Just The Facts: A Closer Look at Duke’s Methane Study

A clear, easily-anticipated pattern has emerged with regard to academic research surrounding Marcellus Shale natural gas development, specifically how the media – perhaps purposefully – neglects details of methodology and other important and proper context, resulting in highly provocative headlines that do not always reflect the actual facts. Such was the case this week, as researchers from Duke University released a study focused on water well quality and the existence of methane – or natural gas – in a non-random sampling of northeastern Pennsylvania private water wells.

Methane and other naturally-occurring contaminants in private water wells is a real, serious and longstanding challenge for many Pennsylvania communities – it’s a well-documented problem that pre-dates Marcellus Shale production activities. In fact, the Center for Rural Pennsylvania – a non-partisan legislative research arm of the Pennsylvania General Assembly – determined that nearly 20 percent of private water wells in the commonwealth contain dissolved methane. The Center also found that  approximately 40 percent of the water wells failed at least one Safe Drinking Water Act water quality standard before gas well drilling occurred.

This from Penn State University researchers, who worked with the Center for Rural Pennsylvania on this study:

  • PSU researchers collected pre- and post-drilling water sample from private wells
  • Collected and analyzed nearly 230 samples within 1,000 feet and within 1 mile of Marcellus wells
  • No significant before/after changes in water quality
  • ~40% of wells fail at least one drinking water standard and background methane found in ~24% of the wells

These are staggering figures, given the fact that there are more than 1 million private water wells in Pennsylvania. What is perhaps most critical in this context – especially as it relates to Duke’s methodology and findings, or lack thereof – is that the Center conducted both pre- and post-drilling analysis. The Duke study, of course, did not – and therefore cannot establish a baseline to scientifically determine whether methane levels in water wells increased, decreased or remained unchanged before or after drilling and fracturing operations.

But there are common sense solutions aimed at addressing decades-old public health challenge. The Marcellus Shale Coalition (MSC) worked with Gov. Rendell’s administration to modernize Chapter 78 of Pa. Code to further strengthen the well construction and casing requirements for Marcellus Shale wells. As a result of these changes, Marcellus wells now have at least four (4) strings of steel casing and layers of cement to protect ground water resources (Click here to learn more about well casing). Further, the MSC produced an educational video (“Methane: An Element of Nature”; available on our YouTube page) as well as several Recommended Practices (available here and here) on this important subject.

As you may know, the Duke study released this week is a follow-up to a paper issued by the same researchers in May 2011. In sum, that paper claimed that the closer a private water well is in proximity to a Marcellus Shale natural gas well in northeast Pennsylvania, the more likely it is that methane is present in said wells. That study sampled 60 private water wells (NOTE: Their sampling was not conducted at random, rather a “sample” water well universe was generated through word of mouth and other non-random means.) and was widely reported by local and national media, spurring more conversation about water well quality in Pennsylvania.

Again, and not to overstate the fact, Duke researchers did not perform pre- and post-drilling/fracturing analysis, thus they cannot – and have not – proved causality between levels of methane present in private water wells and natural gas production activities.

This latest Duke study evaluated 141 non-randomly selected water wells (60 of which were from the 2011 study) across northeastern Pennsylvania, with the researchers conducting several additional tests that were not conducted in their initial study. Their key finding mirrored their previous work: The closer a private water well is in proximity to a Marcellus Shale natural gas well, the more likely it is to contain elevated methane levels.

The additional tests performed focused on determining if the natural gas found in the non-random sampling of private water wells is thermogenic or microbial (i.e. biogenic). Thermogenic gas is created under an enormous amount of pressure over millions of years; this is what Marcellus Shale producers are producing. Microbial gas, however, is created through plant life and waste decay over much shorter periods of time. While both forms can be considered “naturally occurring,” additional research tools – isotopic analysis, for example – are used to fingerprint the gas and ultimately determine the gas’ originating rock formation(s).

Put simply, methane molecules have unique signatures – similar to fingerprints. The narrative advanced by Duke – accompanied by many in the media – is that if the methane found in a water well is thermogenic and its isotopic signature is similar to Marcellus natural gas, then said gas must be from Marcellus Shale wells. This is precisely where the Duke researchers (in addition to their non-random sampling universe, their lack of pre- and post-testing and their failure to release their complete data set) and many in the media get it wrong.

Isotopic reversals – while present in Marcellus natural gas – are commonly found in much shallower formations that also produce thermogenic gas. Having a deep and comprehensive understanding of the unique geology of northeastern Pennsylvania is key to understanding methane migration, as recently outlined in a U.S. Geological Survey (USGS) report of Sullivan Co. water wells. While the northern portion of Sullivan County was part of Duke’s analysis, USGS randomly sampled water wells in central and southern Sullivan Co. – areas without Marcellus development.

