Still Not Worse Than Coal

At the end of last year I examined assertions by a professor from Cornell University, based on his unpublished paper, that leakage from natural gas production and transportation systems in the US resulted in lifecycle emissions for gas that were actually worse than those from coal. From what I saw at the time, I couldn't agree with his conclusions. Now Professor Howarth's paper is apparently about to be published, with a specific focus on shale gas. It has already been leaked via the New York Times and The Hill news site. After seeing the data and calculations supporting its claims, I am still not persuaded, though I would be quick to concede that the subject deserves a more thorough assessment by a body actually equipped to gather the necessary data and process it rigorously.

I don't make a habit of reviewing scientific papers, but this one begs for a critique, for two reasons. First, it's appearing in the middle of a crucial national debate on the potential risks of the techniques involved in unlocking the potentially game-changing shale gas resources that have been found in the US and elsewhere around the world. What better way to make those risks--which I believe to be entirely manageable--seem not worth taking than by portraying shale gas as having more adverse environmental consequences than the chief fuel its supporters see it displacing: coal. So at a minimum the paper demands careful scrutiny because of its potential significance to the debate surrounding the largest energy opportunity the US has uncovered in decades.

In addition, practically every paragraph includes an assumption, simplification or choice by the authors that tends to increase the calculated environmental impact of natural gas. Whether that's the result of bias or merely a series of judgment calls, it undermines confidence in the final conclusions at the same time it amplifies them. I'll focus on the most significant of these decisions and forgo the questioning of many individually less-important, though still cumulatively consequential details for others better equipped to tackle them.

Probably the most significant choice the authors made was to emphasize the global warming impact of methane (the main component of natural gas) over a 20-year period, in preference to than the more commonly used 100-year interval. Then they bypassed the established Global Warming Potential (GWP) factors from the UN IPCCC's Fourth Assessment Report to use much higher factors for methane from a 2009 paper published in Science. I'll leave the angels-on-a-pin debate over this to the climate scientists, but I don't believe you need a Ph.D. in atmospheric physics to understand that if the outcomes of climate change will truly be determined in the next 20 years, we are already cooked. The world can't get global emissions down by enough, fast enough, to solve the problem on that time scale, at least not without a global economic shock that would return hundreds of millions of people to poverty. So when I recalculated the paper's estimate on shale gas emissions, I did so using the consensus 100-year GWP for methane of 25--less than 1/4 of the one on which the paper's scariest results rely.

The other major choice the authors made was to ignore the downstream conversion of gas and coal into electricity. As lifecycle analysis, this earns a failing grade. It's like comparing the overall emissions of a Nissan Leaf and Ford Explorer by focusing only on what happens upstream of the battery charger and the fuel tank. The authors dismiss this by saying that "this does not greatly affect our overall conclusion". That's wrong, not least because it's precisely the comparison of how gas and coal actually compete with each other that matters most here.

On the basis of these two points alone, the paper's conclusions crumble, even with the inclusion of supposed methane leakage rates from shale gas production that would have any engineer worth his or her salt scrambling to redesign the equipment so as to capture so much valuable "lost and unaccounted for" output. So how do shale gas and coal compare, on a full lifecycle basis from well and mine to the power plant bus bar, if 3.6-7.9% of gas actually leaked out during well completion, processing, transportation, storage and distribution, as Dr. Howarth's paper suggests?

Let's start at the power plant and work backwards. A current combined-cycle gas turbine unit requires around 6,700 BTUs of gas to generate a kilowatt-hour (kWh) of electricity. At the rate of 117 lb. of CO2 emissions per million BTUs of gas burned, that yields power plant emissions of 0.78 lb/kWh. But that's on the basis of the gas that reaches the turbine's combustor. We have to gross up that result to account for the emissions that occurred upstream of the plant. At Howarth's estimated leakage midpoint of 5.75%, and using the standard 100-year GWP for methane compared to CO2 on a molar, rather than mass basis, that leakage would add an extra 55% of CO2-equivalent emissions from the well to the combustor, bringing the effective emissions from that combined-cycle plant up to 1.2 lb/kWh. For comparison, the most efficient coal-fired power plant I know of (without carbon capture and sequestration) emits about 1.75 lb/kWh. Only if we included inefficient, simple-cycle gas "peaker" units that don't normally compete with coal would the upstream emissions that Dr. Howarth posits result in lifecycle emissions from gas-fired power worse than the typical coal-fired generation emissions of around 2 lb/kWh. In other words, the gas-fired generation that actually competes with existing coal plants still appears to emit nearly 40% less GHGs than its coal competition, even assuming the shale gas leaks that Dr. Howarth and his contributors reported.

Although my analysis admittedly falls into the back-of-the-envelope category, I'm not sure that the Howarth, et al paper is many notches above that level, given its reliance on non-peer-reviewed sources and its references to irrelevancies like Soviet-era gas systems. All in all, it seems a shaky edifice on which to mount such provocative conclusions. Perhaps all the authors wanted to do was to highlight some areas for the gas industry to investigate further, in order to ensure that methane emissions are kept to a minimum as shale and other unconventional gas deposits are developed. Unfortunately, it seems all too likely that its headline findings will be touted by those who are determined to stop the shale gas revolution in its tracks, or at least delay it for long enough that its utility in addressing our pressing energy problems will be lost. I wonder what Mr. Pickens thinks about all this, given that legislation promoting his plan to convert portions of the US truck fleet to natural gas, which depends on abundant shale gas supplies, has finally attracted bi-partisan support, including from the White House.