Emission of Methane and Heavier Alkanes From the La Brea Tar Pits Seepage Area, Los Angeles

Abstract

Natural hydrocarbon (oil and gas) seeps are widespread in Los Angeles, California, due to gas migration, along faults, from numerous subsurface petroleum fields. These seeps may represent important natural contributors of methane (CH₄) and heavier alkanes (C₂-C₄) to the atmosphere, in addition to anthropogenic fossil fuel and biogenic sources. We measured the CH₄ flux by closed-chamber method from the La Brea Tar Pits park (0.1 km²), one of the largest seepage sites in Los Angeles. The gas seepage occurs throughout the park, not only from visible oil-asphalt seeps but also diffusely from the soil, affecting grass physiology. About 500 kg CH₄ d⁻¹ is emitted from the park, especially along a belt of enhanced degassing that corresponds to the 6th Street Fault. Additional emissions are from bubble plumes in the lake within the park (order of 10²–10³ kg d⁻¹) and at the intersection of Wilshire Boulevard and Curson Avenue (>130 kg d⁻¹), along the same fault. The investigated area has the highest natural gas flux measured thus far for any onshore seepage zone in the USA. Gas migration, oil biodegradation, and secondary methanogenesis altered the molecular composition of the original gas accumulated in the Salt Lake Oil Field (>300 m deep), leading to high C₁/C₂₊ and i-butane/n-butane ratios. These molecular alterations can be important tracers of natural seepage and should be considered in the atmospheric modeling of the relative contribution of fossil fuel (anthropogenic fugitive emission and natural geologic sources) versus biogenic sources of methane, on local and global scales.

Key Points

• Considerable methane emission from the La Brea park in Los Angeles
• Molecular alteration during seepage must be considered in atmospheric hydrocarbon emission modeling
• A fault seems to act as a main pathway of gas migration from the Salt Lake oil reservoir to the surface

Plain Language Summary

The hydrocarbon seepage from subsurface petroleum (gas and oil) accumulations is a major natural source of atmospheric methane (CH₄), ethane (C₂H₆), and propane (C₃H₈), which are powerful greenhouse gases and/or photochemical pollutants. We measured the natural gas flux from one of the largest seepage sites in California, the La Brea Tar Pits park in Los Angeles. The gas emission occurs throughout the park, not only from visible oil-asphalt seeps but also diffusely from the soil, affecting grass physiology. The investigated area has the highest natural gas flux measured thus far for any onshore seepage zone in the USA. Gas at the surface has a specific molecular composition, modified during migration from the original composition in the subsurface reservoir, which can help, through specific atmospheric modeling, to distinguish anthropogenic fugitive emission, natural geologic sources, and biogenic sources of methane, on local and global scales.

1 Introduction

The hydrocarbon seepage from subsurface petroleum (gas and oil) accumulations is a major natural source of atmospheric methane (CH₄), ethane (C₂H₆), and propane (C₃H₈), which are powerful greenhouse gases and/or photochemical pollutants (e.g., Ciais et al., 2013; Etiope, 2015; Etiope et al., 2008; Etiope & Klusman, 2002; Nicewonger et al., 2016). Hydrocarbon seepage may also represent a geologic hazard, especially in urban or industrialized areas, due to the explosive properties of the gas, as documented in Los Angeles and some European towns (e.g., Baciu et al., 2008; Chilingar & Endres, 2005; Etiope et al., 2006). Global geologic CH₄ and C₂H₆ emission estimates from gas seepage in sedimentary, petroleum-bearing basins and geothermal areas, based on bottom-up and top-down methods, are in the order of 50–70 Tg y⁻¹ and 2–4 Tg y⁻¹, respectively (Etiope, 2012; Etiope & Ciccioli, 2009; Etiope et al., 2008; Nicewonger et al., 2016; Schwietzke et al., 2016). These values represent roughly half of the global CH₄ and C₂H₆ emissions due to industrial fossil fuel emissions, which are approximately 100–150 Tg CH₄ yr⁻¹ (Schwietzke et al., 2016) and 6–8 Tg C₂H₆ yr⁻¹ (Nicewonger et al., 2016). In addition to the diffuse microseepage (considered as an areal seepage source) that is widespread in sedimentary basins, several thousand macroseeps (gas seeps, oil seeps, and mud volcanoes) are documented on land (CGG, 2015; Clarke & Cleverly, 1991; Etiope, 2015; Link, 1952). Flux measurements of several hundred seeps worldwide allowed a first assessment of the macroseepage emission factors, a fundamental parameter to derive bottom-up global emission estimates (Etiope, 2015). Extending the seepage flux data set is essential to improve the emission factors and consequent global estimates.

