Database We compiled a global inventory of natural gas molecular and isotopic measurements from the peer-reviewed literature and government reports (Fig. 1). Samples were categorized into conventional oil&gas, coal and shale gas. Recorded parameters include concentrations of C1-C6 alkanes and permanent gases and 𝛿13C and 𝛿D of C1-C6 alkanes.

Global Inventory of Natural Gas Molecular & Isotopic CompositionsOwen Sherwood1, Stefan Schwietzke2, Victoria Arling2, Guiseppe Etiope3, John Miller2

1INSTAAR,  University  of  Colorado,  Boulder,  Colorado  ,  2NOAA  Earth  System  Research  Laboratory,  Global  Monitoring  Division,  Boulder,  Colorado  3Istituto  Nazionale  di  Geofisica  e  Vulcanologia,  Rome,  Italy

Introduction Top-down models of the global atmospheric methane budget use isotopic and/or molecular data to constrain source-specific emissions. These models are sensitive to end-member signatures (𝛿13CCH4, 𝛿DCH4, ethane:methane ratios) for the three main source categories, microbial methanogenesis, biomass burning, and fossil fuels. However, the end-member values are poorly constrained and based on data of unknown or limited sample count, regional extent and global representation. For fossil fuels in particular, few modeling studies reference primary data, despite a vast literature in petroleum geology reporting on the isotopic and molecular composition of natural gas.

References Bousquet et al. (2006) Nature 443:439-443; BP (2014) Statistical Review of World Energy. http://www.bp.com/en/global/corporate/energy-economics/statistical-review-of-world-energy.html; Gupta et al. (1996) J Geophys Res 101:22923-11932; Kirschke et al. (2013) Nat Geosci 6:813-823; Lassey et al. (2007) Atmos Chem Phys 7:2141-2149; Mikaloff et al. (2004) Global Biogeochem Cycles 18:GB4004; Milkov (2011) Org Geochem 42:184-207; Monteil et al (2011) Atmos Chem Phy 11:9141-9153; Neef et al. (2010) Global Biogeochem Cycles 24:GB4024; Quay et al. (1999) Global Biogeochem Cycles 13:445-461; Rice (1983) AAPG Studies in Geology 38:159-184; Rice and Claypool (1983) AAPG Bull 65:5-25; Schwietzke et al. Revised global methane budget and natural gas leakage trends based on extensive isotopic source signatures Nature (submitted); Whiticar (1989) Org Geochem 16:531-547.

Reference Natural Gas 𝜹13CCH4 (‰) Coal 𝜹13CCH4 (‰)

Whiticar 1989 -44 -37

Gupta 1996 -38 -37

Quay 1999 -43 ± 7 -36 ± 7

Mikaloff 2004 -44 -37

Bousquet 2006 -44 -37

Lassey 2007 -40 -35

Neef 2010 -40 -35

Monteil 2011 -40 -35

Kirschke 2013 -55 to -25 N/A

Literature Average: -41.3 ± 2.7 -36.2 ± 0.1

?

?

Conventional

1001−5000501−1000101−50051−10011−501−10

556

2042

402

855

400

114

9236 48

91

24

11

24

27

8

28

31

32

41

176 57

28

28

Coal

1001−5000501−1000101−50051−10011−501−10

722

92 2

195

110

8

5

16

10598

Shale

1001−5000501−1000101−50051−10011−501−10

558

88

Conventional Gas

Coal

Shale Gas

0"5"10"15"20"25"

US"Russian"Federa2on"

Canada"Iran"

Norway"Qatar"Algeria"

Saudi"Arabia"Indonesia"

China"United"Kingdom"

Netherlands"Malaysia"

Uzbekistan"Turkmenistan"

Mexico"United"Arab"Emirates"

Egypt"Australia"Argen2na"Pakistan"

Trinidad"&"Tobago"India"

Nigeria"Venezuela"Thailand"

Oman"Ukraine"

Bangladesh"Other"Asia"Pacific"

Kazakhstan"Germany"

Brazil"Other"Africa"

Kuwait"Bolivia"Brunei"

Romania"Bahrain"

Myanmar"Libya"

Azerbaijan"Italy"

Other"Europe"&"Eurasia"Colombia"Denmark"Vietnam"

Syria"Peru"

Poland"Other"S."&"Cent."America"

Yemen"Other"Middle"East"

Iraq"

0" 5" 10" 15" 20" 25" 30" 35" 40"

US"Russian"Federa4on"

Canada"Iran"

Norway"Qatar"Algeria"

Saudi"Arabia"Indonesia"

China"United"Kingdom"

Netherlands"Malaysia"

Uzbekistan"Turkmenistan"

Mexico"United"Arab"Emirates"

Egypt"Australia"Argen4na"Pakistan"

Trinidad"&"Tobago"India"

Nigeria"Venezuela"Thailand"

Oman"Ukraine"

Bangladesh"Other"Asia"Pacific"

Kazakhstan"Germany"

Brazil"Other"Africa"

Kuwait"Bolivia"Brunei"

Romania"Bahrain"

Myanmar"Libya"

Azerbaijan"Italy"

Other"Europe"&"Eurasia"Colombia"Denmark"Vietnam"

