This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.
Article pubs.acs.org/est
Integrating Source Apportionment Tracers into a Bottom-up Inventory of Methane Emissions in the Barnett Shale Hydraulic Fracturing Region Amy Townsend-Small,*,† Josette E. Marrero,‡ David R. Lyon,§ Isobel J. Simpson,‡ Simone Meinardi,‡ and Donald R. Blake‡ †
University of Cincinnati, Departments of Geology and Geography, Cincinnati, Ohio 45221, United States University of California, Irvine, Department of Chemistry, Irvine, California 92697, United States § Environmental Defense Fund, 301 Congress Ave., Suite 1300, Austin, Texas 78701, United States ‡
S Supporting Information *
ABSTRACT: A growing dependence on natural gas for energy may exacerbate emissions of the greenhouse gas methane (CH4). Identifying fingerprints of these emissions is critical to our understanding of potential impacts. Here, we compare stable isotopic and alkane ratio tracers of natural gas, agricultural, and urban CH4 sources in the Barnett Shale hydraulic fracturing region near Fort Worth, Texas. Thermogenic and biogenic sources were compositionally distinct, and emissions from oil wells were enriched in alkanes and isotopically depleted relative to natural gas wells. Emissions from natural gas production varied in δ13C and alkane ratio composition, with δDCH4 representing the most consistent tracer of natural gas sources. We integrated our data into a bottom-up inventory of CH4 for the region, resulting in an inventory of ethane (C2H6) sources for comparison to top-down estimates of CH4 and C2H6 emissions. Methane emissions in the Barnett are a complex mixture of urban, agricultural, and fossil fuel sources, which makes source apportionment challenging. For example, spatial heterogeneity in gas composition and high C2H6/CH4 ratios in emissions from conventional oil production add uncertainty to top-down models of source apportionment. Future top-down studies may benefit from the addition of δD-CH4 to distinguish thermogenic and biogenic sources.
■
terization of stable isotopic7,18 and alkane ratio19,20 source signatures can help constrain CH4 emissions from oil and gas sources,21 particularly in conjunction with top-down measurements of regional emission rates.22−24 Here, we present measurements of δ13C, δD, and C2−C5 alkane ratios for CH4 sources in the Barnett Shale natural gas producing region in northeast Texas. This region has tens of thousands of hydraulically fractured shale gas wells that can be sources of atmospheric CH4.25 Produced gas (78−97% CH4) is transported by a system of pipelines and gathering compressor stations to processing plants, which remove heavier hydrocarbons to create pipeline quality gas (∼95% CH4). This gas is transported through transmission pipelines and compressor stations to end users, including local industrial, commercial, and residential consumers. Older conventional oil wells in the region can also contribute CH4 to the atmosphere.21 Finally, the Fort Worth−Barnett region is part of the United States’ fourth largest metropolitan area, and there are urban CH4
INTRODUCTION Methane (CH4) is a greenhouse gas with a global warming potential 34 to 86 times greater than carbon dioxide (CO2) on time scales of 100 and 20 years, respectively.1 Production, processing, transmission, and distribution of natural gas, which is comprised primarily of CH4, are, along with agriculture and landfills, among the largest anthropogenic sources of CH4 globally,2−4 nationally,5,6 and regionally.7,8 Natural gas combustion produces less CO2 per unit of energy than coal or petroleum products; it has abundant domestic reserves, and its combustion releases less sulfur, nitrogen, and mercury than coal.9,10 However, among other potential negative impacts, a larger reliance on natural gas for electricity generation and transportation may increase CH4 emissions, potentially overwhelming the climate benefit of natural gas.9,11 Measurements of fugitive emissions of CH4 from along the natural gas supply chain are a critical first step in minimizing this problem.12−14 Because of the large number of anthropogenic and natural sources of CH4, it can be difficult to assess the relative contribution of natural gas sources to CH4 emissions. Uncertainties in activity factors and emissions factors can lead to underestimation of CH4 emission rates using bottom-up approaches regionally,15,16 nationally,17 and globally.11 Charac© 2015 American Chemical Society
Received: Revised: Accepted: Published: 8175
January 6, 2015 April 8, 2015 May 28, 2015 July 7, 2015 DOI: 10.1021/acs.est.5b00057 Environ. Sci. Technol. 2015, 49, 8175−8182
Article
Environmental Science & Technology sources including landfills, wastewater treatment plants, and power plants, as well as agricultural CH4 sources including cattle ranches and feedlots. Our overall goals were (1) to characterize the isotopic and alkane ratios of CH4 sources in the Barnett−Fort Worth region and (2) to integrate these data into a bottom-up CH4 inventory for the region for comparison to top-down studies.
