This week, researchers at University of Texas – Arlington (UTA) released a new study that suggests drilling may be impacting water quality in West Texas’ Cline Shale region. But the data in the study tell a much different story. Moreover, issues with the study’s methodology and lack of baseline data cast additional doubt on the legitimacy of the research team’s claims.
The researchers tested for a range of chemicals and substances in a handful of the region’s water wells – everything from pH level and total dissolved solids to alcohols and heavy metals. Taking multiple samples from 36 to 42 wells during each of the four phases, the researchers then compared the levels of constituents against health guidelines, such as Maximum Contaminant Limits (MCL) and health advisory levels (HAL).
By testing these constituent levels over time, the researchers hoped to show the levels correlated with an increase in oil and natural gas activities. What they found however, was wild swings in chemical levels across the four phases. For example, the percentage of samples that were equal to or above MCL for arsenic declined as the sampling progressed, but total dissolved solid increased. Additionally, for pH, only Phase 2 showed any wells that exceeded the MCL. Iron levels also fell from Phase 1 to 2 and from Phase 2 to 3, but then increased from Phase 3 to 4.
If a connection between drilling activity and water contamination existed, the results would show a general increase in substances over time – not wildly divergent trends, as the data in the UTA study show.
The researchers did acknowledge that these substances, based on the results, could dissipate over time, which is another example of how groundwater testing is complex and often times quite difficult. Nonetheless, such large fluctuations in constituent levels would be unlikely if drilling were playing a significant role in water quality issues.
The researchers partially conceded this point, writing that their data “do not provide a definitive link” between drilling and contamination. They also concluded that if any contamination were occurring, it would be “sporadic” and not “widespread.” As the authors note:
“The ephemeral detections of dichloromethane and various organic molecules with minimal co-variation suggest that contamination events may be variable and sporadic (as opposed to systematic), but also that many of the toxic compounds associated with areas of high UD activity may degrade naturally over relatively short time scales.” (p. 912; emphasis added)
This finding sounds very similar to the conclusion reached by the U.S. Environmental Protection Agency (EPA) in the agency’s five year study on fracking, which found:
“We did not find evidence that these mechanisms have led to widespread, systemic impacts on drinking water resources in the United States.”
Lack of Baseline Data
According to the researchers, the purpose of the study was to analyze the “temporal variation,” or change over time, of water quality in the Cline Shale. Both the press release and the UTA Collaborative Laboratories for Environmental Analysis and Remediation (CLEAR) website state the water wells were sampled before, during and after unconventional drilling and fracking took place. According to the press release:
“The new research…is the first to analyze groundwater quality in the Cline Shale region of West Texas before, during and after the expansion of hydraulic fracturing and horizontal drilling.” [emphasis added]
And, as the CLEAR website states:
“CLEAR is currently concluding an initial time course investigation on the potential impact of unconventional oil extraction on groundwater quality in the Cline [S]hale of [W]est Texas. Approximately 40 water wells were sampled prior to, during, and after the onset of unconventional drilling and hydraulic stimulation.” [emphasis added]
The problem is, they did not take baseline samples (i.e. prior to drilling), and they even admitted it in their paper.
Indeed, there were existing unconventional oil and natural gas development (UD) wells within 5-10 km of their sampling area, with at least one UD well right next to several of the water wells sampled. The authors write:
“Phase 1 water wells were located an average of 5.65 km (range = 0.515-9.59 km) from the nearest UD well.” (p. 908)
If it were a true baseline sampling, with data actually collected “before” or “prior to” drilling and fracking, no UD wells would be in the area.
Contrary to their press release, the authors admit they did not have baseline samples:
“During Phase 1, there was a total of 298 UD wells in the three counties encompassing our study area; with one UD well within 5 km of our study area. We do not consider the first phase to be a true baseline measure of water quality before UD activities, but rather an initial assessment of water quality at the onset of increased UD activities.” (p. 908; emphasis added)
Why is that so important? Because without baseline samples, the researchers can only assume that water quality was different before drilling. Given that aquifers across Texas often contain high levels of the same substances that these researchers found – even in areas without drilling – the lack of pre-drilling data is a huge information gap.
Of course, that’s in addition to the questions it raises about the research team’s public outreach, which included materials distributed to the press that contradict what’s in the actual study!
History of Issues with Study Methodology
The same UTA researchers who contributed to this study also published a debunked report on the Barnett Shale in 2013. Notably, that study also had serious methodological issues.
Specifically, the 2013 Barnett study sampled only 91 private water wells within 3 miles of a gas well, with an additional nine wells sampled outside the Barnett Shale region for a control group. But this methodology prompted criticism from others in the field, such as Stanford’s Rob Jackson (then a researcher at Duke University), who said the research team’s control group was “too small to be statistically relevant,” according to EnergyWire; and Penn State’s Terry Engelder, who critiqued the use of control wells from a completely different geological area.
The Permian study only sampled 43 water wells, and had zero control wells. If using only nine wells was inadequate, then what does it say that the latest study had zero?
Dr. Engelder, in reviewing the team’s latest study, told EnergyWire that the fluctuating measurements raised additional questions, and based on the information in the actual report, a number of causes could explain the findings:
“The authors can’t say that UD [unconventional oil and gas development] had anything to do with the content of groundwater and they might as well have been positing that farmers in West Texas have been cleaning crank cases and radiators of tractors for about a century and dumping the stuff in their back yards less than, say, 100 meters of their water wells,” Engelder wrote in an email. “Note that most UD activity is more than 5 kilometers from the water samples. The authors also admit that ‘the majority of the chemicals used for UD activities are also used in numerous other industrial, agricultural, and residential activities.'”
Continuing on their 2013 research, the same team released a study in 2015, again examining the Barnett Shale. This time, while the researchers analyzed groundwater from a much larger sample size of 550 private and public wells, the study suffered from the same problem as the previous one: no baseline data taken prior to oil and natural gas development. Without baseline data to compare their results, the researchers conceded that they couldn’t identify the direct cause of contamination, as close proximity to the Dallas-Fort Worth metropolitan area leaves groundwater “potentially vulnerable” to contamination from various sources.
In fact, the researchers had to admit that their findings couldn’t link fracking with contamination:
“The detection of numerous volatile organic compounds in aquifers above the Barnett shale does not necessarily implicate unconventional UOG extraction as the source of contamination; however it does provide an impetus for further monitoring and analysis of groundwater quality in this region.” (p. 18; emphasis added)
While the press release strongly suggests the study links drilling activities and degraded groundwater, that’s simply not true. The data show large variations during each sampling phase – and even though the researchers admit their results are all over the map, such results mean a general statement linking drilling to contamination is far from accurate.
Groundwater testing is a complex process, and deriving concrete results from a very limited set of samples is not just difficult; it’s arguably inappropriate.