Earlier this year, headlines across the country blared an alarming finding: methane emissions from oil and gas could be 60 percent higher than what federal data suggest. Coverage from USA TODAY and Time, among many others, painted a picture of shale gas that was far different than what multiple scientific assessments have concluded.
But a new study, published today by the Proceedings of the National Academy of Sciences (PNAS), shows how the earlier finding of higher-than-expected methane leakage could itself be a result of how top-down aircraft measurements are extrapolated to annual averages. The research team was led by Dr. Timothy Vaughn at Colorado State University, who was joined by experts from the University of Colorado, the National Oceanic and Atmospheric Administration, and the National Renewable Energy Laboratory, among others.
The new research has implications for earlier studies that employed what’s known as top-down (TD) inventories, where scientists would fly over producing areas and record how much methane they encountered. Curiously, these types of studies have traditionally found higher emissions than those taking measurements on site, or so-called bottom-up (BU) inventories. Some researchers have suggested that TD inventories are capturing so-called “super-emitters” that cannot be captured in BU inventories. The study that alleged emissions were 60 percent higher than U.S. EPA data used TD measurements.
What today’s study shows, however, is that top-down measurements can significantly inflate methane emissions from oil and gas operations if they are assumed to occur continuously. One reason is because TD inventories often measure emissions during snapshots in time during the afternoon while routine maintenance activities are occurring. Maintenance typically only lasts a few hours, but TD inventories in the past have assumed the measurements are representative of operations all the time.
Methane emissions from oil and gas sites are dynamic and temporal. Measurements taken during flyovers typically occur in the late afternoon, which are often when emissions are “amplified” due to onsite maintenance activities. As a result, those measurements are not necessarily representative of the average emissions profile of a particular operation, much less an entire producing basin.
How do we know that? Because the researchers confirmed it in the study released today. As the study notes:
“Study area total emissions exhibited significant variability throughout the day. On both days of the study period, modeled emissions peaked during mid-afternoon hours due to MLUs performed and recorded by production facility operators.” (emphasis added)
To address this problem, the researchers took concurrent measurements on site in the Fayetteville basin in Arkansas while they were also taking flyover measurements. By comparing the top-down and bottom-up techniques during the same time, the researchers were able to complete a first-of-its-kind reconciliation.
In other words, top-down inventories performed without an accompanying understanding of bottom-up operations aren’t necessarily giving us an accurate picture of average emissions. Rather than confirming the existence of “super-emitters,” TD studies may suggest average emissions are far higher than they actually are. In the study published today, the researchers were able to largely close the gap between their TD and BU measurements by considering the timing of emissions events and not over-extrapolating the TD findings.
There is nothing, then, that requires the presence of “super-emitters” to explain higher emissions in top-down studies for this particular basin.
Getting an Accurate Picture
Today’s study benefited from concurrent onsite measurements, which were done in cooperation with several oil and gas companies operating in the basin. This is what allowed them to reconcile top-down and bottom-up measurements, and thus provide us with a clearer picture of methane emissions from oil and gas sites.
The key takeaway is that industry cooperation is critical to having scientifically valid results. As the researchers showed, top-down measurements may be capturing events that are temporary, particularly daily maintenance activities such as manual liquids unloadings. Not being able to compare those data against data obtained onsite was previously a major limitation of the results obtained during TD studies. Prior TD researches have posed various theories for their higher emission findings; the authors of the paper released today designed a comprehensive, rigorous study to reconcile differences between TD and BU measurements, based on actual events and evidence.
It’s important to stress that today’s study doesn’t necessarily “debunk” prior research, although it does provide a reasonable explanation as to why top-down measurements show methane emissions to be higher than other findings. As a result, TD measurements should be assessed with care and caution, as they could very well be conveying an operational picture that does not exist.
Today’s study also provides strong evidence that similar reconciliations between TD and BU measurements need to occur in other basins to get a better handle on methane emissions.
Are methane emissions higher than EPA estimates? A large body of research already corroborates EPA’s data, but there have been a handful of studies – often widely covered by the media – that have succeeded in sowing doubt about methane emissions. The study released today, however, may provide a scientific blueprint for studying methane in additional basins across the country.
In a dynamic operating environment where workers primarily work during the daytime, the timing of measurements matter. As the authors note: “Careful consideration of all factors influencing methane emissions—including temporal variation—is necessary in scientific and policy discussions to develop effective strategies for mitigating greenhouse gas emissions from natural gas infrastructure.”
This study may also be useful in state and federal regulatory proceedings. If nothing else, what the researchers demonstrated is that the underlying assumptions and extrapolations of prior methane studies – especially those being used to guide public policy – need to be carefully scrutinized.