Methane emissions from oil & gas development

Earlier last year we posted a blog on whether the new natural gas boom, thanks to improved drilling technologies and hydraulic fracturing or “fracking”, was to be considered a boon or bane to Earth’s climate. The boon part comes from the fact that natural gas burns much cleaner and causes roughly a factor of two lower CO2 emissions than the burning of coal. So if the gas were exclusively used in high efficiency gas-fired power plants, or even combined heat and power (CHP) plants to replace coal combustion power plants for electricity production, CO2 emissions reduction would be maximized. The bane part is the fact that mining and use of natural gas does not happen without the inevitable gas leaks, in this case releasing a different, more powerful greenhouse gas: methane.

We concluded that knowledge on leak rates (commonly expressed as a percentage of produced gas), especially for newly developed wells and their infrastructure, was lacking. Some scientific estimates implied rates near or below 2%, while others implied 5% or more. We also pointed out that, regardless of current leak rates from booming oil&gas activities, methane leakage in general is an important issue.

The methane budget

Several recent scientific assessments put current fossil fuel related, “fugitive” methane emissions to the atmosphere at 100 million tons per year, roughly two thirds coming from the oil&gas industry, the remaining third from coal mining. It is useful in this context to realize that humans have roughly tripled the emissions of methane to the atmosphere since the beginning of the industrial revolution. Meaning, nature only provides for one third of atmospheric methane, the other two thirds are from human activities, dominated by domestic ruminants (mostly cows, i.e. the beef you eat) and fossil fuel mining and use. At the same time, nature takes care of all methane removal from the atmosphere, overwhelmingly through its slow atmospheric photo-oxidation. This oxidation is responsible for an atmospheric lifetime of methane of nine years and causes a ripple effect through atmospheric chemistry, such as via producing ozone and carbon monoxide, and via increasing the lifetime of other trace gases, including methane itself.

Inventoring human emissions

Because methane is such a strong greenhouse gas, reducing its emissions has direct benefits for climate stabilization. Methane’s comparatively short atmospheric lifetime would make the effects of emissions reductions measurable in the atmosphere within a decade. Alas, neither the production of beef nor the mining and use of fossil fuels are on the decline. Nevertheless, much ado has been made of EPA’s 2013 US greenhouse gas inventory, in which the agency lowered its estimates of past oil&gas industry related methane emissions to below 2% of produced gas amounts. This change has been misused by “pro-fracking” advocates to again attack the initial Howarth work and argue that methane releases are much lower than presumed, while “anti-fracking” advocates have instead highlighted that methane still contitutes a large fraction of US greenhouse gas emissions.

It is useful, therefore, to remind everyone what an inventory actually is. EPA has been keeping inventories of all regulated air pollutants since the 70ies and of greenhouse gases for approximately two decades. Each is based on knowledge of emission sources, how a source behaves as a function of "activity" or source mechanism (e.g. a combustion versus an evaporation process), external influences on a source, such as temperature, and its potential variation in time and space. This information, usually collected on short time and small space scales, is then extrapolated to longer times and larger scales using metrics such as the number of vehicles and their driving modes, the population density in an area, or the average number of times well sites undergo "liquids unloading". Numerous uncertainties and biases in this inventory development process, from the knowledge of sources to the type and manner of the extrapolation, make the inventory essentially an educated guess, not a reliable number. An example is the emission of carbon monoxide from car traffic, which the EPA has been overestimating for decades [1]. Thus, EPA regularly seeks to validate its pollutant inventory. In the case of methane, EPA’s emissions process knowledge was badly outdated and the agency strongly leaned on detailed information provided by the industry it oversees to update the inventory. No validation activity has yet been undertaken by the agency to verify that its changes (or lack thereof) were justified. So it is a stretch to claim that EPA’s change to the greenhouse gas inventory in itself is evidence of drop of emissions. Such a claim can only be made through independent validation measurements, such as those discussed below.

So are the emissions from the fossil fuel industry in the US increasing due to fracking, or not?

