Unconventional shale gas development may impact surface water quality due to potential effects of construction and contamination by hydrofracking and flowback fluids. To that end, in 2012 we examined water quality and macroinvertebrate indicators as a function of drilling and pipeline construction in 12-15 streams in northeastern Pennsylvania – within the eastern extent of the Marcellus Shale region. Our study sites for this synoptic survey included second and third order streams with drainage areas ranging from 2.2 to 38.0 square miles, representing watersheds with no, low, medium, and high natural gas activities. Using GIS and land use classifications derived from National Land Cover Data (2006), we quantified number of wells, pipeline length and corridor area, land use/land cover, area, and stream classifications, on a watershed basis for each sub-watershed and stream buffer zone (100m) in the study area. Previously reported analyses indicated that macroinvertebrate indicators showed marked seasonal variation in response to land use and shale gas development, with energy development effects being most pronounced in spring and least apparent in fall. More recent multivariate analyses have focused on annual average values of benthic macroinvertebrate indicators vs. a subset of variables encompassing habitat, water chemistry, land use, and shale gas development.
Results/Conclusions
On this annual basis, stepwise regression analysis indicated that percent pipeline in watersheds accounted for significant decreases in EPT Richness (r2 = 0.66, F = 25.61, P < 0.0005), Beck’s Index (r2 = 0.64, F = 23.57, P < 0.0005), and a multi-metric Index for Biological Integrity (r2 = 0.55, F = 16.04, P < 0.001). Although some bivariate correlations were evident between these indicators and water chemistry (r = -0.57, P < 0.05), habitat (r = 0.56, P < 0.05), land cover (e.g., percent forest, r = 0.67, P < 0.01), and percent drilling pads in a watershed (r = -0.61 to -0.71, P < 0.01), multivariate analyses indicated that these variables did not account for any added variability in these benthic indicators; also, composite measures of land use plus combined shale gas development (including both pads and pipelines) only explained 40% of the variability in annual average values of EPT Richness – vs. 66% for percent of pipeline in watersheds. Pipelines cross streams and occur in riparian zones and cross mountain ridges susceptible to runoff and erosion. Two major flood events in our region appear to have exacerbated pipeline effects during our study year.