EPA/600/R-13/183F I August 2013 I www.epa.gov/ncea
United States
Environmental Protection
Agency
The National Dioxin Air Monitoring
Network (NDAMN)
Report of the Results of Atmospheric Measurements of Polychlorinated
Dibenzo-p-Dioxins (PCDDs), Polychlorinated Dibenzofurans (PCDFs),
and Dioxin-Like Polychlorinated Biphenyls (PCBs) in Rural and
Remote Areas of the United States from June 1998 through
November 2004
National Center for Environmental Assessment
Office of Research and Development
-------
EPA/600/R-13/183F
August 2013
The National Dioxin Air Monitoring Network
(NDAMN)
Report of the Results of Atmospheric Measurements of
Polychlorinated Dibenzo-/?-Dioxins (PCDDs), Polychlorinated
Dibenzofurans (PCDFs), and Dioxin-Like Polychlorinated
Biphenyls (PCBs) in Rural and Remote Areas of the United
States from June 1998 through November 2004
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
Preferred Citation:
U.S. Environmental Protection Agency (EPA). (2013) The National Dioxin Air Monitoring
Network (NDAMN): Report of the results of atmospheric measurements of polychlorinated
dibenzo-^-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and dioxin-like
polychlorinated biphenyls (PCBs) in rural and remote areas of the United States from June 1998
through November 2004. National Center for Environmental Assessment, Washington, DC;
EPA/600/R-13/183F. Available at http://epa.gov/ncea.
11
-------
CONTENTS
LIST OF TABLES iv
LIST OF FIGURES v
ABBREVIATIONS AND ACRONYMS vi
PREFACE vii
AUTHORS AND REVIEWERS viii
EXECUTIVE SUMMARY x
1. INTRODUCTION 1-1
1.1. REFERENCES 1-3
2. AMBIENT AIR SAMPLING 2-1
2.1. SAMPLING OBJECTIVE 2-1
2.2. SAMPLING DESIGN 2-1
2.3. AIR SAMPLING PROCEDURES 2-2
2.3.1. Description of the Tisch Environmental® TE1000 PUF (PS-1
sampler) 2-2
2.3.2. Preparation of PUF Sampling Cartridge 2-4
2.3.3. Multipoint Calibration of the PS-1 sampler 2-4
2.3.4. Sample Collection 2-5
2.3.5. Field Quality Assurance (QA)/Quality Control (QC) Procedures 2-6
2.4. REFERENCES 2-7
3. ANALYSIS OF NDAMN SAMPLES 3-1
3.1. OVERVIEW OF THE ANALYTICAL EPA METHOD 3-1
3.2. QUALITY CONTROL (QC) OF LAB SAMPLES AND CALIBRATION 3-3
3.3. REFERENCES 3-4
4. OVERVIEW OF RESULTS 4-1
4.1. OVERVIEW OF NETWORK OPERATION STATUS 4-1
4.2. QUALITY ASSURANCE (QA) MEASURES AND RESULTS 4-2
4.2.1. Blank Samples 4-2
4.2.2. Co-Located Sampler Results 4-6
4.2.3. NDAMN Sample Volume 4-8
4.2.4. A Comment on NDAMN Quality Assurance (QA) 4-11
4.3. OVERVIEW OF FINAL RESULTS FROM NDAMN 4-11
4.4. REFERENCES 4-16
APPENDIX A. DESCRIPTION OF EXCEL WORKBOOK CONTAINING NDAMN
DATA A-l
APPENDIX B. QUALITY ASSURANCE PROJECT PLAN FOR NDAMN B-l
in
-------
LIST OF TABLES
1-1. The toxic equivalency factors (TEFs) for the dioxin-like compounds using the
WHO approach 1-5
2-1. Description of theNDAMN air monitoring stations 2-8
2-2. Dates and seasons of the 29 sampling moments of NDAMN 2-10
3-1. NDAMN analytes measured and their detection limits 3-5
4-1. Operating status of all stations over all moments 4-17
4-2. Comparison of congener-specific results from laboratory blanks with those from
remote, rural, and urban stations of NDAMN 4-19
4-3. Comparison of average concentrations in the NDAMN monitors, #1 and #3, with
the adjacent QA sampler, #2, the correlation coefficient, r, between the set of
measurements, and the average Relative Percent Difference, RPD, between the set
of measurements 4-21
4-4. Comparison of congener concentrations in co-located samplers in Pennsylvania
for two occurrences with discrepencies in volume (Spring and Summer, 2000) and
all co-located samples inPennsyania 4-23
4-5. Survey-wide statistics for all congeners and homologue groups 4-25
4-6. Comparison of the four highest concentrations measured with the station average
where that concentration was measured 4-27
IV
-------
LIST OF FIGURES
Figure 2-1. Geographic locations of the NDAMN monitoring stations 2-11
Figure 2-2. Illustration of the components of the Tisch Environmental® TE-1000 PUF
(PS-1 sampler) 2-11
Figure 4-1. Comparison of control and NDAMN results for 2,3,7,8-TCDD (the perfect
correlation^ = x is shown as dashed line) 4-29
Figure 4-2. Comparison of control and NDAMN results for PCB 156 (the perfect
correlation^ = x is shown as dashed line) 4-30
Figure 4-3. Comparison of control and NDAMN sample volumes (the perfect correlation
y = x is shown as dashed line) 4-31
Figure 4-4. Average TEQ concentrations found at all NDAMN stations 4-31
Figure 4-5. Temporal variability of TEQ concentrations averaged by station
characterization 4-32
-------
ABBREVIATIONS AND ACRONYMS
CAD AMP California Ambient Dioxin Air Monitoring Program
CDEP Connecticut Department of Environmental Protection
DL detection limit
dl-PCBs dioxin-like PCBs
EPA Environmental Protection Agency
OFF glass fiber filter
HpCDD heptachlorodibenzo-p-dioxin
HpCDF heptachlorodibenzofuran
HRGC high-resolution gas chromatography
FIRMS high-resolution mass spectrometry
HxCDD hexachlorodibenzo-p-dioxin
HxCDF hexachlorodibenzofuran
IMPROVE Interagency Monitoring of Protected Visual Environments
NA data not available
NADP National Atmospheric Deposition Program
ND not detected
NDAMN National Dioxin Air Monitoring Network
NTN National Trends Network
OCDD octochlorodibenzo-p-dioxin
OCDF octachlorodibenzofuran
PCB polychlorinated biphenyl
PCDD polychlorinated dibenzo-p-dioxin
PCDF polychlorinated dibenzofuran
PeCB pentachlorobiphenyl
PeCDD pentachlorodibenzo-p-dioxin
PeCDF pentachlorodibenzofuran
PUF polyurethane foam
QA quality assurance
QAPP Quality Assurance Proj ect Plan
QC quality control
QFF quartz fiber filter
RF response factor
RPD relative percent difference
RSD relative standard deviation
S/N signal-to-noise
TCB tetrachlorobenzene
TCDD tetrachlorodibenzo-p-dioxin
TCDF tetrachlorodibenzofuran
TEF toxicity equivalency factor
TEQ toxic equivalent
WHO World Health Organization
VI
-------
PREFACE
To help characterize the ubiquitous presence of dioxins in the environment, the
U.S. Environmental Protection Agency (EPA) established the National Dioxin Air Monitoring
Network (NDAMN) in 1998. The objective of NDAMN was to determine background air
concentrations of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans
(PCDFs), and dioxin-like polychlorinated biphenyls (dl-PCBs) in the United States. NDAMN
began operation on June 16, 1998, with 10 NDAMN sampling stations. Stations were added
over time and composed of 34 by the beginning of 2003. The last sample of NDAMN was taken
in November 2004. The full database is composed of 685 samples, measured for 17 dioxin and
furan congeners, 8 dioxin and furan homologue groups, and 12 dioxin-like polychlorinated
biphenyl (PCB) congeners. The overall average total toxic equivalent (TEQ) concentration was
11.1 fg/m3 with dioxin-like PCBs contributing only 0.8 fg/m3 (7%) of this total. The purpose of
this document is to provide information on the overall purpose, design, implementation,
analytical chemistry, and results of NDAMN. This document also accompanies an NDAMN
database made available now so that others can use the individual sample data for their own
purposes.
This final document reflects a consideration of peer review comments received on March
11, 2013, for an external review draft dated March 2012 (EPA/600/P-04/001 A).
vn
-------
AUTHORS AND REVIEWERS
The National Center for Environmental Assessment, Office of Research and
Development was responsible for preparing this document.
AUTHORS
Matthew Lorber
Exposure and Risk Characterization Group
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Joseph Ferrario
Environmental Chemistry Laboratory
U.S Environmental Protection Agency
John C. Stennis Space Center
Bay St. Louis, MS
Christian Byrne
Environmental Chemistry Laboratory
U.S Environmental Protection Agency
John C. Stennis Space Center
Bay St. Louis, MS
David Cleverly (retired)
Exposure and Risk Characterization Group
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
INTERNAL EPA REVIEWERS
Craig Vigo
Environmental Chemistry Laboratory
U.S Environmental Protection Agency
John C. Stennis Space Center
Bay St. Louis, MS
Vlll
-------
AUTHORS AND REVIEWERS (continued)
The following panel of experts reviewed a draft of this document and the spreadsheet
database, and provided comments.
EXTERNAL PEER REVIEWERS
Stuart Harrad
Professor of Environmental Chemistry
University of Birmingham
Birmingham, UK
Gary T. Hunt
Vice President and Principal Scientist
TRC Corporation
Lowell, MA
Michael C. McCarthy
Senior Atmospheric Scientist
Sonoma Technology, Inc.
Petaluma, CA
ACKNOWLEDGMENTS
The authors recognize the staff of the U.S. EPA Environmental Chemistry Laboratory in
Mississippi for the extraction and analysis of the air samples during the 6 years of sample
collection. The authors also recognize the staff of Battelle of Columbus, Ohio, who was
responsible for the implementation of the field program which entailed siting and maintaining
the air samplers, and shipping air samples to the Environmental Chemistry Laboratory for
analysis. Finally, the authors appreciate the careful review and edit of this report by Terri
Konoza of the National Center for Environmental Assessment.
IX
-------
EXECUTIVE SUMMARY
To help characterize the ubiquitous presence of dioxins in the environment, the
U.S. Environmental Protection Agency (EPA) established the NDAMN in 1998. The objective
of NDAMN was to determine background air concentrations of PCDDs, PCDFs, and dioxin-like
PCBs (dl-PCBs) in the United States. "Background" is defined as areas where there is no
expected influence of nearby known dioxin sources. To meet this objective, sampling focused
on rural and remote areas, although a few stations were added that are closer to urban areas.
NDAMN began operation on June 16, 1998, with 10 NDAMN sampling stations.
NDAMN was expanded to 23 sampling stations by the last sample event in 1999. It was then
expanded to 30 stations at the end of 2000, 32 stations at the end of 2001, 34 stations at the end
of 2002, and finally, 35 at the end of 2003 when the last station was added for the first sampling
event. Sampling concluded in September 2004. The count of 35 stations includes Quality
Assurance/Quality Control (QA/QC) Station 2, which will not be considered as a sampling
station; therefore, the full NDAMN data set is considered to be composed of 34 stations.
Sampling occurred four times per year, roughly corresponding to the four seasons of the year.
Each sampling event was termed a "sampling moment" in which all NDAMN samplers were in
operation, and each event consisted of 20 to 24 days of active sampling over a 28-day period, on
a weekly schedule of 5 or 6 days of continuous operation followed by 1 or 2 days of inactivity.
Sampling was conducted with a Tisch Environmental® TE1000 polyurethane foam (PUF) (PS-1
sampler) in accordance with procedures described in EPA Method TO 9A, as revised in the
Quality Assurance Project Plan for NDAMN. The PS-1 sampler is equipped with a quartz fiber
filter (QFF) and a PUF adsorbent plug for collecting particulate matter and gaseous compounds,
respectively. Each week, the QFF was harvested, and a new QFF was placed in the sampler,
yielding four QFFs per sampling moment. This was done to prevent saturation and clogging of
the filter media with collected particles. With this procedure, each sampling moment entailed a
collection of air mostly in the range of 6,000 to 8,000 m3 of volume.
The harvested samples (PUFs/QFFs) and their associated field blanks were shipped to
EPA's Environmental Chemistry Laboratory in Mississippi for extraction, clean-up, and analysis
by high-resolution gas chromatography coupled with high-resolution mass spectrometry in
accordance with a modification of EPA Method 1613. Four sample sets were generated for each
-------
sampling event at each NDAMN station: (1) one PUF filter from active sampling, (2) one PUF
field blank, (3) one set of four QFFs from active sampling, and (4) one set of four QFF field
blanks. Field blanks were used to determine contamination affecting the active samples (which
are passively exposed only during setup and collection), so field blanks were only exposed to
ambient air during sample setup and collection. Analytes measured include 17 dioxin and furan
congeners, 8 dioxin and furan homologue groups, and 12 polychlorinated biphenyls (PCBs) that
have dioxin-like toxicity. These PCBs are commonly referred to as dl-PCBs. All samples had
seven PCBs, and PCBs 81, 114, 123, 167, and 189 were added in the summer sample of 2002.
The analytical detection limits (DLs) ranged from 0.5 pg for tetra congeners to 20 pg for octa
congeners, and from 1 pg (PCB 69) to 500 pg (PCB 118) for the individual PCBs.
Sample-specific DLs expressed on a concentration basis can be calculated by dividing these
masses by the actual volume for each sampling event.
If all 34 sampling stations operated for all moments following their initially collected
moment, there would be a total of 736 samples. However, only 685 sampling events were
completed. There were 51 sampling events that were not completed and these were
characterized as data not available (NA). Causes for NAs include (1) station not operating
(26 times), (2) QA failure at the lab, all analytes (eight times), (3) QA failure at the lab,
PCDDs/PCDFs only, PCB analysis available but not included in survey results (seven times), (4)
sample volume data lost (two times), and (5) low sample volume (less than 2,000 m3) (eight
times). The protocol to obtain four weeks of air volume guaranteed low DLs and a high
detection frequency. The frequency of positive measurements was mostly above 95% and at
85% for 2,3,7,8-tetrachlorodibenzo-/?-dioxin (TCDD). The lowest detection frequency was 74%
for 1,2,3,7,8,9-hexachlorodibenzofuran (HxCDF). All results in this report have been generated
assuming not detected (ND) = 0 (but a quick check on a few averages showed virtually no
change when assuming ND = /^ DL).
The overall average TEQ concentration was 11.1 fg/m3 with dioxin-like PCBs
contributing only 0.8 fg/m3 (7%) of this total (with PCB 126 explaining most of this PCB
contribution, -88%). The top six contributors explained 67% of the TEQ (1,2,3,7,8-PeCDD at
27.8%; 2,3,4,7,8-PeCDF at 11.4%; 1,2,3,4,6,7,8-HpCDD at 9.1%; 1,2,3,6,7,8-hexachloro-
dibenzo-^-dioxin [HxCDD] at 6.5%; 1,2,3,7,8,9-HxCDD at 6.4%; and PCB 126 at 6.1%). The
compound 2,3,7,8-TCDD contributed 5% to the TEQ. All dioxin-like PCBs, excluding PCB
xi
-------
126, contributed about 0.9% of the TEQ. The archetype dioxin and furan background air
congener profile was seen in the survey averages and in most individual samples. This archetype
profile is characterized by low and similar concentrations for tetra- through hexa dioxins and
furan congeners, with elevations in 1,2,3,4,6,7,8-HpCDD, OCDD, 1,2,3,4,6,7,8-HpCDF, and
OCDF.
Average TEQ concentrations throughout most of the Eastern Seaboard and into the
central part of the United States range between 5 and 20 fg/m3. From the central part of the
United States into the western portion, as well as Alaska, excluding California, the average TEQ
concentration appears to be near or less than 5 fg/m3. Two of the stations on the Western
Seaboard, one in California and one in Oregon, showed average TEQ concentrations just above
20 fg/m3. The other four stations on the Western Seaboard showed average TEQ concentrations
less than 10 fg/m3. Station 20, Fond du Lac Indian Reservation in Minnesota, showed the
highest average TEQ concentration at 47 fg/m3, but this was skewed by a single outlier at 847 fg
TEQ/m3. Without this concentration, the average for the station was 6.9 fg TEQ/m3. Station 28,
Rancho Seco (closed nuclear power plant), was also influenced by a single high concentration,
although not as much. The station average concentration of 36 fg TEQ/m3 was reduced to 21 fg
TEQ/m3 by removing the high concentration of 241 fg TEQ/m3.
