UPDATE (11:39 A.M. ET, 6/25/2013): This post originally suggested the Park Foundation financed the Duke study. Park funded the 2011 Duke study conducted by these same researchers — of which this latest version is an expansion — and also funds Duke’s Nicholas School of the Environment, which released the study yesterday. According to lead researcher Rob Jackson, however, Park did not directly fund this latest study. The post has been updated below to reflect that, and we apologize for the error.

—Original post, June 23, 2013—

In 2011, researchers from Duke University released a report that suggested methane was migrating into drinking water, likely as a result of shale development. The study contained a number of flaws (several conceded by the authors), including a lack of baseline data, the decision not to randomly sample wells, and the presence of high levels of methane in lots of water wells residing nowhere near natural gas wells. A series of peer-reviewed comments took issue with the Duke team’s conclusions, saying there was a “lack of data” to support their conclusions, and that the report “misrepresents potential risks” of shale development.

Subsequent independent studies have also characterized the presence of naturally-occurring methane in water wells throughout Northeast Pennsylvania as “ubiquitous,” further casting doubt on the Duke team’s ongoing campaign to link methane in water to resource development.

Two years removed from that first study, those same researchers from Duke are back on the case with a new study this week – once again suggesting that shale development is somehow contributing to elevated levels of methane in drinking water in Northeast Pa.  But dig a little deeper, and the thing you can’t help but notice is that this latest iteration is really no different from the first.

Here are the four most important things to know about the latest iteration of the Duke paper:

Fact 1: Methane in Water Wells Outside Drilling Proximity

The researchers found methane in 82 percent of the water wells they sampled. More than fifty of those wells – among the 141 total homes sampled – were nowhere near natural gas wells. That’s nearly equal to the number of homes that registered methane concentrations that were in proximity to such wells. At least one of those homes outside the researchers’ established proximity (one kilometer from a natural gas well) actually registered methane above the U.S. Department of the Interior’s remediation limit.

In other words: among the researchers’ samples, drinking water supplies that registered measurable methane concentrations were almost equally likely to be located either close to a natural gas well, or not close. Not exactly a smoking gun.

Fact 2: Thermogenic Fingerprinting Contradicted by USGS

The Duke team suggests that methane concentrations located close to natural gas wells tend to be thermogenic in origin, as opposed to biogenic (microbial). Translation: the Duke folks think that the source of the methane in question is from the Marcellus Shale (though, in fairness, they stop just short of stating that outright).

It’s an interesting theory, but one that’s also directly contradicted by a report released just last week by the U.S. Geological Survey. That report, which examined water wells in Sullivan County, Pa., found lots of thermogenic methane in the samples it collected. The wrinkle? The samples USGS collected were baseline samples, and thus predate drilling activity.
From the USGS report:

“Concentrations of dissolved methane ranged from less than 0.001 to 51.1 mg/L. Methane was not detected in water samples from 13 wells, and the methane concentration was less than 0.07 mg/L in samples from five wells. The highest dissolved methane concentrations were 4.1 and 51.1 mg/L, and the pH of the water from both wells was greater than 8. Water samples from these wells were analyzed for isotopes of carbon and hydrogen in the methane. The isotopic ratio values fell in the range for a thermogenic (natural gas) source.” (p. 1; emphasis added)

If the presence of thermogenic methane is what Duke researchers believe constitutes evidence of drilling impacting groundwater, then how can they explain the presence of thermogenic methane in areas of the exact same region before any drilling has taken place?

Of note: Sullivan County was also one of the areas from which the Duke team took its samples.

Fact 3: Still ‘No Evidence’ of Fracturing Fluids Impacting Drinking Water

One important fact about the first Duke study from 2011 – and something that the media decided was mostly unimportant – was that the researchers found no evidence of hydraulic fracturing fluids migrating from depth to contaminate groundwater. As the researchers noted in that 2011 paper:

“Based on our data (Table 2), we found no evidence for contamination of the shallow wells near active drilling sites from deep brines and/or fracturing fluids.” (p. 4)

In the latest study, the Duke team re-emphasizes that point, noting that there is “no evidence for increased concentrations of salts, metals, or radioactivity in drinking water wells accompanying shale gas extraction” (p. 1).

Fact 4: Still No Random Sample?

One of the quickest ways to undermine your own scientific findings is to engage in selection bias when establishing research parameters. This means, in the simplest terms, that folks attempting to craft a scientific argument should rely on random sampling, rather than carefully selecting one site over another, to remove any suspicion of deliberately using one set of evidence – to prove one particular conclusion – over another.

Which brings us to the “Methods” section of the Duke II paper:

“The samples were obtained from homeowner associations and contacts with the goal of sampling Alluvium, Catskill, and Lock Haven groundwater wells across the region.” (p. 5; emphasis added)

Does this mean the Duke team consciously selected particular water wells (or areas with historically bad water) for the purpose of padding the stats a bit in buttressing their conclusions? Fact is, without knowing who these vaguely described “contacts” are, we can’t really make a determination about the validity of the sample itself. That doesn’t mean the researchers cooked the books and only sought out households that the anti-shale activists in the area identified for them as good targets for the study.  But given the fact that Duke’s Nicholas School — the entity from which this study originated — is funded, in part, by the same organization that finances local “Gasland” screenings across the country, it’s hardly unreasonable to expect these researchers to tell us how they decided to choose one household for the study over another.

So, in the end, the researchers found methane in virtually every water well they sampled, irrespective of proximity to a natural gas well. They suggest these methane concentrations may be linked to Marcellus wells, but pre-drill testing in the exact same part of the state by the federal U.S. Geological Survey directly contradicts that thesis. And, not for nothing, there’s still no evidence of hydraulic fracturing fluids migrating from depth to contaminate aquifers. On that point, at least, we’re happy to agree with them.