A Duke University study that examined the possibility that Marcellus Shale drilling in northeastern Pennsylvania contaminates drinking water concluded that pathways in rock formations that allowed salinated water into shallow aquifers were naturally occurring and not a result of hydraulic fracturing.

Still, the authors warned in the study published Monday in the Proceedings of the National Academy of Sciences that those naturally occurring pathways could allow chemicals and contaminated water caused by fracking also to travel into the drinking water supply.

Avner Vengosh, a Duke University professor of geochemistry and a corresponding author of the paper, characterized it as "good news, bad news."

"We're ruling out [that] this saline water derived from today's shale gas drilling," he said.

But, he continued, "everything is not black and white. We're just in the very beginning of understanding what's going on. The result of this study does not apply to all of Pennsylvania or all areas of the Appalachian Basin. It needs to be duplicated."

Katy Gresh, a spokeswoman with the Pennsylvania Department of Environmental Protection, said her office could not comment on the study because it had just seen it.

"We will review it," she said, adding: "We've never seen any evidence in Pennsylvania of hydraulic fracturing contaminating drinking water supplies."

The paper's conclusions that contamination was not caused by fracking were based on two major points:

The salinated water was not found in proximity to shale-gas wells.

The same type of brine water existed as long ago as the 1980s in samples taken by the U.S. Geological Survey.

Vengosh said the study - only the second peer-review done on drinking water contamination - focused more on water chemistry to understand the source of the water and its evolution. The chemistry worked as a sort of fingerprint or identifier to find its origin.

Researchers focused on a number of valleys in six counties in northeastern Pennsylvania - Bradford, Sullivan, Wyoming, Lackawanna, Susquehanna, and Wayne.

The participants analyzed 426 shallow groundwater samples and analyzed major and trace element geochemistry and compared them to 83 samples from underlying Appalachian brines in deeper formations from the region. The premise was to examine the possibility of fluid migration between the Marcellus foundation and shallow aquifers.

Terry Engelder, a professor of geosciences at Pennsylvania State University, is critical of the study's conclusions.

"I think what the Duke study did in terms of interpretation is overstepped," he said.

In addition, Engelder disagrees with the premise that fluids used in fracking eventually could use the naturally occurring pathways to contaminate shallow aquifers.

"When a well is fractured and starts to flow, that reduces the pressure inside the Marcellus, thus generating a pressure up the well bore," he said. "The gradient is reversed. The well then acts as a safety valve, relieving the pressure."

The act of drilling would preempt the salinated water from coming up.

"They don't have an explanation for that," Engelder said.

He does not disagree that the pathways are naturally occurring and that they have been there for hundreds of millions of years. But, he continued, are they effective for moving brine into the groundwater?

Because of the locations studied, Vengosh said, it is likely that valleys are at a potentially higher risk for contamination.

"We can use these tools for mapping more vulnerable areas where gas and saline water can flow into the aquifer," he said.

The implications of the study, Vengosh continued, are to continue to look for locations where there may be a hydraulic connection and monitor them more closely or avoid them altogether.

Marcellus Shale gas drilling occurs deep in the earth - one to two miles below the surface.

"This kind of depth does not guarantee there would be no hydraulic connection that gas or saline water could flow to a shallow aquifer," Vengosh said.

Kathryn Klaber, Marcellus Shale Coalition president, said the Duke study reinforces what her organization has contended all along.

"This research demonstrates that freshwater aquifers in northeastern Pennsylvania have not been impacted by natural gas development activities," Klaber said.

"And we agree that pre-drill baseline water sampling is critical, which is why our members continue to embrace this commonsense practice across all operations."