Stations were generally categorized as either urban (4 stations), rural (23 stations), or
remote (7 stations). These characterizations were for purposes of this study and should not be
considered representations of any of these three land-use categorizations, particularly for urban.
The average TEQ concentrations over all stations and moments within these categories were
(1) urban at 15.9 fg TEQ/m3, (2) rural at 13.9 fg TEQ/m3, and (3) remote at 1.2 fg TEQ/m3. An
examination of trends over time suggests that the rural stations, as a group, may show elevations
during the fall or winter months as compared to the spring or summer months. Perhaps that
could be said as well for urban stations, but the remote stations appear to show little variation
over the course of a year. Concentrations of dioxin-like compounds appear to be constant
between 1998 and 2004, with no evidence of either a decline or rise in concentrations.
An examination of the four highest measurements reveals some interesting trends. The
locations of these high measurements and the TEQ concentrations are (1) Station 20, Fond du
Lac Indian Reservation in Cloque, MN, at 847 fg TEQ/m3; (2) Station 3, Clinton Crops Research
Station in Clinton, NC, at 292 fg TEQ/m3; (3) Station 28, Rancho Seco in Herald, CA, at 241 fg
xii
-------
TEQ/m3; and (4) Station 29, Hyslop Farm, Oregon Agricultural Experiment Station in Albany,
OR, at 132 fg TEQ/m3. For Stations 3 and 20, the concentrations of all congeners and
homologue groups in the anomalous reading are substantially higher (from 10 to over 100 times
higher) than the station averages. The only pattern for these two stations is that everything
appears elevated. For Stations 28 and 29, a very different picture emerges. The concentrations
for only the dioxin congeners and dioxin homologue groups are 10 to more than 100 times higher
than the station averages. For the furan congeners, furan homologues, and PCBs, the
concentrations are only slightly elevated or less than the station averages. It is also noted that the
station averages of dioxins (congeners and homologues) for Stations 28 and 29 are generally
higher (by a factor of 2) than averages for Stations 3 and 20. Meanwhile, furan and PCB
congener/homologue group averages are about the same for all four stations. These trends
suggest a source near Stations 28 and 29 that might occasionally elevate dioxin concentrations
leading to potentially very high levels. These sources do not appear to influence the general
background of furans and PCBs. This pattern of exaggerated elevation in dioxins with
essentially background levels of furans was also found in ball clay, which was discovered to be a
contaminant in animal feed in the 1990s. Research on dioxins in ball clay from animal feeds
showed TEQ concentrations above 1,500 ppt (for comparison, soil TEQ concentrations are
typically 10 ppt or less), explained in full by elevated dioxins while furan concentrations were
either absent or at least two orders of magnitude lower than dioxin concentrations.
Investigations have not occurred to identify potential sources near Sites 28 and 29; certainly it is
possible that combustion of a product high in PCDDs in comparison to PCDFs, such as clays,
might explain the findings in these monitors. Thermal processes that preferentially emit dioxins
over furans could also be the cause.
xin
-------
1. INTRODUCTION
Polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs),
and dioxin-like polychlorinated biphenyls (dl-PCBs) represent a class of toxic semi-volatile
aromatic compounds commonly referred to as being "dioxin-like." The dioxin-like classification
combines these organic compounds into a single chemical class defined as having analogous
chemical and physical properties, chlorine substitution patterns, a planar molecular orientation, a
common mode of action of toxicity in mammals, and common endpoints or manifestations of
toxicity. The compound 2,3,7,8-tetrachlorodibenzo-/>-dioxin (TCDD) is the prototype
dioxin-like compound and serves as the reference compound to this class.
The PCDDs and PCDFs are aromatic hydrocarbons consisting of a triple-ring structure of
two benzene rings interconnected by a third oxygenated ring. There are eight positions whereby
a chlorine atom(s) can be attached. Theoretically, 75 PCDD and 135 PCDF congeners are
possible, and their physical chemical properties are determined by the number of chlorine atoms
and their respective positions on the molecular nucleus. The environmental effects and
toxicology of PCDDs/PCDFs are largely mediated and controlled by the presence of chlorine
atoms in the 2,3,7,8 positions. There are 7 PCDDs and 10 PCDFs with this substitution pattern.
Polychlorinated biphenyls (PCBs) are aromatic hydrocarbon compounds consisting of
two benzene rings of carbon atoms interconnected by carbon to carbon bonds. The generalized
molecular formula (chemical class) of PCBs is Ci2Hio-nCln, where n is the number of chlorine
atoms (in a range of 1 to 10) substituting for hydrogen atoms on the biphenyl rings. Although
there is the possibility for 209 PCB congeners, this report focuses on the dioxin-like PCBs. The
dioxin-like PCBs are nonortho substituted compounds with chlorine atoms on the para and,
minimally, two meta positions to the molecule. These are considered as having structural
conformity to 2,3,7,8-TCDD. Other dioxin-like congeners include those PCBs having only one
chlorine in the ortho position. There are 12 dioxin-like PCBs, and all were measured in this
program.
In many situations, PCDDs, PCDFs, and dioxin-like PCBs appear as mixtures in
environmental samples. The total toxic equivalent (TEQ) procedure is an accepted convention
for converting the total concentration of toxic congeners within the mixture to an equivalent
concentration of 2,3,7,8-TCDD (the most toxic member of the class). This procedure involves
1-1
-------
assigning individual toxic equivalency factors (TEFs) to the 2,3,7,8-substituted polychlorinated
dibenzo-p-dioxin/polychlorinated dibenzofuran (PCDD/PCDF) congeners and dioxin-like PCBs,
and then summing the product of each congener concentration multiplied by its respective TEF,
as follows:
TEQ = y] t_n (congener\ x TEFt) + (congener'. x TEFj j + (congenern x TEFn)
where:
TEQ = total toxic equivalent of the mixture of PCDDs, PCDFs, and dioxin-like PCBs to
the reference compound 2,3,7,8-TCDD
TEF = toxic equivalency factors assumed for each PCDD, PCDF, and dioxin-like PCB
congener
TEF values are equal to or less than 1.00, with values as low as 0.00001. TEF schemes
have been published in 1994 (Van den Berg et al., 1994), 1998 (Van den Berg et al., 1998), and
2006 (Van den Berg et al., 2006). These publications represent World Health Organization
(WHO) consensus opinions on the final values of the TEFs, and the schemes have been
abbreviated as WHO 1994, 1998, and 2006. In 2010, EPA formally adopted the WHO 2006
TEF scheme (U.S. EPA, 2010). These TEFs for dioxin-like compounds are used in this report
and are provided in Table 1-1.
Dioxin-like compounds are extremely persistent in soils and sediments, bind to organic
carbon, and readily accumulate in fatty tissues of animals. Although there is evidence that
dioxin-like chemicals can be formed in nature, the dominant sources to the environment are
inherently anthropogenic. Combustion-related activities such as incineration of human-
generated waste materials, secondary and primary metal smelting, the production of steel,
backyard trash burning, forest fires, and the combustion of diesel fuel in cars and trucks are all
viewed as sources to the atmosphere. Dioxins in organochlorine products such as the wood
preservative pentachlorophenol can be emitted to air from use of the product. For dioxin-like
PCBs, products still in use such as building caulk which contain PCBs are also thought to be a
source of air emissions. The physical mechanism of atmospheric transport and deposition is
understood to be the primary pathway for the ubiquitous distribution of PCDDs, PCDFs, and dl-
PCBs in terrestrial and aquatic environments. Plants, animals, and ultimately humans
1-2
-------
bioaccumulate these deposited dioxin-like compounds. The contamination of ecological and
terrestrial food chains arises by atmospheric deposition into photosynthesizing plants and grasses
that are eventually consumed by animals.
This paradigm points to the atmosphere as an essential transport media, ultimately
causing environmental exposures to PCDDs, PCDFs, and dl-PCBs, albeit through secondary and
indirect pathways.
To help characterize the ubiquitous presence of dioxins in the environment, the U.S.
Environmental Protection Agency (EPA) established the National Dioxin Air Monitoring
Network (NDAMN) in 1998. Preliminary results from this network have been presented at
several of the annual International Symposia on Halogenated Persistent Organic Pollutants,
commonly referred as the Dioxin Conference (Cleverly et al., 2000, 2002; Riggs et al., 2003;
Byrne et al., 2002), and were published in the peer reviewed literature (Cleverly et al., 2007).
An overview of final results from this network was presented at the Dioxin 2011 symposium
(Lorber et al., 2011). In addition to providing the final NDAMN results, this report identifies
trends associated with land-use type and season, discusses key findings from the QA program,
and reports on anomalous findings from NDAMN. This report accompanies the electronic
version of the data from NDAMN. Chapter 2 reviews the sampling procedures and describes the
various NDAMN sampling stations. Chapter 3 describes the laboratory analytical procedures,
and Chapter 4 provides an overview of the results from the network. Appendix A provides an
overview of the Excel workbook that contains the NDAMN data. Appendix B includes copies of
Quality Assurance Project Plans (QAPPs) for NDAMN that are associated with the field
implementation and laboratory analysis of samples.
1.1. REFERENCES
Byrne, C; Ferrario, J; Cleverly, DH; et al. (2002) Average method blank quantities of dioxin-like congeners and
their relationship to the detection limits: U.S. EPA's National Dioxin Air Monitoring Network (NDAMN).
Organohalogen Compd 59:411-413.
Cleverly, DH; Winters, D; Ferrario, J; et al. (2000) The National Dioxin Air Monitoring Network (NDAMN):
Results of the first year of atmospheric measurements of CDDs, CDFs, and dioxin-like PCBs in rural and
agricultural areas of the United States: June 1998-June 1999. Organohalogen Compd 45:248-251.
Cleverly, DH; Winters, D; Ferrario, J; et al. (2002) The National Dioxin Air Monitoring Network (NDAMN):
Measurement of CDDs, CDFs and coplanar PCBs at 18 rural, 8 national parks, and 2 suburban areas of the United
States: results for the year 2000. Organohalogen Compd 56:437-440.
1-3
-------
Cleverly, DH: Ferrario, J: Byrne, C; et al. (2007) A general indication of the contemporary background levels of
PCDDs, PCDFs, and coplanar PCBs in the ambient air over rural and remote areas of the United States. Environ Sci
Technol 41:1537-1544.
Lorber, M; Ferrario, J; Byrne, C; et al. (2011) The National Dioxin Air Monitoring Network of the United States.
Organo halogen Compd 73:75-78.
Puggs, KB; Cleverly, DH; Hartford, PA; et al. (2003) Anomalous results from National Dioxin Air Monitoring
Network. Organohalogen Compd 60:130-133.
U.S. EPA (Environmental Protection Agency). (2010) Recommended toxicity equivalence factors (TEFs) for human
health risk assessments of 2,3,7,8-tetrachlorodibenzo-p-dioxin and dioxin-like compounds. Risk Assessment
Forum, Washington, DC; EPA/100/R-10/005. Available online at http://www.epa.gov/raf/hhtefguidance/.
Van den Berg, M; De Jongh, J; Poiger, H; et al. (1994) The toxicokinetics and metabolism of polychlorinated
dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) and their relevance for toxicity. Crit Rev Toxicol 24:1-74.
Van den Berg, M; Birnbaum, L; Bosveld, ATC; et al. (1998) Toxic equivalency factors (TEFs) for PCBs, PCDDs,
PCDFs for humans and wildlife. Environ Health Perspect 106(12):775-792.
Van den Berg, M; Birnbaum, LS; Denison, M; et al. (2006) The 2005 World Health Organization re-evaluation of
human and mammalian toxic equivalency factors for dioxins and dioxin-like compounds. Toxicol Sci
93(2):223-241.
1-4
-------
Table 1-1. The toxic equivalency factors (TEFs) for the dioxin-like
compounds using the WHO approach
Dioxin Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
TEF
1.0
1.0
0.1
0.1
0.1
0.01
0.0003
Furan Congener
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
TEF
0.1
0.03
0.3
0.1
0.1
0.1
0.01
0.01
0.0003
Dioxin-Like
PCB
PCB77
PCB 81
PCB 126
PCB 169
PCB 105
PCB 114
PCB 118
PCB 123
PCB 156
PCB 157
PCB 167
PCB 189
TEF
0.0001
0.0003
0.1
0.03
0.00003
0.00003
0.00003
0.00003
0.00003
0.00003
0.00003
0.00003
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.
Source: Van den Berg et al. (2006).
1-5
-------
2. AMBIENT AIR SAMPLING
2.1. SAMPLING OBJECTIVE
The objective of NDAMN was to determine background air concentrations of PCDDs,
PCDFs, and dl-PCBs in rural and remote areas of the United States. "Background" is defined as
areas where there are no expected influences of nearby known dioxin sources. To meet this
objective, sampling focused on rural and remote areas, although a few stations closer to urban
areas were added. The rural NDAMN stations were chosen in order to obtain air concentrations
in areas where crops and livestock are grown. Remote stations were selected on the basis that
they were relatively free of human habitation and greater than 100 km from likely dioxin sources
(i.e., urban, suburban, industrial settings, etc.). The locations of sampling stations covered a
wide range of climate conditions from tropical subhumid to subarctic. The idea behind the
sampling configuration was to provide reasonable geographic coverage of "background areas" of
the United States, limited only by budgetary constraints.
2.2. SAMPLING DESIGN
The locations of sampling stations did not entail a purely random sampling approach. In
order to reduce the costs associated with maintaining air monitoring stations and to ensure access
and security at the stations, most NDAMN stations were co-located on pre-existing nationally
based air monitoring networks. These networks included the National Atmospheric Deposition
Program/National Trends Network (NADP/NTN), and the Interagency Monitoring of Protected
Visual Environments (IMPROVE). These networks were designed to determine the spatial and
temporal measurements of pollutants on a national scale, and to establish time trends of
environmental impacts. The stochastic basis of the design, and other information of the
NADP/NTN and IMPROVE, can be obtained online at http://nadp.sws.uiuc.edu/ and
http://vista.cira.colostate.edu/improve, respectively.
Funding was sufficient for the establishment and maintenance of 35 NDAMN stations for
a period of 6 years. Of the 35 stations, 22 were established at or near existing stations in the
NADP/NTN; one (Station 34) was located at an IMPROVE site. Station 2 was designated as the
QA co-located station. Its purpose was to provide a quality check for the nearby NDAMN
station. For several of the initial moments, Station 2 was set up adjacent to Station 1 at Penn
2-1
-------
Nursery, PA, and later, it was moved to be adjacent to Station 3 at Clinton Crops, NC. The
geographic locations of the NDAMN monitoring stations are illustrated in Figure 2-1. Table 2-1
gives the names, cities, latitude/longitude coordinates, elevation measurements, and
classifications (i.e., remote, rural, or urban) of the NDAMN stations. Table 2-2 lists the
sampling moments, dates and years of sampling, and the season during the sampling moment.
2.3. AIR SAMPLING PROCEDURES
Long-term sampling for dioxin in air was pioneered by the Connecticut Department of
Environmental Protection (CDEP) with their program in the 1990s (Hunt, 2008; Hunt and Lihzis,
2011). They began sampling for dioxin in air in 1987 using a more traditional 48-hour sampling
strategy, but then switched to a 30-day sampling approach in 1993. They reported on the details
of their method's performance at Dioxin '97, the annual international conference on dioxin and
related compounds (Maisel and Hunt, 1997), and presented an overall evaluation of the method
using data from measurements made in Connecticut in the 2000s (Hunt and Lihzis, 2011). EPA
adopted their method for NDAMN and amended it with a strategy of 5 or 6 days of continuous
sampling, followed by 1 or 2 days of down time while the quartz fiber filter (QFF) was harvested
and replaced (see details below). California then adopted this 5 days on, 2 days off, approach to
their 30-day monitoring of dioxins in their California Ambient Dioxin Air Monitoring Program
(CADAMP; http://www.arb.ca.gov/aaqm/qmosopas/dioxins/dioxins.htm). This section provides
an overview of the sampling method. Further details on field implementation can be found in
Appendix B.
