Identification and characterization of high methane-emitting abandoned oil and gas wells

Abstract

Recent measurements of methane emissions from abandoned oil/gas wells show that these wells can be a substantial source of methane to the atmosphere, particularly from a small proportion of high-emitting wells. However, identifying high emitters remains a challenge. We couple 163 well measurements of methane flow rates; ethane, propane, and n-butane concentrations; isotopes of methane; and noble gas concentrations from 88 wells in Pennsylvania with synthesized data from historical documents, field investigations, and state databases. Using our databases, we (i) improve estimates of the number of abandoned wells in Pennsylvania; (ii) characterize key attributes that accompany high emitters, including depth, type, plugging status, and coal area designation; and (iii) estimate attribute-specific and overall methane emissions from abandoned wells. High emitters are best predicted as unplugged gas wells and plugged/vented gas wells in coal areas and appear to be unrelated to the presence of underground natural gas storage areas or unconventional oil/gas production. Repeat measurements over 2 years show that flow rates of high emitters are sustained through time. Our attribute-based methane emission data and our comprehensive estimate of 470,000-750,000 abandoned wells in Pennsylvania result in estimated state-wide emissions of 0.04-0.07 Mt (10¹² g) CH₄ per year. This estimate represents 5-8% of annual anthropogenic methane emissions in Pennsylvania. Our methodology combining new field measurements with data mining of previously unavailable well attributes and numbers of wells can be used to improve methane emission estimates and prioritize cost-effective mitigation strategies for Pennsylvania and beyond.

Keywords: abandoned wells; climate change; high emitters; methane emissions; oil and gas development.

Fig. 1.

The 88 measured abandoned oil and gas wells in Pennsylvania overlaid with conventional oil and gas pools (34), underground natural gas storage fields (34), and workable coal seams within the study area (38).

Fig. 2.

Methane flow rates of 88 abandoned wells in Pennsylvania and the coefficient of variation of methane flow rates measured from 2 to 10 times over 2 years (July of 2013 to June of 2015) at 27 wells. If more than one measurement has been made at the given well, the methane flow rates represent an average of all measurements taken. Plugging status is determined based on field observations, and the well type (gas vs. oil or combined oil and gas) is determined using our database-based estimates of well attributes. Methane flow rates below detection (BD) limits (P values $>$ 0.2) are shown in the gray portion of the plot between the plots of positive and negative flow rates.

Fig. 3.

Carbon and hydrogen isotopes of methane (${\delta }^{13}$C-CH₄ and ${\delta }^2$H-CH₄), hydrocarbon concentration ratios (C₂₋₄/C₁), noble gas data, and methane flow rate data shown colored by well type, circled by plugging status, and marked with green diamond outlines if in a coal area. For repeat measurements, the average of the data for the well is shown. The regions representing thermogenic methane associated and not associated with oil are from ref. .

Fig. 4.

Number of drilled and/or completed oil and gas wells in Pennsylvania from various historical documents and databases (SI Appendix). The thick black lines represent the 1929–2013 data used to estimate the total number of wells (SI Appendix, Table S4, second column). For 1859–1928, we use a total well number provided in ref. , and the curves shown here are not used to estimate well numbers.

Fig. 5.

Number of wells in the PA DEP database (Left) and the corresponding relative methane emissions distribution (Right) based on plugging status, coal area designation, and well type. Each of three attributes is considered independently.