In this work we report the results of a gas seepage flux survey carried out in one of the most active hydrocarbon seeps in North America, the La Brea Tar Pits, in Los Angeles. The La Brea Tar Pits, located in an urban park called Hancock Park, are famous for the numerous oil and bitumen seeps, and the rich discoveries of Pleistocene plant and animal fossils (Stock, 1965). But Hancock Park is also characterized by intense degassing, which is visible as bubbling in oil or asphalt seeps and in the Lake Pit, and is suspected (but never investigated) to occur throughout the grassy ground of the park. Gas leaks also occur in streets adjacent to the park, such as at the intersection between Wilshire Boulevard and Curson Avenue. Previous gas studies in the La Brea area addressed, about 25 years ago, the molecular and isotopic compositions of the gas (Jeffrey et al., 1991). More recently, Weber et al. (2017) measured the gas flux from some oil seeps in the park. Farrell et al. (2013) performed a mobile survey with air sampling and gas chromatographic analyses from a recreational vehicle driven around the La Brea neighborhood. They detected up to 48 ppmv of CH₄ in the air, suggesting significant gas seepage throughout the area. CH₄ sources at and around the La Brea park were also reported by Thorpe et al. (2013) using the Airborne Visible/Infrared Imaging Spectrometer.

In May 2017, we performed gas flux measurements using the closed chamber technique throughout the park, with the objective of assessing the total emissions from the ground to the atmosphere of methane, ethane and propane, and the spatial distribution of the seepage. Methane emission was investigated for both oil-asphalt seeps (macroseepage) and the diffuse degassing from the soil (which is called “miniseepage”; the term “microseepage,” not applicable in our case, refers to diffuse exhalations in a sedimentary basin that are not related to a macroseepage zone; Etiope, 2015; Etiope et al., 2011). New molecular composition analyses of the gas have been carried out to assess, by comparison with those of Jeffrey et al. (1991): (a) possible temporal changes in the gas composition; (b) the methane/(ethane + propane) ratio in the seeping gas, which, due to molecular fractionation and biodegradation and secondary methanogenesis during upward migration, may be different from the ratio in the original reservoir gas; and (c) the isobutane and normal-butane concentrations, whose original relative composition can be altered due to biodegradation along the seepage system. Understanding the compositional ratios between several alkanes, and their modification during seepage, can help in modeling and estimating the relative contribution of fugitive (anthropogenic) fossil fuel leakage versus natural gas seepage on local and global scales (e.g., Hopkins et al., 2016; Nicewonger et al., 2016).

2 Description of the Seepage Area

The city of Los Angeles contains numerous natural hydrocarbon seeps due to prolific petroleum fields and active tectonics, where faults and fractured rocks form preferential conduits for gas and oil migration to the surface (e.g., Chang et al., 2014; Chilingar & Endres, 2005; Gurevich et al., 1993). Hancock Park and the associated Page Museum, where the La Brea Tar Pits are located, is about 0.1 km² in area (roughly 0.3 km by 0.3 km; Figure 1), in the Fairfax District. Approximately 100 visible seeps (bitumen and oil leaks) occur throughout the park, although these are temporal in nature (as also observed by Weber et al., 2017), so new seeps are always forming and existing seeps sometimes cease. Major degassing is observable in the Lake Pit, at the southern side of the park, approximately 3,000 m² in area and 5 m deep, where more than a hundred bubble plumes, with different size and bubble burst frequency, were observed during our survey (Figure 2a). Weaker gas bubbling occurs in some oil seeps emerging from the soil in the park (Figures 2b and 2c). Gardens and soil covered areas often show lack of grass, which is a suspected indication of oxygen depletion in the soil (as observed, e.g., in CO₂-rich mofettes; Rennert & Pfanz, 2016) due to enhanced hydrocarbon concentrations (Figure 2d).

Figure 1

Location and geologic setting of the La Brea Tar Pits redrawn from Antandrus (2009) using additional information from Samuelian (1990).

Figure 2

(a) The Lake pit with bubbling plumes. (b) The high degassing zone within the pits 3-4-61-67. (c) A gas bubble in one of the biggest oil seeps. (d) The closed chamber with portable sensors measuring gas flux from soil. Note the variability of the grass color, which reflects the intensity of the methane seepage (see text for details).

The seeping gas is thermogenic, primarily composed of CH₄ (~80–84 vol%), stemming from sand reservoirs (Upper Miocene to Pliocene Repetto and Puente Formations) of the Salt Lake Oil Field, located approximately from 300 to 1,000 m below the surface (Lang & Dreessen, 1975). The compositions of the seeping and reservoir gases are reported in Table 1. From the reservoirs, the gas naturally migrates to the surface likely through the 6th Street Fault, along the southern border of the Salt Lake Oil Field anticline (Wright, 1987) (Figure 1). A precise mapping of this fault was possible thanks to data from more than 500 wells (Lang & Dreessen, 1975; Samuelian, 1990).