2.3.1. Description of the Tisch Environmental® TE1000 PUF (PS-1 sampler)
Ambient air sampling was conducted with a Tisch Environmental TE1000 PUF (PS-1
sampler) ambient air sampler in accordance with procedures described in EPA Method TO 9A
(U.S. EPA, 1999), as revised in the QAPP for NDAMN (see Appendix B). The PS-1 sampler is
equipped with a QFF and a polyurethane foam (PUF) adsorbent plug for collecting contaminants
bound to total suspended particulates and contaminants in the gaseous phase, respectively.
Initially, glass fiber filters (GFFs) were used to collect the particulate phase. After the
November 1998 sampling moment was completed, QFFs were used instead of GFFs because of
2-2
-------
their durability and low background dioxin levels. The use of QFFs is consistent with EPA
Method TO 9A. Figure 2-2 is an illustration of the components of the PS-1 sampler.
The PS-1 sampler consists of a sampling head, a meter equipped with a magnehelic
gauge to measure air flow, and a blower-type vacuum pump. Sample air flow of 240 liters per
minute is controlled by adjusting the speed of the vacuum blower using a voltage variator. The
sampler is turned on and off using a seven-day timer, and the number of hours that the sampler
operates is recorded with an elapsed time meter. The sampling head assembly consisted of a
10.16 cm (internal diameter) QFF and a 5.85 cm (internal diameter) by 12.7 cm (length) glass
sample cartridge containing a 5.08 cm (length) PUF absorbent plug.
A regulated air flow was drawn through the top of the sampling head assembly with a
vacuum pump, and the particle-bound phase of the contaminants in the air stream was collected
on the filter surface (porosity down to 0.1 jim), while the vapor phase was absorbed into the
PUF. Approximately 6,000 to 8,000 m3 of air passed through the sampling head assembly
during a single sampling moment. Thirty-two percent of the final study samples were outside
this range, but only 7% were lower than 5,000 or more than 9,000 m3, with a low of 2,655 m3
and a high of 13,035 m3. The average volume of air was 6,827 m3, with a standard deviation of
1,074 m3. The purpose of sampling such a large volume of air was to achieve a target detection
limit (DL) of 0.1 fg/m3 for 2,3,7,8-TCDD. Each sampling event consisted of 20 to 24 days of
active sampling over a 28-day period, on a weekly schedule of 5 or 6 days of continuous
operation followed by 1 or 2 days of inactivity. The protocol required that sampling start on a
Wednesday night at midnight, and run approximately 120 hours until midnight of the following
Monday. Then on Tuesday, the QFF was harvested, packaged, and refrigerated until shipment to
the laboratory, and then a new QFF was placed in the sampler. This cycle was to run four times
for each 28-day sampling event, yielding four QFFs. Records were not available to confirm that
all samples were obtained according to this schedule. The practice of harvesting QFFs was done
to prevent saturation and clogging of the filter media with collected particles. Another benefit of
changing the QFFs was the potential to reduce volatile loss of particle-bound dioxin. The PUF
was collected once at the end of the sampling moment. Prior to sampling, the PUFs were
commercially precleaned by heating at 100°C for 16 hours, and then analytically determined to
be free of dioxin contamination. The QFFs were also precleaned to ensure they were free of
dioxin contamination. Two compounds, 13C-labeled 1,2,3,4-tetrachlorodibenzofuran (TCDF)
2O
-3
-------
and PCB 81, were added to the PUF as a QA procedure. Both compounds were selected to
represent the most volatile members of the class of analytes and were intended to gauge the
possibility of any loss of sample during the sampling period.
Prior to the start of a sampling moment, the onsite field technician performed a
multipoint calibration. In addition to initial and final multipoint calibrations, a single-point flow
check was conducted each week during sampling to ensure the accuracy of flow rates for each
sampling station. New motor brushes or a new motor in the PS-1 samplers were required at the
onset of each sampling moment in order to assure the sampler would not fail in the field.
2.3.2. Preparation of PUF Sampling Cartridge
The QFF and PUF sampling media were precleaned prior to set-up in the PS-1 sampler to
ensure that these components were free from contamination of PCDDs, PCDFs, and dl-PCBs.
All QFFs were placed in an oven and baked at 400°C for 5 hours before use. This volatilized
and destroyed any dioxin-like compounds that might have contaminated the filter medium.
PUFs were purchased and certified precleaned from a supplier. The cleanup of new PUFs
involves Soxhlet extraction with acetone for 16 hours at approximately four cycles per hour.
When PUF cartridges are reused, diethyl ether/hexane (5 to 10% volume/volume) is typically
used as the cleanup solvent. As a final step to assure the sample media were free of
contamination, at least 10% of the batch of PUFs and QFFs that were deployed into the field for
the sample moment were tested in the laboratory and certified to be dioxin-free. These steps
were in conformance with the procedures set forth in EPA Method TO 9A (U.S. EPA, 1999).
2.3.3. Multipoint Calibration of the PS-1 sampler
The PS-1 sampler was calibrated prior to the start of the sampling moment using a
calibration kit consisting of a calibration orifice and a water manometer. A post-sampling
calibration was also conducted at the completion of the sampling moment. The sampling head
contains an empty glass cartridge during the calibration process. If an empty glass cartridge is
not available, then a glass cartridge containing the sample PUF for calibration is used. The
NDAMN field operator calibrated the PS-1 sampler using the following procedures as stipulated
in EPA Method TO 9A (U.S. EPA, 1999):
2-4
-------
• Recorded ambient temperature and barometric pressure during calibration.
• Placed an empty glass cartridge in the sampling head. Installed the sampling head onto
the sampler vacuum blower inlet.
• Installed the calibration orifice on the sampling head.
• Installed a manometer from the calibration kit on the front of the air sampler housing.
Opened the shutoff valves on the top of the manometer.
• Adjusted the manometer so the "0" inch mark was on the scale.
• Connected the tubing from one of the manometer inlet ports to the side port on the
calibration orifice.
• Adjusted the sampler airflow using the voltage variator until the sampler magnehelic
gauge indicated a reading of 70 inches.
• Allowed the system to run for approximately 1 minute at this speed. Recorded the
difference in the inches of water from the manometer on the NDAMN field calibration
data form. This was achieved by reading the liquid level on each of the two sides of the
manometer and documenting them on the NDAMN field calibration data forms.
• Readjusted the voltage variator counter-clockwise until the sampler magnehelic indicated
a reading of 60 inches and then repeated the previous step documenting the manometer
readings on the NDAMN field calibration data forms. This step was then repeated for
magnehelic readings of 50, 40, 30, and 20 inches. The PS-1 sampler was turned off at the
completion of the calibration.
• Using the recorded atmospheric temperature and pressure, the operator calculated the
sampler set point using the provided electronic spreadsheet. The spreadsheet calculated
the calibration slope, intercept, and correlation coefficient. All calculations were
recorded at the bottom of the NDAMN field calibration data form. It was required that
the resulting correlation coefficient, R, of this calibration was greater than or equal to
0.98. If R was less than 0.98 (R2 less than 0.96), the calibration procedure was repeated.
The resulting magnehelic set point corresponded to an airflow of 0.24 m3/minute. This
magnehelic set point was the setting at which the PS-1 sampler was operated.
2.3.4. Sample Collection
There were four sample "moments" per year corresponding to the four seasons of the
year. Sampling occurred during a 4-week period during one of the seasons, and all NDAMN
stations were operated at the same time. For example, the "fall" sample of 1998 occurred
between November 24 and December 22, 1998, and all operating NDAMN monitors were
sampling during this time. The PS-1 sampler motor automatically switched off at the completion
of the timed sampling moment. At the end of each 4-week sampling moment, the onsite operator
recorded the flow and elapsed time, collected the QFF and PUF from the PS-1 sampler, and
2-5
-------
performed a multipoint calibration. The sampling head was disassembled, and the glass sample
cartridge was removed, wrapped with aluminum foil, and placed into a sample jar. The caps on
the PUF and filter sample jars were then replaced. All four QFF samples were wrapped in foil
and labeled as "sample" or "field blank." Bubble wrap or a similar material was used to protect
the jars from breakage during shipment. Samples were packed in a shipping container and kept
at 4°C. Samples were shipped to the EPA Environmental Chemistry Laboratory in Mississippi
for chemical analysis.
2.3.5. Field Quality Assurance (QA)/Quality Control (QC) Procedures
Four sample sets were generated for each sampling moment for each NDAMN station:
(1) one PUF filter from active sampling; (2) one PUF field blank; (3) one set of four QFFs from
active sampling; and (4) one set of four QFF field blanks for a total of 10 samples. Field blanks
were used to determine contamination affecting the active samples (which are passively exposed
only during setup and collection), so field blanks were only exposed to ambient air during
sample setup and collection. The PUF field blank remained inside the sampler housing and,
thus, underwent the same environmental conditions (e.g., temperature, pressure, etc.) as the field
samples. The PUF field blank occupied available space, inside the sampler, in a closed jar, and
was exposed to the environment only while the onsite operators were performing sampling
activities. Based on the minimum background contamination detected in most field blanks, EPA
decided that beginning with Moment 9 (November/December 1999), all field blanks did not need
to be analyzed. Therefore, approximately only half of the field blanks collected after that date
were analyzed.
Initially, trip blanks were also part of the protocol, but these were eliminated as a QC
check because analysis of trip blanks collected in Moment 1 (June/July 1998) demonstrated very
low contamination and because trip blanks are not required by EPA Method TO 9A.
At the start of the NDAMN program, two sampling stations were located at Penn
Nursery, PA: Station 1 was the formal NDAMN sampler, and Station 2 was the duplicate
sampler that was maintained for QA/QC purposes. The NDAMN sampler and the PS-1
duplicate sampler were located approximately 10 feet apart. This duplicate sampler operated at
Penn Nursery until the spring 2002 sampling moment, after which time it was moved to the
Clinton Crops, NC location. The presumption for a co-located sampler is that the results from it
2-6
-------
should be similar to the results of the study sampler since they measure essentially the same mass
of air. If the results are not the same, there may be more than one reason: the sampler possibly
drew in different/contaminated air, somehow the sampler was contaminated (either the program
or the co-located sampler), or the sampling matrices were contaminated (again either the
program or the co-located sample matrices).
2.4. REFERENCES
Hunt, GT. (2008) Atmospheric concentrations of PCDDs/PCDFs in Metropolitan Hartford Connecticut - current
levels and historical data. Chemosphere 73:8106-8113.
Hunt, GT; Lihzis, MF. (2011) PCDDs/PCDFs in ambient air (<1 fg m-3) - The CTDEP long term sampling (30d)
method. Chemosphere 85:1664-1671.
Maisel, BE; Hunt, GT. (1997) Long duration measurement of PCDDs/PCDFs in ambient air - method performance
data. Organohalogen Compd 33:128-133.
U.S. EPA (Environmental Protection Agency). (1999) Compendium of methods for the determination of toxic
organic compounds in ambient air second edition compendium method TO 9A. Determination of polychlorinated,
polybrominated and brominated/chlorinated dibenzo-p-dioxins and dibenzofurans in ambient air. Center for
Environmental Research Information, Office of Research and Development, Cincinnati, OH; EPA/625/R-96/010b.
Available online at http://www.epa.gov/ttnamtil/files/ambient/airtox/to-9arr.pdf.
2-7
-------
Table 2-1. Description of the NDAMN air monitoring stations
Station
No.
1
2
3
2
4
5
6
7
8
9
9
10
11
12
13
14
15
16
17
18
Complete Station Name"
State of Pennsylvania Dept. of
Conservation Tree Nursery
(Perm Nursery)
State of Pennsylvania Dept. of
Conservation Tree Nursery
(Perm Nursery) (duplicate
QA/QC sampler)0
Clinton Crops Research
Station (NC3 5)
Clinton Crops Research
Station (NC35) (duplicate
QA/QC sampler)0
Everglades National Park
(FLU)
Lake Dubay State Park)
(WI28)
NW Illinois Agricultural
Center (IL78)
McNay Agricultural Research
Farm, Chariton, IA (IA23)
Lake Scott State Park, KS
(KS32)
Bixby Drinking Water
Treatment Plantd
Lake Keystone State Parkd
Caddo Valley, Arkadelphia
(AR03)
Bennington County Farm
(VT01)
Jasper Farm (NY65)
USDA Agricultural Research
Center
Caldwell Farm (OH49)
Oxford Farm (OH09)
Dixon Springs Agricultural
Center (IL63)
North Florida Research &
Educational Center (FL14)
NASA Stennis Space Center
Nearest City
Potters Mill, PA
Potters Mill, PA
Clinton, NC
Clinton, NC
Florida City, FL
Dancy, WI
Monmouth, IL
Chariton, IA
Scott City, KS
Bixby, OK
Sand Springs, OK
Arkadelphia, AR
Bennington, VT
Jasper, NY
Beltsville, MD
Caldwell, OH
Oxford, OH
Dixon Springs, IL
Quincy, FL
Bay St. Louis, MS
Latitude11
(d/m/s)
40 46 30
40 46 30
350133
350133
25 23 24
44 39 52
40 56 02
40 57 47
384019
360819
36 08 27
34 10 46
42 52 34
42 06 23
390100
394734
393153
37 26 08
303253
30 22 06
Longitude1"
(d/m/s)
77 37 17
77 37 17
781639
781639
80 40 48
893908
90 43 23
93 23 30
100 55 05
96 15 48
96 16 28
93 05 55
73 09 48
77 32 09
76 56 45
81 3152
84 43 27
88 40 19
84 36 03
893701
Elevation
(m)
466
466
40
40
2
350
230
320
863
260
300
71
305
634
46
276
284
161
60
8
Classification
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Rural
Urban
Rural
Rural
Rural
Rural
Rural
2-8
-------
Table 2-1. Description of the NDAMN air monitoring stations (continued)
Station
No.
19
20
21
22
23
24
25
26
27
28
29
29
30
31
32
o o
JJ
34
35
Complete Station Name
Padre Island National
Seashore
Fond du Lac Indian
Reservation (MN05)
North Platte Agricultural
Research Center
Goodwell Agricultural
Research Station (OK29)
Big Bend National Park
(TX04)
Grand Canyon National Park
(AZ03)
Theodore Roosevelt National
Park
Craters of the Moon National
Park (ID03)
Chiricahua National
Monument (AZ98)
Rancho Seco (closed nuclear
power plant)
Hyslop Farm, OR
Agricultural Experiment
Station6 (OR97)
Marval Ranch (cattle ranch)6
Lake Ozette , Olympia
National Park
Fort Cronkhite National
Monument
EPA Ecological Research
Laboratory, Newport, OR
Craig
Denali National Park
(IMPROVE)
Yaquina Head State Park
Nearest City
Corpus Christi,
TX
Cloque, MN
North Platte, NE
Goodwell, OK
Alpine, TX
Tuba City, AZ
Medora, ND
Hailey, ID
Willcox, AZ
Herald, CA
Albany, OR
Corvallis, OR
Ozette, WA
San Francisco,
CA
Newport, OR
Craig, AK
Trapper Creek,
AK
Newport, OR
Latitude11
(d/m/s)
27 25 37
46 42 47
410333
36 35 27
291808
360335
465341
432741
320035
382036
44 38 05
4437 11
48 05 45
37 50 03
44 37 18
55 27 07
62 18 57
44 40 30
Longitude1"
(d/m/s)
97 17 55
92 30 39
100 44 47
1013703
103 1038
1121101
103 22 40
11333 17
109 23 20
1210627
123 1124
123 33 36
124 37 48
1223154
124 02 35
133 05 17
150 1842
124 03 56
Elevation
(m)
8
390
919
999
1,056
2,152
841
1,807
1,570
64
69
190
69
30
30
5
646
39
Classification
Rural
Rural
Rural
Rural
Remote
Remote
Remote
Remote
Remote
Urban
Rural
Rural
Remote
Urban
Urban
Rural
Remote
Rural
2-9
-------
Table 2-1. Description of the NDAMN air monitoring stations (continued)
a Stations that are co-located at NADP/NTN monitoring sites have the NADP/NTN identification noted in
parentheses.
bLatitude and longitude are reported as degrees, minutes, and seconds.
