FEASIBILITY OF USING SPECIFIC CONDUCTANCE FOR MONITORING STREAMS FOR MARCELLUS SHALE HYDROFRACTURING WATERS

Hydrofracturing for exploration and production of natural gas from the Marcellus Shale in PA, OH, WV and NY produces large volumes of very saline Na/Ca/Cl dominated “flowback” water. Agencies and citizens groups have proposed monitoring potential receiving streams using specific conductance (SC).Such monitoring would be most valuable if “red flags” were noticeable in real-time, without requiring follow-up sampling and analysis that would delay remedial action. We used GIS to map locations of the 210 US Geological Survey (USGS) gaging stations in Pennsylvania by county and by hydrologic unit along with locations of Marcellus well permits and major streams. In the most active region for Marcellus drilling, 8 of 99 gaging stations measure SC. Two years of hydrographs from 9 (7 in the Marcellus region) USGS gaging stations that collect SC data showed gaps in SC data between 4 and 71% of the time for individual stations. Plots of SC versus discharge for several time periods show considerable and variable hysteresis effects dependent upon the stream chosen and antecedent hydrologic conditions, suggesting that using SC monitoring for Marcellus “leaks” may be more difficult than simple visual inspection of SC/discharge/time data.

Hydrographs can be affected by some combination of precipitation, runoff, interflow, groundwater, changes in acid mine drainage (AMD) discharges, and potential Marcellus “leaks.” We developed a mixing model to examine the impact of different hydrologic conditions and different Marcellus flowback “leak” scenarios. Abnormally high SC stream events could easily be discerned during near-base flow conditions in a small stream. However, such an event cannot be distinguished from a road salt “leak” or a change in an AMD discharge without additional chemical data (especially SO4 data for AMD) that would have to be collected after a potential “leak” was flagged. In contrast, an abnormally high specific conductance stream event could easily be missed during a storm event. Such events could also be missed in larger or wide shallow streams, due to incomplete mixing. Cl-selective electrodes might be more appropriate for monitoring for Marcellus “leaks,” but most ion-selective electrodes are not recommended for long-term deployment by manufacturers.