Th-BA-14
Conductivity Thresholds For Abundance Of Imperiled Stream Fishes

Thursday, September 12, 2013: 1:00 PM
Marriott Ballroom A (The Marriott Little Rock)
Nathaniel P. Hitt , Leetown Science Center, U.S. Geological Survey, Kearneysville, WV
Mary Rockey , Leetown Science Center, US Geological Survey, Kearneysville, WV
Michael Compton , Kentucky State Nature Preserves Commission, Frankfort, KY
Michael Floyd , Kentucky Ecological Services Field Office, US Fish and Wildlife Service, Frankfort
Kenneth McDonald , Tennessee Ecological Services Field Office, US Fish and Wildlife Service, Cookeville
We investigated the relationship between stream conductivity and abundance of two imperiled fishes endemic to Appalachian headwater streams: blackside dace (Chrosomus cumberlandensis, BSD) and Kentucky arrow darter (Etheostoma spilotum, KAD).  We evaluated co-located fish and water quality records collected by state and federal agencies within the native range of each species (n=308 and 126, respectively).  Both taxa exhibited significant declines in abundance with increases in conductivity.  Segmented linear regression techniques indicated breakpoints in abundance associated with conductivity levels of 343 µS/cm (BSD) and 258 µS/cm (KAD).  Bootstrapped 95% confidence intervals bounded breakpoints between 155-590 (BSD) and 123-632 µS/cm (KAD).  Boosted regression techniques also identified a nonlinear conductivity-abundance relationship and indicated the relative importance of conductivity over other water quality parameters (dissolved oxygen, pH, temperature) for BSD abundance.  The observed abundance-conductivity relationships were robust to autecological differences in local abundances (high for BSD and low for KAD), habitat use preferences (pelagic for BSD and benthic KAD), and feeding behaviors (invertivore/algivore/detritivore for BSD and obligate invertivore for KAD).  We discuss possible direct and indirect mechanistic pathways to explain the observed patterns and consider the potential effects of imperfect detection using simulation techniques.