Some Thoughts on the Howarth Shale Gas Paper

A new paper by Cornell’s Robert Howarth, which claims that shale gas is worse for greenhouse gas emissions than coal, has been getting a lot of attention in the popular press. Howarth’s basic question is an important one: what happens to the claimed emissions benefits of natural gas once you include the methane leaked in its production and transport? Alas, his analysis is based on extremely weak data, and also has a severe methodological flaw (plus some other questionable decisions), all of which means that his bottom line conclusions shouldn’t carry weight. But someone else, with better data and more careful calculations, ought to address this important set of questions that he raises properly.

I won’t catalogue every problem with the study – it’s more useful, I think, to flag the biggest issues. I see four.

First, the data for leakage from well completions and pipelines, which is where he’s finding most of his methane leaks, is really bad. Howarth used what he could get – figures for well completion leakage from a few isolated cases reported in industry magazines, and numbers for pipeline leakage from long-distance pipelines in Russia – but what he could get was very thin. There is simply no way to know (without access to much more data) if the numbers he uses are at all representative of reality.

Second, Howarth’s gas-to-coal comparisons are all done on a per energy unit basis. That means that he compares the amount of emissions involved in producing a gigajoule of coal with the amount involved in producing a gigajoule of gas. (Don’t worry if you don’t know what a gigajoule is – it doesn’t really matter.) Here’s the thing: modern gas power generation technology is a lot more efficient than modern coal generation, so a gigajoule of gas produces a lot more electricity than a gigajoule of coal. The per kWh comparison is the correct one, but Howarth doesn’t do it. This is an unforgivable methodological flaw; correcting for it strongly tilts Howarth’s calculations back toward gas, even if you accept everything else he says.

Third, the problems with gas that Howarth flags have cheap technological fixes (green well completion techniques, better pipeline care), though there may be institutional barriers to implementing them. If we scale up gas and realize we have an emissions problem, there are things we can do. The only technological fix for coal, in contrast, is CCS, which isn’t commercial yet; if we decide we want to fix our coal problem, it’s not clear we have any options.

The fourth question, which has been getting a lot of attention (perhaps most of it), is Howarth’s decision to use 20 year global warming potentials (GWPs) to compare coal with gas, rather than the customary 100 year figures. Basically, the purportedly high lifecycle emissions of gas are due to methane leakage. Methane is a potent greenhouse gas, but decays in the atmosphere on scales of decades, in contrast with carbon dioxide, which decays on the century scale. This means that if you average the impact of GHG emissions over 20 years instead of 100, it boosts the relative influence of methane, and hence the downsides to gas.

Lots of people have been going after Howarth for his unorthodox approach on this front. I have to admit, though, that this is one where I’m not so sure that he’s as wrong as his critics claim. I’ve been a big fan, for example, of efforts to cut near term emissions by targeting black carbon and ozone – but you can’t really justify the value of such measures unless you look at short term GWPs, since these are very short lived species. And given a lot of the rhetoric out there about nearish-term tipping points, it also isn’t entirely clear to me that it’s consistent to turn around and say that we should only look at impacts averaged over a hundred years. Over at NRDC’s blog, Dan Lashof (who I hear wrote his PhD dissertation on GWPs) suggests that a 50-year timeframe is best. This makes as much sense as anything else. [UPDATE: In response to some email feedback, let me clarify: I’m not saying that a 20 year GWP is wise, or that a 50 year one is. All I’m saying is that this is a judgment call, and that one shouldn’t assume that a 100-year time frame is the best way of looking at every problem.]

One last comment: I worry about what this paper says about the peer review process and the way the press treats it. This article was published in a peer-reviewed journal that’s edited by talented academics. It presumably got a couple good reviews, since its time from submission to publication was quite short. These reviewers don’t appear to have been on the ball. Alas, this sort of thing is inevitable in academic publishing. It’s a useful caution, though, against treating peer review as a mark of infallibility, as too many in the climate debate – both media and advocates – have done.