The boom in US natural gas production made possible by fracking techniques has raised an awkward question: how much is leaking to the atmosphere before reaching a power plant turbine or your furnace? Natural gas power plants are more efficient than coal-burning plants and emit much less CO2. But methane is a potent, though short-lived, greenhouse gas, so the exact benefit of that trade off depends on the level of leaks from wells and pipelines.

The EPA produces estimates of leakage calculated using limited measurements of typical equipment and production practices. Those estimates put natural gas leakage in the neighborhood of one percent of production— low enough to ensure that the shale gas (fracking) boom is a net positive in terms of climate-changing emissions. A major study sampling new shale gas wells showed that the EPA estimates for well leakage did a pretty good job—at least for those newer wells.

Much has been made, however, of several studies that took a different approach and got very different results. Those studies used methane measurements made from a NOAA airplane upwind and downwind of shale gas fields. At a field outside Denver, that yielded an estimate of 3.1 to 5.3 percent leakage. At a Utah field, leakage was estimated at between 6.2 and 11.7 percent. Near Los Angeles, a leakage rate of 12-22 percent was calculated.

So what gives? Are the EPA numbers greatly underestimating leakage? Is shale gas actually worse than coal? To make sense of this, it’s important to keep in mind that the three shale gas fields sampled, together, accounted for just about two percent of US natural gas production. The big fields had yet to be sampled.

A new NOAA study remedies that. In the summer of 2013, sampling flights were sent around the Haynesville Shale field in Texas and Louisiana, the Fayetteville Shale in Arkansas, and part of the Marcellus Shale area in northeastern Pennsylvania. Together, these three areas account for about half of shale gas production in the US, and about 20 percent of the total natural gas production.

By measuring methane concentrations upwind and downwind of the fields, the researchers are able to calculate an in/out mass balance, and compare it to well production records. These areas contain some gas processing and pipelines, so the leakage rates include more than just the wells themselves. There are also some other sources of methane, like landfills, livestock, and coal mines, though they were minor in comparison to the natural gas activities. The researchers tried to account for those other sources in the error bars on their natural gas leakage estimates.

Around the Haynesville in Texas and Louisiana, they estimated leakage at 1.0 to 2.1 percent of production. Leakage in the Fayetteville region in Arkansas was similar, between 1.0 and 2.8 percent. (A 2008 Arkansas Department of Environmental Quality report also estimated leakage at 2.0 percent.) And in northeast Pennsylvania, leakage from Marcellus production was well under one percent (0.18-0.41 percent).

Averaged together, the estimated leakage is 1.1 percent, in close agreement with the EPA’s estimates.

Why the differences—sometimes large—from one region to another? The researchers say that it’s partly down to the character of the gas coming up. In some fields, like the Denver, Utah, and Los Angeles ones mentioned earlier, there’s more oil, and the gas has to be handled a little differently. Another important factor is that improvements in drill rig efficiency mean recent efforts have a leg up on older wells.

Monitoring methane leakage from natural gas production, transport, and use continues to be important, because less is always better. Leakage contributes to anthropogenic greenhouse gas emissions, and there isn’t any benefit to letting the methane go to waste. Now that drilling activities are well monitored, attention can shift downstream towards consumption Apart from some early work in a few cities, there’s plenty to be done monitoring leaks that occur as the gas is distributed and used. But it still appears that the climate impacts of natural gas beat using coal.