v>EPA
United States
Environmental Protection
Agency
EPA 600/R-17/224 | September 2016 | https://www.epa.gov/research
Advancing Understanding of Emissions
from Oil and Natural Gas Production
Operations to Support EPA's Air Quality
Modeling of Ozone Non-Attainment Areas:
Final Summary Report
Office of Research and Development
-------
Advancing Understanding of Emissions from
Oil and Natural Gas Production Operations
to Support EPA's Air Quality Modeling
of Ozone Non-Attainment Areas
FINAL SUMMARY REPORT
Regional Applied Research Effort Program
Project Period: June 2014 to February 2016
Rebecca Matichuk and Gail Tonnesen
EPA Region 8
Officeof Partnerships and Regulatory Assistance
Air Program, Indoor Air, Toxics, andTransportation Unit
Denver, CO
Adam Eisele
EPA Region 8
Enforcement, Complianceand Environmental Justice
Air and Toxics Technical Enforcement Program, Toxics Enforcement Unit
Denver, CO
Eben Thoma and Michael Kosusko
Officeof Research & Development
National Risk Management Research Laboratory
Research Triangle Park, NC
Madeleine Strum
Officeof Air Quality Planning and Standards
Air Quality Assessment Division
Research Triangle Park, NC
Cindy Beeler
EPA Region 8
Officeof the Regional Administrator
Denver, CO
r
September 6, 2016
Photo Credits: Top Photo: 2013 Uinta Basin Winter Ozone Study (March 2014), Chapter 8: Tethered Ozonesonde
and Surface Ozone Measurements in the Uinta Basin, Winter 2013; NOAA Earth System Research Laboratory,
Boulder, CO; CIRES - University of Colorado, Boulder, CO; Science and Technology Coiporation, and University
of Colorado, Boulder, CO. Bottom Photo: Captured by the contractors during the fiscal year 2014 to 2016 Regional
Applied Research Effort -Enclosed CombustorDevices field campaign, September 2014, Weld County, Colorado.
-------
Acknowledgments and Disclaimer:
This project was funded by the EPA Office of Research and Development (ORD) Regional
Applied Research Effort (RARE) program, administered by the Office of Science Policy (OSP),
with supplemental funding from ORD's ACE Climate and Energy (ACE) Program, Task EM
1.2, "Next generation air measurements for fugitive, area source, and fence line applications."
We would like to thank Mark Modrak (formerly with Arcadis) who managed the field study
portion of the project under EPA Contract EP-C-09-027, and by extension, the scientists with
IMACC and Telops who participated in the pilot study. We also thank Ying Hsu and Frank
Divita, with Abt Associates for their efforts on the SPECIATE portion of the project under EPA
Contract No. EP-D-08-100.
The views expressed in this report are those of the authors and do not necessarily represent the
views or policies of the U.S. Environmental Protection Agency. Mention of any products or trade
names does not constitute endorsement.
-------
Table of Contents: Page
EXECUTIVE SUMMARY 1
1. BACKGROUND 3
2. PROJECT OVERVIEW 6
3. FIELD CAMPAIGN 7
3.1. Background 7
3.2. Site Description 8
3.3. Campaign Schedule 9
3.4. Experimental Methods 9
3.4.1. IMACC Passive Fourier Transform Infrared Radiometer i PI- IIR t 9
3.4.2. TELOPS Mid-Wave Infrared Hyper-Spectral Imager (HSI) 10
3.4.3. Optical Gas Imaging Camera 10
3.4.4. Instrument Calibration Procedures 10
3.4.5. Combustion Efficiency Calculations 12
3.5. Data Products and Reports 14
3.6. Results I I
3.6.1. Site #5 Emissions Data 15
3.6.2. Combustion Efficiency Estimates at Various Sites 17
3.6.3. Evaluation of PFTIR and His Sensing Approaches for ECD Assessment 19
4. EPA SPECIATE DATABASE 20
4.1. Background 20
4.2. SPECIATE Database Description 20
4.3. SPECIATE Data Processing and Entry Approach 23
4.3.1. Data Collection 23
4.3.2. Documentation 24
4.3.3. Data Format 24
4.3.4. Speciation Data Quality 24
4.4. Study or Measurement Datasets 26
4.4.1. RARE ECD Study 26
4.4.1.1. Background 26
4.4.1.2. Results 26
4.4.2. WRAP Phase III VOC Speciation Profiles 27
4.4.2.1. Background 27
4.4.2.2. Methodology 27
4.4.2.3. Results 28
4.4.3. Uintah and Ouray Indian Reservation Tribal Minor Source Registration 30
4.4.3.1. Background 30
4.4.3.2. Methodology 31
4.4.3.3. Results 34
4.4.4. Denver-Julesburg Basin Direct Measurement Study 35
4.4.4.1. Background 35
4.4.4.2. Methodology 36
4.4.4.3. Results 37
4.4.5. East Texas Oil Field Speciation Data 37
4.4.5.1. Background 37
4.4.5.2. Methodology 38
4.4.5.3. Results 38
4.4.6. San Joaquin, California Oil and Gas Speciation Data 39
4.4.6.1. Background 39
4.4.6.2. Methodology 40
4.4.6.3. Results 41
iii
-------
4.5. Results of SPECIATE Work 42
5. SUMMARY 44
6. WORK PRODUCTS and FUTURE WORK 46
APPENDIX A: PHOTOGRAPHS OF MEASUREMENT SITES 48
APPENDIX B: SPECIATION PROFILES 55
APPENDIX C: LIST OF PRODUCERS IN UINTAH and OURAY RESERVATIONS 62
Figures: Page
Figure 1: IMACC PFTIR Radiometer 10
Figure 2: TELOPS Mid-Wave Hyper-Spectral Imager 11
Figure 3: Overhead View of Site#5 with Measurement Configurations 15
Figure 4: TELOPS Chemical Map Example Images 17
Figure 5: Daily Calculated Combustion Efficiency from PFTIR 18
Figure 6: Overlay of WRAP Basins 27
Tables: Page
Table 1: Initial List of Target Compounds for Study 8
Table 2: Latitude, Longitude, and Date of Data Acquisition of Eiach Measurement Site 9
Table 3. Summary of PFTIR Concentration Determinations (ppm-m) at Site #5 16
Table 4: Number of Individual Profiles to Develop Composite Profile 28
Table 5: Number of Different Samples for each Operator Profile 33
Table 6: Number of Profiles for Eiach Oil and Natural Gas Production Emission Source 34
Table 7: Number of Profiles for Eiach Oil and Natural Gas Production Emission Source 37
Table 8: Number of Profiles for Eiach Oil and Natural Gas Production Emission Source 39
Table 9: Number of Profiles Collected from each California Study 40
Table 10: Number of Profiles for Eiach Oil andNatural Gas Production Emission Source 40
Table 11: Number of Individual Profiles Developed from Eiach Study Reviewed for This Project 43
IV
-------
Abbreviations:
A&WMA
Air and Waste Management Association
ACE
Air, Climate, and Energy Research Program
API
American Petroleum Institute
ASTM
American Society for Testing and Materials
bbl
barrels of oil
BPA
Beaumont-Port Arthur
BTEX
Benzene, Toluene, Ethyl-benzene and Xylenes
CARB
California Air Resources Board
C02
Carbon Dioxide
CO
Carbon Monoxide
CAS
Chemical Abstract System
CAA
Clean Air Act
CHIEF
Clearing House for Inventories and Emission Factors
CBM
Coal-Bed Methane
CFR
Code of Federal Regulations
CE
Combustion efficiency
DFW
Dallas-Fort Worth
DJ
Denver-Mesb urg
ECDs
Enclosed Combustor Devices
EPA
Environmental Protection Agency
E&P
Extraction and Production
FIP
Federal Implementation Plan
FY
Fiscal Year
FID
Flame Ionization Detection
FTIR
Fourier Transform Infrared
GC
Gas Chromatography
GOR
Gas Oil Ratio
GRI
Gas Research Institute
GPS
Global Positioning System
GHGs
Greenhouse Gases
HAPs
Hazardous Air Pollutants
HVS
High Volume Sampler
HGB
Houston-Galveston-Brazoria
H2S
Hydrogen Sulfide
HSI
Hyper-Spectral Imaging
IMACC
Industrial Monitor and Control Corporation
IPAMS
Independent Petroleum Association of Mountain States
IR
Infrared
IRMS
Isotope Ratio Mass Spectrometer
MS
Mass Spectrometry
ch4
Methane
NAAQS
National Ambient Air Quality Standards
NEI
National Emissions Inventory
NESHAP
National Emission Standards for Hazardous Air Pollutants
v
-------
NSPS
New Source Performance Standards
NSR
New Source Review
NOx
Nitrogen Oxides
NAICS
North American Industry Classification System
OAQPS
Office of Air Quality Planning and Standards
OIG
Office of Inspector General
ORD
Office of Research and Development
OSP
Office of Science and Policy
OGI
Optical Gas Imaging
PM
Particulate Matter
PMio
Particulate Matter less than 10 microns
PM2.5
Particulate Matter less than 2.5 microns
ppb
parts per billion
ppm
parts per million
ppmv
parts per million by volume
PFTIR
Passive Fourier Transform Infrared
PAMS
Photochemical Assessment Monitoring System
PECASE
Presidential Early Career Award for Scientists and Engineers
QA
Quality Assurance
ROG
Reactive Organic Gases
RARE
Regional Applied Research Effort
SBIR
Small Business Innovation Research
SO2
Sulfur Dioxide
TCEQ
Texas Commission on Environmental Quality
TERC
Texas Environmental Research Consortium
TOG
Total Organic Gas
TMS
Tribal Minor Source
U.S.
United States
VMT
Vehicle-Miles Traveled
VOCs
Volatile Organic Compounds
WEA
Western Energy Alliance
WRAP
Western Regional Air Partnership
-------
Executive Summary
Environmentally responsible development of national energy assets requires well-developed
emissions inventories and measurement techniques to verily emissions and to understand the
effectiveness of emissions control strategies. To properly model the energy production sector
impacts on air quality, it is also critical to have accurate activity data, emission factors, and
chemical speciation profiles for volatile organic compounds (VOCs) and nitrogen oxides (NOx).
The upstream oil and gas sector, specifically well pad operations, presents many challenges in
this regard. The vast number and variety of potential emission sources, coupled with company-
specific differences in engineering and maintenance practices, and the natural variability in
product composition make it difficult to understand and properly represent emissions from well
pads. Advancements in understanding this source sector require both speciated inventory
development and measurement techniques that can assess the emissions in the field.
This report describes an United States (U.S.) Environmental Protection Agency (EPA) effort that
aimed to improve the understanding of well pad emissions, the capability of different
measurement methods, and identify areas where future work is needed. Funded through the EPA
Office of Research and Development (ORD) Regional Applied Research Effort (RARE)
program (awarded 2014 to 2016), EPA Region 8, ORD, and the Office of Air Quality Planning
and Standards (OAQPS) conducted a two-phase project to explore a novel measurement
approach for a potentially important oil and gas source, and gather source emission information
to advance the EPA SPECIATE database. The SPECIATE database is a key tool used by air
quality modelers to predict areas of ozone non-attainment. The emphasis of this RARE project
was on product-related VOC emissions from on-going well pad operations.
Currently, there is little information on the potential emissions from non-optimally operating
emissions control systems, called enclosed combustion devices (ECDs), on well pads, and how
often the ECDs do not function properly. Fugitive emissions of VOCs can originate from leaks
and from potentially ineffective control systems, such as ECDs. In the case of ECDs, it is
possible that byproducts of incomplete combustion may produce more highly reactive ozone
precursor species. As a result, the first phase of this RARE project consisted of a limited-scope
field effort that investigated new remote sensing techniques for off-site assessment of ECDs on
well pads.
For both compliance and scientific purposes, the ability to quickly and easily assess ECD
operations from off-site vantage points is potentially important. The exploratory methods study
described in this report represents the first attempt to assess well pad ECD emissions using an
optical remote sensing approach. The limited-scope field demonstration was executed over a
five-day period in September 2014 and produced observations on 10 well pads in the Denver-
Mesburg (DJ) Basin Colorado. The demonstration showed that it may be possible to effectively
assess the operational states of ECDs using remote sensing approaches for compliance purposes,
but improved technologies and further method development are necessary. Of the 10 well pads
investigated, at least one demonstrated clear evidence of improper ECD operations and these
measurements are described in this report. The demonstration also indicated the challenges in
measuring and collecting emissions from these sources that are suitable for use in the SPECATE
database and representing the emissions for air quality models to predict air quality impacts. In
1
-------
particular, the data acquired from the field study were not of high enough quality or quantity for
this purpose. Although this methods demonstration project was not successful in robust
quantification of emissions of highly reactive VOCs from ECDs with current remote sensing
systems, it identified a need for additional research to more efficiently measure emissions from
ECD operations and update emission speciation profiles in SPECIATE for both properly
operating and malfunctioning ECD systems and to improve our understanding of the prevalence
of the latter.
The second phase of this RARE project focused on data synthesis work from existing oil and gas
studies to improve speciated emissions information for this sector. This component of the
project contributed significantly to improvements to EPA's SPECIATE database. This database
is used to develop emissions inventories and is a key tool for air quality modelers to predict non-
attainment areas that may be impacted by oil and gas development. The data synthesis portion
gathered information on VOC emissions from multiple internal and external projects. These data
were utilized to develop process-related VOC speciation profiles that were incorporated into
EPA's SPECIATE database for multiple basins and to improve SPECIATE's representation of
oil and gas processes. This report and associated Excel spreadsheet provides background
information on these efforts, summarizes results, and provides information on areas for potential
future work. While this project produced over 90 new VOC speciation profiles for various oil
and gas processes, it identified gaps in understanding how to link the profiles to inventories or
whether there is a potential need to develop and utilize site- or process-specific VOC speciation
profiles for interpreting ambient measurement data and creating model-ready emissions for
photochemical modeling applications.
2
-------
1 BACKGROUND
The establishment of accurate emissions inventories and the development of field observation
techniques that can help verify emissions and the effectiveness control strategies are key factors
that support environmental responsible development national energy assets. In particular, well-
developed speciated source emission profiles and source activity data that includes knowledge of
the malfunction frequency of control strategies assist modeling efforts that investigate energy
production sector impacts on air quality. The upstream oil and gas sector, specifically well pad
operations, presents many challenges in this regard. The vast number and variety of potential
emission sources, coupled with company-specific differences in engineering and maintenance
practices, and the natural variability in product composition make it a difficult to understand and
properly represent emissions from well pads. Advancements in understanding this source sector
require both speciated inventory development and measurement techniques that can assess
emissions in the field.
The Intermountain West is an important source of domestic energy resources. Operations
involved in the extraction, production, and distribution of oil and natural gas have significant
environmental impacts. One of the primary environmental impacts associated with oil and
natural gas production is related to air emission releases of a number of air pollutants. The
primary air pollutants released include NOx, VOCs, particulate matter (PM), sulfur dioxide
(SO2), Greenhouse Gases (GHGs), and Hazardous Air Pollutants (HAPs). In the case of NOx
and VOCs, these pollutants are important precursors to the formation of ground-level ozone.
Ozone is one of the criteria pollutants regulated by EPA under the Clean Air Act (CAA). Studies
have indicated that the emissions associated with the oil and natural gas sector are highly
uncertain because little information exists on the emissions from this source category. In
particular, a recent EPA Office of Inspector General (OIG) Report found that EPA has limited
directly-measured air emissions data for oil and natural gas sources, and approximately half of
EPA's oil and natural gas emission factors are rated below average or unrated because of
insufficient or low quality data.1
This report describes a U.S. EPA effort that aimed to improve information on emissions and
measurement methods and identify areas where future work is needed. Funded through the EPA
RARE program (awarded 2014 to 2016), EPA Region 8, ORD, and OAQPS conducted a two-
phase project to explore a novel measurement approach for a potentially important oil and gas
source and gather source emission information to advance the EPA SPECIATE database. The
SPECIATE database is a key tool used by air quality modelers to predict areas of ozone non-
attainment. The emphasis of this RARE project was on product-related VOC emissions from on-
going well pad operations.
To reduce the emissions released during the production of oil and gas, most well pads in EPA
Region 8 use control devices, such as ECDs. These control devices are generally assumed to
1 EPA Needs to Improve Air Emissions Data for the Oil and Natural Gas Production Section, Report No. 13-P-0161,
February 20, 2013, Office of Inspector General, EPA.
3
-------
have collection and control efficiencies in excess of 95 percent.2 Properly maintained and
controlled oil and gas extraction and production (E&P) processes are important for limiting the
amount of pollutants released into the air, thereby protecting human health and the environment.
However, recent observations, including visibly smoking ECDs and infrared (IR) camera
footage, suggest that under some operating conditions, capture and control efficiency can be
substantially lower than 95 percent. As a result, it is possible that these poorly maintained E&P
processes can emit VOCs, including air toxics and precursors to ozone, at potentially significant
levels. For ECDs that may not be operating at designed destruction efficiencies, there is also
uncertainty in the composition of combustion byproducts. The composition of the combustion
byproducts or the speciation of the VOC emissions from these sources, as well as other oil and
gas operations, are not only important for understanding the magnitudes of the individual VOC
chemical species being emitted into the air, but are also important for interpreting ambient
measurement data and creating speciated emission inventories for regional haze, climate, and
photochemical air quality models. Currently, there is little information on the potential emissions
and frequency of occurrence of non-optimally operating ECDs and the representation of ECD
emissions. Therefore, research on emissions from well pad ECDs and other oil and gas
operations is needed to improve emissions inventories, to better understand highly reactive VOC
species for air quality modeling, and to aid in the design of fiiture field measurement studies for
the oil and gas sector.