Subsamples of each canister were transferred to evacuated 12 mL glass vials (Exetainers, Labco Ltd., Buckinghamshire, UK) for analysis of δ13C and δD of CH4 via isotope ratio mass spectrometry at the University of Cincinnati.30 The instrument was calibrated with CH4 standards matched to sample concentrations to eliminate any possible linearity problems. The reproducibility of δ13C and δD measurements using this method is 0.2‰ and 4‰, respectively,30 and daily analysis of multiple replicates of CH4 standards met or exceeded this. Data Analysis. As gas concentrations for each sample type varied significantly, we used Keeling plots, where the isotopic ratio is regressed against 1/[CH4] for each sample type and the y-intercept is the composition of excess CH4 in the data set (Figures 2 and 3; Table 1).31−33 Concentrations of C2H6,
■
MATERIALS AND METHODS Sample Collection. Whole air samples (n = 119) were collected in 2 L stainless steel canisters in October 2013 near Fort Worth, Texas during the Barnett Coordinated Campaign.26 Most samples were collected downwind of CH4 sources, including cattle feedlots, landfills, natural gas wells, and conventional oil wells (Figure 1). Five out of 119 samples
Figure 1. Map of sampling locations in the Barnett Shale−Fort Worth region. ng = natural gas. Dotted lines are the “core” Barnett Shale counties.
were collected directly at gas distribution and compressor station sources. We took 25 samples of background air upwind of the region. Location, time, type, and composition of all samples can be found in the Supporting Information. Prior to sampling, canisters were heated to 150 °C and then evacuated to 10−2 Torr.27 Detailed canister preparation procedures have been described previously.27 Sampling was guided by a Picarro Instruments G2301 CO2/CH4/H2O analyzer powered by the vehicle alternator.28 This instrument was used to detect CH4 enhancement only: canisters were not filled from the instrument exhaust or inlet. Canisters for source analysis were filled where CH4 was elevated by at least 50 ppb, the minimum enhancement needed to detect isotopic source signatures.7 All samples were collected at ambient pressure by opening the canister valve for 60 s above and upwind of the sampling technician’s head. Sample Analysis. The whole-air samples were analyzed at the University of California, Irvine via flame ionization detector (FID) for concentrations of CH4, ethane (C2H6; or C2), propane (C3H8; C3), n-butane (n-C4H10; C4), and n-pentane (n-C5H12; C5).29 The CH4 precision was 0.1% with an accuracy of 1%. The detection limit for C2−C5 alkanes was 3 parts per trillion by volume (pptv). The analytical precision ranged from 1% to 3%, and the accuracy was 5% (% relative standard deviation). Standards are traceable to NIST and subject to frequent intercalibration.27
Figure 2. Keeling plots of (a) δ13C-CH4 and (b) δD-CH4 of biological methane sources sampled in the Barnett Shale region.
C3H8, n-C4H10, and n-C5H12 were plotted against the concentration of CH4 for each source type, such that the slope gives the average ratio of each alkane to CH4 (Figures 3 and 4; Table 2).19 Isotope and alkane ratios for natural gas production sites were calculated by correcting for background air because our samples were taken downwind of production sites (Tables 3 and 4). For δ13C and δD of CH4 from individual samples, this correction was done according to ref 7 using the background air composition described below. For alkane ratios, we subtracted the background concentration of each alkane from the concentration measured in each source sample and then calculated the ratio of each C2−C5 alkane to CH4. Both 8176
DOI: 10.1021/acs.est.5b00057 Environ. Sci. Technol. 2015, 49, 8175−8182
Article
Environmental Science & Technology
Figure 3. Composition of methane from natural gas sources in the Barnett Shale region. (a) Keeling plot of δ13C-CH4 vs 1/[CH4]; (b) δD-CH4 vs 1/[CH4]; (c) [C2H6] vs [CH4]; (d) [C3H8] vs [CH4]; (e) [n-C4H10] vs [CH4]; and (f) [n-C5H12] vs [CH4].