'Sampling Station 2 was designated as a QA station where a duplicate sampler was located . The duplicate sampler
was run at Perm Nursery, PA, for sampling Moments 1-19. The duplicate sampler was then moved to the Clinton
Crops, NC, location. The duplicate sampler operated at Clinton Crops, NC, for sampling Moments 20-29.
dSampling Station 9at the Bixby Drinking Water Treatment Plant, Bixby, OK, was shut down after sampling
Moment 17 due to technical difficulties. The sampling station was then set up and operated at Lake Keystone State
Park, Sand Springs, OK, for sampling Moments 18-29. Taken together, Station 9 operated a total of 29 sampling
moments.
Sampling Station 29 at Hyslop Farm (dairy farm), Albany, OR, was shut down after sampling Moment 25 due to
technical difficulties. The sampling station was then set up and operated at Marval Ranch, Benton County, OR, for
sampling Moments 26-29. Taken together, Station 29 operated a total of 20 sampling moments.
Table 2-2. Dates and seasons of the 29 sampling moments of NDAMN
Sampling
Moment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Dates
06/16-07/14
08/18-09/15
11/24-12/22
01/26-02/23
03/23-04/20
05/18-06/15
07/13-08/10
08/24-09/21
11/09-12/07
01/18-02/15
04/04-05/02
08/22-09/19
11/22-12/19
01/31-02/26
05/03-05/28
Year
1998
1998
1998
1999
1999
1999
1999
1999
1999
2000
2000
2000
2000
2001
2001
Season
Summer
Summer
Fall
Winter
Spring
Spring
Summer
Summer
Fall
Winter
Spring
Summer
Fall
Winter
Spring
Sampling
Moment
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Dates
08/02-08/27
11/01-11/26
02/07-03/04
05/02-05/27
08/01-08/26
10/31-11/25
02/13-03/10
05/01-05/26
07/31-08/25
11/06-12/01
02/26-03/16
05/25-06/21
08/04-08/31
11/02-11/30
Year
2001
2001
2002
2002
2002
2002
2003
2003
2003
2003
2004
2004
2004
2004
Season
Summer
Fall
Winter
Spring
Summer
Fall
Winter
Spring
Summer
Fall
Winter
Spring
Summer
Fall
2-10
-------
30
35
32
Figure 2-1. Geographic locations of the NDAMN monitoring stations.
MagneMc Gauge
0-100 in.
7-Day Timer
Source: U.S. EPA (1999).
Figure 2-2. Illustration of the components of the Tisch Environmental15
TE-1000 PUF (PS-1 sampler).
2-11
-------
3. ANALYSIS OF NDAMN SAMPLES
This chapter briefly describes the analytical procedures used to analyze the NDAMN
samples retrieved from each sampling moment. Further details are provided in the laboratory
QAPP for this sampling program (U.S. EPA, 1999).
3.1. OVERVIEW OF THE ANALYTICAL EPA METHOD
The harvested samples (PUF/QFFs) and their associated field blanks were shipped to
EPA's Environmental Chemistry Laboratory for extraction, clean-up, and analysis by high-
resolution gas chromatography/high-resolution mass spectrometry (FIRGC/FIRMS) in
accordance with a modification of EPA Method 1613 (U.S. EPA, 1994). The analytes measured
in the study are shown in Table 3-1. They include 17 dioxin and furan congeners, 12 PCBs, and
8 dioxin and furan homologue groups. All samples had measurements of seven PCBs; it is noted
that five dioxin-like PCBs were added in the summer of 2002: PCBs 81, 114, 123, 167, and 189.
The combined PUF and QFFs of the samples and field blanks were extracted with
benzene or toluene using a Soxhlet apparatus. Prior to the initiation of the extraction period, the
PUF was spiked with 100-400 pg of 13Ci2-labeled analogs, one dioxin and one PCB analog, to
monitor losses during the sampling period. The extract was collected and stirred with acidified
silica gel, followed by acid/base silica gel clean-up and alumina and carbon chromatography.
The final extract was concentrated to approximately 10:1 and fortified with 13Ci2 internal
standards of all analytes prior to FIRGC/FIRMS analysis. The chromatographic separation was
achieved on a DB-5MS capillary column, and the mass spectrometer was operated in the lock
mass drift correction mode at a resolution of 10,000 atm/z. A set of samples consisted of 12
samples including: 10 field samples and/or field blanks, one method blank, and one laboratory
control spiked sample fortified with native target analytes at twice the limit of quantitation (these
limits are provided at the end of this chapter). All reagents were prepared according to
procedures detailed in EPA Method 1613 (U.S. EPA, 1994), and the analyses and QA/QC
procedures and thresholds were consistent with those described in EPA Method 1613, with
several notable exceptions:
5-1
-------
1. Standard calibration solutions were prepared at lower concentrations. For the NDAMN,
the lowest calibration standard contained 50 fg of TCDD and 5 pg/uL of 13Ci2-labeled
surrogates; EPA Method 1613 's lowest calibration standard contains 500 fg and
100 pg/uL of surrogates. The samples in this study were fortified to deliver 5-20 pg/uL
(EPA Method 1613 delivers 100 pg/uL from the same 20 uL final volume). The lower
13Ci2 surrogate fortification level allowed for a more realistic approximation of the actual
recovery of native analytes at the subparts-per-trillion level and better approximated the
behavior of trace levels of natives during sample processing and analyses.
2. A DB-5MS column was used in place of the DB-5 specified by EPA Method 1613. The
DB-5MS has superior separation of the 2,3,7,8-TCDD from the other tetra isomers and
better resolves the 2,3,7,8-Cl-substituted dioxins and furans.
3. EPA Method 1613 specifies an AX21/Celite® mixture of graphitized carbon where the
NDAMN procedure used a mixture consisting of 0.5 g of BioSil-A® silica gel and 0.5 g
of Amoco PX-21® carbon. The eluting solvents and fractionization are also different.
The column was conditioned with 10 mL of 50/50 benzene/methylene chloride (MeCl),
10 mL toluene, and 5 mL hexane. The sample was added to the column in 0.5 mL
hexane and following two 0.5 mL rinses of the sample. Fraction 1, containing most of
the ortho PCBs, was eluted with 4.5 mL of 25/75 MeCl/hexane. Fraction 2, collected in
one vessel, consisted of 5.5 mL of MeCl and contained the mono-ortho PCBs and 11.5
mL of benzene/MeCl, which contained the nonortho PCBs. The column was then
reversed, and the dioxins and furans were collected with 13 mL of toluene. Fractions
were reduced to less than 10 uL, and solvent was exchanged with hexane and stored in
the freezer until analyzed. All analyses were performed on either a Kratos Concept® or a
Micromass Autospec® high resolution mass spectrometer using isotope dilution. The
FIRMS was operated in the electron impact ionization mode using selected ion
monitoring. Chromatographic separations were achieved using a Hewlett Packard 6890
Series II high-resolution gas chromatograph, utilizing a 60 m x 0.32 mm (0.25 um film
thicknesses) DB-5MS capillary column. The gas chromatography conditions were
optimized to completely separate the various 2,3,7,8-Cl-substituted dioxins/furans: initial
oven temperature, 130°C; injector temperature, 270°C; interface temperature, 275°C;
temperature programming, time 1, 1.0 minute, rate 1, 5°C/minute, time 2, 15.0 minute,
rate 2, 6°C/minute; temperature 3, 295°C; injector, splitless, 1.0 minute; split flow,
30-40 mL/minute; purge flow, 1-2 mL/minute; and temperature equilibration time,
2 minute. A combination of 23 psi (constant pressure) and 1.5 mL/min (constant flow)
were used throughout the project resulting in similar chromatography and retention times
on the same column. The mass spectrometer was tuned and calibrated prior to all
analyses. It was tuned to a minimum resolution of 10,000 ppm (10% valley) using
m/z = 330.9792 (or any suitable reference peak) at full accelerating voltage of 8,000 V.
Pertinent mass spectroscopy parameters were as follows: cycle time for each congener
group, -1.0 s; electrostatic analyzer sweep (analytes), 10 ppm; native ion dwell, -100-
200 ms; 13C-labeled ion dwell, -30-66 ms; lock mass sweep, 200 ppm; lock mass dwell,
50 ms; ionization voltage, -35 eV; source temperature, 250°C; accelerating voltage,
8,000 V; and trap current, 500-700 uA.
5-2
-------
3.2. QUALITY CONTROL (QC) OF LAB SAMPLES AND CALIBRATION
Between four and six calibration standards with native analyte concentrations bracketing
the expected analyte concentrations were analyzed prior to analyzing samples. The analyses of
calibration standards permitted the response factors (RFs) to be determined as a function of
concentration using linear regression. The RF for each native analyte at each concentration was
calculated relative to its 13C-labeled analog. The relative standard deviation (RSD) for the
average RF for each of the native analytes had to be <20%. Similarly, the RF for each 13Ci2
recovery surrogate relative to the appropriate internal standard was also calculated. The RSD for
the average RF for each labeled surrogate had to be <35%. The calibration curves were
considered linear under these conditions, and the analytical system was considered calibrated
when these conditions had been satisfied. If these conditions could not be satisfied, corrective
actions were taken. The average RFs were used for subsequent quantitations. Prior to sample
analysis, the linearity of the calibration curve was verified by analyzing calibration solution 2
(200 fg of TCDD) and calculating the RF as described previously. The percentage difference
between the new RF and the average had to be <20% for the native analytes and <35% for the
13Ci2 recovery surrogates.
The chromatogram was also examined to ensure that all 2,3,7,8-Cl-substituted congeners
were clearly separated. If the signal-to-noise (S/N) ratio values were >10, the ion abundance
ratios were +15% of the theoretical (this ratio was relaxed to + 25% if the mass quantified was
less than 100 fg), and the RF and isomer separations were within specified limits, then sample
analyses proceeded. Corrective actions were initiated if specified control limits were exceeded.
These corrective actions included returning of the mass spectrometer and/or new calibration
standards prepared and re-analyzed. On the days that samples were analyzed, 10 uL of the
internal standard solution (20 pg/uL) was added to each sample, and the final volume was
adjusted to 20 uL.
Once all QA/QC parameters had been verified to be within specified limits, sample
analyses proceeded. The mass spectrometer was operated in a mass drift correction mode using
perfluorokerosene to provide lock masses. The selected ion current profile areas for the
characteristic ions for each native and labeled analyte were measured. Native analyte
concentrations were determined by isotope dilution. Peak areas from the characteristic ions for
each native analyte and its 13C-labeled analog were used in conjunction with RFs from the
-------
internal calibration data to determine concentrations directly. Labeled surrogate concentrations
(expressed as percentage recovery) were similarly calculated using an internal standard EPA
method. Samples were organized and analyzed in sets: method blank, laboratory control spike,
and the ten field samples. Peak identification criteria were as follows: S/N > 3.5; the isotope
ratio of the two characteristic ions for each congener class within 15% of the theoretical value;
the peak maxima for the molecular cluster ions coincide within 2 seconds; and native analytes
elute within ± 3 seconds of their corresponding 13Ci2-labeled analogs. Method blanks were
examined for the presence of interfering background. For furans, an ion for the appropriate
chlorinated diphenyl ether was monitored, and the ion chromatogram was examined to ensure the
absence of chlorinated diphenyl ether contamination. The amount of any native analyte detected
was listed on the quantitation report, along with the recovery of its labeled analog. Recoveries of
13C-labeled analogs for the samples were between 25 and 150%. Sample sets were reviewed by
the QA/QC officer to ensure compliance with QA/QC guidelines/criteria.
The analytical DLs ranged from 0.5 pg for TCDD/TCDF to 20 pg for octochlorodibenzo-
/7-dioxin/octachlorodibenzofuran (OCDD/OCDF), and from 1 pg (PCB 169) to 500 pg (PCB
118) for the individual PCBs. Analyte-specific DLs are shown in Table 3-1. DLs expressed on a
concentration basis are a function of the volume of the sample, and thus varied by sample. With
a sample volume average of 6,827 m3, DLs were less than 1 fg/m3 for all PCDDs/PCDFs except
OCDD at 3 fg/m3, and were generally higher for PCBs than PCDDs/PCDFs, with a high of about
70 fg/m3 for PCB 118.
3.3. REFERENCES
U.S. EPA (Environmental Protection Agency). (1994) Method 1613. Tetra- through octa-chlorinated dioxins and
furans by isotope dilution HRGC/HRMS. Office of Water Engineering and Analysis Division, Washington, DC.
Available online at
http://water.epa.gov/scitech/methods/cwa/organics/dioxins/upload/2007 07 10 methods method dioxins 1613.pdf
U.S. EPA (Environmental Protection Agency). (1999) Compendium of methods for the determination of toxic
organic compounds in ambient air second edition compendium method TO 9A. Determination of polychlorinated,
polybrominated and brominated/chlorinated dibenzo-p-dioxins and dibenzofurans in ambient air. Center for
Environmental Research Information, Office of Research and Development, Cincinnati, OH; EPA/625/R-96/010b.
Available online at http://www.epa.gov/ttnamtil/files/ambient/airtox/to-9arr.pdf.
5-4
-------
Table 3-1. NDAMN analytes measured and their detection limits (DL, in pg)
PCDD
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
Total TCDD
Total PeCDD
Total HxCDD
Total HpCDD
DL,pg
0.5
1.5
2.5
2.5
2.5
2.5
20.0
0.5
1.5
2.5
2.5
PCDF
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Total TCDF
Total PeCDF
Total HxCDF
Total HpCDF
DL,pg
0.5
1.5
1.5
1.5
1.5
1.5
1.5
2.5
2.5
4.0
0.5
1.5
1.5
2.5
PCB (IUPAC #)
3,3',4,4'-TCB (PCB 77)
3,4,4,5-TCB (PCB 81)
3,3',4,4',5-PeCB (PCB 126)
3,3',4,4',5,5'-HxCB (PCB 169)
2,3,3',4,4'-PeCB (PCB 105)
2,3,4,4',5-PeCB (PCB 114)
2,3',4,4',5-PeCB (PCB 118)
2',3,4,4',5-PeCB (PCB 123)
2,3,3',4,4',5-HxCB (PCB 156)
2,3,3',4,4',5'-HxCB (PCB 157)
2,3',4,4',5,5'-HxCB (PCB 167)
2,3,3',4,4',5,5'-HpCB (PCB 189)
DL,pg
20
2
2
1
300
20
500
10
80
20
10
2
Notes: PCDD = polychlorinated dibenzo-p-dioxin; PCDF = polychlorinated dibenzofuran; PCB = polychlorinated
biphenyl; TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.
5-5
-------
4. OVERVIEW OF RESULTS
This section provides a summary of results of measuring atmospheric PCDDs, PCDFs,
and dl-PCBs from June 1998 through November 2004 at 35 NDAMN stations throughout the
United States. Results are presented on a congener-specific basis and also for TEQ
concentrations. The intent of this chapter is to present an overview of the results, rather than a
comprehensive interpretive analysis. Researchers are encouraged to apply their own statistical
models to the data provided in the NDAMN database that is released along with this report for
more in-depth analysis.
Section 4.1 describes the breadth of the final data set. Section 4.2 provides a summary of
the QA results for the study. Section 4.3 provides a summary of the results, with a look at some
basic trends, including concentrations found, trends over time and land use, and some of the high
concentrations found.
4.1. OVERVIEW OF NETWORK OPERATION STATUS
On June 16, 1998, 10 NDAMN stations became operational in the field at 9 rural
locations, with Station 2 acting as a duplicate QA/QC sampler operating adjacent to Station 1.
NDAMN was expanded to 23 sampling stations by the last sample moment in 1999, 30 stations
at the end of 2000, 32 stations at the end of 2001, 34 stations at the end of 2002, and finally, 35
stations at the end of 2003 when the last one was added for the first sampling moment in 2003.
The count of 35 stations includes QA/QC Station 2. Because this QA/QC station did not obtain
data for the program but rather was a duplicate sampler serving as a quality measure, the full
NDAMN data set is considered to be composed of 34 stations. A comparison of the results from
this duplicate sampler and the regular NDAMN samplers is provided in Section 4.2.
