Subsurface Wastewater Injection Sites Shown to be More Susceptible to Remote Earthquake Triggering

Underground fluid injection can induce earthquakes by increasing pressure on the geologic features deep below the surface of the site. This induced seismicity has affected a growing number of sites in the past decade due to an increase in subsurface wastewater disposal associated with natural gas extraction. The increased seismicity of a place may not appear for a few days or up to many years, and there currently is no way to diagnose underground damage near fluid injection sites that might lead to increased seismicity. Van der Elst et al. (2013) investigate the possibility that dynamic triggering from large, distant earthquakes may be an indicator of stress in sites of wastewater injection. The triggers in this study are three large earthquakes (over 8 moment magnitude, or Mw) between 2010 and 2012. The authors find three sites with a history of fluid injection that respond to these triggers with patterns of seismicity indicative of underground damage.—Shannon Julius
Van der Elst, Nicholas J., et al. “Enhanced remote earthquake triggering at fluid-injection sites in the midwestern United States.” Science 341.6142 (2013): 164-167.

Van der Elst et al. (2013) examined earthquake data to look for sites responsive to three large trigger earthquakes. The earthquakes were: a February 2010 8.8 Mw in in Maule, Chile; a March 2011 9.1 Mw earthquake in Tohoku-oki; and April 2012 8.6 Mw in Sumatra. In order to judge whether induced seismicity was promoted by anthropogenic activity, the authors found data for all earthquakes 3 Mw and higher in the central U.S. within 10 days of each event. For sites that had a clear response to the trigger earthquake, the authors studied earthquake records for months and years following the triggering event. In particular, the authors looked for sites that had low seismicity before the initial trigger and uncharacteristic seismic events afterwards.  
When the authors mapped all 3 Mw and higher earthquakes that happened within 10 days of a trigger earthquake, they found that triggering happened almost exclusively at three sites: Prague, Oklahoma; Snyder, Texas; and Trinidad, Colorado. Each of these sites had relatively low seismicity before the first trigger, at least one fairly large earthquake in response to the trigger, and a delayed earthquake “swarm” months after the trigger. Additionally, each site had a strong history of wastewater injection within 10 kilometers of the induced earthquake activity.
The first site was the Cogdell oil field near Snyder, Texas. This site had a number of earthquakes in response to the March 2011 event in Tohoku-oki. The largest earthquake was a 3.8 Mw event that happened two and a half days after smaller events. A few months later, in September 2011, the site had a seismic “swarm” that included a 4.3 Mw main shock. The rate of earthquakes was higher at the site in the 10 days after the earthquake at Tohoku-oki and immediately after the September swarm than at other any time from February 2009 to the present.
The 2010 Maule event triggered a series of earthquakes in an area near a fluid injection site in Prague, Oklahoma. The largest event was 4.6 Mw, and it occurred only 16 hours after the Maule quake. There was a very low rate of earthquakes at the site prior to the trigger and no activity measured in the 4 months before the Maule event. These triggered earthquakes were suggestively located near the epicenter of a 5.0 Mw earthquake that occurred the following year, in November 2011. This event led to two more earthquakes with a magnitude greater than 5.0 Mw. The largest of these, a 5.7 Mw in November, could possibly be the largest earthquake associated with wastewater disposal. As a sign of continued seismicity, the 2012 earthquake in Sumatra also caused a small amount of activity near the edge of the swarm in 2011. 
Trinidad, Colorado also experienced seismic activity as a result of the Maule earthquake, again a small response near the site of future activity. There were only four events in the day after the earthquake in Maule, but the site only had five earthquakes in the entire previous year, so this result is significant. The delayed swarm occurred in August 2011 and included a 5.3 Mw main shock. As in Oklahoma, the Sumatra earthquake caused a small boost in earthquake activity near the site of the previous earthquake swarm.

