Fish assemblage composition along a gradient of unconventional natural gas development in north-central Arkansas streams

Natural gas extraction from shale basins has expanded exponentially due to technological advances in both horizontal drilling and hydraulic fracturing. The Fayetteville shale gas play, located in north-central Arkansas, has experienced rapid unconventional natural gas development (UNGD) over the last decade in areas that have not historically been exposed to large-scale drilling operations. Ongoing research conducted on actively developed gas plays has identified impacts of UNGD on the terrestrial landscape including conversion of forest land cover, fragmentation of habitat, and the reduction of riparian buffers. However, there remains a paucity of research investigating the impact(s) of UNGD on aquatic ecosystems. UNGD is a multi-stage drilling process that presents different environmental impacts to surface and ground waters at each stage of development. Potential impacts include: increases in suspended sediments from construction of the well and supporting infrastructure, alterations to the natural flow regime from water withdrawal, and contamination from fracking fluids.;We investigated fish assemblage composition in thirteen stream catchments in spring 2013 and eleven stream catchments in spring 2014 across a gradient of gas well activity (0 -- 4.88 gas wells/km2). Using a suite of Index of Biotic Integrity metrics we evaluated differences in fish assemblage composition related to ongoing UNGD and other dominant land use variables within the catchment. In 2013, we observed significant negative correlations for relative abundance of simple lithophilic spawners (rho=-0.70, P=0.008), invertivores (rho=-0.60, P=0.030), sensitive individuals (rho=-0.70, P=0.008), darters and madtoms (rho=-0.71, P=0.007) with gas well density (GWD; wells/km2) as well as a significant relationship between relative abundance of tolerant individuals ( Lepomis cyanellus, Lepomis macrochirus, and Ameiurus natalis ) with GWD (rho=0.85, P<0.001). These correlations were not observed between these metrics and other land use variables which included percent pasture/woody herbaceous land cover and catchment area (km2). Correlation analyses using densities for fish assemblage metrics found non-significant negative trends for benthic invertivores (rho=-0.48, P=0.093), benthic individuals (rho=-0.43, P=0.138), and darters and madtoms (rho=-0.48, P=0.093) with GWD, and a significant positive association between tolerant individual density and GWD (rho=0.78, P=0.002). Our data provided evidence for differences in fish assemblage composition along a gradient of UNGD that appeared to be driven by declining abundances of darter and madtom species and increasing abundances of green sunfish and bluegill with increasing GWD.