Syria"Peru"

Poland"Other"S."&"Cent."America"

Yemen"Other"Middle"East"

Iraq"

0"10"20"30"40"50"

China"US"

India"Australia"

Russian"Federa:on"South"Africa"Indonesia"Germany"Poland"

Kazakhstan"Other"Europe"&"Eurasia"

Ukraine"Canada"

Colombia"Turkey"Greece"

Czech"Republic"Other"Asia"Pacific"

Vietnam"Romania"Bulgaria"

United"Kingdom"Thailand"

Spain"Mexico"

Hungary"Brazil"

Venezuela"New"Zealand"

Pakistan"Uzbekistan"Zimbabwe"

South"Korea"Other"Africa"

Japan"France"

Other"S."&"Cent."America"

0" 10" 20" 30" 40" 50" 60"

China"US"India"Australia"Russian"Federa;on"South"Africa"Indonesia"Germany"Poland"Kazakhstan"Other"Europe"&"Eurasia"Ukraine"Canada"Colombia"Turkey"Greece"Czech"Republic"Other"Asia"Pacific"Vietnam"Romania"Bulgaria"United"Kingdom"Thailand"Spain"Mexico"Hungary"Brazil"Venezuela"New"Zealand"Pakistan"Uzbekistan"Zimbabwe"South"Korea"Other"Africa"Japan"France"Other"S."&"Cent."America"

Conventional Gas% global production

(2000-2014) % sample count % global production (2000-2014) % sample count

Coal

Global Representation The inventory contains data from 45 countries, 179 basins, >597 geological formations, and 8734 unique samples. On a country-level basis, the data represent 82.5% of world natural gas production and 80.2 % of world coal production over the period 2000-2014 (BP 2014). Note over-representation of some countries (e.g. USA, China, Australia) and under-representation of others (e.g. Norway, Qatar, Algeria, Saudi Arabia, Turkmenistan).

Caveats • Focus was on publications having isotope data; much more gas

concentration data exist in the literature/reports. • No way to verify isotopic calibration/standardization (esp. Soviet era

papers).

Fig. 2: Comparison of country-specific production (as percent of global production) vs. sample count for conventional gas (left) and coal (right).

Fig. 3: Distributions of 𝛿13CCH4, 𝛿DCH4 and C2:C1 ratios for conventional gas, coal, and shale gas. Note skewed conventional gas distributions and bimodal coal distributions.

mean: -44.0median: -42.2

mean: -49.5median: -49.8

mean: -42.5median: -41.1

mean: -194median: -186

mean: -232median: -215

mean: -167median: -145

mean: .074median: .045

mean: .032median: .004

mean: .048median: .016

𝛿13CCH4 𝛿DCH4 C2:C1 ratio

Coa

lSh

ale

Gas

Con

vent

iona

l Gas

Implications for Top-Down Models Data distributions are shown in Fig. 3. Note significantly lower mean and median 𝛿13CCH4 values for conventional gas and especially coal, than typically used in top-down models (Table 1). This is due to the importance of 1) primary and secondary microbial methanogenesis in approximately 20% of global oil and gas reservoirs (Rice & Claypool 1981; Milkov 2011) and in coal (Rice 1993) and 2) isotopically light signatures in low-maturity, oil-associated gas (Fig. 4). For example, giant Cenomanian gas pools of western Siberia, which account for 17% of global gas production, have mean 𝛿13CCH4 = -51 ‰. Revision of 𝛿13CCH4 end-member signatures has a significant effect on modeled fossil fuel contributions to the global methane budget (Schwietzke et al. submitted).

Fig. 4: Plot of 𝛿13CCH4 vs. 𝛿DCH4 for conventional gas, coal, and shale gas, with density distributions and genetic domains. Microbial methanogenesis (via fermentation and microbial CO2 pathways) in conventional gas reservoirs and coal accounts for negatively-skewed 𝛿13CCH4 distributions.

Table 1: Natural gas and coal 𝛿13CCH4 source signatures used in some top-down models of the global methane budget. Citation pathways indicated by arrows. Only one study (Whiticar 1989) is based on a large empirical dataset; however, the data were proprietary and therefore of unknown global or regional representation.

Fig. 1: Maps showing country-level sample counts for conventional natural gas, coal and shale gas data compiled from the literature and government reports. Shale gas data exist for USA and Canada only.

Conclusions • Most complete database of natural gas compositions ever compiled.• Includes isotopic (𝛿13C & 𝛿D) and molecular compositions.• Data represent ~80% of global natural gas and coal production.• At global level, 𝛿13C and 𝛿D values are skewed to lower values because

of microbial methanogenesis and low thermal maturity oil-associated gases.

• Previous inversion studies may have underestimated the fossil fuel flux due to use of end-member values that were too high.

• Database will be published as a standalone manuscript and is available by download through:

Sherwood, O., Schwietzke, S., Arling, V. & Etiope, G. Global Inventory of Fossil and Non-fossil Methane δ13C Source Signature Measurements for Improved Atmospheric Modeling. (2016). doi:http://ezid.cdlib.org/id/doi:10.15138/G37P4D.

𝛿13 C

CH

4

𝛿DCH4

Atmosphere