Long-awaited research results suggest methane leakage small.
The rise of two particular techniques -- horizontal drilling and hydraulic fracturing or fracking -- to extract natural gas from layers of shale buried deep beneath the earth has engendered broad concern and debate about the environmental damage fracking can bring. Key among these concerns has been the possibility that fracking causes contamination of surface waters and drinking water.
From the data I have seen, it seems pretty clear that while not always the case, fracking operations have led to water contamination in many instances. (See here, here, here.) What such contamination implies about the safety of fracking remains an open and hotly debated issue. But that's not the only issue nipping at the heels of fracking enthusiasts.
Another Concern Relates to Climate Change
It has long been argued that natural gas is a good bridge fuel for helping to wean us away from coal while we develop more viable renewable energy sources. The premise for this argument is that methane, the major component of natural gas, is relatively clean; deriving one British Thermal Unit (BTU) of energy from burning natural gas produces about half as much carbon dioxide (CO2) as a BTU from coal. And so, substituting natural gas for coal would seemingly slow the emissions of CO2 and thus slow global warming.
But as Bob Howarth of Cornell University has pointed out, such an accounting may be wrong. While clean as a fuel, methane is a powerful greenhouse gas, some 21 times more powerful than carbon dioxide. If a large amount of natural gas leaks into the atmosphere during fracking and transport, the benefit of using natural gas instead of coal disappears. And with too much leakage (in a paper I co-authored we estimated the threshold to be about 3.2 percent of total gas production), natural gas may actually be worse from a climate perspective than coal.
So How Much Is the Leakage?
The critical question: how much natural gas leaks into the atmosphere during production? Easy to ask, but answering has been difficult because there is so little data. The latest estimate from the U.S. Environmental Protection Agency is a leakage rate of less than 1 percent. Other estimates, like Howarth's, are much larger, perhaps closer to 4 percent.
To try to help settle the debate, a team of scientists funded by a consortium of oil and gas companies and Environmental Defense Fund have undertaken an ambitious program of 16 studies to directly sample natural gas leakage rates from natural gas wells. David Allen of the University of Texas, Austin, led the first study, the results of which were published online yesterday in the Proceedings of the National Academy of Sciences. (Full disclosure: I, along with a half dozen other scientists, served on an external science advisory panel as the project plans were developed. I received no compensation, and was not involved in the collection of data or its analysis and interpretation.)
The study involved emission measurements at 190 natural gas sites in four regions of the country where the authors were given access. Of those sites, the authors measured emission from:
- 150 sites in production with a combined total of 489 wells;
- 27 well completion flowbacks (period between drilling and production);
- nine well unloadings (where excess liquids are removed during production); note that these data, due to high variability between wells and small sample size, not included in the national estimate; and
- four well workovers (when a well is re-completed); because this category had so few data points and accounted for such a small percentage of emissions, the data from these events were not included in the analysis.
The bottom line: 0.42 percent of the gross natural gas produced leaks into the atmosphere -- a percentage that is slightly below that of the EPA and low enough to suggest that methane leakage from natural gas production is not a major concern when it come to climate change. (Keep in mind, however, that this percentage only pertains to emissions from the production part of the system. According to EPA, less than 40 percent of leakage [pdf] is from field production. Significant leakage can also occur elsewhere in the system, for example, from the pipelines that transport natural gas.)
One surprising find of the study is that more methane leaks from onsite equipment than has been previously accounted for. In particular, pneumatics, which control basic well-site operations, had up to 270 percent higher leak rates than EPA's estimates, with Gulf Coast pneumatics having the highest rates and Rocky Mountain pneumatics the lowest (the authors hypothesize these are low as a result of regulations to control ozone).
There are caveats to the study, of course. For example, it looked at only a tiny fraction (between a few tenths and a few hundredths percent) of sources of methane emissions at natural gas sites around the country. A fraction so small, one can hardly consider the results to be statistically definitive, especially given the fact that there is a lot of variability in the emission rates the authors measured between individual wells.
Second, it included only nine companies out of the thousands of producers in the United States. As explained on EDF's website:
“In 2011, the participants accounted for roughly 12% of all U.S. gas wells, 16% of gross gas production and almost half of all new well completions. In 2012, the 150 production sites UT visited include 478 wells, or about 0.1% of the national total of 446,745 gas wells.”
It could be that this subset of companies -- the ones willing to have their sites measured -- are the ones that are most careful to limit leakage and thus are not representative of the industry.
Third, it is hard to rationalize these results with other recent studies that have used a completely different approach to calculate basin-wide leakage rates. These studies suggest that methane leakage from natural gas production sites is considerably larger: on the order of some four percent larger to as much as 6-12 percent.
Still, as the most comprehensive study on leakage from active sites to date, this study has to be taken seriously.
Slings and Arrows
Anything that has to do with fracking is bound to engender hot debate and the Allen et al. paper is no exception. There are those out there who are questioning the work by questioning the links between the sponsors of the study and the industry it is studying. I am uncomfortable when arguments are made to question funding sources as a way to question a study's validity because it ultimately questions the scientists' integrity. Let's stick to the substance of the work and leave the other stuff out.
A criticism I find somewhat more plausible is that the results are not reliable because the oil and gas companies chose the sites to be studied. Isn't it possible that the companies chose "model sites" in order to make things look a lot better than the actual are? This issue is addressed to some extent at the UT-Austin website where it is stated that the team got to choose the sites from a list and got to choose the time:
“[T]he UT-led study team selected times and general locations for sampling activities, and companies provided access to completions that occurred during those periods. Production sites near the completions were selected by the study team for sampling based on lists of available sites in the region provided by the participating companies.”
Here's my two cents: all this debate about statistical representativeness, small datasets, only a handful of companies, and sites being chosen by the gas companies misses a key result -- perhaps the most important one. What the Allen et al. study clearly demonstrates is that there are sites out there with very low leakage rates; in other words, the oil and gas industry is capable of running a fracking operation with very low amounts of natural gas leakage. Maybe this is typical of most wells and maybe not. Regardless, it can inform the way wells are operated in the future: Let's find out what the wells with low leakage rates are doing, translate those procedures into best practices, and require all wells to operate that way, and provide for stiff fines if they screw up. And while we're at it, let's require that they install monitors to track methane leakage in real time at their wells.
Maybe, just maybe, we've made a little bit of progress.