One difficulty with assessment of upstream oil and gas sources is the large number of well pads
in a given basin and the cost and complexity of executing on-site field measurements, which
typically require prearranged site access with the oil and gas operator. This mode of operation
can allow for effective investigation of nominally operating systems, but may obstruct the study
of "as-encountered" systems for the purpose of establishing control effectiveness and the
frequency of process malfunctions. Many well pads are close to public roadways and in theory
can be somewhat eflectivity assessed with regard to gross malfunctions using remote observation
approaches. For both compliance and scientific purposes, the ability to quickly and easily assess
ECD operations from off-site vantage points is potentially important.
The first phase of this RARE project consisted of a limited-scope, exploratory methods study
that represents the first attempt to assess well pad ECD emissions using an optical remote
sensing approach. In addition to methods development objectives, a goal of the field study was
to produce emissions measurements suitable for use in the SPECATE database and representing
the emissions for air quality models to predict air quality impacts.
The second phase of this RARE project focused on data synthesis work from existing oil and gas
studies to improve EPA's SPECIATE database. The SPECIATE database is a key tool for air
quality modelers to assess impacts of oil and gas development in non-attainment areas.
Photochemical air quality models are used to simulate the transport of air pollution and are
important tools in the regulatory process. Within these models, the predictions of major
pollutants, such as ozone, NOx, VOCs, and PM, are represented using simplified chemical
mechanisms and emissions inventories. The common chemical mechanisms within models either
2 Control Techniques Guidelines for the Oil and Natural Gas Industry (Draft), EPA-453/P-15-001, U.S.
Environmental Protection Agency, Office of Air and Radiation, Office of Air Quality Planning and Standards,
Sector Policies and Programs Division, Research Triangle Park, North Carolina, August 2015.
4
-------
group compounds based on reactivity with hydroxyl radicals or break compounds into functional
groups. Further, the emissions inventories are based on the EPA's National Emissions Inventory
(NEI), which contains estimates of total anthropogenic emissions of NOx, VOCs, PM, and other
pollutants in the United States. To utilize the NEI for the models, speciated emission profiles
found in the SPECIATE database are routinely used to convert the total emissions from specific
sources in the emissions inventory into the speciated emissions needed for models.
To improve the predictions of air quality models, the second phase of this RARE effort reviewed
and expanded upon the Total Organic Gas (TOG) speciation profiles stored in EPA's SPECIATE
Database for oil and gas sources to ensure that the most recent and representative profiles are
available to the community.
5
-------
2 PROJECT OVERVIEW
The objective of this project was to advance understanding of speciated VOC emissions from
upstream oil and gas production operations. Specifically, the project aimed to advance
information on methods for off-site remote assessment ofECDs and fugitive emissions, and
improve EPA's SPECIATE database for this sector. These objectives were accomplished
through a collaborative effort with EPA Region 8, ORD, and OAQPS to build upon VOC
measurement and database development projects currently in progress at EPA. In particular, the
project leveraged and built upon a previous effort in EPA Region 8, using funds and the time of
one of EPA's winners of the Presidential Early Career Award for Scientists and Engineers
(PECASE). The PECASE project entitled "Detection and Quantification of Component Level
Emissions at Oil and Natural Gas Production Well Pads Using Remote/Direct Measurements"
fed into the design of a field campaign, while the SPECIATE program3 provided the basis for
developing and adding oil and gas VOC speciation profiles to EPA's SPECIATE Database.
The project occurred in two phases. The first phase of the project consisted of a limited-scope
field campaign in the DJ Basin that provided the first demonstration of two experimental off-site
remote sensing measurements to assess control efficiency and emissions of highly reactive VOCs
from ECDs at well pads. The two instrument used were: (1) a Passive Fourier Transform Infrared
(PFTTR) radiometer (IMACC, LLC, Round Rock, TX, USA), and (2) a mid-wave infrared hyper-
spectral imaging (HSI) camera (Telops, Quebec City, QC, CANADA). The field campaign was
conducted over a course of five days in September 2014. The second phase of the project utilized
the information gathered from multiple oil and gas measurement studies to improve the VOC
speciation profiles used for emissions inventory development and within air quality models.
EPA's SPECIATE program develops and maintains a repository (i.e., SPECIATE Database) of
VOCs and PM speciation profiles of air pollution sources or weight fractions of chemical species
of both VOCs and PM.
This report summarizes the activities and results from this two-phase RARE project.
6
-------
3 FIELD CAMPAIGN
3.1 Background
For the first phase of this project, EPA Region 8 and ORD conducted a pilot demonstration field
campaign in the DJ Basin to investigate remote assessment of ECD performance using two off-
site observation approaches. Previous EPA field studies have shown that methane and VOCs
can be emitted from well pad sources that are improperly maintained or controlled.3'4'5'6 ECDs
are commonly used as control devices to control VOC emissions from well pad sources such as
atmospheric storage tanks. To support energy development practices with minimal
environmental impacts, it is important to develop easy-to-use and reproducible measurement
techniques that can verify the effectiveness of ECD operation in the field. Remote sensing
systems may provide a means to improve understanding of ECD operation without direct onsite
sampling of the combustion plume.
The goals of the field campaign were to demonstrate and evaluate the performance of
measurement technologies to characterize emissions from ECDs at upstream oil and gas
production sites, and to the extent feasible, provide speciated emissions information from the
ECDs, and assess the combustion efficiency of the ECDs. The initial list of target compounds for
the campaign is presented in Table 1.
The pilot campaign was executed over a course of five days in September 2014 and aimed to
collect emissions data for the compounds listed in Table 1 from ECDs at multiple well site
locations in an active natural gas field in Weld County, Colorado. The campaign utilized two
primary instruments that could characterize ECD performance on well pads from remote vantage
points, including (1) aPFTTR radiometer (IMACC, LLC, Round Rock, TX, USA), and (2) a
mid-wave infrared HSI camera (Telops, Quebec City, QC, Canada). The two technologies have
been used in previous studies to characterize emissions from industrial flares.7 An optical gas
3 Brantley, H.L.; Thoma, E.D.; Squier, W.C.; Guven, B.B.; Lyon, D. Assessment of Methane Emissions from Oil
and Gas Production Pads Using Mobile Measurements; Environ. Sci. Technol., 2014, 48, 14508-1451, doi:
10.1021/es503070q.
4 Modrak, M.T.; Amin, M.; Ibanez, J.; Lehmann, C.; Harris, B.; Ranum, D.; Thoma, E.D. ; Squier, B.C.
Understanding Direct Emission Measurement Approaches for Upstream Oil and Gas Production Operations, Control
No. 2012-A-411-A&WMA, Proceedings of the 105 th Annual Conference of the Air & Waste Management
Association, June 19-22, 2012, San Antonio, Texas.
5 Thoma, E.D.; Squier, B.; Olson, D.; Eisele, A.; DeWees, J.; Segall, R; Amin, M.; Modrak, M. 2012, Assessment
of Methane and VOC Emissions from Select Upstream Oil and Gas Production Operations Using Remote
Measurements; Control No. 2012-A-21-A&WMA, Proceeding of the 105th Annual Conference of the Air & Waste
Management Association, June 19-22, 2012, San Antonio, Texas.
6 Brantley, H.L.; Thoma, E.D.; Squier, W.C.; Eisele, A.P. Remote and Onsite Direct Measurements of Emissions
from Oil and Natural Gas Production, Abstract #551, Proceedings of the 108th Annual Conference of the Air &
Waste Management Association, June 23-26, 2015, Raleigh, North Carolina.
7 Texas Commission on Environmental Quality, PGA No. 582-8-862-45-F Y09-04, Tracking No. 2008-81 with
Supplemental Support from the Air Qaulity Research Program TCEQ Grant No. 582-10-94300, TCEQ 2010Flare
Study Final Report. David Allen, Vincent Torres, University of Texas at Austin, The Center for Energy and
Environmental Resources, August 1,2011 (accessed January, 2015).
http://www.tceq.texas.gOv/assets/public/implementation/air/rules/Flare/201Qflarestudy/2010-flare-study-final-
report.pdf. Last Accessed: June 2016
7
-------
imaging (OGI) camera (GF-320, FLIR Systems, Inc., Boston, MA, USA) was also deployed to
provide complementary qualitative data on the ECD emissions. Additionally, measurements of
ECD temperature were collected remotely using the OGI camera and a hand-held infrared
thermometer (Fluke Corporation, Everett, WA, USA).
Table 1. Initial List of Target Compounds for Study
hi lid
Ins Iriiiiioul
Butane
Passive FTIR
Ethene
Passive FTIR
Propene
Passive FTIR
Methane
Both1
Benzene
Mid-W ave Hyper-Spectral Imager
Toluene
Mid-W ave Hyper-Spectral Imager
Xylenes
Mid-W ave Hyper-Spectral Imager
Ethylbenzene
Mid-W ave Hyper-Spectral Imager
Total Hydrocarbons
(aggregate of compounds with carbon number greaterthan 4)
Both
Carbon Monoxide
Both
Carbon Dioxide
Both
Water Vapor
Both
Ammonia
Passive FTIR
Nitrous Oxide
Both
Nitrogen Dioxide
Both
^oth = Passive FTIR and Mid-Wave Hyper-Spectral Imager
Remote sensing measurements were conducted from safe and appropriate offsite observing
locations on the side of public roadways, and the sites were selected based on a combination of
factors including ease of observation and the results of OGI imaging that indicated the potential
presence of ECD operational issues. This research effort was not part of any enforcement or
compliance activity.
The subsequent sections describe the measurement technologies, including details of instrument
operation, deployment, and data analysis methods, as well as a discussion of data results from the
campaign.
3.2 Site Description
Measurements were conducted in Weld County, Colorado, which is located approximately 50
miles northeast of the Denver Metropolitan area. A total of ten well pads were surveyed during
the campaign. Both instruments were not able to collect data at all ten sites because of time and
resource limitations. Both instruments were only employed together at three of the ten sites. The
general locations of the sites were in areas with a large density of active well sites.
Measurements were collected from the side of public roadways adjacent to the emission source,
with approximately four well sites surveyed per day. Table 2 outlines the location of each
8
-------
measurement site and information related to the data collected by each of the instruments. The
rows filled in yellow represent sites that had data from both instruments. Photographs of the site
locations are included in Appendix A.
Table 2. Latitude, Longitude, and Date of Data Acquisition o:
Site
Latitude
Longitude
Date
Time
[Locall
Number of ECDs
at Site
Instrument
Site #1
40.3197 N
-104.5688 W
09/08/2014
13:00
1
PFTIR
Site #2
40.2040 N
-104.8147 W
09/08/2014
15:30
2
PFTIR
Site #3
40.1889 N
-104.7582 W
09/09/2014
08:00
3
PFTIR
Site #4
40.3343 N
-104.6232 W
09/10/2014
10:00
2
PFTIR/HSI
Site #5
40.1314 N
-104.6896 W
09/11/2014
09:30
2
PFTIR/HSI
Site #6
40.1168 N
-104.7078 W
09/11/2014
12:00
4
PFTIR
Site #7
40.1314 N
-104.6896 W
09/11/2014
12:30
10
HSI
Site #8
40.1167 N
-104.6922 W
09/11/2014
14:30
4
PFTIR/HSI
Site #9
40.1221 N
-104.6973 W
09/11/2014
16:00
1
HSI
Site #10
40.3385 N
-104.7934 W
09/12/2014
09:00
2
PFTIR
Each Measurement Site.
3.3 Campaign Schedule
The field campaign was conducted from September 8-12,2014 in Weld County, Colorado. Data
were collected with the PFTTR during each day of the campaign, and with the mid-wave infrared
HSI on September 10-11, 2014. Table 2 outlines the date and time of the data acquisition for
each instrument over the course of the campaign.
3.4 Experimental Methods
The instruments utilized during the field campaign, calibration procedures and combustion
efficiency calculations are discussed in the following sub-sections.
3.4.1 IMACC Passive Fourier Transform Infrared Radiometer (PFTIR)
The Industrial Monitor and Control Corporation (IMACC) was responsible for all on-site
measurement activities using the PFTIR. The PFTTR radiometer analyzes thermal radiation
emitted by hot gases in the ECD plume. During the measurement process, the instrument does
not transmit an infrared light source through the measurement plane. Instead, infrared energy
emitted from the hot gases from the source is the infrared signal, and the instrument acts as a
receiver. This approach is possible because the emissions spectra of hot gases are very similar to
their absorption spectra, and can therefore be used for identification and quantification of species
through emission spectroscopy, just as with absorption spectroscopy. The PFTTR has been used
successfully on large open industrial flares, but this was the first attempt to use the approach on
smaller (lower temperature) ECDs with high duty cycles. The PFTTR was chosen for the current
study instead of active open-path Fourier Transform Infrared (FTTR) monitoring (where the
instrument transmits and receives an infrared source) because of potential difficulties of
transmitting and receiving an infrared source through a small, elevated plume, and the need for
site access and support infrastructure to position the required beam retroreflectors.
9
-------
Data from the EV1ACC PFTTR radiometer were used primarily to determine the combustion
effic iency of the flares at each site. Data from the PFTIR were also used to provide information
on the constituents present in the flare plumes. However, the PFTIR measurement approach
cannot provide actual gas concentrations. To obtain actual gas concentrations, accurate
knowledge of both the path length through the plane and plume temperature are required. For
this reason, gas concentrations determined for the project are considered estimated values
obtained using estimated values of path length through the plume and plume temperature. The
raw PFTTR data were post-processed using IMACC proprietary software.
The instrument was deployed from the back of a field trailer, and was mounted on a mechanical
positional scanner. At each measurement site, the trailer was oriented to provide a clear line of
sight between the PFTTR and the emissions source detected during the pre-measurement OGI
camera survey. Measurements were collected at 0.5 per centimeter (cnr1) spectral resolution, and
the instrument field of view was approximately 14 inches in diameter. Each data point was
averaged for 30 seconds, with analyte concentrations in units of parts per million-meter (ppm-
m). Because many of the plumes measured during the campaign showed weak infrared signal
due to overall low combustion throughput, often times it was not possible to make a valid
measurement. A data filter was developed to eliminate data points with insufficient infrared
signal. Due to the lack of sustained infrared signal from many ECDs, the PFTIR was not as
effective as compared to industrial flare applications. The PFTTR was deployed during each day
of the field campaign at a total of eight well pad sites. Figure 1 presents the IMACC PFTTR
radiometer.
Figure 1. IMACC PFTIR Radiometer.
3.4.2 TF.LOPS Mid-Wave Infrared Hyper-Spectral Imager (HSI)
TELOPS was responsible for all on-site measurement activities using the HSI. The mid-wave
infrared HSI is a standoff instrument that uses FTTR technology. Incoming infrared radiation
from the vicinity of the source being monitored is modulated using a Michelson interferometer
located inside the instrument. A high-resolution spectium is then recorded tor each pixel of a
focal plane array detector. By comparing the measured spectrum to a series of reference spectra
of known gases, the constituent species can be identified and quantified. The instrument has a
nominal spectral range of 3 to 5 microns, with spectral resolution of 0.25 cm-1 wavenumber. The
instrument field of view consists of 128 by 128 pixels, with individual pixel size ranging from
10
-------
approximately 5 to 166 cm2, depending on the distance from the instalment to the source. Data
collected with the instrument are used to create quantitative chemical imaging sequences,
showing the column density, in units of ppm-m, for detected compounds in the instalment field
of view. Figure 2 shows the Telops mid-wave hyper-spectral imager, and an example chemical
map showing carbon monoxide column density.
CO Quantitative Chemical Imaging toap
Figure 2. TELOPS Mid-Wave Hyper-Spectral Imager.
The instrument was mounted on a heavy-duty tripod, and was deployed from the back of a field
vehicle, which housed the data acquisition/control computer. The instrument was deployed at
each measurement site in a location that provided a clear line of sight to the emissions source, as
determined during the pre-measurement infrared camera survey. Measurements were collected
with the mid-wave hyper-spectral imager during two days of the field campaign (September 10-
11) at a total of five well pad sites.
3.4.3 Optical Gas Imaging Camera
The OGI camera was deployed to provide complementary, qualitative data on ECD flares. The
camera was deployed at each site concurrently with the PFTIR and US I. and operated by EPA
Region 8 personnel. The OGI was used to confirm that hydrocarbon8 emissions were present in
the flare plume, and utilized to document any potential changes in the plume characteristics
during the measurement period. Although the camera was used at times to collect videos
simultaneously while data were being collected with the other instrumentation, the camera was
not operated continuously due to limitations in the capacity of the camera flash drive. Data
collected with the OGI were considered non-critical for this project.