Table 1. Molecular Composition (vol %) of the Gas and Isotopic Ratios of CH₄ and CO₂ (‰, C-VPDB; H-VSMOW) Seeping at La Brea Tar Pits and in the Underlying Salt Lake Oil Field Reservoir

Sample	CH4	C2H6	C3H8	C1/(C2 + C3)	i-C4	n-C4	i-C5	n-C5	CO2	N2	δ^13C-CH4	δ^2H-CH4	δ^13C-CO2
Circular sidewalk degassing soil (no. 160) a	68.5	0.1	0.02	523	0.013	0.002	Nd	Nd	11.4
Circular sidewalk degassing soil (no. 160) b (ppmv)	6377	51	0.6	124	1.2	<0.1	1.1	<0.001
Bubbling Oil Seep (no 198) a	75.0	0.4	0.1	150	0.017	0.001	0.004	0.002	12.2
Wilshire-Curson corner a	78.0	1.3	0.5	43	0.02	0.03	0.008	0.006	13.5
Wilshire-Curson corner c	8.93	0.112	0.096	44	0.016	0.022	0.007	0.002
Average oil seep by Weber et al, (2017) b^,d	19.63	0.153	0.067	89	0.021	0.002	0.001	<0.001			−38.9
Sample 1,060 by Jeffrey et al. (1991) e	79.9	0.72	0.18	89	0.023	0.002	0.004	Tr.	15.6	3.5	−42.4	−178	25.9
Sample 1,061 by Jeffrey et al. (1991) e	84.5	0.69	0.17	98	0.021	n.d.	n.d.	n.d.	14.8	Tr.	−42.4	−183	25.4
Salt Lake Reservoir, Puente F. (Jeffrey et al., 1991)	80.5	1.59	0.60	37	0.23	0.38	0.17	0.07	14.4	0.07	−42.5	−181	15.6

• Note. tr., traces; n.d., not detected.
• ^a Accumulated in 10 L chamber after atmospheric air flushing (no. 160 and Wilshire-Curson corner) or through inverted funnel (no. 198) and analyzed by FTIR (not corrected for air contamination);
• ^b Collected in 2 L canister after accumulation in 125 L chamber (without atmospheric air flushing) and analyzed by GC (not corrected for air contamination);
• ^c Collected in 2 L canister and analyzed by GC (not corrected for air contamination);
• ^d Molecular composition represents an average of 102 samples, and isotopic ratio represents an average of 2 samples; and
• ^e Values corrected for air contamination.

Gas also seeps at several points just outside the park. A particularly important case is the intense gas leak at the intersection of Wilshire Boulevard and Curson Avenue (Figure 3), where in 1999 city authorities had to develop a permanent ventilation system. The reason for this leak is not clear. Chilingar and Endres (2005) mentioned the possible presence of an abandoned oil well, but the oil well database of the Department of Conservation's Division of Oil, Gas, and Geothermal Resources of California does not report any well at Wilshire Boulevard and Curson Avenue. More recently, Los Angeles city officials considered the gas leak a natural seepage phenomenon (e.g., Blevins, 2013). Furthermore, additional gas seepage has been observed at other locations above the Salt Lake Oil Field. Most notably, a department store exploded in 1985 due to a buildup of methane seepage at the corner of 3rd Street and Ogden Drive (about 0.5 km north of the tar pits). This event led to revised building codes, which now require methane mitigation systems in neighborhood buildings in order to vent seeping gases (Chilingar & Endres, 2005).

Figure 3

Distribution of the gas flux measurement points in correspondence with oil-asphalt seeps and over the soil.

3 Methodology

3.1 Gas Sampling and Analysis

Molecular composition of the gas exhaling at the La Brea park was analyzed at three sites: (a) from a bubbling oil seep (point no. 198; southeast corner of the park, Figure 3); (b) from the grassy ground inside the circular sidewalk near the central area of the park (point no. 160; corresponding to Figure 2d; location shown in Figure 3); and (c) from the gas leak at the Wilshire-Curson intersection. Gas was sampled using an inverted funnel from bubbling seep no. 198 and from a 10 L closed chamber on the degassing soil and at the Wilshire-Curson leak. Funnel and chamber were flushed by using a three-way valve, tubing, and a 50 mL syringe to reduce air contamination.

The molecular analysis of samples no. 198, no. 160, and the Wilshire-Curson leak was performed using a Fourier Transform Infrared spectrometer (FTIR DX-4030, Gasmet, Finland) with a standard spectra library to rapidly and simultaneously determine 10 gases (CH₄, CO₂, CO, C₂H₄, C₂H₆, C₃H₈, n-C₄H₁₀, i-C₄H₁₀, n-C₅H₁₂, and i-C₅H₁₂) with accuracy around 10% and typical detection limits of 2–3 ppmv.

For comparison, high precision gas chromatographic (GC) analyses of methane, ethane, propane, i-butane, n-butane, i-pentane, and n-pentane were also carried out for two locations. For point no. 160 (inside the central circular walkway) this comparison utilized 2 L canister samples connected to a 125 L closed-chamber where gas accumulated, without flushing with atmospheric air. A 2 L canister was also collected approximately 10 cm above one of the vents at the Wilshire-Curson intersection. Canisters were analyzed at Loyola Marymount University using a Bruker 430 GC equipped with a flame ionization detector (FID). C₂–C₅ alkanes were analyzed using a Varian 3400 GC-FID that utilized cryogenic concentration for samples. Analytical details were reported in Weber et al. (2017).

3.2 Gas Flux Measurements

We performed direct measurements of the flux of methane and carbon dioxide (CO₂) from the ground using the closed-chamber method and portable sensors and estimated the emission of ethane and propane on the basis of their relative concentration to CH₄ in the seeping gas.

