United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/8-91/007
March 1991
Feasibility of Environmental
Monitoring and Exposure
Assessment for a Municipal
Waste Oombustor: Rutland,
Vermont Pilot Study
_
V^if*^.—^ -^.ii-.iTpi^a-'" #»ViFH,,5
-------
-------
EPA/600/8-91/007
Feasibility of Environmental Monitoring and
Exposure Assessment for a Municipal Waste Combustor:
Rutland, Vermont Pilot Study
Environmental Criteria and Assessment Office
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268 . . .. •
Printed on Recycled Paper
-------
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.
11
-------
PREFACE
In response to a Congressional mandate, a study was undertaken
by the Office of Research and Development, to monitor several metal
and organic-pollutants in air and other environmental media near
the Rutland, Vermont Municipal Waste Combustor (MWC) facility and
to estimate the magnitude of any increases in health risk. As data
became available, it became apparent that there was no obvious
relationship between the operation of the MWC and ambient air
pollution levels. Therefore, the focus of the study shifted from
one of health risk assessment to one of more sophisticated
statistical analysis to determine whether any influence of the MWC
was detectable.
This final report is intended as a summary of the study
undertaken in Rutland, Vermont and some practical applications of
the feasibility of conducting environmental monitoring and exposure
assessment of such facilities. A companion report will be prepared
as a guidance manual utilizing the findings summarized in this
report to provide a "blueprint" for other long-term, multimedia and
multipollutant monitoring studies that States or permit applicants
may elect to undertake to address questions of impact associated
with municipal waste combustors.
This report has been peer reviewed by scientists within and
external to the Agency culminating in a workshop which was held in
February, 1990. The discussions held at the workshop resulted in
this final report and the future direction of the development of a
companion guidance manual. This study was undertaken under
Cooperative Agreement No. CX184651-01 with the State of Vermont.
For more information, please contact Cynthia Sonich-Mullin,
Environmental Criteria and Assessment Off ice, U.S. EPA, Cincinnati,
Ohio 45268.
ill
-------
DOCUMENT DEVELOPMENT
Authors and Contributors
C. Sonich-Mullin, Document Co-Manager
R.J.F. Bruins, Document Co-Manager
Environmental Criteria and
Assessment Office
Office of Health and Environ-
mental Assessment
U.S. Environmental Protection
Agency
Cincinnati, OH 45268
L. Fradkin
Office of Technology Transfer
and Regulatory Support
U.S. Environmental Protection
Agency
Cincinnati, OH 45268
P.M. McGinnis
M.A. Eichelberger
D.A. Gray
Chemical Hazard Assessment Division
Syracuse Research Corporation
Cincinnati, OH 45206.
Syracuse, NY 13210
M. Callahan
Exposure Assessment Group
Office of Health and
Environmental Assessment .
U.S. Environmental
Protection Agency
Washington, D.c. 20013
G.K. Moss
Office of Air Quality,
Planning and Standards
U.S. Environmental
Protection Agency
Research Triangle Park,
NC 27711
B.J. Fitzgerald
H. Garabedian
G.A. Hall
R.A. Valentinetti
Air Pollution Control
Division
Agency of Natural Resources
State of Vermont
Waterbury, VT 05676
T.C. Lawless
R.L. Harless
T.A. Hartlage
J.F. Walling
Atmospheric Research and
Assessment Laboratory
U.S. Environmental
Protection Agency
Research Triangle Park,
NC 27711
D. DeMarini
R.R. Watts
Environmental Health
Research and Testing
Laboratory
U.S. Environmental
Protection Agency
Research Triangle Park,
NC 27709
T.S. Sander
PEI Associates, Inc.
Cincinnati, OH 45246
P. Cramer
Midwest Research
Institute
Kansas City, MO 64110
D. McDaniel
Environmental Chemistry
'Section
U.S. Environmental
Protection Agency
Stennis Space Center, MO
39529
IV
-------
DOCUMENT DEVELOPMENT (cont.)
External Reviewers
H. Ozkaynak
Energy and Environmental Policy
Center
Harvard University
Cambridge, MA 02138
S.S. Que Hee
Environmental and Occupational
Health Sciences
University of California
Los .Angeles, CA 90024
T.O. Tiernan
Toxic Contaminant Research Program
Wright State University
Dayton, OH 45435
Technical Publications Editor
J. Olsen
Environmental Criteria and Assessment
Office
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
-------
TABLE OF CONTENTS
Page
1. INTRODUCTION 1-1
1.1. PROJECT OBJECTIVE 1-1
1.2. THE RUTLAND RESOURCE RECOVERY FACILITY....... 1-2
1.3. STUDY APPROACH. 1-9
2. SITE SELECTION, SAMPLING AND ANALYSIS 2-1
2.1. AIR DISPERSION MODELING FOR SELECTION
OF MONITORING SITES 2-1
2.2. SAMPLING AND ANALYSIS 2-4
2.2.1. Ambient Air Sampling. 2-7
2.2.2. Meteorologic Information 2-9
2.2.3 Ambient Air Analyses............... 2-12
2.2.4. Environmental Media Sampling 2-19
2.2.5. Environmental Media Analysis 2-23
3. MEASURED CONCENTRATIONS IN AMBIENT AIR AND
ENVIRONMENTAL MEDIA 3-1
3.1. RESULTS OF MONITORED CONCENTRATIONS IN
AMBIENT AIR . . . , ... ........ 3-1
3.1.1. Metal concentrations 3-2
3il.2. Benzo(a)pyrene 3-8
3.1.3. PCB Concentrations 3-9
3.1.4. PCDD/PCDF 3-9
3.2. ENVIRONMENTAL MEDIA 3-20
3.2.1* Metals 3-21
3.2.2. PCBs... 3-21
3.2.3. PCDD/PCDF 3-21
4. AIR DISPERSION MODELING. 4-1
4.1. METEOROLOGIC RESULTS 4-1
4.1.1. SLAMS Site . i 4-15
4.1.2. Watkins Avenue Site 4-15
4.1.3. River Street Site. 4-16
4.1.4. Conclusion.... 4-16
4.2. MODELING METHODOLOGY. . . „ 4-17
4.2.1. Stack Emission Testing 4-19
vi i
-------
TABLE pF CONTENTS , (cont.)
4.3. PROBLEMS AND UNCERTAINTIES ASSOCIATED
WITH THE MODELING. . . . .'. . 4-20 :
4.4. ISCST MODELING RESULTS FOR RUTLAND , 4-26
5. APPROACHES FOR ANALYSIS OF SOURCE CONTRIBUTION.... 5-1
5.1. AMBIENT AIR APPROACHES. . ........... . . 5-1
5.1.2. Qualitative Approaches to
Analyzing Air .Source Contribution. . 5-2
5.1.3. Quantitative Approaches to
Analyzing Ambient Air Source
Contribution. . ..'. 5-4
5.2. ENVIRONMENTAL MEDIA - . . . 5-8
,«
6. CORRELATION OF TONS OF WASTE .BURNED TO
PARTICULATE CONCENTRATION. 6-1
7. MUTAGENICITY . . ; 7-1
8. AMBIENT AIR PCDD/PCDF CONGENER PROFILES 8-1
9. ANALYSIS OF MODELED AND MEASURED AMBIENT
AIR CONCENTRATIONS 9-1
9.1. COMPARISON OF MEASURED AND MODELED.LEAD 9-4
9.1.1. Modified Sign Test Analysis
for Lead. . 9-4-
9.1.2. Friedman Nonparametric ANOVA
for Lead . . . .... ; 9-10
9.2. COMPARISON OF MODELED AND MEASURED
PCDD/PCDF. 9-14
9.2.1. Modified Sign Test Analysis for
PCDD/PCDF 9-16
9.2.2. Friedman Nonparametric ANOVA for
PCDD/PCDF . 9-19
9.3. CONCLUSION , . 9-23
10. LONG-.TERM AIR DISPERSION MODELING. . . . . 10-1
10.1. MODELING METHODOLOGY. . . 10-1
10.2. ISCLT RESULTS 10-3
i i
-------
TABLE OF CONTENTS (corit. )'
11,
12,
1.3.
10.2.1. PoIiutaht^Specif ic.,
Co,ncentr alb ions.l,'... .
10 . 3 . CONCLUSION. .,
ENVIRONMENTAL MEDIA RESULTS,
11.1,
11.2
11.3
METALS,
11.1.1.
11.1.2.
11.1.3.
11.1.4
Produce and Forage
Milk.
Water, Sediment and Soil,
Conclusion..............
PCB.
11,2.1.
11.2.2.
11.2.3.
PCDD/PCDF.
Produce and Forage..
Milk, Sediment and Soil,
Conclusion * ,
11.4.
CONCLUSION.
REFERENCES >
11.3*1
11.3.2
11.3.3
SUMMARY.
Produce and Forage ,
Milk, Sediment and Soil,
Conclusion.
Page
10-6
10-9
11-1
11-3
11-3
11-21
11-23
11-36
11-36
11-36
11-41
11-44
11-44
11-50
11-50
11-52
11-53
12-1
13-1
IX
-------
LIST OF TABLES
Table Page
1-1 Emission Standards Allowed in Attended
Air Pollution Control Permit. "1-5
1-2 Source Characteristics of the Vicon MWC in
Rutland, Vermont..............;.... . ........... 1-7
2-1 Sampling Sites in Rutland, Vermont. 2-5
2-2 Equipment at the Ambient Air Monitoring
Sites in Rutland, Vermont 2-10
2-3 Ambient Air Analysis Analytical. Procedure
and Laboratory 2-14
2-4 Sampling Distribution for Environmental
Media 2-21
2-5 Method of Analysis:for Pollutants in
Environmental Media .' 2-24
3-1 The Sampling Period; Detection Limits and
the Number of Concentrations Detectable
for Each Pollutant 3-3
3-2 Occurrence of Detectable Pollutant Concentrations
in Ambient Air................................ 3-4
3-3 Proportionality Factors for PCDD/PCDF
Derived from Rutland, Vermont Ambient
Air Data 3-15
3-4 Toxic Equivalency Factors ,(TEFs) of
the Congeners of PCDD/PCDF.. 3-18
3-5 2,3,7,8-TCDD Equivalent Concentrations (pg/m3)
in Rutland, Vermont. *. .-.-'*.; 3-19
4-1 PCDD/PCDF, in Stack Emissions of Rutland
Incinerator (ng)..*............. ........ ...... 4-21-
4-2 Stack emission Rate of Metals (g/s)........... 4-22
'v • t
4-3 Dates Modeled Using SLAMS Meteorologic
Data and Associated Missing Data.............. • 4-24
4-4 Dates Modeled Using River Street Meteorologic
Data and Associated Missing Information....... 4-25
-------
LIST OF TABLES (cont.)
Table • Page
4-5 Predicted Concentrations at the 4 Monitoring
Sites and the Polar Receptor(s) With the
Greatest Concentration Based on Unit Emissions
(l g/s) and SLAMS Meteorologic Data 4-27
47-6 Predicted Concentrations at the 4 Monitoring
Sites and the Polar Receptor with the
Greatest Concentration Based on Unit
Emissions (1 g/s) and River Street Meteorologic
Data ...... 4-28
6-1 P-values and R-square Values for Regression
Analysis According to Site. 6-10
9-1 Occurrence of Maximum Detectable Concentration
in Ambient Air..-..;. 9-2
9-2 Ranks for the Four Sampling Sites Based on
Both Measured and Modeled Lead
Concentrations 9-5
9-3 Ranks for Three Sampling Sites,(Slams Excluded)
Based on Both Measured and Modeled Lead
Concentrations 9-8
9-4 Average Ranks of Lead Concentrations for
Four Sampling Sites 9-12
9-5 Average Ranks of Lead Concentrations for
Three Sampling Sites (Excluding Slams)........ 9-13
9-6 Ranks for Four Sampling Sites Based on Both
Measured and Modeled 2,3,7,8-TCDD
Equivalent Concentrations 9-17
9-7 Ranks for Four Sampling Sites Based on Both
. ; Measured and Modeled OCDD Concentrations...... 9-19
9-8 Average Ranks of 2,3,1,8-TCDD Equivalent
; Concentrations for Four Sampling Sites........ 9-21
9—9 Average Ranks of OCDD for Four Sampling
Sites. 9-22
'. . * - , '*
lOf-1 Results of Site-Specific ISCLT Modeling. ...... 10-4
XT
-------
LIST OF TABLES (cont.)
Table Page
10-2 Five-Highest Predicted Concentrations from
ISCLT for the Three Meteorologic Collection
Sites ................. ............. . .......... 10-7
10-3 The Highest Modeled Ambient Air Concentrations
for the Three Rutland Sites. . . . . . ............. 10-8
10-4 Maximum Predicted Annual -Average Concentration
and Analytical Limit of Detection for Each
Pollutant ....................... .............. 10-10
11-1 Metal Concentrations in Milk, Produce and Forage
October and November 1987 and June 1988 ....... 11-20
11-2 Concentration of Metals in Milk (p-g/g) ........ 11-24
11-3 Metal Concentrations in Water, Sediment and Soil
October and November 1987 and June 1988 ....... 11-25
11-4 Background Level Concentrations of Metals in
Water (Mg/L) .......... ................. ....... 11-27
11-5 Concentration of Metals in Soil (Atg/g) ....... 11-32
11-6 PCB Concentrations (Total) in Environmental
Media (X ± SD) (pg/g) ............. ............ 11-37
11-7 2,3,7,8-TCDD Equivalent Concentrations in
Environmental Media (X + SD) (pg/g) . .......... 11-45
11-8 Octachlorodibenzo-p-dioxin (OCDD)
Concentrations in Environmental Media ......... 11-49
xii
-------
LIST OF FIGURES
Figure Page
1-1 Location of Rutland/ Vermont ................... 1-3
1-2 Diagram of the Rutland Resource
Recovery Facility ... ...... ... .... 1-8
1-3 Summary of Rutland MWC Operations 1-10
2-1 Location of Monitoring Stations in Rutland,
Vermont. ,.'....' 2-3
2-2 Sampling Periods of the Rutland, Vermont Study.. 2-8
3-1 Ambient Air PCDD/PCDF Concentrations (pg/m5)
for the Duplicate Samples Collected at
Watkins Avenue 3-12
3-2 Approaches Used for Estimating 2,3,7,8-TCDD
Equivalent Concentrations 3-13
4-1 Bar Graphs of Monthly Rutland Wind Data from
January 1988 to June 1988 for SLAMS '. 4-4
4-2 Bar Graphs of Monthly Rutland Wind Data from
July 1988 to December 1988 for SLAMS ... ......... 4-5
4-3 Bar Graphs of Monthly Rutland Wind Data from
January 1989 to June 1989 for SLAMS 4-6
4-4 Bar Graphs of Monthly Rutland Wind Data for
July and August 1989 and all months for SLAMS .. 4-7
4-5 Bar Graphs of Monthly Rutland Wind Data from
January 1988 to June 1988 for Watkins Avenue ... 4-8
4-6 Bar Graphs of Monthly Rutland Wind Data from
July 1988 to December 1988 for Watkins Avenue... 4-9
4-7 Bar Graphs of Monthly Rutland Wind Data for
all montlis for Watkins Avenue 4-10
4-8 Bar Graphs of Monthly Rutland Wind Data from
January 1988 to June 1988 for River Street 4-11
4-9 Bar Graphs of Monthly Rutland Wind Data from
July 1988 to December 1988 for River Street .... 4-12
xlii
-------
LIST OF FIGURES (cont.)
Figure Page
4-10 Bar Graphs of Monthly Rutland Wind,Data from
January 1989 to June 1989 for River Street ..... 4-13
4-11 Bar Graphs of Monthly Rutland Wind Data for :
July and August 1989 and all months for
River Street • • • 4-14
4-12 Windrose for January 16, 1988 in Rutland, VT
based on the SLAMS meteorologic data ........... 4-30,
4-13 Windrose for January 28, 1988 in Rutland, VT
based on the SLAMS meteorologic data ............ 4-31
4-14 Windrose for February 21, 1988 in Rutland, VT .
based on the SLAMS meteorologic data ........... 4-32
4-15 Windrose for March 4, 1988 in Rutland, VT
based on the SLAMS meteorologic data ..... . .,.,.. . 4-33
4-16 Windrose for March 16, 1988 in Rutland, VT
based on the SLAMS meteorologic data ............ 4-34
4-17 Windrose for April 21, 1988 in Rutland, VT
based on the SLAMS meteorologic data 4-35
4-18 Windrose for May 3, 1988 in Rutland, VT
based on the SLAMS meteorologic data ........... 4-36
4-19 Windrose for May 27, 1988 in Rutland, VT
based on the SLAMS meteorologic data ....1...... 4-37
4-20 Windrose for June 8, 1988 in Rutland, VT ., ' :
based on the SLAMS meteorologic data ........... 4-38
4-21 Windrose for June 20, 1988 in Rutland, VT .
based on the SLAMS meteorologic data ............ 4-39'
4-22 Windrose for July 14, 1988 in Rutland, VT
based on the SLAMS meteorologic data ............ 4-4Q_,
4-23 Windrose for July 26, 1988 in Rutland, VT ,
based on the SLAMS meteorologic data .,...,..,''. .. 4-41
4-24 Windrose for August 7, 1988 in Rutland', . VT ,,, ,, ,,
based on the SLAMS meteorologic data ....... .?".. 4-42
4-25 Windrose for May 27, 1988 in Rutland, VT
based on River Street meteorologic data 4-43
xiv
-------
LIST OF FIGURES (cont.)
-.-',' f ,'• ' . ' '
Figure . Paqe
4-26 Windrose for June 20, 1988 in Rutland, VT
based on River Street meteorologic data 4-44
4-27 Windrose for July 14, 1988 in Rutland, VT
based on River Street meteorologic data•"... 4-45
4-28 Windrose for August 7, 1988 in Rutland, VT
based on River Street meteorologic data 4-46
4-29 Windrose for August 19, 1988 in Rutland, VT
based on River Street meteorologic data 4-47
6-1 Particulate Concentration (/ig/m3) and Amount of
Waste Burned per Day (tpd) on November 5, 1987
Through March 26, 1988 • • • 6-2
6-2 Particulate Concentration (jug/m3) and Amount
' Waste Burned per Day (tpd) on April 4 Through
, October 6, 1988 ....... -•- 6-3
6-3 Correlation Between PM-10 Particle
Concentration (jug/m3) at SLAMS and Amount
of Waste Burned (tpd) '..... 6-4
6-4 Correlation Between Particulate Concentration
(Mg/m3) of the Duplicate Sample Collected at
SLAMS and Amount of Waste Burned (tpd) .......... 6-5
6-5 Correlation Between Particle Concentration
(jug/m3) . and Amount of Waste Burned . 6-6
6-6 Correlation Between PM-10 Particle
Concentration (/Lsg/m3) at Route 4 and Amount
of Waste Burned (tpd) • • • • 6-7
6-6 Correlation Between PM-10 Particle
*''"'• Concentration (/ng/m ) at River Street and Amount
of Waste Burned (tpd) , . 6-8
7-1 Correlation Between PM-10 Particle Concentration
in Ambient Air (Mg/™3) and Indirect Mutagenic
Activity (revertants/m ) for Ambient Air Samples
Collected 11/17/87 to 3/16/88 7-2
7-2 Mutagen Concentration in Ambient Air Compared to
Tons of Waste Burned for the Sampling Period
11/17/87 to 3/16/88 7-4
xv
-------
LIST OF FIGURES (cont.)
Figure
8-1
8-2
8-3
8-4
8-5
8-6
8-7
Ambient
Ambient
1/16/88
Ambient
1/16/88
Ambient
1/16/88
Ambient
Ambient
2/21/88
Ambient
2/21/88
Air
Air
Air
Air
Air
Air
Air
Congener
Congener
Congener
Congener
Congener
Congener
Congener
Profiles
Profiles'
Profiles
Profiles
Profiles
Profiles
Profiles
for
for=
for
for
for
for
for
SLAMS,
River
Route
1/16/88.
St.,
4,
Wat-kin's, -
SLAMS,
River
Route
2/21/88.
St. ,
4,
Page
8-r3
8-4
8-5
8-6
8-7
8-8
8-9
8-8 Ambient Air Congener Profiles for Watkins,
2/21/88 8-10
8-9 Ambient Air Congener Profiles for River St.,
3/04/88 8-11
8-10 Ambient Air Congener -Profiles for Route 4,
3/04/88 8-12
8-11 Ambient Air Congener Profiles for Watkins,
3/04/88 8-13
8-12 Ambient Air Congener Profiles for SLAMS, 4/21/88..8-14
8-13 Ambient Air Congener Profiles for River St.,
4/21/88 8-15
8-14 Ambient Air Congener Profiles for Route 4, >
4/21/88 . ... . , 8-16
8-15 Ambient Air Congener Profiles for Watkins, '
4/21/88 •... i..'..• 8-17
8-16 Ambient Air Congener Profiles for River St.,
5/27/88 .... 8-18
8-17 Ambient Air Congener Profiles for Route 4, "
5/27/88 ,"•....;•. . .. 8-19
XVI
-------
LIST OF FIGURES (cont.)
Figure Page
8-18 Ambient Air Congener Profiles for Watkins,
5/27/88 ............>,.,......,... 8-20
8-19 Ambient Air Congener Profiles'for SLAMS, '6/20/88. 8-21
8-20 . Ambient Air Congener Profiles for River St.,
6/20/88 ». . .... , . . • » • 8-22
8-21 Ambient Air Congener Profiles for Watkins,
6/20/88 8-23
8-22 Ambient Air Congener Profiles for SLAMS, 7/26/88. 8-24
8-23 Ambient Air Congener Profiles for River St.,
7/26/88 . . '..;.., •• 8-25
8-24 Ambient Air Congener Profiles for Route 4,
7/26/88 8-26
8-25 Ambient Air Congener Profiles for Watkins,
7/26/88 8-27
8-26 Congener Profiles of Chimney Soot From Wood
Oven 8-29
8-27 Congener Profiles of MWC Stack Emissions Tested
on March 8, 1988 ^. . . . . 8-31
8-28 Congener Profiles of MWC Stack Emissions Tested
on March 9, 1988 ........... 8-32
8-29 Congener Profiles of MWC Stack Emissions Tested
on March 10, 1988 .... ... • • • • 8~33
11-1 Chromium Concentrations in Milk Samples in
- Rutland, Vermont ... • 11-4
11-2 Lead Concentrations in Milk Samples in
Rutland, Vermont 11-5
11-3 Chromium Concentrations in Water Samples in
Rutland, Vermont 11-6
11-4 Lead"Concentrations in Water Samples in
Rutland, Vermont 11-7
11-5 Arsenic Concentrations in Sediment Samples
in Rutland, Vermont ......;.. 11-8
xv 11
-------
LIST OF FIGURES (cont.)
Figure
11-6
11-7
11-8
11-9
11-10
11-11
11-12
11-13
11-14
11-15
11-16
11-17
11-18
11-19
11-20
11-21
Page
Beryllium Concentrations in Sediment Samples
in Rutland, Vermont .*........ . . 11-9
Chromium Concentrations iri': Sediment Samples ;
in Rutland, Vermont f 11-10
Lead Concentrations in Sediment Samples
in Rutland, Vermont n-n
Nickel Concentrations in Sediment Samples
in Rutland, Vermont ... 11-12
Arsenic Concentrations in Soil Samples
in Rutland, Vermont n-13
Beryllium Concentrations in Soil Samples
in Rutland, Vermont 11-14
Cadmium Concentrations in Soil Samples
in Rutland, Vermont . . . 11-15
Chromium Concentrations in Soil Samples
in Rutland, Vermont 11-16
Lead Concentrations in Soil Samples
in Rutland, Vermont . „'. 11-17
Mercury Concentrations in Soil Samples
in Rutland, Vermont ; 11-18
Nickel Concentrations in Soil Samples
in Rutland, Vermont ; 11-19
PCB Concentrations in Milk Samples
in Rutland, Vermont 11-38
Total PCB Concentrations in Sediment
Samples in Rutland, Vermont 11-39
PCB Concentrations in Soil Samples
in Rutland, Vermont 11-40
TCDD Equivalent Concentrations in Milk
Samples in Rutland, Vermont 11-46
TCDD Equivalent Concentrations in Sediment
Samples in Rutland, Vermont 11-47
xvm
-------
Figure
LIST OF FIGURES (cont.)
