In 2015, the U.S. Environmental Protection Agency released a report finding 457 fracking-related spills in eight states between 2006 and 2012. Last month, a new study tallied more than 6,600 fracking spills in just four states between 2005 and 2014. But, as usual, the numbers only tell part of the story.

Not every spill counted in that new number represents a spill of potentially harmful materials or even a spill that made contact with the environment. In fact, the study’s goal wasn’t to tally an absolute number of fracking spills. Instead, researchers set out to collect available spill data and then drill down (no pun intended) into the details to unearth common patterns and characteristics. And it’s those commonalities that can reveal the larger story of how to prevent such spills — which often contain health-harming chemicals — from happening in the first place.

“When you look across spills, what are the risk factors, in what stage of a well’s life are you most likely to see a spill, are we more likely to see a spill from a well that’s already experienced one, are there changes in the law or in enforcement that drive more spills,” asked study co-author Kate Konschnik, founding director of the Harvard Law School’s Environmental Policy Initiative. “We wanted to see what the larger story told us about risk.”

The study, which was published in February in Environmental Science & Technology, comes from a working group convened by the Science for Nature and People Partnership and is part of a larger line of research trying to assess the risks that unconventional oil and gas (UOG) production, commonly referred to as fracking, pose to water resources. For instance, Konschnik and her colleagues published a different study in December that assessed the environmental risk of UOG spills by determining their distances from nearby streams. In the more recent study, Konschnik told me that there were two overriding goals: to look for trends in spill data and to tease out what kinds of spill data may be most useful in making UOG development safer.

To conduct the study, Konschnik and colleagues analyzed spill data from 2005 to 2014 at more than 31,400 UOG wells in Colorado, New Mexico, North Dakota and Pennsylvania. They included spill data related to the full UOG production cycle, including storage and transportation, rather than focusing only on the fracturing stage. The study only included UOG production wells, not fracking disposal wells, which are used to store the often chemical-laden wastewater that comes back up to the surface during drilling. Researchers tried to include 11 other states in the study, but the data was either incomplete or too difficult to access.

Overall, researchers found 6,648 spills across the four states during the nine-year study period. That number exceeds the EPA findings by so much because the study included the entire fracking life cycle, whereas EPA only examined spills explicitly related to the fracturing stage. (“UOG is growing in scale and intensity…and a fairer examination of risk is to look at releases throughout (a well’s) entire life,” Konschnik told me.) North Dakota reported the most spills and the highest overall spill rate at about 12 percent. Pennsylvania reported 1,293 spills (4.3 percent), New Mexico had 426 (3.1 percent) and Colorado had 476 (1.1 percent). Across the four states, wells that experienced multiple spills contributed to a larger proportion of spills, indicating that prior spills may be an indicator of future spills.

Researchers also examined yearly spill rates, finding that fluctuations were “likely” shaped by changes in state reporting requirements, “demonstrating how state policies directly impact efforts to identify and accurately assess UOG risk, their causes and potential mitigating remedies.” For example, when North Dakota switched reporting requirements from verbal to written, spill rates increased by up to 4 percent. And in Pennsylvania, annual spill rates increased as more inspectors were hired, the study found. Across all four states, the greatest spill risk occurred during the first three years of a well’s existence.

Spill volumes ranged from 1 gallon to up to 991,000 gallons. In addition to 46 freshwater spills, the total volume of spills associated with fracking chemicals, solutions and flowback ranged from more than 99,000 gallons in Pennsylvania to more than 203,000 gallons in Colorado. The most common pathways for spills, according to the study, were storage tanks and pits as well as flowlines. Spills related to transportation were also prevalent across the four states, with most associated with loading and unloading. As for what caused the spills, only Colorado and New Mexico explicitly asked for such information during reporting. In examining the available causal data, researchers found that human error and equipment failures were the most common culprits.

One of the study’s biggest takeaways was the importance of data reporting as well as the challenge of varying reporting requirements. For example, in Colorado, reporting requirements are triggered for any fracking spill of 42 gallons or more that escapes secondary containment. While in New Mexico, reports are required for spills greater than 25 barrels or if an operator thinks a spill might endanger water quality or public health. Study co-authors Konschnik, Lauren Patterson, Hannah Wiseman, Joseph Fargione, Kelly Maloney, Joseph Kiesecker, Jean-Philippe Nicot, Sharon Baruch-Mordo, Sally Entrekin, Anne Trainor and James Saiers write:

Further improving reporting requirements and processes for reporting will facilitate states’ and companies’ efforts to identify risks for certain types of spills and take action to mitigate some of the identified risk factors. To the extent that this information is publicly available and searchable, operators can use it to remove or mitigate risk factors to improve environmental performance and avoid higher insurance premiums.

Assembling these data electronically within a centralized database would allow state regulators and other stakeholders to identify trends, including the most common spill pathways and causes, as well as identify the wells or operators associated with unusually high spill rates. Making this information publicly available and providing it in an easy, usable format would allow operators, insurance companies, and citizen monitoring groups 
to assess the largest and most prevalent risks and respond accordingly. This paper illustrates the benefits of having 
available and accessible data.

Konschnik said that “without question,” the study reveals that many spills are likely preventable. For example, she said enhanced training or simple reminder signage could help prevent the human errors underscoring a significant portion of spills identified in the study. Other interventions are even simpler. For instance, she said the study’s data indicated that wildlife caused some of the spills, which could mean operators simply need to fence off areas to reduce spill risks.

As for the health hazards of such spills, this study doesn’t address that question. But Konschnik did say that current data — and, of course, more robust datasets — could help pinponit areas where public health monitoring is needed.

“Our data can be used as an indicator of where more research can be done,” she said. “If we had more robust data that was publicly available, you could dig much deeper…this is one piece of the puzzle in which a more granular view of spills data married with some community health assessment data and monitoring data could help determine whether or when there are risks to exposure.”

For residents living in fracking regions, finding spill data on one’s own can be quite difficult, Konschnik said. As such, she and her colleagues created an interactive website anyone can use to learn more about fracking spills and their causes — check it out here.

“UOG really is the wave of the future — that’s where we’ll see growth,” Konschnik told me. “And so these spills might be more representative of what we’ll see in the future.”

For a full copy of the study, visit Environmental Science & Technology.

Kim Krisberg is a freelance public health writer living in Austin, Texas, and has been writing about public health for 15 years.