Unfortunately, this question was not answered in 2013, despite a number of new publications shining a light on the question through actual measurements. In August, a publication by CIRES and NOAA researchers [2] showed that methane emissions from a large oil&gas exploration field in Utah may be 6% to 12% (one standard deviation range) of average production rates, far exceeding the less than 1-2% claimed by the industry. In November, another study by NOAA [3] revealed that methane emissions nationwide appear to be significantly underestimated (by the inventory!) with respect to both of the large man-made sources, beef production and fossil fuel mining.

On the other hand, the large University of Texas (UT) study already mentioned in our earlier post published its first research paper [4] showing that industrial gas well operations carried out by several large companies produced highly variable methane emissions depending on what process was studied, where it was studied, and what equipment was used. Overall though, it did not show enhanced fugitive methane emissions over the inventory estimate. The UT study, sponsored by the industry, who selected well sites and times for measurements, did the Herculean task of on-the-ground individual well-site measurements to evaluate the leak rates of step-by-step well development and individual gas handling equipment. Most importantly, it found that measures to mitigate fugitive methane emissions, such as collecting back-flow fracking fluid, separating it from the methane it contains, then flaring said methane, do work in practice. It provided partial inventory validation data. But it was criticized for potentially being biased as it measured what would be expected without showing that these practices are representative or implemented nationwide. Nevertheless, the authors extrapolated nationwide emissions from their data, finding that EPA’s current inventory, which is based on similar data provided by the industry, is compatible with their data.

Apples and Oranges?

It is important to realize that the bottom-up study by UT researchers cannot directly be compared to the top-down studies carried out by NOAA researchers. The former is a single source study typically used by regulatory agencies, such as the EPA, to extrapolate to nationwide estimates. In contrast, the NOAA-led studies integrated over a larger region and do not provide the temporal and spatial resolution the UT-led study provided. Such top-down studies, however, include large numbers of well sites irrespective of their operation and irrespective of well type, in addition to sources not directly related to the oil&gas industry. Thus, they address integrated methane emissions from a region, and are not limited to fracking activities. As a result, critics have speculated that there exist a varying amount of well sites per region that have very high fugitive methane emissions, skewing the distribution of leak rates. If so, only a representative sampling of these sites could potentially reconcile the differences between the UT/industry and NOAA studies.

There are, however, other possible explanations. Methane emissions from co-located non-oil&gas sources could be underestimated, methane could also seep from old and abandoned wells or occur as "natural seepage", and oil well sites are often interspersed with gas well sites causing potential double counting of sources. This is because produced raw oil is separated from its gaseous components, such as methane and ethane, but if those gases are not marketable, for example because there was or is no pipeline infrastructure in place, they were either “vented” (i.e. released to the atmosphere) or, more commonly flared. Texas wasted the equivalent of 1% of its annual gas consumption to flaring in 2011/12.

And (uncontrolled) flaring becomes a source of other pollutants, such as soot and more reactive volatiles that go on to contribute to regional ground-level ozone formation, which has become a major air quality concern. It can therefore be argued that at least some fraction of the air pollution reduced by burning natural gas instead of coal for electricity production [5] is re-introduced locally via flaring unwanted volatiles at well sites. Permit applications for flaring at Texas well sites have increased dramatically since 2008, so was venting the dominant process before that time? Could this be an alternative explanation for the higher than inventoried central Texas 2008 methane emissions outlined by the second NOAA study? What is the role of oil and gas producing versus only gas-producing wells?

The way ahead

These and other questions are currently explored by various researchers. At the AGU Fall Meeting 2013, similar to 2012, several sessions addressed the topic or gas leakage, and we summarize some highlights here.

1. Leak rates are indeed highly variable

The hypothesis that leak rates are even more variable than indicated by the UT study is supported by mobile measurements carried out in the Barnett Shale area (presentations A44A-07, A53H-03, A44F-05). Researchers surveyed 275 well sites, of which 77% were found emitting methane. Among 52 well sites studied in detail, they found emission rates between several liters per minute (lpm) up to several hundred lpm per well site. Even among a group of nearly identical well sites drilled by the same company, a very large range of well-site emissions was observed. Methane in areas of intensive oil&gas development, such as the Barnett shale in Texas, showed a clear fossil fuel carbon isotopic signature (Fig. 1), and closely associated ethane with methane (Fig. 2). It also showed the city of Dallas as a significant methane source.

graph from presentation A53H-03

Figure 1 (original text): (left panel) Distribution of oil and gas well pads (yellow) and landfills (blue) in the Dallas / Ft. Worth area. Mobile nocturnal measurements of methane are shown in red, indicating a strong degree of source heterogeneity. (right panel) Histogram of individual isotopic source signatures, showing distinct signatures for landfills (red) and oil and gas sources (green).