Page
11-22 TCDD Equivalent Concentrations in Soil
Samples in Rutland, Vermont .., .. * .• • 11-48
xix
-------
LIST OF ABBREVIATIONS
AA
acfm
ANOVA
As
B[a]P
Be
Cd
CDD
CDF
Cr
DMSO
ECL
EOM
ESP
F.G.R.
fps
GC-ECD
GPPAA
Hg
Hi-Vol
HpCDD
HpCDF
HRGC
Direct aspiration atomic absorption
spectrometry
Atmoshperic cubic feet per minute
Analysis of variance
Arsenic
Benzo[a]pyrene
Beryllium
Cadmium,
Chlorinated dibenzo-p-dioxin
Chlorinated dibenzofuran
Chromium
Dimethylsulfoxide :
Environmental Chemistry Laboratory
Extractable organic mass
Electrostatic preeiptator
Flue Gas Return
Feet per second
Gas chromatography with electron capture
detection
Graphite furnace atomic absorption
spectrometry
Mercury
High-volume
Heptachlorinated dibenzo-p-dioxin
Heptachlorinated dibenzofuran
High resolution gas chromatography
xx
-------
LIST OF ABBREVIATIONS (cont.)
HRMS
HxCDD
HxCDF
h/yr
ICP-AES
ISCST
Me
Mo
MLD
m.S.1.
MWC
NAA
Ni
OCDD
OCDF
Pb
PGB
PCDD
PCDF
PeCDD
PeCDF
PM-10
PS-1
PUF
High resolution mass spectrometry
Hexachlorinated dibenzo-p-dioxin
Hexachlorinated dibenzofuran
Hours per year
Inductively coupled plasma-atomic emission
spectrometry
Industrial Source Complex Short-Term
Measured
Modeled
Minimal limits of detection
Mean sea level
Municipal waste combustor
Neutron activation analysis_
Nickel
Octachlorinated dibenzo-p-dioxin
Octachlorinated dibenzofuran
Lead
Polychlorinated biphenyls
Polychlorinated dibenzo-p-dioxin
Polychlorinated dibenzofuran
Pentachlorinated dibenzo-p-dioxin
Pentachlorinated dibenzofuran
Particulate matter < 10 /j,
Particulate sampler
Polyurethane foam
xxi
-------
SD
SLAMS
TCDD
TCDF
TEF
TLC
tpd
TSP
UTM
VAPCD
LIST OF ABBREVIATIONS (cont.)
Standard deviation
State and Local Air Monitoring Station
Tetrachlorinated dibenzo-p-dioxin
Tetrachlorinated dibenzofuran
Toxic Equivalency Factor
Thin-layer chromatography
Tons per day
Total Suspended Particulate
Universal Transverse Mercator
Vermont Air Pollution Control Division
xxn
-------
EXECUTIVE SUMMARY
This report describes a multipollutant, multimedia study
designed to determine levels of contaminants in the ambient air,
soil, sediment, water, and agricultural products (carrots,
potatoes, milk, and grass hay) surrounding a municipal waste
combustor (MWC) in Rutland, Vermont. The study was initiated to
provide a preliminary determination of human exposure resulting
from the MWC emissions. The study procedures and analytical
results are detailed for samples collected between October 1987 and
, February 1989.
The levels of selected pollutants were measured in the ambient
-n .
air and environmental media at or near predicted sites of maximum
deposition surrounding the MWC. Air dispersion modeling of ittdfc
emissions from the MWC prior to its operation was conducted to
select appropriate locations to place ambient air monitors and to
collect .environmental media samples. As a result, a four-statioft
ambient air monitoring network was established for collection of
samples to measure ground-level ambient air concentrations of
pollutants from the incinerator emissions. The monitors were
placed at Watkins Avenue, River Street, Route 4, and the Rutland,
Vermont State and Local Air Monitoring Station (SLAMS).
Ambient air samples were analyzed for the following
pollutants: arsenic and chromium (by neutron activation analysis);
beryllium, cadmium, lead, and nickel (by Inductively Coupled
Plasma-Atomic Emission Spectrometry); • mercury (by pyrolyzer-
xxiii
-------
dosimeter) ;. benzo(a)p;yrene '(by thin-layer chromatography) ; PCBs (by
gas chromatography with electron capture detection); and PCpp/PCDFs
(by high resolution , gas chromatography-high resolution mass
»••-., , '~j; >;.*.-.. .••-,/,.,., :,,i,y, . ;•„ Ci..^ .'.."•..•'. '• -. ,• ... •••'.• '• •,'.' • , .'."'• f:'-. \.-,l.
spectrometry). Particulaties, were^ examined for mutagenic activity
by the reverse mutation assay
Wind speed, wind direction, temperature, relative humidity,
and solar radiation.were continuously monitored and recorded at
three sites: SLAMS, River Street, and Watkins Avenue. Rainfall
intensity and atmospheric pressure were also collected at the
SLAMS.
Environmental media samples, except water, were analyzed for
the following pollutants; arsenic (by graphite furnace.atomic
absorption spectrometry); beryllium, cadmium, chromium, lead, and
nickel (by direct aspiration atomic absorption spectrometry);
mercury (by the cold vapor technique, of direct aspiration atomic
absorption spectrometry); and PCBs and PCDD/PCDFs (by, high
resolution gas chromatography-high, resolution mass spectrometry).
Water samples were analyzed for the following pollutants:,
arsenic and beryllium (by graphite furnace, atomic .absorption ,
spectrometry); cadmium, chromium, lead,.and .nickel (by .direct s t
aspiration atomic absorption spectrometry) ; and mercury (by the ,,....
cold vapor technique of direct aspiration atomic .absorption
spectrometry).
** •* ' •• " .-• ' _• .......... .- .. _ ,- _: . .-. ',"-' .-,;• - . - - -.'".< .-_ _j ;,_.-_ j.l f. trf •*£>{';;; ?> £
Most metals were measured above the detection limit in only,,,..,,,,
a few ambient air. samples^. Arsenic was, measured ,. above,, its,,, ,
•i. . -. . " „ ., -.-... -.,'.,- -iv •>-•-, .,, . S • t t J, > », •'" ', „ , „ ., ..- : V ,,in-^ ',- ' -ti •r.J'. .'., ', '.I/ K^Ji
detection limit of 0.0046-0.0047 jug/fti3 in 7 of 98 samples.
xxiv
-------
Beryllium was ^measured' above It s:'detect Ion limit of 't).*224"3' 'ng/m3
in 4 of 122 samples. Cadmium wai measured afebve its detection,
limit of 0.0009-0.0014 /Ltg/ms iii ;2'of'i^2 samples. Chromium was
measured above its detection limit'of'~Q.QQ65*-Q'.0069 Mg/m3 in 1 of
98 samples. Lead was measured above its detection limit of 0.0061
iug/m3 in 108 of 122 samples. Nickel was measured above its
detection limit of 0.0038-0.0077 /lg'/m3 in 3 of 122 samples.
Benzo(a)pyrene was measured above its detection limit of 0.3348 ,
ng/m3 in 43 of 131 samples. No ~.PCBs were measured above the
detection limit of 0.7-0.8 ng/m3 in any samples collected.
Total congener and specific 2", 3,7,8-chlorine substituted
isomeric concentrations in ambient air samples were determined.
When the reported concentration of a 2,3,7,^-substituted isomer in
a particular homologous series was nondetectable, the concentration
was assumed to be a proportion-of the total isomeric concentration
of the homologues in. the series."' "For example, if the 2,3,7,8-TCDD
concentration emitted from the incinerator was approximately 5% of
the total emitted TCDD concentration, a proportionality constant
of 0.05 was used to estimate the concentration of 2,3,7,8-TCDD in
that air sample. The proportionality factors were determined from '
actual samples. " ' ! ~
Once 'the proportion of each 2,3,7,8-chlorine substituted
isomer was estimated, the concentrations were converted to 2,3,7,8-
TCDD 6quivalehfes using TEFs. Total 2,3,7,8-TCDD equivalent
concentratibhs in ambient air samples ranged frbm 0.011-5.39 pg/m3i.
xxv
-------
The Industrial. Source Complex Short-Term (ISCST} model was
run, using Rutland meteorologic data, to predict the ground-level
awbi«nt air concentrations of pollutants in Rutland for the same
"day* on which the ambient air was sampled at the four monitoring
§it«s. The goal of the modeling procedure was to predict the 24-
hour average ambient air concentrations,at each monitoring site
for each sampling day assuming one unit emission. This would
enable these concentrations to be used later for the comparison of
th« measured and predicted concentrations.
The concentrations predicted to occur at the monitoring sites,
assuming one unit emission, ranged from 0-5.22 /jg/m3, using
meteorologic data from the SLAMS, and 0-4.782 jug/m3, using
meteorologic data from River Street.
Analysis of the incinerator as a source for the measured
' . ,-'•.,' ' " ' • f. . • - " '
pollutants in ambient air encompassed four approaches: (1) the
daily tons of waste burned in the MWC were compared to measured
particulate matter (PM-10) concentrations, (2) mutagenic activity
compared to PM-10 concentrations and tons of waste burned, (3)
congener profiles of measured PCDD/PCDF in Rutland ambient air
compared to those of potential sources, and (4) daily ambient
air concentrations of pollutants that were predicted from air
di§p«rsion modeling were compared to the measured pollutant
concentrations.
The approach for the analysis of environmental media was
qualitative, comparing concentrations between the various sampling
xxyr .
-------
periods and comparing pollutant concentrations detected in Rutland
with those described for other geographical regions.
The first approach to assessing the contribution of the MWC
emissions to the pollutant concentration in Rutland ambient air
was to attempt to correlate the amount of waste burned by the
incinerator each day with the particulate matter (PM-10 fraction)
concentrations. A correlation between tons of waste burned and.
PM-10 concentration would suggest that the MWC was the primary
source of pollutants in the air. No correlation by regression
analysis between the amount of waste burned daily and ambient air
particulate concentration at any of the sites was found to exist.
This suggests that the MWC .is not the sole source of particulates
in the Rutland ambient air.
The reverse mutation assay was used to determine the levels
of mutagenic activity associated with particles from ambient air
collected surrounding the Rutland MWC. A positive correlation
between particle concentration and mutagenic activity was observed
v-, fr , . „ :.-...» •• , ...':•-,.'•*..-
at all four sampling sites. There was, however, no correlation
between the number of'tons of waste burned and mutagenic activity
..'•'', , . •'•• : ' ' -.'••' >•-".)-':' . - "'•,.".•" ' ' 'I ..
at any of the sites. This suggests that other sources are
responsible for particles in ambient air that induce mutagenic
activity in Rutland. .
The PCDD/PCDF congener concentrations of the ambient air
samples were used to make graphic displays of the distribution
patterns of the homologues. The purpose of the congener profiles
was to compare the pattern of the PCDD/PCDF congeners in 'the
xxv
-------
samples with the patterns of congeners from potential sources.
The PCDD/PCDF concentrations and distribution patterns for the same
day/ and also on different days, differed among monitoring sites.,
indicating that local sources (i.e., sources very close to each
monitoring site) influence • the concentrations and distribution
patterns at each site. The PCDD/PCDF concentrations and
distribution patterns of homologues vary between days and different
sampling intervals, suggesting that PCDD/PCDF sources may change
with time.
The congener profiles of ambient air were compared to the
congener profiles of the stack emission from the MWC and chimney
soot. In general, the congener profiles of the ambient air samples
collected on two winter days do not resemble those of chimney soot.
Congener profiles were developed for the MWC stack emissions
measured on three days by the MWC contract laboratory. The stack
testing was performed on different days than the ambient air
sampling. The profiles of stack emission have similar PCDD/PCDF
distribution patterns. When the congener profiles of the ambient
air collected at one specific site are compared to the profiles of
the stack emissions, the PCDF congener patterns show a resemblance,
but the PCDD congener patterns do not. In general, the ambient air
samples have higher HxCDD and OCDD relative percentages than the
^.
stack emissions.
xxviii
-------
Because of the variations detected in concentrations and
congener profiles between sites, days, and weeks, it is unlikely
that the PCDD/PCDFs were from wood burning or the MWC alone, but
from a variety of sources.
The pollutant concentrations measured in Rutland ambient air
when the incinerator was in operation represented the total
concentration of each pollutant from both the incinerator and other
sources. In order to determine if the concentrations of measured
pollutants were primarily from the MWC, the proportion of the
pollutants attributable to other sources needed to be assessed.
Since an inventory of other sources for the measured pollutants was
not available, source apportionment was assessed1 by statistically
comparing measured and predicted ambient air concentrations.
Lead concentrations were compared using two nonparametric
methods* the modified sign test and the Friedman nonparametric
ANOVA. From the modified sign tests'it was determined that there
was no evidence for a correlation between the measured lead
concentrations and the lead concentrations predicted by the
dispersion model. From the Friedman nonparametric ANOVA tests, it
was determined that the pattern of lead concentrations (highest to
lowest concentration) differed between the modeled and measured
concentrations.
The statistical comparison of the measured and modeled
concentrations of PCDD/PCDFs involved , the conversion of the
PCDD/PCDF isomer concentrations to 2,3,7,8-TCDD equivalents. As
with lead, PCDD/PCDF concentrations were compared using the
xxix
-------
modified sign test and the Friedman nonparametric, ANOVA. The
analyses were performed for both the 2,3,7,8-TCDD equivalent
concentrations and the OCDD concentrations. The modified sign test
using OCDD indicated no correlation between measured and predicted
OCDD concentrations. The results of the Friedman analyses using
either the 2,3,7,8-TCDD equivalent or the OCDD concentrations
indicate that there is no statistically significant difference in
the measured or modeled concentrations between the four ambient air
monitoring sites.
The statistical analyses of the measured and predicted lead
and PCDD/PCDF data suggest that there are other sources
contributing to these measured levels and that the MWC was not the
primary source of the pollutants.
Additional air dispersion modeling was performed to predict
annual-average concentrations. Using site-specific Rutland data,
the ISCLT results confirmed the initial'modeling efforts used to
locate the ambient air monitoring sites. Assuming the maximum
stack emission rates of the 3 stack testing runs, the.majority of
the pollutant levels attributable to the MWC (with the exceptions
of PCDD/PCDFs and lead) may not be measurable using the current
analytical techniques. The predicted concentrations of some
pollutants were orders of magnitude less than the analytical limit
of detection. Consequently, the pollutant ambient air
concentrations emitted by the MWC generally could not have been
measured.. -
XXX-
-------
, , Concentrations of .arsenic^ beryllium, chromium, lead, mercury,
and nickel in both produce ancl, fprage; were nondetectable. The mean
.concentration p.f. cadmium,, which was detectable in produce, was 0.2
.. and 0.3, ,mg/kg in October and November 1987., respectively. The
cpncentration of ,cadmium in forage.was detectable (0.1 mg/kg) in
one of two samples in November ,1.9,87. and was nondetectable in all
other produce and forage samples for both sampling rounds.
, , Concentrations, of beryllium ^,in milk were nondetectable for
all sampling periods and sites. Chromium and. lead concentrations
„ were found in milk in measurable quantities at several sites in
October and November 1987, but were below the detection limit
during the incinerator's operational period (June, 1988).
Water concentrations of arsenic, beryllium, and nickel were
nondetectable at all sites for al.]L sampling periods. Cadmium and
t .merqury concentrations in water were detectable at one site during
one sampling period, but the measured cpncentration was equal to
,t:he detection -limit. Arsenic, . beryllium, cadmium, and nickel
,(^cpncentratipns in water were at,pr,e:qual to the detection limits.
^.Chrpmium, and lead .concentrations in water, exceeded the detection
, ...limit .in several samples .collected ,in. the sampling periods when the
, .incinerator was, pre-operational, (October and November 1987).
.,,->-• .,>-. ,A11 metals except, cadmium an4 mercury were found to be present
in sediment in detectable concentrations. Only one sample each of
cadmium and mercury were detectable.
xxxi
-------
Overall, these results indicate that there were no apparent
increases in metal concentrations in the environmental media during
the period when the Rutland MWC was operational relative to the
period prior to combustor operation.
The concentrations of PCB in the produce and forage ranged
from 1.86xl03 (carrot) to 6.18xi03 (potato) pg/g. The produce PCB
concentrations in Rutland are similar to those found elsewhere.
The results of the milk, sediment, and soil sample analyses do not
indicate that PCB concentrations in these environmental media have
increased because of deposition of PCBs from the stack emissions,
but indicate the concentrations are similar to those found
elsewhere.
The effect of incinerator emissions on total PCB
concentrations in forage and produce could not be determined, since
these media were only sampled prior to MWC operations. No
difference in total PCB concentrations was found in milk, sediment,
or soil sampled both before and during incinerator emissions.
Most of the 2,3,7,8-TCDD equivalent average concentrations
were derived from values that .were nondetectable but were
conservatively set equal to the detection limit. The average
2,3,7,8-TCDD equivalent concentrations in the produce and forage
ranged from 4.88-11.1 pg/g.
The majority of PCDD/PCDF isomer concentrations in milk,
sediment, and soil were non-detectable, and were set equal to the
detection limit for the purpose of calculating average 2,3,7,8-
TCDD equivalent concentrations.
xxxii
-------
Since samples of forage and produce were only collected prior
to commencement of operations of the MWC, it was not possible to
determine whether concentrations of PCDD/PCDFs in these media were
altered because of combustor emissions. In samples of milk,
sediment, and soil, there were no statistically significant
increases in 2,3,7,8-TCDD. equivalent concentrations in samples
collected after commencement of operations of the MWC, when
compared to samples taken prior to operation.
"The measured concentrations of metals, PCB, or PCDD/PCDF in
produce, forage, milk, soil, sediments, or water (metals only) are
within the range of background concentrations found in other
geographical areas.
The objective of this study was to determine if there were
human health risks attributable to the operation of this
incinerator. This objective could not be attained because the
majority of pollutants in the ambient air and environmental media
were not present in concentrations that could be detected by the
analytical, methods employed. This made a direct determination of
the contribution of the incinerator to the measurable concentration
of pollutants not possible. Therefore, an analysis of the
likelihood that the incinerator was a primary contributor to the
measured pollutant concentrations was assessed using several
alternative approaches.
The conclusion reached by evaluation of the collected field
samples is that the measured concentrations of the pollutants in
the ambient air and environmental media cannot be correlated with
xxxm
-------
the emissions or operation of the MWC. The MWC does not appear to
be the primary source of these pollutants. Evidence for this
conclusion comes from both qualitative and quantitative evaluation
of the measured pollutant concentrations in the .ambient air and
environmental media, as well as .comparison with predicted ambient
air concentrations of the pollutants using local meteorologic
information. e :, _„ .V,.,.J:- -.; ,,'- •.. ;;, r -. ; • „ .•;.;?.••,
While this field, study did nqt show that the MWC was a primary
contributor to the measured levels of pollutants, .the results
contain information about the background levels ,of pollutants,and
the contribution of other sources to,the, Rutland, Vermont area.
Contained in the accompanying appendices is information
relevant to this pilot study. The Quality Assurance/Quality
Control Plans, the analytical results, the environmental modeling
and the statistical analyses are.presented. . .
xxx TV
-------
1. INTRODUCTION
1.1. PROJECT OBJECTIVE
This report describes a multipollutant, multimedia study
designed to determine levels of contaminants in the ambient air,
soil, sediment, water and agricultural products surrounding a
municipal waste combustor (MWC), The project, coordinated by the
Environmental Criteria and Assessment Office in Cincinnati (U.S.
EPA, Office of Research and Development, Office of Health and
Environmental Assessment), was initiated to provide a preliminary
determination of human exposure resulting from MWC emissions for
use by Agency personnel.
The U.S. EPA entered into a cooperative agreement with the
State of Vermont to perform environmental monitoring of the MWC at
Rutland, Vermont (Vermont Air Pollution Control Division, Agency
of Natural Resources, 1987a). Although similar studies have been
conducted in Europe (i.e., Yasuhara et al., 1987; Morita et al.,
1987), this was one of the first multipollutant, multimedia
investigations of municipal waste combustion in the United States,
In the past, other field investigations of pollutants emitted from
MWCs have primarily focused on quantifying one or a few classes of
chemicals (e.g., polychlorinated dibenzo-p-dioxins and
dibenzofurans, or metals) in a few environmental samples (e.g.,
air, milk or soil). This study measured pollutants in ambient air
and various environmental media so that indirect routes of exposure
1-1
-------
in addition to the direct inhalation route (U.S. EPA, 1987a) could
be considered. This study may also serve as a protocol for future
multipollutant, multimedia field assessments of other MWCs.
This report details the study procedures and analytical
results for samples collected between October 1987 and February
1989. An assessment of whether the measured concentrations in the
environmental samples can be attributed to the MWC is presented.
The report summarizes the uncertainties associated with the study
design and collection and analysis of the data, and discusses the
implications of these uncertainties in the interpretation of the
data. Several issues that complicate the use of these data are
also discussed.
1.2. THE RUTLAND RESOURCE RECOVERY FACILITY
The Rutland Resource Recovery Facility is located in Rutland,
Vermont, a city with a population of approximately 18,000 (Figure
1-1). Rutland has an average yearly temperature of 46.3°F.
Rutland is situated in west-central Vermont in Rutland County. The
town is in a mountain valley, with ridges to the east and west
rising over 1000 feet above the valley floor. Hills rising to over
1000 ft. mean sea level (m.s.l.) are present to the immediate
north-northwest and south-southwest. Elevations over 2000 ft.
m.s.l. are found 7 km to the east. The seasonal rainfall for
Rutland is 33.62 inches and the seasonal snowfall is 62.8 inches.
1-2
-------
CANADA
NEW YORK
NEW HAMPSHIRE
Interstate Highway
— US Highway
CHy Population
A Less thir. 15,000
O lS.OOl-Zi.KX3
D 25.001-SO.OOO
• S0.001OOO.OOO
• Mote tha-. '100.000
JO 30
MASSACHUSETTS
Figure 1-1. Location of Rutland, Vermont
1-3
-------
Rutland is designated as an attainment area and the area within a
40 km radius of the facility is designated as either attainment or
unclassified for all criteria pollutants (Agency of Environmental
Conservation, State of Vermont, 1986).
In accordance with Vermont Air Pollution Control Division
(VAPCD) Regulations, a permit was issued to the Rutland Resource
Recovery Facility, manufactured and operated by Vicon Resource
Recovery Systems (Butler, New Jersey), on March 20, 1984. The
permit was reopened due to concerns over dioxin and acid gas
emissions. The incinerator was redesigned to include additional
pollution control equipment, which changed the stack parameters.
An amended permit was issued on September 11, 1986. Table 1-1
lists the emission standards allowed under the amended air
pollution controltpermit. •
The facility is ~2 km west of the downtown center of Rutland
on a site bounded on the north by U.S. Route 4 and on the south by
Otter Creek. It is located on flat terrain at an elevation of 554
ft. m.s.l. The MWC consists of two mass-fired incinerators, each
consisting of a refractory lined furnace and a separate waste heat,
boiler (modular burners) (Vermont Air Pollution Control Division,
Agency of Natural Resources, 1985). Each of the two incinerators
at the facility is limited to its maximum design capacity of 120
tons of municipal solid waste per day (total of 240 tpd) and the
entire facility cannot combust more than 80,000 tons per year of
refuse based on a 91% availability factor (Agency of Environmental
1-4
-------
- • • Emission Standards Allowed in •' •'
Amended Air Pollution Control Permit
Pollutants • •'•• ^
Particulate Matter
Sulfur Dioxide
Nitrogen Oxides
Carbon Monoxide
Lead • * ' •.'-•-'.-' . ' • ••-•'•* -•
Sulfuric Acid
Total Fluorides ".i .. ' i^;".: -'.:-" •-'':• \
0.018
0.0006 •''•"••
4.6
,• . --:, .j' " •-,,•"« ; .-. •••
Significant ; '
Emissions
(Tons/ Year )b
„ :', •• '. 25 ' ; ; " '-
40 .