Table 4-1 shows when each of the study stations were operational and when data were
"not available," NA. If all 34 study stations operated for all moments following their initially
collected sample, there would be a total of 736 samples. However, there were only
685 completed samples, and the remaining 51 samples were NA for the following reasons:
4-1
-------
1. Station Not Operating: This was the most common cause, and it occurred 26 times.
2. QA Failure at the Lab, All Analytes: The station was operating and a volume in the
sampler was recorded, but the laboratory recorded a QA Failure and did not develop any
data for the sample. This occurred eight times, and seven of those occurrences were for
Moment 28 samples, for Stations 12 and 14 through 19.
3. QA Failure at the Lab, PCDD/PCDFs: There were occasions where the laboratory
reported QA failure for measurements of all dioxins and furans but did report
measurements for the dioxin-like PCBs (or most dioxins and furans; one sample had
reported some but not all dioxins and furans). Rather than count the PCBs in any of the
results compilation for this chapter, none of the data from these samplers was used for the
results generated in this chapter. The PCB data are available on the spreadsheet
accompanying this report, if others wish to analyze it. This occurred seven times.
4. Volume Lost: The station was operating, and the laboratory reported measurements of
analyte mass. However, the volume in the sampler was missing, so concentrations could
not be developed. This occurred two times.
5. Low Volume: The station was operating, and the laboratory reported measurements of
analyte mass. However, the volume was less than 2,000 m3. According to the protocol
for sampling (see Appendix B), a sample with less than this volume was to be rejected.
While this rejection did not occur during the NDAMN program, these samples will not be
included in the generation of results for this chapter. The results are reported in the raw
data file, however, should others wish to analyze them. This occurred eight times in the
NDAMN sample and once for the QA sampler.
The spreadsheet with final NDAMN results identified the cause for each of the NA
samples. It is noted that the following dioxin-like PCBs were only first measured in the summer
of 2002 (Moment 20): PCBs 81, 114, 123, 167, and 189. While the total "»" for all other analyte
measurements was 685, the number of measurements per analyte for these congeners was 317.
4.2. QUALITY ASSURANCE (QA) MEASURES AND RESULTS
4.2.1. Blank Samples
There were three types of blanks employed in NDAMN: trip blanks, field blanks, and
laboratory blanks. Trip blanks were found to be minimally contaminated during the first
moment of sampling and were discontinued after that. Another reason trip blanks were
discontinued was that they are not specified as part of Method TO 9A. Field blanks were only
exposed to ambient air during sample setup and harvesting. Based on minimum background
contamination found in field blanks, EPA decided that beginning with Moment 9
4-2
-------
(November/December 1999), all field blanks did not need to be analyzed. From that point
forward, at least one field blank per station per year was analyzed. Method blanks were used
during the entire program consistently—one method blank per nine NDAMN samples. Their
purpose, of course, was to determine if there was contamination of laboratory equipment or
reagents.
Subsets of field and laboratory method blank results were used for analysis in this
section. All field blanks through Moment 3 were collected, along with all field blanks for
Moments 5, 7, and 9. This subset entailed 56 samples and was judged to be sufficient to
characterize the NDAMN field blank results. All laboratory method blank results through
Moment 13 were used in this analysis, along with all method blanks every other moment
thereafter (Moments 15, 17, and 19 through 29). By this selection, there were 108 method blank
results in the analysis, and this was judged as sufficient to characterize the overall results of
method blanks.
Results comparing the average mass of congener found in the blanks with average mass
of congener found in NDAMN study samples are shown in Table 4-2. All average masses were
derived with non-detects set equal to 0.0. For NDAMN, average congener masses were
presented for the three types of stations—remote, rural, and urban. This delineation was
described in Chapter 2. Six stations were characterized as "remote" with 153 samples, 4 stations
characterized as "urban" with 69 samples, and the remaining 24 stations of NDAMN were
characterized as "rural" with 463 samples. Along with congener-specific average masses found
in the blank samples (again counting non-detects as 0.0), Table 4-2 shows the program-specific
"target" method DLs and the percent positive quantified congener in the blank samples. These
targets were determined prior to the beginning of NDAMN. If a sample had less than this
amount, in a sample or in a blank, generally speaking the result was characterized as ND. An
examination of the blank results did show, however, that some blank samples had congener-
specific measurements above the target levels, as well as some below the target levels, and then
some were simply described as ND. In other words, these "targets" were not very rigidly
observed, as the laboratory did quantify some congener masses from blank samples that were
lower than these targets.
-------
Several observations are made based on the results from Table 4-2:
1. The NDAMN categorization of remote, rural, and urban did appear to capture some
differences in the levels of dioxins, furans, and PCB congeners in the air, particularly in
that masses in the remote samplers were much lower than the other two categories. This
is discussed later in Section 4.3.
2. The average mass of all tetra- through hexa dioxin and furan congeners in the blank
samples, both the field and laboratory blanks, were significantly lower than the target
levels, and similarly significantly lower than measured even in the remote samplers. The
fact that it is lower than the target is because most samples were characterized as NDs,
and NDs were counted as zero in the calculation of averages. Specifically, the percents
of positive quantifications in the blanks were under 10% most of the time. The one
exception to this generalization was 2,3,4,6,7,8-hexachlorodibenzofuran (HxCDF), found
in 40% of the field blanks (and in 84% of the study samples). Also, the amount of this
congener in the blanks was closer to the amount in the remote samples; the average of 1.1
pg in the laboratory blank was about one-fifth of the average amount found in the remote
sampler, at about 5 pg. Other than this hexa furan congener, the average amounts of all
other congeners found in the blank samples were an order of magnitude or more lower
than found in the remote samples, and two or more orders of magnitude lower than in the
rural or urban samplers.
3. The hepta- and octa dioxin and furan congeners were quantified more frequently in blank
samples, with four of five such congeners found in 88% or more of the laboratory blanks
samples. But, like the tetra- through hexa congeners, the average mass found in the field
and laboratory blank samples was about an order of magnitude lower than the amount
found, on average, in the remote samplers, and about two orders of magnitude lower than
the amounts found in the rural and urban samplers.
4. Other than PCB 169, all PCB congeners were found in blank samples at high
percentages, most above 90%, and for five of the seven congeners, it was found at either
98 or 99% of the blank samples. The average masses found in the blanks were close to
expectations based on the target levels set, except for PCB 118. For that congener, the
average amount found in the field blanks was about three times the target, and the amount
in the method blanks was about twice the target.
5. For all PCBs, the average masses found in the field blanks were less than, but close to,
the masses found in the remote samples, approximately lower by a factor of 2 (half as
much in the blank as the remote sampler) to lower by a factor of 5. Comparing to the
rural averages, the amount found in the blanks ranged from about a factor of 3 less to a
factor of 10 less.
It can be concluded that PCBs are ubiquitous in the environment, as well as in the
laboratory, as seen by their presence in the blank samples. The same might be said of the higher
chlorinated dioxins and furans, however, not the lower chlorinated dioxins and furans, whose
4-4
-------
presence was quantified in blank samples mostly less than 10% of the time. Even though found
in blanks, the masses of higher chlorinated dioxins and furans are still well below levels found in
the NDAMN samples: near an order of magnitude lower than samples in the remote areas and
two orders of magnitude lower than levels found in the rural or urban samplers. The same
cannot be said for PCBs. As noted above, levels of PCBs in the blanks were within factors of
two and five of levels in the remote area and within a factor of 10 and less for levels in the rural
areas. PCBs have been identified as a laboratory contaminant in environmental measurement
studies (Ferrario et al., 1997). The fact that they were present was not unexpected in NDAMN,
as evidenced by the DL targets established before the program began. But it is also clear that
PCB results may be problematic for proper interpretation of NDAMN results.
One way that similar issues are dealt with is through subtraction of chemicals found in
blanks from the amount found in study samples. This is termed, blank subtraction. However,
there was no blank subtractions in the NDAMN samples, for either dioxins/furans or the PCBs.
Blank subtraction is often a judgment call and not a required protocol, particularly for dioxin-like
compounds. The exact protocol for blank subtraction is not established for dioxin-like
compounds. One approach is to subtract the method blank concentration values within a single
batch of samples. Another is to make corrections post-survey based on overall method blank
statistics, such as subtracting the mean amount found in all study method blanks from all study
samples. Subtracting PCB levels in blanks with either of these methods would have resulted in a
large number of final concentrations being reported at values less than the stated DL, which is
problematic for obvious reasons. Ideally, measured concentrations would be significantly larger
than concentrations found in method blanks, and that appears to be the case with dioxin and
furan congeners, but not with PCB congeners.
It might be observed, however, that the most toxic PCB congeners, PCB 126 and 169,
appeared to be measured with the least laboratory contamination issues. The target DLs of 2 and
1 pg of mass for PCB 126 and 169, respectively, were achieved in both the field and method
blanks. The average masses in remote samples were relatively close to these target DLs, with
averages of 8 and 0 pg of mass for PCB 126 and 169, respectively. The rural and urban
concentrations were over an order of magnitude higher, at 32 and 68 pg for PCB 126, and though
not an order of magnitude higher, still much higher for PCB 169 at 5 and 6 pg. As discussed in
Section 4.3, PCB 126 drives the PCB portion of the TEQ concentration, with PCB 169 the
4-5
-------
second most toxic congener, even though it did not influence the TEQ as did PCB 126. From
these perspectives, it is concluded that the PCB 126 and 169 measurements were reliable and
useful.
4.2.2. Co-Located Sampler Results
Station 2 was designated as a QA station where measurements were taken concurrently in
order to provide a duplicate field test of the sampling and analytical EPA methods. The
sampling apparatus and procedures were exactly the same for this station as all regular NDAMN
stations, and the operators were instructed to process the samples in the same manner. For the
first 19 moments, this station was located adjacent to the Station 1 monitor at Penn Nursery, and
for moments 20-29 (not including Moment 28, when the station was not operating) the
monitoring equipment was moved to North Carolina at Clinton Crops, Station 3. Ideally, results
from the side-by-side monitors should be similar. Table 4-3 provides congener averages,
correlation coefficients, and relative percent differences (RPDs), in the side-by-side
measurements. These results do not include a side-by-side measurement for moment 10, the
winter sample of 1999-2000, when the volume on the QA sampler was 1,198 m2. The
implications of this low sample volume on the QA sampler are discussed later in this section.
The RPD is a common method to characterize precision in co-located samples and has been used
to evaluate co-located dioxin samplings in the long-term air sampling program in Connecticut
(described previously in Chapter 2; Hunt, 2008). The RPD for each pair of (congener NDAMN
sample, congener QA sample) is calculated as:
RPD = [ABS (CNDAMN - CQA)/AVG (CNDAMN, CQA)] * 100%
where:
RPD = relative percent difference, %
CNDAMN = concentration of congener in an NDAMN sample
CQA = concentration of corresponding congener in Q A sample
ABS = absolute value function
AVG = average of the two concentrations
4-6
-------
Figures 4-1 and 4-2 graph the concentrations of specific congeners found in the side-by-side
monitors to display the comparison between the NDAMN and the co-located samples.
For the most part, there was a high correlation between NDAMN station and QA station
results for dioxins and furans (see Table 4-3 and Figure 4-la). Correlation coefficients for the
PCDDs/PCDFs were mostly in the 0.80 to 0.99 range, although a few lower correlations, even at
negative values, occurred for the tetra- and penta dioxin congeners at the Clinton Crops station.
The RPDs listed in Table 4-3 were the average of each congener pair; for example the RPD of
17% for 2,3,7,8-TCDD at Penn Nursery was the average of 18 individual RPDs for that station.
These average RPDs were approximately 20% or less for all PCDD/PCDF congeners at both
Penn Nursery and at Clinton Crops, except for OCDD which was above 30% at Penn Nursery.
The congener-specific average RPD varied between 9% and 36% for the PCDDs and PCDFs,
with an overall average of 18.9%. This is a bit higher than the 12.5% overall found by Hunt
(2008) in co-located air samplings of dioxins. That program in Connecticut involved 24-day
sampling, similar to NDAMN, and though the precision they achieved was superior to that found
here, NDAMN is nonetheless judged to show adequate precision for the PCDDs/PCDFs in co-
located samples. The Penn Nursery found consistently higher concentrations of dioxins as
compared to Clinton Crops, and this was seen in the average concentrations in the two sets of
results from Penn Nursery compared to the two sets of results from Clinton Crops. Also,
interestingly, the reverse was mostly true for furans; they were uniformly higher in the Clinton
Crops station, except for OCDF which appeared similar at the two locations. Figure 4-la shows
the high correlation in 2,3,7,8-TCDD concentrations and the nearness to the perfect correlation
line of_y = x, with, importantly, a uniform finding of a high concentration of about 3 fg/m3 found
in both the NDAMN and QA sampler. Figure 4-lb shows an example where there was a low
correlation in the NDAMN and QA stations for a congener, also 2,3,7,8-TCDD. It is seen that
most of the samples were similar, but there were two instances where a disparity was seen in
2,3,7,8-TCDD measurements. In one case, the NDAMN sampler showed a higher concentration,
and in the second case, the QA sampler had the higher concentrations. These two results explain
the low correlation coefficient of- 0.10. Overall, the average concentrations and the positive,
mostly high, correlations speak to the ability of the protocol to obtain excellent duplication of
results for dioxins and furans.
4-7
-------
The same cannot be said about dioxin-like PCBs. There was a trend that the QA sampler
consistently had higher results as compared to the NDAMN sampler (within but sometimes even
higher than a factor of 2). The correlation coefficients were often less than 0.50 and the RPDs
were also consistently higher for the PCBs as compared to the PCDDs/PCDFs, often in the 40-
50% range. There did not seem to be any apparent trends in overall PCB concentrations between
geographic locations, as there was for dioxins and furans (as discussed above). This is seen in
Table 4-3, where similar concentrations were found in both locations from an examination of just
the NDAMN sampler. Figure 4-2 shows the common trends found in the comparison between
the NDAMN and QA sampler. In Figure 4-2a, most of the QA results were higher than the
NDAMN results, with some exceeding NDAMN results by a factor of 10. In other instances,
there appeared to be a higher correlation, but still a trend of higher concentrations was found in
the QA sampler. This is seen in Figure 4-2b, where again, the majority of samples were higher
in the QA sampler, but at least the highest concentrations found in the QA sampler were matched
with the highest samples found in the NDAMN sampler. Although a high correlation coefficient,
it is clear that even for this case, a different set of data was obtained from the QA sampler. It is
not known why there is this difference. The same equipment was used at the two stations—the
QA monitoring equipment was transferred from Station 1 to Station 3 starting at Moment 20.
This might suggest some internal PCB contamination of the QA sampler. In any case, the reason
for this discrepancy was never identified.
4.2.3. NDAMN Sample Volume
The target volume for NDAMN, based on the protocol, was 6,000 to 8,000 m3;
unfortunately, that target was not always met. Specifically, 32% of the final study samples were
outside of this range, but only 7% were lower than 5,000 or more than 9,000 m3, with a low of
2,655 m3 and a high of 13,035 m3. The average volume of air was 6,827 m3, with a standard
deviation of 1,074 m3. As noted earlier, there were eight study samples with volumes less than
2,000 m3, and these were not included in the final study results (although the raw data of
NDAMN for these samples with low volumes are available on the spreadsheet and can be
evaluated by others).
The issue of sample volume was evaluated with the co-located samplers. Ideally, the
volumes in the NDAMN and co-located samples should be very similar, event-to-event,
4-8
-------
however, this did not occur in all cases. Figure 4-3 compares the volume in co-located samples
with the NDAMN samples for the 19 events when the co-located sampler was located at Penn
Nursery in Pennsylvania. One of the co-located samples had a very low volume of 1,198 m3, so
it would have been rejected from the NDAMN data. Not counting the pair with this low volume,
the average of the other 18 samples was 6,530 m3 for NDAMN samples and 7,170 m3 for the co-
located sampler. The t-test for the 18 paired values had ap-va\ue of 0.13, indicating the means
were not significantly different. From Figure 4-3, it is seen that 11 of the 18 pairs were very
close to the perfect correlation line, y = x, while 7 were somewhat outside of this range. The
causes for the differences in volume are not known. Detailed records from the stations, which
would have included instrument readings and perhaps other information to explain the
differences, were not available for this report. One would suspect or presume that the individual
cooperating with EPA at this research nursery harvested the GFF and PUF from both the
NDAMN and the QA sampler at the same time. Differences therefore, could only be explained
by different air flow rates between the samplers.
