Mesh size affects macroinvertebrate descriptions in large rivers: examples from the Savannah and Mississippi Rivers

Abstract

We explored how mesh size can affect the description of the macroinvertebrate community in large rivers using data from artificial substrate samplers in the main channel of the Savannah River near Augusta, Georgia, and from fine sediments in the main channel of the Upper Mississippi River near Cape Girardeau, Missouri. Samples from the Savannah River were collected on five occasions between 2000–2003 and processed through coarse-mesh (1.8 mm) and fine-mesh (0.5 mm) sieves. Samples from the Mississippi River were collected annually in 2002–2004 and processed through coarse-mesh (1.18 mm) and fine-mesh (0.355 mm) sieves. These mesh sizes contrast procedures associated with long-term studies (coarse mesh) and procedures that are frequently used or recommended now (fine mesh). In both rivers, coarse mesh greatly underestimated densities, capturing only 35% of the total macroinvertebrates in the Savannah River and 20% of the total numbers in the Mississippi River relative to the fine mesh. As a result, the density and relative abundance of dominant taxonomic groups differed between mesh sizes and among sampling dates. The differences for relative abundance, assemblage structure, and biometrics between the fine and coarse meshes were not consistent between rivers. Non-metric Multidimensional Scaling indicated in the Savannah River that the overall macroinvertebrate assemblage structure differed based on year and site but not mesh size, whereas assemblage structure in the Mississippi River differed based on mesh size. Similarly, biometrics from data with coarse and fine meshes combined implied better water quality than the coarse mesh alone in the Savannah River but lower water quality in the Mississippi River. These results indicate that mesh size can have a significant impact on ecological studies of macroinvertebrates in large rivers, and suggest that the finer mesh produces a more accurate estimate of the structure and density of the macroinvertebrate community. Our results also suggest caution when different mesh sizes are involved in a large river study—it may be inappropriate to contrast data produced with different mesh sizes or to combine data from different mesh sizes to create a long-term perspective.