3.4.4 Instrument Calibration Procedures
8 "Hydrocarbons" are VOC species that consist entirely of hydrogen and carb on and include most of the primary
VOC emissions species from oil and gas Operations, including alkanes and aromatic species. Hydrocarbon doesnot
include oxygenated VOC species that are typically formed via oxidation of hydrocarbons, however, Some oil and
gas production activities also emit oxygenated VOC directly, e.g., formaldehyde is formed as a contamination
product from methanol, and both formaldehyde and methanol are emitted directly to the atmosphere. The
instruments used in this study were not able to detect oxygenated VOC emissions.
11
-------
After the instrumentation was deployed, calibration procedures were conducted prior to
collection of measurements. At the beginning of each day, the PFTIR team performed a series of
calibrations using a custom-designed calibration cart that contained a telescope and various
calibration materials. The PFTIR team performed a black body calibration using a black body
with an infrared source of known spectral radiance, an infrared source calibration to determine
atmospheric transmission loss between the flare plume and the PFTIR, a cold source calibration
to determine radiance generated by the atmosphere between the flare plume and the PFTIR, and
a sky background calibration to determine the background radiance from the sky. The
calibrations were typically done once per day (at the first site of each day), although the sky
background calibration was conducted more frequently when sky conditions varied during the
measurement period. Deployment of the PFTIR and completion of the pre-measurement
calibrations were completed at each site in one to two hours.
The HSI camera also required preliminary system setup and calibration. The initial system setup
and calibration were completed at the first site where the HSI was deployed in approximately 90
minutes. System setup and calibration at subsequent sites were generally completed in 5 to 10
minutes.
After deployment and system calibrations were completed, information collected during apre-
deployment OGI camera survey was used to locate the ECD emissions, and align the PFTIR and
mid-wave HSI on the source. Measurements were collected at each site for one to four hours,
depending on the frequency and duration of emissions observed from the ECD. The data
collected were analyzed for pollutants listed in Table 1, along with combustion efficiencies using
data from the PFTIR.
3.4.5 Combustion Efficiency Calculations
Several EPA regulations, including New Source Performance Standards (NSPS)9 and National
Emission Standards for Hazardous Air Pollutants (NESHAP),10 require facilities to use good air
pollution control practices to minimize the emissions released into the environment. Because not
all waste gas emissions can be prevented or recovered, various control technologies are used to
reduce the impact of these waste streams. For instance, EPA regulations require facilities to
install control devices, such as flares, on operations that emit waste gases into the atmosphere. A
flare is a mechanical device used to combust and thereby destroy volatile organic compounds,
toxic compounds, and other pollutants from oil and gas operations, refineries, and other
industrial facilities. Many flares employ steam or air as assist gases to improve flame stability
and to promote mixing of oxygen within the vent gas to ensure combustion occurs without
smoke. Completeness of combustion in a flare is governed by flame temperature, residence time
in the combustion zone, turbulent mixing of the gas stream components to complete the
oxidation reaction, and available oxygen for free radical formation. Combustion is complete if all
VOCs are converted to carbon dioxide and water. Incomplete combustion results in some of the
VOCs being unaltered or converted toother organic compounds such as aldehydes or acids.
9 EPA NSPS: https://wmm'.epa.gov/conpliance/denionstrating-conB3liance4ieW"Source-perfoniiance-standards-and-
s tate-impfomentation -plans. Last Access Date: June 2016.
10 EPA NESHAP: https://www.epa.g-ov/coni3liance/national-eniission-standards4iazardous-air-pollutantS"
colli)liance-nionitonng. Last Access Date: June 2016
12
-------
There are many parameters that affect the combustion efficiency of a flare. One important
parameter is the heating value of the gases that are to be combusted, often measured in British
Thermal Units per standard cubic foot (BTU/sci). Generally, it is easier to maintain a stable
flame and achieve high efficiency for gas streams with higher heating values. The NSPSand
NESHAP requirements regulate the net heating value and require that gases contain at least 300
BTU/scf if they are being combusted in an air- or steam-assisted flare. If this heating value
minimum cannot be met by the vent gases alone, then supplemental gas, such as natural gas,
must be added.
The current regulations require control devices to achieve 98 percent destruction efficiency.
However, because most of the flare data is reported in terms of combustion efficiency, it was
necessary to estimate a combustion efficiency equivalent to 98 percent destruction efficiency as a
means for determining good performance for flares. Note that destruction efficiency is a measure
of how much of the hydrocarbon is either fully or partially oxidized; and combustion efficiency
is a measure of how much of the hydrocarbon is fully oxidized to yield carbon dioxide and water
vapor.
Based on a series of flare performance studies conducted in the early 1980s, the EPA concluded
that properly designed and operated flares achieve good combustion efficiency (e.g., greater than
98 percent conversion of organic compounds to carbon dioxide).11 However, flares operating
outside "their stable flame envelope" produced flames that were not stable or would rapidly
destabilize, causing a decrease in both combustion and destruction efficiency (i.e., 40 Code of
Federal Regulations (CFR) 60.18 and 40 CFR 63.11(b)). Other EPA studies also found that the
combustion efficiency will always be less than or equal to the destruction efficiency; and a flare
operating with a combustion efficiency of 98 percent can achieve a destruction efficiency in
excess of 99.5 percent. 12 The relationship between destruction and combustion efficiency is not
constant and changes with different compounds. However, these studies estimate that a 1.5%
difference is a reasonable assumption. 13
One of the goals of the project was to evaluate the combustion efficiency of ECDs at oil and gas
well pads using data collected with the PFTTR. When ECDs are operating properly, efficient
combustion is achieved by converting hydrocarbons to carbon dioxide and water. However,
inefficient combustion occurs when the oxygen supply to the ECD is insufficient, forming
products of incomplete combustion such as carbon monoxide, intermediate hydrocarbons, and
carbonyls. Carbonyls include formaldehyde and other aldehydes that are HAPs and are also
highly reactive precursors of ozone. Combustion efficiency (CE) is defined as the ratio of the
mass concentration of carbon dioxide to the sum of the concentrations of carbon dioxide, carbon
monoxide, and total hydrocarbons in the plume and is expressed using the following equation
(1):
^ i , sn/\ [Carbon Dioxide]
Combustion Efficiency (/o) (1)
[Carbon Monoxide]+[Carbon Dioxide]+\Total Hydrocarbons]
11 EPA Flare Report: https://www3.epa.g-ov/airtoxics/flare/2012flaretechreport.pdf. Last Access Date: June 2016.
12 EPA Flare Report: https://www3.epa.gov/airtoxics/flaiig/2012flaretechreport.pdf. Last Access Date: June 2016.
13 EPA Flare Report: https;//www3.epa.gov/airto^cs/flare/2012flaretechreport.pdf. Last Access Date: June 2016.
13
-------
The fundamental output of the remote sensing instalments used in this project was the gas
concentration times the path length of the gas, or ppm-m. Since the path length of all gases in the
plume is the same, the path length cancels in the ratio given in the equation above. Consequently,
for the CE calculation, knowledge of the actual path length through the measured plume is not
necessary. At this time, only the PFTTR had the capability of generating CE values but the HSI
protocols for determination of CE values are under development by the manufacturer.
3.5 Data Products and Reports
Reporting requirements for this project included relative concentrations of the target compounds
listed in Table 1 measured with the PFTTR, estimated concentrations of the target compounds
measured with the PFTTR, concentrations of the target compounds measured with the HSI,
combustion efficiency calculations for each site using data from the PFTTR instrument, Global
Positioning System (GPS) coordinates of each site and measurement locations, and videos from
each site collected with the OGI. Personnel from IMACC and TELOPS were responsible for all
data acquisition with the PFTTR and HSI respectively, with oversight by Arcadis. Personnel from
EPA Region 8 were responsible for data acquisition with the OGI. Deliverables included:
Interferograms collected with the PFTTR;
Interferograms collected with the HSI;
GPS coordinates of each site and each measurement location; and
OGI videos.
After the data were analyzed and validated, IMACC and Telops submitted a short-form report of
the results and a project data package to Arcadis. Arcadis compiled the information and
submitted a report and project data package to EPA. Additionally, the descriptions of the
measurement technology and results of the field campaign were reported and presented at the
June 2015 Air and Waste Management Association (A&WMA) Conference in Raleigh, North
Carolina14.
3.6 Results
The instruments were deployed at a total of ten representative well pads during the field
campaign (see Table 2). Both instruments collected data at three of the 10 sites, otherwise data
were collected by only one of the instruments. As a general observation, it was apparent to the
field teams during deployment that the emission signals observable by the remote sensing
instruments were fairly weak at most of the well pads. The amount of gas flowing to and
combusted by the ECD raises the temperature of the ECD stack and the emitted plume. If the
temperature of the emitted plume becomes too low, the radiated signal to the PFTTR or HSI
becomes insufficient for acquisition of usable data. Since the load on the ECD is time-dependent
(increases during periodic separator dumps), the signal available to the remote sensing
equipment changes with time. One result of the pilot study is the observation that remote
assessment of lower temperature ECDs is a more difficult task than higher temperature facility
14 https://cfpub.epa.gov/si/si_public_record_report.cfm?dirEntryId=309972&subject=Health%2520Research
&showCriteria=0&searchAll=Environmental%2520Justice&sortBy=revisionDate [Access Date: June 2016],
14
-------
flares. For the equipment configurations and source strengths encountered in this pilot study, the
emission signals were marginal in many cases. For this reason, the effective detection limits of
specific compounds are variable and the quantity of usable data was generally low.
At Sites 1 and 10, the ECDs were not sufficiently active during the observation period to execute
the measurement (although emissions were detected during the pre-deployment gas imaging
camera survey). As a result, no usable data were collected fr om the PFTIR. Analysis of data
collected with the PFTIR indicated that most of the ECDs showed relatively high combustion
efficiency values (close to, or exceeding 0.95), with little to no detected hydrocarbon emissions.
However, data collected with both instruments at Site #5 indicated emissions of hydrocarbons
greater than emissions found at the other sites. A summary of the data collected at Site #5 is
included below.
3.6.1 Site #5 Emissions Data
Data were collected at Site #5 with the PFTIR and HSI for approximately 2.5 hours on
September If, 3014. The PFTIR and HSI were deployed approximately 45 and 78 meters,
respectively, from an active ECD stack at the site. Figure 3 presents an overhead view of Site #5,
showing the location of the PFTIR and HSI during the measurements.
HSI Camera Location
Distance to Source= 78 m
PFTIR Location
Distance to Source= 45 m
Figure 3. Overhead view of Site #5 with measurement configurations.
The PFTIR detected emissions of several hydrocarbons from the stack. A summary of path-
integrated concentrations determined from data collected with the PFTIR is presented in Table 3.
The table shows that most of the lighter alkanes and alkenes were detected in emissions from the
site, with relatively higher concentrations of methane, ethane, and pentane.
15
-------
Measurements collected with the HSI also indicated the presence of hydrocarbons in emissions
from Site #5. Specifically, analysis of the data detected methane, propane, and butane, as well as
carbon monoxide and formaldehyde. Figure 4 presents quantitative chemical map sequences for
several compounds detected at Site #5. The top of the ECD stack is located in the lower right
hand portion of each image. The maps present a spatial distribution of path-integrated
concentrations (in units of ppm-m) across the ECD plume. The figures in the left column
represent the path-integrated concentrations during the burst and the figures in the right column
represent the path-integrated concentrations during the fade out combustion stages. The burst and
fade correspond to flash gases that were routed to an ECD as result of a separator dumping
liquids to atmospheric storage tanks.
Table 3. Summary of PFTTR concentration determinations (ppm-m) at Site #5.
('<>1111X11111(1
Mi milium
M;i\i mum
\UTilJiC
Carbon Dioxide
2,960
61,500
18,300
Carbon Monoxide
0
510
184
Methane
92.0
485
258
Ethane
0
224
58.3
Propane
0
36.4
10.8
Pentane
0
197
50.7
Ethene
0
80.0
40.4
Propene
0
53.2
21.0
Cyclopentene
0
57.1
35.8
Total Hydrocarbons1
92.0
3140
1170
Computed as carbon-weighted sum of all CI through C5 hydrocarbons
16
-------
Figure 4. Quantitative chemical map sequences of carbon monoxide, methane, propane, butane,
and formaldehyde emissions from enclosed combustion device at Site #5 during combustion
stages, burst of heavy fuel load (left) and as the fuel load fades out (right).
3.6.2 Combustion Efficiency Estimates at Various Sites
The combustion efficiency of the ECDs at each site was determined from the field measurements
using Equation 1. Daily plots of CE values for data that passed quality assurance (Q A)
acceptance criteria from four of the six sites measured using PFTIR are presented in Figur e 5.
The preliminary QA acceptance criteria were based on an analysis of sufficient signal level from
the source as determined by the strength and spectral analysis fit quality of the observed plume
compound (CO for example), compared to the noise level in the analysis region. The
determination (amount of CO) must exceed the residual noise by several factors to be considered
a robust detection. Insuffi cient signal can be caused by number of factor such as misali gnment
17
-------
of the observation region of the instrument (not looking at the plume), insufficient plume size
and temperature, or the absence of the compound of interest in the plume. Standard definitions of
PFTIR detection sensitivity and QA acceptance criteria for small intermittent sources such as
ECDs is the subject of future work.
Calculated Combustion Efficiency Values: 09/08/2014
A
Calculated Combustion Efficiency Values: 09/09/2014
B
~
~ ~
?06
?06
i0-6
ฃ
1
E
!ฐ's
E 0.4
|0-5
J 0.4
e u-*
0J
0.1
A
3:28:48
3:16:00 3.43:12 3:50:24 3:5736 4:04:48 4:1200
Time
8:24:00
9:36.00 10:48:00 12:004)0 13:12:00 14:24:00 1S:36:00
16:48:00
Calculated Combustion Efficiency Values: 09/10/2014
Calculated Combustion Efficiency Values: 09/11/2014
D
1
A
n % t *
c
0.9
0.9
~ ~
04
~
0.8
0.7
~ ~ *
0.7
8"
I 0.6
i
i
S o.s
1"
ฃ
| 0.4
0.3
Sitp
#5
0J
0.1
9:364)0
10:4860 12:00:00 13:12:00 14:24.-00 15:36:00
16:48.00
9:364)0
10:48:00 124)0.-00 13:12:00 14:24:00 15:36:00
Tlmt
16:48:00
Figure 5. Daily calculated combustion efficiency values from PFTIR data. Plots shown for
September 8 (A), September 9 (B), September 10 (C), and September 11 (D).
For the study sites, the majority of the calculated CE values are close to, or greater than 98%.
There are periods of time where the CE appears to fall below 90% and this is most obvious at
Site 5 (Figure 5D). These time periods of relatively lower CE values directly correspond with the
observation of increased measurable hydrocarbon emissions from this site (see Table 3). From
simultaneous OGI observations, these time periods could potentially be related to flash emission
events. In general, when high CE values are registered by the PFTIR, the presence of speciated
hydrocarbon emissions could not be confirmed (e.g., Table 3), as concentrations were most
likely below instrument detection limits. Regarding accuracy of CE estimates for ECDs,
significant uncertainty exists from a method development standpoint. Due to the small flare size
and low temperatures encountered, the determination of CE is generally more difficult than for
larger flare systems.15 Additional method development work would be required to determine the
15 Texas Commission on Environmental Quality, TCEQ 2010 Flare Study Final Report. See:
http ://www.tceq.texas. gov/assets/pub lie/imp lementation/air/rules/Flare/201Qflarestudy/2010-flare -Study -final-
report.pdf (accessed January, 2015).
18
-------
absolute accuracy of the CE measurement for ECDs. The apparent time dependency of the CE is
believed to be real, however the magnitude of these drops and the accuracy with which it is
determined carries significant uncertainty. Future studies conducted on the use of PFTTR for
remote ECD CE assessments would focus on the relationship between the ECD temperature,
plume size, and variability, with respect to PFTTR effective field of view.
3.6.3 Evaluation ofPFTIR and HSI Remote Sensing Approaches for ECD Assessment
In general, the remote sensing approaches used in the field campaign were found to be
potentially useful for offsite observation of ECD operation for research purposes, if direct onsite
measures were not available. However, clear improvements in the instrumentations and methods
would be required to improve signal strength understanding of measurement uncertainty since
the signals from ECDs are not strong and are temporally variable. Limitations were found in
ease of execution (setup and use of equipment), data analysis throughput, and observable ECD
temperature ranges. Due to the lack of sustained infrared signal from many ECDs, the PFTTR
was not as effective as compared to industrial flare applications. The accuracy of CE
determination with the PFTTR approach requires additional investigation and may be
complicated by the small size and variability of the ECD plume. As evidenced from Figure 4, the
HSI approach provides a superior diagnostic of plume heterogeneity compared to the single
element (non-imaging) PFTTR. However, the vast amount of data provided by the HSI approach
make even simple determination of CE challenging, and requires significant method
development work. Both techniques are best characterized as high-asset value research tools
requiring significant set up time and data processing resources, making them relatively
impractical for routine use. In the future, other types of emerging multi-channel remote sensing
approaches may provide what is essentially a combination of aspects of the two instruments used
in the campaign, but in a more implementable form. Development of one multi-channel remote
sensing approach for flare CE measurements is the subject of a recently announced EPA ORD
Phase II Small Business Innovative Research Award.16'17
16 U.S. EPA Phasell SBIR award; Development ofReal-Time Flare Combustion Efficiency Monitor, 2014,
Providence Photonics, LLC: Baton Rouge, LA, USA (Accessed February 2015)
http://cfpub.epa.gov/ncer_abstracts/indexcfm/fuseaction/outlinks.sbir/fullList/Yes/showYear/current.