Table 1. Stable Isotopic Endmembers for CH4 Sources in the Barnett Shale Regiona δD
a
δ13C
sample type
n
source endmember (‰)
p
n
source endmember (‰)
p
cattle landfills natural gas well pads transmission compressor stations distribution systems conventional oil wells
7 8 35 11 5 7
−283 −260 −152 −145 −133 −170
Article pubs.acs.org/est
Integrating Source Apportionment Tracers into a Bottom-up Inventory of Methane Emissions in the Barnett Shale Hydraulic Fracturing Region Amy Townsend-Small,*,† Josette E. Marrero,‡ David R. Lyon,§ Isobel J. Simpson,‡ Simone Meinardi,‡ and Donald R. Blake‡ †
University of Cincinnati, Departments of Geology and Geography, Cincinnati, Ohio 45221, United States University of California, Irvine, Department of Chemistry, Irvine, California 92697, United States § Environmental Defense Fund, 301 Congress Ave., Suite 1300, Austin, Texas 78701, United States ‡
S Supporting Information *
ABSTRACT: A growing dependence on natural gas for energy may exacerbate emissions of the greenhouse gas methane (CH4). Identifying fingerprints of these emissions is critical to our understanding of potential impacts. Here, we compare stable isotopic and alkane ratio tracers of natural gas, agricultural, and urban CH4 sources in the Barnett Shale hydraulic fracturing region near Fort Worth, Texas. Thermogenic and biogenic sources were compositionally distinct, and emissions from oil wells were enriched in alkanes and isotopically depleted relative to natural gas wells. Emissions from natural gas production varied in δ13C and alkane ratio composition, with δDCH4 representing the most consistent tracer of natural gas sources. We integrated our data into a bottom-up inventory of CH4 for the region, resulting in an inventory of ethane (C2H6) sources for comparison to top-down estimates of CH4 and C2H6 emissions. Methane emissions in the Barnett are a complex mixture of urban, agricultural, and fossil fuel sources, which makes source apportionment challenging. For example, spatial heterogeneity in gas composition and high C2H6/CH4 ratios in emissions from conventional oil production add uncertainty to top-down models of source apportionment. Future top-down studies may benefit from the addition of δD-CH4 to distinguish thermogenic and biogenic sources.
■
terization of stable isotopic7,18 and alkane ratio19,20 source signatures can help constrain CH4 emissions from oil and gas sources,21 particularly in conjunction with top-down measurements of regional emission rates.22−24 Here, we present measurements of δ13C, δD, and C2−C5 alkane ratios for CH4 sources in the Barnett Shale natural gas producing region in northeast Texas. This region has tens of thousands of hydraulically fractured shale gas wells that can be sources of atmospheric CH4.25 Produced gas (78−97% CH4) is transported by a system of pipelines and gathering compressor stations to processing plants, which remove heavier hydrocarbons to create pipeline quality gas (∼95% CH4). This gas is transported through transmission pipelines and compressor stations to end users, including local industrial, commercial, and residential consumers. Older conventional oil wells in the region can also contribute CH4 to the atmosphere.21 Finally, the Fort Worth−Barnett region is part of the United States’ fourth largest metropolitan area, and there are urban CH4
INTRODUCTION Methane (CH4) is a greenhouse gas with a global warming potential 34 to 86 times greater than carbon dioxide (CO2) on time scales of 100 and 20 years, respectively.1 Production, processing, transmission, and distribution of natural gas, which is comprised primarily of CH4, are, along with agriculture and landfills, among the largest anthropogenic sources of CH4 globally,2−4 nationally,5,6 and regionally.7,8 Natural gas combustion produces less CO2 per unit of energy than coal or petroleum products; it has abundant domestic reserves, and its combustion releases less sulfur, nitrogen, and mercury than coal.9,10 However, among other potential negative impacts, a larger reliance on natural gas for electricity generation and transportation may increase CH4 emissions, potentially overwhelming the climate benefit of natural gas.9,11 Measurements of fugitive emissions of CH4 from along the natural gas supply chain are a critical first step in minimizing this problem.12−14 Because of the large number of anthropogenic and natural sources of CH4, it can be difficult to assess the relative contribution of natural gas sources to CH4 emissions. Uncertainties in activity factors and emissions factors can lead to underestimation of CH4 emission rates using bottom-up approaches regionally,15,16 nationally,17 and globally.11 Charac© 2015 American Chemical Society
Received: Revised: Accepted: Published: 8175
January 6, 2015 April 8, 2015 May 28, 2015 July 7, 2015 DOI: 10.1021/acs.est.5b00057 Environ. Sci. Technol. 2015, 49, 8175−8182
Article
Environmental Science & Technology sources including landfills, wastewater treatment plants, and power plants, as well as agricultural CH4 sources including cattle ranches and feedlots. Our overall goals were (1) to characterize the isotopic and alkane ratios of CH4 sources in the Barnett−Fort Worth region and (2) to integrate these data into a bottom-up CH4 inventory for the region for comparison to top-down studies.