Methane and CO₂ fluxes were measured at 214 points throughout Hancock Park, distributed over an area of 69,400 m² (Figure 3). These samples included 26 points over visible oil and bitumen seeps and 188 points that were on soil or grassy areas, either surrounding or far from the seeps. Three additional flux measurements were performed in correspondence with the Wilshire/Curson gas leak, from manhole covers and sidewalk cracks, a few meters outside the park (Figure 3). It is recognized that pavement can focus leakage toward existing cracks and openings.

Flux measurements were performed using widely tested instrumentation (e.g., Etiope, Christodoulou, et al., 2013; Etiope, Drobniak, et al., 2013), based on an aluminum, 10 L closed chamber (Figure 2d; West Systems srl, Italy) connected to portable CH₄ and CO₂ sensors: a Tunable Diode Laser Absorption Spectroscopy CH₄ detector (accuracy 0.1 ppmv, lower detection limit 0.1 ppmv) and a double beam infrared CO₂ sensor (Licor; accuracy 2%, repeatability ±5 ppmv). The data are transmitted via Bluetooth communication to an Android smartphone. The gas flux is calculated through a linear regression of the gas concentration buildup (ppmv/s) in the chamber; for high fluxes, where the flux curve tends to become convex due to advection decrease as the gas pressure within the chamber increases, the first linear part of the curve is considered. Fluxes are expressed as mg CH₄ m⁻² d⁻¹ and g CO₂ m⁻² d⁻¹, analogous with previous works on gas seepage and biologic fluxes from soils.

Carbon dioxide, a primary greenhouse gas, generally is not a major component of the petroleum seepage (CO₂ is more important in volcanoes and geothermal manifestations). However, the Salt Lake Oil Field gas has about 15 vol % CO₂ (Table 1). Carbon dioxide fluxes from the ground are expected to be comparable, and mixed, with CO₂ from soil respiration, except in high flux gas vents, where the CH₄/CO₂ ratio may be similar to the reservoir gas. Simultaneous CH₄ and CO₂ measurements in hydrocarbon seeps are additionally useful to verify that the chamber is correctly positioned on the soil surface, without leakages; this reduces considerably the uncertainty of the flux measurement. The accuracy of the flux measurement indicated by laboratory calibration tests is ±10% (www.westsystems.com), and field tests performed on various seepage sites indicated reproducibility <20% (Etiope, unpublished data). In general, due to perturbations of the pressure gradient in the soil, accumulation chambers may underestimate the gas flux by 0 to 25% (Evans et al., 2001).

The total CH₄ output from diffuse degassing (“miniseepage,” as described in section 1) was estimated using a “Natural Neighbor” spatial interpolation between individual gas measurements (Sibson, 1981). This method of spatial interpolation provides the best contour estimates for sampling of irregularly spaced points avoiding the assignment of flux values to sectors that were not actually sampled. Total gas output from visible seeps/vents was calculated by summing the emission measured in each vent. Within the Pit 3-4-61-67 enclosure (600 m²; Figure 2b, location indicated in Figure 3) the diffuse CH₄ flux values were very high, reaching on the order of 10⁵ mg m⁻² d⁻¹, distinctly higher than those measured outside the enclosure. The gas output from this area was calculated (by Natural Neighbor interpolation) independently to avoid overestimated extrapolations for the surrounding area outside of the enclosure. Accordingly, total gas output from the La Brea park is the sum of the emission from (a) all vents, (b) the diffuse miniseepage, and (c) the high degassing Pit 3-4-61-67.

Gas flux from the numerous bubble plumes in the Lake Pit (Figure 2a) was not measured, but the order of magnitude of the gas output was roughly estimated on the basis of bubble plume number, sizes, and bursting frequency, which were observed by video recording (up to 10 min long, for a total recording of 14 min). Bubble plume features were compared with experimental flux data and images of bubble plumes obtained in other seepage areas (Etiope, Christodoulou, et al., 2013 and unpublished data archive), in conditions similar to those of the Lake Pit (water depth of 7 m and CH₄ concentration in bubbles of ~80 vol %). These measurements allowed us to determine a flux of methane, in terms of order of magnitude, as a function of the size and frequency of individual bubble trains, which was consistent with theoretical models (Etiope et al., 2004). The emission of heavier alkanes, ethane, and propane was estimated considering their molecular ratio to methane (C₂/C₁ and C₃/C₁) observed in the oil seeps and in the gas exhaling from the soil, assuming that the relative composition of the gases in oil seeps and soil degassing does not significantly change over the seepage area.

4 Results

4.1 Molecular Composition of the Gas

As expected, the seeping gas is dominated by methane (Jeffrey et al., 1991), with concentrations observed in the collecting chamber or funnel (not corrected for air contamination as O₂ was not analyzed) up to 78 vol % (Table 1). Carbon dioxide is the second most abundant gas, exceeding 10 vol %, as reported in previous analyses (Jeffrey et al., 1991). The CH₄/CO₂ ratio in the gas seeping in the park is around 6, similar to the ratio at the Wilshire-Curson leak (5.8) and in the underground reservoir (5.6; Jeffrey et al., 1991). The C₁/(C₂ + C₃) ratio was much higher in one of the degassing soil samples (523) compared to the bubbling seep (150) and especially compared to the Wilshire-Curson leak (~45), which exhibited a ratio close to that of the reservoir (37). The isobutane/normal-butane volumetric ratio of the gas seeping in the park (>1) is also different from the original gas in the reservoir (<1), while the ratio in the Wilshire-Curson leak is similar to the reservoir gas.

