Aquatic Invertebrate Assemblages: Spatial and Temporal Differences among Sampling Protocols

Abstract

Listen

Translate article

Sampling protocols (different sampling gears used in different habitats) often vary among invertebrate studies. Several questions commonly arise: Which invertebrate attributes differ between pools and riffles? Are pool and riffle invertebrate assemblages differentially sensitive to the influences of human society? Can qualitative samples provide as much discriminating ability as quantitative samples? How can we best evaluate the status (biological condition) of the resident biota? We examined three sampling protocols: replicate Hess samples taken in pools, replicate Surber samples taken in riffles, and one qualitative sample from all habitats. Differences between Hess and Surber samples probably reflect differences in organisms found in pool and riffle habitats, whereas the qualitative protocol differed from the others in both habitats sampled and sampling effort. We determined differences among protocols for 18 attributes of invertebrate assemblages using data from yearly benthic samples (1986-1989) in three Tennessee River tributaries: Clinch, Powell, and Sequatchie rivers. Conditions differed among those rivers as reflected by fish assemblages (using the index of biotic integrity--IBI). We found distinct differences among the protocols and among many of the attributes selected for evaluation. Ten attributes either did not differ (stonefly and intolerant snail and mussel taxa richnesses, dominance, and proportions of omnivores, gatherers, grazers, and predators), or exhibited consistent differences (proportions of filterers, shredders, and chironomids), between riffles and pools. Of these ten attributes, six (stonefly taxa richness, dominance, and proportions of chironomids, filterers, omnivores, and predators) exhibited trends that coincided with our expectations based on fish assemblage condition. Eight attributes exhibited spatial and/or temporal trends that differed between riffles and pools (total, mayfly, caddisfly, and sediment-surface taxa richnesses, proportions of Corbicula, detritivores, and oligochaetes, and total abundance). When interyear differences among protocols were detected, trends among sites in pools, but not riffles, matched our expectations based on the biological condition of the fish assemblages (exception--the proportion of detritivores). Thus, monitoring programs may reach conclusions that are biased if biological condition is assessed with many commonly used attributes (e.g., total and mayfly taxa richnesses) and sampling is restricted to only one habitat. None of the taxa richness attributes (e.g., total and mayfly) distinguished differences among rivers in the multihabitat, qualitative protocol even though taxa counts were generally higher than in samples from either pools or riffles. Consequently, unreplicated, qualitative sampling may not have sufficient statistical power to detect subtle differences among streams. Replicated, quantitative sampling in both riffles and pools interpreted through a variety of biological attributes provides the strongest assessment of biological condition based on invertebrates.