As I noted in last Wednesday's posting, one of the questions that came up in a webinar on shale gas in which I participated concerned the climate consequences of higher recent estimates of methane leakage from US natural gas systems. In reading further comments and blog postings on this subject, I was surprised to see assertions that went beyond drawing attention to the importance of the leakage of a high-value, high global-warming-impact gas, to suggest that the apparent rate of leakage renders the lifecycle emissions from natural gas as bad as those from coal, or worse. If that were true, it would have significant implications not only for the development of shale and other natural gas resources, but also for our entire emissions reduction strategy. From what I can tell, however, such claims have not been substantiated by current studies.
Several comments I received in email or on the posting pointed to the work of Professor Robert Howarth of Cornell University, and specifically to a press release describing a paper he has apparently submitted addressing the climate impact of methane leaks from shale gas production, transportation and storage. Until the details of the paper are available, the information provided in the press release simply doesn't stand on its own or merit further analysis. In the meantime, a recent EPA report evaluating greenhouse gas emissions from the oil and gas industry identifies significantly higher estimates for methane emissions from natural gas systems than those incorporated into that agency's most recent US Greenhouse Gas Inventory. I became aware of the EPA report in the course of reading one of the blog postings I alluded to above.
The EPA estimated the total CO2-equivalent methane leakage from the production, processing, transportation, storage and distribution of natural gas in the US in 2006 at 261 million tons per year. That amounts to more than 4% of total net US emissions for that year, so it is hardly insignificant. It's also about 2.5 times the figure reported in the agency's latest GHG inventory. Converting that quantity back into natural gas at normal conditions yields 656 billion cubic feet of gas, or 3.4% of marketed US natural gas production in 2006. That's a lot higher than typical leakage estimates of less than 1%, as David Lewis notes in his blog. The question is whether this higher level of leaks, or some even higher notional level of leaks proposed by other critics, would be sufficient to make the emissions from gas worse than those from coal.
To understand why that might even be possible, you have to know something about the relative strength of different greenhouse gases (GHGs). While much of the public's attention has been focused on CO2, the most prevalent man-made GHG, other gases have dozens or hundreds of times the impact on climate, per ton. Because of the way it decays in the atmosphere, methane's global warming potential (GWP) starts high and diminishes over longer time spans. Most reports, including the EPA's, use a 100-year GWP estimate indicating methane is around 21 times worse than CO2.
However, it's not correct to infer from that that upstream leaks of 3.4% of all natural gas must therefore inflate the lifecycle emissions of the gas we consume by 21 times 3.4%, or 71%. That's because a ton of methane doesn't convert to a ton of CO2 when burned; it yields 2.75 tons, as a result of basic high school chemistry:
CH4 + 2O2 --> CO2 + 2H20
So for each ton of natural gas, it's roughly 7.6 time worse for it to be vented or leaked than burned, after adjusting methane's standard GWP for the ratio of molecular weights from the above reaction equation. In fact, when I added the EPA's latest methane emissions estimates to their figures for indirect and direct CO2 emissions from natural gas in the GHG inventory, the result was very close to the 26% increase you'd get from multiplying 3.4% by 7.6. As a result, although the emissions advantage of natural gas over coal is less than it would be without such a high rate of leakage, gas still emits 35% less CO2 equivalent per BTU over its lifecycle than coal, on average.
When you consider how natural gas actually competes with coal, its effective emissions advantage should be larger than that. Even after accounting for upstream emissions (including leakage) that add 30% to its CO2 emissions from combustion, an efficient combined-cycle power plant still generates electricity with emissions per kilowatt-hour that are more than 40% lower than those from a highly-efficient coal plant. That's because the combined cycle turbine converts more than half the BTUs in its fuel into electricity, while the coal plant converts less than 40% of coal's BTUs into power. Fewer BTUs for the same output results in fewer emissions.
I don't claim my back-of-the-envelope analysis is definitive, but it certainly doesn't support the notion that gas is worse than coal. Barring conclusive evidence of a much higher level of upstream natural gas leakage than indicated by the EPA's latest work on the subject, natural gas--even with existing infrastructure--could reduce the emissions associated with coal use in power generation by at least a third, and by much more than that depending on the specific generating facilities involved. At the same time, that shouldn't be read as excusing avoidable leaks of gas. If that 3% figure is accurate or low, then several billion dollars worth of gas--even at today's depressed prices--is escaping into the atmosphere rather than being captured and turned into useful energy by gas customers. That sounds like the epitome of low-hanging fruit to me.