4.2 Methane Flux

The gas flux measurements exhibited a widespread emission of methane throughout the park. Methane fluxes from oil-bitumen and bubbling seeps range from 6,900 to approximately 53,800,000 mg m⁻² d⁻¹ (supporting information S1), with the highest values (orders of 10⁶–10⁷ mg m⁻² d⁻¹) corresponding to seeps with visible gas bubbling. Methane flux from the grassy soil ranges from −3 mg m⁻² d⁻¹ (a normal value in dry soil in temperate climates) to more than 9 × 10⁶ mg m⁻² d⁻¹. Interestingly, the highest flux values were systematically observed in correspondence with grass characterized by a dull green color (Figure 2d); medium-high values were detected where grass was yellowish or not growing and, lower values where the grass had a more vivid green color. These are likely indications of plant stress and response to high hydrocarbon content in the soil, coupled with low oxygen. The reason why yellowish grass, or its absence, exhibited lower CH₄ flux than the dull green grass is unknown; we suspect that gas seepage changes its position over time (channeled gas flow may migrate horizontally), and yellowish grass represents the site where gas seepage was very high until recently; the current higher flux is starting to affect the grass producing a dull green color, which will become yellowish later. Specific biochemical analyses in the future may reveal the mechanism. Very high CH₄ fluxes were concentrated within the Tar Pits 3-4-61-67 area (Figure 2b), ranging from 882 to 789,000 mg m⁻² d⁻¹, over several small seeps and fractured soil. In this area a vigorous bubbling seep was responsible for the highest absolute gas flux (53,800,000 mg m⁻² d⁻¹).

As outlined in the section 3, total gas output from the La Brea park is calculated as the sum of the emission from (a) all vents (96 kg d⁻¹), (b) the diffuse miniseepage (177 kg d⁻¹, covering 69,400 m² of the park, estimated via Natural Neighbor interpolation), and (c) the high degassing Pit 3-4-61-67 (220 kg d⁻¹ over 500 m², via Natural Neighbor interpolation and adding separately the most intense seep, so as not to distort neighboring points of measurement). Total methane emission from the park is 493 kg d⁻¹ (180 t yr⁻¹). The uncertainty of the natural neighbor interpolation estimate of the diffuse emission, assessed by twofold cross-validation technique (e.g., Ciotoli et al., 2015), is within ±40 kg d⁻¹.

About 30 major bubble plumes, with different size and bursting frequency, could be detected in the La Brea Lake through video recording. Additionally, ~ 70–80 small and episodic bubble trains, not recorded by the camera, were observed near the banks of the lake. Three plumes were continuously active over 10 min of video recording, with a diameter of about 0.5–1 m. By comparison with experimental data, as described in section 3, we estimated that these gas plumes may emit individually an amount of CH₄ in the order of 10–100 kg CH₄ d⁻¹. Other plumes were episodic, with diameters from about 5 to 30 cm, and bubbling events lasting from a few seconds to 1 min. A group of 10 more active plumes may emit, individually, about 1–10 kg CH₄ d⁻¹. Seventeen other minor plumes (recorded by the camera), and those (>70) observed at the lake banks, could emit orders of 0.1–1 kg CH₄ d⁻¹. The order of magnitude of the total CH₄ output from the bubbling lake may then range from 10² to 10³ kg d⁻¹. However, these are first-order semiquantitative estimates awaiting verification by direct flux measurements.

The spatial distribution of the methane seepage (Figure 4) clearly shows a linear trend crossing diagonally the park from NW to SE and a second N-S trend on the eastern boundary of the park. The bubbling plumes in the lake are in positions consistent with the NW-SE seepage lineament.

Figure 4

Contour map of methane seepage in the La Brea park (iso-flux lines in mg m⁻² d⁻¹; methane flux is shown in different colors and its intensity increases from white, yellow, green, and orange to dark pink) and photos of shattered pavement, due to gas seepage, outside the southeastern corner of the park and at the Wilshire-Curson intersection.

Carbon dioxide fluxes from the seeps range from 16 to 9,400 g m⁻² d⁻¹ (supporting information). Carbon dioxide flux from the soil was, for most sites, within the range of normal soil respiration (typically 10–50 g m⁻² d⁻¹), and it exceeds 100 g m⁻² d⁻¹ only in association to the highest CH₄ fluxes. Because CO₂ in the seeping gas does not exceed 15 vol % (Table 1), in locations where the seepage is low the biologic soil respiration may prevail over the geologic CO₂ input, making it difficult to distinguish between the two CO₂ sources. In these cases, the CH₄/CO₂ mol/mol ratio in the flux data is much lower (often <1) than the original CH₄/CO₂ ratio of the seeping gas (which is around 5–6), due to the excess biologic CO₂. At high seepage points, the CH₄/CO₂ ratio ranges from 3 to 9, which is consistent with the molecular ratio of the seeping gas (Table 1). This is an indication of the good quality of the flux data.