Subsamples of each canister were transferred to evacuated 12 mL glass vials (Exetainers, Labco Ltd., Buckinghamshire, UK) for analysis of δ13C and δD of CH4 via isotope ratio mass spectrometry at the University of Cincinnati.30 The instrument was calibrated with CH4 standards matched to sample concentrations to eliminate any possible linearity problems. The reproducibility of δ13C and δD measurements using this method is 0.2‰ and 4‰, respectively,30 and daily analysis of multiple replicates of CH4 standards met or exceeded this. Data Analysis. As gas concentrations for each sample type varied significantly, we used Keeling plots, where the isotopic ratio is regressed against 1/[CH4] for each sample type and the y-intercept is the composition of excess CH4 in the data set (Figures 2 and 3; Table 1).31−33 Concentrations of C2H6,
■
MATERIALS AND METHODS Sample Collection. Whole air samples (n = 119) were collected in 2 L stainless steel canisters in October 2013 near Fort Worth, Texas during the Barnett Coordinated Campaign.26 Most samples were collected downwind of CH4 sources, including cattle feedlots, landfills, natural gas wells, and conventional oil wells (Figure 1). Five out of 119 samples
Figure 1. Map of sampling locations in the Barnett Shale−Fort Worth region. ng = natural gas. Dotted lines are the “core” Barnett Shale counties.
were collected directly at gas distribution and compressor station sources. We took 25 samples of background air upwind of the region. Location, time, type, and composition of all samples can be found in the Supporting Information. Prior to sampling, canisters were heated to 150 °C and then evacuated to 10−2 Torr.27 Detailed canister preparation procedures have been described previously.27 Sampling was guided by a Picarro Instruments G2301 CO2/CH4/H2O analyzer powered by the vehicle alternator.28 This instrument was used to detect CH4 enhancement only: canisters were not filled from the instrument exhaust or inlet. Canisters for source analysis were filled where CH4 was elevated by at least 50 ppb, the minimum enhancement needed to detect isotopic source signatures.7 All samples were collected at ambient pressure by opening the canister valve for 60 s above and upwind of the sampling technician’s head. Sample Analysis. The whole-air samples were analyzed at the University of California, Irvine via flame ionization detector (FID) for concentrations of CH4, ethane (C2H6; or C2), propane (C3H8; C3), n-butane (n-C4H10; C4), and n-pentane (n-C5H12; C5).29 The CH4 precision was 0.1% with an accuracy of 1%. The detection limit for C2−C5 alkanes was 3 parts per trillion by volume (pptv). The analytical precision ranged from 1% to 3%, and the accuracy was 5% (% relative standard deviation). Standards are traceable to NIST and subject to frequent intercalibration.27
Figure 2. Keeling plots of (a) δ13C-CH4 and (b) δD-CH4 of biological methane sources sampled in the Barnett Shale region.
C3H8, n-C4H10, and n-C5H12 were plotted against the concentration of CH4 for each source type, such that the slope gives the average ratio of each alkane to CH4 (Figures 3 and 4; Table 2).19 Isotope and alkane ratios for natural gas production sites were calculated by correcting for background air because our samples were taken downwind of production sites (Tables 3 and 4). For δ13C and δD of CH4 from individual samples, this correction was done according to ref 7 using the background air composition described below. For alkane ratios, we subtracted the background concentration of each alkane from the concentration measured in each source sample and then calculated the ratio of each C2−C5 alkane to CH4. Both 8176
DOI: 10.1021/acs.est.5b00057 Environ. Sci. Technol. 2015, 49, 8175−8182
Article
Environmental Science & Technology
Figure 3. Composition of methane from natural gas sources in the Barnett Shale region. (a) Keeling plot of δ13C-CH4 vs 1/[CH4]; (b) δD-CH4 vs 1/[CH4]; (c) [C2H6] vs [CH4]; (d) [C3H8] vs [CH4]; (e) [n-C4H10] vs [CH4]; and (f) [n-C5H12] vs [CH4].
Table 1. Stable Isotopic Endmembers for CH4 Sources in the Barnett Shale Regiona δD
a
δ13C
sample type
n
source endmember (‰)
p
n
source endmember (‰)
p
cattle landfills natural gas well pads transmission compressor stations distribution systems conventional oil wells
7 8 35 11 5 7
−283 −260 −152 −145 −133 −170