; . •; " '..- --40' -• '•"'••• ! •'•
50
:" "" •:.' " ',/'.\ o:i'6;:' '-"""•••
7
.•..-; ' •:."•:..• . • 3 ... • -\ "•* '
0.1
' 0.0004 ^ -'
40
, " , «. • 1 . • • .• . • • . ; .. .•- '-'
aBased on two incinerator units operating at their maximum rated-
capacity, a 91% availability factor (iie. , 80,000' -tons-'-per ydar
of refuse) and meeting the limits prescribed in the permit.
Significant emission rate as promulgated by the Agency for
Environmental Conservation,' State of Vermont •''• !- ' :.-?:•
Source:
1986.
Agency of Environmental Conservation, State of Vermont,
1-5
-------
Conservation, State of Vermont, 1986). A summary of the source
characteristics is presented in Table 1-2 and a diagram of the
facility is presented in Figure 1-2.
The Rutland facility is designed such that solid waste is
• ' ' ,
dumped into an enclosed tipping floor having a storage capacity of
400 tons (Agency of Environmental Conservation, State of Vermont,
1984). The refuse is transferred from the tipping floor, to a
loader, and then to the feed hoppers. From the feed hoppers, the
solid waste enters the furnace by means of a hydraulic ram that
pushes it into the primary combustion chamber. The burning waste
travels through the furnace down a series of fixed refractory
hearths. The hot gases from the primary combustion chamber enter
a secondary combustion chamber where combustion is completed, and
then pass through a tertiary (mixing) chamber. The gas is passed
through the boilers, producing superheated process steam, and
through an economizer that preheats boiler feedwater. Gases
exiting the economizers enter an electrostatic precipitator (ESP,
one unit per furnace) for the removal of particulate matter, pass
through a condensing heat exchanger, and finally pass through wet
scrubbers for the removal of acid gases prior to release to the
atmosphere. Emissions from the two units are vented to separate
flues within the same 50-rmeter high stack. The steam produced in
the waste heat boilers is used to generate electricity. Although
no auxiliary fuel is required to maintain the flame in the mass fed
furnaces, each furnace has an auxiliary burner capable of burning
1-6
-------
TABLE 1-2
Source Characteristics of the Vicon MWC in Rutland, Vermont
Source location: 4,829,700 m north, UTM
. t '* *' ' :.."=, - •'" '' '• -''-'; '/ ''""'•£•?"(' •' ' •;{*-.' /' ; ' *''''• ° ~ - ::: •"-* - "'- >:; . , *£ .',"'*•. ' -
661,700 m east, UTM
Source elevation: 554 ft msl (169 m)
Stack height: 165 ft (50.3 m)
Stack diameter: 3.4 ft (1.04 m)
Exhaust temperature: 130"F (327,6 K)
Exhaust velocity: 50 fps (15.24 m/s)
Exhaust flow: 27,566 acfm (13.0 m/s)
Cross-sectional area of structure:
Building height: 36 ft (11.0 m)
Building length: 240 ft (73.2 m)
Building width: 160 ft (48.8 m)
Emission factor: Unity factors (i.o g/s)
Particulate size distribution: Assumed gaseous
Number of stacks: One (two flues in one stack)
Number of incineration units:, Two (mass burn)
Daily capacity of each unit: 120 tpd
Expected operational time: 8,000 h/yr (modeled at 8,760 h/yr)
Control equipment: Four-field ESP followed by condensing heat
exchanger followed by wet scrubber (packed
tower type)
Source: Vermont Air Pollution Control Division, Agency of Natural
Resources, 1985. ,,;;; ,
1-7
-------
R
HIGH PRESS.
-0 ™RBINE HIGH PRESSURE S
5
^
^
j , '
^ ^\
,v^
2
Oi
u.
\
— \
5
"ij
t/>
U
as
OL
c
\
S
n
i
p
CONDENSER^
LOV PRESSURE STEAM ^^
ii ii ^
••^
i L
1 '
V
r~X
EATERS K
r, r
UJ
K
/
I
s
\ \
ECONOMIZER
NO. Z
a-
UI
_j
S
m
/
/
^
CONTROL
»->
T
T
CV1
c5
«.
/
c
I
?
1
c
•^
II
*
^
ECONOMIZER
NO. I
cp-
j
j
3
a
— r
-T
1 — • — • — r
aBHHVHD
AavilJJ31
1
tJOOIJ DNIddlJL
D
z
»—
V
^
V r
\
M
K
Ul
H.
< --
3
H
b i
, ,, ,,, -, ,--
,£ CXJ ^
s°;L~"
uz /
( - . .' • ., ^
OL -"
S^3*" >;
^2 /
fc
H SECONDARY^
CHAMBERS/
f
1
j(
L^-T;
TP^
siy
BUSTOR
a:
a
o
cu
Rfflgj
nrvm
ii 1 1 1
5 ,
(L
1
LOADERS
I
Is
J t;
. -i
X
.. c
(.
. . c
1, 1
«*%
§«£
• 1 -^*- i — i
1 • -*•
-QJ2
: >-. gg"2
1
F.G.R. DUCT-^
^
SNVJ 'iCD'J
3«iJa3aNn
Q SECONDARY^
"CHAMBERSy
\±J
mag^
nfwn
1 1 11
B-
1 LOADERS
/
-fp ,
// I1' '' •' '
3VERRRE F.GJ?. 1
- OVERHRE
AIR. FAN
IS
Til,
W -2
gS|
j
: ~dz
: g££
|l
1
C£
a
.,->-
S ijJ
(/> ^
< -^
x
a
-'.. U,
u-
•H
H
•H
O
(0
t
Q)
O
Q)
Q),
O
O
(0
0)
c
nJ
rH
•H
Q
CM
I
rH
Q)
.1-8
-------
natural gas or oil. The auxiliary burner is designed for start-
up use and load stabilization (Agency of Environmental
Conservation, State of Vermont, 1984).
The incinerator began burning solid waste in November 1987
and continued operating until August 1988. The facility spent a
significant amount of time either shut down or operating at half
capacity (see Figure 1-3) (Fitzgerald, 1990). The incinerator was
shut down on December 13-23, 1987; January 3*12, 1988; and April
8-11 and 21, 1988. In addition, the facility was operating at half
capacity (only one unit operating) on November 5 and 17, 1987;
December 11, 1987; January 18 to February 7, 1988;, February 13-
21, 1988; and April 5-7 and 22, 1988.
1.3. STUDY APPROACH , , r
In order to accomplish the objective of this project, levels
of selected pollutants were measured in environmental media before
the Rutland MWC began operating and in both ambient air and
environmental media after the MWC began operating. The VAPCD
identified several pollutants to be monitored in ambient air during
this project:
Arsenic (As)
Beryllium (Be)
Cadmium (Cd)
Chromium (Cr)
Lead (Pb)
Mercury (Hg)
Nickel (Ni.)
Benzo[a]pyrene (B[a]P)
Pplychlorinated dibenzodioxin (PCDD)
Polychlorinated dibenzofuran (PCDF)
Polychlorinated biphehyl (PCS)
Mutagenic Orgariics
1-9
-------
I-l
0
OS
a
R
S
ir
s
n
f\
c\
fH
(N
O
O»
cn
H
CO
H
^
r-t
\0
I
in
«r
rH
n
"~*
C4
1-1
^
O
c\
CO
t^
KO
tn
«r
n
«
r-i
i-i
X
1
I
0 1 1 |
01 | |
0 II |
0 1 1 |
O 1 1 1
O 1 1 |
01 II
0 1 1 |
0 1 | |
Olio
00 IX
O O 1 1
0011
0011
1 0 1 |
1 O II
1011
•1 0 1 1
1 O 1 1
XI 1 |
XI IX
X 1 1 X
X 1 1 X
XI IX
X 0 1 0
X 0 1 0
x o i o
X 0 1 |
x o ii
01 1
0 1 1
CO CO CO CO
CO CO CO CO
o o o o
t!l
c
•H
M
M
•H
n
10
c
o
c 'ra
0 0
•H «H
<0 *H
01 W
§• 5
o
c
s ^
§ -g
u)
to C 1
^J
(d
to
*
1
H
a)
OS ._!
o\ .P
H pL(
1— I
tf
0)
F
•H
s
g
U)
1-10
-------
Except for benzo[a]pyrene and mutagenic organics, the above
pollutants were also measured in soil, water, sediment and
agricultural products.
The location of the sampling sites was determined by air
dispersion modeling prior to commencement of operation of the
incinerator. The MWC was also required to stack test for a number
of pollutants after incineration commenced (Agency of Environmental
Conservation, State of Vermont, 1986). The results of these stack
tests were used in addition to air dispersion modeling for
examination of the contribution of the MWC to ambient air
concentration of pollutants. The results of both air dispersion
modeling and the stack testing are presented in Chapter 4.
The VAPCD, the Vermont Water Quality Laboratory of the Vermont
Water Quality Control Division and U.S. EPA laboratories (Office
of Modeling, Monitoring Systems and Quality Assurance,
Environmental Chemistry Laboratory, Health Effects Research
Laboratory) were responsible for collection and analysis of the
contaminants. The actual sampling and analyses used for ambient
air and environmental media are discussed in Chapter 2, while the
analytical results are summarized in Chapter 3. Chapter 5 presents
approaches used to determine the contribution of the MWC to these
measured concentrations found in ambient air and other
environmental media. Chapters 6 through 9 present the results of
the analyses used for the determination of attribution of the MWC
to the pollutants in the ambient air. Chapter 10 presents
1-11
-------
additional air dispersion modeling to determine the magnitude of
the long-term ambient air concentration in Rutland. Chapter 11
focuses on the results of the analyses performed on the
environmental media concentrations. The report is concluded with
a summary of the findings and a discussion of the lessons learned
for completing a multimedia, multipollutant field assessment of a
MWC.
1-12
-------
2. SITE SELECTION, SAMPLING AND ANALYSIS
The levels of selected pollutants were measured in the ambient
air, soil-, water, sediment, produce arid forage'samples at or near
predicted sites of maximum deposition surrounding the Rutland MWC ;•
This chapter summarizes the ambient-air model used to predict the
sites of maximum deposition of pollutants. The sampling techniques
and analytical methods used for quantifying each pollutant in these
environmental media are also detailed.
2.1. AIR DISPERSION MODELING FOR SELECTION OF MONITORING SITES
Air dispersion modeling analysis of normalized (i.e., unit
emissions of 1 g/s) stack emissions from the MWC in Rutland,
Vermont was conducted to select appropriate locations for placement
of ambient air monitors to measure groundrlevel ambient air
concentration of pollutants due to the incinerator emissions.
These dispersion models considered source characteristics, terrain,
meteorologic data and receptor location. Both the UNAMAP 6 version
of the Industrial Source Complex Long-Term (ISCLT) Model (U.S. EPA,
1986a) and the LONGZ Model (U.S. EPA, 1982a) were used to predict
long-term average annual air concentrations of pollutants in the
vicinity of the MWC. Both models were run using polar grid
receptors as well as discrete individual receptors. Maximum annual
average ground-level concentrations at receptor sites were
estimated for 16 wind directions beginning with north and spaced
2-1
-------
every 22.5° along the polar azimuth and at radial distances of 0.2,
0.5, 1.0, 2.0, 5.0, 10, 20, 30, 40 arid 50 km from the MWC for a
total of 160 receptors. In addition, a total of 59 discrete
receptors were sited. These discrete receptors were placed at
*'- "" • ' • , . "''•"• ' -'.
points clustered around points of maximum concentration as
predicted by the polar grid model. A few discrete receptors were
also placed at points that represented facilities used by certain
sensitive segments of the population (e.g., schools and hospitals) .
Five years of meteorologic data (1970-1974) from the National
Weather Service Station in Albany, New York were used as input into
the models since this station had the most recent available
meteorologic data for several years in an area with some
*".•-'
topographical similarities to Rutland. Specific meteorologic data
for Rutland, Vermont were not used because the data were not
available at the time of modeling. Modeling was repeated using
limited data from one site, the Rutland, Vermont State and Local
Air Monitoring Station (SLAMS), and from cloud cover observations
from Burlington, Vermont, recorded during 1980 (U.S. EPA, 1987b).
Results using the Rutland-Burlington data were similar to those
obtained using the Albany data. Dispersion modeling showed the
areas for maximum impact lie within a 1-km radius from the MWC
stack.
Based on the results of the air dispersion modeling, a four-
station ambient air network was established for collection of
samples (Figure 2-1). The stations were located on accessible
2-2
-------
Figure 2-1,
Location of Monitoring Stations in Rutland, Vermont.
(See Table 2-1 for identification of sampling sites.)
2-3
-------
public property in primarily residential areas. Three of the
stations were either near the modeled sites of highest estimated
annual average concentration of pollutant emissions (within a 1 km
radius of the stack) or close to areas of topographical importance;
these sites were located on Watkins Avenue, Route 4, and River
Street (See Table 2-1). The fourth station was the existing SLAMS.
Water, soil, sediment, food and forage samples were also collected,
and the collection points for these samples are also given in Table
2-1. Some of these sites were at distances >2.0 km and are not
shown in Figure 2-1.
The Watkins Avenue monitoring site was located 0.37 km north-
northeast of the MWC on the property of the Havenwood School. The
Route 4 monitoring site was located 0.40 km west-southwest of the
MWC, next to the Evergreen Cemetery and the Rutland municipal
building. Residential homes in the area were not located as close
to the Route ,4 "monitoring site as the other sites. The River
Street monitoring site was located by the River Street Pumping
Station and across the street from an athletic field, 0.59 km from
the MWC.
2.2. SAMPLING AND ANALYSIS
The target pollutants for this study were listed in Chapter
1. The methods used for the collection and chemical analysis of
samples are described in separate sections because of the
difference in these methods for air samples and the environmental
2-4
-------
TABLE 2-1
Sampling Sites in Rutland,. Vermont
Site
Location Relative to MWC
Media
Sampled
MWC
SLAMS
Watkins
Avenue
Route 4
River
Street
Route 3
Quarter1ine/
Boardman Hill
Roads
Creek Road
Route 133
Route 100
Rutland City
Reservoir
Rocky Pond
Adjacent to MWC
1.1 km east-northeast
0.37 km north-northeast
0.40 km west-southwest
0.59 km south-southeast
1.7 km west-northwest
2 . 2 3cm south-southwest
2.8 km south
4.6 km west-southwest
West Rutland
Westfield, Vermont
6.4 km northeast; this
is the primary drinking
water source for Rutland
2.4 km north
Soil (1)
Air (2)
Air (3)
Soil
Air (4)
Soil
Air(5)
Soil
Milk (6a)
Forage (6b)
Soil (6b)
Milk (7a)
Potato (7b)
Soil (7b)
Milk (8)
Forage
Soil .. .
Carrot
Soil
Milk
Surface water
Sediment
Surface water
Sediment
2-5
-------
TABLE 2-1 (continued)
Site
Location Relative to MWC
Media
Sampled
Junction of
East and
Otter Creeks
Otter Creek
at Rutland
Town/City
Line
Otter Creek
at Junction
of Routes 3
and 4
0.42 south-southeast
2 km west, downstream
of the Rutland Waste
Water Treatment Plant (RWTP)
2 km west, downstream of
both the Rutland City
and RWTP
Surface water (9)
Sediment
Surf ace water (10)
Sediment
Surface water (11)
Sediment
Site location on the map is indicated by number in parenthesis.
Sites not located on the map are not numbered.
2-6
-------
media. Figure 2-2 displays the time periods when the ambient air
and environmental media samples were collected.
2.2.1. Ambient Air Sampling. The selection of sites for ambient
air-was based on air dispersion modeling as discussed in Section
2.1. Since the same sample collection method could not be used for
all selected pollutants, four-different techniques were used.
Standard mass flow Total Suspended Particulate*(TSP) high-volume
(Hi-Vol) samplers were used to collect samples for the later
determination of mutagenicity of the total suspended particles in
the air. The PS-1 PUF samplers, detailed in Compendium of Methods
for the Determination of Toxic Organic Compounds in Ambient Air
(U.S. EPA, 1984a) , were • used for the determination of total
(suspended and vapor phase) PCDD/PCDFs, total PCBs and the
mutagenic activity in air. The inhalable arsenic, beryllium,
cadmium, chromium, lead, nickel, silver and B[a]P in the air were
collected by PM-10 critical-flow Hi-Vol samplers. Ambient air
samples for mercury were collected by low volume vacuum samplers
with controlled mass flow. The mercury samples were collected only
at the SLAMS site because the sampler required a controlled
environment (Vermont Air Pollution Control Division, Agency of
Natural Resources, 1987b).
Each ambient air monitoring site was equipped with at least
two General Metal Works PS-1 samplers, one standard mass flow Total
Suspended Particulate (TSP) Hi-Vol sampler and one Wedding PM-10
2-7
-------
a
00
3 JJ
BO'S
E'S
£
.1?
CO
I
O
O
S1 «S"
IS1 I §"
i g " g
S- |S
!t It
Ii«
i — O
I '§ 8
I a1! a
S S a cJa
O «< CO CU CO
II
CO
CL,
3
CQ
O
o «
§•
W CO
2-8
-------
critical flow Hi-Vol sampler. Two ambient air monitoring stations
were designed as co-located sites for quality assurcince purposes
(Vermont Air Pollution Control Division, Agency of Natural
Resources, 1987b). A co-located site is a monitoring site equipped
with 2 of the same samplers so that duplicate samples can be
collected and the overall precision of the sample collectors can
be evaluated. The SLAMS was the co-located site for the TSP and
PM-10 samplers (i.e., the site has 2 TSP and 2 PM-10 samplers).
The Watkins Avenue site was the co-located site for the PS-1 PUF
sampler. Table 2-2 lists the air sampling equipment located at
each site.
The PS-1, the TSP Hi-Vol, and the PM-10 Hi-Vpl samplers were
run for one 24-hour period every 12 days; this frequency produces
-150 air samples annually for each metal, B[a]P, PCDD/PCDFs, PCBs
and mutagenicity analysis (a total of 1400 samples per year).
Sample collection occurred during the same 24-hour interval for
each monitor- and site. No ambient air samples were .collected
before the start of the MWC in November 1987; the first samples
s " ' " -
were collected in November 1987.
2.2.2. Meteorologic Information. Wind speed, wind direction,
temperature, relative humidity and solar radiation were
continuously monitored and recorded at three sites, the SLAMS,
River Street and Watkins Avenue, using Climatronics Electronic
Weather stations. Additionally, the SLAMS collected rainfall
2-9
-------
. .. TABLE 2-2r . _ ,,,.:;;%,,; .. .,„,... ...... ...,._•
Equipment at the Ambient. Air Monitoring Sites in Rutland, Vermont
Site
Co-located Equipment
Equipment8'0''
SLAMS
TSP and PM-10
Watkins PS-1 PUF samplers
Avenue •
Route 4 Not a co-located site
River Not a co-located site
Street
• ,,2: JPS-l^PUF
'2 TSP "
..2 PM-10 .-, , v^:.«,,
VAPCD #"ld"'
Low volume, ,
vacuum6
4 PS-1 PUF
1 PM-10
-VAPCD #2f, ,.
2 PS-1 PUF .
1 TSP
.1 PM-10 ,
., .,2 ,-PS-l ,PU;F
" "l 'TSP
- , .1 ,PM-10 . , i .
VAPCD #3f
aPS-l PUF samplers collected samples for . PCDD/PCDFs ,, PC^s .and ,
mutagenic activity.
''TSP samplers collected particulates for the mutagenicity bioassay.
°PM-10 samplers collected B[a]P, arsenic, beryllium, cadmium, chromium,
lead, nickel and silver. , , ,. •:-,-• ; -? • - -, , •-. ,, -;--;r ,;;
fl collected meteorologic information: wirid speed, and, direction
at 10 meters elevation, temperature, rainfall intensity, relative
humidity, atmospheric pressure ; and , solar, radiation. , «,,,.,
°Low-volume vacuum sampler with a mass flow controller collected
air samples for mercury analysis.
fVAPCD #2 and #3 collected meteorologic information: 'wind speed, wind
direction (at 2.5 meters elevation), and temperature.
2-10
-------
intensity and atmospheric pressure. The SLAMS began measuring
these parameters for this study on October 5, 1987. The Watkins
Avenue and River Street sites began monitoring on January 1, 1988
and May 19, 1988, respectively. A total of twenty months from the
SLAMS, ten months from Watkins Avenue site (November 1988 is
unavailable), and sixteen months of data from the River Street site
are currently available. The measurement principles used for each
of the meteorologic parameters are discussed below (Vermont Air
Pollution Control Division, Agency of Natural Resources, 1987b).
Wind speed was measured using a three-cup anemometer. The
rotation of the cup was converted into an electrical signal by a
phototransistor and light source. The frequency of the electrical
signal produced was proportional to the wind speed. The signal was
amplified and transmitted to a translator for conversion into a DC
voltage.
A vane was used to determine the wind direction. The position
of the vane was converted to an electrical signal by a low-torque
potentiometer and then sent to a translator. The translator
converted the signal to a DC voltage output.
Temperature was determined by a thermistor network. As the
temperature of the thermistor changed, the resistance of the
network changed. The change that occurred in the network was then
converted to a DC voltage output.
Relative humidity sensors detected moisture by the
hydromechanical stress of small cellulose crystallite structures
2-11
-------
acting on a pair of thermally-matched silicon strain gauges
connected by a half wheatstone bridge. The strain gauges converted
the strain into electrical signals that were amplified by the
translator into an electrical voltage analog of the relative
humidity.
The relative humidity and temperature sensors were housed in
a mechanically aspirated radiation shield to reduce error caused
by solar heating. Ambient air was drawn across the sensors by an
electric fan. The exterior housing of the shield was painted white
to reflect radiation. The shield was mounted horizontally with the
air intake facing north to eliminate solar heating during sampling.
The solar radiation sensor was a temperature compensated
silicon photovoltaic cell mounted under a pyrex dome. The signal
from the cell was proportional to the intensity of sunlight
striking it. The radiation translator converted the output of the
cell to a DC voltage.
2.2.3. Ambient Air Analyses. Four analytical techniques were used
to quantify the concentration of the pollutants in collected
ambient air samples: neutron activation, inductively coupled
plasma emission spectrometry, thin-layer chromatography and high
resolution gas chromatography-high resolution mass spectrometry.
Particulates were examined for mutagenic activity by the reverse
2-12
-------
mutation assay (Maron and Ames, 1983). Table 2-3 summarizes each
method and the laboratories that conducted the analysis. Details
about the methods used for each pollutant are described below.
Arsenic and chromium were analyzed as total metals by neutron
activation analysis (NAA) using the procedure described in Standard
Operating Procedure for NAA Determination of Trace Elements in
Suspended Particulate Matter Collected on Glass-Fiber Filters (U.S.
EPA, 1984b). Two circles were removed from each glass-fiber filter
and irradiated by neutrons. A gamma-ray spectrum of the irradiated
material was obtained by a high-resolution large volume germanium
detector. The spectral data were compared to spectral data of
known standards for quantification. A blind repliceite, solutions
of four working standards, a quality control standard and fifteen
samples comprised a group of samples irradiated and analyzed
together.
