Assessing impacts of the Roanoke River Flood Reduction Project on the endangered Roanoke logperch (Percina rex): Summary of Construction-Phase Monitoring
Roberts, James H.
Anderson, Gregory B.
Angermeier, Paul L.
MetadataShow full item record
The United States Army Corps of Engineers (USACE) has partnered with the City of Roanoke to carry out the Roanoke River Flood Reduction Project (RRFRP), a suite of floodplain modifications to the Roanoke River. The incidental take permit issued to USACE requires that USACE monitor populations of the federally endangered Roanoke logperch (Percina rex) prior to (Phase A), during (Phase B), and after (Phase C) construction, to estimate effects of incidental take during the course of the project. This report summarizes logperch relative abundance, suitable habitat, and water quality conditions across Phase B and compares these data to Phase A. We also conducted additional research in 2011 to assess the representativeness of permanent monitoring sites, estimate the sampling efficiency of the electrofishing methodology, and evaluate the statistical power and appropriateness of alternative impact-detection methods. Despite substantial fluctuation of the relative abundance of adult logperch over the course of monitoring, we found no statistical evidence for an impact of RRFRP construction. Thus USACE has maintained compliance with its incidental take permit during Phase B. Young-of-year logperch density, habitat availability, and water quality conditions also varied considerably over time and space during Phase B, but not in ways that could be attributable to the RRFRP. We found that permanent sites were representative of reach-wide conditions, suggesting that our findings can reasonably be extrapolated to the entire study area. The sampling efficiency of our standard electrofishing method was estimated to be low (~ 11%), yet our assessment method produced indices of abundance that were strongly correlated with true population estimates. Herein we demonstrate a new, generalized linear modeling approach to impact assessment that should provide greater insight and statistical rigor than the traditional t-test approach.