Experts Report Fracking Study Is Fatally Flawed

Reflecting clear biases and cherry picking of information, fracking study is "industry garbage in, industry garbage out" that does not hold up to close review.
From Physicians Scientists and Engineers for Healthy Energy
June 18, 2012
Add to My MSN

Slightly more than half of the homes in the United States use natural gas as their main heating fuel.

Content Tools

The following article is posted with permission from Physicians Scientists & Engineers for Healthy Energy.

The nonprofit Physicians Scientists & Engineers for Healthy Energy (PSE) today issued the following joint statement by Profs. Anthony Ingraffea and Robert Howarth, and research technician Renee Santoro of Cornell University: “We have analyzed the widely publicized report from the American Petroleum Institute (API) and American Natural Gas Alliance (ANGA) which asserts that methane emissions from the natural gas sector are 50 percent lower than US Environmental Protection Agency (EPA) estimates.

The study relies on a critically flawed survey design, completely ignores many other recent studies, and would not have passed peer-review in a scientific journal. In contrast to this API/ANGA report, a recent and objective study which measured the entire rate of methane emissions from an unconventional gas field, the first of what should become one of many such studies, demonstrated emissions that were higher than EPA estimates (Petron et al. 2012).

Fatally Flawed Survey Design

At its core, the API/ANGA study is based on a biased survey. Twenty oil and gas corporations provided answers to questions on the details of the procedures they use to develop and maintain gas wells. Amazingly, the survey in effect coached those being questioned, indicating what answers were sought. The cover-page instructions stated that the purpose of the survey was to evaluate EPA emission estimates. The survey then gave the EPA emission estimates so that respondents could gauge their answers. Given the huge economic interest of industry in projecting low methane emissions, a valid study would have taken steps to detect and counter-act potential bias, not blatantly encourage it.

Further, to be valid, a survey must be based on random and representative sampling, but there is no evidence that such an approach was followed. This is an elementary statistical error and the study would not have passed peer review had this report been submitted to a rigorous scientific journal. Only a small proportion of the wells included in the report were shale-gas wells using the current technology of high volume, slick-water, horizontal hydraulic fracturing of multiple long horizontal well legs.

Cherry Picking of Information 

Beyond the biased and non-random sampling of the API/ANGA study, the report is highly selective in the information used. Many times information is ignored, generally resulting in underestimated methane emissions. For example, the report cites data for the rate of completed wells drilled in 2010, indicating completion rates greater than those used by EPA, but then failed to use this higher rate in their emissions accounting. Had they used these higher completion rates, API/ANGA would have estimated a 64 percent greater rate of methane emissions at completion than EPA has estimated.

Another example of bias is the handling of well “workovers” by API/ANGA. The productivity of shale gas and other unconventional natural gas wells falls quickly over time after initial well completion, requiring repeated high volume hydraulic fracturing – a term called “workover.”  API/ANGA claim a 72 percent reduction in methane emissions from workovers compared to EPA estimates. To reach this conclusion, the report used a lower rate of workovers than that considered by EPA, based on the API/ANGA survey data. However, their survey data also indicated some 8,000 hydraulically fracked gas wells in the Gulf Coast and northeastern US regions which EPA had not included, and the API/ANGA study failed to correct for this counter-acting piece of information.

Methane is also emitted during “liquid unloading,” the process of removing from a well the fluids that can build up over time and reduce gas production. Here, the API/ANGA report claims an 86 percent decrease in emissions relative to EPA 2010 estimates, based on their industry survey data for the frequency of liquid unloading and an assumed rate of emission per unloading event. However, they do not document the derivation of this assumed rate. Using the rate from a 2010 study from the Governmental Accountability Office, we calculate methane emissions during liquid unloading that are about 50 percenrt higher than the API/ANGA estimate.

Although the API/ANGA study purports to present methane emissions for the entire lifecycle of natural gas, from well development to final use, they actually focus almost entirely on just the “upstream” emissions (those that occur at the well site). A complete analysis would also consider the “downstream” emissions that occur as gas is transported through high pressure pipelines, stored, and distributed through lower pressure urban and suburban gas-lines. For these downstream emissions, the API/ANGA report relies on a 1996 joint EPA-industry study. EPA too continues to rely on this study for their estimate of downstream emissions in the US national greenhouse gas inventory, although they are actively considering reanalyzing this as they have done for all other parts of the natural gas lifecycle emissions over the past 18 months.

Several studies have indicated that these downstream emission estimates are too low. In our analysis, we concluded – based on more recent information than was available at the time of 1996 EPA-industry study -- that the downstream methane emissions are at least 50 percent greater and perhaps 400 percent greater than in the EPA inventory (Howarth et al. 2011). If so, the downstream emissions equal or exceed the upstream emissions, suggesting that natural gas is a very poor energy choice from the standpoint of global warming unless society is willing to invest a huge amount of capital in tightening up gas transmission and distribution systems. The API/ANGA report failed to even discuss such a possibility,

Moving Forward

Most past efforts at quantifying methane emissions from the natural gas industry have relied on estimating venting and leakage for each process and summing the total, an inherently uncertain approach. The API/ANGA study followed this methodology, erroneously claiming to use more accurate data than EPA has.  Unfortunately, the bias underlying the study makes it almost useless. But fortunately, scientists are rapidly developing and applying more objective approaches. An excellent example is the 2012 peer-reviewed study by Petron and colleagues at NOAA and the University of Colorado. They measured the actual flux of methane from an entire unconventional gas field, demonstrating that the flux was somewhat higher than the EPA estimates, and considerably higher than most industry-connected reports, including the API/ANGA study.

More such studies are needed, evaluating emissions from both conventional and unconventional gas fields, and the downstream emissions from pipelines and urban distribution systems. Nonetheless, the best available science indicates that natural gas is not a bridge fuel. The global warming consequences of the resulting methane emissions are too great, giving methane a larger greenhouse gas footprint than any other fossil fuel.”


The nonprofit Physicians Scientists & Engineers for Healthy Energy (PSE) is dedicated to supplying vetted, evidence-based, scientific information and resources on unconventional gas development (high-volume hydrofracking) and other novel energy production methods. PSE's mission is to bring transparency to the important public policy issues surrounding such methods, helping to level the playing field for citizens, advocacy groups, media, policy-makers and politicians. For more information, go to on the Web.

Previous | 1 | 2 | 3 | Next

Subscribe Today!

Pay Now & Save 67% Off the Cover Price

(* indicates a required item)
Canadian subs: 1 year, (includes postage & GST). Foreign subs: 1 year, . U.S. funds.
Canadian Subscribers - Click Here
Non US and Canadian Subscribers - Click Here