Beryllium, cadmium, lead and nickel were analyzed by
Inductively Coupled Plasma-Atomic Emission Spectrometry (ICP-AES)
(U.S. EPA, 1983a). Metals collected on the glass-fiber filter
were dissolved in a mixture' of nitric and hydrochloric acid by
ultrasonication and centrifugation. The metal concentrations were
determined after dilution of the sample into the concentration
range of the ICP-AES. Working standards, dilutions of the working
standards, quality control solutions (high and low concentrations),
and filter and reagent (acid matrix) blanks were analyzed for
quality assurance purposes.
2-13
-------
TABLE 2-3
Ambient Air Analysis Analytical Procedure and Laboratory
Pollutant
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
Nickel
B[a]P
PCDD/PCDF
PCB
Mutagenic
activity
(from TSP
and PUP)
Analytical Method
NAAa
ICP-AES0
ICP-AES
NAA
ICP-AES
Pyrolyzer-
dosimeter
ICP-AES
TLCe- fluorescence
Preparation
HRGC-HRMS9
GC-ECDh
Reverse mutation
Laboratory
EPA-ORD/OMMSQAb
EPA-ORD/OMMSQA
EPA-ORD/OMMSQA
EPA-ORD/OMMSQA
EPA-ORD/OMMSQA
VAPCDd
EPA-ORD/OMMSQA
EPA-ORD/OMMSQA
EPA-OPP/ECLf
EPA-ORD/OMMSQA
EPA-ORD/OMMSQA
EPA-ORD/OHR'
Reference
(U.S. EPA, . 1984b)
(U.S. EPA, 1983a)
(U.S. EPA, 1983a)
(U.S. EPA, '1984b)
(U.S. EPA, 1983a)
(Spittler, 1973)
(U.S. EPA, 1983a) -.
(U.S. EPA, 1986b)
(Harless and
McDaniel, 1988)
(U.S. EPA, 1984a)
(Maron and Ames, 1983)
aNeutron activation analysis
bOffice of Modeling, Monitoring Systems and Quality Assurance,
U.S. EPA Office of Research and Development
°Inductively coupled plasma-atomic emission spectrometry
Vermont Air Pollution Control Division
°Thin-layer chromatography
Environmental Chemistry Laboratory, U.S. EPA Office of Pesticides and
Toxic Substances
sHigh resolution gas chromatography-high resolution mass spectrometry
hGas chromatography with electron capture detection
'Office of Health Research, U.S. EPA Office of Research and Development
2-14
-------
Mercury was to be analyzed using the methods described in "A
System for Collection and Measurement of Elemental and Total
Mercury in Ambient Air over a Concentration Range of 0.004 to 25
/^g/m3" (Spittler, 1973) . However, because of quality assurance
problems, mercury concentrations were not reported (Fitzgerald,
1990).
Benzo[a]pyrene samples were analyzed according to the
procedure described in Standard Operating Procedure for ultrasonic
Extraction and Analysis of Residual Benzora^pyrene from Hi-Vol
Filters via Thin-Laver Chromatoorraphy (U.S. EPA, 1986b) . A portion
of the glass-fiber filter was immersed in cyclohexane and sonicated
to extract the B[a]P. An aliquot was spotted on a. thin-layer
chromatography plate and developed in an ethanol/methylene chloride
solvent mixture. Ultraviolet fluorescence spectrometry was used
for quantification.
PCBs in ambient air were analyzed using a modified version of
EPA Method TO4 detailed in Compendium of Methods for the
Determination of Toxic Organic Compounds in Ambient Air (U.S. EPA,
1984a). The glass filters and PUF cartridges were subjected to
Soxhlet extraction; each extract was concentrated using the
Kuderna-Danish techniques and cleaned-up with alumina column
chromatography. The PCBs were quantified using gas chromatography
with electron capture detection according to EPA Method 608 (U.S.
EPA, 1984a). The system was calibrated using a 50:50 mixture of
Aroclors 1242 and 1260 for PCB identification and quantification.
2-15
-------
PCDD/PCDF collection and retention efficiency of air samplers
were verified by a field spike. An 800 pg aliquot of the
analytical standard 13C12-1,2,3,4-TCDD was spiked onto the center
two-inch area of the fiberglass filter, directly above the PUF plug
on the field blank and all field sampling cartridges before
sampling. No significant loss of the 13C12-1,2,3,4-TCDD was
observed, indicating that volatilization loss of the PCDD/PCDF was
not significant during sample collection, transport or storage
(Harless and McDaniel, 1988).
Sample preparation and analysis of PCDD/PCDF concentrations
were performed on "sets" of 12 samples consisting of nine test
samples, a method blank, field blank(s) and a laboratory method
spike. The filters and PUF plugs from each ambient air monitor
were combined prior to extraction. An aliquot of a spike solution
containing 1.0 ng each of 13C12-labeled PCDD/PCDF internal standards
(described below) was spiked into each sample immediately before
Soxhlet extraction for 16 hours with benzene. Cleanup of extracts
was accomplished using an acid/base procedure and a micro-silica
gel column, and a micro-alumina column followed by a micro-carbon
column. An aliquot of a solution containing 0.5 ng 37Cl4-2,3,7,8-
TCDD was spiked into each extract prior to final concentration to
60 /il for analysis. The extracts were fire sealed in glass tubes
and shipped to the U.S. EPA laboratory for analysis in a blind
manner, i.e., test samples and QA samples were not identified as
such.
2-16
-------
All samples were analyzed using a Finnigan MAT 311A and a
Finnigan MAT 90 HEMS system operating in the electron ionization
and multiple ion detection mode at 8000-10000 mass resolution and
equipped, with a 30m x 0.25 mm i.d. SE-54 fused silica capillary
column and a 60m x 0.24 i.d. SP-2331 fused silica capillary column.
The areas of exact masses of the mplecular ion clusters of 37C14 and
13C12-labeled and nonlabeled PCDDs and PCDFs and respective response
factors were used for quantification purposes. The 37Cl4-2,3,7,8-
TCDD was used to determine the method efficiency for 13C12-labeled
PCDD/PCDF internal standards. Respective 13C12-labeled PCDD/PCDF
internal standards were used for quantification of respective
nonlabeled PCDDS and PCDFs and for determination of the minimum
limits of detection (MLDs) with two exceptions, 13C12-l,abeled HpCDD
was used for HpCDF and 13C12-OCDD was used for OCDF. The 13C12-
labeled 1,2,3,4-TCDD was used to determine PS-1 .air sampler
collection and retention efficiency. Total congener concentrations
and isomer-specifie concentrations were reported in pg/m3.
The HRGC-HRMS analytical criteria used for confirmation of
PCDDs and PCDFs were: chlorine isotope ratios of molecular ions
(±15% of theoretical values, tetra - 0.77, penta - 1.55, hexa -
1.24, hepta - 1.04, and octa - 0.89); simultaneous responses (+3
sec) for exact masses of ,13C12-labeled and nonlabeled 2,3,7,8
chlorine-substituted congeners on a known isomer-specifie
column(5); resolution of PCDDs and PCDFs on the SP-2331 isomer-
2-17
-------
specific column demonstrated and confirmed using a standard
containing all tetra- through hexa- PCDD/PCDF isomers; analysis
that confirms the absence of respective chlorinated diphenyl
ethers; HRGC-HRMS peak' matching analysis of exact masses if
necessary, and responses of nonlabeled PCDD/PCDF masses must be
greater than 2.5 x area of the noise level.
The data from a "set" of 12 samples were evaluated using the
analytical criteria and following QA/QC requirements: method
recovery efficiency for 13C12-labeled tetra-, penta- and hexa-CDDs
and CDFs, 50 to 120%, hepta- and octa-CDDs, 40-120%; satisfaction
of the analytical criteria described for PCDDs/PCDFs; accuracy and
precision achieved for laboratory method spike(s) at 0.5 pg/m3 to
2.0 pg/m3, +50%; and method blank and field blank free of
significant PCDD/PCDF contamination at the MLDs . required for
generation of useful and meaningful data, usually in the range of
0.03 to 0.3 pg/m3 to tetra-, penta- and hexa-CDDs and CDFs. The
analytical procedures and QA/QC used in this study are fully
described elsewhere (Harless and McDaniel, 1988).
The samples collected between November 5, 1987 and February
9, 1988 were analyzed on the 311A HRMS system for 2,3,7,8-TCDD,
2,3,7,8-TCDF and total tetra-, penta-, hexa-, hepta- and octa-CDDs
and CDFs. The samples collected after February 9, 1988 were
analyzed on the more sensitive MAT 90 HRMS system for all 2,3,7,8-
chlorine substituted isomers and total tetra-, penta-, hexa-,
hepta- and octa-CDDs and CDFs.
2-18
-------
TSP Hi-Vol and PUF filters were extracted with dichloromethane
'(Williams et al., 1988). The resulting extract was concentrated
by rotary vacuum evaporation and redissolved in a final volume of
10 ml. Aliquots were subject to gravimetric analysis for
determination of extractable organic mass (EOM). Samples with
sufficient EOM were assayed for mutagenic activity using a reverse
mutation assay (Maron and Ames, 1983; U.S. EPA, 1987c) in
triplicate at a minimum of five doses with and without Aroclor-
induced rat liver metabolic activation (+S9 and -S9, respectively).
Solvent (DMSO) and positive controls (2-anthramine and 2-
nitrofluorene, with and without activation, respectively) were
tested concurrently with each assay. Statistical analysis of the
mutagenicity data was conducted according to the method of
Bernstein et al. (1982) . The slope values (revertants/jug) from the
dose-response analyses were converted to revertants/m of air to
reflect the concentrations of mutagens in the ambient air samples.
Chapter 3 briefly describes the results of the analyses.
Chapter 5 describes how these pollutant concentrations were
analyzed to determine the attribution of the MWC. Chapters 6
through 9 present the results of the analyses.
2.2.4. Environmental Media Sampling. Results of dispersion
modeling of projected emissions from the Rutland MWC -prior to
operation of the incinerator indicated that the greatest impact
2-19
-------
from the MWC would lie within a 1-km radius of the facility. Using
this dispersion modeling, general-locations for collecting water,
sediment, soil and agricultural products were identified (Table 2-
1 and Figure 2-1) and were located within 6.5 km of the MWC. The
VAPCD was responsible for sampling, the coordination of handling
and shipping of all samples to the respective laboratories. In
addition, the VAPCD compiled all related sample collection data and
results of chemical analysis.
Table 2-4 outlines the schedule followed for sampling of
water, sediment, soil, food and forage throughout the project year
\ t
1987-1988. Water, sediment, soil and milk samples were taken twice
prior to full operation of the facility and once after the
combustor was operational. Potato and forage were sampled twice,
and one carrot was sampled only once, before commencement of MWC
operation. Procedures for collection of samples in the various
environmental compartments are described in sections 2.2.4.1 -
2.2.4.3.
2.2.4.1. Surface Water and Sediments. For water and sediment
sample analyses, a total of fifteen samples, five per sampling
round, were collected and a representative composite of the samples
was used (Vermont Air Pollution Control Division, Agency of Natural
Resources, 1987b). Ten samples were taken before and five after
the initiation of MWC operations. One surface water sample per
site was collected with a water column sampler from the deepest
2-20
-------
TABLE 2-4 ,
Sampling Distribution for Environmental Media
Media
Water
Sediments
Soils
f
Milk
Produce
(Potato/
Carrot)
Forage
Pollutants
Metal sb
Metals
PCDD/PCDF
PCBs
Metals
PCDD/PCDF
PCBs
Metals
PCDD/PCDF
PCBs
Metals
PCDD/PCDF
PCBs
Metals
PCDD/PCDF
PCBs • .
No. Sample
Periods8
3
3
3
. - -3 : , '
3
3
3
. • , 3
3
' . '. . . 3 • . .• ,
2d ' ' ^
,:• --.'. 2d , .
2d
" " '"' ' ' 2* ' '
2d
2d
No. Samples
Per Period
, - . • 5
5
5
/ :• .5 •
•:••.:• 8
6
6
3C
3C
3C
2
,. :.. . 2
2 . ' .
2
"• •• 2
2
aThe sampling dates were mid-October 1987, early November 1987 and
late June 1988.
bArsenic, beryllium, cadmium, chromium, lead, mercury, nickel and
silver
was sampled at Quarter line, Route 3 and Creek Road in mid-
October 1987, and at these three sites and Route 100 in November
1987 and June 1988.
dProduce and forage were sampled in October 1987 and November 1987.
Source: Vermont Air Pollution Control Division, Agency of Natural Resources,
1987b
2-21
-------
part of the water. Sediment samples were collected along the
stream bank using a brass dredge (Vermont Air Pollution Control
Division, Agency of Natural Resources, I987b).
2.2.4.2. Food and Forage. Four milk, one carrot, one potato
and two forage (grass hay) samples were collected for each sampling
round from various farms in the area surrounding the facility.
Milk was sampled from bulk storage tanks at three different dairy
farms in the area surrounding the MWC. The carrot, forage, and
potatoes were collected directly from the field (Vermont Air
Pollution Control Division, Agency of Natural Resources, I987b).
For use as a background sample, one milk sample was also collected
from a bulk storage tank in Westfield, Vermont, an area -123 km
away from the MWC with no obvious source of external,pollution.
2.2.4.3. Soil. Four of the sampling sites were located
• ' J
within the area of expected maximum deposition (~l-km radius) . The
remaining sites were located at a distance >1 km. Systematic grid
sampling was used at all the sites to obtain a representative
sample from the area. Grid samples were collected and then
consolidated into one representative sample for each site. Samples
were collected from 1-6 cm deep for undisturbed soil and from 1-
15 cm deep for tilled soils using a thin-walled stainless steel
corer. Soil sampling procedures followed protocols specified in
U.S. EPA (1986c).
2-22
-------
2.2.5. Environmental Media Analyses. The water, sediment, soil,
food and forage samples were analyzed by the State of Vermont using
U.S. EPA standard operating procedures for the appropriate matrix
and pollutant. Internal quality control for extraction and
analysis of samples consisted of laboratory analysis of field and
laboratory blanks (minimum of 10% of total number of samples
collected) , duplicate pic split samples (10% of total number of
samples collected) and spiked samples (decided by the laboratory
performing the analysis). Spiked samples analyzed along with
unspiked samples provided an estimate of accuracy and precision of
chemical analysis. Table 2-5 lists the methods of analysis for the
f
pollutants in the these media.
Surface water samples were prepared for analysis by acidifying
with nitric acid, heating with hydrochloric acid, and filtering to
remove silicates and other insoluble materials. Soil, sediment and
agricultural samples were digested in nitric acid and hydrogen
peroxide and refluxed with either hydrochloric acid (beryllium,
cadmium, chromium, lead and nickel) or nitric acid (arsenic).
Metal analyses in medium other than water were conducted using
either direct aspiration (flame) atomic absorption for cadmium,
chromium, lead, mercury, nickel and silver or graphite furnace
technique for arsenic and beryllium. The graphite furnace was used
for all metals in water samples (U.S. EPA, 1979; U.S. EPA, 1983b).
2-23
-------
TABLE 2-5
Method of Analysis for Pollutants
in Environmental Media
Pollutant
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
Nickel
PCS
PCDD/PCDF
"Graphite furnace
Analytical Method
Water
GFFAA8
GFFAA
AA
AA
AA
AAC
AA
d ' •
d
atomic absorption spectrometry
Soil, Sediment, Food
and Forage
GFFAA
AAb
AA
AA
AA
AAC
AA
HRGC-HRMS6
HRGC-HRMS
vapor technique
Pollutant concentration not measured in sample
°High resolution gas chromatography-high resolution mass spectrometry
2-24
-------
Levels of PCBs in solid matrices (soil, sediment) were
determined using a modification of EPA Method 608 (U.S. EPA,
1982b). The samples were homogenized with sodium sulfate, spiked
with 13C-labeled surrogates.and Soxhlet extracted with toluene. The
extracts were solvent exchanged with hexane, acid/base washed with
concentrated sulfuric acid and potassium hydroxide and further
purified using a neutral/acid silica gel column. The resulting
extract was split into equal volumes for PCB and PCDD/PCDF
determination.
PCDD/PCDF and PCBs in milk were extracted using the procedures
of Rappe et al. (1987a) by Midwest Research Institute under
contract to the State of Vermont. Each milk sample was initially
fortified with 13C-labeled internal standards, then aqueous sodium
oxalate, ethanol and diethyl ether were added sequentially. The
mixture was extracted with hexane and back-extracted with water.
The resulting extract was slurried with acid silica gel and
decanted onto a neutral/acid silica gel column identical to that
used for the solid matrix samples. The extract was then carried
through the remainder of the clean-up as described above for the
solid sample matrices. . '" . ,
Prior to quantification, the PCB split extract was evaporated
and spiked with internal standards in tridecane. The PCDD/PCDF
split extract was further cleaned using a neutral alumina column
and a carbopak C/Celite 545 column. The final PCDD/PCDF extract
was reduced and spiked with internal standards in tridecane.
2-25
-------
High resolution gas chromatography-high resolution mass
spectrometry of the extracts was initially conducted in two phases.
Mono- through tri-substituted PCB isomers were determined on a
Finnigan MAT/Varian 311-A high resolution mass spectrometer using
a 60-m DB-5 fused silica capillary column, then the remaining PCB
isomers were determined using a Kratos MS50-TC mass spectrometer.
This two-phase technique was used for only the six samples
collected in 1987. The remaining 1987 samples were analyzed in one
phase using the Kratos MS50-TC mass spectrometer, which was
sensitive for all PCB isomer levels. The PCB extract splits of
1988 were analyzed on a Kratos MS50-TC using 30-m DB-5 fused silica
capillary column. With both PCB and PCDD/PCDF analyses, method
blanks were used to determine accuracy. The method blank
determined the concentration of the pollutant in the reagents,
glassware, and instrument used during the foretreatment of samples
prior to actual quantification. ,
Concentrations of all contaminants in soil, sediment, food and
forage samples (excluding milk) were calculated on a dry weight
basis. Metal concentrations in liquids were expressed as
weight/volume of sample. Concentrations of PCBs and PCDD/PCDFs in
milk were expressed as weight/weight of sample on a whole milk
basis. Chapter 4 describes how the measured pollutant
concentrations are used . in the determination of possible human
health effects. Chapter 11 presents the results of the exposure
assessment of the MWC.
2-26
-------
3. MEASURED CONCENTRATIONS IN AMBIENT AIR
AND ENVIRONMENTAL MEDIA
The data collected in this study as described in Chapters 1
and 2 were analyzed by several approaches to determine if the
source of these pollutants could be the Rutland incinerator. The
first step in ascertaining the source of the pollutants was a
qualitative review/analysis of the data, i.e., concentration of the
pollutants in the ambient air and environmental media, received
from the analytical laboratories. Several approaches for analyzing
the contribution of the incinerator to the measured levels of the
pollutants in both ambient air and environmental media were then
undertaken and are described in Chapter 5.
This section presents the ambient air and environmental media
monitoring data. The determination of the ambient air
concentrations from the air dispersion modeling of Rutland is
presented in Chapter 4. Chapter 5 describes the qualitative and
quantitative approaches used to discriminate the contribution of
the incinerator to the concentrations measured in Rutland. The
approach comparing the measured concentrations (from this section)
with the modeled concentrations (from,Chapter 4) is described in
Chapter 5.
3.1. RESULTS OP MONITORED CONCENTRATIONS IN AMBIENT AIR
The ambient air samples were analyzed for arsenic, beryllium,
cadmium, chromium, lead, nickel, benzo[a]pyrene, PCBs, and
PCDD/PCDFs. The time periods during which samples for each of
3-1
-------
these pollutants were collected varied slightly for several
reasons, including replacement of analytical equipment, inability
to detect any measurable pollutant concentrations, or the lack of
precision in the analytical procedure. The time periods of the
samples and the detection limits for each pollutant are presented
in Table 3-1. For risk assessment purposes, all pollutants except
PCDD/PCDF, the concentrations that are not detectable in the field
samples were assumed equal to the detection limit as determined by
the analytical laboratory. This conservative assumption was
applied since the sample concentration is known to be either less
than or equal to the specified detection limit. The assumptions
applied to the PCDD/PCDF field samples are described in Section
3.1.4."
Table 3-2 displays the sites at which the pollutant
concentrations were detectable. It should be noted that pollutants
were not detectable at any specific site for each day; the sites
varied. For example, on March 4, lead was detected at all four
sites, whereas B[a]P was detected at SLAMS. Beryllium and cadmium
were detected at Watkins Avenue, whereas chromium and arsenic were
detected at River Street and Route 4, respectively. The dates and
sites where all PCDD and PCDF congeners had detectable
concentrations are indicated with "PCDD/PCDF".
3.1.1. Metal Concentrations. The concentrations of metals
measured In Rutland ambient air were reported by the analytical
laboratory as ^g/m3, with the exception of beryllium that was
reported as ng/m3' The analytical laboratory adjusted the filter
3-2
-------
(0
H
CO
C£l
J
CQ
<
EH
W
CJ
O
•H
-P
rd
JLj
-p
c • •
0)
o
c
0
u
1-1
o
M
Q)
g^
I C
55$
(!) ^
X< ,_(
"^ O
£x)
rrj 0
« M
-P w
"e ^
•H £
C _1 '
Q O
tJ*
|JL
°Q
O
-H
/it
w
ft
tn
C
-H
rH
O_i
g
(TJ
w
Q)
EH
13 c
o , j>
•H Q
|t "-
td |
M jj
-M r^
•s I
0 W
{3
O W
U 0)
rH
Q) Qj
rH g
X) (0
(0 W
OH ,_,
Q) O ^3
•P ?
0) rH §
Q 0) .°
•S 4J
^~* S frt
os £
S el)
Q) rH §*
"g -p
3 O
55 EH
—.
g
>-
0 3
-p ^^
O 4J
Q) -H
-p g
0) -H
Q HH
in
rH
g
OTA
WJ
W T3
(U Q)
4J N
Q rH
(0
c
»
CM •*
v^^ ^^
CO CM
•in in
c^ t^*
"^^ *^^
"* CM
VO
O
O
0
c^ co
^j* ^
0 CM
O CM
O O
CO CO
CO CO
VO >
0 H
O CM
1 1
CO CO
in in
O H
*•< *•«,'
H H
H^
n
(0
g
O -H
•H . rH
C rH
Q) >t
in JH
M Q)
< CQ
r^
^j*
^^
o
in
f^
•^^,
CM
H
O
O
O
en
o
o
o
o
CO
CO
H
CM
1
CO
in
o
»^
H
g
3
-H
g
-0
crj
U
CO
CM
•^
O
in
f^
*^
H
in
VO
0
o
o
en
vo
o
o
o
CO
CO
vo
o
o
1
CO
in
o
-^
H
g
3
•H
g
O
JH
3
^
^>
*~^
CO
•>*
in
r->
^^^
in
vo
H
VO
o
o
o
CO
CO
H
CM
1
CO
in
o
^»»
H
rrj
(0
0)
KH
r>
^
*^
H
in
^^
••-^
oa
[^
^^
0
o
' O
CO
fO
o
o
o
CO
CO
H
CM
1
CO
in
o
-^.
H
rH
0)
X
o
•H
55
VO
in
^s.
CM
CM
in
r^
^^^
H
CM
CO
n
CO
o
en
CO
in
H
CM
1
CO
in
o
~^
H
CD
ft
r~~t
(d
m
tH
CO
**>^
O
t"»
f^
•*s^.
o
^f
•<*
vo
1
CO
0
CO
CO
CO
CM
H
1
CO
in
o
^.
H
ID
O -P
c fJ
«|j
"g Q) S
Tri O C
c in o
•H O
t. r* Q)
m ^
QJ *St (ti
l\ Jj
nj cu o
0? -P
rQ Q) "O
C "H rrt
g>X!