Key USGS findings:

  • 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. The water samples from these two wells had the highest concentrations of arsenic, boron, bromide, chloride, fluoride, lithium, molybdenum, and sodium of the 20 wells sampled. 
  • This study, done in cooperation with the Pennsylvania Department of Conservation and Natural Resources, Bureau of Topographic and Geologic Survey (Pennsylvania Geological Survey), provides a groundwater-quality baseline for central and southern Sullivan County prior to drilling for natural gas in the Marcellus Shale.

And late last year, a separate USGS study conducted in New York State’s Southern Tier – where horizontal drilling and hydraulic fracturing remains banned – also found high levels of methane in private water wells. Importantly, that USGS study – like the more recent one in Sullivan Co. – was based on randomly selected water wells, unlike Duke’s research.

Another peer-reviewed analysis (available in full here) of Susquehanna Co., Pa. water wells – published in May 2013 in Groundwater – found:

  • Testing of 1,701 water wells in northeastern Pennsylvania shows that methane is ubiquitous in groundwater, with higher concentrations observed in valleys vs. upland areas and in association with calcium-sodiumbicarbonate, sodium-bicarbonate, and sodium-chloride rich waters—indicating that, on a regional scale, methane concentrations are best correlated to topographic and hydrogeologic features, rather than shale-gas extraction.
  • In addition, our assessment of isotopic and molecular analyses of hydrocarbon gases in the Dimock Township suggest that gases present in local water wells are most consistent with Middle and Upper Devonian gases sampled in the annular spaces of local gas wells, as opposed to Marcellus Production gas. 
  • Combined, these findings suggest that the methane concentrations in Susquehanna County water wells can be explained without the migration of Marcellus shale gas through fractures, an observation that has important implications for understanding the nature of risks associated with shale-gas extraction.

To their credit, some in the media took the necessary time to fully evaluate Duke’s study and provide their readers with proper context. Here are several fact-based examples.

  • Eighty percent of all the water wells they tested contained some level of methane, including many with no nearby drilling. … The Duke researchers haven’t found any evidence that chemicals from hydraulic fracturing, or fracking, have contaminated water wells. … The situation is complicated because Pennsylvania has many layers of oil, gas, and coal-bearing rock as well as natural faults. All those can enable gas to seep naturally into water wells, even in areas without drilling. … Fred Baldassare, who worked for the Pa. Department of Environmental Protection for 25 years, said the study doesn’t present an accurate picture of the whole state because the Duke team went to areas where residents had complained about drilling contamination, rather than doing a random sample. (Associated Press, 6/24/13)
  • The second water study was published online last week by the U.S. Geological Survey. It found that some Pennsylvania water wells in areas with no nearby drilling are naturally contaminated with high levels of methane. It also found that 85 percent of the samples had radon levels higher than federal safe limits. One well sample, taken at a hunting club, had such high natural methane levels, it could have been flammable, said hydrologist Ronald Sloto. “They knew they had a major water quality problem, they didn’t know what it was,” Sloto said. The USGS took samples from 20 wells in Sullivan County, in northeastern Pennsylvania, in order to establish a pre-drilling baseline for water quality. … “Once you have drilling you can’t get a baseline, it’s too late” to determine if drilling caused water problems or if they were already there naturally, Sloto said. [NOTE: Duke, again, did not pre-test and therefore did not establish a baseline to make a categorical determinations.] (Ibid, 6/24/13)
  • Researchers have also detailed the widespread natural occurrence of methane in the region’s water wells in places where drilling had not yet begun. Last week, the U.S. Geological Survey released a study of 20 randomly selected water wells in an area of Sullivan County without shale drilling and found that seven contained some naturally occurring dissolved methane with one containing enough of the gas – 51.1 milligrams per liter – to create a potential explosion risk. The methane in two wells was thermogenic – gas that comes from deep underground, not from the breakdown of biological matter near the surface. A recently updated study by Cabot Oil and Gas Corp. contractors and employees found detectable levels of methane in 78 percent of 1,700 water wells it sampled before drilling nearby in Susquehanna County. … It also found that high methane concentrations in water wells generally correspond to how close they are to valleys, not their proximity to gas wells. (Scranton Times-Tribune, 6/25/13)
  • The [Duke] research did not find evidence of contamination from the gas extraction process of hydraulic fracturing, or fracking. Fred Baldassare, the state’s former stray gas inspector, said the most notable aspect of the Duke paper is the identification of “post-mature” thermogenic gas – deeply buried gases similar to those produced from the Marcellus Shale – in some of the sampled water wells. Most cases of drilling-related methane contamination have traced the thermogenic gas in water wells to shallower rock layers, not the target shale. But he said that the Duke researchers picked water wells to sample in areas with known drilling-related gas contamination so it would be wrong to conclude from their data that water wells in close proximity to gas wells are likely to be tainted by gas. “That’s not taking into consideration the thousands of gas wells that have been drilled in that area,” he said. “The percentage that don’t have gas well migration incidents is far higher than those that do. If you read their paper you wouldn’t know that.” (Ibid)

While the natural gas industry absolutely welcomes and encourages broad scientific and academic research, certain “findings” – upon further and more detailed review – are not always as they as may seem or as newspaper headlines suggest. Such was the case (again) this week.