Research Proposal
Estimation of Greenhouse Gas Emissions from
Unconventional Natural Gas Wells in Pennsylvania
Abstract
Natural gas from hydraulically fractured wells is a new source of energy for the United States, but the full greenhouse gas impact from the new techniques is not yet determined. Specifically, estimates on the quantity of gas that can escape during well completion differ significantly, which means the net impact on the environment and climate is not certain. Passive remote sensing methods using the principle of differential absorption, tuned to the infrared spectrum of methane, provide the opportunity to establish reasonable estimates of the volume of gas emitted during flow-back on the wells. Quantitative estimates of gasses in different situations using active methods or passive methods combined with models have been examined, but they often assume particular models. Moreover, time and spatial variation are not studied as often as instantaneous, point measurements. Passive methods with differential absorption have not been fully developed for this application and will provide quantitative estimates with spatial and temporal information.
Introduction
Currently, there is a movement to use unconventional drilling techniques, especially high-volume, slick-water horizontal hydraulic fracturing (fracking), to obtain natural gas from tight formations such as low-permeability shale. There is contention about the risks and benefits of using this new technology, but one crucial issue is the net impact on greenhouse gas (GHG) emissions of developing the natural gas resources for energy production versus continued use of coal power plants. Much of the disagreement about the GHG footprint derives from different estimates of the amount of natural gas, mainly methane, which is leaked from well sites over the course of completion of the well.
Natural gas can leak at many points in the production and transportation process, but the primary time of interest for shale gas is during well completion. During the time after fracturing the well and before production, flow-back, the fracturing fluid that returns to the surface from the well, can bring a substantial quantity of natural gas with it (Howarth et al., 2011); however, the estimate of this quantity varies widely in the literature (Howarth et al., 2011; Cathles et al., 2011; Weber & Clavin, 2012). Developing quantitative estimates for leaks based on data is very important because methane is a more potent GHG over shorter time periods, so uncertainty in estimates is magnified for short-run evaluation of net GHG impacts.
The objective of this pilot study is to demonstrate that remote sensing technologies can be used to refine estimates of the emissions from unconventional natural gas wells during the flow-back period. Differential absorption is a standard remote sensing technique to determine concentration levels of a gas species without physically sampling from the gas plume. Applying differential absorption methods requires taking two measurements: one on the peak absorption band and another just outside of the absorption region. The difference in the two values is proportional to the quantity of gas over observation path. A differential absorption image will show the distribution of the target gas in the plume, which means the total volume of gas can by determined by summing all differences over an image. Many volatile organic carbons (VOCs), including methane, the main component of natural gas, have peak absorption at 3.2 μm. This characteristic makes monitoring of wells using cameras an attractive option.
Literature Review
Much of the existing research can be divided into three categories: detection of gas emissions sources, quantification of emissions using models, and quantification of gasses using differential absorption.
The principle of passive detection of gas emissions based on single, narrow-band infrared (IR) wavelength images and monitoring has already been demonstrated, but this is not the same principle of differential absorption because it is single band, not multiple bands, and does not provide quantitative estimates of concentration. One particular area of interest has been natural gas leaks in pipelines, and an early study showed that the narrow absorption band of VOCs could be used to detect natural gas leaks from buried pipes (Gross et al., 1998). IR cameras, coupled with dispersion modeling have already been used to qualitatively estimate the dispersion of liquid natural gas spill plumes (Safitri & Mamman, 2010). It should be noted that this study documents a strong dependence of absorption/emissivity on the temperature of the gas. The effect occurs at different rates over the midwave region. This will complicate the quantification of the gas if there is a significant temperature variation of the gas within the scene. Although the natural gas industry is already using similar technologies to identify leaks, there are problems in quantifying leaks accurately outside of laboratory conditions when environmental factors must be taken into account (Safitri et al., 2011).One study demonstrated that passive IR scanning from aircraft could determine methanol and ethanol emissions, and this was part of a more extensive project that included image processing procedures and methods (Wabomba et al., 2007).
Some attempts have been made to quantify the emissions rates using passive remote sensing technology. One study used Gaussian dispersion modeling, with the known minimum detectable concentration, to estimate the mass flow rate leaving a leaky pipe (Safitri et al., 2011). Another method that has been used to estimate the gas flow rate is tracking different sections of the plume from one image to the next, over short time intervals, and inferring the speed of the plume from the shift in the critical points (Sandsten & Andersson, 2012). Once the concentration and velocity are known, a gas flow rate can be computed (Sandsten & Andersson, 2012).