17Zeng, Y. 2012. White Paper on A Calibration/Verification Device for Gas Imaging Infrared Cameras. Providence,
June 25, 2012. http://www.providenceeng.com/services/technology/Optical-Gas-Imaging (accessedFebruary 2015).
19
-------
4 EPA SPECIATE DATABASE
4.1 Background
For the second phase of this project, EPA Region 8, ORD, and OAQPS utilized information from
multiple oil and gas measurement studies, to evaluate total organic gas (TOG) speciation profiles
associated with oil and natural gas sector. The TOG speciation profiles are important for
interpreting ambient measurement data and developing model-ready emissions for
photochemical modeling applications. This modeling provides the foundation for air quality
management decisions and is a critical input to air quality models used to demonstrate attainment
of the National Ambient Air Quality Standards (NAAQS) or the prediction of air quality
impacts.
Photochemical air quality models are used to simulate the transport of air pollution and are
important tools in the regulatory process. Within these models, the predictions of major
pollutants, such as ozone, NOx, VOCs, and PM, are represented using simplified chemical
mechanisms and emissions inventories. The common chemical mechanisms within models either
group compounds based on reactivity with hydroxyl radicals or break compounds into functional
groups. Further, the emissions inventories are based on the EPA's National Emissions Inventory
(NEI), which contains estimates of total anthropogenic emissions of NOx, VOCs, PM, and other
pollutants in the United States. To utilize the NEI for the models, speciated emission profiles are
routinely used to convert the total emissions from specific sources in the emissions inventory
into the speciated emissions needed for models.
To improve the predictions of air quality models, a goal of this project included reviewing and
expanding upon the TOG speciation profiles stored in EPA's SPECIATE Database18 for oil and
gas sources to ensure that the most recent and representative profiles are available to the
community. SPECIATE is a key tool used to develop speciated emission inventories for regional
haze, PM, GHGs, and photochemical air quality modeling. The Database can also be used for
estimating hazardous and toxic air pollutant emissions from PM and organic gas primary
emissions. It should be noted that most HAPs emissions are developed using test data or
emissions factors that are more specific to a detailed process and pollutant.
Prior to this project, the SPECIATE database contained non-location-specific speciation profiles
for the oil and gas sector. Many of these profiles were based on test data collected in 1989, 2000,
or 2004, and added to the SPECIATE Database in 1989, 1999, 2007, 2010, or 2013. In general,
the speciation profiles covered the following oil and gas areas or processes (year included either
represents date of test data (DT) or date added to SPECIATE (DA)):
External Combustion Boiler (DA=1989): Residual Oil; Distillate Oil; Natural Gas;
Refinery Gas
Natural Gas: Production (DA=2013); Transmission (DA=2013); Distribution (DT=2004);
Extraction (DA=2013); Turbine (DA=1989), Flares (DA=1989); Internal Combustion
Engine (DA=1989)
18
20
-------
Crude Oil Production (DA=1989): Fixed Roof Tank; Gathering Tanks; Storage Tanks
Composite
Oil Field (DT=1989/DA=2013): Pipeline Tanks; Extraction Wells; Gage Tank; Shipping
Tank; Surge Tank; Compressor; Separator; Dehydration Tank; Vapor Recovery; Sump
Oil and Gas Production (Fugitives) (DA=1989): Unclassified; Valves and Fittings -
Liquid Service; Valves and Fittings - Gas Service
Oil and Gas Extraction (DT=2004): Conventional; Non-Conventional; Services
Well Heads (Water Flood) Composite (DA=1989)
Petroleum Storage Facilities Composite (DA=1999)
Liquefied Petroleum Gas Composition (DT=2000)
Well Heads (Gas Drive) (DA=1989)
Reciprocating Diesel Engine (DA=1989)
Initially, a large number of the non-basin specific profiles were used in the 2011 NEI.19'20
However, improvements were made in more recent versions of the 2011 NEI to associate sources
to more basin-specific speciation profiles. Given this information, it is important to develop more
recent and process-specific speciation profiles for the oil and gas sector.
This project amended an existing EPA Work Assignment for the analysis and additions to the
SPECIATE Database specifically related to TOG emissions from oil and gas operations. With
contract support, this project surveyed the community for measurement data, reports, and
publications associated with TOG emissions associated with these sources. This included
information and data from:
1. RARE ECD Study;
2. WRAP Phase in Speciation Profiles;
3. Uintah and Ouray Indian Reservations Tribal Minor Source Registrations;
4. Denver-Julesburg Basin Direct Measurement Study;
5. East Texas Oil Field Speciation Data; and
6. San Joaquin, California Oil and Gas Speciation Data.
The data were consolidated into EXCEL spreadsheets, developed into speciation profiles,
reviewed for completeness and representativeness, analyzed for differences among the available
profiles, and prepared for entry into the SPECIATE Database. Other than the RARE ECD Study,
these profiles will be included in the SPECIATE version 4.5 public release, planned for fall of
2016. The RARE ECD Study did not provide usable profiles because it was difficult to
determine whether all of the targeted compounds were properly measured by the instruments.
This was a result of either the emission signals observable by the remote sensing instruments
being too weak at the well pads, or the limitation of the measurements.
19 2011 NEI Emissions Inventory Technical Support Document:
http://www.epa.gov/ttn/chief/enich/201 lv6/201 !v6.1_ '2SJjaseJEmisMod_TSI)jov2014jv6, pdf. Last
Accessed: June 2016.
20 2011 v6.1 Emissions Modeling Platform Technical Support Document:
https://www.epa.gOv/sites/production/files/2015-08/docunients/201 lv6.1_2018_2025 base_eniisniod tsd
nov2014 v6.ptif. Last Accessed: June 2016.
21
-------
The subsequent sections describe the SPECIATE Database, the datasets utilized for this project,
the data processing and SPECIATE Database entry methodology, and an overview of the TOG
speciation profiles developed from this project.
4.2 SPECIATE Database Description
EPA develops and maintains a repository (i.e., SPECIATE Database) of TOG and PM speciation
profiles of air pollution sources or weight fractions of chemical species of both TOGs (e.g.,
VOCs) and PM. The SPECIATE Database was computerized in 1988 and the first electronic
version was distributed to the user community in 1993. The development and continuous update
of the SPECIATE Database support EPA's ORD Air, Climate, and Energy Research Program
(ACE).21 In particular, this effort supports the assessment of impacts associated with air
pollutants at various spatial scales, and provides data and tools to develop and evaluate
approaches to prevent and reduce emissions of pollutants to the atmosphere.22
The SPECIATE Database is available to the public through EPA's Clearing House for
Inventories and Emission Factors (CHIEF) website. In general, the most recent version of the
database, SPECIATE 4.4, includes comprehensive speciation of TOG profiles from oil and gas
fugitive emissions, gasoline vehicle exhaust, VOC emissions from the dairy industry (including
silages, other feedstuffs, and animal waste), gasoline vapor from enclosed fuel tanks, PM profiles
from the Kansas City Light-Duty Vehicle Emissions Study, outdoor wood boiler aerosol
emissions, and commercial aircraft jet engine PM emission profiles. The SPECIATE 4.4
Database contains the following total number of profiles and unique species:
3,600 PM profiles;
1,879 organic gas profiles;
249 Other Gases profiles;
2,346 unique species; and
Composite profiles for 58 (47 PM and 11 TOG) source categories.
The SPECIATE version 4.5 is expected to be available to the public in fall of 2016. The
SPECIATE Database is used in conjunction with an inventory of VOC and PM, such as the NEI,
in order to provide the model-ready species required for air quality modeling. The species
needed are dependent on the chemical mechanism and air quality model. The NEI provides
emissions of criteria air pollutants and their precursors. The NEI also contains HAPs, specific
pollutants (such as benzene), and classes of pollutants (such as PM, VOC, and NOx) that may be
determined through source-specific emission measurements, mass balance, source-specific
models, emissions models, or emission factors.
21 ACE Action Plan: htips://www.epa.gov/sites/production/files/2014-06/documents/strap-ace2012.pdf. Last
Accessed: June 2016.
22 ACE Overview:
https://vosemite.epa.gov/sab/sabproduct.nsf/CB83B5741E45F33085257A3300579242/$File/ACE+for+SAB BOSC
ial.pdf. Last Accessed: June 2016.
22
-------
The data from SPECIATE may also be used to estimate HAPs in the inventory, or may be used
as inputs to models such as in EPA's Oil and Gas Tool.23 For example, a speciation profile
containing benzene, toluene, ethyl benzene and xylenes (BTEX) can be used to estimate these
HAPs from VOC for a particular process by computing VOC to BTEX factors from the TOG
speciation profile and applying them to the VOC emissions for the county and process. The data
from SPECIATE also provide an easy way to develop emissions inventories, to quickly analyze
and determine source sectors that are major contributors to nationwide emissions of specific
VOC compounds that are important for ozone formation and toxics exposure, and to provide the
information needed to conduct air quality modeling.24
4.3 SPECIATE Data Processing and Entry Approach
EPA has developed some guidance documents outlining a quality assurance plan and procedures
for collecting and presenting source profile data to assess whether the data should be
incorporated into the SPECIATE Database.25'26'27 EPA is currently working to update the
guidance in a Standard Operating Procedures Document.28 The following sections explain the
general data processing and entry approach for the development of speciation profiles for the
SPECIATE Database.
4.3.1 Data Collection
In general, profiles are defined as the mass fractions of chemical species that make up a source-
specific emission stream. The VOC profiles should include the mass fractions of each of the
species present, including species that cannot be identified. When all organic gas species are
present (e.g. methane), these profiles are referred to as TOG profiles.
Profile data must contain information on the chemical abundance of each species noted above.
These data can be defined as the fraction of mass emissions of VOC/TOG or the mass emission
rate of each species (e.g. pounds per ton (lb/ton); grams per vehicle-miles traveled (g/VMT),
etc.). In addition to the estimate of central tendency for each species (e.g. mean, median), an
estimate of the variability of each species should also be provided (e.g. standard deviation).
Available information on the analytical uncertainty for individual test profiles should be
identified and described separately.
23 https://www.ei3a.gov/air-eniissions-inventones/oil-aiid-gas-101-overview-oil-aiid-gas-upstreani--activities-and-
using-epas. Last Accessed: June 2016.
24 Simon, et. al., "The Development and Uses of EPA's SPECIATE Database", Atmospheric Pollution Research 1
(2010) 196-06 (http://www3.epa.gov/ttnchiel/software/speciate/atmospheric.pdf).
25 SPECIATE Version 4.4 Quality Assurance Project Plan (QAPP), EPA Contract No. EP-D-08-100, WA4-02, July
15, 2013.
26 Protocol for Expansion of the SPECIATE Database, EPA Contract No. 68-D-00-265, WANo. 4-46, May 30,
2005.
27 SPECIATE Version 4.4 Database Development Documentation, EPA Contract No. EP-D-08-100, WA No. 4-14,
February 19, 2014. https://www3.epa.gov/ttn/chief/software/speciate/speciate_version4_4_finalreport.pdf
28 SPECIATE Version 4.4 Database Development Documentation, EPA Contract No. EP-D-08-100, WANo. 4-14,
February 19, 2014. https://www3.epa.gov/ttn/chief/software/speciate/speciate_version4_4_finalreport.pdf. Last
Accessed: June 2016.
23
-------
The specific details of the data processing for each dataset are discussed below in section 4.4 of
this report.
4.3.2 Documentation
The primary reference for the profile is cited as the source of documentation. The "document"
column in the "Reference Table" of the SPECIATE database is used to store this information.
The "notes" column in the Reference Table of the SPECIATE database also contains additional
descriptive information on the profile.
4.3.3 Data Format
The data were formatted into a template developed by EPA and available on the SPECIATE
Database Documentation website.29 The template allows for the data to be easily added to the
SPECIATE database. All the information requested in the template was provided by EPA or
developed by the SPECIATE contractor, including references, test methods, analytical methods,
Chemical Abstract System (CAS) numbers, data quality ratings, normalization basis, etc.
4.3.4 Speciation Data Quality
Recommendations for or against inclusion of profiles into SPECIATE are based on a number of
factors including whether there any available data for a particular process in the database. In
some situations, the perceived overall quality of the profiles is used. There are no simple criteria
that can be set to scrutinize speciation data for inclusion in the SPECIATE database. The
supporting information (metadata) housed within SPECIATE is therefore critically important.
Also, communication with the principal investigator involved in producing the data is essential.
The SPECIATE database provides structure sufficient to thoroughly document profiles and the
underlying analyses. EPA guidance recommends that data housed in the SPECIATE Database
originate from one of the following sources:
Peer-reviewed data appearing in journal articles;
Products of other EPA projects; or
A select group of expert scientists in consultation to the EPA.
In addition, EPA guidance provides a profile rating criteria to assist is determining the overall
quality of the profiles for inclusion into the SPECIATE Database. Each profile will have a
quality rating that is assigned by the profile developer. The quality rating protocol is documented
on SPECIATE Database Documentation website.30 The profile ratings developed for newly
added source profiles are based on the following criteria:
V-rating (profile vintage): The vintage of the profile reflects measurement technology
and methodology. For profiles before year 1980, score = 1; 1980-1990, score = 2; 1991 -
2000, score = 3; 2001-2005, score = 4; and after year 2006, score = 5.
29 http://cfpub.epa.gov/si/speciate/ehpa_speciate_documentation.cfm. Last Accessed: June 2016.
30 http://cfpub.epa.gov/si/speciate/ehpa_speciate_documentation.cfm Last Accessed: June2016.
24
-------
D-rating (number of samples): This category is rated based on the number of samples: #
of samples > 10, score = 4; 5-9 samples, score = 3; 3-4 and composite samples, score = 2;
and 1-2 or unknown # of samples, score = 1.
Quality Score: V-rating x D-rating.
J-rating (expert judgment): Given a "1" (poor) to "5" (excellent) rating. This value is
based on the information underlying each profile including, but not limited to:
o Profile composition;
o Relative ratios of species within the profile;
o Sum of the speciated mass fractions; and
o Supporting documentation.
It should be noted that speciation profiles can be based on an individual sample or an average of
samples. Averaging speciation profiles is generally based on weighting and normalizing the
individual profiles to generate a composite speciation profile. Composite profiles are needed
because tests are often performed on different emission sources that represent the same type of
operation, but perhaps at different times or different locations. When using the profiles in
applications such as modeling, it is more practical to take an average of the different tests, rather
than having to choose a single test.
The Quality Score is basically an objective rating and the J-rating is a subjective rating.
Additional consideration in deciding whether to include the profile in SPECIATE includes:
Appropriate Method - Reviewers experienced in analytical methods and application of
speciation profiles will need to determine if characteristic compounds are present and
properly measured. Sampling and analytical procedures need to be specific to the source
and documented as thoroughly as possible.
Measurement Precision - Low precision is expected for certain species; the data quality
ratings should reflect this issue. In cases where the sampling or analytical methods are
found to be wholly inappropriate for a given species, these data should not be included in
SPECIATE.
Overall Test Program Confidence -Results obtained from the test program should be
consistent with expectations for that source, and if not, the differences should be
sufficiently accounted for.
Source Category-specific Considerations -For certain source categories, such as the
pulp and paper industry, oxygenated compounds contribute significantly to organic gas
emissions, thereby interfering with the proper characterization of the total TOG or VOC
emissions. The solution is to collect fully speciated data using appropriate methods and to
consolidate all organic gases into a total organic gas profile for normalization.
The overall rating and constituent ratings, as well as the expert judgment rating, are available to
the user and auditor for consideration. Users may consider the ratings as well as the reference
and summary information about the profiles housed in the profile tables to determine the
suitability of a profile to the community's needs.
25
-------
4.4 Study or Measurement Datasets
The following sections discuss the studies and datasets reviewed for this project to support or
develop VOC speciation profiles for the SPECIATE Database. Either a quality assurance review
of these datasets were conducted to determine whether the data were sufficiently supported for
the entry into the SPECIATE Database, or the datasets were developed into profiles that could
then be utilized in the SPECIATE Database.
4.4.1 RARE ECD Study
4.4.1.1 Background
As discussed in the previous sections, the instruments deployed for the RARE ECD study
collected data at a total of ten representative well pads in Weld County to characterize emissions
from ECDs at upstream oil and gas production sites. Unfortunately, the emission signals
observable by the remote sensing instruments were fairly weak at most of the well pads. This
could have been a result of the limitation of the measurements or simply the absence of the
compounds (i.e., ECDs were functioning properly). As a result, it was difficult to determine
whether all of the targeted compounds were properly measured by the instruments. In addition,
the field campaign focused on ambient measurements, as opposed to source-specific
measurements. Given that inventories are for specific sources at an oil and gas well pad (i.e.,
condensate tanks, dehydrator vents, pneumatic devices), source-specific speciation data is
needed for use in speciating the source-specific VOC emissions.