In any case, these differences provided an opportunity to look at the influence of sample
volume on sample results. If, in a pair of these samples, the one with lower volume had
proportionally lower mass measured, and hence the same concentration, then perhaps volume
perturbations would not be a cause to invalidate NDAMN measurements. Two sample pairs
were culled from this group of 19 for further analysis. One of them included the co-located
sample with a volume under 2,000 m3 and the second had a co-located volume that was about
2,000 m3 less than the study sample. Table 4-4 compares the concentrations between these two
pairs, the RPDs between these two pairs, and then the average concentrations and average RPDs
for all other 17 co-located samples at the Penn Nursery. Looking at the raw data for mass of
congeners (not shown on Table 4-4), it is seen that lower masses are found in the co-located QA
sampler as compared to the study sample. This is expected—lower masses should be associated
with lower volumes. However, the question is, are these lower masses proportionally lower
considering the difference in volume? For the Spring 2000 sample, the NDAMN sample had a
volume of 4,859 m3, and the co-located sampler had about one-quarter as much volume at 1,198
m3. For the concentrations to have been similar in this sampler, the masses would similarly need
to be about one-fourth as much. In fact, they were not—they were more than one-fourth as
much, and subsequently, the concentrations were higher for the co-located sampler. They are
4-9
-------
about 30% higher for the lower chlorinated dioxins, and then about twice as high or more for the
higher chlorinated dioxins, the furans, and the PCBs. The average RPD for all congeners within
this sample was 66%. For the second sample pair shown in Table 4-4, the co-located sample
concentrations were much more similar to the study samples, with an average RPD of 20%,
which is similar to the RPD average for all other Penn Nursery samples. Some of the co-located
sample concentrations were higher and some lower than the corresponding NDAMN sample.
This analysis suggests two conclusions: (1) that there quite possibly was a QA problem,
in general, with samples that had low volume and their removal from the overall NDAMN
sample was justified, and (2) samplers with lower (or possibly higher due to extended periods)
volumes than the target volume would still accurately be characterizing the air concentration
during the period of time they were operating.
On average, the volume of air collected in NDAMN samples was consistent with a
second major use of this long-term sampling method for dioxin and furans. The CDEP
monitoring program consists of 30-day sampling events (Hunt and Lihzis, 2011). Sample
volumes during the 10-year period of 2002-2012 averaged 7,659 m3. Sample collection flows
ranged narrowly from 165 to 182 liters per minute, with an average flow of 177 liters per minute
(telephone conversation between Gary Hunt and Matthew Lorber, March 2013). These volumes
are lower than suggested by EPA TO 9A, which specifies a sample air volume range of 325 to
400 m3 per 24-hour sampling period. This equates to a sample collection flow ranging from 225
to 275 liters per minute but the latter assumes a 24-hour sampling event. Lower flows are to be
expected for a 30-day sampling period (telephone conversation between Gary Hunt and Matthew
Lorber, March 2013). While the average 30-day volume in the CDEP program at 7,659 m3 was
higher than the NDAMN average of 6,827 m3, a difference is to be expected as the monitor in the
CDEP program ran for 30 continuous days, while in NDAMN, the monitor was shut down for 1
to 2 days per week per every 28-day sampling period. The reason that one-third of all NDAMN
samples were outside the target range of 6,000 to 8,000 m3 (although only 7% were outside the
5,000 to 9,000 m3 range) remains unknown. Chapter 2 discussed the implementation of
NDAMN, where monitors were located near existing networks, such that EPA could rely on
operators to maintain and implement the NDAMN protocol for sampling. It seems possible that
these cooperators could have deviated from the protocol, harvested the GFFs earlier or later than
the 5- to 6-day protocol, to cause either higher or lower air volumes. It is also possible that log
4-10
-------
sheets provided to EPA used to calculate final volumes contained missing or incorrect values, on
occasion. These possibilities could not be evaluated. However, the analysis above with the co-
located sampler showing similarity in concentrations despite a 2,000 m3 volume difference
suggested that volumes were generally valid. A scan of other stations in NDAMN shows
differences in sample volumes, but similarities in concentration, supporting this finding.
4.2.4. A Comment on NDAMN Quality Assurance (QA)
Sections above addressed results from sample blanks (field and method), results from the
co-located samples, and sample volumes. Quality issues were identified, but any judgment
regarding overall validity of the NDAMN results should consider the context of the data and the
intended uses. As a research project without a regulatory mandate or intended regulatory
purposes, funding was limited. This resulted in some restrictions and had design implications.
The decision was made to only characterize background air concentrations, and not venture into
urban centers to any extent. Samplers were added to the program as funding permitted, and the
program closed perhaps earlier than ideal. Cooperators were sought to implement EPA's
sampling protocol rather than have the samplers consistently manned under a single contract.
Issues were found with PCB measurements, both in terms of laboratory performance, and in the
co-located samplers. In comparison, the CDEP program, which was operated by a single
contractor over time, had sample volumes that were within a more desired narrow range. By
contrast, NDAMN appeared to have a wider range in sample volumes than desired.
Still, measurements of dioxin and furan congeners appeared to have met QA expectations
with regard to blank results and co-located comparisons. PCB results should be used cautiously,
but it does appear that PCBs 126 and 169, the most toxic of the PCB congeners, appear to have
been measured within target DLs and with a minimum of external contamination. For its
intended purpose, which was to establish background levels of dioxin concentrations for research
purposes, it is concluded that the NDAMN program performed adequately from a QA
perspective.
4.3. OVERVIEW OF FINAL RESULTS FROM NDAMN
Table 4-5 provides congener- and homologue-specific survey-wide statistics. Several
observations are made from this table:
4-11
-------
1. Frequency of occurrence was very high, mostly above 95% and at 85% for
2,3,7,8-TCDD. The lowest detection frequency was 74% for 1,2,3,7,8,9-HxCDF. The
protocol to obtain 4 weeks of air volume guaranteed low DLs and a high detection
frequency. All results in this chapter have been generated assuming ND = 0, but a quick
check on a few averages showed virtually no change if instead assuming ND = /^ DL.
Sample-specific, congener-specific DLs can be generated by using the cogener-specific
mass DLs provided in Table 3-1 with the volume in the sampler. For example, the
congener-specific mass DL for 2,3,7,8-TCDD was 0.5 pg, and assuming a volume was
7,000 m3 (as noted in Chapter 2, volumes typically range between 6,000 and 8,000 m3),
the concentration-based DL would be 0.07 fg/m3.
2. The archetype dioxin and furan background air congener profile was seen in the survey
averages and in most individual samples. Discussions below show that some of the
higher samples did not follow this pattern. This archetype profile is characterized by low
and similar concentrations for tetra through hexa dioxins and furan congeners, with
elevations in 1,2,3,4,6,7,8-HpCDD, OCDD, 1,2,3,4,6,7,8-HpCDF, and OCDF. Lorber
et al. (1998) discuss this profile and show it to be present in urban as well as rural
settings. A similar archetype profile of dioxin-like PCBs in air has not been elucidated in
the literature, but the values found here could serve that purpose. The highest
concentrations were found for PCBs 118 and 105, with concentrations in the hundreds to
thousands of fg/m3; the lowest mean concentrations, at less than 10 fg/m3, were seen for
PCBs 126, 169, and 189. PCB 126 is the most lexicologically significant of the
dioxin-like PCBs, with a TEF at 0.1. It was detected 100% of the time at an average
concentration of 6.9 fg/m3.
3. The overall average TEQ was 11.1 fg/m3 with dioxin-like PCBs contributing only
0.8 fg/m3 (7%) of this total and with PCB 126 explaining most of this contribution (about
88%). The top six contributors explained 67% of the TEQ, and their percentage
contributions to this TEQ were 1,2,3,7,8-PeCDD at 27.8%, 2,3,4,7,8-PeCDF at 11.4%,
1,2,3,4,6,7,8-HpCDD at 9.1%, 1,2,3,6,7,8-HxCDD at 6.5%, 1,2,3,7,8,9-HxCDD at 6.4%,
and PCB 126 at 6.1%. The congener, 2,3,7,8-TCDD, contributed 5% to the TEQ. All
dioxin-like PCBs excluding PCB 126 contributed about 1% to the TEQ.
Figure 4-4 shows the average TEQ concentrations for all NDAMN stations. Average
TEQ concentrations throughout most of the Eastern Seaboard, into the central part of the United
States, range between 5 and 20 fg/m3. In the central part of the United States and into the
western portion, as well as Alaska, excluding California, the average TEQ concentration appears
to be near or less than 5 fg/m3. Two of the stations in the Western Seaboard, one in California
and one in Oregon, showed average TEQ concentrations just above 20 fg/m3. The other four
stations on the Western Seaboard showed average TEQ concentrations less than 10 fg/m3.
Station 20, Fond du Lac Indian Reservation in Minnesota, showed the highest average
concentration at 47 fg TEQ/m3, but this was skewed by a single outlier at 847 fg TEQ/m3.
4-12
-------
Without this concentration, the average for the station was 6.9 fg TEQ/m3. Station 28, Rancho
Seco (closed nuclear power plant), was also influenced by a single high concentration, although
not as much. The station average concentration of 36 fg TEQ/m3 was reduced to 21 fg TEQ/m3
if the high concentration of 241 fg TEQ/m3 is not considered in calculating the average. Two
other stations had a single measured concentration above 100 fg TEQ/m3, and these
measurements will be discussed shortly.
As discussed in Chapter 2, stations were generally categorized as either urban
(4 stations), rural (23 stations), or remote (7 stations). These characterizations were for purposes
of this study and should not be considered representations of any of these three land-use
categorizations, particularly for urban. For example, Station 13 in Beltsville, Maryland, is
considered "urban" because it is near the Washington, DC urban area. Cities such as Chicago or
New York would be expected to have higher air concentrations than were characterized by the
urban sites in NDAMN due to proximity of air emission sources (industrial incinerator sources,
vehicular emissions, other air sources). EPA (2003) summarized the literature on dioxins in air
in the United States pertinent to the time frame of the late 1980s into the 1990s. While this time
frame is a bit earlier than the time frame of NDAMN, nonetheless it reports on numerous studies,
including those in California, Ohio, and New York, which show average air concentrations
exceeding 100 fg TEQ/m3, where "average" is over time for some temporal sampling or over
several different monitors. As noted below, recent urban measurements in California averaged
30 fg TEQ/m3. In any case, the average TEQ concentrations over all stations and moments
within these categories were (1) urban at 15.9 fg TEQ/m3, (2) rural at 13.9 fg TEQ/m3, and (3)
remote at 1.2 fg TEQ/m3. Figure 4-5 shows the land-use averages over time. The difference
between remote areas and the other two areas are clear from this figure. Otherwise, no
unambiguous trends emerge from a visual examination of this figure. It might be observed that
the rural stations, as a group, may show elevations during the fall or winter months as compared
to the spring or summer months. Perhaps that could be said as well for urban stations, but the
remote stations appear to show little variation during the course of a year. Over the 6-year
period of NDAMN, there does not appear to be any significant upward or downward change that
would occur as a result of a meaningful increase or decline in dioxin source emissions.
Concentrations found in NDAMN are comparable to or lower than similar studies
undertaken within the United States and around the world. Among the early sampling
4-13
-------
campaigns was one undertaken to study PCBs, polycyclic aromatic hydrocarbons, and
PCDD/PCDFs in urban air in the United Kingdom (UK) in the early 1990s (Coleman et al.,
1997). Between 1991 and 1995, they measured these semi-volatile compounds at sites in
London and Manchester. Samples were taken every two weeks by hi-volume samplers.
Declines in concentrations were observed at both locations between 1991 and 1994, with an
upturn in 1995. Concentrations in Manchester ranged from about 100 to approximately 450 fg
TEQ/m3, while concentrations in London were between 50 and 150 fg TEQ/m3. It is presumed
that the values were determined using the 1994 TEF values as described in Van den Berg et al.
(1994), although the authors did not provide a reference for their assignment of TEFs. A ten-
year program monitoring dioxins in air in Catalonia, Spain, was reported by Abad et al. (2007).
A total of 174 samples were taken between 1994 and 2004. Concentrations were highest in an
industrial area, with a range of 50 to 1,196 fg TEQ/m3 and a mean of 140 fg TEQ/m3. They were
second highest in a high traffic area, with a range of 10 to 357 fg TEQ/m3 and a mean of 72 fg
TEQ/m3, and they were lowest in a rural area with a narrow range of 5 to 45 fg TEQ/m3 and a
mean of 28 fg TEQ/m3. This is a bit higher than the rural results of NDAMN which averaged
13.9 fg TEQ/m3. In the United States, California's Air Resources Board conducted a CAD AMP
consisting of 12 sites which operated from 2002 to 2006 (CADAMP, 2010). Measuring 30
dioxin-like congeners including PCDDs, PCDFs, and PCBs, their average concentration for
urban sites was just above 30 fg TEQ/m3, which was about twice the total TEQ found at the
single rural site of the program. Their program was modeled after EPA's NDAMN program and
also included measurements of polybrominated diphenyl ethers. Another study with longer-term
air measurement similar to NDAMN and CAD AMP occurred between November 2004 and
December 2007, and entailed four sites around the Great Lakes of the United States (Venier et
al., 2009). Individual samples were obtained over a 24-day continuous period for three rural and
remote sites and for a 48-hour period for an urban site near Chicago, Illinois. A total of 185
samples showed similar concentrations as the California program, and a similar difference
between the urban and remote sites: the average for Chicago was 35 fg TEQ/m3 while it was 2.3,
7.4, and 13 fg TEQ/m3 for the three remote and rural sites. The state of Connecticut has
similarly conducted long-term air modeling for dioxin-like compounds, with results in rural areas
in the low fg TEQ/m3 as found in NDAMN (Hunt and Lihzis, 2011).
4-14
-------
Table 4-5 shows a comparison of the four highest measurements with the station average,
which is calculated for all other moments not including the high measurement. A very
interesting trend emerges here. For Stations 3 and 20, the concentrations of all congeners and
homologue groups in the anomalous reading are substantially higher than the station average,
from 10 times higher to over 100 times higher. The only pattern for these two stations is that
everything appears elevated. For Stations 28 and 29, a very different picture emerges. Here it is
only the dioxin congeners and dioxin homologue group concentrations that are 10 to more than
100 times higher than the station average. For the furan congeners, furan homologues, and
PCBs, the concentrations are only slightly elevated and even less than the station average. Upon
further inspection, it is noted that the station averages of dioxins (congeners and homologues) in
Stations 28 and 29 are also generally higher, by about a factor of 2, than station averages for
Stations 3 and 20. Meanwhile, furan and PCB congener/homologue group averages are all about
the same for all four stations. These trends suggest a source near Stations 28 and 29 that might
generally elevate dioxin concentrations with the potential for very high dioxin concentrations
occasionally. These sources would not appear to influence the general background of furans and
PCBs. This pattern of exaggerated elevation in dioxins, while essentially background levels of
furans, was also found in ball clay, which was discovered to be a contaminant in animal feed in
the 1990s. As described in Ferario et al. (2000), samples of ball clay from animal feeds showed
TEQ concentrations above 1,500 ppt (for comparison, soil concentrations are typically 10 ppt
TEQ or less), explained in full by elevated dioxins while furan concentrations were either absent
or at least 2 orders of magnitude lower than dioxin concentrations. Recently, Chinese
investigators have observed this pattern of high PCDD and reduced or absent PCDF
concentrations in air samples collected near six ceramic plants in China. Analyses of the exhaust
gases indicate that these plants might be contributing significant concentrations of dioxin TEQ to
the environment in China. Kaolinitic clays are the source materials for this manufacturing
process (Lu et al., 2012). Investigations have not occurred to identify potential sources near
Sites 28 and 29; certainly it is possible that combustion of a product high in PCDDs in
comparison to PCDFs such as clays might explain the findings in these monitors. Thermal
processes that preferentially emit dioxins over furans could also be the cause.