•H C -P
rH O (0
rH -H X!
I. r\
a) a} in
X) -P
-------
TABLE 3-2
Occurrence of Detectable Pollutant Concentrations in Ambient Air
*ll/05/87
*ll/17/87
*ll/29/87
*12/ll/87
12/23/87
01/04/88
*01/16/88
*01/28/88
*02/09/88
*02/21/88
*03/04/88
*03/16/88
*03/28/88
04/09/88
*04/21/88
*05/03/88
*05/27/88
*06/08/88
*06/20/88
*07/14/88
*07/26/88
*08/07/88
*08/19/88
08/31/88
09/24/88
10/06/88
10/18/88
SLAMS
-NA-
Pb BaP
Pb BaP
Pb Ni BaP
Pb
Pb Ni BaP
Pb BaP
Pb BaP
Pb
Pb BaP
Pb
Pb BaP
Pb
Pb BaP
Pb
Pb
Pb
Pb
Pb
Pb
Pb
Pb
As Pb
Pb
Pb BaP
Pb
Watkins Ave.
Pb
Pb
Pb BaP
Pb BaP
As Pb
Pb Ni BaP
PCDD/PCDF
Pb BaP
Pb BaP
Be Cd Pb
As
As Pb
Pb
Pb
-NA-
Pb
Pb
Pb
As Pb
Pb
Pb
Pb
Pb
Pb
River St .
Pb BaP :
Pb BaP
Pb BaP
Pb
Pb BaP
Pb BaP
Pb BaP
Cr Pb
PCDD/PCDF
Pb
Be Pb
Pb
PCDD/PCDF
Pb
Pb
Pb
Pb '
Pb
Pb
Pb
Pb
Pb
Be Pb .
Route 4
-NA-
Pb
Be Cd Pb
Pb BaP
Pb BaP
Pb
Pb BaP
Pb BaP
Pb BaP
As Pb
Pb
-NA-
As Pb
Pb
PCDD/PCDF
Pb
Pb
-NA-
Pb
Pb
Pb
Pb
Pb
NA
Pb
Pb
Pb
3-4
-------
TABLE 3-2 (continued)
10/30/88
11/11/88
11/23/88
12/05/88
12/17/88
01/22/89
02/03/89
02/15/89
SLAMS
Pb
Pb
Pb BaP
PCDD/PCDF
Pb BaP
Pb BaP
PCDD/PCDF
BaP
BaP
Watkins Ave.
Pb
Pb BaP
PCDD/PCDF
Pb BaP
PCDD/PCDF
Pb BaP
PCDD/PCDF
PCDD/PCDF
BaP
PCDD/PCDF
BaP
River St.
Pb
Pb
Pb BaP
Pb BaP
Pb BaP
PCDD/PCDF
Route 4
Pb
Pb BaP
Pb BaP
Pb BaP
PCDD/PCDF
BaP
PCDD/PCDF
* = Combustor operating
3-5
-------
concentration for the volume of the air sample for each filter
(amount of air drawn through the sampling apparatus) and also for
blanks. Minimal limits of detection (MLD) were reported for each
metal, and the accuracy of the method was determined as described
by Harper et al. (1983). Samples without detectable concentrations
were assumed to have concentrations equal to the MLD reported by
the analytical laboratory.
As shown in Table 3-1, arsenic was measured above its
detection limit of 0.0046-0.0047 jug/m3 in 7/98 samples. The
measured concentrations ranged from 0.0061-0.0080 jug/m3. One
sample above the detection limit was collected from SLAMS, four
from Watkins Avenue, and two from Route 4. The highest detected
concentration was located at Route 4 and was collected during a
period when the incinerator was in operation. Beryllium was
measured above the detection limit of 0.2243 ng/ra3 in 4/122
samples. The detectable concentrations ranged from 0.3361-0.4618
ng/m3. One of the 'samples with a detectable concentration was
collected at Watkins Avenue, two at River Street, and one at Route
4. The sample with the highest detectable concentration was
collected when the incinerator was operating.
Cadmium was measured above its detection limit of 0.0009-
0.0014 Mg/m3 in only 2/122 samples. One sample, with a
concentration of 0.0022 ng/m3 was collected at Watkins Avenue when
the incinerator was operating. The ' other sample, with a
concentration of 0.0013 jug/m3, was collected at Route 4 when the
incinerator was operating.
3-6
-------
Chromium was measured above its detection limit of 0.0065-
0.0069 MS/™3 in only 1/98 samples. This sample was collected from
River Street when the incinerator was operating; the concentration
was 0.0113 jLtg/m3.
Lead was measured above its detection limit of 0.0061 Atg/m in
108/122 samples. All samples at SLAMS were above the detection
limit with a concentration range of 0.0084-0.0958 jug/m3. The
sample with the highest concentration was collected when the
incinerator was not operating. At Watkins Avenue, all but six
samples were above the detection limit. The concentrations ranged
from 0.0070-0.0529 jug/m3. At River Street, all but six samples were
above the detection limit and the concentrations ranged from
0.0072-0.0438 Aig/m3. At Route 4, all but two samples were above
the detection limit with concentrations ranging from 0.0070-0.0450
jLtg/m3. For Watkins Avenue, River Street and Route 4, the samples
with the highest concentrations were all collected on-the same day,
January 16, 1988, when the incinerator was operating.
Nickel was detected above its detection limit of 0.0038-
0.0077 |tg/m3 in only 3/122 samples. The concentrations ranged from
0.0086-0.0096 jig/m3. Two samples above the detection limit were
collected at SLAMS and one was collected at Watkins Avenue. The
sample with the highest concentration was collected at SLAMS when
the incinerator was not operating.
Samples for mercury analysis were collected at SLAMS.
However, precision of the collected samples was unacceptable (i.e.,
3-7
-------
the QA objectives were not met) and, while the problem was not
resolved, the mass flow controllers were suspected of being the
source (Fitzgerald, 1990).
3.1.2. Benzo[a]pyrene. The concentrations of benzo[a]pyrene
measured in Rutland ambient air were reported by the analytical
laboratory as ng/m3' The analytical laboratory adjusted the filter
concentration for the volume of the air sample and also for blanks.
The minimal limits of detection (MLD) was reported and samples
without detectable concentrations were assumed to have
concentrations equal to the MLD reported by the analytical
laboratory.
Benzo[a]pyrene was detected above its detection limit of
0.3348 ng/m in 43/131 samples.- These concentrations, however,
may not reflect the total B[a]P concentrations due to losses (of
10-90%) incurred by the sampling method for collecting polycyclic
aromatic hydrocarbons in suspended particulate matter (Peters and
Seifert, 1980). The concentrations ranged from 0.3755-6.391 ng/m3,
and samples with concentrations above the detection limit were
evenly distributed among the four sites. The sample with the
highest concentration was collected at SLAMS when the incinerator
was not operating. B[a]P was detected at all four sites with the
highest detectable levels in January- March 1988. and October 1988-
February 1989, which may have occurred due to increased wood and
fossil fuel burning. The levels of B[a]P during March-September
3-8
-------
1988 were either nondetectable or near ;the detection limit. The
increase in B[a] P" levels during winter months and the decrease
during the summer months indicate a seasonal fluctuation.^
3.1.3. PCB Concentrations. Total-PCB concentrations were adjusted
by the volume of the air sample for each filter arid reported as
hg/m3i No PCBs were measured above the detection limit in any
samples collected. The detection limits generally ranged from 0.7
to 0.8 ng/m3. However) two samples deviated from this range with
-detection limits of 12.10 and 1.13 .ng/m3. These two detection
limits are high because the samples had low total air flow drawn
through the sampling cartridge. Detection limits were derived by
dividing the total amount of PCB measured in each cartridge (<3 jug
for ali: cartridges) by-the total air-flow. Therefore, samples with
low air flow had higher than average detection limits (Sander,
1989) .,'•-'• , " -•-•--.'• ' - - - . -..-•,-• .:••... •-•-- > >• • •-' '•
3.1.4. PCDD/PCDF. Field blanks and field samples were collected
at the monitoring sites as described in Chapter 2 and analyzed for
PCDD/PCDFs. Each~ field blank consisted of a cartridge and PUF,
which were taken into the field; placed in , the equipment, and
handled in the same manner as:the field samples without having air
drawn throughi (Vermont Air Pollution Control Division, Agency of
Natural Resources, 1987b) .- The concentrations detected in the
field blanks represented contamination from sampling and analytical
techniques. The field samples were assumed to have the same level
of contamination as the field blanks.
3-9
-------
PUFs from two vendors, Supelco and GMW, were used in the study
(Harless, 1989). As the study progressed, concentrations of
several TCDF isomers, including 2,3,7,8-TCDF, began to be routinely
detected in field blanks and samples that had been collected with
the Supelco PUFs. These isomers were not detected in GMW PUF
filters or method blanks. Comprehensive HRGC-HRMS analyses
performed on 60 m SP-2331 and 50 m DB-5 Dioxin fused silica
capillary columns suggested that these TCDF isomers may have been
adsorbed from material used to package the PUF. However, this was
not confirmed, and the source of the isomers was never conclusively
identified. Since the distribution of TCDF isomers was
recognizable in the samples and field blanks collected with Supelco
PUF, corrections were made by the analytical laboratory by
subtracting the concentrations detected in the respective field
blanks from those detected ^n the field samples.
In addition to TCDFs being detected in samples using the
Supelco PUF, low levels of HpCDDs and OCDD in the range of 0.1 to
0.3 pg/m3 were consistently detected in method blanks, field blanks
and samples throughout the study, regardless pf the type of PUF
used during sampling. The elevated levels of HpCDD and OCDD were
due' to contamination from reagents, glassware and analytical
procedures. No corrections for HpCDDs or OCDD were made to sample
data by the analytical laboratory because there were no significant
differences in the minimum limits of detection.
Quantification of PCDDs/PCDFs in samples collected prior to
February 9, 1988 was performed on a 311A HRGC-HRMS system. Results
were reported for 2,3,7,8-TCDD, 2,3,7,8-TCDF and total tetra-,
3-10
-------
penta-, hexa-, hepta-, and octa-CDDs/CDFs. Quantification of
PCDDs/PCDFs in samples collected after February 9, 1988 were
performed on a more sensitive MAT 90 HRGC-HRMS system (See Section
2.2.3). Results were reported for total congener and all 2,3,7,8-
chlorine substituted isomers. The analytical laboratory reported
the concentrations as pg/m3 ambient air.
Watkins Avenue was the co-located site for the PCDD/PCDF
sampling. Concentrations for the duplicate samples were averaged
for reporting of sample concentrations for a particular day.
Figure 3-1 displays the precision achieved by the sampling and
analytical methods for the data from January 16, 1988. The
precision achieved throughout the study was very good except in a
>
few cases where the concentrations were very low.
For the purposes of the human health evaluation, the
concentrations reported by the analytical laboratory were further
adjusted so that the TEF. approach (described below) could be
applied. Figure 3-2 shows the decision tree used for these
adjustments. If the concentrations of the total congener and
2,3,7,8-isomer were detectable, the non-2,3,7,8-isomeric
concentration for the specific congener was determined by
subtracting the adjusted 2,3,7,8-isomer concentration from the
adjusted total congener concentration. However, if the
concentrations of 2,3,7,8-isomer were nondetectable, certain
assumptions were applied to,the total congener concentration so
that the 2,3,7,8-isomer portion could be estimated. For example,
if the 2,3,7,8-TCDD concentration emitted from the incinerator is
-5% of the total emitted TCDD concentration, a proportionality
3-11
-------
I
M
a)
tr>
•H
3-12
-------
. Ambient Air
Concentrations
Sampled
before
2/9/88?
Is total
2,3,7,8-
detectable?
Multiply total congener
by proportion factor
= 2.3.7.8-
Areall
2,3,7,8-for
a particular
congener
detectable?
Multiply congener
concentration by
proportion = total
2.3.7,8-
Sum 2,3,7,8-for
particular congener
= total 2,3.7,8-
Subtract total 2,3,7,8-
from total congener
concentration =
non-2,3,7,8-
Figure 3-2,
Approaches Used for Estimating 2,3,7,8-TCDD
Equivalent Concentrations. (See Section 3.1.4.)
3-13
-------
constant of 0.05 was used to estimate the concentration of 2,3,7,8-
TCDD in that air sample. The 2,3,7,8-isomeric concentration can
be computed as follows:
Total Cone, x Proportionality = 2,3,7,8-conc. for
of congener Constant that specific congener
Equation (3-1)
The proportionality factors used in this study were obtained
from two sources: the Rutland ambient air data and the interim
TEF method of U.S. EPA (1989). The values for the concentration
ratios of 2,3,7,8-substituted isomers to total homologue for the
PCDF series were obtained from the mean values of the detectable
field sample concentrations collected from Watkins Avenue (1/16/88,
12/05/88, 12/17/88, 1/22/89), River Street (12/05/89) and SLAMS
(12/17/89). These data were the only samples collected during the
study period that contained detectable isomer-specific PCDD/PCDF
concentrations. The proportionality of the 2,3,7,8-substituted
isomers in these samples is assumed to be representative of Rutland
ambient air. The proportionality factors that were estimated from
these data for PCDDs and PCDFs are listed in Table 3-3.
For the PCDD/PCDF samples collected after February 9, 1988,
each 2,3,7,8-isomer was analytically separated and quantified so
that the total 2,3,7,8-isomeric concentrations could be computed.
If the 2,3,7,8-isomeric concentrations for a particular congener
were all detectable, the concentrations were summed to equal the
3-14
-------
TABLE 3-3
Proportionality Factors for PCDD/PCDF Derived from
Rutland, Vermont Ambient Air Data
Proportionality Factor* ± SD
PCDD .%
2,3,7,8-TCDD/Total TCDD 0.05 ±0.05
2,3,7,8-PeCDD/Total PeCDD 0.06 ± 0.01
2,3,7,8-HxCDD/Total HxCDD 0.18 ± 0.01
2,3,7,8-HpCDD/Total HpCDD 0.51 ± 0.05
PCDF
2,3,7,8-TCDF/Total TCDF ' 0.04 ± 0.03
2,3,7,8-PeCDF/Total PeCDF 0.13 ±0.03
1,2,3,7,8-PeCDF/Total PeCDF 0.06 + 0.01
2,3,4,7,8-PeCDF/Total PeCDF 0.07 + 0.01
2,3,7,8-HxCDF/Total HxCDF ^0.36 ± 0.04
2,3,7,8-HpCDF/Total HpCDF 0.68 ± 0.11
Proportionality factor derived from Rutland ambient air data,
i.e., derived from 6 samples wherein all isomers were detectable.
3-15
-------
total 2,3,7,8-isomer concentration. However, if any of the
2,3,7,8-isomers were not detectable, then the proportionality
factors were applied as described in Equation 3-1. The product of
the proportionality factor and total congener concentration should
be less than or equal to the sum of the 2,3,7,8-isomeric
concentrations. If this product was greater than the sum of the
2,3,7,8-isomers, then this sum was used since in this case the
product would have overestimated the total concentration.
Once the total 2,3,7,8-isomeric concentration was computed,
the non-2,3,7,8-isomeric concentration for each congener was
calculated by:
Total congener - 2,3,7,8-isomeric = non-2,3,7,8-isomeric
cone. cone. cone.
Equation (3-2)
This computed concentration was then multiplied by the
appropriate TEF to estimate the 2,3,7,8-TCDD equivalent
concentration for all samples, according to Equation 3~-3.
The PCDD/PCDF isomers and congeners have different toxicities
depending primarily on the positions of the chlorine substitution
(U.S. EPA, 1989). In general, substitution at the 2,3,7,8-
positions gives rise. to greater potency. Thus, to relate the
different isomeric and congener concentrations of the samples, the
isomeric and congener concentrations were converted to 2,3,7,8-
TCDD equivalent concentrations by using the toxic equivalency
3-16
-------
factors (TEFs). The TEFs relate the potency of the different
congeners to the potency of 2,3,7,8-TCDD, the most potent congener.
The TEFs of the congeners are presented in Table 3-4.
The concentrations of PCDD/PCDF congeners were converted to
a total 2,3,7,8-TCDD equivalent concentration by applying
individual TEFs according to the following equation (U.S. EPA,
198.9) :
2, 3,7,8-TCDD
equivalent
cone.
S(TEF x cone, of + S(TEF x cone, of each
each 2,3,7,8-CDD/CDF . non-2,3,7,8-CDD/CDF
congener) congener)
Equation '(3-3)
Once the 2,3,7,8-TCDD equivalent concentration was estimated
for each sample, the 2,3,7,8-TCDD equivalent concentrations were
compared with the modeled concentrations using the same statistical.
tests as described above.
Total 2,3,7,8-TCDD equivalent concentrations in Rutland
measured ambient air samples ranged from 0.011 to 5.39 pg/m3.
Table 3-5 shows these concentrations. The highest concentrations
were measured during the time when the MWC was shut-down. The
highest detected 2,3,7,8-TCDD equivalent concentration of 5.39
pg/m3 was measured in January 1989 after the MWC was shut-down.
The fluctuation in the PCDD/PCDF concentrations and the high
concentrations during !the shut-down period indicate input from
other sources (such as automobiles or wood burning) or- meteorologic
changes (i.e., temperature inversion). The data in Table 3-5 also
3-17
-------
TABLE 3-4
Toxic Equivalency Factors (TEFs) of the Congeners
of PCDD/PCDF i
Isomer
TEF
PCDDs
PCDFs
2,3,7,8,-TCDD
All other TCDDs
2,3,7,8-substituted PeCDD
All other PeCDDs
2,3,7,8-substituted HxCDD
All other HxCDDs
2,3,7,8-substituted HpCDD
All other HpCDDs
OCDD
2,3,7,8-TCDF
All other TCDFs
1,2,3,7,8-PeCDF
2,3,4,7,8-PeCDF
All other PeCDFs -
2,3,7,8-substituted HxCDF
All other HxCDFs
2,3,7,8-substituted HpCDF
All other HpCDFs
OCDF
loO
0
0.5
0
0.1
0
0.01
0
0.001
0.1
0.001
0.05
0.5
0
0.1
0
0.01
0
0.001
The symbols T, Pe, Hx, Hp, and O are abbreviations for tetra-,
penta-, hexa-, hepta-, and octa-, respectively.
Source: U.S. EPA, 1989
3-18
-------
TABLE 3-5
,2,3,7,8-TCDD Equivalent Concentrations (pg/m3)
in Rutland, Vermont
Monitoring
Date
11/05/87
11/17/87
11/29/87
12/11/87
12/23/87
01/04/88
01/16/88
02/09/88
02/21/88
03/04/88
03/16/88
03/28/88
04/21/88
05/03/88
05/.27/88
06/20/88
07/26/88
08/07/88
08/31/88
09/24/88
10/06/88
10/18/88
10/30/88
11/11/88
11/23/88
12/05'/88
12/17/88
01/22/89
02/03/89 .
02/15/89
SLAMS
0.02
0.02
0.03
,0.14
0.06
0.03
0.84
,0.61
. , 0.04
0.02
0.02
0.06
0.07
0.04
. 0.03
0.03
0.04
! 0.03
0.03
0.04
... 0.18
0.04
0.02
0.01
0.09
0.08
0.13
0.06
0.07
0.07
Watkins
Duplicate Samples
0.02
0.03
0.02
0.04
0.04
0.03
1.31 , ,
0.39
0.06
0.04
0.06
0.05
0.06
0.10
0.05
0.04
0.02
0.03
0.02
0.03
0.04
0.02
0.02
0.01
0.03
5.04
,0.15
5.20
0.07
0.07
0.02
0.02
0.02
0.03
0.04
0.03
1.04
0.29
0.03
0.07
0.08
0.06
0.06
0.08
0.04
0.07
0.02
0.03
0.04
0.03
.0.03
0.01 ,
0.02
0.01
0.03
5.04
0.15
5.59
0.06
0.09
Site
River St.
0.02
0.02
0.02
0.04
0.03
0.17
0.96,, -
0.04 ,
0.04
0.22
0.05
0.04
0 . 09
0.06
0.03
0.03
0.03
0.02
0.03
0.03
0.06
0.02
0.02
0.03
0.09
0.42
0.06
0.07
0.06
0.11
Rte. 4
0.02
0..02
, 0.02
0.03
0.03
0.02
0.16
0.03
0.05
0.07
0.02
0.01
0.07
0.02
0.04
0.02
0.02
0.03
0.03
NA
0.02
0.04
0.02
0.01
0.04
0.03
0.07
0.49
0.05
0.08
NA = Sample concentration was not available,
3-19
-------
indicate that atmospheric transport is a major mode for dispersal
of there compounds throughout the environment and provides an
explanation for the routine detection of trace levels. For
example, high concentrations of PCDDs/PCDFs on 01/16/88, 12/05/88
and 01/22/89 were rapidly dispersed in the atmosphere, and only
elevated background levels could be detected in the next sampling
periods, on 02/09/88, 12/17/88 and 02/03/89.
3.2. ENVIRONMENTAL MEDIA ,,,
Environmental media were sampled in areas surrounding the
Rutland MWC during the project . period. Three rounds of
environmental sampling were conducted: October and November 1987,
and June 1988. Water, sediment, soil and milk samples were
collected twice before and once after the incinerator began
operating. Potato and forage were sampled twice, and one carrot
was sampled prior to MWC operation. The sampling procedures have
been described in Section 2.2. The environmental media were
analyzed for the following pollutants: arsenic, beryllium,
cadmium,chromium, lead, mercury, nickel, PCB (except water) and
PCDD/PCDF (except water). Table 2-5 shows the analytical methods
for these pollutants.Samples collected in 1987 prior to operation
of the Rutland MWC represent background levels for comparison with
those samples taken during MWC operations.. The primary objective
of sampling both before and during operation was to show the
incremental increase of pollutant concentrations in environmental
media, if any, caused by emissions from the facility.
3-20
-------
Concentrations of all contaminants were calculated on a dry
weight basis for soil, -sediment and agricultural products
(excluding milk). Liquid concentrations of metals were expressed
as weight/volume of sample. Concentrations of PCBs and PCDD/PCDFs
in milk were expressed as weight of sample on a whole milk basis.
3.2.1. Metals. The metals concentrations were used in the
statistical comparisons as reported by the analytical laboratory
without any further corrections. The results of the monitored
concentrations are reported in Chapter 11.
3.2.2. PCBs. Concentrations of PCBs were reported by the
analytical laboratory as individual congener concentrations. To
account for any contamination that occurred during the laboratory
handling and analysis, the detectable method blank concentrations
were subtracted from' the respective field sample concentrations.
This "adjusted" concentration represents the PCB concentration
present in the environmental media. Following correction of the
concentrations, the congener concentrations for each sample were
summed to calculate the total PCB concentration for each sample.
Statistical analyses were performed with the total PCB
concentration as described in Chapter 11.
3.2.3. PCDD/PCDF. For the PCDD/PCDFs, the laboratory analysis
provided the results for each 2,3,7,8-isomer and total congener of
each field sample and method blank. Samples were corrected for
possible analytical and handling contamination by the method blank
3-21
-------
concentrations. The field samples were corrected for contamination
by subtracting detectable method blank concentrations from the
corresponding isomer and total congener concentrations in the field
samples. If the method.blank concentrations were non-detectable,
they were assumed to be zero and no correction was made to the
isomer or total congener PCDD/PCDF concentrations in the "field
samples. If the method blank and sample were both non-detectable,
then the sample was set equal to the detection limit. If the
sample was non-detectable but the method blank was detectable, then
the method blank was subtracted from the sample, which had been set
equal to its detection limit to account for contamination due to
the analytical methodology. This procedure resulted in a
conservative estimate of the PCDD/PCDF isomer and total congener
concentrations, as the actual concentrations were less than or
equal to the detection limit.