4.4.1.2 Results
Until additional data can be collected for these sources, the speciated emissions collected from
this campaign will not be utilized to develop profiles for the SPECIATE Database. Future work
could include:
Developing a method to utilize the data collected from the RARE field campaign;
Determining whether the data collected from the RARE field campaign represented all of
the pollutants released from the sources;
Investigating the speciated emissions profiles of properly maintained and controlled oil
and gas E&P processes;
Investigating the speciated emissions profiles of poorly maintained and controlled E&P
processes.
This work would assist in determining whether the speciated emissions collected from the RARE
field campaign are potentially representative of these types of sources. Further, this work could
be compared to the current flare profile included in the SPECIATE Database that was based on
data collected in 1989 to determine its representativeness.
26
-------
4.4.2 WRAP Phase HI VOC SPECIATION Profiles
4.4.2.1 Background
The Western Regional Air Partnership (WRAP) and the Western Energy Alliance (WEA),
formerly the Independent Petroleum Association of Mountain States (IPAMS), sponsored the
development of a Phase HI regional oil and gas emission inventory for the Inter-Mountain
West.31 This effort focused on creating a comprehensive criteria pollutant emissions inventory
for activities associated with oil and gas field operations in the basins throughout the study
region for a baseline year (2006 for most basins) as well as future projection years. The
inventory includes all point and area sources related to the oil and gas industry exploration and
production operations at well sites and midstream (primarily compressors station and gas plants)
sources, known through states' inventory efforts or disclosed by operators for the first time in the
project data collection effort. Figure 6 shows oil and gas basins covered by the WRAP Phase III
work with the state and county boundaries overlaid.
Big Horn Basin
Wy o
Southwest Wyoming Basin
Figure 6. Overlay of the WRAP basins with state and county boundaries.
4.4.2.2 Methodology
To convert the raw emissions data into an emissions inventory that could support air quality
activities, this effort collected gas composition analyses and developed oil and gas speciation
profiles for different well types, processes, and basins in the Rocky Mountain States.32 The gas
31 WRAP Website: http://www.wrapair2.org/emissions.aspx Last Accessed: June2016.
32 WRAP Phase III oil and gas speciation profiles Memorandum, Ramboll ENVIRON, Revised August 27, 2015
27
-------
composition analyses were collected through operator surveys as part of the WRAP Phase HI
project. The data are based on oil and gas companies taking Gas Chromatography/Mass
Spectrometry (GC/MS) analyses of their produced gas or in some cases running models such as
E&P TANK using input measured compositions (again derived from GC/MS tests of
hydrocarbon liquids). The gas composition data were gathered to develop 2006 base year oil and
gas inventories, but may not necessarily reflect samples collected in 2006. While the samples
could be from a different year, an assumption was made that the gas composition was not
expected to vary much in time for a given basin. Survey respondents were instructed to provide
"representative" gas compositions. However, no provisions were made to ensure that the
collected compositions were statistically valid. Nevertheless, these data represent actual gas
compositions collected by multiple companies in each of the WRAP basins and therefore
represent an improvement over other potentially available data sources.
Each survey respondent's gas compositions were averaged to obtain a representative operator-
specific gas composition. Then, the composite weighted averaged profiles were developed by
taking a weighted average of all operator specific compositions using the fraction of gas
production ownership for each operator as the weighting factor. Table 4 shows the number of
individual profiles across all survey respondents that were used to create the weighted average
composite profile. Note that in some cases a straight average was calculated instead of a
production weighted average (e.g., DJFLA) because information was not available to estimate a
weighted average profile.
Table 4. Num
)er of individual profiles averaged to develop composite profile.
PNUMBER
Name
Number
of Profiles
SSJCB
South San Juan Basin Produced Gas Composition from Coal-Bed Methane (CBM)
Wells
4
SSJCO
South San Juan Basin Produced Gas Composition from Non-CBM Gas Wells
15
WRBCO
Wind River Basin Produced Gas Composition from Non-CBM Gas Wells
7
PRBCB
Powder River Basin Produced Gas Composition from CBM Wells
8
PRBCO
Powder River Basin Produced Gas Composition from Non-CBM Wells
11
DJFLA
Denver-Julesburg Basin Flashing Gas Composition for Condensate Tanks
16
DJVNT
Denver-Julesburg Basin Produced Gas Composition from Non-CBM Gas Wells
13
UNT01
Uinta Basin Produced Gas Composition from CBM Wells
3
UNT02
Uinta Basin Produced Gas Composition from Non-CBM Wells
28
UNT03
Uinta Basin Flash Gas Composition from Oil Tanks
1
UNT04
Uinta Basin Flash Gas Composition from Condensate Tanks
5
PNC01
Piceance Basin Produced Gas Composition from Non-CBM Gas Wells
20
PNC02
Piceance Basin Produced Gas Composition from Oil Wells
1
PNC03
Piceance Basin Flash Gas Composition for Condensate Tank
5
SWFLA
SW Wyoming Basin Flash Gas Composition for Condensate Tanks
6
SWVNT
SW Wyoming Basin Produced Gas Composition from Non-CBM Wells
23
PRM01
Permian Basin Produced Gas Composition for Non-CBM Wells
4
The composite weighted average profiles were normalized for organic gaseous species,
excluding inorganic gases (e.g., carbon dioxide (CO2) and hydrogen sulfide (H2S)), to develop
the profiles for the SPECIATE Database and Speciation Tool. SPECIATE species identification
numbers were assigned to each profile based on name and engineering judgment. When there
28
-------
were unknown groups (e.g., CIO compounds) in the profile, the species identification number for
the lowest carbon number species in the group (i.e. CIO compounds) was applied to the profile.
Finally, the WRAP Phase in basin-specific oil and gas speciation profiles were converted into
the SPECIATE Database format.
4.4.2.3 Results
A total of 17 profiles related to oil and gas sector gaseous emissions were developed through this
project (see Table 4). The specific profiles are provided in the accompanying EXCEL workbook,
and referenced as "WRAP" in the worksheets. Appendix B also presents the final WRAP Phase
III oil and gas speciation profiles. The SPECIATE profiles provided in Appendix B differ by
basin and within basin by the type of emission stream.
The WRAP study suggested applications of the different types of profiles to associated source
categories:
Produced Gas Composition from non-CBM Wells: Applied to vented source emissions
from non-CBM oil and gas wells in a basin for source categories such as completions,
blowdowns, pneumatic controllers, pneumatic pumps, and fugitive leaks. This type of
profile should not be applied to CBM well or tank emissions.
Produced Gas Composition from non-CBM Gas Wells: Applied to vented source
emissions from non-CBM gas wells in a basin for source categories such as completions,
blowdowns, pneumatic controllers, pneumatic pumps, and fugitive leaks. This type of
profile should not be applied to emissions from CBM wells, oil wells, or tanks.
Produced Gas Composition from non-CBM Oil Wells: Applied to vented source
emissions from non-CBM oil wells in a basin for source categories such as completions,
blowdowns, pneumatic controllers, pneumatic pumps, casing head gas venting, and
fugitive leaks. This type of profile should not be applied to emissions from CBM wells,
gas wells, or tanks.
Produced Gas Composition from CBM Wells: Applied to vented source emissions from
CBM gas wells in a basin for source categories such as completions, blowdowns,
pneumatic controllers, and fugitive leaks. This type of profile should not be applied to
emissions from non-CBM wells or tanks.
Flashing Gas Composition from Condensate Tanks: Applied to emissions in a basin from
condensate tanks. This profile should not be applied to vented source emissions such as
completions, blowdowns, pneumatic controllers, pneumatic pumps, and fugitive leaks.
Flashing Gas Composition from Oil Tanks: Applied to emissions in a basin from oil
tanks. This profile should not be applied to vented source emissions such as completions,
blowdowns, pneumatic controllers, pneumatic pumps, and fugitive leaks.
29
-------
4.4.3 Uintah and Ouray Indian Reservations Tribal Minor Source Registrations
4.4.3.1 Background
In 2011, EPA promulgated the Indian Country Minor New Source Review Rule or Tribal Minor
Source (TMS) New Source Review (NSR)Rule [40 CFR49.151].33'34 The Rule is a
preconstruction permitting program that serves two important purposes. First, it ensures that air
quality in reservation areas of Indian country is not significantly degraded from the addition of
new and modified sources of air pollution, such as factories, industrial boilers and power plants.
In areas with unhealthy air, the Rule assures that new emissions do not slow progress toward
cleaner air, or that new emissions do not significantly worsen air quality. Second, the Rule
assures people that any new or modified industrial source in their neighborhoods will be as clean
as possible, and that advances in pollution control occur concurrently with industrial expansion.
Tribal minor sources are defined in attainment areas as those sources with the potential to emit
less than major source preconstruction permitting thresholds, but more than:
10 tons per year of carbon monoxide (CO), NOx, SO2, or PM, or
5 tons per year of VOCs, or
5 tons per year of particulate matter less than 10 microns (PM10), or
3 tons per year of particulate matter less than 2.5 microns (PM2.5), or
0.1 tons per year of lead, or
1 ton per year of fluorides, or
2 tons per year of H2S.
The TMS N SR. rule, when promulgated, required registration of existing and new minor sources
in reservation areas of Indian country until September 2, 2014, at which time new sources were
required to obtain a minor source permit prior to construction [40 CFR 49.160],35 EPA made a
number of revisions to this rulemaking to extend the permitting deadline for new and modified
minor sources in the oil and natural gas sector operating or proposing to operate in reservation
areas of Indian country and other areas of Indian country for which tribal jurisdiction has been
demonstrated.36 The most recent February 24, 2016 deadline has been extended to October 3,
2016 for obtaining a permit. The revisions were necessary to avoid the potentially unnecessary
burden of sources in the oil and natural gas sector needing to obtain source-specific permits
while EPA develops a streamlined permitting solution for the source category, as contemplated
in the Rule for certain source categories expected to be common and widespread throughout
Indian country. The revisions also provided a level of certainty to the regulated industry, tribes
and other parties pending final action on the streamlined permitting solution. Each registration
requires the following information, as applicable:
Identifying information;
33 https://www3.epa.gov/air/tribal/tribalnsr.html. Last Accessed: June2016.
34 http://www.ecfr.gov/cgi-bin/text-idx?SID=b3c384e3b0e2672b9d 150flfflbf4483&mc =
true&node=se40.1.49_1151&rgn=div8. Last Accessed: June 2016.
35 http://www.ecfr.gov/cgi-bin/text-idx?SID=b3c384e3b0e2672b9d 150flfflbf4483&mc=true&node
=se40.1.49_1160&rgn=div8. Last Accessed: June 2016.
36 76 FR 38788, July 1, 2011, as amended at 79 FR 31045, May 30, 2014; 79 FR 34239, June 16, 2014.
30
-------
A description of source's processes and products;
A list of all emissions units;
Allowable and estimated actual annual emissions of each regulated NSR pollutant in tons
per year (tpy) for each emissions unit listed (estimates must be based on actual test data
or acceptable procedures, including source-specific emissions tests, mass balance
calculations, published emission factors, or other engineering calculations);
Information on fuels, fuel use, raw materials, production rates and operating schedules;
Identification and description of any existing air pollution control equipment and
compliance monitoring devices or activities; and
Any existing limitations on source operation affecting emissions or any work practice
standards for all NSR-regulated pollutants at the source.
The registrations are publically available for review. The EPA proposed a Federal
Implementation Plan (F1P) under the Rule for oil and natural gas sector in September 2015 that,
if finalized, will take the place of individual source permitting, unless a source opts for a site-
specific permit. The proposed FIP, if finalized as proposed, would require new oil and natural
gas production sources to comply with a suite of cost-effective regulations, as individually
applicable, that reduce harmful air pollution from the oil and natural gas industry, in lieu of
obtaining a site-specific permit prior to construction or modification.
4.4.3.2 Methodology
EPA Region 8 surveyed the IMS NSR registration data for oil and natural gas operations located
on the Uintah and Ouray Indian Reservations. The registrations surveyed consisted of existing oil
and natural gas sources with registration submissions between August 2011 and March 2015.
Any registration data submitted for new sources or sources created or modified after this period
were not included in this survey. It should also be noted that while the survey reviewed 2011 to
2015 submissions, the data may not always reflect emissions from these years. IMS NSR
registration data do not need to be specific to a particular year, only representative of the
expected emissions from the specific sources.
Approximately 5,200 registrations were surveyed for VOC emissions associated with oil and gas
operations located on the Uintah and Ouray Indian Reservations. These registrations represented
23 different operators or companies, and consisted of crude petroleum and natural gas extraction
facilities (North American Industry Classification System (NAICS) 211111), natural gas liquid
extraction facilities (NAICS 211112), and support activities for oil and gas operations facilities
(NAICS 213112). About 63 percent of registrations consisted of crude petroleum and natural gas
extraction, about 37 percent consisted of natural gas liquid extraction, and less than 1 percent
consisted of support activities for oil and gas operations.
The registrations provided extended composition analyses of the untreated natural gas stream,
the tank emission stream, and the glycol dehydrator regenerator emission stream associated with
each facility. The composition of each untreated natural gas stream was determined by lab
analysis. The composition of each tank emission stream was determined using an emission
model programs, such as the American Petroleum Institute's (API's) E&P Tanks or an analysis
of the Gas Oil Ratio (GOR). These emission models require operating inputs that include
31
-------
separator pressure and temperature and extended hydrocarbon analyses of pressurized liquid
samples. The composition of the glycol dehydrator regenerator emission stream was determined
using the Gas Research Institute (GRI-GlyCalc) emission model.37 This emission model requires
operating inputs that include the absorber tower pressure and temperature, the glycol circulation
rate, and extended gas analysis of the input "raw" gas stream. The outputs of all these models
provide speciated emission streams. No provisions were made to assess the quality or timing of
the necessary pressurized liquid and raw gas stream lab analyses necessary to run these models.
Further, no analyses were performed to assess the statistical "representativeness," or ensure that
the collected compositions were statistically valid. The estimated speciated emission streams
were also based on emission calculations that did not account for controls being implemented.
The majority of minor registered sources are uncontrolled at this time.
The registration forms do not provide details on the methods used in collecting and analyzing the
samples for the oil and gas components. Nevertheless, these data represent actual gas
compositions collected by multiple companies in the Uinta Basin, and are routinely collected
across the oil and gas fields. Therefore, these data represent an improvement over other
potentially available data sources and the current information available to the community.
An average composition for the untreated "raw" natural gas emission stream, the oil tank
emission stream, the condensate tank emission stream, and the glycol dehydrator regenerator
(also referred to as still vent) emission stream was determined for each operator using the data
provided in the registrations. The average operator-specific profile calculated for each emission
stream was then normalized for each operator. Where emission stream outputs were in terms of
percentage by mole (mol%), the data were converted to percentage by weight (wt%) and
normalized for TOG profiles (e.g., removing H2S, CO2 or Nitrogen). This resulted in 37 average
operator-specific composition profiles, which consisted of 15 untreated "raw" natural gas
emission stream profiles, three oil tank emission stream profiles, 11 condensate tank emission
stream profiles, and eight glycol dehydrator regenerator emission stream profiles. Table 5
presents the number of different samples used to calculate the operator-specific composition
profiles for each emission stream.
The average operator-specific composition profiles developed for each operator and emission
stream were then renamed into the classification established in the SPECIATE Database. For
instance, "isomers of octane" from SPECIATE cover "n-Octane (C8)" output in the tank
emission model. The re-classification was applied to methane, n-hexane, isomers of hexane,
isomers of heptane, ethane, propane, isobutene, n-butane, isopentane, and n-pentane.
With assistance from Abt Associates, four "composite" speciated VOC profiles were developed
using the 37 operator-specific composition profiles. A composite VOC profile was developed for
each process emission source, including a composite profile for the untreated "raw" natural gas
emission stream, the oil tank emission stream, the condensate tank emission stream, and the
glycol dehydrator regenerator emission stream. Before creating the composite VOC profiles, the
data were converted from volume percent (vol%) or mol% to wt%. The composite profiles were
developed by combining the operator-specific profiles for a particular emission source and
calculating the weighted average using oil/condensate production for tank emissions and natural
37 Gas Technology Institute, Des Plaines, USA; ILhttp://sales.gastechnology.org/000102.html).
32
-------
gas production for untreated "raw" natural gas and glycol dehydrators. The production data are
based on information collected during 2014 for the top 20 producers in the Uintah and Ouray
reservations. Appendix C outlines the data used for weighting the profiles.
Table 5. Number of Different Samples for Each Averaged
Profile
Emission Stream
Number of Unique
Profiles
95336
Untreated Natural Gas
1
95337
Untreated Natural Gas
9
95338
Untreated Natural Gas
1
95339
Untreated Natural Gas
1
95340
Untreated Natural Gas
2
95341
Untreated Natural Gas
2
95342
Untreated Natural Gas
2
95343
Untreated Natural Gas
2
95344
Untreated Natural Gas
4
95345
Untreated Natural Gas
3
95346
Untreated Natural Gas
2
95347
Untreated Natural Gas
3
95348
Untreated Natural Gas
1
95349
Untreated Natural Gas
1
95350
Untreated Natural Gas
25
95351
Oil Tank Vent Gas
1
95352
Oil Tank Vent Gas
23
95362
Oil Tank Vent Gas
42
95353
Condensate Tank Vent Gas
2
95354
Condensate Tank Vent Gas
6
95355
Condensate Tank Vent Gas
7
95356
Condensate Tank Vent Gas
6
95357
Condensate Tank Vent Gas
2
95358
Condensate Tank Vent Gas
29
95359
Condensate Tank Vent Gas
59
95360
Condensate Tank Vent Gas
4
95361
Condensate Tank Vent Gas
4
95363
Condensate Tank Vent Gas
2
95364
Condensate Tank Vent Gas
6
95409
Glycol Dehydrator
7
95410
Glycol Dehydrator
2
95411
Glycol Dehydrator
3
95412
Glycol Dehydrator
3
95413
Glycol Dehydrator
2
95414
Glycol Dehydrator
1
Operator-Specific Profile.