4-15
-------
4.4. REFERENCES
Abad, E; Martinez, K; Gustems, L; et al. (2007) Ten years measuring PCDDs/PCDFs in ambient air in Catalonia
(Spain). Chemosphere 67:1709-1714.
CAD AMP (California Ambient Dioxin Air Monitoring Program). (2010) California Ambient Dioxin Air Monitoring
Program 2002 to 2006 data analysis of dioxins, furans, biphenyls, and diphenylethers. Final report STI-907024.05-
3292-FR2. Prepared for the Monitoring and Laboratory Division Air Resources Board, California Environmental
Protection Agency, Sacramento, CA. Available online at
http://www.arb.ca.gov/aaqm/qmosopas/dioxins/dioxins.htm (Accessed July 2012).
Coleman, PJ; Lee, ROM; Alcock, RE; et al. (1997) Observations on PAH, PCB, and PCDD/F trends in UK urban
air, 1991-1995. Environ Sci Technol 31:2120-2124.
Ferrario, J; Cyrne, C; Dupuy, AE, Jr. (1997) Background contamination by coplanar polychlorinated biphenyls
(PCBs) in trace level high resolution gas chromatography/high resolution mass spectrometry (HRGC/HRMS)
analytical procedures. Chemosphere 34:2451-2465.
Ferario, J; Byrne, CJ; Cleverly, DH. (2000) 2,3,7,8-Dibenzo-p-dioxins in mined clay products from the United
States: evidence for possible natural origin. Environ Sci Technol 34(21):4524-4532.
Hunt, GT. (2008) Atmospheric concentrations of PCDDs/PCDFs in Metropolitan Hartford Connecticut-current
levels and historical data. Chemosphere 73:8106-8113.
Hunt, GT; Lihzis, MF. (2011) PCDDs/PCDFs in ambient air (<1 fg m-3) - the CTDEP long term sampling (30d)
method. Chemosphere 85:1664-1671.
Lorber, M; Pinsky, P; Gehring, P; et al. (1998) Relationships between dioxins in soil, air, ash, and emissions from a
municipal solid waste incinerator emitting large amounts of dioxins. Chemosphere 37(9-12):2173-2197.
Lu, M; Wang, G; Su, Y. (2012) Characterization and inventory of PCDD/F emission for the ceramic industry in
China. Environ Sci Technol 46:4159-4165.
U.S. EPA (Environmental Protection Agency). (2003) Exposure and human health reassessment of 2,3,7,8-
tetrachlorodibenzo-p-dioxin (TCDD) and related compounds. [National Academy Sciences Review Draft] National
Center for Environmental Assessment, Washington, DC; EPA/600/P-00/001Cb. Available at
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=87843.
Van den Berg, M; De Jongh, J; Poiger, H; et al. (1994) The toxicokinetics and metabolism of polychlorinated
dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) and their relevance for toxicity. Crit Rev Toxicol 24:1-74.
Venier, M; Ferrario, J; Kites, RA. (2009) Polychlorinated dibenzo-p-dioxins and dibenzofurans in the atmosphere
around the Great Lakes. Environ Sci Technol 43:1036-1041.
4-16
-------
Table 4-1. Operating status of all stations over all moments
Station
Number
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Moment Number
1
V
NA
V
V
V
V
V
V
2
V
V
V
V
V
V
V
V
V
3
V
V
V
V
V
V
V
V
V
4
V
V
NA
V
NA
V
V
V
V
5
V
NA
V
V
V
V
V
V
V
6
V
V
V
/
NA
/
I
NA
I
V
7
V
V
NA
/
NA
/
I
NA
I
V
V
V
V
8
V
NA
V
V
NA
V
NA
NA
V
V
V
V
V
V
9
V
V
V
V
V
V
V
V
NA
V
V
V
V
V
V
V
V
V
10
V
V
V
V
V
V
V
NA
V
NA
NA
V
V
V
NA
V
V
V
11
V
V
V
NA
/
/
/
I
1
V
NA
V
NA
V
V
V
/
1
1
12
V
V
V
1
1
1
1
1
1
V
V
V
NA
V
V
V
/
1
1
13
V
V
V
1
1
1
1
NA
/
V
V
V
V
V
V
V
1
1
1
1
14
V
V
V
V
V
V
V
V
V
NA
V
V
V
V
V
V
V
V
V
V
15
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
16
V
V
V
V
V
V
V
NA
V
V
V
V
V
V
V
V
V
V
V
V
17
V
V
V
V
V
V
V
NA
V
V
V
V
V
V
V
V
NA
V
V
V
18
V
V
V
1
1
1
1
1
1
V
V
V
V
V
V
V
1
1
1
1
19
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
20
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
NA
V
V
21
V
V
V
/
1
1
1
1
1
V
V
V
V
V
V
V
/
1
1
1
22
V
V
V
V
V
V
V
V
V
V
NA
V
V
V
V
V
V
V
V
V
23
V
V
NA
V
V
V
V
V
V
V
V
V
V
V
V
V
V
NA
V
V
24
V
V
V
/
/
/
/
/
/
V
V
V
V
V
V
V
/
/
/
/
25
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
26
V
V
V
1
1
1
1
1
1
V
V
V
V
V
V
V
/
/
1
1
27
V
V
NA
1
1
1
1
1
1
V
NA
V
V
V
V
V
1
1
1
1
28
V
NA
NA
1
1
1
1
1
1
V
NA
NA
NA
NA
NA
NA
NA
NA
1
1
29
V
V
NA
1
NA
1
1
1
1
V
V
NA
V
V
V
V
1
1
1
1
TOTAL V
29
25
22
28
24
29
28
22
28
20
18
22
20
22
19
18
15
18
21
21
-------
Table 4-1. Operating status of all stations over all moments (continued)
Station
Number
22
23
24
25
26
27
28
29
30
31
32
33
34
35
TOTAL
Moment Number
1
7
2
9
3
9
4
7
5
8
6
8
7
10
8
10
9
V
V
V
V
21
10
V
V
I
1
1
1
1
21
11
V
V
1
1
1
1
1
23
12
V
V
1
1
1
1
1
V
26
13
V
V
1
1
1
1
1
V
V
28
14
V
V
/
1
1
1
1
V
V
28
15
V
V
V
V
V
V
V
V
V
V
V
31
16
V
V
1
1
1
NA
1
1
V
V
V
29
17
V
V
1
1
1
1
1
1
V
V
V
29
18
V
V
V
V
V
V
V
V
V
V
V
V
V
33
19
V
V
V
V
V
V
V
V
V
V
V
V
V
33
20
V
V
/
1
1
1
1
1
V
V
V
NA
V
31
21
V
V
V
V
V
V
V
V
V
V
V
V
V
33
22
V
V
1
1
1
1
1
NA
V
V
V
V
V
V
32
23
V
V
V
V
V
V
V
V
V
V
V
V
V
V
32
24
V
V
1
1
1
1
1
1
V
V
V
V
V
V
34
25
V
V
/
1
1
1
1
1
NA
V
V
V
V
V
33
26
V
V
/
/
/
/
/
/
V
V
V
V
V
V
34
27
V
V
V
V
V
V
V
V
V
V
V
V
V
NA
31
28
V
V
1
1
1
1
1
1
V
V
V
V
V
V
24
29
V
V
1
1
1
1
1
1
V
V
V
V
V
V
31
TOTAL V
20
20
21
21
21
20
15
19
17
17
15
11
12
7
685
oo
Notes: Station 2 was a QA/QC station, located adjacent to Station 1 for most of the sampling program. Results for Station 2 not included on table.
Key: "V" = station operating; " " (blank) = station not yet operating; NA = data not available, due mostly to the station not operating but also, on occasion, to
QA failures at the lab or lost sample (see text).
-------
Table 4-2. Comparison of congener-specific results from laboratory blanks
with those from remote, rural, and urban stations of NDAMN
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
PCB77
PCB 105
PCB 118
PCB 126
PCB 156
PCB 157
PCB 169
NDAMN Results, pg1
Remote
(n=153)
0.4
2
3
6
5
76
265
2
2
o
J
o
J
o
J
5
1
20
2
29
238
1,260
3,316
8
214
48
0
Rural
(n=463)
4
20
27
50
49
718
2,521
14
14
24
31
28
37
5
158
20
139
448
2,401
6,650
32
375
80
5
Urban
(n=69)
5
27
31
56
54
716
2,356
21
20
32
38
40
46
15
187
18
119
942
6,408
17,191
68
895
191
6
Blank Results, pg
Target2
0.5
1.5
2.5
2.5
2.5
2.5
20.0
0.5
1.5
1.5
1.5
1.5
1.5
1.5
2.5
2.5
4.0
20.0
500.0
300.0
2.0
80.0
20.0
1.0
Field
Actual3
(n=56)
0.1
0.1
0.1
0.1
0.1
5.1
32.1
0.1
0.1
0.2
0.3
0.3
0.8
0.1
1.3
0.1
1.5
43.1
518.6
1,017.3
2.5
124.1
26.6
<0.1
Percent
Positive4
2
2
2
4
4
85
6
2
2
2
7
4
39
6
20
2
37
98
98
98
65
98
98
6
Laboratory
Actual3
(n=108)
0.1
0.2
0.3
0.5
0.5
10.2
43.3
0.1
0.1
0.2
0.3
0.3
1.1
0.1
1.6
0.2
3.3
31.7
271.6
556.8
1.3
72.6
15.7
0.2
Percent
Positive4
2
3
9
6
4
93
99
7
5
7
19
15
84
6
94
8
88
99
99
99
71
99
99
8
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.
4-19
-------
Table 4-2. Comparison of congener-specific results from laboratory blanks
with those from remote, rural, and urban stations of NDAMN (continued)
Average measurement in pg, independent of volume (see text for more detail), assuming non-detects equal to 0.
2"Target" levels were the method detection limit levels determined in method development prior to NDAMN.
3"Actual" levels were the average mass measured in 56 field and 108 laboratory method blank samples during
NDAMN, with non-detects equal to 0.
4"Percent Positive" was the percentage of field and method blanks with quantified measurements.
4-20
-------
Table 4-3. Comparison of average concentrations in the NDAMN monitors,
#1 and #3, with the adjacent QA sampler, #2, the correlation coefficient, r,
between the set of measurements, and the average Relative Percent
Difference, RPD, between the set of measurements (concentrations in fg/m3;
NA = no data available)
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
PCB77
PCB81
PCB 105
PCB 114
PCB 118
PCB 123
PCB 126
PCB 156
PCB 157
PCB 167
PCB 169
PCB 189
Penn Nursery (n = 18)
St#l
0.88
4.56
5.02
8.30
7.41
118.72
533.86
2.30
2.19
3.53
4.35
4.42
5.16
0.93
28.74
2.70
29.32
58.38
NA
267.41
NA
761.72
NA
7.14
48.28
10.85
NA
0.85
ND
St#2
0.82
4.33
4.96
8.19
7.38
113.17
417.23
2.40
2.20
3.64
4.23
4.39
5.06
0.99
26.97
2.69
26.41
81.15
NA
500.91
NA
1,318.11
NA
8.46
90.75
20.38
NA
0.82
ND
R
0.95
0.98
0.99
0.97
0.97
0.98
0.88
0.84
0.83
0.91
0.85
0.72
0.87
0.96
0.81
0.93
0.93
0.61
0.27
0.15
0.92
0.22
0.25
0.94
— -
RPD
17
18
16
19
20
23
36
18
16
16
16
16
16
21
17
17
24
33
39
39
23
39
38
38
Clinton Crops (n = 9)
St#3
0.52
2.97
3.34
6.32
5.73
71.44
253.37
3.12
3.62
6.69
8.28
7.72
10.31
0.82
37.78
4.63
27.16
93.93
7.17
377.50
31.06
877.40
17.82
7.04
52.24
11.49
21.58
1.61
4.24
St#2
0.58
3.21
3.19
5.99
5.37
69.59
282.76
2.87
3.32
5.89
7.38
5.85
9.30
0.72
35.70
4.19
27.15
102.17
6.68
487.73
35.48
1,141.23
21.16
17.91
86.37
19.03
48.19
2.70
7.01
R
-0.10
-0.05
0.69
0.68
0.80
0.98
0.88
0.74
0.75
0.81
0.84
0.59
0.83
0.96
0.85
0.92
0.97
0.89
0.57
0.79
0.87
0.87
0.87
-0.26
0.85
0.83
0.57
-0.24
-0.15
RPD
31
24
20
21
18
9
16
15
13
17
17
16
18
33
16
20
14
24
20
16
25
30
27
44
45
46
56
37
40
4-21
-------
Table 4-3. Comparison of average concentrations in the NDAMN monitors,
#1 and #3, with the adjacent QA sampler, #2, the correlation coefficient, r,
between the set of measurements, and the average Relative Percent
Difference, RPD, between the set of measurements (concentrations in fg/m3;
NA = no data available) (continued)
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.
4-22
-------
Table 4-4. Comparison of congener concentrations in co-located samplers in
Pennsylvania for two occurrences with discrepencies in volume (Spring and
Summer, 2000) and all co-located samples in Pennsyania (concentrations in
fg/m3; ND = no data, NA = not applicable, RPD = relative percent difference)
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
PCB77
PCB 105
PCB 118
PCB 126
PCB 156
PCB 157
PCB 169
PCB 189
Average RPD
Volume, m3
Spring 2000
St#l
1.5
9.4
12.9
22.8
22.0
318.4
1,325.9
5.6
4.7
8.6
8.2
7.4
10.1
0.9
41.2
5.7
46.5
36.1
205.8
548.3
8.0
46.8
11.1
1.5
1.5
4,859
St#2
1.5
10.6
15.5
30.1
28.7
606.6
3,303.8
11.1
9.4
16.8
16.7
14.4
21.3
1.8
87.9
11.9
106.2
82.3
683.8
1,846.9
17.0
147.2
32.6
2.0
1.5
1,198
RPD
o
J
12
19
28
27
62
85
66
66
65
68
64
71
59
72
70
78
78
107
108
72
104
98
28
3
66
Summer 2000
St#l
0.33
1.70
2.11
5.14
3.67
89.09
507.08
2.82
2.23
3.38
3.94
3.52
4.93
1.18
23.22
2.48
28.42
46.67
252.58
745.44
6.03
53.17
13.00
0.76
0.33
5,564
St#2
ND
1.31
1.56
3.95
3.24
63.39
287.25
3.02
1.78
2.98
3.05
3.05
4.44
1.23
17.73
1.73
21.01
45.24
280.96
790.33
5.73
64.97
15.94
0.64
ND
3,474
RPD
NA
26
30
26
12
34
55
7
23
13
25
14
11
4
27
36
30
o
J
11
6
5
20
20
18
NA
20
All Other Samples (n=17)
St#l
0.9
4.7
5.2
8.5
7.6
120.5
535.4
2.3
2.2
3.5
4.4
4.5
5.2
0.9
29.1
2.7
29.4
59.1
268.3
762.7
7.2
48.0
10.7
0.9
0.9
6,585
St#2
0.9
4.5
5.2
8.4
7.6
116.1
424.9
2.4
2.2
3.7
4.3
4.5
5.1
1.0
27.5
2.7
26.7
83.3
513.9
1,349.2
8.6
92.3
19.9
0.8
0.9
7,368
RPD
17
18
16
19
20
22
35
19
16
17
16
16
17
33
17
16
24
34
41
41
24
40
39
39
17
25
4-23
-------
Table 4-4. Comparison of congener concentrations in co-located samplers in
Pennsylvania for two occurrences with discrepencies in volume (Spring and
Summer, 2000) and all co-located samples in Pennsyania (concentrations in
fg/m3; ND = no data, NA = not applicable, RPD = relative percent difference)
(continued)
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.