The authors are concerned that these findings indicate the likelihood of future induced seismic activity at these damaged sites, and they suggest that improved seismic monitoring should occur in areas of subsurface wastewater injection.

Hydrophobic Amino Acids as a New Class of Kinetic Inhibitors for Gas Hydrate Formation

Sa et al. investigated the effects of amino acids as Kinetic Hydrate Inhibiters (KHIs) on the initial formation, the continued growth, and the structure of hydrate blockages in natural gas and oil pipelines. They found that hydrophobic amino acids were more effective KHIs than the more commonly used polyvinyl pyrrolidone (PVP). It was also observed that in general, amino acids with shorter alkyl chains were more effective KHIs than those with longer alkyl chains, with glycine and L-alanine being the most effective KHIs. Examining how various KHI’s impacted the rate of growth of hydrate blockages, it was found that both PVP and glycine as KHIs caused a decrease in the rate of formation of the hydrogen blockage. When comparing amino acids with varying alkyl chain lengths, it was observed that as the length of the alkyl chain increased, its ability to act as an effective KHI decreased. It was found that the crystal structure of hydrates formed did not change in the presence of the amino acid KHIs. However, it was found that all amino acids, regardless of their hydrophobicity, were effective in inhibiting hydrate blockages once the blockages had begun to form, as seen by the increased number of ice crystals in the hydrate in the presence of glycine.—Allison Kerley

Sa, J., Kwak, G., Lee, B., Park, D., Han, K., Lee, K., 2013. Hydrophobic amino acids as a new class of kinetic inhibitors for gas hydrate formation. Scientific Reports 3, 2428.

                  Using nucleation kinetics measurements to observe the onset of hydrate blockage formation, Sa et al. examined the effects of different amino acids and PVP on hydrate formation in fresh water and memory water. The “memory effect” of memory water is a phenomenon in which hydrates form more easily in gas and water that has formed hydrates in the past. While PVP did not display any effect on the inhibition of hydrates, glycine (at an increased concentration of 1.0% mol) slowed the formation of hydrates.
                  Synchrotron powder X-ray diffraction (PXRD) was used to identify the structure of the hydrate blockages, enabling Sa et al. to determine whether KHIs affected the structural makeup of hydrate blockages. It was found that in the presence of glycine, hydrate blockages displayed more ice crystals, which was attributed to water molecules freezing instead of forming hydrates.

Allison Kerley

According to Life Cycle Assessment, Shale Gas Produces Half the GHG Emissions and Consumes Half the Freshwater of Coal

The increase in shale gas production in the United States has led to an interest in the environmental impacts of this unconventional and largely unstudied source of natural gas. Ian Laurenzi and Jersey Gilbert of the ExxonMobile Research and Engineering Company present a life cycle assessment (LCA) of  both greenhouse gas (GHG) emissions and freshwater consumption of Marcellus shale gas. This assessment includes processes from drilling the gas well to power generation. Using their elaborated system boundaries, the authors found that a Marcellus shale gas life cycle releases 466 kg of carbon equivalent units per megawatt hour of power produced (kg CO2eq/MWh) and consumes 224 gallons of freshwater per megawatt hour of power produced (gal/MWh). The biggest contributor to both GHG emissions and freshwater consumption is the power plant. The results are similar to previous LCAs of conventional and shale gas and are far lower than results from LCAs of coal. Even considering factors that can increase total results, this study shows that average GHG emissions from shale gas are 53% lower and freshwater consumption is 50% lower than required for an average coal life cycle. —Shannon Julius
                  Laurenzi, Ian J., and Gilbert R. Jersey. “Life Cycle Greenhouse Gas Emissions and Freshwater Consumption of Marcellus Shale Gas.” Environmental science & technology 47.9 (2013): 4896-4903.