33
-------
95415
Glycol Dehydrator
2
95416
Glycol Dehydrator
27
4.4.3.3 Results
A total of 37 operator-specific speciation profiles and four composite profiles representing
weighted-average profiles for each of the emission streams/sources were developed from the
Uintah Basin and Ouray Tribal Minor Source Registrations. Table 6 outlines the number of
profiles for each oil and natural gas production emission source based on the data from the IMS
NSR registrations. The specific profiles are provided in the accompanying EXCEL workbook,
and referenced as "TMSR" in the worksheets. Appendix B also presents the final speciation
profiles.
Table 6. Number of profiles for each oil and natural gas production emission source.
Kmissiun Source S|K'ei;ili<>n
NuihIkt of
Profiles
Untreated "Raw" Natural Gas
15
Oil Tank
3
Condens ate T ank
11
Glycol Dehydrator Regenerator
8
Composite of Each Emission Stream
4
Total
41
The analysis of the registration data allowed for the development ofVOC speciation profiles for
oil and natural gas sources located on the Uintah and Ouray Indian Reservations. These
speciation profiles have the following suggested applications:
Untreated "Raw" Natural Gas Emission Stream: This profile should be applied to
emission sources such as fugitive leaks, pneumatic controllers, and pneumatic pumps.
Oil Tank Emission Stream: This profile should be applied to emissions sources from oil
tanks (those with API Gravity of the sales oil <40 degrees) and include flash,
working/breathing/standing emissions.
Condensate Tank Emission Stream: This profile should be applied to emissions sources
from condensate tanks (those with API Gravity of the sales oil >40 degrees) and include
flash, working/breathing/standing emissions.
Glycol Dehydrators Regenerator (Still Vent) Emission Stream: This profile should be
applied to emissions sources from glycol dehydrator regenerator/still vents.
Given the similarity in approaches used for these and the WRAP HI profiles, a comparison was
done in species concentrations across similar emission sources for the individual and composite
TMS NSR profiles and the Uintah Basin WRAP in profiles. The comparison is provided in the
accompanying EXCEL workbook, and referenced as "Uintah Profile Comparison" in the
worksheets. The comparison, provided in Appendix B, shows varying agreement by source type
and species, with the raw gas profiles being the most consistent.
34
-------
4.4.4 Denver-Julesburg Basin Direct Measurement Study
4.4.4.1 Background
In July 2011, EPA, with contract support from Arcadis, conducted a direct measurement study of
production pad emissions in Weld County, Colorado.38'39 This effort was coordinated with Sage
Environmental Consulting (Sage) and several industry operators. The study focused on the
determination of instantaneous VOC and methane (CH4) emissions from production pads, with
emphasis on oil and condensate tank emissions, using non-invasive measurement techniques,
such as infrared video and real-time leak measurements coupled with subsequent laboratory
analysis of acquired canisters. The general goals for the study were to improve the understanding
of component-level emissions and speciation profiles from production pads using non-invasive
measurement approaches. Another goal of the study was to improve the understanding of the
performance of high volume sampling equipment for emissions that are VOC rich (defined here
as combustible vapor less than -95% CH4).40'41
A total of 23 sites within Weld County, Colorado were selected for sampling. For each well pad,
leak inspections were performed to identify emission points, and the identified emission points
were sampled to determine the emissions rate and to estimate the mass emission rates of
individual organic compounds. From the largest emission point on each well pad, at least one
sample was acquired at the exit of the High Volume Sampler (HVS) using a leak-free, sub-
atmospheric 6-liter stainless steel canister with a valve and passivated interior. The canister-
derived concentration values were used with the measured HVS flow rates to calculate emission
rates for individual and groups of compounds, including the US EPA Photochemical Assessment
Monitoring (PAMS) Target (VOC, as well as percent level CH4, ethane, ethene, propene, and
propane. The concentrations of total and speciated non-methane volatile organic compounds
were determined using Gas Chromatography with Flame Ionization Detection (GC/FID) as
described in EPA/600-R-98/16142 coupled with American Society for Testing and Materials
(ASTM) 1946/D1945 analysis43 of methane, ethane, and propane. The calculation of the
emission rate of total combustibles and total VOC emissions was accomplished by summing the
38 Understanding Direct Emission Measurement Approaches for Upstream Oil and Gas Production Operations, M.
Modrak, M. Shahrooz, J.Ibanez, C. Lehmann, B. Harris, D. Ranum, E. Thoma, B. Squier, Air & Waste Management
Association, 105th Annual Conference and Exhibition, June 19-22, 2012, Texas.
39 Assessment of VOC andHAP Emissions from Oil and Natural Gas Well Pads Using Mobile Remote and Onsite
Direct Measurements, Halley L. Brantley, E. D. Thoma, A.P. Eisele (2015):, Journal of the Air & Waste
Management Association,DOI: 10.1080/10962247.2015.10568.
40 EPA Report, Oil and Gas Production Pad Air Emission Study, Weld County, Colorado, prepared by Arcadis
underEP-C-09-027, (in preparation).
41 U.S. EPA. Greenhouse Gas Emissions Reporting From the Petroleum andNatural Gas Industry, Background
Technical Supporting Document, supporting 40 CFR Part 98.230, 77 FR 11039Subpart W - Petroleum andNatural
Gas Systems, at web site: http://www.epa.gov/climatechange/eniissions/downloads 10/Subpart-W_TSD.pdf. Last
Accessed: June 2016.
42 Technical Assistance Document for Sampling and Analysis of Ozone Precursors, EPA National Exposure
Research Laboratory, Research Triangle Park. NC, September 1998.
https://www3.epa.g-ov/ttnaniti 1/files/ambient/pams/newtad.pdf (Accessed: June 2016).
43 American Society for Testing and Materials (ASTM). 2010. ASTM D1945-03. Standard Test Method for
Analysis of Natural Gas by Gas Chromatography. West Conshohocken,PA: ASTM International, www.astm.org.
doi:10.1520/D1945-03R10.
35
-------
concentrations of individual measured species to achieve a total measured pollutant vol%, which
was then multiplied by the total gas flow rate (converted to standard conditions). The calculation
of speciated mass emissions was accomplished by first converting the VOC concentration results
from parts per million by volume (ppmv) to units of micrograms per cubic meter (|j,g/m3) and
converting the gas flow rate to standard gas flow.
From the 23 sites, a total of 106 emission points were measured by the instruments. Samples
were acquired from condensate tank thief hatch leaks or other emission points prior to the control
device. The average production pad consisted of five wells, 258 valves, 2,583 connectors, three
condensate tanks, one produced water tank, four thief hatches, five pressure relief devices, three
separators, and all sites contained one flare or combustor. One production pad contained a
dehydration unit, and four each contained one vapor recovery unit. All sites were fitted with one
ECD as per current State of Colorado requirements. Of the 23 sites surveyed, 19 processed field
gas by a single stage three phase separator and four utilized a two stage separation process to
further recover natural gas by reducing the net pressure by approximately 25 percent of the liquid
sent to the condensate tanks via a buffer tank.
4.4.4.2 Methodology
The direct measurement study conducted in the Denver-Mesburg Basin collected data from
produced water tanks, separators, well heads, dehydrators, and condensate tanks from oil and
natural gas production operations.37'38'44 However, only the data from the condensate tanks were
used in this work because of the amount and quality of the available samples. There was not
enough confidence in the data from the other sources because only a limited number of samples
were collected. For instance, data referenced as separator emissions may be a pneumatic
controller leak that happened to be near the separator. Additionally, the produced water tank only
had one sample. Therefore, this work focused on the data collected from the condensate tanks.
The raw speciation data collected from the canister samples were used to develop VOC
speciation profiles. The data were available in concentrations (ppb). Because the sum of adding
methane, ethane, and all other measured species was so large relative to the unknown species, it
was reasonable to use the sum of measured or known species as the normalization basis. Based
on ideal gas law, for each canister sample (i.e., fixed volume), volume fraction (ppb) is
equivalent to mole fraction for each species. By multiplying molecular weight to concentration
(ppb), each species was converted into mass. The sum of these speciated masses was the
normalization basis for each profile. Each speciated mass was divided by the sum of speciated
masses to calculate the weight fraction, then converted to wt% by multiplying 100.
The composite profile was developed by taking the mean of the individual condensate tank
profiles. By comparing the composite profile compositions using "mean" and "median" of each
44 U.S. EPA. "Technical Assistance Document for Sampling and Analysis of Ozone Precursors. (1998), At web site
https://www3.epa.gov/ttnamtil/files/ambient/pams/newtad.pdf (accessed June 22, 2016).
American Society for Testing andMaterials (ASTM). ASTM D1945-03, "Standard Test Method for Analysis of
Natural Gas by Gas Chromatography" West Conshohocken,PA: ASTM International, www.astmorg. (2010)
doi:10.1520/D1945-03R10.
36
-------
species, it was found that both methods resulted in very similar composite profile compositions.
As a result, this work elected to use the "mean" to composite the condensate tank profiles.
4.4.4.3
Results
A total of 27 speciation profiles and one composite profile for the condensate tank emissions
were developed from the measurements collected in the DJ Basin. Table 7 presents the number
of oil and natural gas production profiles for each emission source. The specific profiles are
provided in the accompanying EXCEL workbook, and referenced as "DJBasinDMS" in the
worksheets. Appendix B also presents the final speciation profiles.
Tabic 7 \'Limber of profiles for each oil and natural gas production emission source.
NuihIkt iปl' Profili's
Condens ate T ank
27
Composite
1
Total
28
The analysis of the measurement data allowed for the development of VOC speciation profiles
for oil and natural gas sources located in the DJ Basin. These speciation profiles have the
following suggested applications:
Condensate Tank Emission Stream: This profile should be applied to emissions sources
from condensate tanks. Samples were acquired from condensate tank thief hatch leaks or
other emission point prior to the control device. All sites were fitted with one ECD as per
current State of Colorado requirements.
4.4.5 East Texas Oil Field Speciation Data
4.4.5.1 Background
Measurements conducted by the National Oceanic and Atmospheric Administration (NOAA)
and other research organizations during the 2000 Texas Air Quality Study (TexAQS 2000)
suggested that the levels of VOC found in ambient air could not all be accounted for based on
reported emissions estimates. As a result, the Texas Commission on Environmental Quality
(TCEQ) began an intensive effort to identify, quantify, and reduce VOC emissions that had been
underestimated in the past. In this effort, TCEQ identified oil and condensate storage tanks,
through the use of remote sensing measurements, as a source category for potentially
underestimated emissions (TCEQ, 2005). Oil and condensate storage tank emissions at wellhead
and gathering sites are composed of working losses, breathing losses, and flashing losses.
Working losses are vapors that are displaced from a tank during the filling cycle and breathing
losses are vapors that are produced in response to diurnal temperature changes. Flashing losses
are vapors that are released when a liquid with entrained gases experiences a pressure drop, as
during the transfer of liquid hydrocarbons from a wellhead or separator to a storage tank that is
vented to the atmosphere. This effort also resulted in a study that investigated and developed
speciated VOC profiles and emissions factors for condensate storage tanks in Texas. The study is
summarized below.
37
-------
To support the TCEQ efforts, the Texas Environmental Research Consortium (TERC) study was
conducted in 2005 to evaluate ozone control strategies for Dallas-Fort Worth (DFW), Houston-
Galveston-Brazoria (HGB), and Beaumont-Port Arthur (BPA) by measuring speciated VOCs
and developing average emission factors (in units of pounds of VOC per barrel of oil or
condensate produced (pound per barrels of oil (lb/bbl)) from direct measurements of vent gas
flow rates and chemical composition.45 The measurements were made by directly monitoring the
flow rates of gases escaping from storage tank vents and sampling the vent gases for chemical
composition. Producers of oil and condensate from seven companies allowed the emission
measurements at one or more wellhead or gathering sites. The storage tank battery sites generally
consisted of one or more wellheads, one or more high pressure separators, and two or more
storage tanks containing either water or liquid hydrocarbon (oil or condensate). The approximate
age of the inspected tank batteries ranged from two to more than 50 years. The conditions of the
storage tank batteries were found to vary quite a bit, with some older tanks being of bolted
construction and the newer tanks being of welded construction. The welded tank batteries
generally had piping for vent gas consolidation to a common vent. The storage tank capacities
ranged from 300 to 500 barrels except for at one gathering station, which had tank capacities
ranging from 5,000 to 10,000 barrels. Thirty-three tank batteries met the criteria for sampling
vent gas emissions. Of the 33 tank batteries that were sampled, 27 transferred its liquid product
by tanker truck, five by pipeline, and one by barge. The 2005 TERC study reported
measurements of speciated VOC emissions made at 11 oil and 22 condensate tank battery sites in
the BP A, DFW, and HGB areas during May-July, 2006.
4.4.5.2 Methodology
Prior to the commencement of this RARE project, EPA had collected the speciated VOC data
from the 2005 Texas study. A total of 33 profiles were obtained from the TCEQ Texas study,
including 11 oil tank battery vent gas profiles and 22 condensate tank battery vent gas profiles.46
However, composite profiles of the oil tank and condensate tank profiles were not developed for
the SPECIATE Database. As a result, the RARE project prepared two composite speciation
profiles for entry into SPECIATE.
The Texas study had already calculated a mean speciation profile for the oil tank battery vent gas
and condensate tank battery vent gas. However, the mean speciation profiles provided by the
Texas study included the wt% of some measured gases that are not needed for the SPECIATE
Database, including nitrogen and carbon dioxide. As a result, this work removed the nitrogen and
carbon dioxide contributions from each mean profile and re-normalized the profiles by the sum
of the remaining gases.
4.4.5.3 Results
A total of two composite profiles for tank batteries were developed from the measurements
collected in the Texas study. Table 8 presents the number of oil and natural gas production
45 2005 TERC Study Report: http://files.harc.edu/Projects/AirQuality/Projects/H051C/H051CFinalReport.pdf. Last
Accessed: June 2016.
46 2005 TERC Study Report: http://files.harc.edu/Projects/AirQuality/Projects/H051C/H051CFinalReport.pdf: See
Table 3-4 and Table 3-5 on pages 3-6 to 3-10. Last Accessed: June 2016.
38
-------
profiles for each emission source. The specific profiles are provided in the accompanying
EXCEL workbook, and referenced as "Texas" in the worksheets. Appendix B also presents the
final speciation profiles.
The analysis of the measurement data allowed for the development of VOC speciation profiles
for oil and natural gas sources located in East Texas. These speciation profiles have the
following suggested applications:
Oil Tank Battery Vent Gas
Condensate Tank Battery Vent Gas
Table 8. Number of new composite profiles for each oil and natural gas production emission
source.
Numk'r of
Profiles
Oil l ank Battery Vent Oas
1
Condensate Tank Battery Vent Gas
1
Total
2
4.4.6 San Joaquin, California Oil and Gas Speciation Data
4.4.6.1 Background
California is an important region for oil and natural gas production in the United States. These
sources are prominent in the California Air Resources Board (CARB) emission inventory of
reactive organic gases (ROG) in the San Joaquin Valley.47 Petroleum operations include
extraction, storage, transport, and processing; all of which can have varying degrees of fugitive
emissions of methane and other gas-phase organic carbon, including VOCs.
In accordance with the Request for Proposal issued by CARB, a study was completed in 1991 to
characterize fugitive emissions from oil fields in California.48 The Ventura, Elk Hills, and
Wilmington fields in California were targeted for this study. At the time of the study, the major
oil producers in this area included Shell, Chevron, Bechtel, Texaco, Union Oil, and THUMS
Long Beach Company. The study investigated three source categories from California oil
production facilities, including oil production fugitive emissions, utility engine exhaust, and farm
and heavy-duty engine exhaust. Only the oil production fugitive emissions were reviewed for
this project. The components of the oil production fugitive emissions sampled in this study
included wellheads, pipelines, processing, and storage tanks. Samples from two secondary sumps
were also collected from a flux chamber in SUMMA electro-polished, evacuated stainless steel
canisters. Storage tank headspace samples were collect in evacuated steel canisters. Samples
from other components were obtained by isolating the selected components) with a Teflonฎ
47 Emissions of organic carbon and methane from petroleum and dairy operations in California's San Joaquin
Valley, Atmos. Chem Phys., 14, 4955-4978, 2014, www.atmos-chem-phys.net/14/4955/2014/, doi:10.5194/acp-14-
4955-2014.
48 Final Report on Development of Species Profiles for Selected Organic Emission Sources, Volume I: Oil Field
Fugitive Emissions, Prepared by Albert C. Censullo, California Polytechnic State University, Contract No. A832-
059, Prepared for California Air Resource Board, April 30, 1991.
https://inis.iaea.org/search/search.aspx?orig_q=RN:23071542. Last Accessed: June 2016.
