasi-experiment that assessed the relative role of spatial gradients and fishing methods on fish diversity and composition. Active (cast nets, electrofishing and seines) and passive (gillnets and longlines) fishing gears were employed in lotic, transitional and lentic habitats of the reservoir and three large tributaries. In total 6281 individuals of 101 species were captured. Sampling methods and spatial gradients had an influence on species composition, with significant interactions between these factors. In addition, results produced by a variation partitioning analysis suggested that 21 % of the species richness was explained by the sampling methods and 7 % by the longitudinal gradient; whereas for species composition, 12 and 3 % corresponded to sampling methods and transversal gradients, respectively. Overall, our results suggest that variations among fishing gears is greater than across strong spatial gradients, emphasizing the importance of using multiple and complementary fishing methods in reservoirs for characterization of fish assemblages. The interactions between spatial gradients and sampling methods were higher with transversal than with longitudinal gradients, suggesting that substantial differences in species richness and composition among sampling gears depend to some extent on habitat. The approach used here would be applicable to any large zoned inland waterscape.