Even though burning natural gas produces approximately half the carbon dioxide emissions of burning coal, a new study suggests that replacing 50 percent of the coal burned for electricity with natural gas by 2050 would have little effect on the world’s climate for many decades. The study, conducted by the U.S. National Center for Atmospheric Research (NCAR), will be published next month in Climatic Change Letters.

The key point is that carbon dioxide emissions, which trap heat, are not the whole climate story. While burning less coal would indeed reduce these emissions, it would also reduce emissions of other pollutants, like sulfur dioxide, which cool the planet.

Then there is methane. Natural gas, for the most part, is methane, itself a greenhouse gas over 20 times more potent than carbon dioxide at trapping heat in the atmosphere. Recent analyses, including by the U.S. Environmental Protection Agency, have suggested that the amount of methane that leaks into the atmosphere during unconventional gas production has been generally underestimated. Just how much methane escapes from production facilities is currently a subject of scientific debate, which has produced widely varying estimates.

In the new study, author Tom Wigley, an NCAR senior research associate, ran a series of computer simulations, accounting for a range of leakage scenarios that span the recent published estimates. He found that even if gas production facilities leaked no methane, temperature would rise, during the shift from coal to gas, until about 2050. If larger estimates of leakage are correct, the temperature could keep rising until several decades into the 22nd century. In all cases, according to the simulations, the changeover to gas will eventually begin to reverse the warming trend, but only at small rate.

In interpreting this study, it is important to keep in mind that a large amount of uncertainty clouds estimates and forecasts of methane emissions from unconventional natural gas production. Wigley also admits that it was difficult to quantify the long-term effect of sulfur dioxide emissions, given widely varying sulfur contents in coal, and because of current and future laws regulating such pollution.

Even so, the results are a useful reminder that climate science is more complicated than which fuel emits less carbon dioxide. A large-scale replacement of coal for gas would likely result in both less warming and less cooling—regardless of how clean natural gas is.