4.3 Ethane and Propane Fluxes

Knowing the CH₄ flux and the C₁/C₂ and C₁/C₃ molecular ratio observed in the seeping gas (C₁/C₂: 685 and C₁/C₃: 3425 for soil degassing; 187 and 750 for bubbling seeps; 60 and 156 for the Wilshire-Curson leak; Table 1), and considering the molecular mass correction, the total emission of ethane and propane from the park was estimated in the order of 4 and 1 kg d⁻¹, respectively (Table 2). These estimates assume that the relative composition of the gas does not significantly change over the seepage area. Our partial measurements of the Wilshire-Curson leak, performed at only 3 points, showed C₂-C₃ emissions equivalent to those of the entire park. The uncertainty of these emission estimates derives from the methane flux uncertainty (discussed above).

Table 2. Total Emission of Methane, Ethane, and Propane (kg d⁻¹) From the La Brea Park Area

Seepage type or sector (measured area)	CH4	C2H6	C3H8
Diffuse degassing (69,400 m^2)	177	0.5	0.14
Pits 3-4-61-67 (500 m^2)	220	2.20	0.8
Oil-gas seeps (28)	96	1	0.3
Total emission from the park ground	493	3.7	1.2
First-order emission estimate from the lake (3,000 m^2)	10^2–10^3	10^0–10^1	10^0–10^1
Wilshire-Curson corner leak (3 m^2)	>130	4	2.3
Total output (order of magnitude)	10^3	>10	>5

5 Discussion

5.1 Postgenetic Alteration of Gas Composition: Molecular Fractionation and Biodegradation

Jeffrey et al. (1991) showed that La Brea seeping CH₄ and CO₂ has the same isotopic composition as the reservoir gas (thermogenic CH₄ signature and positive δ¹³C-CO₂ values, suggesting oil biodegradation; Table 1) but a slightly different concentration of several alkanes. This suggests that there is no isotopic fractionation during gas migration to the surface, as typically observed in many seeps worldwide (e.g., Baciu et al., 2017; Etiope et al., 2009). Instead, the C₁/(C₂ + C₃) ratio of the reservoir is 37 and increases to 98 in a seep (Jeffrey et al., 1991). We have observed similar values and trends in our seeping gas data. The Wilshire-Curson leak displays a C₁/(C₂ + C₃) ratio (~43) very close to the ratio of the Salt Lake Oil Field reservoir (Table 1). The ratio increases in the bubbling seep (150) and even more in one of the degassing soil samples (523). Basically, the C₁/(C₂ + C₃) ratio increases as the gas flux decreases, which is typical of gas seeps (Etiope et al., 2009, 2011). This phenomenon can be attributed to molecular fractionation during the advective gas migration from the reservoir to the surface. The process is a type of distillation, or differential segregation of light hydrocarbon molecules as a function of their adsorption and solubility properties, so that seeping gas is dryer (higher C₁/C₂₊ ratio) than the original reservoir gas (e.g., Brown, 2014; Etiope et al., 2009; Waseda & Iwano, 2008). Therefore, the lower the flux, the higher the residence time in the migration pathway system, and the higher the fractionation.

It is likely, however, that the C₁/C₂₊ ratio also increases due to addition of secondary microbial methane, following oil biodegradation: the reservoir is biodegraded (δ¹³C-CO₂: +15.6‰; Table 1) but additional biodegradation and methanogenesis may take place along the seepage systems. This is suggested by the higher δ¹³C-CO₂ value in the seeping gas, which is up to +25.9‰ (Jeffrey et al., 1991) and, in particular, by the ratio between isobutane and normal butane. Isobutane to n-butane ratios (iC₄/nC₄) in nonbiodegraded reservoirs are typically around 0.1–1.5 (e.g., Larter & di Primio, 2005; Pallasser, 2000), and the ratio generally increases with the maturity of the source rock (e.g., Prinzhofer et al., 2000). The iC₄/nC₄ ratio may, however, increase substantially, exceeding 5–10, due to preferential microbial degradation of n-alkanes versus i-alkanes (Pallasser, 2000). The ratios measured by us in La Brea are 17 in the bubbling oil seep (no. 198) and from 6 to 250 in the degassing soil site (no. 160). These elevated ratios are consistent with those previously observed in seeps by Jeffrey et al. (1991) and Weber et al. (2017). The ratio is much lower (<1) in the reservoir and in the gas of the Wilshire-Curson leak, which, due to the intense and rapid flow, does not exhibit alteration or fractionation from the reservoir (as also indicated by the C₁/C₂₊ ratio; Table 1). The fact that the reservoir appears to be biodegraded (positive δ¹³C-CO₂), but without significant n-C₄ consumption, may suggest that oil biodegradation in the reservoir did not perturb the iC₄/nC₄ ratio. Alternatively, the biodegradation was simultaneous to a progressive accumulation (from deeper reservoirs or source rocks) of wet gas with an unaltered iC₄/nC₄ ratio. Combining δ¹³C-CO₂ and iC₄/nC₄, it is clear that seeping gas is more biodegraded than the reservoir gas. Though Weber et al. (2017) suggested that the biological degradation is occurring in the wet asphalt, the increased ratio at the site of soil degassing no. 160 (a “dry” surface vent with no associated asphalt) suggests that soil microbes are perhaps an additional, or more likely, cause of nC₄ consumption.