4-24
-------
Table 4-5. Survey-wide statistics for all congeners and homologue groups
(concentrations in fg/m3)
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Total TCDF
Total TCDD
Total PeCDF
Total PeCDD
Total HxCDF
Total HxCDD
Total HpCDF
Total HpCDD
PCB77
PCB81
PCB 105
PCB 114
PCB 118
PCB 123
Percentage
Detected
85
89
94
97
96
100
100
96
94
96
98
98
99
74
100
91
99
99
98
98
94
99
99
99
98
100
100
99
100
99
100
Mean
0.6
3.1
4.2
7.3
7.2
102.3
352.8
2.1
2.4
4.3
5.6
4.9
6.4
1.5
27.3
3.5
21.9
75.4
18.4
57.2
40.0
58.2
102.1
43.9
241.6
157.2
12.5
629.8
47.4
1,430.3
32.8
SD
1.2
5.9
10.4
15.3
15.3
243.6
973.4
9.6
14.1
28.8
41.4
31.1
41.3
22.3
178.1
25.2
142.8
263.2
69.9
303.4
132.5
262.7
220.2
232.7
520.0
1,286.7
104.8
3,601.8
375.3
6,248.5
273.8
95% CI
0.5-0.7
2.7-3.6
3.4-5.0
6.1-8.4
6.1-8.3
85.4-120.5
28-425
1.4-2.8
1.3-3.4
2.1-6.4
2.6-8.7
2.6-7.2
3.3-9.5
0-3.1
14.0-40.6
1.6-5.4
11.2-32.5
55.7-94.9
13.2-23.6
34.6-79.8
30.1-49.8
38.9-77.8
85.7-118.5
26.6-61.3
202.8-281.3
61.3-253.0
1.0-24.1
361-898
6.4-88.9
965-1,896
2.7-62.9
Median
0.3
1.7
2.1
3.8
3.5
52.8
187.4
1.1
1.1
1.7
2.2
2.0
2.6
0.3
11.3
1.2
9.8
44.1
9.0
26.6
18.1
26.9
52.3
19.2
131.1
36.9
2.9
188.3
13.9
489.5
9.1
Max
23
87
209
257
305
5,487
23,953
249
361
738
1,056
787
1,031
597
4,498
644
3,721
6,300
1,732
7,619
2,962
6,467
3,293
5,735
10,975
31,167
1,539
80,653
6,895
134,846
4,923
4-25
-------
Table 4-5. Survey-wide statistics for all congeners and homologue
groups (concentrations in fg/m3) (continued)
Congener
PCB 126
PCB 156
PCB 157
PCB 167
PCB 169
PCB 189
TEQDF
TEQP
TEQ DFP
Percentage
Detected
100
99
99
100
83
100
Mean
6.9
67.7
14.9
22.2
0.9
2.7
10.5
0.8
11.3
SD
32.7
168.6
37.9
67.4
9.7
4.6
33.2
3.7
36.1
95% CI
4.5-9.3
55.1-80.2
12.1-17.7
14.8-29.6
0.2-1.7
2.2-3.2
8.1-12.9
0.6-1.1
8.7-13.9
Median
3.0
30.2
6.8
9.9
0.3
1.7
5.9
0.4
6.5
Max
758
2,633
590
1,083
260
50
773
84
857
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl.; DF=dioxins and furans when used in TEQ DF; P = PCBs when used in TEQ P; DFP =
dioxins, furans, and PCBs when used in TEQ DFP.
n = 685 except for PCBs 81, 114, 123, 167, and 189, where n = 318. See Section 3.1 for more detail.
4-26
-------
Table 4-6. Comparison of the four highest concentrations measured with the
station average where that concentration was measured (for each station, the
date given as month/year—1/2000—is compared to the station
average—AVG; all concentrations in fg/m )
Congener
2,3,7,8-TCDD
1,2,3,7,8-PeCDD
1,2,3,4,7,8-HxCDD
1,2,3,6,7,8-HxCDD
1,2,3,7,8,9-HxCDD
1,2,3,4,6,7,8-HpCDD
OCDD
2,3,7,8-TCDF
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
1,2,3,4,7,8-HxCDF
1,2,3,6,7,8-HxCDF
2,3,4,6,7,8-HxCDF
1,2,3,7,8,9-HxCDF
1,2,3,4,6,7,8-HpCDF
1,2,3,4,7,8,9-HpCDF
OCDF
Total TCDF
Total TCDD
Total PeCDF
Total PeCDD
Total HxCDF
Total HxCDD
Total HpCDF
Total HpCDD
Station 20, MN
1/2000
5.6
55.9
118.9
191.4
98.4
1,336.8
2,892.5
249.0
361.3
738.0
1,055.9
786.7
1,030.5
596.9
4,498.2
644.4
3,721.4
6,299.9
327.9
7,619.2
1,153.5
6,467.3
2,235.0
5,735.3
2,907.2
AVG
0.3
1.8
2.9
4.4
4.9
73.6
297.4
2.0
1.7
3.0
3.6
3.0
3.9
0.8
16.1
2.2
12.7
61.0
13.9
36.6
22.7
34.0
63.7
26.0
166.0
Station 3, NC
11/2001
6.3
42.5
57.6
125.0
105.5
860.7
1,185.3
48.1
95.2
213.3
330.4
256.4
386.5
43.9
1,532.3
193.1
815.8
2,749.3
1,732.2
2,557.0
2,962.4
2,363.8
3,293.2
2,224.8
2,302.0
AVG
0.5
2.8
3.4
6.0
5.6
69.8
235.0
3.2
3.5
6.6
7.9
7.3
10.2
1.6
39.3
4.6
29.1
132.9
32.7
91.9
51.9
89.3
100.1
58.0
171.1
Station 29, OR
9/2000
22.6
68.3
55.8
100.6
99.3
1,046.5
2,627.0
2.1
2.6
5.0
6.7
4.8
5.8
1.7
37.8
3.5
24.4
91.9
441.1
63.8
968.2
96.2
1,819.5
103.2
2,997.1
AVG
1.2
8.5
13.6
22.5
22.2
321.6
914.0
1.9
1.8
3.4
3.9
3.6
4.5
0.6
20.0
2.5
20.5
63.2
27.2
41.5
97.7
53.2
333.1
41.3
799.6
Station 28, CA
2/2003
7.0
86.5
209.0
257.0
304.9
5,487.4
23,953.0
2.5
5.5
7.4
13.8
14.2
12.9
2.7
74.0
28.8
85.4
110.3
120.5
118.8
643.1
130.2
2,758.2
164.0
10,974.8
AVG
2.1
9.7
8.2
14.1
13.8
180.7
582.3
2.0
1.8
3.0
3.8
3.7
4.6
0.3
21.5
2.1
16.8
65.7
20.4
55.1
57.3
66.9
164.7
40.8
392.1
4-27
-------
Table 4-6. Comparison of the four highest concentrations measured
with the station average where that concentration was measured (for
each station, the date given as month/year—1/2000—is compared to the
station average—AVG; all concentrations in fg/m3) (continued)
Congener
PCB77
PCB81
PCB 105
PCB 114
PCB 118
PCB 123
PCB 126
PCB 156
PCB 157
PCB 167
PCB 169
PCB 189
TEQDF
TEQP
TEQ DFP
Station 20, MN
1/2000
452.9
702.7
967.7
758.2
724.6
262.5
260.4
773.3
83.8
857.1
AVG
27.0
124.3
304.0
3.2
23.6
5.4
0.6
6.6
0.4
6.9
Station 3, NC
11/2001
357.0
3,017.0
8,207.5
144.0
599.0
156.2
ND
277.5
14.8
292.3
AVG
91.8
447.1
1,096.6
7.6
72.1
16.1
1.4
11.1
0.9
11.9
Station 29, OR
12/2000
65.8
618.1
2,443.6
6.2
131.3
27.7
0.0
0.3
131.8
0.7
132.6
AVG
52.6
513.1
1,313.0
7.0
4.4
124.2
26.9
12.8
0.4
1.0
21.8
0.5
22.3
Station 28, CA
8/2001
58.0
3.8
238.1
21.2
736.0
12.0
4.0
32.3
7.3
13.7
0.4
1.9
240.7
0.4
241.2
AVG
93.2
3.0
580.3
29.7
1,659.6
17.4
8.6
89.6
20.0
26.2
0.5
2.0
20.0
1.0
21.0
Notes: TCDD = tetachlorodibenzo-p-dioxin; PeCDD = pentachlorodibenzo-p-dioxin;
HxCDD = hexachlorodibenzo-p-dioxin; HpCDD = heptachlorodibenzo-p-dioxin;
OCDD = octochlorodibenzo-p-dioxin; TCDF = tetrachlorodibenzofuran; PeCDF = pentachlorodibenzofuran;
HxCDF = hexachlorodibenzofuran; HpCDF = heptachlorodibenzofuran; OCDF = octachlorodibenzofuran;
TCB = tetrachlorobiphenyl; PeCB = pentachlorobiphenyl; HxCB = hexachlorobiphenyl; HpCB =
heptachlorobiphenyl; DF=dioxins and furans when used in TEQ DF; P = PCBs when used in TEQ P; DFP =
dioxins, furans, and PCBs when used in TEQ DFP.
4-28
-------
a. Moments 1 through 19 at Penn Nursery in Pennsylvania.
c
o
u
OJ
4-J
i/1
0.5
0.0 4- »—i-
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Sitel, PANDAMN,fg/m3
b. Moments 20 through 29 (excluding 28) at Clinton Crops, NC.
2
4->
O
CM
(D
2,3,7,8-TCDD
r = -0.10
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
SiteS, NCNDAMN, fg/m3
Figure 4-1. Comparison of control and NDAMN results for 2,3,7,8-TCDD
(the perfect correlation j = x is shown as dashed line).
4-29
-------
a. Moments 1 through 19 at Penn Nursery in Pennsylvania.
0
u
IN
OJ
200
r = 0.25
100 200
500 600
Site 1, PAN DAMN, fg/m3
b. Moments 20 through 29 (excluding 28) at Clinton Crops, NC.
Ł
150
100
PCB156
r = 0.85
y
s •
50 100 160 200
SiteS, NCNDAMN, fg/m3
Figure 4-2. Comparison of control and NDAMN results for PCB 156 (the
perfect correlation y =x is shown as dashed line).
4-30
-------
10000
.5 7500
o
01
T3
0)
4-1
CD
(J
^g
6
u
5000
2500
V
X
0 2500 5000 7500 10000
NDAMN Sample Volume, m3
Figure 4-3. Comparison of control and NDAMN sample volumes (the perfect
correlation y = x is shown as dashed line).
Figure 4-4. Average TEQ concentrations found at all NDAMN stations.
4-31
-------
•Rural
•Remote -*-Urban
100.0
Figure 4-5. Temporal variability of TEQ concentrations averaged by station
characterization (note: calendar year 1999 had six sample moments, 1998
had three moments, and all other years had four moments. Each year
identified on the x-axis corresponds to January of that year).
4-32
-------
APPENDIX A. DESCRIPTION OF EXCEL WORKBOOK CONTAINING NDAMN
DATA
The NDAMN Workbook contains three worksheets: (1) Title Page, (2) All NDAMN
Raw Data 1998 to 2004, and (3) NDAMN concentrations at non-detect (ND) equal to zero. The
following comment and column descriptions are included for each worksheet:
Title Page: This contains a brief description of the database, a map showing station locations, a
legend to the map providing location names, the full citation of the report, and contact
information.
All NDAMN Raw Data: This worksheet contains the raw data, including sample identification
information, sample volumes, and all analytical data as reported by the laboratory for each
sample. The column definitions, mostly self explanatory, are as follows:
Column A: Station name
Column B: Nearest town
Column C: Type of station, characterized as either "rural," "urban," or "remote." See Chapter 2
for further discussion on the study design.
Column D: Latitude
Column E: Longitude
Column F: Station number
Column G: Sampling moment. There were 29 sampling moments in the study. See Chapter 2
for further discussion on the study design.
Column H: Dates of sampling including month/day to month/day (06/16-07/14); formatted as
text.
Column I: Year of sampling
Column!: Season of sampling
Column K: Operating Status. There are three conditions of operation: (1) "Operating" meaning
that the sampler was fully operational and samples were delivered to the laboratory, (2) "Not
Operating" meaning that for some reason, the sampler did not obtain a sample for analysis (it
may have been operating a part of the time but then a failure resulted; for logistical or other
reasons, it may not have been operating at all during the sampling moment, or some other
reason), and (3) "QA Failure" meaning a sample was obtained and shipped to the laboratory
where a quality assurance failure resulted in no or only partial data being developed. The
analytical result fields provide more information on the status of measurements.
Column L: Sample volume, m3
Columns M-AW: Laboratory analytical results for: 7 dioxin congeners, 10 furan congeners, 8
dioxin and furan homologue group concentrations, and 12 dioxin-like PCB measurements. See
Table 3-1 for a complete list of the analytes measured in the study. The information in each cell
can take these forms: (1) picograms of analyte measured by the laboratory (reported in 2 or 3
places after the decimal point), the concentration is simply these picograms divided by the
sample volume; (2) blank—in some cases, this analyte was not measured in the sample—this
was the case for several PCB congeners which were measured starting later in the program. In
other cases, the sampler was not operating so there is nothing to report, (3) ND—this analyte not
detected in the sample (see Table 3-2 for the analyte-specific detection limit in pg). To get the
A-l
-------
sample- and analyte-specific detection limit, simply divide the analyte specific detection in Table
3-2 with the sample-specific volume provided in Column L; and (4) QA Failure—no value can
be reported due to a quality assurance failure at the laboratory.
NDAMN Concentrations, ND = 0: This is the complete set of data where concentrations were
derived by dividing the picograms reported by the laboratory by the sample volumes, with zero
values substituted for non-detects. This was the data set upon which all of the results in this
report were derived. The column definitions, mostly self-explanatory, are as follows:
Column A: Station name
Column B: Station number
Column C: Sample moment number
Column D: Dates of sampling
Column E: Year of sampling
Column F: Season of sampling
Columns G through AQ: Final concentrations, in fg/m3, for each of 7 dioxin congeners, 10 furan
congeners, 8 dioxin and furan homologue group concentrations, and 12 dioxin-like PCB
measurements. The information in each cell can take these forms: (1) the concentration
calculated as the mass of analyte reported by the laboratory, in picograms, divided by the sample
volume (both of these on the previous worksheet), with conversion to arrive at the concentration
in fg/m3. The concentration is formatted to report concentrations two places after the decimal
point; (2) blank—there are three possible reasons for a blank: (a) the analyte was not measured in
the sample while others may have been measured, (b) the sampler was not operating so no
analytes were measured in the sample, and (c) there was a QA failure for this analyte (in some
cases, the QA failure pertained to all analytes, but in others, a portion of the analytes were
measured and reported); and (3) the value "0." Non-detects were replaced by zeros.
Column AR, AS, and AT: It is noted that row 9, columns G through AI, contain TEF values for
each of the congeners in columns G through AI. There is no TEF value associated with the
homologue group concentrations reported in columns AJ through AQ. Rows AR, AS, and AT
provide calculations for TEQ concentrations of dioxins and furans (Column AR), dioxin-like
PCBs (Column AS), and Total (sum of columns AR and AS).
A-2
-------
APPENDIX B. QUALITY ASSURANCE PROJECT PLAN FOR NDAMN
Two Quality Assurance Project Plans (QAPPs) are included in this Appendix:
1. A QAPP for field implementation procedures prepared by Battelle dated April 2001 and titled
"Dioxin Exposure Initiative Implementation, Operation, and Maintenance of the National Dioxin
Air Monitoring Network (NDAMN)" that was signed by principals from EPA and Battelle.
2. A QAPP for the analytical methods prepared by the Environmental Chemistry Laboratory of
EPA dated July 2001 and titled "Quality Assurance Project Plan for the Dioxin Exposure
Initiative: National Dioxin Air Monitoring Network" that had signature blocks but was not
signed.
It should be noted that these are provided here as examples of the documents developed
over the course of NDAMN that were used in the implementation of the study. For example,
before Battelle was the primary contractor for NDAMN implementation, that task was performed
by Versar, Inc, of Springfield, Virginia. There were documents associated with siting of the
samplers on public lands and maintenance of samplers. There were sample transmittal forms and
lab receipt information. There were also documents associated with shutdown of the field
samplers and disposition of equipment. These examples are provided simply to give readers a
sense of the complexity and the procedures used in NDAMN for sampling and analysis.
B-l
-------
United States
Environmental Protection
Agency
National Center for Environmental Assessment
Office of Research and Development
Washington, DC 20460
------- |