                  Laurenzi and Jersey use a “from well to wire” approach to study the carbon and water footprints of Marcellus gas. In this study, the shale gas life cycle is defined to include drilling, well completion, wastewater disposal, transportation of gas from well via gathering pipelines, treatment and processing, transmission, and power generation. It also includes consideration of water consumed for hydraulic fracturing and evaporative cooling at the power plant. Excluded from the study are gas distribution networks, which deliver gas for purposes other than electricity. The GHG emissions are expressed in units of CO2 equivalents based on an IPCC specification. The idea behind this unit is to make the “global warming potential” for all green house gases comparable. This study used 100-year global warming potential values of 25 kg CO2eq/kg CH4 (methane) and 298 kg CO2eq/kg N2O (nitrous oxide). The functional units for the whole study were kg CO2eq/MWh (amount of gas released per unit of power produced) and gal/MWh (gallons of water consumed per unit of power produced).  The authors used data from over 200 Marcellus shale wells in West Virginia and Pennsylvania. Their information largely came from XTO Energy, a subsidiary of ExxonMobil, and where data was not available  they used established standards from different regulatory agencies or publicly available data. Modeling of the power generation stage used a combined cycle gas turbine power plant operating at 50.2% efficiency.
                  The authors’ calculations revealed that the total life cycle GHG emissions of Marcellus shale gasses are 466 kg carbon equivalent units per MWh of power produced. The majority (almost 78%) of emissions occur at the power plant. The second most significant source of GHG emissions are the gas engines that drive the gathering system compressors, which are part of the system that transports gas from the well to a central location. Hydraulic fracturing activities are only responsible for 1.2% of the lifecycle GHG emissions. Only 1.17% of total GHG emissions are specific to Marcellus shale gas production and processing, making the difference between Marcellus shale gas and conventional gas statistically insignificant. Some other sources of emissions are: transmission compressors, transmission losses, processing plant compressors, processing losses, pneumatic devices and chemical injection pumps, and road transportation for well maintenance.
                  The total life cycle water consumption is 224 gallons of freshwater per MWh of power produced, with 93.3% of that total occurring at the power plant. Of the remaining water consumed, 91% (13.7 gal/MWh) goes towards hydraulic fracturing operations. That figure includes water used in the life cycles of gasoline or diesel used to power the fracturing process or for transportation. Water is also consumed during drilling, casing manufacture, and road transportation for well maintenance.
                  The results of this assessment are dependent on the particular boundaries chosen to represent the life cycle of shale gas. The most important parameter is the expected ultimate recovery (EUR) of natural gas from a well, since there will be more greenhouse gases released per unit of power produced if more wells are needed to yield the same amount of natural gas.  The GHG emissions associated with life stages besides drilling and completion are independent of EUR. Still, there is a strong inverse relationship between EUR and total lifecycle GHG emissions. Other important parameters are pipeline length, gas composition, water scarcity in the region, and other infrastructural elements.
                  Another factor that could greatly change the final result is the efficiency of the power plant. This assessment used an efficiency of 50.2%, which is relatively consistent with the 80% of U.S. power plants that operate within the range of 42-48% efficiency. When the authors present a separate GHG distribution using data from the less efficient, currently operating U.S. power plants, they get a distribution that  is wider with a higher average of GHG emissions. Even so, the highest possible level of Marcellus shale gas emissions from this higher life cycle distribution is lower than the lowest possible GHG emission for an average coal life cycle.
                  Despite the potential for variability of results due to the previously stated factors, the result of 466 CO2eq/MWh is consistent with other published life cycle assessments for conventional and shale gas, and almost all of the 14 studies fall within the 10%-90% range of 450-567 CO2eq/MWh. 

