ProPublica’s Fuzzy Math
Absent any data, advocacy outlet attempts to conjure up its own lifecycle analysis of natural gas – EID takes a look at the method
In a debate that over time has seen its fair share of twists, turns and temporary detours from fact, the rejection on the part of shale gas opponents of the basic science supporting natural gas’s status as an efficient, clean-burning fuel source stands out as a particularly troubling case of ideology as a substitute for methodology. Indeed, that natural gas emits less carbon dioxide per unit of energy produced than all other fossil fuels isn’t an assertion of faith, policy or politics – it’s a reality of thermodynamics. And it’s a fact corroborated by EPA, EIA, MIT and even the Union of Concerned Scientists, just to name a few.
Of course, as one anti-shale activist cautioned in a recent email exchange, “we should not, and cannot solely depend upon science to guide us.” And on that point, it’s apparent the advocacy journalism outlet ProPublica heartily agrees. Earlier this week, ProPublica author Abrahm Lustgarten published a piece putting forth a series of back-of-the-envelope calculations on the “lifecycle” greenhouse gas (GHG) impact of natural gas. His contention? That natural gas might not be as “clean” as previously thought. And further: that “new research” from EPA actually confirms it.
Except that it doesn’t — precisely because no new EPA research on this question actually exists. So how exactly did ProPublica conjure an analysis without the service of a single new data point upon which to structure its case? Well, it’s complicated. But we’ll do our best to unravel it some.
Let’s start here: In April 2010, EPA published a paper that, among other things, tracks the steady decline of methane emissions from natural gas systems over the past 20 years – even as thousands of new wells (particularly shale wells) were drilled, and billions of additional cubic feet of natural gas (particularly shale gas) were delivered to consumers. The key table from the report is excerpted below, and to its credit, ProPublica at least includes mention of this study in the 35th paragraph of its 44-paragraph piece:
Candidly, this table is, and has remained, a constant thorn in the side of those who believe expanded shale gas development is deleterious to our climate. The reason? Previously, the only credible way to advance such an argument was to suggest the volume of natural gas leaked (or fugitively emitted) into the atmosphere during the development and transportation process is much greater than we thought. EPA, for its part, estimates that 1.4 percent of U.S. natural gas produced is lost somewhere between the wellsite and the customer. Most producers, which have a strong economic incentive to minimize such losses, believe that number’s too high. But whatever the actual figure, the table above indicates that from the well head, to the pipeline, to the processing unit, to the distribution network – the volume of methane that escapes into the air pursuant to natural gas operations in the United States continues to go down. By a lot.
Undeterred by the science, activists such as Robert Howarth, a professor at Cornell, have structured their entire case against shale gas atop the presumption that the published figures for methane leakage/fugitive emissions are too low. So how much methane (and therefore money) does Prof. Howarth believe producers of natural gas fritter away into the ether, to the detriment of both the environment and their bottom line? His lips had been sealed on this for the better part of the past year, but earlier this week, the professor quietly posted an updated two-pager on his website positing his theory that as much as 7.9 percent (!) of this cash-crop is wasted away each day. Of course, without the use of that staggeringly inflated figure, Howarth has no ability to argue that natural gas has a similar GHG profile as coal. And he knows it.
ProPublica knows it too – which is probably why it tacked in a different direction in advancing its cause this week. Working backward from the predetermined conclusion that natural gas isn’t as clean as scientists believe, ProPublica bases its lifecycle GHG analysis of natural gas upon a “technical support document” assembled by EPA as part of the agency’s new proposed reporting standards for oil and natural gas producers. Only one problem: Though the EPA document runs 144 pages long, the word “lifecycle” is found nowhere in the text. In other words, this EPA report, the source most critical to ProPublica’s case against natural gas on the basis of its lifecycle GHG performance, doesn’t include any lifecycle GHG data itself. Nor does it include any new data on the small percentage of methane lost or leaked along the bit-to-burner development and delivery process.
Here’s the thing about lifecycle analyses: Putting together a credible one is hard work. Just ask our friends in the corn business. More than five years into the process, with tens of thousands of man-hours committed and at least 40 separate technical tomes on the subject published, and EPA is still making tweaks to its lifecycle GHG read-out for ethanol.