39
-------
shroud, and collecting the shroud effluent in evacuated steel canisters. Additional samples from
several sources were taken by direct connection of the evacuated canisters to pipe fittings in the
distribution lines using Teflon tubing. Analysis for desired hydrocarbon constituents were
performed using a variety of validated chromatographic methods. This study collected a total of
38 samples from the oil production facilities.
In 2003, the U.S. Geological Survey (USGS) completed an assessment of the oil and gas
resource potential of the San Joaquin Basin Province of California. 49 Part of this assessment
included the identification and characterization of natural gas types in the San Joaquin Basin. 50
To accomplish this goal, 66 gas samples were analyzed for CI to C7 hydrocarbons, CO2, CO,
Nitrogen, Oxygen, Argon, Helium, Hydrogen, and H2S using a Wasson gas analyzer, and a
customized Hewlett-Packard gas chromatograph. Stable carbon isotopes of methane, ethane,
propane, butane and carbon dioxide were measured using a Hewlett-Packard gas chromatograph
interfaced with a Micromass Optima continuous-flow isotope ratio mass spectrometer (IRMS).
The results were also combined with analyses of 15 gas samples from previous studies. For the
purpose of this resource assessment, each gas type was assigned to the most likely petroleum
system. Three general gas types were identified on the basis of bulk and stable carbon isotopic
composition, including thermogenic dry, thermogenic wet and biogenic. About 75 percent of the
gas samples in this study were taken from oil fields, with the rest taken from gas fields, generally
reflecting the dominance of oil and associated gas production over non-associated gas production
in the San Joaquin Basin Province.
4.4.6.2 Methodology
Prior to the commencement of this RARE project, EPA had collected the speciated VOC data
from the 1991 and 2003 California studies discussed above. A total of 39 profiles were obtained
from the 1991 study51 and 77 profiles were obtained from the 2003 study.52 Table 9 outlines the
numbers of profiles for each emission source from each study. However, composite profiles for
each source type were not developed for the SPECIATE Database. As a result, the RARE project
prepared the composite speciation profiles for entry into SPECIATE.
The data from the 1991 study were already presented in wt%. As a result, no manipulations were
applied to the data. The composite profiles using the 1991 study were based on the mean of the
original profiles. The mean was applied rather than the median for two reasons, including:
49 Petroleum Systems and Geologic Assessment of Oil and Gas in the San Joaquin Basin Province, California,
USGS, Energy Resource Program, edited by Allegra Hosford Scheirer, 2007. http://pubs.usgs.gOv/pp/ppl713/.Last
Accessed: June 2016.
50 Chapter 10, Petroleum Systems of the San Joaquin Basin Province - Geochemical Characteristics of Gas Types,
Petroleum Systems and Geologic Assessment of Oil and Gas in the San Joaquin Basin Province, California, USGS,
Lillis et al., 2007. http://pubs.usgs.gov/pp/ppl713/10/ppl713_chl0.pdf. Last Accessed: June2016.
51 Final Report on Development of Species Profiles for Selected Organic Emission Sources, Volume I: Oil Field
Fugitive Emissions, Prepared by Albert C. Censullo, California Polytechnic State University, Contract No. A832-
059, Prepared for California Air Resource Board, April 30, 1991. See Table 7 to Table 13 on pages 37 to 43.
52 Chapter 10, Petroleum Systems of the San Joaquin Basin Province - Geochemical Characteristics of Gas Types,
Petroleum Systems and Geologic Assessment of Oil and Gas in the San Joaquin Basin Province, California, USGS,
Lillis et al., 2007. See Table 10.2A and Table 10.2B on pages 24 to 29.
40
-------
1. Some of the composites only had three to five individual profiles available. Therefore, no
meaningful median profiles could be calculated; and
2. For those composites that were based on more than 20 individual profiles, it was found
that the composites based on mean and median were very similar.
The data from the 2003 study were presented in mol%. As a result, the data were converted to
wt% using the molecular weight of each species, and then normalized by the sum of the species.
Multiple species were also renamed into the classification established in the SPECIATE
Database. For instance, "CI" from the data was renamed to methane. The re-classification was
applied to ethane, propane; isobutene; n-butane; isopentane (2-methyl butane); n-pentane;
neopentane (2,2-dimethyl propane); n-hexane; n-heptane.
Table 9. Number of profiles collected from each California Study.
NuihIkt of Profiles
1991 Study
Oil Wells
9
Oil Tanks
17
Oil Separators
3
Oil Vapor Recovery
5
Oil Field - Dehydration Tank
1
Oil Field - Gage Tank
1
Oil Field - Sump, inlet end
1
Oil Field - Sump, outlet end
1
Oil Field - Sump
1
Total
39
2003 Stndv
Gas Wells
20
Oil Wells
40
Oil and Gas Separators
9
Oil Well Tanks
3
Oil Well Casings
2
Gas and Oil Condensate Wells
3
Total
77
4.4.6.3 Results
A total often composite profiles were developed from the measurements collected from the 1991
and 2003 California studies. Table 10 presents the number of oil and natural gas production
profiles for each emission source. The specific profiles are provided in the accompanying
EXCEL workbook, and referenced as "California" in the worksheets. Appendix B also presents
the final speciation profiles.
The analysis of the measurement data allowed for the development of VOC speciation profiles
for oil and natural gas sources located in San Joaquin, California. These speciation profiles have
the following suggested applications: gas wells, oil wells, oil and gas separators, oil well tanks,
oil well casings, gas and oil condensate wells, and oil vapor recovery operations.
41
-------
Table 10. Number of composite profiles for each oi
NuihIkt of I'mlllcs
Oil Wells
2
Oil Tanks
1
Oil Separators
1
Oil Vapor Recovery
1
Gas Wells
1
Oil and Gas Separators
1
Oil Well Tanks
1
Oil Well Casings
1
Gas and Oil Condensate Wells
1
Total
10
and natural gas production emission source.
4.5 Results of SPECIATE Work
For the second phase of this project, EPA Region 8, ORD, and OAQPS, with contract support
from Abt Associates, utilized the information gathered from the RARE ECD field campaign, as
well as other oil and gas measurement studies, to expand upon the VOC speciation profiles
associated with the oil and natural gas sector stored in EPA's SPECIATE Database. This project
reviewed data from the six different studies summarized above, and generated a total of 98
profiles associated with various oil and gas operations. These profiles will be included in the next
release of the SPECIATE Database (version 4.5). Table 11 presents the number of individual
profiles developed from each study reviewed for this project.
In a preliminary review of the data, the VOC speciated profiles vary by oil and gas basin, region
of the United States, the type of pollutants emitted from the sources, and ratio of the pollutants
emitted from the sources. It should be noted that all of these profiles represent uncontrolled
sources. Table 11 presents the number of different species (including groups like "CIO
compounds") that comprised the profiles measured or developed from each of the studies. The
number of species measured or used to speciate the VOC profiles ranged from 24 species to 59
species. A total of 87 different pollutants/groups were obtained among all the profiles developed
for this project. Table 11 also presents the range of methane and highly reactive VOC
contributions for each dataset. The specific profiles are provided in the accompanying EXCEL
workbook. Appendix B also presents the percent contribution by weight of the various pollutants
by profile developed for this project. The results presented in Appendix B are also grouped by
various oil and gas components.
Based on these results, care should be taken in selecting the speciation profiles to apply to
various source categories and basin to basin. When possible, the SPECIATE profiles specifically
associated with operations or sources should be used to speciate the VOC emissions. The types
of controls should also be carefully considered when selecting a speciation profile to develop
speciated emissions for the oil and gas sector. For instance, a speciated profile associated with
the control device should only be used to characterize the controlled portion of the VOC emitted
from the emission stream. Profile developers should also use care and be specific when labeling
and describing the profiles for SPECIATE to assist users in linking the profiles to inventory
emissions. However, additional work is needed by EPA to develop a common set of terminology
and metadata that could be used by profile developers for profile labels, assignments, and
descriptions.
42
-------
Table 11. Number of individual profiles developed from each study reviewed for this project.
Siuih
T\|*'s <>f Sources
#of
Profiles
# ol'
S|K'l'il's
of Mi'llmnc
C oniriluilion |% |
of Ui'iicliu'
VO( Coniriluilion
i% r
RARE ECD Study
Enclosed Combustor Devices
0
0
-
-
WRAP Study
Composite Profiles - Produced
Gas for CBM Wells, Non-CBM
Gas Wells and Oil Wells;
Flashing Gas for Condensate
Tanks and Oil Tanks
17
28
0.008-99.9
0.12-36.02
DJ Basin Direct
Measurement Study
Individual Profiles of Oil and
Natural Gas Production
Condensate Tanks; and One
Composite Profile
28
58
3.08-28.8
26.1-33.23
Utah Indian
Reservations Tribal
Minor Source
Registrations
Individual Profiles of Oil and Gas
Production - Untreated Natural
Gas; Oil Tank Vent Gas;
Condensate Tank Vent Gas;
Glycol Dehydrator; and Four
Composite Profiles
41
24
0.02-89.6
1.6-74.13
Texas Study
Composite Profiles of Oil Tank
Battery Vent Gas and Condensate
Tank Battery Vent Gas
2
33
15.9-37.3
17.42-22.2
California Studies
Composite Profiles of Gas Wells,
Oil Wells, Oil and Gas
Separators, Oil Well Tanks, Oil
Well Casings, Gas and Oil
Condensate Wells, and Oil Vapor
Recovery Operations
10
59
30.7-96.3
2.3-24.1
Total
98
87 2
1 Represents the sum of benzene, ethane, xylenes (xylene/m&p-^lene/m-^lene/o-xylene/p-^lene), toluene, n-
butane,2,2,4-trimethylpentane, and ethylbenzene.
2 This number represents the total number of different species among all of the profiles. Because the species overlap
among the profiles or not all of the same species are covered in all profiles, this number will not represent of sum of
the values in this column.
43
-------
5 SUMMARY
This Fiscal Year (FY) 2014 to FY 2016 RARE project was a collaborative research effort
between EPA Region 8, ORD and OAQPS to improve our understanding of remote measurement
methods to assess VOC control devices and VOC emissions emitted from various oil and gas
operations. The project also utilized information from multiple measurement projects to improve
EPA's SPECIATE database. This report summarized the Region 8 RARE effort that aimed to
improve information on upstream oil and production emissions and identify areas where future
work is needed for this source sector.
The project occurred in two phases. The first phase of the project consisted of a pilot field
campaign in the DJ Basin using off-site remote measurements to assess VOC control efficiency
and emissions of highly reactive VOCs from ECDs at well pads. The goals of the campaign were
to evaluate the performance of measurement technologies to characterize emissions from ECDs
at upstream oil and gas production sites, to provide speciated emissions information from the
ECDs, and to assess the combustion efficiency of the ECDs. The campaign was executed over a
course of five days in September 2014 to collect emissions data from ECDs at multiple well site
locations in a natural gas field of Weld County, Colorado. The campaign utilized two primary
instruments, including aPFTTR radiometer (IMACC, LLC, Round Rock, TX, USA), and a mid-
wave infrared HSI camera (Telops, Quebec City, QC, Canada).
The results of the pilot field campaign indicated that it is important to develop easy-to-use
remote approaches for assessment of ECD operational states, as high combustion efficiencies
cannot be assumed under all conditions. In general, the remote sensing approaches evaluated in
this project were found to be potentially useful as research tools for off-site observation of ECD
operation if more direct on-site measures are not available. However, clear improvements in the
instrumentations and methods would be required to improve the signal strength of the
measurement uncertainty, given that the signals from ECDs are not strong and are temporally
variable. Limitations were found in ease of execution, data analysis throughput, and observable
ECD temperature ranges. Due to the lack of sustained infrared signal from many ECDs, the
PFTIR was not as effective as compared to industrial flare applications. The results of the field
campaign also developed some limited knowledge on potentially emitted byproducts from
improperly operated combustors and their associated VOC reactivity, and collected
measurements to determine component-level emissions from oil and gas control-related devices.
Often well pads investigated, at least one demonstrated evidence of improper ECD operations.
However, the field study was not able to quantify the emissions of highly reactive VOC from
these devices with a high-level of certainty.
The second phase of the project utilized information from oil and gas measurement studies to
improve the VOC speciation profiles for multiple oil and gas basins used for emissions inventory
development and air quality models. The goals of this phase of the project included reviewing
and expanding upon the VOC speciation profiles stored in EPA's SPECIATE Database for oil
and gas sources to ensure that the most recent and representative profiles are available to the
community. The current version of the SPECIATE Database (v4.4) contains non-location-
specific speciation profiles for the oil and gas sector. Many of these profiles were based on test
data collected between 1989 and 2004. Further, the speciation profile generally used for oil and
gas is based on data from 1989 and assumes that ethane and propane are 30 percent of the VOCs,
44
-------
and formaldehyde and methane are 20 percent. This weighted profile of highly reactive VOCs
may not be appropriate to use to generate speciated VOC emissions for most oil and gas
operations because current data and research studies suggest that the speciation of VOC
emissions can vary significantly among oil and gas operations and basins. Without more
representative VOC speciation profiles to generate improved VOC emissions from these types of
sources, the air quality model will have difficulties predicting air quality impacts from this
source sector. As a result, this project reviewed data from a total of six different studies to
develop more recent and process-specific speciation profiles for the oil and gas sector. The data
were consolidated into EXCEL spreadsheets, developed into speciation profiles, reviewed for
completeness and representativeness, analyzed for differences among the available profiles, and
prepared for entry into the SPECIATE Database.
The results of the SPECIATE work generated a total of 98 profiles associated with various oil
and gas operations within multiple oil and gas basins. These profiles will be included in the next
release of the SPECIATE Database (version 4.5). In a preliminary review of the data, the VOC
speciated profiles vary by oil and gas basin, region of the United States, the type of pollutants
emitted from the sources, and ratio of the pollutants emitted from the sources. Based on these
results, care should be taken in selecting the speciation profiles to apply to various source
categories. When possible, the SPECIATE profiles specifically associated with operations or
source should be used to speciate the VOC emissions. The types of controls should also be
carefully considered when selecting a speciation profile to develop speciated emissions for the
oil and gas sector. For instance, a speciated profile associated with the control device should only
be used to characterize the controlled portion of the VOC emitted from the emission stream.
Further, profile developers should use care and be specific when labeling and describing the
profiles for SPECIATE to assist users in linking the profiles to inventory emissions. However,
additional work is needed by EPA to develop a common set of terminology and metadata that
could be used by profile developers for profile labels, assignments, and descriptions.
Better measurements and models not only help protect the environment, but also help facilitate
efficient resource development by alleviating concerns where appropriate. Air emissions from oil
and gas production sites vary based on a number of factors including the composition of the oil
and gas product, age of well, production equipment designs, control devices, and equipment
maintenance states. There is an ongoing need to improve emissions estimates, as well as to
facilitate identification and remediation of compliance issues related to air quality.
45
-------
6
WORK PRODUCTS AND FUTURE WORK
EPA Region 8, ORD, and OAQPS plan to continue this collaborative effort to further investigate
the VOC emissions associated with oil and gas operations and cost-effective measurement tools
that can facilitate leak detection and repair to protect human health and the environment.
Improving our understanding of the emissions from the oil and gas sector is important for
environmentally responsible development of this energy sector. To accurately model the oil and
gas sector impacts on air quality, it is also critical to have accurate activity data, emission factors
and chemical speciation profiles for VOCs and NOx.
The results of this project will support EPA's priorities related to air quality and all three
research themes outlined in ORD's ACE Strategic Research Action Plan 2012-2016, including
assessing impacts associated with air pollutants, providing data and tools to prevent and reduce
emissions of pollutants into the atmosphere, and developing improved air quality models. The
results will also be widely applicable to state/tribal agencies and EPA regions with oil and
natural gas production. The results of the project will be transferred to groups through journal
articles and presentations produced by EPA and collaborators.
The information gathered from this project will be a step towards achieving effective monitoring,
model predictions, and controls of oil and gas emissions, particularly for upstream oil and
production emissions on well pads. However, the work from this project identified the need to
continue this investigation in order to make additional improvements and to fill in data and
measurement tools gaps, including:
Importance to develop easy-to-use remote approaches for assessment of ECD operational
states, as high CE cannot be assumed. In general, the remote sensing approaches used
were found to be potentially useful as research tools for offsite observation of ECD
operation if more direct onsite measures are not available. However, the limitations of the
off-site remote sensing tools investigated for this project were included the ease of
execution, data analysis throughput, and observable ECD temperature ranges.
Importance to develop and utilize site- or process-specific VOC speciation profiles for
interpreting ambient measurement data and creating model-ready emissions for
photochemical modeling applications. In the past, speciation profiles generally used for
oil and gas were based on older data and were not representative of the multiple
processes that exist in the oil and gas operations. While this project develop a number of
speciation profiles to represent a variety of oil and gas processes across multiple basins,
limitations included labeling and describing the profiles for SPECIATE to assist users in
linking the profiles to inventory emissions, and the need to develop a common set of
terminology and meta data that could be used by profile developers for profile labels,
assignments, and descriptions.
Importance to properly account for the impact of controls on VOC speciation.
Conduct photochemical grid modeling with the updated VOC speciation profiles to
evaluate the model performance and understand the model's sensitivity to the new
profiles.
