Pop-sci: Natural Gas Power Plants Emit up to 120 Times More Methane Than Previously Estimated (https://www.desmogblog.com/2017/03/20/natural-gas-power-plants-fracking-methane)
Researchers at Purdue University and the Environmental Defense Fund have concluded in a recent study that natural gas power plants release 21-120 times more methane than earlier estimates.
Published in the journal Environmental Science and Technology, the study also found that for oil refineries, emission rates were 11-90 times more than initial estimates. Natural gas, long touted as a cleaner and more climate-friendly alternative to burning coal, is obtained in the U.S. mostly via the controversial horizontal drilling method known as hydraulic fracturing ("fracking").
The scientists measured air emissions at three natural gas-fired power plants and three refineries in Utah, Indiana, and Illinois using Purdue's flying chemistry lab, the Airborne Laboratory for Atmospheric Research (ALAR). They compared their results to data from the U.S. Environmental Protection Agency's (EPA) Greenhouse Gas Reporting Program.
"[Methane is] a better fuel all around as long as you don't spill it," Paul Shepson, an atmospheric chemistry professor at Purdue, said in a press release. "But it doesn't take much methane leakage to ruin your whole day if you care about climate change."
The researchers were careful to differentiate between emissions related to natural gas combustion versus leakage, with the latter found to be the primary source of methane emissions in this small, preliminary study. Previous estimates of methane emissions were reported to the EPA from the facilities themselves and were restricted to what came out of the smokestack, which means they excluded leaks from equipment such as steam turbines and compressors.
The study was done as part of EDF's ongoing series of studies measuring methane emissions and leakage throughout the U.S. natural gas supply chain. EDF said in its press release that the Purdue scientists plan to follow up with research at additional oil refineries and power plants. Purdue stated in a press release that support for the research also came from the National Science Foundation (NSF).
Paper: Assessing the Methane Emissions from Natural Gas-Fired Power Plants and Oil Refineries (http://pubs.acs.org/doi/full/10.1021/acs.est.6b05531) (open access).
Methane combustion is ~120 lbs CO2/MMBtu. Coal varies but is ~225 lbs CO2/MMBtu. However, direct Methane emissions to the atmosphere are 20x to 25x (depending on whose number you use) the effect of CO2.
Since spilled coal doesn't contribute to emissions but spilled Methane does there is a certain loss rate where Methane emissions (combustion + leaks) equal coal, and that is around 5%.
The EPA doesn't calculate spills/leaks in their emission profile unless it's reported, and there is no real rule to report minor leaks. The emission of methane from combustion is pretty small (part per million) so a UAV doing sniff tests above an operating turbine WILL pick up leaks and that is why you have this large multiplier for methane emissions (120x reported). I would guess a leak rate of 0.5% to 1.0% total in a large plant that isn't well maintained, but you can easily cut that rate by 50% to 75% with good maintenance. If you have mercaptin traced natural gas then detection is on the order of 0.05% (500 parts per million) versus a non-scented gas stream where a gas detector is 10x less sensative.
The point (IMO) isn't that some methane will leak. It's obviously going to be impossible to have exactly 0% methane leakage. There will always be some. What's more concerning is that the figures previously only included exhaust gas from the turbines, and didn't cover the rest of the system, so weren't telling the whole story. Then you have Trump wanting to cut funding and monitoring for this sort of thing, which is obviously going to result in lower standards (because it always does) and therefore more leakage.