Now let’s take a look at the ingredients involved in ProPublica’s lifecycle analysis of natural gas: Weeks of work instead of years, zero original data either acquired or introduced, an EPA technical document that at no point addresses the lifecycle question itself, and a six-page pamphlet by a researcher at Carnegie Mellon.
Now, in ProPublica’s defense, the EPA report at the center of its investigation isn’t entirely silent on the question of methane emissions from unconventional energy development – but the sources the agency uses to justify an increase in emission factors are, in fact, quite amazing. As per EPA’s explanation in Appendix B of this document (page 84), instead of gathering new data itself, the agency used a 2004 industry PowerPoint presentation to find a single data point (from 2002) from a single operator to set its baseline volume for the total amount of methane lost annually in the U.S. as part of the well completion and workover process. Once it had that, EPA simply multiplied that figure by the number of unconventional wells it assumes have been drilled (fracked, and re-fracked) over the past several years. Through that process, it eventually arrives at its new emission factors. But not without reminding the reader of this:
The estimated activity factors were multiplied by the associated emission factors to estimate the total emissions from well completions and workovers in the U.S. for 2007. This does not reflect reductions due to control technologies such as flares or bringing portable treatment units onsite to perform a practice called “reduced emission completions.”
In other words, the new nationwide emission factors put forth by EPA in November and then converted this week by ProPublica into a back-of-the-envelope lifecycle assessment don’t take into account the technologies currently being used by operators to reduce the amount of methane emitted into the atmosphere (and remember: they’ve got one hell of an incentive to do just that). And again, as EPA admits, the new figures don’t even attempt to characterize what’s happened in this arena since 2007 (or even 2002, since that’s where its data is derived from). For that information, we must again refer to the April 2010 EPA report – the one ProPublica doesn’t like – in which you’ll find the following statement:
The U.S. natural gas system encompasses hundreds of thousands of wells, hundreds of processing facilities, and over a million miles of transmission and distribution pipelines. Overall, natural gas systems emitted 96.4 Tg CO2 Eq. (4,591 Gg) of CH4 in 2008, a 26 percent decrease over 1990 emissions, and 30.0 Tg CO2 Eq. (29,973 Gg) of non-combustion CO2 in 2008, a 20 percent decrease over 1990 emissions.
Improvements in management practices and technology, along with the replacement of older equipment, have helped to stabilize emissions. Methane emissions decreased since 2007 despite an increase in production and production wells …
So, to recap: In the article published this week, ProPublica shares the results of its own lifecycle GHG analysis of natural gas based on 1) no new original data; 2) no new information on methane leakage and fugitive emissions; 3) an EPA report that does not, and was never intended to, speak to the question of lifecycle; 4) “updated” emissions factor research based on a single data point from 2002 which is then extrapolated out to characterize the entire U.S., and finally; 5) a six-page article from a Carnegie Mellon researcher. And you know what? Even with the aid of its Rube Goldberg analysis machine, ProPublica still can’t come up with a conclusion suggesting natural gas is anything less than between 25 percent and 40 percent cleaner than other fossil fuels – directly contradicting the work of Prof. Howarth at Cornell.
As we know, nothing truly transformational occurs in this world without picking up a few gainsayers along the way – and the more transformational the event, the more powerful the forces that are bound to ally against it. Thankfully, facts still mean something in this debate, and method still counts as a means of establishing credibility. You can’t structure a lifecycle analysis on the back of a napkin. But as we saw this week, that doesn’t mean you can’t try.
UPDATE (1/31/11; 4:31p EST): EPA releases statement on ProPublica piece
EPA has not conducted an analysis of coal versus natural gas, and there is no new report. The information referred to in the article was developed based on information from a Technical Support Document, however, which was developed as support for the Greenhouse Gas Reporting Program. The reporter used that data and did his own calculations to arrive at the figures used in the article.
The document above does not estimate emissions from the gas industry and the emissions estimates in the article were not developed by EPA. EPA has not reviewed the analysis described in the article in detail, but we have not seen any indication that the benefits of natural gas have been called into question. Available data demonstrate that switching from another fossil fuel to natural gas reduces emissions of carbon pollution and other harmful pollutants that threaten Americans’ health.
Director of Communications
Climate Change Division
U.S. Environmental Protection Agency