This week the U.S. Congress heard testimony supporting a bill that would push to replace diesel with natural gas in heavy vehicles. It’s an attempt to cut oil imports, and at the same time reduce greenhouse gas emissions. Part of the argument is that natural gas is substantially cleaner than diesel, and results in the emission of about 25 percent less greenhouse gas.

But experts are warning that natural gas might not be as clean as it seems.

In fact, using natural gas rather than diesel in vehicles could actually increase climate change, says Robert Howarth, professor of ecology and environmental biology at Cornell University. “You’re aggravating global warming more if you switch,” he says.

Howarth is basing his conclusion on a preliminary analysis that includes not only the amount of carbon dioxide that comes out of a tailpipe when you burn diesel and natural gas, but also the impact of natural gas leaks. Methane, the main component of natural gas, is much more effective at trapping heat than carbon dioxide, so even small amounts of it contribute significantly to global warming. When you factor this in, natural gas could be significantly worse than diesel, he says. Using natural gas would emit the equivalent of 33 grams of carbon dioxide per megajoule. Using petroleum fuels would emit the equivalent of just 20 grams of carbon dioxide per megajoule.

Howarth goes further, suggesting that natural gas could even rival greenhouse gas emissions from mining and burning coal–the dirtiest of fossil fuels. He says it’s “not significantly better than coal in terms of the consequences of global warming” and is calling for a moratorium on extracting natural gas from shale, which requires more energy (and so emits more greenhouse gases) than extracting it from conventional natural gas sources.

Howarth’s analysis, however, is just a preliminary one. He’s already found one major error in his original calculations. “I blew it,” he says, by not including the impact of methane leaks from coal mining. (Here’s a link to his original, which contains the error; and here’s the updated version). But he still says the gap between coal and natural gas is far smaller than generally thought. And his numbers are significantly different than those researchers at MIT came up with a year ago. (On a CO2 equivalent grams per megajoule basis, they scored diesel at 10.7 and gasoline at 14.4, with natural gas splitting the difference at 12.5). The two studies make different assumptions about the strength of methane as a greenhouse gas, and the amount of methane leakage, for example. A complete analysis should also look at the different efficiencies of natural gas and gasoline or diesel vehicles. The MIT study concludes that there is a benefit from switching to natural gas, all told, but it might not be worth the cost or the hassle. Making more efficient gasoline and diesel vehicles might work better, and be a faster way to reduce greenhouse emissions, it suggests.

But for all the shortcomings of Howarth’s analysis, it points to a real need. Before Congress passes any bill promoting natural gas, a thorough study of the potential impact needs to be taken into account, including the energy it takes to obtain it, and the impact of methane leaks.

Otherwise the U.S. might end up subsidizing something that does little to reduce carbon dioxide emissions–as happened with corn ethanol.