For comparison between the sampling periods, the adjusted
concentrations were converted to 2,3,7,8-TCDD equivalent
concentrations by using Equation 3-3. If the concentrations of
the total congener and 2,3,7,8-isomeric concentrations were
detectable, the non-2,3,7,8-isomeric concentration for the specific
congener was determined by subtracting the adjusted 2,3,7,8-isomer
concentration from the adjusted total congener concentration. If
the concentrations of the 2,3,7,8-isomer(s) were nondetectable,
they were assumed to be equal to the method detection limit.
However, if this value exceeded that for the total congener
concentration (e.g., when both the concentrations of the 2,3,7,8-
isomer and total congener were nondetectable, but with different
3-22
-------
limits of detection), the concentration of the 2,3,7,8-isomer(s)
was set equal to that of the total congener concentration. This
would result in a non-2,3,7,8-isomer concentration of zero. For
the PeCDFs, different TEFs for the 1,2,3,7,8- and 2,3,4,7,8-
isomers were used (U.S. EPA, 1989). Therefore, in cases where the
2.,3,7,8-PeCDF concentrations were nondetectable, but exceeded the
total PeCDF concentration, the concentration of the more potent of
the two, the 2,3,4,7,8-isomer, was set equal to the total PeCDF
congener concentration andrthe 1,2,3,7,8-isomer concentration was
set at zero. Results are shown in Chapter 11.
3-23
-------
-------
4. AIR DISPERSION MODELING
The Industrial Source Complex Short-Term (ISCST) model was
run to predict the ground-level ambient air concentrations of
pollutants in Rutland for the same days at which the ambient air
was sampled at the four monitoring sites. These predicted
concentrations were 24-hour average ambient concentrations, and
were later compared with the measured ambient air concentrations
(also 24-hour concentrations). The comparison of the measured and
predicted ambient air concentrations was an approach to examining
the contribution of the MWC to the pollutants in Rutland. This
comparison is discussed in Chapter 5.
Prior to the modeling of the emissions, the wind speed and
wind direction data' that were collected at the monitoring sites
(i.e., SLAMS, River Street, and Watkins Avenue) were evaluated to
determine the more appropriate data to use for the modeling.
This chapter describes the wind speed and wind direction data
j •
collected at the three monitoring sites, the modeling procedure and
parameters'used to model the stack emissions from the Rutland MWC,
the uncertainty associated with the modeling results, and the ISCST
model results.
4.1. METEOROLOGIC RESULTS
Data were collected at the SLAMS, Watkins Avenue and River
Street sites. Twenty months of data were available from the SLAMS,
10 months of data were available from the Watkins Avenue site, and
4-1
-------
16 months of data were available from the River Street site. The
meteorologic recording period for each site is as follows:
Site
SLAMS
Watkins Avenue
River Street
Start Date
October 1987
January 1988
May 1988
Stop Date
August 1989
October 1988
August 1989
Data before October 1, 1987 were available for the SLAMS site
only. However, these data were flawed because no wind data were
recorded for the south to west quadrant (bearings >180°, due south,
to <270', due west). Therefore, the data collected before October
1, 1987 could not used for the air dispersion modeling.
The SLAMS site was situated in a parking lot in downtown
Rutland on a 10 meter tower 1300 meters northeast of the
incinerator. The site was near office buildings that may have had
some effect on the recorded wind direction. The Watkins Avenue
site was 250 meters north of the incinerator, 3 meters above the
ground, and was near some trees that may have affected the wind
speed and possibly the wind direction. Any effect on the wind
speed and direction would probably be minimal in the winter months,
but is more pronounced in the late spring, summer and early fall
when leaves were on the trees. The River Street site anemometer
was 3 meters above ground in an athletic field ~400 meters south
southeast of the incinerator and was probably unaffected by local
buildings or trees. ' ,
4-2
-------
Wind direction and speed were recorded electrpnically every
hour at each site; the data were transferred to the State of
Vermont computer. The wind direction data were collapsed into the
16 wind direction sectors by combining the exact wind directions
recorded at each site into categories of 22.5° intervals from 0°
to 337.5°. Speed data were collapsed into 6 classes (0-1, 1.1-2,
2.1-3, 3.1-4, 4.1-6, and >6 meters per second). These categories
were used so that subtle differences could be detected.
Frequencies of detecting hourly speed/direction combinations were
then generated by counting those data points that had both
direction and speed data since, for many hours, data were available
for only one of the two parameters. Each data set or point
represented both a wind direction and a wind speed measurement.
The number of data points available for any one month varied from
162 to 744 (672 data points are possible for a 28 day month, 720
»-" " ~ '.' > ' ' ' , . , ' , " •' " '""'•,.
for a 30 day month, and 744 for a 31 day month).
The wind speeds were grouped into the following categories:
data for each site by month, site total (all months), monthly data
across all sites, and all data. The analysis was performed in this
way so that variations in monthly wind patterns at each site and
between sites could be assessed and the change in overall patterns
made by combining site data sets could be estimated.
The data from River Street and SLAMS are graphically presented
in Figures 4-1 to 4-11 as three-dimensional bar graphs. For all
graphs, the bars located in the back row (criss-cross pattern) of
the graph represent the total wind in each direction; -bars nearer
4-3
-------
CO
*i
CO
8
1
H
(0
g
(0
h>
I
-------
-S
**********
1 09 O
********* J):
i^^i-eoiociiaocQO
CM N Ol »- •- i-
aaqoq.oo
aaquieoaa
jequiaq.de s
c
o
01
fi
tn
eg
I
Q>
i.
•H
4-5
-------
O4 O4 CV* v «- v
eunr
**********
"> O
***********
4-6
-------
(0
43
-P
C
o
a\
CO
ui
3
I
i—r
***** *#'•** *
r^*-*-coiocja>c
e
>1
•O
C
•H
A
4->
a
i
4-7
-------
aunr
C
•H
^
4J
O
CO
CO
01
H
(0
I
(ti
§
n)
•o
Tf
§
tn
in
i
4-8
0)
I
•H
-------
II
CO
CO
0)
d)
o
0)
Q
O
-p
00
CO
W N Ol "^ *• T"
I
aequraosa
(0
-P
CM (MM •- *•*-
c
•H
•o
c
(0
A
-P
I
ft •
(fl 0)
n c
CQ
vo
0)
4-9
-------
VI
G
•H
X
-p
to
II
N CM CM »- f t-
o
-------
•p
CO
0)
***********
J5 S 5 5 £ 5 « • s s
*****
*****
naono
CO
CO
0)
c
O
-P
CO
CO
&
-------
oo
CO
0)
0)
o
a)
Q
CO
CO
CTl
H
-P
(0
c
•H
(0
'
*****.*****
&
>t
H
|
O
g
O.4J
(0 W
(U
>
4-12
-------
« W CM »••-•-
*********£
&-<-T-coioMo>«no
« « •• ^ »•
01
CO
a\
H
I
O
-p
CO
o>
H
(0
g
(0
aunr
•g'
•H
I
>1
rH
|
O
g
o
in
ft-p
(0 Wl
m
•
o
H
-*
0)
I
0)
>
-H
4-13
-------
'•a
c
.to
en
CO
en
H
4J
in
a
**********
• - Natomo
(0
JJ
(0
•d
•s
"S
(0
X! 4J
-P W
C
O H
fi Q)
O
CO
(0 O
M S
s
4-14
-------
the front of the graph represent sequentially increasing wind
speed. All bars represent wind coming from the direction specified
on the X-axis.
4.1.1. SLAMS Site. The dominant wind directions (the five
directions with the highest percentage of data points) at the SLAMS
site were from the south southwest (14%), north northeast (12%),
southwest (11%), north (10%), and west southwest (8%) (summarized
in Figure 4-4). Wind was almost totally absent in the east
northeast through south southeast directions, which may .be the
result of wind channeling from buildings located in the general
area. The absence of wind in these directions is a contrast to the
data from the other sites. At the Watkins Avenue and River Street
sites, the percentage of data points was more evenly distributed
over the 1/6 wind directions. The yearly summary of wind speed data
at the SLAMS site shows that 80% of the time the wind was < 2 m/s
and only 2% of the time it was > 4 m/s.
4.1.2. Watkins Avenue Site. The dominant wind directions
for the Watkins Avenue site were west northwest (14%) ,, northwest
(11%), northeast (10%), west (9%), and south southwest (6%)
(summarized in Figure 4-7). The yearly summary of wind speed data
shows that ~95% of the time the wind speed at Watkins Avenue was
< 2 m/s and only 0.1% of the time it was > 4 m/s. The much lower
wind speeds seen at the Watkins Avenue site, particularly during
the summer months, compared with the SLAMS site may be the result
4-15
-------
of both the height of the recording station (3 m versus 10 m at
SLAMS) and the close proximity of trees (see Figures 4-2 and 4-
4.1.3. River Street Site. The dominant wind directions at the
River Street site were south southeast (12%) , southeast (12%) ,
northwest (11%) , 'west northwest (7%) , and south (7%) (summarized
in Figure 4-11) . Wind direction datav for this site "were similar
to the Watkins "Avenue data for May /'but the 'data' were not similar
to either of the two sites for tlie remaining months.
The yearly summary of wind speed data at River Sti-eet shows
that 84% of the time wind "was < 2 m/s and 2% of the time it was
> 4 m/s. Wind speeds at the River Street site were slower than at
the SLAMS site (probably because the anemometers are different
heights) although these data appear to : be more similar than are
the Watkins Avenue and the SLAMS data "for the months June, July,
and August. .-••'-. • -' •••••• • -"•.''• ;.'---'- ........ : ....... .-•••'•' ..... ••••.•'•'±~:,- ••. • .-..-•. •-
4.1.4. Conclusion. Due to 'the apparent •Variability in the wind
speeds measured at the Watkins Avenue site, these meteqrolbgic data
were not used for modeling. The, wind speeds appeared to be
affected by the surrounding' barrier since they were slower during
the summer when there was foliage on the trees.
4-16
-------
4.2, MODELING METHODOLOGY ! / ;>: '^f
Twenty-four hour average amb ient , air? concentrations were
predicted for the Rutland area using the Industrial Source Complex
Short-Term (ISCST) model in the Urban 3 Mode (U.S. EPA, 1986a).
The Urban 3 Mode, an option of the ISCST used to describe the
surrounding, topography, was selected because the incinerator was
located in a rural area with complex terrain. , The model was run
for each date for which there, was adequate .meteorolpgic data,
ambient air samples were collected, and the MWC was in operation.
The output from each ISCST modeling run was a ground-level .ambient
air concentration at designated receptors. The ISCST ,w7as run
using both discrete and polar receptors. The discrete receptors
corresponded to the locations of the four monitoring sites by using
their Universal Transverse Mercatpr ,(UTM) coordinates.. The polar
receptors represented the intersects, of the 16...wind directions
beginning with north and spaced every 22.5 degrees; along the, .polar
azimuth at distances of 0.2, 0.5, 1.0,2.0, 5.0,1020, 30, 40 and
50 km from the MWC (for a total of 160 receptors) . An emission
rate of 1.0 g/s was used since the stack emission, rates were not
available for each sampling day. ... , , , ...... ....
The source parameters, described in Section 1.3, consisted of,
general information, about the MWC. Exhaust from the incinerator
was vented from a single stack, which was 1.040 m in diameter and,
50.3 m high. The exhaust gas exited at a temperature of 327.6 K
• ' i , . *^ •
and a velocity of 15.24 m/s.
4-17 '
-------
Hourly meteorologic inputs required, by the ISCST included mean
wind speed, the direction to which the wind was blowing, ambient
air temperature, the Pasquill stability category, the mixing layer
height, the vertical potential temperature gradient and the wind-
profile exponent. The only input parameters available for Rutland
were wind speed, wind direction and ambient air temperature. Cloud
cover information from Glens Falls, NY was used to predict
stability categories because no such information was available for
Rutland. Glens Falls has the closest National Weather Service
Station and has similar topography to Rutland; both cities have
valleys oriented north-south. Hourly mixing height was not
available for Rutland, so morning and afternoon mixing height data
were developed by the National Climatic Data Center based on
Albany, NY and Burlington, VT data (U.S. Department of Commerce,
National Oceanic and Atmospheric Administration, 1990). Since
hourly mixing heights and stability categories were not available,
the RAMMET preprocessor program was used to develop hourly mixing
heights and Pasquill stability categories from the surface and
upper-air meteorologic data.
Wind speed and wind direction data were collected at three
monitoring sites in Rutland (as discussed in Chapter 2): SLAMS,
River Street and Watkins Avenue. The ISCST was run using the data
and anemometer heights for SLAMS and River Street. The data from
Watkins Avenue were not modeled because the wind speeds observed
during the summer months were much lower than that observed during
the other months.
4-18
-------
The modeling results represent the ground-3,evel ambient air
"concentrations of the pollutants assuming one unit emission. These
concentrations do not represent the actual concentrations
attributable to the MWC for each sampling day because the actual
stack emission rates were not incorporated into the model; these
daily stack emission rates were not available. To determine an
estimate of the magnitude of the pollutant-specific ground-level
ambient air concentrations, the predicted concentrations at each
receptor (assuming 1 g/s) can be multiplied by the measured stack
emission rate of the pollutant that was measured during the stack
emission testing, which was required permitting. However, these
pollutant-specific concentrations do not represent the actual daily
concentrations since the daily stack emissions were not
incorporated.
4.2.1. Stack Emission Testing.' Stack emission testing of the
MWC was required under the Air Pollution Control Permit for the
State of Vermont (Agency of Environmental Conservation, State of
Vermont, 1986). The emission concentration of each pollutant was
sampled for four hours on three days in March 1988. Lead, arsenic,
mercury, beryllium, cadmium, chromium and nickel were collected on
a heat filter and in a series of impingers on March 2, 3 and 14 and
were analyzed by inductively-coupled argon plasma spectroscopy and
atomic absorption spectroscopy using the proposed Methodology for
the Determination of Trace Metal Emissions in Exhaust Gases from
Stationary Source Combustion Processes (Lodi, 1988). PCDD/PCDF
4-19
-------
stack samples were isokinetically collected by the MM-5 Sampling
Train method of the U.S. EPA (Lodi, 1988) on March 8, 9 and 10.
The PCDD/PCDF were trapped in a glass fiber filter and XAD-2 resin
of the sampling train and were analyzed using high resolution mass
spectrometry (Lodi, 1988). Method blanks were also analyzed. The
concentrations of PCDD/PCDF in the three stack samples from the
incinerator are presented in Table 4-1.
Measured stack concentrations of each PCDD/PCDF isomer were
corrected by the respective blank concentrations. The corrected
concentrations were then converted into an overall 2,3,7,8-TCDD
equivalent concentration by the TEF method (U.S. EPA, 1989) using
the TEFs listed in Table 3-4. The 2,3,7,8-TCDD equivalent emission
rates from the Rutland municipal combustor stack for the three
days were 5.22xlO"8, 6.78xlO"8, and 9.16xlO~8 g/s. The results of
the stack emission testing for all pollutants are shown in Table
4-2.
4.3. PROBLEMS AND UNCERTAINTIES ASSOCIATED WITH THE MODELING
The goal of the modeling procedure was to predict the
concentrations at each monitoring site for each sampling day
assuming one unit emission so that these concentrations could later
be used for the comparison of the measured and predicted
concentrations. However, because of the lack of meteorologic data,
only thirteen of the sampling days were modeled using the data from
SLAMS, and five days were modeled using River Street data.
4-20
-------
TABLE 4-1
PCDD/PCDF in Stack Emissions of Rutland Incinerator
(ng)
Compound
2,3,7,8-TCDD
Other TCDD
1,2,3,7,8-PeCDD
Other PeCDD
2,3,7,8-HxCDD
Other HxCDD
2,3,7,8-HpCDD
Other HpCDD
2,3,7,8-TCDF
Other TCDF
1,2,3,7,8-PeCDF
Other PeCDF
2,3,7,8-HxCDF
Other HxCDF
2,3,7,8-HpCDF
Other HpCDF
Sample
Run 1
0.117
4.403
0.341
6.886
2.266
9.27
4,107
5.762
5.929
29.741
3.793
27.017
1.1.347
14.962
126.884
25.131
Collection
Run 2
0.198
7.222
0.559
11.214
4.486
16.02
7.051
7.89
9.904
51.203
6.307
40.922
15.31
21.062
13.238
5.942
Run
Run 3
0.222
7.759
0.798
14.739
6.134
22.499
10.959
11.831
11.422
47.734
7.401
45.188,
19.157
22.896
15.646
6.947
Blank
0.048
1.495
0.139
2.658
0.967
1.624
1.703
1.874
2.593
, 11.295
1.867
9.524
3.591
4.058
2.729
0.908
Source: Lodi, 1988
4-21
-------
Ci
1
•V*
3
s
^
^
w
^
Vs.
c^
rHI
«3
-P
o
E
'•w
O
o
«
_^
issior
e
a
t^
Q
B
W
0
V>
Q)
ea
45 W
•P
•H -H
rij C
^^ irt
0}
CO
C C
P4 P4
C
O
•H
ts
0)
rH CM
rH
O C
Q)
rH
ft
(C
to
H
1
PS
C
•H
•P W
O Q)
Q) -P
rH Q
o
0
•g
(0 '
4J
rH
rH
O
CM
1
X
CO
CO
+1
•o
'o
H
X
**
CM
N.
i
O
X
H
0
\0
N.
'o
X
a\
*#
*Q
i
,30X10
VO
CO
CO
•v.
s^
CO
CO
Vs.
CO
CM
Vs
CO
o
•H
C
0)
w
a
in
i
o
H
X
^0
en
in
-H
i ^
3 o
KJ H
o X
° 0
r- \o
^
0
*Q X
5z co
"CM
H
\Q \Q
'o 'o
X X
O CM
\Q CO
r~ co
in
'o
* ' H
Q X
a CM
•«*
^0
CO CO
CO CO
Vs. Vs
Vs. Vs.
CO CO
CO CO
\ "V
CO CO
CM CM
X. S».
CO CO
g
•H e
rH 0
iH -H
r*i 3
w 13
m u
10
i
0
H
X
CM
H
H
+1
m
'o
H
X
CO
H
in
t
O
X
0
CO
CM
1
o
X
CO
o
p*
T •
,43X10
CM
CO
CO
Vs.
Vs.
CO
CO
Vs.
CO
CM
V^
CO
g
rj
•H
O
6
l
o
H
X
H
O
CM
H-l
S.J.
1
O
H
X
O
\O
-4*
-4-
i
O
X
CM
t~-
CM
t
o
X
H
H
CO
t
, 95X10
p-
co
CO
Vs.
CO
CO
Vs.
CO
CM
>s.
CO
T3
(0
a
I
O
H
X
O
CO
H
+1
sj.
• | --.
O
H
X
CO
,a\
H
-*
i
0
X
in
, rH,
•
Q
*
co"
CO
H
,rf
•o
ti
lj
w
V^
w
rH
'*
O
ft
r*i
ui
0)
-P
.. .2
C
w
ui
-rj
m
0)
5
S'
O
-------
The modeling incorporated Rutland-specific meteorologic data
along with mixing height data based on the meteorology of
Burlington, Vermont and Albany, New York and cloud cover
information from Glens Falls, New York. If any of the data were
missing, the missing information ,was estimated from the existing
data. .' If a data point (such as a temperature reading, wind
direction or wind speed) was missing, the proceeding and following
hourly observations were averaged; this average was assumed to
equal the missing datum. Tables 4-3 and 4-4 'indicate which
sampling dates were modeled and any missing data.
Uncertainty was introduced into the modeling by using
incomplete data files and meteor.ologic data from other national
weathei: service stations (i.e.,- Albany, Burlington and Glens
Falls)'. The extrapolation of a mixing height introduces
uncertainty into the concentrations. RAMMET uses the sampling
day's morning and afternoon mixing height observations, and the
following morning's observations to predict the hourly mixing
height observations for the sampling day. If the missing mixing
height is estimated to be lower than the actual mixing height, the
pollutants would not be estimated to be transported as far.
The ISCST model for stacks uses the Gaussian plume equation
(U.S. EPA, 1986a) where the ground-level ambient air concentration
is inversely proportional to the mean wind speed at the stack. If
the missing wind speed is estimated to be less than what it
actually is, the concentration at a, point downwind may be
overpredicted. If the wind direction is incorrectly assigned, the
4-23
-------
TABLE 4-3
Dates Modeled Using SLAMS Meteorologic Data and
Associated Missing Data
Date
Data Information
01/16/88
01/28/88*
02/09/88
02/21/88*
03/04/88*
03/16/88*
03/28/88
04/21/88*
05/03/88*
05/27/88*
06/08/88*
06/20/88*
07/14/88*
07/26/88*
08/07/88*
08/19/88
Missing 1 wind direction
Missing temperatures and wind directions from 000
to 800 hours
Missing the afternoon mixing height
Missing the next day's morning mixing height
Missing that day's mixing heights and 1 wind
direction
No available wind speed or wind direction data
Missing that day's morning mixing height
Missing that day's morning mixing height
Missing the next day's morning mixing height, 2 wind
directions and 2 temperature observations
Missing wind direction data for 000 - 1000 .hours
8 An asterisk
date.
(*) indicates that modeling was completed for this
4-24
-------
Date
TABLE 4-4
Dates Modeled Using River Street Meteorologic Data and
Associated Missing Information
Data Information
05/27/88
06/08/88
06/20/88*
07/14/88*
07/26/88
08/07/88*
08/19/88*
Missing 1 wind direction
Missing that day's morning mixing height and 4 wind
directions
Missing the next day's morning mixing height
Missing wind directions from 000 - 1200 hours
a An asterisk (*) indicates that modeling was completed,for this
date.
4-25
-------
concentrations predicted to be downwind may be overpredicted, while
the concentrations at other points (that is, those points towards
which the wind was actually blowing) may be underpredicted.
4.4. ISCST MODELING RESULTS FOR RUTLAND
The ISCST model was run two separate times for each sampling
day, once using the wind direction and speed data from SLAMS and,
a second time using the River Street data. The wind speed and
direction data from Watkins Avenue were not used for modeling
because of the low wind speeds observed during the summer, and
therefore may not reflect the actual wind conditions in Rutland.
For each sampling day, ambient air concentrations were
predicted at the four monitoring sites as well as at the polar
receptors. The polar receptors were used as quality assurance;
the precision of the modeling could be checked by comparing the
predicted concentrations of the polar and discrete receptors. The
modeled concentrations based on one unit emission at the monitoring
sites and the maximum concentrations with the corresponding polar
receptor using the SLAMS meteorologic data and the River Street
meteorologic data are shown in Tables 4-5 and 4-6.
The concentrations predicted to occur at the monitoring sites
using the SLAMS meteorologic data ranged from 0 jig/m3 to 5.22 fj>g/m.
assuming one unit emission. The Watkins site was predicted to
receive the highest concentrations compared with the other
monitoring sites. The prominent wind directions from which the
wind was blowing for the days modeled occurred in the southwest
4-26
-------
in
l
W
m
^c
o
•rH
-«
ai o
° Tj
a °
ai
M £
0
0,
w
•H
•a
a aiai
01 > k
O -H 4J
s <*w
01
•P in
0) CO)
r< -H 3
0 Ai C
If) 4-> 01
•H (0 >
Q 3 <
0)
4J
Q
O
O
CM
in
*d*
o
o
in
CO
o
o
0
co
o
1-1
o
o
CO
f
in
«•
CO
co
rH
o
CM
CO
i-l
in
•
CM
CM
O
O
in
o
o
o
o
CM
i-i
n
CO
rH
O
H
in
r^
o
01/28/88
CO
CM
r-l
in
f
CM
rH
rH
O
O
in
o
o
0
o
vo
n
0
co
i-i
o
CO
o
o
02/21/88
rH
in
CM
in
eg
CM
o
o
in
1-1
o
o
0
o
0
o
o
o
o
o
o
o
o
o
o
03/04/88
r~ oS
CM O
CM CM
"
in
o
CM CO
CM r-l
O. 0
0 0
in in
CM O
H O
; o o
o o
.. , ...
