Natural gas releases less CO2, but it releases more methane in the shipment process, which is far more potent as a climate change agent than CO2. Depending on which estimates you buy into on the amount of methane lost, natural gas may be worse than coal for climate change.
It's not quite that simple because the overall impact of an individual compound relies on both its atmospheric residence time and infrared absorption. This article[1] covers the basics.
That doesn't really contradict adrianN's point in any way. If you integrate methane's heat retention over now to 100 years from now it still beats carbon dioxide. That doesn't mean that if you integrate from 50 years from now until 100 years from now it would still win. Much less so if you look at a geological time period in the future.
What do you think happens to the methane in the atmosphere? It doesn't vanish. Most of it reacts in the statosphere and troposphere and oxidizes into ... you guessed it: Carbon Dioxide.
10% gets eaten by organisms in the soil, but at least 90% turns into CO2.
Of course, Natural Gas thermal combined-cycle electricity generation is MUCH more efficient than Coal, no matter which way you look at it. (Unless you have an old already-paid-for coal plant next to a coal mine and only care about dollars)
There's still other reasons natural gas is a better fuel, though. Plants come on & offline more quickly, there's no need for emissions scrubbers & no truly noxious combustion byproducts. There's also no steam output or water consumption, though as there is still waste heat maybe that doesn't matter.
There's also potential to get natural gas plants up to 60-80% efficient, which means still less CO2 production compared to a 30-40% for coal.
The other thing about combined cycle plants is they are cheap. A big worry for anyone building a power plant is the long term economics, especially capital costs. Obviously you want the plant to run profitably long enough to pay off the bonds.
For Combined cycle low capital costs means the financial risk involved is much lower. Much more likely it'll pay off the capital costs. After that if the economics get dodgy it can be run intermittently or mothballed. IE, the utility isn't stuck paying off bonds on a plant that is shut down.
Flip side is coal fired plants have higher capital costs, longer payback and high financial risk. If carbon taxes make your plant uneconomic to run before the capital costs are paid off, the owners take a bath. (Nuclear and solar plants have the same issues, very high initial capital costs)
For apple to apple comparison Green House Gases values are converted to CO2e (CO2 equivalent). This was a standard practice when I was in the industry. The article it self does not mention it, But I am pretty sure the graphs have CO2e values not CO2.