Most active detection with remote sensing is limited to differential absorption applied to light detection and ranging (lidar) measurements. A differential absorption lidar (DIAL) requires a dual wavelength lidar with one wavelength tuned to the absorption band and one wavelength just outside the absorption region. The lidar pulse is sent in both wavelengths and the returned radiation is measured for both wavelengths. A difference between the two returns means that the species of interest is present and the concentration can be computed based on the magnitude of the difference. One study compared DIAL to four other methods to determine the methane fluxes for a landfill (Babilotte et al., 2010).
Differential absorption techniques have been applied in determining SO2 and NO2 emissions from ships. One study showed that applying differential absorption to sunlight that passed through the ship plume and was then reflected off the surface of the water to the aircraft sensor allowed estimates within the 45% uncertainty for the validation case (Berg et al., 2012). Another study monitored several ships passing through a channel for SO2 and NO2 emissions using the differential absorption techniques with the empty channel as the background image and incorporating meteorological data (Premuda et al., 2011).
Employing differential absorption to passive monitoring techniques for unconventional natural gas well emissions is reasonable based on previous work and will provide new applications for remote sensing methods to this crucial problem.
Technical Method & Testing
A passive means of monitoring the emissions is proposed using differential absorption. The camera can sense the emission of thermal radiation in the mid-wavelength IR spectrum. Both the gas and the background objects in an image will emit in this range, but the gas emission will absorb at a specific wavelength as a property of the material. Outside the narrow absorption band the gas and background will emit radiation at close to the same level, so they will look the same. If two images are taken, one at the peak absorption for the gas and one farther away, the peak absorption image will be darker where there is gas absorbing. The gas acts as a filter and removes the background radiation, which reduces the thermal radiation detected by the camera. The more gas that is present in the area covered by the pixel the darker the pixel will look. Moreover, if there is no gas present the gas will have no impact and the two images will look the same. For the two images the difference in each pixel can be determined, and this difference will give a distribution of gas in the plume. For each pixel, the difference will be proportional to the quantity of gas in the plume covered by that pixel.
Methane, the gas of interest, has a strong absorption band around 3.2 μm (3020 cm-1), as shown in the figure below; however, many VOCs have similar peak absorption spectra. Thus, the images alone cannot determine what gas is present, but they can show that some gas is absorbing at the specified wavelength. Therefore, paired images can be employed to estimate the quantity of VOCs at a single instant. The difference between the image collected at 3.2 μm and a reference image at approximately 4.2 or 2.5 μm will give a pixel-by-pixel representation of the gas quantity. The more gas in each pixel, the greater the difference. This, combined with gathered on-site weather conditions, will allow the quantity and flow rate of the gas to be determined, similar to the method used for the concentration of gasses on the ship plumes (Premuda et al., 2011). A visual camera will be used as well to monitor the site and check that nothing is hindering data collection.
A FLIR Orion SC7000, which can take images in up to eight bands, will be used to monitor the wells.[1] This camera has sensors for the 1.5~5.1μm range and a filter wheel that can spin at up to 400 Hz.[2] Only four filters are intended for this project: two at the peak absorption for VOCs (3.2 μm) and two off peak absorption (2.5 and 4.2 μm). Consequently, for each rotation of the filter wheel four images will be taken. Of these four images two will be background images and two peak absorption images. Therefore, two differential absorption calculations are possible because there are off-peak absorption images at both shorter and longer wavelengths. Both sets of images will give an estimate of the quantity of VOCs present. Because the rotation is 400 Hz the images will be close enough to consider the gas cloud essentially stationary and no further corrections are necessary.
The limits of the system under expected ranges of field conditions will be determined through controlled experiments before any field measurements are completed. All system limitations, especially related to concentration and distance, will be tested to aid in the optimal placement in the field. Also, since VOCs cannot necessarily be separated from methane, a remote methane detector will be included to help substantiate the differential absorption measurements are for methane and not some other substance.
Absorption Spectrum of Methane ( 3500 cm-1 = 2.9 μm, 3000 cm-1 = 3.3 μm, 2500 cm-1 = 4.0 μm, 2000 cm-1 = 5.0 μm,
Field Work
Several challenges, including location, timing, and weather, will be considered during fieldwork. Bradford County, PA, is the proposed location for all fieldwork. It has a high gas well density and reports of methane entering the groundwater, a sign that could imply methane emissions from wells. A well site is approximately 3.5 and 1.0 acres during the development and production phases, respectively (NYS DEC, 2011). Given the size of each site, it is possible to get a field unit within a reasonable distance of the wellhead.