 

2. NOAA investigated more shale areas using the mass balance technique

NOAA researchers presented data from research aircraft flights (similar to Karion et al., 2013 [2]) in the Barnett, Haynesville, Fayetteville, and Marcellus shale areas (abstracts A44A07-06, A44F-05, and A53H-02). Except for the Marcellus shale flight, clear-cut methane plumes were encountered in all regions, and co-emitted ethane was used to distinguish sources (Fig. 2). After analyzing all these data, methane emissions estimates from roughly one third of all new oil&gas development areas will become available.

NOAA graphics from session A44A

Figure 2 (original text): Flight track colored by methane (CH4, left) and ethane (C2H6, right) mole fraction. A three-hour back trajectory (red line) constructed from lidar wind measurements passes over the Barnett natural gas well locations (gray points) prior to reaching the location on the flight path indicated by the red star

 

3. Air pollution impacts are getting stronger recognition

For years now, ground level ozone in parts of Utah (Uintah basin), especially in winter, has been extremely high at times (>100 ppb), clearly affected by the extensive regional oil&gas development. Numerous posters and presentations showed that the hydrocarbons co-emitted with methane cause both particle and ozone pollution, and that winter snow cover is critical in enhancing the smog. Regional ozone enhancements due to oil&gas developments since 2007/08 are now measurable also in Texas.

Some conclusions

The presented data, including both well-site and atmospheric measurements, can be expected to enter the peer-reviewed literature later in 2014. In addition, now that the scientific community has developed an arsenal of investigation techniques, more measurements are likely to be carried out in the coming years, some already announced at the AGU Fall Meeting.

While not yet resolved, recent well-site and atmospheric methane measurements and associated emissions calculations suggest that the current inventory numbers for emissions from the oil&gas industry are likely underestimates. A continental scale estimate presented at AGU (abstract A44A-08) suggests that actual, nationwide emissions are more likely in the 3-5% range of produced natural gas. Emission rates may be significantly higher or lower locally, creating a large range of well-site variability, making upscaling from well-site measurements difficult and noisy at best.

At these leak rate levels, natural gas still holds a greenhouse gas advantage over coal combustion for electricity production in the long run. However, such leak rates are higher than claimed by the industry, and co-emitted or flared hydrocarbons produce locally and regionally recognized air pollution that needs to be addressed. While methane leaks may not have been on everybody's radar in the past, they have always mattered. Or as one AGU abstract (A53A-0138) elegantly summarized:

"There is increasing recognition that minimising methane emissions from the oil and gas sector is a key step in reducing global greenhouse gas emissions in the near term. Atmospheric monitoring techniques are likely to play an important future role in measuring the extent of existing emissions and verifying emission reductions."

 

References

1.  Parrish, D.D., Critical evaluation of US on-road vehicle emission inventories. Atmospheric Environment, 2006. 40(13): p. 2288-2300.

2.  Karion, A., et al., Methane emissions estimate from airborne measurements over a western United States natural gas field. Geophysical Research Letters, 2013. 40(16): p. 4393-4397.

3.  Miller, S.M., et al., Anthropogenic emissions of methane in the United States. Proceedings of the National Academy of Sciences, 25 Nov. 2013, doi: 10.1073/pnas.1314392110.

4.  Allen, D.T., et al., Measurements of methane emissions at natural gas production sites in the United States. Proceedings of the National Academy of Sciences, 16. Sept. 2013, doi: 10.1073/pnas.1304880110.

5.  Pacsi, A.P., et al., Regional Air Quality Impacts of Increased Natural Gas Production and Use in Texas. Environmental Science & Technology, 2013. 47(7): p. 3521-3527.

Posted by gws on Thursday, 2 January, 2014


Creative Commons License The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License.