O CO
n in
o o
o o
o <^
O rf
O CM
0 O
n o
CM 0
r- o
o o
03/16/88
04/21/88
- .
' ffj *3>
V£> rH
r-l CM
O
co in
H *r
o o
o o
in CM
H CO
O CM
0 0
CO 0\
rH in
O rH
o p,
in en
O VO
o n
o o
in \o
co i-i
[^ rH
O rH
05/03/88
.05/27/88
VD
rH
n
in
CM
o
CM
o
o
o
rH
O
O
O
I.
(o
rH
o
O
i-H
T
CO
O
in
o
0
o
06/08/88
\O M1
r~- I-H
n n
in
- .
r- in
vo rt
o o
,o o
CM CM
O O
O O
O O
O O
,r- CM
in co
rH O
o co
o o
.0 o
p o
CO CO
CO '^*
i-H O
06/20/88
07/14/88
^
co
. in
in
CM
CM
o
0
CN
o
0
o ,
o
rH
CO
O
CO
O
O
O
in
CM
CM
in
07/26/88
cv
o
CM
in
•<*
o
0
in
CM
0
o
o
'co
o
OY
in
o
co
rH
08/07/88
•
s
0) (0
>H Ul
Hi 4J
•a «
— 43
•p
•P C
I-l O
o
C iJ
o
B -P
o a
J-i 0)
01
a) ^H
Ul
3 10
X rH .
O O
o a
rH
O >H
' O
01
M C
3 O
Ul -H
(0 .p
0) 10
CP C
C 01
"H O
RJ O
0) O
XI
8 Az imuth
b Greatest
4-27
-------
w
3
a
^
«_.
W
jj
o.
0)
0
(!)
05
.
rH
o
Hi
0)
•p
•a
(0
o
10
S1
•rt"
to
4J
•r(
o
£
•*•
JC
33
4J
(0
w
o
centrati
o
CJ
dieted
2
10
Q
U
"en
0
r~t
o
to
O
Q)
(11
W
£
(1)
0)
M
W
rl
(U
^
a
TJ
(0
,— ^
u>
>x
"
(fl
o
Emissi
4J
C
•^
o
•a
0)
U)
(0
CO
o
•H
4J
(0
4J
0)
U
o
CJ
4J
0)
+J
m
0)
**
•G
4*
"•»
^
i-i
0
4J
U
a
P5
to
(0
»~ t
0
a,
^.^
MB
o
:entrati
\j
o
CJ
to
O
4J
a
a)
K
Lscrete
Q
""c
0
4J
to
H
4J
in
c
o
CJ
to
1
•rt
a
•H
s
K»
s
a.
a
O
•rl
-P
O
0)
>-l
•H
Q
1?
M
3
•H
•a
(0
a
^
0)
jj
I
CO
CO
4J
0)
s
CO
(1)
1
CO
0
n
00
r>
co
00
oJ
In"
o
en
01
*o
01
in
r-
o
o
01
o
o
o
o
CO
CO
01
o
O!
CO
rH
O
CO
in
o
CO
00
0)
o
CO
^3«
rH
in
in
O!
01
o
o
01
01
o
o
rH
rH
O
in
o
n
o
01
co
^r
CO
CO
.
rH
\.
O
CA
^»
co
01
in
rH'
CO
o
o
in
o
o
o
H
H
O]
o
0
o
in
in
H
O
co
co
0
co
o
o
Oi,
o
01
in
OJ
01
o
o
01
CO
CO
\o
rH
<£>
in
o
o
CTi
H
in
o
in
in
rH
0
01
co
CO
"•x
01
H
CO
O
B5
•a
•H
rH
a
a
a
w
4J
(0
4J
1°
O
e -P
o a
to a)
C
C 0)
•rl O
l-l C
10 O
a) o
ja
43 W
•P o>
3 -P
H (rf
•H S
< O
a JQ
4-28
-------
quadrant, thus the Watkins Avenue site was downwind from the MWC
for a majority of sampling days. Figures 4-12 through 4-24 display
the windrose for each sampling day based on the SLAMS meteorologic
data.
On July 26, the Watkins Avenue site was predicted to have the
largest concentration of all the sampling sites for all the modeled
days. The maximum concentration in Rutland was predicted to be
very close to this monitoring site. The prominent wind directions
were south southwest and southwest (See Figure 4-221) .
On March 4 all of the monitoring sites were predicted to have
approximately zero concentrations. For this day, the wind was.
blowing from the northeast and north northeast, so none of the
sites were located downwind from the MWC on this day. The maximum
concentration modeled at a polar receptor was predicted to be 2.51
/ig/m3 at 5OO meters southwest of the MWC.
r
The concentrations predicted to occur at the monitoring sites
using the River Street meteorologic data ranged from 0 jig/m3 . to
4.782 jug/m3 assuming one unit emission. As with the SLAMS
meteorologic data, the Watkins Avenue site was predicted to receive
the highest concentrations compared with the other monitoring
sites. The directions from which the wind was blowing for the days
modeled were more variable than that observed at the SLAMS, but the
wind blew most frequently from the southwest. Figures 4-25 through
4-29 display the windrose for 'each sampling day based on the River
Street meteorologic data.
4-29
-------
— E
S '
0-1 1.1-2 2.1-3 3.1-4 li-.>
Wind Speed Classes
(meters/second)
4.1-6
6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From, which the wind is blowing.
Figure 4-12,
Windrose for January 16, 1988 in Rutland, VT .based
on the SLAMS meteorologic data.
4-30
-------
N
\ \ £5 30%
1^ 2.0
_ 1J .
LO
I
/
- E
r~4
0-1 1.1-2 2.1-3 3.1-4 11— >
4.1-6
6
Wind Speed Classes
(meters/second)
NOTES: . :' ' ' •••••'.
Diagram o!f the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-13. Windrose for January 28, 1988 in Rutland, VT based
on the SLAMS meteorologic :data.
4-31
-------
30%
— E
I
O-i 1.1-2 2.1-3 3.1-4 I r- >
Wind Speed Classes
(meters/second)
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-14. Windrose for February 21, 1988 in Rutland, VT based
on the SLAMS meteorologic data.
4-32
-------
w—
- — E
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-15. Windrose for March 4, 1988 In Rutland, VT based on
the SLAMS meteorologic data.
4-33
-------
I
-1 1.1-2 2.1-3 3.1-4 I r- >
Wind Speed Classes
(meters/second)
NOTES: .
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-16. Windrose for.'March 16, 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-34'
-------
40%
— E
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
• NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
Prom which the wind is blowing.
Figure 4-17.
Windrose for April 21, 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-35
-------
4096
— E
.1-3 3.1-4 i- >
Wind Speed Classes
(meters/second)
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-18. Windrose for May 3, 1988 in Rutland, VT based on the
SLAMS.meteorologic data.
4-36
-------
w—
s
r~
0-1 1.1-2 2.1-3 3.1-4 Ir- >
Wind Speed Classes
4.1-6
6
(meters/second)
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-19. Windrose for May 27, 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-37
-------
N
W —
40%
— E
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram, of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-20. Windrose for June 8, 1988 in Rutland, VT based
the SLAMS meteorologic data.
on
4-38
-------
w —
s
0-1 1.1-2 2.1-3 3.1-4
-Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-21. Windrose for June 20, 1988 in Rutland, VT based on
the SLAMS taeteorologic data.
4-39
-------
N
r~
)-i 1.1-2 2.1-3 3.1-4 I r- >
Wind Speed Classes
(meters/second)
4.1-6
6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-22. Windrose for July 14,. 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-40
-------
w—
— E
I
-1 1.1-2 2.1-3 3.1-4 Ir- >
Wind Speed Classes
(meters/second)
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-23. Windrose for July 26, 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-41
-------
\
Lo '
;
\
.5 2
'
/
,o 2
/
5 30
56
T E
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-24. Windrose for August 7, 1988 in Rutland, VT based on
the SLAMS meteorologic data.
4-42
-------
N
W —
— E
S .
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
'NOTES: ; • " : -
Diagram of the frequency of
• Occurrence for each wind direction.
Wind direction is the direction
• f , ' • " '
From which the wind is blowing.
Figure 4-25. Windrose for May 27, 1988 in Rutland, VT based on
River Street meteorologies data.
4-43
-------
25 30*
— E
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
'.NOTE'S:
Diagram of the frequency of
•Occurrence for each wind direction.
Wind direction is the direction
From, which the wind is blowing.
Figure 4-26. Windrose for June 20, 1988 in Rutland, VT based on
River Street meteorologic data.
4-44
-------
w—
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram, of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-27. Windrose for July 14, 1988 in Rutland, VT based on
River Street meteorologic data.
4-45
-------
E
0-1 1.1-2 2.1-3 3.1-4 11- >
Wind Speed Classes
(meters/second)
4.1-6
6
NOTES:,
: Diagram of the frequency of
' Occurrence for each wind direction.
Wind direction is the direction
., From which the wind is blowing.
Figure 4-28. Windrose for August 7, 1988 in Rutland, VT based on
River Street meteorologic data.
4-46
-------
,N ,
W —
0-1 1.1-2 2.1-3 3.1-4
Wind Speed Classes
(meters/second)
4.1-6
> 6
NOTES:
Diagram of the frequency of
Occurrence for each wind direction.
Wind direction is the direction
From which the wind is blowing.
Figure 4-29. Windrose for August 19, 1988 in Rutland, VT based
on River Street meteorologic data.
4-47
-------
These modeling results indicate ground-level ambient air
concentrations of the pollutants emitted from the stack at a rate
of 1 g/s. Because these concentrations do not represent the actual
concentrations attributable to the MWC (since the stack emission
rates were not incorporated into the model) , the results were used
in the nonparametric statistical tests described in Chapter 5. The
results of the statistical analyses are described in Chapter 6.
4-48
-------
5. APPROACHES FOR ANALYSIS OP SOURCE CONTRIBUTION
The purpose of this report is to determine the human exposure
to the pollutants emitted from the Rutland MWC. This chapter
describes the methods used to determine the contribution of the MWC
to the measured pollutants in the ambient air and environmental
media.
Both qualitative and quantitative approaches were used for
analysis of ambient air concentrations of the pollutants; -only a
qualitative approach was used for the environmental media. The
approach for the analysis of environmental media was qualitative,
involving a comparison of concentrations between the various
sampling periods and a comparison with pollutant concentrations in
other geographical regions.
5.1. AMBIENT AIR APPROACHES
Analysis of the incinerator as a source for the measured
pollutants in ambient air encpmpassed four approaches: (1) the tons
of waste burned by the MWC were compared with measured particulate
matter (PM-10) concentrations, (2) mutagenic activity was compared
with amount of waste burned and PM-10 concentration, (3) the
congener profiles of measured PCDD/PCDF in Rutland ambient air were
compared with those of potential sources, and (4) daily ambient air
concentrations of pollutants that were predicted from air
dispersion modeling (ISCST) were compared with the measured
pollutant concentrations. One quantitative approcich that could
not be conducted due to limitations in the data was the comparison
5-1
-------
of ambient air samples collected during operation with those
collected while the incinerator was nonoperational (or shut-down).
The majority of the shut-down (August 1988 - February 1989) and
operational samples (December 1987 - August 1988) were collected
during different seasons, precluding a direct comparison of
operational and non-operational (or shut down) measured pollutant
concentrations. Kniep et al. (1970) has reported on the seasonal
patterns of metals in ambient air that are dependent on
temperature, wind speed and sources.
5.1.2. Qualitative Approaches to Analyzing Ambient Air Source
Contribution.
5.1.2.1. CORRELATION OF TONS OF WASTE BURNED TO PARTICULATE
CONCENTRATION — The TSP Hi-Vol glass-fiber filters and PUF
samples were analyzed for both PM-10 (particulate matter < 10 p)
concentration and mutagenicity (see Section 5.1.2.2.)- °ne
approach to analyze the concentration of pollutants in ambient air
was to determine if there was a relationship between the amount of
particulate (PM-10 concentration) and the amount of waste (as tons
per day) burned by the incinerator. This relationship was
investigated since many pollutants adhere to particulate matter and
because a possible correlation may not be apparent between the
individual pollutants since many of the concentrations were not
detectable, but might exist if total particulate were examined.
A significant positive correlation between the tons of waste burned
per day and the PM-10 concentration would support the MWC as the
5-2
-------
source . for these particulates. The statistical analyses were
performed on Statgraphics 3.0. The results are discussed in
Chapter 6.
5.1.2.2. CORRELATION OP MUTAGENIC ACTIVITY TO TONS OF WASTE
BURNED AND PARTICULATE CONCENTRATION — A relationship between the
amount of waste burned daily and mutagenicity of collected filters
was conducted because emissions of organic mutagens result from
incomplete combustion of municipal waste (Watts et al., 1989). A
positive significant correlation would support the Incinerator as
a possible source for mutagenicity in Rutland ambient air. This
analysis is discussed in Chapter 7. .
5.1.2.3. COMPARISON OP PCDD/PCDF CONGENER, PROFILES—
Ballschmiter et al. (1986) have suggested that the distribution
patterns of congener profiles may indicate the nature of PCDD/PCDF
sources. The congener profiles of the samples collected on January
16, February 21 and July 26, 1988 were compared to determine,if the
profiles varied between sites, days within the same season and
seasons of the year. The differences in these daily profiles could
represent contributions from different sources. The ambient air
profiles were also compared with those of potential sources (i.e.,
chimney soot and the emissions from the MWC). If the congener
profile of the MWC resembled that of ambient air on a particular
sampling day, then the MWC may have been the main source of
PCDD/PCDFs in the ambient air. Results are shown in Chapter 8.
5-3
-------
5.1.3. Quantitative Approaches to Analyzing Ambient Air Source
Contribution. The concentrations of the pollutants measured in the
ambient air (described in Chapter 3) were compared with the
concentrations predicted by the ISCST air dispersion model (as
described in Chapter 4) using the meteorologic data collected at
SLAMS, and the concentrations predicted using the meteorologic data
collected at River Street. If the MWC is the primary source for
the pollutants measured at the four ambient air monitoring sites,
then the concentrations predicted by the air dispersion modeling
(ISCST) for these sites should correspond to the measured
concentrations. The relationship between the predicted
concentrations and measured concentrations in Rutland ambient air
was analyzed using two nonparametric statistical methods.
Since the predicted concentrations from the dispersion
modeling were based on unit emission (refer to Chapter 4), they
could not be used to predict absolute ambient air concentrations.
Instead/ the model results were used to indicate the relative
ordertt or ranking, of the ambient air concentrations for the four
monitoring sites on a particular day.
In the nonparametric procedures, the actual ambient air
concentrations were replaced by their rank, in order of decreasing
concentrations within a day, with the highest predicted
concentrations getting the highest rank. The same concentrations
received a "tied" ranking. Modeled and measured concentrations
were ranked separately, then the ranks compared statistically.
5-4
-------
If the ranking of the measured concentrations for a
particular day corresponded with the ranking of the predicted
concentrations from the dispersion model for the same day, the
hypothesis that the" pollutant(s) originated at the stack would be
supported. Conversely, a difference between the order of the
measured ranks and the order, of ranks predicted by the dispersion
model would indicate either that the MWC was not the sole source
of the pollutants or that the dispersion model was inaccurate.
Ambient air concentrations of many of the pollutants could
not be quantified, as concentrations were below the limit of
detection. In the nonparametric procedures, the, impact of the
values belpw the detection limit is minimized since the analysis
is based on the ranking of the data and not the actual numerical
value. Having one value below the detection limit on a given day
would have no effect on the analysis since that site would be
identified with the lowest rank. ; When two or more values were
below the detection limit, they were treated as tied (for lowest
rank). If, on a particular day, all of the sites had values below
the detection limit, ranks could not be assigned and statistical
analysis could not be completed. For a nonparametric test based
only on the position (location) of the maximum concentration (such
as the modified sign test described below), only one of the four
sites needed to have a detectable concentration.
All statistics were conducted using Statgraphics, Statistical
Graphics System (version 3.0). The nonparametric tests used to
5-5
-------
examine the relationships between the measured and predicted
concentrations were a modified;sign test and the Friedman Two-way
Analysis of Variance. "•'-' • - -•"" ' . ,-
5.1.3.1. MODIFIED SIGN TEST — The modified sign test
compares the location of the maximum measured concentration with
the location of the maximum predicted value. The sign test is a
nonparametric test for comparing two paired samples (i.e., the
modeled and measured concentrations) . The null hypothesis was that
the maximum predicted arid maximum measured concentrations occur at
the same location (i.e., same monitoring site) on a particular day.
This test examined whether there was a direct link between the
highest modeled and measured concentrations that would be expected
if the MWC was the primary source contributing to the measured
levels in the ambient air^ -
A criteria for sufficient data to conduct this test for a
particular pollutant on "a" particular day was at lea'st one
detectable concentration among the four sites and also modeled
concentrations for the four sites when the MWC Was in operation.
To conduct this test, a plus sign was assigned for each day when
predicted values were-available from the dispersion model arid the
maximum predicted value occurred at the same location as the
measured maximum for that day. If not, a negative sign was
recorded. ''••'-- - "'-•'.'". ::'•••• •: " r • • ' ' •":
5-6
-------
If no relationship between the location of the predicted
maximum and the actual measured maximum existed for a particular
day, a "match" was expected due to chance, variation with a
probability of 0.25.
If the dispersion model did correctly identify the location
of .the highest actual .concentration significantly more than, 25% of
the time, some .correlation between the MWC stack output and the
measured ambient air levels existed.
The computation of a p-value for the hypothesis that there
was no relationship between the locations of predicted and observed
maximums was, based on the binomial distribution, as with the
ordinary sign, test, except that the.binomial parameter representing
probability of "success" was 0.25 instead of 0.5.
5.1.3.2. FRIEDMAN TWO-WAY ANALYSIS OP VARIANCE —This test
was used to analyze, the pattern of .occurrence pf the measured
concentrations and of the concentrations predicted with the
.dispersion model. It would be expected that the meteorplogic
conditions and.spatial arrangement of the sampling sites would be
such that,,one or more of the sites would receive a greater amount
of the pollutants than the others. While the actual measured
.concentrations could be analyzed by a parametric analysis of
variance (ANOVA), only relative rankings were available for the
predicted concentrations obtained from the dispersion model making
the Friedman nonparametric ANOVA the appropriate statistical test.
5-7
-------
The Friedman test is the nonparametric counterpart to the ANOVA
for a randomized complete block design. For this analysis, days
are blocks and sites are levels within the block.
Values below the detection limit were not a limitation as this
test accounts for "ties". If, on a particular day, two sites were
below the detection limit, they were considered to be tied and both
were assigned a rank of 1.5 indicating that they shared first and
second place in the ranking. (A low number meant a low rank). If
there were more than 2 sites with nondetectable concentrations,
this test could not be conducted.
In this analysis, the two data sets (measured and predicted)
were considered separately to determine how the four sites differed
in their ranking with respect to level of a pollutant. The pattern
of the rankings of the measured concentrations was compared with
the pattern of the rankings of the modeled concentrations. Finding
the same pattern of ranking for' both data sets suggested 'the
possibility that the MWC was the primary contributor to the
measured ambient air concentrations.
5.2. ENVIRONMENTAL MEDIA
Environmental media were sampled in areas surrounding the
Rutland MWC during the project period. Three rounds of
environmental sampling were conducted: October and November 1987
and June 1988. Samples collected in 1987 prior to operation of
the Rutland MWC represent background levels for comparison with
those samples taken during MWC operations. The primary objective
5-8
-------
of sampling both before and during operation was to show the-
V
incremental increase of pollutant concentrations in environmental
media, if any, caused by emissions from the facility.
The environmental assessment was qualitative and took several
approaches. Samples of the same media (e.g., soil) were pooled
across the various sites for each sampling round. The
concentrations of each pollutant for each sampling round (i.e.,
October 1987, November 1987 and June 1988) were compared using a
one-way analysis of; variance (ANOVA, a= 0.05) to determine if
pollutant concentrations differed. If the concentrations of the
sampling rounds were significantly different by the ANOVA, the
Scheffe multiple range, test was performed to determine which of
the sampling periods differed. ;;v If there ; was ; no statistically
significant difference between October and November, a two-sample
(pooled) t-test was conducted/comparing the combined pollutant
concentrations for the sampling rounds prior to commencement of
operation (i.e., background; October and November 1987) with those
from the sampling round during incinerator operation (June 1988).
To assess the validity of pooling the various sites for each
sampling period, the pollutant concentrations for each sampling
round were also compared using the" Kruskal-Wallis nonparametric
analysis of variance. This procedure applies a rank transformation
of the data (i.e., replacing the data by their ranks) and then
conducts a parametric analysis of variance on the ranks of the data
(rather than on the numerical value of the data) (Conover, 1971).
If the two procedures give nearly identical results, then the
assumptions underlying the parametric analysis of variance (i.e.,
5-9
-------
normally distributed data, equal variances) are likely to be valid,
and the pooling of the sampling sites is acceptable. However, if
the two procedures give different results, more weight is given to
the results of .the Kruskal-Wallis test, since the nonparametric
procedure is less sensitive to the effect of outliers (observations
that are unusually large or small compared with the bulk of the
data) or very nonsymmetric distributions (Conover, 1971),.
Rutland environmental media concentrations were also compared
• - ' . s' ' I >'\ ' „ ""* ' - "" *, ,' , " .,„ '„ * I . . . "J , ' . i ' "i , -. i '" " - - '
with pollutant concentrations measured at other sites within the
United States and Europe. These data from other locations were
used to assess whether the magnitude of pollutant concentrations
found in Rutland during operation of the MWC fell within the range
of concentrations found elsewhere.
5-10
-------
6. CORRELATION OF TONS OF WASTE BURNED
TO PARTICULATE CONCENTRATION
The first approach to assessing the contribution of the MWG
emissions to the pollutant concentration in Rutland ambient air
was to attempt to correlate the amount of waste burned by the
>• . • - . -
incinerator each day with the particulate matter (PM-10 fraction)
concentrations for the period of November 5, 1987 through October
6, 1988. A correlation between tons of waste burned and PM-10
concentration would suggest that the MWC was the primary source of
pollutants in the air.
Since many pollutants adhere to particulate matter and many
of the pollutant concentrations were not detectable (i.e., less
than or equal to the detection limit), the PM-10 concentrations
were compared with the tons of waste burned for each sampling day
(tpd) to determine if there was a relationship. Figures 6-1 and
6-2 display the amount of waste burned and PM-10 concentration for
each day. Simple linear regression analyses were performed to
correlate the PM-10 concentration of each site for the samples
collected from November 5, 1987 through October 6,. 1988 to the
amount of waste burned (tpd). The regression of PM-10 versus tons
of waste burned per day for each site is presented in Figures 6-
3 through 6-7. Since the SLAMS co-located site was the site for
PM-10 samples, the regression analysis was performed on both
duplicate samples.
The regression analyses indicate that PM-10 concentration is
not linearly related to the amount of waste burned. Very little
6-1
-------
6-2
-------
|
•1 Route 4 (ug/m3) iH SLAMS (ug/m 9 Hi SLAMS (ug/ms)
__ ^ O airomiiiFfn frteoooctlicoa
River St. (ug/m jliiiiiil Watkins (ug/m ) Tons burned/day
..-
'
'
'
• -
' • • * - "•
•
rrmWfflffi
KX\\\\\\\\V
,KZ%%
{TOVVyW
pn:;:;
WSS/SSs
\\Vv\\\\\VVV
•
t%#%
fess^^&
I^^u^g
•
I
ggSjfesgS^SgJSESSPftff
1
IA\\X\\\\\\\\^W\\\
^^^^^•"^""^
KSftJ^SS
CO
-o -o
o ®
- 5 1
s o>s
1— 4-» OU
CO CO O)
j; CO ^
* ^5 -
O > (Q
0) «*-
- !s ° 0>
00 •»-» jQ
r^° § o
5 00
"% <2
CM rvj -a O)
£ co S o
^ o uj *->:£
^ "5 OC^E ^.