Since flow-back can last for several weeks (Howarth et al., 2011), each site will be monitored for approximately two to three weeks. Timing the monitoring of each well to correspond with the flow-back period is important, but the high density in the area should make these logistics somewhat easier. In total, six to eight wells will be monitored.
At the well site, a generator will run all equipment including cameras, a computer, and a weather station. Most of the electronics will be inside a small shed and protected from the elements, but the generator and weather station will be outside. The FLIR camera will view the site through an optical sapphire window, which has 90% transmittance in the range of interest. Similar material has been used before (Gross et al., 1998).
All equipment will be controlled remotely using a wireless connection to the Internet through a computer. Additionally, a camera will stream visible data from the site to determine if any changes are needed based on activities at the site. When the structure is moved the information will be transferred for further analysis.
Expected Results
The main result of this pilot study should be determining the limits of a passive infrared differential absorption as a measurement method for evaluating the VOC leaks from unconventional natural gas wells. Secondary results include analysis of the best weather conditions for such measurements and possible correlations between any of the measured environmental parameters and the measured gas emissions.
Timeline
A rough timeline is given below. One critical component to the success of the project is finding new wells drilled in an appropriate timeframe so that the equipment can be used fully. Image analysis should be relatively simple and can be started before all wells are monitored because the data is transferred after each well is finished.
Months1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12
Task 1: Equipment & Testing
1.1 / Ordering Equipment
1.2 / Laboratory Testing
1.3 / Field Unit Construction
Task 2: Field Experiments
2.1 / Negotiations with Companies & Land Owners
2.2 / Monitoring of wells
Task 3: Data Analysis
Task 4: Final Report
Conclusion
Understanding an energy source’s impact on greenhouse gas emissions is vital for understanding which resources should be developed. Remote sensing provides a reasonable option to estimate emissions through differential absorption and gain a better sense of the real impacts of the quickly growing hydraulic fracturing industry.
References
Babilotte, A., Lagier, T., Fiani, E., & Taramini, V. (2010). Fugitive methane emissions from landfills: field comparison of five methods on a French landfill. Journal of Environmental Engineering 136:777-784.
doi: 10.1061/(ASCE)EE.1943-7870.0000260
Berg, N., Mellqvist, J., Jalkanen, J.-P., & Balzani, J. (2012). Ship emissions of SO2 and NO2: DOAS measurements from airborne platforms. Atmospheric Measurement Techniques 5:1085-1098.
doi: 10.5194/amt-5-1085-2012
Cathles, L. M. III, Brown, L., Milton, T., & Hunter, A. (2011). A commentary on “The greenhouse-gas footprint of natural gas in shale formations” by R.W. Howarth, R. Santoro, and Anthony Ingraffea. Climatic Change 113:525-535
doi: 10.1007/s10584-011-0333-0
Gross, W., Hierl, T., Scheuerpflug, H., Schirl, U., Schreer, O., & Schulz, M. (1998). Localization of methane distributions by spectrally tuned infrared imaging. SPIE 3533:234-240.
Howarth, R., Santoro, R., & Ingraffea, A. (2011). Methane and the greenhouse-gas footprint of natural gas shale formations: A letter. Climatic Change 106:679-690. doi: 10.1007/s10584-011-0061-5
New York State Department of Environmental Conservation (2011). Revised Draft Supplemental Generic Environmental Impact Statement: Potential Environmental Impacts.
Premuda, M., Masieri, S., Bortoli, D., Kostadinov, I., Petritoli, A., & Giovanelli, G. (2011). Evaluation of vessel emissions in a lagoon area with ground based Multi axis DOAS measurements. Atmospheric Environment 45:5212-5219. doi:10.1016/j.atmosenv.2011.05.067
Safitri, A., & Mannan, M.S. (2010). Methane gas visualization using infrared imaging system and evaluation of temperature dependence of methane gas emissivity. Industrial Engineering & Chemistry Research 49:3926-3935.
doi: 10.1021/ie901340g
Safitri, A., Gao, X., & Mannan, M.S. (2011). Dispersion modeling approach for quantification of methane emission rates from natural gas fugitive leaks detected by infrared imaging technique. Journal of Loss Prevention in the Process Industries 24:138-145. doi:10.1016/j.jlp.2010.11.007
Sandsten, J. & Andersson, M. (2012). Volume flow calculations on gas leaks imaged with infrared gas-correlation. Optics Express 20:20318-20329.
doi: 10.1364/OE.20.020318.
Wabomba, M.J., Sulub, Y., & Small, G.W. (2007). Remote Detection of Volatile Organic Compounds by Passive Multispectral Infrared Imaging Measurements. Applied Spectroscopy 61:349-358.
