This study was published in 2011 in the Proceedings of the National Academy of the Sciences and is the most comprehensive and independent study on the impacts of hydraulic fracturing to date.
Although it eventually gets a little technical, the study demonstrates a direct correlation (pg. 2) between methane concentration and distance to a hydraulic fracturing operation.
Methane concentrations on average were 17 times higher when water samples were taken from a water well near hydraulic fracturing than water samples from wells out of extraction areas.
The researchers also concluded from isotopic analyses that methane in gas wells was NOT naturally occurring. Methane was directly linked to the hydraulic fracturing process through isotopic analysis (pg. 3-4).
A total of 68 drinking water samples were collected from two states; drilling was ocurring for both Marcellus and Utica shale.
The study corrected for error by sampling geologically identical areas and using a random well sample.
Although so called "pre-drilling" and "post-drilling" tests were not done per se, they were unnecessary because of the extensive isotopic analyses.
The researchers concluded that hydraulic fracturing could be a moderately safe process, but inadequate regulations have allowed the industry to cut corners. They recommend closing the "Halliburton loophole" in the Energy Policy Act of 2005.
The researchers have defended their study against industry officials here and here.
This is an open letter to regulators in response to shale gas being touted as a "transitional fuel."
The researchers report that shale gas emissions are at least 30% more, or could be as much as twice as much, as conventional gas. The gas itself is responsible due, in large part, to different isotopes of methane. The process of hydraulic fracturing also allows gas to flow into the atmosphere during and after fracturing. This increases dramatically the total environmental cost of the drilling practice.
Compared to coal, shale gas is at least 20% dirtier, and could have a footprint as much as twice that of coal. When the results are projected out 100 years or so, shale gas and coal become more comparable due to coal miners resorting to unconventional mining and the closure of shale gas wells.
Specifically, the GHG footprint of shale gas was analyzed. Over the life of the well, methane is constantly escaping so that the shale gas can be simultaneously released. Flowback water also contains large amounts of methane that are released due to fracturing.
Routine venting, equipment leaks, and gas released during transport and refining also contribute to shale gas having a larger environmental footprint than coal.
The industry must vent gas from wells and equipment, and during transport and processing.
Current pipeline systems and transport and refining infrastructure is inadequate for shale gas, so much of the gas is lost along the way, increasing the footprint.
Distribution leaks also contribute to losses along the transport system. This is particularly alarming because an antiquated system would be leaking shale gas near residential, commercial, and industrial areas.
The researchers concluded that, over the life of a shale gas well, 3.6% to 7.9% of the total production is lost and emitted to the atmosphere without energy gain.
In the twenty year time horizon, shale gas' high estimate footprint is much higher than any other fuel including coal, oil, and conventional gas. The low estimate for shale gas was higher than any other fuel except for the high estimate of conventional gas.
In the one hundred year time horizon, shale gas still has the highest footprint for both high and low estimates, though it's lead is not as large.
The researchers concluded that shale gas' environmental implications do not make it a "bridge fuel." If the goal is to reduce global climate change, shale gas is neither the solution, nor part of the solution.
This is the first time that the EPA has studied hydraulic fracturing in detail and its relation to groundwater or aquifer contamination. The study supports Duke's conclusions.
Methodology is on par with other independent studies.
Four sampling phases were conducted between March 2009 and April 2011.
35 domestic wells and 2 municipal wells were sampled. A random sample is claimed.
High levels of methane detected in Phases I and II prompted the installation of 2 deep monitoring wells (Phases III and IV) that are described in detail on pg. 9-10.
A closed system and later borehole data were incorporated into the methodology of Phase IV.
Part One Conclusions: Hydraulic fracturing contaminates aquifers for the following reasons:
High pH values (Very basic samples): The deep monitoring wells' data show that potassium hydroxide (KOH) had contaminated aquifers, suggesting that drilling additives may leak into water supplies from aquifers.
Data indicate that potassium and chloride levels were higher in aquifers than expected, suggesting that drilling additives may leak into water supplies from aquifers.
Synthetic organic compounds were detected in aquifers. Synthetic compounds cannot occur naturally; analysis results indicate that the compounds came from drilling fluid.
Data show the detection of petroleum based compounds in aquifers that are only used in fracturing fluid.
Products of the breakdown of previously discussed compounds were also found.
Data indicate that production casings were often completed inadequately.
Geologic analysis indicates that additives and production fluids could easily migrate to aquifers relatively quickly.
Part Two Conclusions: Hydraulic fracturing contaminates groundwater for the following reasons:
Hydrocarbon and isotopic analysis of gas: Similar procedures to the Duke study were used; similar results were reported.
Elevation of dissolved methane: methane concentration increases with decreasing distance to a gas well.
Geologic feature near a specific well well cited as cause of extremely high methane concentrations. Geologic analysis revealed easier methane migration to groundwater.
Production casing elevation: Production casings did not extend low enough to shield fractures from paths for migration to groundwater.
Domestic complaints of water contamination are uniformly consistent.
Final Conclusion: "A lines of reasoning approach utilized at this site best supports an explanation that inorganic and organic constituents associated with hydraulic fracturing have contaminated ground water at and below the depth used for domestic water supply" (39).
Recommendations include further study and regulations that ensure that natural gas extraction is done safely, as current regulations are inadequate.
This study's conclusions support some of the Duke study's conclusions, but contradict the Duke study in some key areas. The study was conducted in 2010 and 2011 by the Center for Rural Pennsylvania. Thank you to Scott Smith for this tip [Smith wishes to declare a conflict of interest, working for the oil and gas industry.]
This study's methodology calls into question any results it declares. The researchers were underfunded and did not test for all variables for which the Duke researchers tested.
Wells were not taken as a random sample.
The researchers sampled wells all over the state; the only criterion for testing was the testing data from before drilling began.
The testing was done by different testing firms, so it's impossible to derive what their methodology was and it's impossible to adequately compare the results of different tests.
Tests also were done for different chemicals, and isotopic analysis was not done. Additionally, the wells tested were chosen by the gas industry, and pre-drilling tests were done by industry hired testing firms.
Wells that were selected were concentrated in certain distance brackets from drilling, making some results void by lack of sufficient data.
Pre-drilling tests were taken over a long period of time, making comparisons and any conclusions void.
Wells were not compared based on geological activity.
Overall, the methodology makes any results void by lack of adequate or sufficient information, or by lack of comparison or causal link tests. Therefore, take any results with a grain of salt!
The researchers tested for some chemicals that were common in pre-testing, but, for many, data was not available.
Isotopic analysis was not done; no causal link can be established.
Statistical analysis showed no major correlation between distance from a gas well and methane concentration.
Correlations were concluded between distance from a gas well and other chemicals' presence and concentration.
The researchers also concluded that the "distance from a gas well" parameter in other analyses should be extended from 1000 feet to 3000 feet. This conclusion also recommended a policy change that would include extending the "distance for presumed responsibility" to 3000 feet.
Recommendations include further testing and large scale analyses with better methodology (see: Duke study or EPA report).
Sandra Steingraber, Ph. D Lecture to the New York Legislature