A HYDROCHEMICAL FACIES MODEL FOR DISSECTED, COAL-BEARING ROCKS OVERLYING THE MARCELLUS SHALE: IMPLICATIONS FOR MAKING SENSIBLE INTERPRETATIONS OF BACKGROUND WATER CHEMISTRY RELATED TO OIL AND GAS PRODUCTION

Approximately 60 percent of the area that defines the Marcellus and Devonian shale basin in the Appalachian Plateau is capped by Pennsylvanian coal-bearing rocks. By and large, these rocks provide the majority of groundwater to water wells in this region. For example, the median depth for water wells in the Pennsylvanian Pottsville group is 100 feet, while the entire Pennsylvanian section in Pennsylvanian exceeds 1,000 feet in many areas. Understanding the hydrogeology and groundwater geochemistry is important for making sensible interpretations of background chemistry or potentially impacted groundwater from gas production from the underlying shales. A hydrochemical faces model developed for the dissected, coal-bearing rocks in the Appalachian plateau provides an understanding of the evolution of major ion geochemistry, and for classification of water types and minor elements of interest, such as barium and fluoride. Groundwater from coal seems contains a low pH, and is predominantly a Ca-Mg-HCO3 water type. Groundwater from fractured rock along hill slopes and stress-relief zones varied between Ca-HCO3 and Mg-SO4 water types. Groundwater found deep in the interior of ridges or in minimally dissected uplands is typically old groundwater with a Na-HCO3 water type, and often has a high pH (> 8.0) and high fluoride concentrations. Anomalously high barium concentrations were associated with Na-Cl groundwater common in discharge areas in valley bottoms of third order or higher streams. SIRA analysis of sulfur species indicate that sulfate reduction along with cation exchange controls the occurrence of barium in groundwater. Differential head measurements, packer tests, and tritium age dating were used to define groundwater flow paths. Reaction path modeling was used to identify a plausible set of water-rock interactions that were consistent with the observed chemical evolution of groundwater in the region. Groundwater data from several additional sites in Kentucky, Ohio, and Pennsylvania is consistent with these observations, leading to the development of a regional hydrochemical model that should be useful for predicting groundwater types likely to be encountered as a function of well depth, geology, physiographic location, and a well’s vertical and spatial relationship to local, major drainage valleys.