The process of hydraulic fracturing (or “fracking”) shales to extract oil and natural gas has lowered prices and displaced some coal with cleaner-burning natural gas in the US. However, some of the methane we're extracting also escapes from oil and gas wells and heads straight to the atmosphere, where it is a potent greenhouse gas.

That leakage is harmful to the climate, a wasted resource, and lost profit for natural gas producers, so researchers are working hard to find out just how much is leaking. If enough of it gets loose, natural gas can even lose its carbon emissions advantage over coal, despite its cleaner-burning nature.

Many different natural gas fields have been investigated using different methods. Some estimates are “top-down,” using measurements from aircraft circling well fields to estimate how much is coming out of wells and pipelines. Other estimates are “bottom-up,” relying on measurements on the ground at individual sites and scaling them up to the total number of sites. Top-down techniques often yield larger estimates, and leakage rates can vary widely from one gas field to another. It’s complicated.

In the fall of 2013, a large group of researchers collaborated on a unique effort to throw the kitchen sink at the gas production around the Dallas-Fort Worth area (from the Barnett Shale). Many techniques—some top-down, others bottom-up—were independently employed simultaneously to complement each other. The result was a pile of papers published together in the journalEnvironmental Science & Technology detailing every aspect of leakage from the Barnett.

Leakage inequality

An important point has become increasingly clear through this kind of research: the majority of methane leakage from these gas fields often comes from a small minority of point sources. That was true here as well. One group sampled 186 sites where a well or pipeline equipment was located. Just five percent of them accounted for over half of the total methane leakage. At 30 percent of the sites, on the other hand, their instruments detected no leakage at all.

There are at least two obvious ways to get one of these “super-emitters”: either a moderate flow of gas through the site is simply escaping completely, or a much smaller percentage is leaking from a piece of equipment hosting a much higher volume of flow. This difference is important if you want to actually do something about that leakage. Where most of the flow is escaping, there’s often simply something wrong with the equipment—a valve stuck open, perhaps—that could be remedied.

Another paper laid out a functional definition of “super-emitter” designed to highlight those avoidable equipment malfunctions. They separated the samples sites into four groups based on the amount of gas moving through them. Any site above the 85th percentile for leakage (based on the portion leaking rather than the absolute amount) in each group was flagged as a super-emitter. If you’re looking for broken equipment, that’s your list of suspects.

Even within that list, the worst of the worst separated themselves. Between the 95th percentile and the very top, leakage rates could jump from about 10 percent loss to nearly 100 percent.

The researchers calculated that if you could bring all those super-emitters down to the average leakage rate of similar sites, the total amount of methane being lost to the atmosphere could be cut by 73 percent.

Give me a number

Some early aircraft-based leakage estimates for smaller gas fields in the US yielded shocking amounts of leakage, ranging from around four percent of production to as much 20 percent—much higher than the EPA’s national estimate of around 1.5 percent. However, a recent study covering much larger gas fields calculated leakage rates closer to one percent.

So how does the Barnett Shale region, which contributes about eight percent of natural gas production in the US, compare? Another top-down, aircraft-based study estimated leakage at 1.3 to 1.9 percent. A bottom-up study pulling together the on-the-ground measurements put it at 1.2 percent. Those numbers are low enough to ensure that burning natural gas in power plants has a smaller climate impact than coal.

While this compares well to the EPA’s national estimate, it’s actually significantly larger than indicated by several inventories used to track local greenhouse gas emissions, including the EPA’s. So even if it’s not a big number, it’s bigger than expected.

One interesting detail from the bottom-up work is that the biggest discrepancy in the inventories wasn’t leakage from the gas wells themselves. It was leakage from the “gathering stations” that compress gas coming from the wells and pump it into the backbone pipelines that transport it. Those stations accounted for a whopping 40 percent of the total leakage. And once again, a few particularly leaky gathering stations provided an outsized share.

These studies continue to fill out our understanding of how much natural gas is leaking, where it’s leaking, and why. Armed with that information, officials can craft good policies (and industry can take effective actions) that plug the most glaring leaks. In this case, minimizing environmental impact should even coincide with maximizing profits.