Stray Gases from Shale Gas Extraction Contaminate Drinking Water in Pennsylvania

Shale gas is an unconventional source of natural gas recently made accessible by horizontal drilling and hydraulic fracturing. Shale and other unconventional sources of natural gas have caused overall U.S. production of methane to increase 30% since 2005. Despite their increasing importance, the environmental implications of producing unconventional natural gas have not yet been studied extensively. Jackson et al. explored the possibility of stray gas contamination by testing for concentrations of methane, ethane, and propane in drinking water wells near homes in the Marcellus shale region of Pennsylvania. In general, they found higher amounts of dissolved gases in sources less than one kilometer from a natural gas well. Statistical analysis showed that distance from gas wells was a more significant factor for raised levels of natural gas than other potential sources of contamination. Closer analysis into the chemistry of the samples showed that at least some of the natural gases present in drinking water wells came from a thermogenic source, which includes gas wells. The authors suggest that the stray gases could be due to wells with faulty steel casings or cement sealing.—Shannon Julius
                  Jackson, R.B., Vengosh, A., Darrah, T.H., Warner, N.R., Down, A., Poreda, R.J., Osborn, S.G., Zhao, K., Karr, J.D., 2013. Increased stray gas abundance in a subset of drinking water wells near Marcellus shale gas extraction. Proceedings of the National Academy of Sciences 110, 11250-11255

                  Jackson et al. sampled 81 drinking water wells and combined their results with information from 60 previously-collected samples. They measured the concentrations of dissolved methane, ethane, and propane in the water samples and the distance to the nearest gas well from each sample. Other possible sources of natural gas contamination—valley bottom streams and the Appalachian Structural Front—were ruled out using multiple regression analysis, linear regression, and Pearson and Spearman coefficients. In addition, the authors tested to see if the gas came from biogenic sources, i.e. produced by microorganisms, or from thermogenic sources, i.e. with a potential connection to shale gas production. Indicators of thermogenic gas include the presence of ethane or propane, certain isotopic signatures in methane (δ13C-CH4), and the ratio of helium isotope (4He) to methane.
                  The study found dissolved methane at 115 of 141 homes (82%), ethane at 40 of 133 homes (30%), and propane at 10 of 133 homes (8%). Methane had a far higher average than other natural gases in all cases, but homes within one kilometer of a natural gas well had 6 times the amount of methane as homes farther away. The 12 highest concentrations of methane were above the U.S. Department of Interior level for hazard mitigation, and 11 of  those houses were within one kilometer of a gas well. In addition, homes within one kilometer of a gas well had 23 times the amount of ethane as homes farther away, and propane was only detected at homes within one kilometer of a gas well. Ethane and propane only derive from thermogentic sources, so their presence is evidence that the natural gas contamination is likely from a gas well.
                  Another way to determine if gas is from a thermogenic source is to look at the isotopic signature. Yet again the strongest evidence for thermogenic sources (the most δ13C-CH4 signatures greater than –40‰) were within one kilometer from natural gas wells. There is also a trend of shale gas in which isotopes (δ13C) of methane become heavier than those of ethane, though in other cases it is the reverse. Six out of 11 houses where sampling was possible showed this shale gas trend, indicating that those gas samples came from shale gas production.
                  The helium isotope 4He is a component of thermogenic natural gas. The ratio of this isotope to methane (4He to CH4) in the dataset was fairly consistent, except for the points with elevated levels of methane. These had a ratio of 4He to CH4 that was consistant with Marcellus production gases, somewhat lower than normal drinking water levels.
                  The authors contend that poor well construction led to this contamination of drinking water. In particular, stray gases could have escaped through faulty protective steel casings or from imperfections in the cement sealer between the casings and rock outside the well. Faulty steel casings would lead to gas from inside the well leaking out to the surroundings, followed by metals, fracking fluids, or other evidence of gas extraction. Faulty cement would lead to any gas in the spaces around the well to escape upwards into drinking water, meaning the gases would not be easily identifiable with the well itself.

                  The authors would like to see further research to understand more about how drinking water near the Marcellus shale gas production area compares to drinking water near other shale gas sources. They also suggest gathering predrilling data, even making detailed studies of water quality before, during, and after drilling and hydraulic fracturing.