STREAM FLOW AND FRAC-WATER WITHDRAWALS IN PENNSYLVANIA

High volume slick-water hydraulic fracturing (HVHF) is used to extract gas and oil from low-permeability geologic formations. HVHF involves injecting a frac fluid mixture, primarily composed of a large volume of freshwater, into wells drilled kilometers beneath the surface to fracture and increase the permeability of hydrocarbon-bearing formations. Following HVHF, water that returns to the surface is disposed of via underground injection, treated and reused for HVHF, or treated and released into the environment. Despite increases in recycling rates in some shale plays, freshwater consumption for HVHF remains a concern, especially in regions where the majority of injected freshwater remains deep beneath the ground, and is effectively removed from the hydrologic cycle. In the Marcellus Shale region, streams and rivers serve as the primary sources for freshwater used in HVHF, and over half of frac fluid remains underground following well stimulation. Frac-water withdrawals in Pennsylvania occur predominantly from streams in ungaged basins or at significant distances from USGS stream-gaging stations. Consequently, the effects of freshwater appropriation for HVHF on stream flow regimes remain poorly understood. To address this issue, we compared permitted stream-water withdrawals for HVHF to estimates of stream flow indices at 350 permitted frac-water withdrawal sites. We began our analysis by using USGS stream-flow data to estimate discharge exceedance curves for 150 gaged watersheds in Pennylvania’s Susquehanna and Ohio River basins. The discharge-exceedance probabilities derived from these curves were quantified in terms of watershed properties through multiple-linear regression, whereupon the regression equations were used to predict discharge-exceedance probabilities for ungaged stream sites where frac-water withdrawals occur. Our findings suggest that, in the majority of cases, frac-water withdrawals represent a small fraction of stream discharge, even during precipitation-free periods of low stream flow. But, for a small number of first- and second-order streams, maximum permitted withdrawals are within an order of magnitude of stream-baseflow discharges, which suggests that frac-water withdrawals from this subset of streams should be avoided or more carefully monitored.