Qualitative and quantitative sampling of lake fish communities

Abstract

We demonstrated that estimates of the relative abundance and patterns of covariation for fish species captured in baited minnow traps, plastic traps, fine- and coarse-mesh trap nets, and multimesh gill nets from 43 lakes differed greatly among the gears; contradictory results about species' relative abundance were often provided. We conclude that simple approaches to integrate catches from gears providing an overall estimate of species abundance in communities across lakes are compromised given the inconsistency in estimates of abundance and covariation. We recommend an increased effort towards the use of presence-absence data and their adequate sampling within lake community studies. We used a resampling approach to demonstrate the relative ability of the different gears to detect the presence of species within lakes under various levels of sampling intensity. The variation among lakes and constituent species necessitates the use of multiple gears to adequately detect species composition. Many studies probably underestimate the species composition within lakes given insufficient effort and too restricted use of different sampling gears. The need for increased sampling effort is of particular concern in areas focusing on biodiversity (e.g., species losses) and ecosystem health (e.g., ecological integrity) where failing to detect species may have substantial economic and management implications.