5.2 Seepage Distribution

The NW-SE seepage lineament (Figures 4 and 5) has the same orientation and approximate position of the 6th Street Fault (Figure 1). The Wilshire-Curson leak is also positioned in line with the seepage lineament and the fault. These data support the hypothesis that the 6th Street Fault is actually the main preferential pathway of migration of gas from the Salt Lake Oil Field to the surface. Our spatial survey suggests however the possible presence of a secondary N-S pathway, transversal to and intercepting the 6th Street Fault near the southeastern corner of the park, at roughly the Wilshire-Curson intersection. The strong emission of the Wilshire-Curson leak could then be explained by the occurrence of a very permeable gas migration pathway, an enhanced fracture zone, created by the intersection of two faults. Large seeps are in fact frequently observed at the crossing of two or more faults (e.g., Etiope, 2015). The possible existence of a N-S fault should be examined reviewing available borehole and geophysical prospection data. However, as discussed above, we cannot exclude that the Wilshire-Curson gas is leaking from an abandoned, yet unknown (non-inventoried) well, as some authors hypothesized (Chilingar & Endres, 2005).

Figure 5

The 3-D plot of methane emission in the La Brea park with approximate position of the 6th Street Fault, based on detailed geologic map after Wright (1987).

5.3 Significance of Methane and C₂₊ Emissions

Individual methane fluxes measured as high as 9.0 × 10⁶ mg m⁻² d⁻¹ for miniseepage measurements made over soil/grass and 5.4 × 10⁷ mg m⁻² d⁻¹ for visible oil/asphalt seeps. These values are similar to fluxes observed in/around other seeps (typically in the order of 10⁵–10⁷ mg m⁻² d⁻¹; e.g., Etiope et al., 2011; Etiope, 2015). The fluxes measured in this study from visible seeps are more than an order of magnitude higher than those reported by Weber et al. (2017), which were considered as the lower limit of emissions because of sealing problems with the flux chamber. Thus, the results of this study seem to confirm that the Weber et al. (2017) study underestimated emissions from visible seeps.

The estimated overall CH₄ output from the La Brea park (including the measured emission of 493 kg d⁻¹ and the estimated order of magnitude of emission from the lake, 10²–10³ kg d⁻¹) is comparable with the emission of large gas seeps and mud volcanoes studied in Europe and Asia (Etiope, 2015, and references therein). If the Wilshire-Curson leak can be considered a natural seep, then the overall natural CH₄ emission from the La Brea area may exceed 10³ kg d⁻¹. Regardless, the gas emissions from the ground within the park (~180 t yr⁻¹) is the highest measured so far in onshore seeps in the USA. Smaller seepage fluxes were reported by Duffy et al. (2007) just to the west in Ventura County, California, Walter Anthony et al. (2012) in Alaska, LTE (LT Environmental, Inc.) (2007) in Colorado, and Etiope, Drobniak, et al. (2013) in New York State. A technical report by Exploration Technologies, Inc. (2001) reported methane flux data from the Playa Vista seepage zone in Los Angeles; applying the Natural Neighbor interpolation method on the flux data indicated in the report, we estimated a total output of about 160 kg d⁻¹, much lower than at La Brea.

The La Brea CH₄ emission factor (total emission divided by area) is on the order of 7,000 t km⁻² yr⁻¹ (considering park and lake emissions; Table 2). This is higher than the global average emission factor estimated for mud volcanoes (about 3,100 t km⁻² yr⁻¹; Etiope, 2015) and similar, in terms of order of magnitude, to the emission factor of the offshore Coal Oil Point seepage zone, near Santa Barbara (California), which releases about 30,000 t CH₄ yr⁻¹ over a relatively large area, ~ 3 km² (Hornafius, et al., 1999).

Through a mobile survey, Farrell et al. (2013) estimated a CH₄ output of about “1/6 of a kiloton per day” for the entire La Brea neighborhood. By comparison with our estimated order of magnitude of 10³ kg d⁻¹ for the park only, the Farrell et al. (2013) data would suggest that considerable CH₄ emissions occur in the area around the park. It is likely that significant gas leaks, similar to the one partially measured by us at the Wilshire-Curson intersection, occur elsewhere at one or more locations in the neighborhood.

Ethane and propane emissions, estimated in total at least 10 and 5 kg d⁻¹, respectively, are not especially large when compared to seeps measured in other regions. Much smaller, individual vents but with higher relative C₂₊ content release similar total C₂-C₃ amounts (e.g., Etiope, Drobniak, et al., 2013). Clearly, the molecular fractionation and biodegradation during the seepage, leading to high C₁/C₂₊, makes the La Brea park seepage a less significant source of heavier alkanes. The Wilshire-Curson leak, where gas is not fractionated during migration, is a relatively larger ethane-propane emitter. A more complete and exhaustive set of flux measurements should be performed at this site to better evaluate the overall C₁-C₃ output.