The results of this project will also inform and aid EPA in the design of future field studies and
emissions inventories in areas where oil and natural gas contribute significantly to regional air
46
-------
quality issues. This project also directly addresses the OIG report's statements regarding the
uncertainty and limitations of EPA's air emissions factors and data for oil and natural gas
sources. Specifically, the project improved our knowledge of the emissions uncertainties
associated with ozone precursors for oil and natural gas production sources.
-------
APPENDIX A: PHOTOGRAPHS OF MEASUREMENT SITES
SITE #1
FTIR Location
Distance to Source= 56 m
SITE #2
FTIR Location
Distance to Sources 196 m
48
-------
SITE #3
-------
SITE #4
FTIR Location
Distance to Source= 48 m
-------
SITE #5
FTIR Location
Distance to Source- 45 m
51
-------
SITE #6
--:, " ->s' %'5(S*-
\. ' y"" ฆฆ' 2* " -. *'
v , - -. - - -*
SITE #7
Jffltlh
; *
inr
H
fc
4
n
v
W
FTIR Location
Distance to Source= 90 m
-
52
-------
SITE #8
W ฃ
FTIR Location
Distance to Source= 56 m
ฆ >DซI|QU
-------
SITE #9
SITE #10
FTIR Location
Distance to Source= 40 m
-------
APPENDIX B:
Table B.l: Percent Contribution by Weight of each Chemical Species per Profile Developedfrom this Project
FY14 RARE ECD Project - Oil and Gas VOC Speciation Profiles
I
cjuootor^r^r^
~Q>1 I CO CO CO CO
rv'Sl'Z.'Z.lDlDlDiri
cococococococococococococo^"^"^"
ProfileName
cn ro
^I CH
o
LO I
cn lo
cn
o o
"nT "nT
LO LO
CD (J)
I Unknown
I trans-2-butene
I t-butyl benzene
p-xylene
Propane
I n-undecane
I n-pentane
I n-nonane
I n-heptane
n-decane
I n-xylene
I Methylcyclohexane
I Methane
I Isopropylbenzene
I Isopentane
I Isomers of octane
I Isomers of hexane
Isomers of decane
Isobutane
Ethylene
I Ethane
I Cydohexane
I cis-2-butene
I C8 Paraffin
C-7 Cycloparaffinss
C-6 Compounds
I C-4 Compounds
I C-10 Compounds
I Benzene
I 3-methylpentane
3- methylheptane
2- methylpropyl-benzene
I 2-methylhexane
I 2,4-dimethylpentane
I 2,3-dimethylpentane
I 2,3-dimethylbutane
2,2- di methylpropane
I 2,2,4-trimethylpentane
I l-Methyl-4-ethylbenzene
I l-Methyl-2-ethylbenzene
11-butene
11,3-diethylbenzene
11,2-diethylbenzene
11,2,3 -tri met h yl be nz ene
I trans-2-pentene
Tolu ene
Styrene
Propylene
I o-xylene
I n-propylbenzene
I n-octane
I n-hexane
n-dodecane
In-butane
I Methylcyclopentane
I Methyl alcohol
I m & p-xylene
I Isoprene
Isomers of xylene
I Isomers of nonane
Isomers of heptane
Isomers of butylbenzene
Isobutane
I Ethylbenzene
I Cydopentane
I cis-2-pentene
C-9 Compounds
C-8 Compounds
C-7 Compounds
I C-5 Compounds
I C-ll Compounds
I Butylbenzene
Acetylene
3- methylhexane
3-ethylhexane
I 2-methylpentane
I 2-methylheptane
I 2,4-di methylhexane
I 2,3-di methylhexane
2,3,4-trimethylpentane
I 2,2-dimethylbutane
11-pentene
I l-Methyl-3-ethylbenzene
11- hexe ne
11,4-diet hylbenzene
1,3,5 -tri met h yl be nz ene
11,2,4-tri methyl benzene
55
-------
Table B.2:
Project
Percent Contribution by Weight of Highly Reactive VOCs per Profile for Gas Condensate Tanks Developed from this
FY14 RARE ECD Project- Oil and Gas VOCSpeciation Profiles for GasCondensateTanks
100
I n-Pentane
90
80
70
u
"5 60
50
40
30
20
10
I Isopentane
I Isobutane
I Ethylbenzene
I 2,2,4-Trimethylpentane
I n-Butane
I Toluene
Xyl en es
Ethane
I Benzene
I Methane
CO LO
CO (B CO
cocococococooocoooo^cncnaiCDcTiCDCDCD
LDLnLDLnLnLnLnCD^D^D
LDLnLnLnLnLnLnLnLnLnLnLnLDLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLnLn
cTicT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)a)O^CT)CT)CT)CT)O^CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)CT)a)O^CT)CT)Lno^
< <
Q Q Q
Q Q Q
Q Q
Q Q
QQQQQQQQQQQQQQQQQQQ
QQQQQQQQQQQQQQQQQQQ
Q Q Q
Q Q Q
Profile Na me
56
-------
Table B.3: Percent Contribution by Weight of Highly Reactive VOCs per Profile for Oil Tanks Developed from this Project
FY14RARE ECD Project-Oil andGasVOCSpeciationProfilesforOilTanks
100
90
80
70
<-> 60
50
40
30
20
10
I n-Pentane
I Isopentane
I Isobutane
I Ethylbenzene
I 2,2,4-Trimethylpentane
I n-Butane
I Toluene
Xylenes
Ethane
I Benzene
I Methane
5 -2
O
H
Z>
O
H
Z>
a z> w
ฆ o
H
Z>
Profile Name
' o
ฃ
<
U
<
u
57
-------
Table B.4: Percent Contribution by Weight of Highly Reactive VOCs per Profile for Gas Condensate Tanks Developed from this
Project
100
90
O
&_
Q_
Q_
Q.
00
_C
U
80
70
60
50
D
_Q
C
o
(J
-M
c
Q)
Q_
30
20
10
FY14 RARE ECD Project - Oil and Gas VOC Speciation Profiles for Gas Wells
I n-Pentane
I Isopentane
I Isobutane
I Ethylbenzene
I 2,2,4-
Trimethylpentane
I n-Butane
I Toluene
Xylenes
I Ethane
I Benzene
I Methane
J.
CO
CO
1
H
T1
(J
u
O
o
CO
00
OT,
H
>
(J
CC
Q_
3
Q
Q_
Q_
Q_
<
CC
Q_
Q_
Q_
<
<
<
<
CC
5
5
CC
5
CC
5
CC
5
O
u
Q.
<
DC
o
u
CO
CC
Q_
<
CC
O
u
CO
CC
Q_
Q_
<
CC
t1
H
rsi
ID
Is*
00
CD
5
ti
rsi
ro
5
LO
KD
Is*
00
CD
o
Is*
ro
o
;z:
o
CO
ro
ro
ro
LO
ti
O
CC
>
H
CO
ro
ro
ro
ro
ro
ro
ro
ro
ro
ro
ro
ro
ro
ro
z.
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
LO
CD
CT)
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
CD
Q_
00
n
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
CC
<
Q_
Q_
<
q:
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
U
<
q:
5
<
CC
5
5
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Profile Name
58
-------
Table B.5: Percent Contribution by Weight Highly Reactive VOCs per Profile for Oil Wells Developed from this Project
FY14 RARE ECD Project - Oil and Gas VOC Speciation Profiles for Oil Wells
O
i_
Q_
U >
S O
< 2
2 e-
ง o
u
o
rn fx
LO
+J
'to
o
Q.
E
o
(J
CD
<
U
O
ฐ u
LO
CD '
- (U
<
CJ
(J
Profile Name
59
-------
Table B.6: Percent Contribution by Weight of Highly Reactive VOCs per Profile for Separators/Vapor Recovery/Well Casings
Developed from this Project
FY14 RARE ECD Project - Oil and Gas VOC Speciation Profiles for Separators/Vapor Recovery/Well Casings
100
n-Pentane
o
Q_
ru
i_
ro
^ Q.
(U
O oo
^ <
S s
51 i
(U
<
(j
o
Q.
E
o
(J
^ Q.
CM
<
>.
r---
o
U
i
OJ
>
o
o
LO
<
(J
to
CUD
_C
+-ป
u
en
(U
to
'to
O
(U
ro
cd
<
'to
o
U
Q.
u
E
Q.
o
u
E
o
u
I Isopentane
I Isobutane
I Ethylbenzene
I 2,2,4-
Trimethylpentane
In-Butane
I Toluene
Xylenes
Ethane
I Benzene
I Methane
Profile Name
60
-------
Table B.7: Percent Contribution by Weight of Highly Reactive VOCs per Profile for Gas Condensate Tanks Developedfrom this
Project
100
FY14 RARE ECD Project - Oil and Gas VOC Speciation Profiles for Dehydrators
n-Pentane
o
Q_
+->
ru
n:
CD
O
~G
>
<
O
l
~G
>
<
I
I
LO
(D
O
LO
Q
LO
51
o>
cc
cc
cc
00
o
00
o
00
>.
>.
H
(J
H
(J
H
ฆa
>
_c
(D
Q
"o
u
_>-
(J
ฆa
>
ฃ S
ฃ 8
s >-
H (J
1
o
o
+->
+->
ru
n:
ro
ti
~G
>
<
I
~G
>
<
LO
1
LO
(D
O
LO
(U
Q
LO
51
51
51
cc
cc
cc
00
o
00
o
00
>.
>.
H
(J
H
(J
H
1
o
I
O
o
4'
+->
o
ru
ru
_>*
~G
s
ฆa
O
s_
>
_C
_c
dJ
rsi
LO
H
Z)
o
+->
n:
Q
G)
Q
51
i
(U
ฆa
~o
CC
00
~o
cc
>
_c
u
o
00
'to
(U
Q
>.
>.
O
(J
H
(J
H
Q.
E
o
u
H
H
Z)
Z)
I Isopentane
I Isobutane
I Ethylbenzene
I 2,2,4-
Trimethylpentane
I n-Butane
I Toluene
Xylenes
Ethane
I Benzene
I Methane
Profile Name
61
-------
Table B.8: Percent Contribution by Weight of each Chemical Species per Profile for EPA Region 8-Specific Profiles Developed
from this Project
100
90
80
70
60
50
o 40
30
20
10
0 I-
o
3
DJ Bssin, > Wyoming
FY14 RARE ECD Project-Oil and GasVOCSpeciationProfiles
CO n3
Z> H
CO >
5 Uinta Basin, Utah
'Profile Namef
I Unknown
I trans-2-butene
I t-butyl benzene
p-xyl en e
P ro pa ne
I n-undecane
I n-pentane
I n-nonane
I n-heptane
n-decane
I n-xylene
I Methylcyclohexane
I Methane
I Isopropylbenzene
I Isopentane
I Isomers of octane
I Isomers of hexane
Isomers of decane
Isobutane
Ethylene
I Ethane
I Cyclohexane
I cis-2-butene
I C8 Paraffin
C-7 Cycloparaffinss
C-6 Compounds
I C-4 Compounds
I C-10 Compounds
I Benzene
I 3-methylpentane
3- methylheptane
2- methylpropyl-benzene
I 2-methylhexane
I 2,4-dimethylpentane
I 2,3-dimethylpentane
I 2,3-dimethylbutane
2,2-dimethylpropane
I 2,2,4-trimethylpentane
I l-Methyl-4-ethylbenzene
I l-Methyl-2-ethylbenzene
11-butene
11,3-diethyl benzene
11,2-diethyl benzene
11,2,3-trimethyl benzene
I trans-2-pentene
Toluene
Styren e
Propylene
I o-xylene
I n-propylbenzene
I n-octane
I n-hexane
n-dodecane
I n-butane
I Methylcyclopentane
I Methyl alcohol
I m & p-xylene
I Isoprene
Isomers of xylene
I Isomers of nonane
Isomers of heptane
Isomers of butylbenzene
Isobutane
I Ethyl benzene
I Cyclopentane
I cis-2-pentene
C-9 Compounds
C-8 Compounds
C-7 Compounds
I C-5 Compounds
I C-ll Compounds
I Butylbenzene
Acetylene
3-methylhexane
3-ethylhexane
I 2-methylpentane
I 2-methylheptane
I 2,4-dimethylhexane
I 2,3-dimethylhexane
2.3.4-trimethylpentane
I 2,2-dimethylbutane
I 1-pentene
I l-Methyl-3-ethylbenzene
I 1-hexene
I 1,4-diethylbenzene
1.3.5-trimethyl benzene
I 1,2,4-trimethyl benzene
62
-------
Table B.9: Percent Contribution by Weight of Highly Reactive VOCs per Profile for EPA Region 8-Specific Profiles Developed
from this Project
FY14 RARE - O&G VOC Speciation Profiles
100
90
80
70
60
50
40
30
20
10
I n-pentane
I Isopentane
I Ethylbenz
I 2,2,4 TMP
I n-butane
I Toluene
Xylene
Ethane
I Benzene
I Methane
in
in
c
in
in
in
in
in
in
X
c
~a3
o
"a3
~aj
c
c
~Z.
c
c
Q
I
ฃ
ti
5
I
D
I
ฃ
cc
u
~o
5
5
O
u
CC
u
O
LO
CQ
o
CQ
CQ
LO
Q_
cl
U
U
Q_
Q_
CC
cc
H
<
i
i
~Z.
<
<
H
LO
co
cc
-z.
Q_
<
CC
CO
Q_
cc
CC
Q
<
CC
WRAP
5
H
H
DJ Basin, g Wyoming
Profile Name
Uinta Basin, Utah
63
-------
APPENDIX C: Top 20 Producers in the Uintah and Ouray Reservations -2014
Profile #
Name
2014 Data
Production
Mcf
95336
Natural Gas - Untreated - Oil Well - Axia Energy II, LLC
1406147
95337
Natural Gas - Untreated - Oil Well - Bill Barrett Corporation
11389526
95338
Natural Gas - Untreated - Oil Well - Crescent Point Energy U.S. Corp
2854439
95339
Natural Gas - Untreated - Oil Well - Enduring Resources, LLC
1213888
95340
Natural Gas - Untreated - Oil Well -EOG Resources, Inc.
35547477
95341
Natural Gas - Untreated - Oil Well - Koch Exploration Company
422013
95342
Natural Gas - Untreated - Oil Well -Newfield Production Company
11888643
95343
Natural Gas - Untreated - Oil Well - QEP Energy Company
24401675
95344
Natural Gas - Untreated - Oil Well - QEP Field Services Company
0
95345
Natural Gas - Untreated - Oil Well - Red Rock Gathering Company, LLC
0
95346
Natural Gas - Untreated - Oil Well - Rosewood Resources, Inc.
596571
95347
Natural Gas - Untreated - Oil Well - Ultra Resources, Inc.
1450111
95348
Natural Gas - Untreated - Oil Well -Ute Energy, LLC
0
95349
Natural Gas - Untreated - Oil Well - Whiting Petroleum Company
3664200
95350
Natural Gas - Untreated - Oil Well -XLO Energy, Inc
10740094
95409
Oil Field - Glycol Dehydrator - EOG Resources, Inc.
655458
95410
Oil Field - Glycol Dehydrator - Koch Exploration Company
3436
95411
Oil Field - Glycol Dehydrator - Newfield Production Company
7043408
95412
Oil Field - Glycol Dehydrator - QEP Energy Company
1074473
95413
Oil Field - Glycol Dehydrator - QEP Field Services Company
0
95414
Oil Field - Glycol Dehydrator - Ultra Resources, Inc.
1397871
95415
Oil Field - Glycol Dehydrator - Whiting Petroleum Company
9011
95416
Oil Field - Glycol Dehydrator - XLO Energy, Inc
97486
Oil Produced
Barrels
95353
Oil Field - Condensate Lank Battery Vent Gas - El Paso Midstream Group, Inc
0
95354
Oil Field - Condensate Lank Battery Vent Gas - Enduring Resources, LLC
10641
95355
Oil Field - Condensate Lank Battery Vent Gas - EOG Resources, Inc.
655458
95356
Oil Field - Condensate Lank Battery Vent Gas - Gasco Energy, Inc
74545
95357
Oil Field - Condensate Lank Battery Vent Gas - Kerr-McGee Oil and Gas Onshore LP
1042197
95358
Oil Field - Condensate Lank Battery Vent Gas - Koch Exploration Company
3436
95359
Oil Field - Condensate Lank Battery Vent Gas - QEP Energy Company
1074473
95360
Oil Field - Condensate Lank Battery Vent Gas - QEP Field Services Company
0
95361
Oil Field - Condensate Lank Battery Vent Gas - Rosewood Resources, Inc.
8091
95363
Oil Field - Condensate Lank Battery Vent Gas - Whiting Petroleum Company
9011
95364
Oil Field - Condensate Lank Battery Vent Gas - XLO Energy, Inc
97486
95351
Oil Field - Oil Lank Battery Vent Gas - Axia Energy II, LLC
1106170
95352
Oil Field - Oil Lank Battery Vent Gas - Bill Barrett Corporation
3151243
95362
Oil Field - Oil Lank Battery Vent Gas - Ultra Resources, Inc.
1397871
65
-------
SEPA
United States
Environmental Protection
Agency
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Printed on 100% recycled/recyclable paper
with a minimum 50% post-consumer
fiber using vegetable-based ink.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
------- |