GSA Annual Meeting in Denver, Colorado, USA - 2016

Paper No. 299-10
Presentation Time: 3:50 PM


BURGESS, Crystal S.1, LAUTZ, Laura K.2, CHIEN, Nathaniel Patrick2, HOKE, Gregory D.2, LEONTE, Mihai3, KESSLER, J.D.3, CHRISTIAN, Kayla2, SIEGEL, Donald I.2 and LU, Zunli2, (1)Department of Geology and Environmental, Alfred University, 1 Saxon Drive, Alfred, NY 14802; Research Experiences in Solid Earth Sciences for Students, UNAVCO, 6350 Nautilus Drive, Boulder, CO 80301, (2)Department of Earth Sciences*, Syracuse University, 204 Heroy Geology Laboratory, Syracuse, NY 13244, (3)Department of Earth and Environmental Sciences, University of Rochester, Rochester, NY 14627,

Naturally occurring methane is commonly observed in shallow groundwater aquifers that have three major characteristics: Na-rich water chemistry, topographic lows such as valleys, and/or aquifer confinement. Under the Shale-Water Interaction Forensic Tools project (Project SWIFT) at Syracuse University, we analyzed the correlation between these three aquifer characteristics and methane concentrations in domestic drinking water wells throughout the Southern Tier of New York. The Southern Tier of New York overlies the gas-bearing Marcellus Shale, and lies just north of regions of Pennsylvania where shale gas extraction from the Marcellus is common. We developed a classification model to evaluate the probability of methane occurrence in different classes of wells using the data from the initial year of this study (2013) and similar studies in New York and Pennsylvania prior to hydraulic fracturing. Based on the model, groundwater samples characterized as Na+dominant and sourced from bedrock aquifers and/or valley settings had a >50% probability of naturally occurring methane levels above 1 mg/L.

In 2014 and again in 2016 we resampled 26 wells with among the highest methane concentrations (>0.01 mg/L) in the original sample of 137 wells. Using the 2013, 2014, and 2016 methane concentrations, we developed a time series of methane levels as a baseline record prior to unconventional drilling in New York. Methane concentrations between 2013, 2014 and 2016 were all positively correlated with correlation coefficients > 0.8. In all three years, >10% of observed methane concentrations (in the subsample of 26) were above warning levels (>10 mg/L).

The classification system and time series can be used to predict natural methane occurrence in New York state as well as regions with similar hydrogeologic settings, such as Pennsylvania, where unconventional natural gas extraction is common and expanding. Understanding the natural conditions favorable to methane occurrence will help identify the differences between methane migration due to unconventional drilling versus natural occurrence.