Use of fish community data to evaluate restoration success of a riparian stream

From 1985 to 1988, stream and riparian habitats in Pen branch and Four Mile branch began recovering from deforestation caused by the previous release of hot water from nuclear reactors. The Pen branch corridor was replanted w ith wetland trees in 1995 to expedite recovery and restore the Pen branch e cosystem. Pen branch, Four Mile branch, and two relatively undisturbed stre ams were electrofished in 1995/1996 to determine how fish assemblages diffe red between the previously disturbed and undisturbed streams and whether su ch difference could be used to measure restoration success in Pen branch. F ish assemblages were analyzed using nonparametric multivariate statistical methods and the index of biotic integrity (IBI), a bioassessment method bas ed on measurement of ecologically sensitive characteristics of fish assembl ages. Many aspects of fish assemblage structure (e.g. species richness, dis ease incidence, taxonomic composition at the family level) did not differ b etween disturbed and undisturbed streams; however, the disturbed streams we re characterized by higher densities of a number of species. These differen ces were successfully detected with the multivariate statistical methods, w hereas, the IBI did not differ between most recovering and undisturbed samp ling sites. Because fish assemblages are strongly influenced by instream ha bitat, and because instream habitat is strongly influenced by the riparian zone, fish assemblages can be used to measure restoration success. Nonparam etric ordination methods may provide the most sensitive measure of progress towards restoration goals, although the IBI can be used during early stage s of recovery to indicate when certain ecologically important aspects of st ructure and function in recovering streams have reached levels typical of u ndisturbed streams. (C) 2000 Elsevier Science B.V. All rights reserved.