The variability of the C₁/C₂ and C₁/C₃ ratios, due to molecular fractionation during seepage, is an important process that should be taken into account in the attribution of gas sources (natural versus anthropogenic; fossil fuels versus biogenic), as performed by Peischl et al. (2013) and Hopkins et al. (2016) for the Los Angeles area. Our data confirm that the C₁/C₂₊ ratio of geologic seepage (a natural fossil fuel source) can be lower than that of leaks from boreholes and pipelines (i.e., fugitive emissions and anthropogenic fossil fuels) where gas composition is practically the same as the subsurface reservoir. Seepage C₁/C₂₊ ratios can even be similar to those of biogenic sources (>500) so that, in the measurements performed in the atmosphere, geologic seeps may be confounded for biogenic sources if the underground source is not known, as suggested by Hopkins et al. (2016).

A similar discussion can be made concerning the use of isobutane to normal-butane ratio for source attribution (e.g., Peischl et al., 2013). Our data show that biodegraded seeps may release gas with a very high iC₄/nC₄ ratio (>6), higher than that of the natural gas distribution network or vehicles (traffic), which are typically ≤1 (Peischl et al., 2013). We note that Peischl et al. (2013), in their Figure 6f, considered seeps from Los Angeles (data from Jeffrey et al., 1991) releasing gas that is scarcely or nonbiodegraded, thus with iC₄/nC₄ ratios ≤1. We have redrawn Figure 6f of Peischl et al. (2013), adding all available La Brea data, including those of Jeffrey et al. (1991), Weber et al. (2017), and those from the present work (Figure 6). Clearly, seepage of biodegraded reservoir gas provides an iC₄/nC₄ signal in the atmosphere that is completely different from that of anthropogenic sources. Accordingly, the iC₄/nC₄ ratio can be used as a specific tracer of a natural source of fossil fuel (i.e., seepage), but geochemical analyses of both seeping and reservoir gas are necessary to evaluate if and where molecular alteration by biodegradation occurred. In Los Angeles, if biodegraded seepage of gas, with high iC₄/nC₄ ratio, is occurring over a large enough area, it could account for some of the missing isobutane in the local budget estimates made by Peischl et al. (2013).

Figure 6

Isobutane versus normal-butane plot for La Brea seeping gas, compared with anthropogenic fossil fuel data, petroleum reservoirs, and nonbiodegraded seeps in Los Angeles, after Peischl et al. (2013).

6 Conclusions

Thermogenic methane and minor amounts of heavier alkanes (ethane, propane, and butane) are emitted from the La Brea Tar Pits from visible oil-asphalt and bubbling seeps and, diffusely, from the soil, throughout the entire area of the park. Gas seepage clearly affects grass physiology, as evidenced by changes in the plant color and growing patterns in relation to the measured CH₄ flux. This phenomenon would deserve a specific biochemical study. We estimated a total methane emission from the ground of about 0.5 t d⁻¹, of which ~35% is from soil degassing, ~20% from 26 active oil/asphalt seeps, and ~45% from a small area with a high degassing rate, the Pits 3-4-61-67 enclosure (including soil degassing and seeps). Additional significant emissions, roughly estimated in the order of 10²–10³ kg d⁻¹, are from bubble plumes in the lake located on the southern margin of the park. Another important gas emission occurs just outside the park, at the intersection of Wilshire Boulevard and Curson Avenue (>130 kg d⁻¹ measured at only 3 points). La Brea park represents the onshore seepage zone with the highest methane output measured in the USA, to this point in time.

The higher gas fluxes are distributed along a NW-SE lineament that has the same orientation and position as the 6th Street Fault. As suggested by previous geologic surveys, this fault seems to act as a main pathway of gas migration from the Salt Lake oil reservoir (>300 m deep) to the surface. Our data suggest, however, that the gas seepage reaches beyond the boundaries of the park, and the total gas output we estimated is a fraction of the actual seepage in the La Brea-Hancock area. Further gas flux measurements should be carried out at the Wilshire-Curson intersection and in all streets surrounding the park.

The seeping gas shows geochemical features typical of oil biodegradation (positive δ¹³C-CO₂ values and high i-butane/n-butane ratio). The gas has also a higher C₁/C₂₊ ratio compared to the reservoir gas, which can be due to molecular fractionation and secondary methanogenesis along the seepage system. These molecular alterations are important tracers of natural seepage and should be considered in future studies and models aimed at assessing the fossil fuel (anthropogenic versus natural) or biogenic origin of atmospheric methane over a given region.

Acknowledgments

The authors would like to thank Aisling Farrell, from the Page Museum at the La Brea Tar Pits, for her assistance, including access to sampling areas not open to the public, and Giancarlo Ciotoli for consultancy on geospatial analysis of the data. The data used are listed in the tables and supporting information. Three anonymous reviewers provided valuable suggestions that improved the manuscript.