Development of an Index of Biotic Integrity for the Sand Hills Ecoregion of the Southeastern United States

Abstract We developed an index of biotic integrity (IBI) for the Sand Hills ecoregion within the upper Coastal Plain of the southeastern United States. This ecoregion exhibits differences in species composition among river basins and possesses a relatively depauperate fish fauna that has been affected by faunal homogenization. We also investigated whether this IBI could be improved by adding benthic macroinvertebrate information. Data collected from 70 wadeable stream sites were used to calculate 45 fish assemblage metrics. Twelve metrics within the categories of species richness, species composition, habitat guild, trophic guild, tolerance level, alien/invasive species, and individual fish condition were retained after screening all metrics for sensitivity to disturbance, redundancy, and ecological content. Heterogeneity of species composition among basins led to a greater reliance on richness, indicator species, trophic guild, and habitat guild metrics rather than on more geographically variable species composition metrics. Characteristic headwater assemblages provided the basis for some metrics, and invader species metrics were included to address faunal homogenization. The Sand Hills IBI (SHIBI) was more accurate ecoregion‐wide than IBIs that were developed for portions of the Sand Hills, correctly classifying about 60% of the disturbed sites compared with less than 30% for the other IBIs. However, the SHIBI was not highly responsive to gradients in site quality, as indicated by an R 2 of 0.30 between SHIBI scores and instream habitat quality scores. This R 2 value was improved to 0.49 by supplementing the SHIBI with a benthic macroinvertebrate index (the Ephemeroptera, Plecoptera, and Trichoptera index). We conclude that ecoregion‐wide IBIs should (1) be based on ecoregion‐wide data that characterize responses to disturbance and natural habitat gradients, (2) maintain a substantial ecological basis by including several metric classes, and (3) include metrics that address faunal homogenization where disturbance has increased the dispersion of nonnative species and native generalist species. In addition, combining information from different assemblages can enhance IBI responsiveness to disturbance and improve performance. Received May 5, 2016; accepted September 19, 2016 Published online December 8, 2016