Evaluating biotic integrity and spatio-temporal variation in tributary fish communities

Tennessee is home to some of the most diverse freshwater fish communities in the world. The Caney Fork River below Center Hill Dam in middle Tennessee has three tributaries, a 3rd order spring-fed stream (Indian Creek), a 4th-order stream (Smith Fork Creek) and a 3rd-order stream (Hickman Creek); thus, there was an opportunity to evaluate differences in biotic integrity among three tributaries of a regulated river. The purpose of this study was to calculate and compare indices of biotic integrity (IBI) in those three tributaries over two seasons using two sampling protocols. The protocol developed by the Tennessee Valley Authority (TVA) relies on species-accumulation curves to dictate sampling effort. The second protocol developed for Wisconsin streams requires sampling over a reach that is 35 times the mean stream width (MSW). At five sites in those three tributaries over two seasons in 2010, more than 27,000 fish were collected representing 53 fish species. The IBI scores were “good to excellent” at three of the five sites, “good” at one site, and “fair” at one site. More species were collected at most sites for both seasons when following the 35 MSW protocol. Index of biotic integrity scores varied temporally more than spatially.;Plymouth routines in multivariate ecological research (PRIMER) software, including analysis of similarity (ANOSIM) and similarity percentage breakdown (SIMPER) procedures, were originally developed to analyze changes in marine communities. The ANOSIM procedure utilizes a Bray-Curtis similarity matrix based on species abundance data to determine if there is a significant difference between factor groups. The SIMPER analysis determines which species were most influential in separating factor groups. Based on the ANOSIM analysis, all factors tested (tributary, site, season and macrohabitat) significantly affected the fish assemblages in the three tributaries of the Caney Fork River. Banded sculpin Cottus carolinae (67.11%), central stoneroller Campostoma anomalum (24.01%) and bluntnose minnow Pimephales notatus (44.11%) were the leading contributors to similarity between samples within Indian Creek, Smith Fork Creek and Hickman Creek, respectively. Comparisons between seasonal samples were more similar than comparisons among samples from different macrohabitat types. The fish assemblages in run riffle and pool macrohabitat were all significantly different from each other.