Are natural gas wells in northeastern Pennsylvania contaminating the groundwater supply? Some researchers think so.

Scientists at Duke University detected elevated levels of methane, ethane and propane in groundwater samples near active fracking sites. The scientists conclude that the gasses come from the wells, not natural sources, but that the problem could be solved with better-designed casings. "We think there's a well-integrity problem in this part of the Marcellus,” says Robert Jackson, a professor at Duke and lead author on the paper describing the findings. "And well problems are relatively easily fixed. They’re especially easier to fix than if there's some fundamental problem with fracking."

Hydraulic fracturing—more commonly referred to as fracking—extracts natural gas reserves that are unreachable by conventional techniques. The process requires drilling wells thousands of meters down and then blasting them with a mix of water, sand and chemicals to induce fractures in the underlying rock. Gas seeps through these fractures and flows back up the well where it is captured.

But if the wells aren’t properly sealed, then gas can leak into the groundwater. The wells are lined with metal casings that prevent extracted gas and contaminated water from leaching into the surrounding rock. To block gas from flowing up the outside of the well shaft, engineers pour cement around the outer casing to plug any gaps. If the cement or casing isn’t properly set, then gas from deep shale deposits can find its way in to shallow groundwater. If the casing ruptures, fracking chemicals can also enter the water supply.

The researchers sampled well water from 141 homes in six counties. Many of the samples contained methane, but those wells within one kilometer of a gas well showed concentrations six times higher than average. Ethane in those nearby homes was 23 times above that of homes farther away. Ten homes also showed traces of propane.

The paper, published in Proceedings of the National Academy of Sciences on Monday, builds on a previous study done in 2011. The researchers added 81 homes to their data and expanded their analysis to identify the source of the stray gases.

There are many natural sources of methane, none of which have to do with fracking. Microbes, for example, produce methane as they break down organic matter. Shallow gas pockets also can find their way through underground fissures and into water supplies. But methane has a way of telling researchers where it came from. Its molecule consists of one carbon atom attached to four hydrogen atoms. But not all carbon atoms are alike; some are slightly heavier than others. These heavier flavors, or isotopes, of methane can tell scientists where the gas has been. Microbes, for example, favor lighter carbon atoms. Deep natural gas reserves, like the Marcellus shale, tend to be heavier. Isotope analysis showed gases that look like they come from the Marcellus Formation.

The appearance of ethane and propane also rules out microbes as a source in some cases. Biological sources of methane do not produce ethane and propane. Beyond one kilometer from the gas wells the background level of ethane and propane is almost zero. But within that radius, the level jumps dramatically. “Where we find higher methane concentrations,” Jackson explains, “sometimes it looks like natural background, sometimes it looks like Marcellus gas. When it’s Marcellus gas, that points to a problem with well casings.” He points out that the elevated gas level is not a criticism of fracking itself, but it does suggest that there are well problems and that these need to be addressed. “The place to start is well integrity.”

Fred Baldassare, owner and senior geochemist at ECHELON Applied Geosciences Consulting, disagrees. “I don’t get a really strong sense of where they collected the water samples from,” he says. Without that information, he explained, it is difficult to independently assess the Duke study’s conclusions. ECHELON investigates stray gas incidents for clients such as the Marcellus Shale Coalition and the Pennsylvania Department of Environmental Protection.

Lisa Molofsky, an environmental geologist at GSI Environmental, a consulting group that works with industry to manage and assess environmental risk, echoes Baldassare’s concerns. "How were these water wells selected?" she asks. "Are these from people who are already concerned? If so, there is a high potential for selection bias."

John Connor, president of GSI, also cautions against extrapolating to a wider problem. "They're making very broad statements about things that could be very local," he says, referring to other studies that do not show evidence for methane contamination on a wider scale. "If you happen to pick the one place where a meteor hit the planet, you'll think meteors hit all the time. Maybe they're just standing on the edge of a smoking crater."

Connor and Molofsky wrote a report in 2011 for the Oil & Gas Journal that challenges Jackson’s findings. They find methane to be not only common in groundwater in northeastern Pennsylvania, but also more concentrated in valleys as opposed to higher elevations. They argue that geology, not compromised well casings, is to blame for methane contamination. Houston-based Cabot Oil & Gas Corp., owner of numerous wells in the Marcellus shale region, supported the research. "It's very important to talk about this candidly," Baldassare says. "There have been incidents related to gas well drilling, I'll be the first to admit that.” In 2009, for example, the town of Dimock, Pa, garnered attention when one water well exploded after an electric spark ignited methane in the water. Thirteen wells in the area showed dangerously high levels of a wide range of contaminants, including methane. "But from 2009 to 2013 the way the industry operates is like day and night. They're working hard to improve well integrity."

Both Connor and Molofsky stress the importance of understanding the local geology and getting better baseline data. Whereas the ethane and propane data do seem to rule out a biological source, there are many ways natural gas can seep into the water supply. "It’s hard to say which hat a rabbit came from when all hats contain rabbits," Connor says. "We don't want to bash these guys,” Connor adds. “No one has this really figured out. Papers like this are important in moving the ball forward. They're posing questions, important questions. It's helpful to the overall process."

To move forward, researchers need water samples taken before wells are drilled. Pre-drilling data will establish a baseline for naturally occurring gases and help determine whether or not drilling is introducing methane into the water supply. Some of the homes in the Duke study were selected because they are in areas slated for wells in the next few years. "My motivation is to figure out why sometimes things go wrong and how to fix it," Jackson says. "We're not interested in shutting an industry down, we're interested in solving problems. We just want to make this better."