Weber, C.L. & Clavin, C. (2012). Life cycle carbon footprint of shale gas: review of evidence and implications. Environmental Science & Technology 46:5688-5695. doi:10/1021/es300375n
Budget: Justification of Expenses
The university dictates most of the graduate student and faculty benefits and expenses. This budget provides the student with ample supervision from the principal investigator over the course of the year.
The main component of the budget is the FLIR camera and accessories. This camera is very expensive but is unique in its ability to capture multiple images at the different wavelengths, making the differential absorption observation possible. The other equipment specific to the camera is costly because of the specialized nature.
Most of the field instruments besides the camera are necessary for a complete analysis of the data. On-site weather information is critical for understanding the limitations of this method in fieldwork. Additionally, the data could be useful for developing further models for the dispersion of the gasses. Software to record and store field measurements is necessary for later analysis of the data.
Transportation of the field unit, which is a shed with all other instruments attached, requires a trailer and a truck. For security reasons the shed will be secured to the trailer for the full duration in the field.
Research Proposal
PROPOSAL BUDGETYear 1 / increment / 0
Jul-13 / 1.05 / # months committed
Jun-14 / Year 1
Jul-13
1. / Salaries / Base Salary / Jun-14
Principal Investigator / AY / (E) / 1,000 / 10,000 /mo / Ac. Yr / 0.1
summer / (E) / 1,000 / 10,000 /mo / summer / 0.1
Graduate Students (1)
Ac. Yr. stipend / (E) / 23,335 / Ac. Yr. / 9
summer stipend / (E) / 8,449 / summer / 3
health insurance / (E) / 2,248 / Health Ins. / 12
GRA exclusion / (E) / 14,750
Subtotal / 50,782
tuition / 29,500
2. / Fringe Benefits / rates / 1.02 / Ac. Yr. / 23,335
Acad. Yr. (E): / 07/01/13 / - / 06/30/15 / .360 / 360 / 1 / summer / 8,449
summer (E): / 07/01/13 / - / 06/30/15 / .360 / 360 / 1.125 / health ins. / 2,248
Subtotal / 720
% of tuition / 0.50
# grad students / 1
3. / Supplies and Miscellaneous
Vehicle & Trailer Rental / 3,100
Vehicle / 700 / 7 days of truck rentals for moving field unit, $100 per day
Trailer / 2,400 / Open trailer rented for $20/day for 120 days
General supplies / 500
Subtotal / 3,600
4. / Equipment & Software
FLIR IR Camera & Equipment Total / 129,000
Camera / 110,500 / FLIR SC7000 Orion camera for images in differing bands
Lens & Filters / 5,000 / Lens for camera and filters for rotating wheel
Maintenance / 12,500 / Maintenance package for cameras after use
Software / 1,000 / Software to gather and analyze collected images
Generator & Fuel / 1,750
Generator / 150 / 1000 Watt gas generator to run camera and computer
Fuel / 1,600 / 4 gallons of fuel per day, 100 days, $4 per gallon
Other Field Unit Costs / 3,850
Storage Shed / 300 / Basis for field unit structure to protect equipment
Computer / 400 / Computer to store data from field measurements
Weather Station / 1,500 / Collects on-site weather information to aid in analysis
Web Camera / 50 / Allows visual monitoring of site
IR Window / 600 / Optical Sapphire material is 90% transparent in range of interest
Remote Methane Detector / 1000 / Allows verification that gas is methane
Storage cases / 300 / Cases to protect camera
Software / 5,000 / LabView software for field unit
Wireless / 60 / Allows remote monitoring
Subtotal / 139,960
5. / Travel
Domestic / 5,000 / Travel and lodging for field work and final presentation of results
Subtotal / 5,000
Total Direct Cost / 202,762
Exclusions / These items are excluded from the overhead costs
GRA Exclusion / 14,750
Health Insurance / 2,248
Capital Equipment / 129,000 / Has a value of $5,000 or more/unit, with useful life of at least 2 years
Subtotal / 145,998
Modified Total Direct Cost (MTDC) / 56,764 / = Total Direct Cost - Exclusions
Indirect Cost (% of MTDC)
07/01/13 / - / 06/30/14 / 0.60 / 34,058 / Indirect Cost is the overhead charged by the University to cover general expenses: electricity, heat, maintenance, administration, secretarial services, etc.
Project Total = / 235,820
[1] FLIR Systems, Inc. (2012). FLIR Orion SC7000 Series: Gasses