The goals of this project are to determine the leakage rates of methane and ozone-forming volatile organic carbons (VOCs) and the emission rates of air toxics from Marcellus Shale gas activities at a process level. Methane emissions in the Marcellus Shale region shall be differentiated between “newer” sources associated with shale gas development and “older” sources associated with coal or conventional natural gas exploration.
Carnegie Mellon University, Pittsburgh, PA 15213
The extensive development of shale gas exploration and production in the Marcellus region has caused concern about potential environmental impacts on water resources, land, and the atmosphere. The overall effects of shale gas development have not been extensively characterized, including the nature and magnitude of the following pollutant emissions:
Emission and leak rates of methane and other gases from shale gas activities are in many cases estimated rather than directly measured. Quantifying the magnitude of these emissions and understanding their sources at a process level is important to ensure industrial compliance with regulatory agencies, promote public health, and improve public understanding of the risks associated with the shale gas industry. The magnitude of methane emissions from Marcellus shale gas operations should be compared to other possible sources of fugitive methane such as landfills, agricultural sources, abandoned coal mines, and gas wells. Quantifying the methane emissions from other regional sources will aid in understanding the net influence of shale gas operations on methane emissions.
The project will directly address uncertainties in the emissions of methane and other VOCs from shale gas operations. Analysis of direct emission rates will improve overall understanding of how shale gas development affects air quality. The analyses of methane, VOCs, and other air toxics will improve the characterization of air emissions across the spectrum of technologies and processes used in the Marcellus Shale region. Project results will help determine how fugitive emissions vary based on different technologies (i.e., “green” versus standard well completions) and if emissions vary with the age of the facility. The project will therefore provide a dataset of emission rates for use in lifecycle analyses, chemical transport models, and policy decision-making.
Concentrations of methane and the tracer were measured by analyzing data for individual plumes at 30 sites. Multiple plumes were detected for each location and the average emission for each site was determined by computing the WAFLER (weighted average facility level emission rate). One challenge of the tracer flux method is to separate intentional plumes (e.g., the released tracers and methane from the sampling site) from any other potential plumes. This is accomplished by releasing two tracers, and identifying concurrent plumes of both tracers with the correct mass ratio of the tracers. The team also verified the location of the plumes with measured wind speed and direction data. The largest emissions were observed at a compressor station during a blow down event.
Results from field sampling showed that conventional wells tend to have methane emissions twice that of unconventional wells. The older equipment associated with conventional wells may be one reason for the higher rate, as newer equipment and detection methods are utilized for the new shale gas wells. A poster discussing the VOCs emissions from wells in Pennsylvania and West Virginia was presented at the December 2015 AGU conference.
“Methane Emissions from Conventional and Unconventional Natural Gas Production Sites in the Marcellus Shale Basin” was published online by the journal Environmental Science and Technology: http://pubs.acs.org/doi/pdf/10.1021/acs.est.5b05503.
During field site sampling, VOC concentrations were often low, even when sampled in the center of a methane plume. The low VOCs are largely consistent with other studies that have observed lower emissions of VOCs from Marcellus shale regions than other basins in the western U.S. CMU is building a database of all sites sampled during this project and related projects. The database will include API numbers, equipment inventory, production data, and measured emissions data.
Based on samples taken, researchers see higher methane emissions at flowback sites than producing wells (either shale or conventional). Emissions are also slightly higher at shale wells than conventional wells, however the raw emission rate does not account for differences in production between the shale and conventional sites. In general, the shale gas wells are larger (multi-well pads) than conventional wells, and also have large gas production. When normalized against production, shale gas wells (both flowback and producing) emit < ~1% of total production.
Field measurements are beginning to reflect the distribution of emissions across sites, with higher emissions at a few wells and many sites with low emissions. The largest emissions were observed at a compressor station during a blow down event. Total leakage at each site appears to be correlated with facility size, however this relationship only describes 55% of the variability between sites. It appears that other factors influence methane leakage rates as well.
Seasonal sampling of wells continued in the 2015 spring and summer months. The team will combine emission data with well population data in order to estimate emission contributions from shale and non-shale sources.
The team sampled active and abandoned wells sites in Pennsylvania and will study the results based on well age and seasonality. The latter will be tested by identifying five wells that will be sampled quarterly over the next year.
Project results were presented at the American Geophysical Union (AGU) annual meeting on December 15, 2014.
Researchers have performed sampling at 31 facilities with tracer testing. Methane emissions at producing wells were estimated at one percent of production levels. Emission measurements were highest at flowback stages than during production.
Researchers conducted calibration experiments of the tracer release technique to determine uncertainty bounds for the field measurements
Researchers conducted a literature search to compile existing data on methane emissions from traditional (non-shale) natural gas sources and shale gas operations and have begun to compile emissions data from non-gas sources of methane, including those from landfills and agriculture. Field measurements began in spring 2014. The tracer flux experiment utilizing the mobile laboratory has been planned for the spring at shale and non-shale sites. Measurements were also gathered from landfill locations.
Researchers sampled ten shale gas sites in Pennsylvania in collaboration with an industry partner. (Seven of the wells were producing and three wells were in the flowback stage.) Methane emissions were higher at flowback sites than at the producing wells. Future reports will normalize the methane emissions rates to total natural gas production at each site in order to determine the fractional leakage of methane. In addition, five conventional gas wells were sampled in West Virginia.
Researchers have contacted industry partners to identify specific unit operations at different stages of shale gas development and are arranging for on-site access to conduct direct measurements of methane emissions and tracer release experiments. A literature search of published emission rates of methane, ozone-forming VOCs, and air toxics from oil and gas activities will begin and include public data submitted to the Environmental Protection Agency or state agencies for regulatory purposes, and the resulting data will be published in the scientific literature. Researchers will obtain methane and VOC emissions data from industry partners.
In Southeast PA and WV, total methane production is dominated by newer, high producing shale wells, but emissions are almost equally split between new wells and older, less productive conventional wells. When normalized against production, shale gas wells (both flow back and producing) emit <~1% of total production.
In Northeast PA (NE PA), the normalized emission rate was typically small with 87% of sites having normalized emissions less than 1% of production, similar to the emission rates observed at shale gas wells in the southeastern portion of the Marcellus Shale in PA and WV. Oil and gas operations in NE PA are dominated by newer shale gas wells. Thus, unlike the southwestern portion of the Marcellus shale, there is not a significant contribution from emissions from older conventional wells.
Emission rates of non-methane VOCs from shale and non-shale sources were measured in the Marcellus Shale and other basins. Overall, it was observed that small emission rates that are not well correlated with methane emissions, indicating that the sources of VOC emissions are often distinct from the sources of methane emissions (e.g.,tank venting versus process leaks). The ethane emission rates in NE PA are slightly lower and less variable compared with the Denver-Julesburg Basin and the Uintah Basin, but for all other VOC species, the emission rates in NE PA are much lower.