£ a z> ^.^
0 ^ 3 ~
1 CM u. — ' Q.
~ ' » S**
m° ~ «=
S 150
~^ *- /-^
(O *- -o
O. C 0
s s5
"<,§ l«
S ° Q
-5 CD i_
10 *; o)
0 «Q_
5 0
•^ ••-•
0 k.
g £
o
1 -
DOOOOOOO
O ^ O CO . CM CO '•*•
M CM CM T- T-
6-3
-------
ID
(U
-------
CD
Tt
ru
— (0
- w
Q
§
W
to
o
EH
0)
I
vo
0)
(£ui/6ry) -
OT-Wd
6-5
-------
S>
S>
(XI
ffi
(0
(9
ru
IB
oa
S>
Tt
tn
a.
o
•H
•P
a)
O •
O 'B
Oft
Q) —
rH
o -d
•H 0)
•P C
ftX)
C Q)
0) -P
0) U)
0)
O -P
•H C
-P !3
(0 O
rH g
Q) (0
O fi
O rt
in
i
a)
&
•H
'ONOO OT-Wd
6-6
-------
-
-
-
-
-
_'.-
1 . 1 n
\
\
\
\
\ •
\
\
\
\
'
"
:,' °
*
i i i
] 1 — n — ^~i — -
i
\
\
\
\
\
\
i
\ °
\
\
\
1
'"i ; ;
i
i
\
i
i
1
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
— _J 1 i
-— — i — — i r—
i
,\
i
\
o
I
\
\
\
\
\
V
I
" V
i '.
\
V
\
'l
1
. i ' I
" i
n
a
i
i
i
, ;
i
I
1
I
: l
\ i i
i 1 1 :•
i
i
f
t
t
i
i
' I
1
I
I
I
I
i
;
r
/
;
1
i .
i
;
i
•• j"'°
o|
i
i
i
i
i
\
i
i
i
i
i
I
i
V
\
\
\
\
\
\
\
II 1 1
|
.. —
-
"
"
~
o>
o
ni
a>
•H
si
ll
09
01
H
a>
(D
§
Q
W
• s
H
EH
1
EM
0
cn
|
y
°3
p ro
cct
f»ro i
0 =t
U — '
VD
vo
(1)
-------
—- r
09
0)
m
IB
01
Q
W
pa
'w
EH
CO
ft,
o
1
(9
(0
O)
OJ
•o
<- ft
O 3.-P
CJ '>*-^"v*»
I
vo
0)
•ONOO OT-Wd
6-8
-------
of the variability (R-square values) in PM-10 concentrations is
explained by volume of waste burned per day. Table 6-1 shows the
statistical analyses of these data.
In summary, no correlation between the amount of waste burned
daily" and ambient air particulate concentration at any of the sites
was found to exist. This result suggests that the MWC is not the
sole source of particulates in the Rutland ambient air.
6-9
-------
TABLE 6-1
P-values and R-square values for Regression Analysis
According to Site
Monitoring
Site
P
Value
R
Square
Watkins Avenue
River Street
SLAMS
SLAMS (duplicate sample)
Route 4
0.21
0.38
0.25
0.26
0.55
6.3%
3.2%
5.4%
5.2%
1.5%
6-10
-------
7. MUTAGENICITY
Each of the 12 sampling periods between November 5, 1987 and
March 16, 1988 generated five TSP and five PM-10 filters from the
four ambient air monitoring stations. Only one PUF sample was
collected during.the collection time. Materials collected on the
TSP high-volume fiber filters were assayed for mutagenic activity.
Particulate concentrations were determined gravimetrically from
materials on the PM-10 filters.
The Ames Salmonella typhimirium histidine reversion assay with
strain TA98 (Maron and Ames, 1983; U.S. EPA, 1987c) was used to
determine the levels of mutagenic activity associated- with
particles from ambient air collected surrounding the Rutland MWC.
This Salmonella strain detects frameshift mutagens and historically
has been found to be the most efficient strain in detecting
mutagenicity associated with an urban air environment (Sandhu and
Lower, 1987). Dose response data were generated and mutagenicity
concentrations were calculated using the statistical method of
Bernstein et al. (1982).
The positive correlation between PM-10 particle concentration
and indirect mutagenic activity (+S9) is shown in Figure 7-1.
Statistical analysis of the data yields a slope of 0.37,
corresponding to 0.37 revertants/jug of extractable o>rganic mass,
. and a correlation coefficient of 0.74. The slope values
(revertants//ng) for dose response determinations were converted to
revertants/m3 of air. These values reflect the concentration of
7-1
-------
20 30 40
PARTICLE CONG, (/zg/m3)
50
60
Figure 7-1.
Correlation between PM-10 particle concentration in
ambient air (ug/xn) and indirect mutagenic activity
(revertants/m ) for ambient air samples collected
11/17/87 to 3/16/88. Slope = 0.37; r?= 0.74.
Source: Watts et al. (1989)
7-2
-------
mutagens found in ambient air. The co-located PM-10 samplers at
the SLAMS site show the highest concentration of particles (0-10
microns). The TSP samplers from that site show correspondingly
higher concentrations of both direct (-S9) and indirect (+S9)
mutagens. The mutagenic activities of samples from the SLAMS site
are consistently higher than those from the Watkins Avenue site.
While these results represent 12 samples collected during a three
month time period, this finding is not consistent with the initial
air dispersion modeling that had predicted that if the source of
mutagenic activity was deposition, from the incinerator, the Watkins
Avenue site, because of its proximity to the point of maximum
deposition, would have the highest amount of activity. The SLAMS
site was farthest from the incinerator but closest to the town
center and likely to be contaminated by city combustion sources.
Mutagenic activity does not correlate with the number of tons
of municipal waste burned for any sampling period (Figure 7-2).
The largest amount of waste was burned on March 16, 1988, but the
indirect mutagenic activity of the samples collected that day was
relatively small. The sampling day on which no waste was burned
(December 23, 1987) resulted in samples with relatively high levels
of indirect mutagenic activity. The data suggest a seasonal
fluctuation of both particulate concentration and mutagenic
activity from low levels in November to peak amounts in January to
low,levels in mid-March.
PM-10 particle concentrations were compared with mutagenic
activity of the samples collected on the PS-1 PUF samplers at each
7-3
-------
f. IU -
200 -
190 -
180 -
170 -
160 -
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
n -
'••'.•""- • ' .-,..,„
'. = '- •:.:>-' ' ..: •. • ,,- - • ..,'• • ' •• •-.
i ':;'». « - - .'I . - • , ,
P
tJss.li
- ° " r r
No Burn Day .
\
„
nn R llinO
\\\A \l\\\
• f-
. ».- . , *
''• :'
•;,.."•-.. "i:.
!. n
U
lj|
1 1
.
rtt tt gfl
SB
«*.
lllll
*
• , : .':
• Pr-
1/17 11/29 12/11 12/23 1/4 1/16
DATE
TONS BURNED
1/28 2/9 2/21 3/4 3/16
REV./ltT
Figure 7-2. Mutagen concentration in ambient air compared with
tons of waste burned for sampling period 11/17/87-
3/16/88.
Source: Watts et al. (1989)
7-4
-------
monitoring site. Both the pre-PUF particle filter (consisting of
the glass cartridge filter) and the PUF plug, used to collect semi-
volatile organics, were compared with the PM-10 particle
concentrations. The data from three sites (Route 4, SLAMS, River
Street) suggest that mutagenicity is primarily associated with
particle-bound organics because the PUF plug mutagenic activities
were very similar to those seen in the PUF blank. The Watkins
Avenue site, however, shows levels of semi-volatile mutagens equal
to the amount seen in pre-PUF particulate samples.
In summary, a positive correlation was seen between particle
concentration and mutagenic activity at all four sampling sites
but there was no correlation between the number of tons of waste
burned, and mutagenic activity at any of the sites. This suggests
that other sources are responsible for the mutagenic activity
observed in particles from ambient air in Rutland.
7-5
-------
-------
8. AMBIENT AIR PCDD/PCDF CONGENER PROFILES
The congeners and isomers of polychlorinated dibenzo-p-
dioxins and dibenzofurans (PCDD and PCDFs) were analyzed in ambient
air samples collected from November 1987 through February 1989 by
high , resolution gas chromatography-high resolution mass
spectroscopy (HRGC/HRMS). The congener concentrations of the
samples in ambient, air were used to make graphical displays of the
distribution patterns of the homologues. The purpose of the
congener profiles was to compare the pattern of the PCDD/PCDF
congeners between samples and potential sources. The congener
profiles, therefore were displayed both as concentrations and
relative percentages.
Distribution patterns of congeners have been used to indicate
PCDD/PCDF sources. Ballschmiter et al. (1986) determined the
existence of widespread .sources (e.g., automobiles and MWC) of
PCDD/PCDF in the environment. Tiernan et al. (1988) concluded that
PCDD/PCDFs in metropolitan areas (industrialized regions) appear
to originate from a combination of sources including MWC and motor
vehicles using profiles. The patterns of the homologue ratios for
ambient air samples collected at each site in Rutland on January
16, February 21, March 3, April 21, May 27, June 20 and July 26,
1988 were compared with each other and to homologous patterns of
potential sources (i.e., wood burning and MWC) in an attempt to
identify possible sources of the PCDD/PCDFs. Relative percentages
were used as a basis of comparison since a sample collected close
to a source could have concentrations greater than a sample
8-1
-------
collected further away, yet the pattern of congener profiles would
appear to be the same and their relative percentages would not
change because the PCDD/PCDFs originated from the same source. The
congener with the maximum concentration of each sample has a
relative percentage of 100%. Figures 8-1 through 8-25 display the
congener profiles in ambient air. The ambient air concentrations
were just above the minimum limits of detection on 2/21/88, 3/4/88,
4/21/88'and 5/27/88.
The PCDD/PCDF distribution patterns for the same day differed
among monitoring sites indicating that local sources (i.e., sources
in very close proximity to each monitoring site) influence the
distribution pattern at each site. For example, on January 16,,
1988 the relative percentages and concentrations of OCDF and PeCDF
varied greatly. The relative percentages of OCDD ranged from 23%
at Watkins Avenue to 100% at SLAMS, whereas the relative
percentages of PeCDF ranged from 0% at Route 4 (where it was not
detectable) to 100% at River Street. Occasionally, the congener
profiles for the same day at different sites resembled each other
indicating that the sites may be influenced by the same or similar
source(s) that "override" the local sources in close proximity.
On February 21, 1988, the patterns of the congener profiles
resemble each other since HpCDD, OCDD, and TCDF were predominately
the congeners with detectable concentrations (Figures 8-5 through
8-8) .
The PCDD/PCDF distribution patterns of homologues vary between
days suggesting that PCDD/PCDF sources may change with time. At
8-2
-------
(0
0)
oo
»—
c o
o *-
oo
8-3
-------
to
O
200
Q.OO
CM
-------
CJ
W
0>
2co
a. oo
cp a5<2
CO O ^
III °* -
"J C Tf
oc o
C O
a> *•
8-5
-------
(0
0)
2co
Q. 00
cc o
.
u. <
o
E
O
8-6
-------
a?
o>
0)
cc
E
3
0
O)
*—
C
0
s
I
0
c
o
O
1 ;l •''• ' ' - ^^^^
I i j
i . ' •
•HHi
i ^
I
i 1
*
' ' 1
"^
o
Q
j { • ^H.
j 04
i 2 .
_
o
~
0 '
a
i
1 1 . — 1
™
I f™
00 fl> ^J
Sfg
^0^
£ tj
LL. < CO
*- w.
CO
g> •*-
!5
E
8-7
-------
CM
co
CD
CD CD CM
CO CD ^
UJ gj
OC O CO
^o ^
O _ o
8-8
-------
CVJ
(0
CD
2co
O. 00
N- o> w
CO
-------
8-10
-------
o
a
o
o
o
o
a
I
a
o
X
z:
u_
a
o
H—
O 00
o_ J~
1- "fr
O) CD O
I C ^x
CO 0) CO
UJ c J
S3?
~~ •— >
u- < ~
C w
o o
!5 **~
E
CM
8-11
-------
(0
0)
Q.OO
?is
00 ® CO
m 5 M-"
oc o^
-300
~ +••
E
8-12
-------
W
a)
0(0
T-
T- 0) O
i c «^
oo o eo
gig
I*
JQ
E
8-13
-------
CM
CO
0)
2 co
Q. oo
< CO
•+-• l_
n o
Q) *"
!o
E
CM
8-14
-------
CVJ
(0
0)
o oo
co cv
LJJ c +:
oc o co
Z)0 -,.
g.h ®
Is
JQ
E
o
8-15
-------
CM
(0
o
2co
Q. 00
oo CD
o>
in c
CC o
Z)0
o
E
8-16
.
-------
CO
0)
2 co
Q. CO
I C X
00 "*
O) „
UJ C CO
cc o c
8-17
-------
CM
8-18
-------
W
-------
CO
0>
N-
-------
(0
Q>
2oo
Q_ 00
CO <">
z>0i
O.v_<
U. < CO
= 0
Q) i-
o
8-21
-------
CM
CM
to
0>
O GO
»_ O
0) CM
oo »-
LLJ C +J
.0) ,0
la
E
8-22
-------
(0
o
T- *- O
CM ® CM
I C ««x
(O (O
O) '
w c w"
.
ja
8-23
-------
(0
-------
CO
0)
O 00
^s.
eo u. <0
Sȣ
iS":
UJ C +1
= oco
0) O
8-25
-------
CO
-------
(0
10
CM
I
co
LLJ
CC
»- (O
ocv;
o> r*-
O)
2^3
8-27
-------
River Street, the TCDF and PeCDF have relative percentages of 88
and 100% on January 16 (Figure 8-2) , respectively, but on June 20
the relative percentages decrease to 0% (Figure 8-20).
The congener profiles of ambient air were compared with the
congener profiles based on the stack emission of the MWC and the
emissions from wood burning systems. Emissions from wood burning
systems have been included for the purpose of possible
identification of source contribution, because the air monitoring
sites in Rutland encompass residential wood burning in the
proximity of the MWC. Because of the.lack of Rutland-specific data
on the PCDD/PCDF concentrations in fly ash from residential wood
burning, the arithmetic mean of the PCDD/PCDF concentration of nine
chimney soot samples from wood burning home heating systems in
Germany (Thoma, 1988) was used. The.congener profile for.chimney
soot is displayed -in Figure 8-26.
The PCDD/PCDF congener profiles of the ambient air samples
collected during the winter months were compared with the congener
profile of chimney soot. The congener profile of Watkins data on
January 16, 1988 does resemble the profile of the soot with the
exception of PeCDDs. However, the other congener profiles of the
ambient air samples collected on January 16 and February 21 do not
resemble those of chimney soot. For many of the air samples, the
OCDDs have high relative percentages while for the chimney soot,
OCDD has a low relative percentage. The relative percentage of
PeCDF of many ambient air samples was low (range 0-75%) while the
relative percentage of PeCDF of chimney soot was high (100%).
8-28
-------
o
o
CO
•c
e c
£ 9>
o
c
o
O)
c
o
O
8-29
-------
Congener profiles were developed'for the MWC stack emissions
that were measured on March 8, 9 and 10, 1988. *»Stack emission
testing of the MWC was required under the Air Pollution Control
Permit for the State of Vermont (Agency of Environmental
Conservation, State of Vermont, 1986). The emission concentration
of PCDD/PCDFs was one of many pollutants that was sampled for
fourhours on three days as discussed in Chapter 4. The congener
profiles of the stack emissions from March 8, 9, and 10 are
displayed in Figures 8-27 though 8-29.
The profiles of stack emission have similar PCDD/PCDF
distribution patterns. The profile for March 8, 1988 (Figure 8-
27) differs from the other two in the HpCDF. and OCDF relative
percentages, but the reason for this is unknown and may be due to
a change in operation parameters. The concentrations of HpCDF and
OCDF are greater than that of the congeners for the stack emissions
collected on both March 9 and 10 (Figures 8-28 and 8-29).
The congener profiles of the stack,emissions were compared
with profiles of the ambient air samples collected at Watkins
Avenue on May 27, June 20 and July 26, 1988. The ambient air
samples collected on May 27 and June 20 were chosen for comparison
because they were the sites predicted by the ISCST twice using the
SLAMS and River Street meteorologic data to receive more of the MWC
pollutants than the other sites for these days. The Watkins Avenue
ambient air sample collected on July 26 was compared because it was
predicted to receive the greatest concentration for all sites for
all sampling ctays. When me congener profiles of the ambient air
are compared with the profiles of the stack emissions, the PCDF
8-30
-------
CM
8-31
-------
CO
CO
CO
£
UJ
co
i
co
UJ
o
CO
o
- CO
.— TJ
*~ CO
o -S
-------
CO
c
o
'
CO
E
UJ CO
O)
CM
-------
congener patterns show a resemblance but the PCDD congener patterns
do not. In general, the ambient air samples have higher HxCDD and
OCDD relative percentages than the stack emissions.
The comparison of the ambient air congener profiles between
each site indicates that there is not a specific distribution
pattern between sites (i.e., the profiles vary between sites). The
ambient air profiles also vary for each day. Because of the
variations occurring between sites, days and sources, it is
unlikely that the PCDD/PCDFs were from wood burning or the MWC
alone, but a variety of sources.
Uncertainty was introduced into the interpretation of these
congener profiles due to the lack of MWC emission data. Since the
tons of waste burned fluctuated between days and the MWC stack
emissions were tested only on three days, it is not known if the
»
profiles of the emissions changed over time. Therefore, the graphs
that were used as the basis of comparison to determine if the MWC
was the major source of PCDD/PCDFs may not have been accurate.
8-34
-------
9. ANALYSIS OF MODELED AND MEASURED
, • .-, - AMBIENT AIR CONCENTRATIONS
., . . Several, approaches were used to estimate human exposure to
•the,pollutants emitted from the MWC. The pollutant concentrations
measured in Rutland; . ambient air when the incinerator was in
.operation represented the total concentration of each pollutant
from both the incinerator and other .sources. In order to determine
if the concentrations of measured pollutants were primarily from
the MWC, the proportion of the pollutants attributable to other
sources needed to be .assessed. This chapter presents the results
of the statistical comparison of measured and predicted ambient air
concentrations of the pollutants as a way of assessing source
apportionment,-,since an inventory of other sources for the measured
pollutants .was not available. Statistical results for the
environmental media are reported in Chapter 10.
As discussed in Chapter 3, few of the pollutants had
detectable concentrations. Table 9-1 shows the occurrence of the
maximum detectable concentrations of the various metals and B[a]P
that were detectable at the four ambient air monitoring sites.
PCDD/PCDFs on Table 9-1 indicate the maximum concentration at.the
monitorings when all of the congeners hdd detectable
concentrations. While B[a]P had a large percentage of samples
above the detection limit (43/131), only a few (3/43) occurred on
days when meteorologic data needed for dispersion modeling were
available, precluding a statistical analysis of the data. There
were sufficient data in the detectable range for lead to enable
9-1
-------
TABLE 9-1
Occurrence of Maximum Detectable Concentrations , in Ambient Air
*ll/05/87
*ll/17/87
*ll/29/87
*12/ll/87
12/23/87
01/04/88
*01/16/88
*01/28/88
*02/09/88
*02/21/88
*03/04/88
*03/16/88
*03/28/88
04/09/88
*04/21/88
*05/03/88
*05/27/88
*06/08/88
*06/20/88
*07/14/88
*07/26/88
*08/07/88
*08/19/88
08/31/88
09/24/88
10/06/88
10/18/88
SLAMS
-NA-
BaP
Pb BaP
Pb Ni BaP
Pb
Pb BaP
Pb BaP
Pb BaP
Pb
BaP
Pb
Pb BaP
Pb BaP
Pb
Pb
Pb
Pb
Pb
As Pb
BaP
Pb
Watkins Ave.
Pb
As
Ni -
PCDD/PCDF "
Be Cd
AS '
As
~NA-
pb,- ; *
Pb , : * v
AS '""'
f
-'-^,
River St.
BaP
,
Cr
PCDD/PCDF
Be Pb
PCDD/PCDF
H , , ,,
Pb
Be
Route 4
-NA-
Pb
Be Cd
As Pb
-NA-
As
-NA-
Pb
-NA-
Pb
9-2
-------
TABLE 9^1 (continued)
10/30/88
11/11/88
,11/23/88
12/05/88
12/17/88
02/03/89
02/15/89
SLAMS .
Pb .....' '
Pb .
.BaP
PCDD/PCDF
Pb BaP
Pb BaP
BaP
BaP
Watkins Aye.
PCDD/PCDF
PCDD/PCDF
River St.
Pb
-
PCDD/PCDF
Route 4
* = Combustor operating ,
Shaded cells indicate locations of maximum modeled concentrations
using SLAMS meteorlogic data.
9-3
-------
detailed statistical analysis (the criterion for sufficient data
is discussed in Section 5.1.3.)- PCDD/PCDFs were statistically
analyzed as 2,3,7,8-TCDD equivalent concentrations.
9.1. COMPARISON OF MEASURED AND MODELED LEAD
As discussed in Chapter 5, predicted and measured ambient air
concentrations of lead were statistically compared using two
nonparametric methods, the modified sign test and the Friedman
nonparametric ANOVA.
9.1.1. Modified Sign Test Analysis for Lead. This test was
conducted twice; once, comparing the measured ambient air
concentrations with the concentrations predicted by the dispersion
model using meteorologic data collected at the SLAMS, and again
comparing the measured ambient air concentrations with the
concentrations predicted by the dispersion model using meteorologic
data collected at River Street. The two different meteorologic
data sets were used for this statistical analysis to assure that
the results obtained when the SLAMS data were used were not
compromised in any way due to the limitations in the collection of
the SLAMS data, as described in Section 4.1.
There were eleven days for which there were both dispersion
model data for the SLAMS and a complete set of measured lead
concentration data. These days are listed in Table 9-2 (08/19/88
is not included in the first analysis as there were no meteorologic
data for the SLAMS on this day) . Any day for which lead
concentration was not available for one or more of the monitoring
9-4
-------
TABLE 9-2
Ranks for the Four Sampling Sites Based on Both Measured8 and
Modeled1* Lead Concentrations
Date
01/16/88
01/28/88
02/21/88
03/04/88
03/16/88
04/21/88
05/27/88
06/20/88
07/14/88
07/26/88
08/07/88
08/19/88
SLAMS
Mo
3
4
3
2
3
3
1(1)
3(3)
3(2)
3
2(3)
(1)
Me
4
4
4
2
4
4
3
3
4
4
3
4
Watkins Ave.
Mo
4
3
2
2
4
1.5
4(4)
4(4).
4(4)
4
4(2)
(4)
Me
3
3
2
3
1.5
1
4
4
1
3
2
2
River St.
Mo
2
2
4 >
2
1
4
3(2)
1.5(2)
2(3)
2
3(4)
(2)
Me
1
1
2
1
.1.5
2
2
2
3
1
1
1
Route 4
Mo
1
1
1
4
2
1.5
2(3)
1.5(1)
1(1)
1
1(1)
(3)
Me
2
2
2
4
3
3
1
1
2
2
4
3
Me": Ranks based on measured concentration data
Mob: Ranks based on dispersion model using SLAMS meteorologic data.
Ranks based on dispersion model using River St. meteorologic
data are in the parentheses.
9-5
-------
sites was eliminated since the modified sign test compares the
highest predicted and highest ,observed concentrations and missing
data precluded the determination of« "highest". Values that were
not detected could still be analyzed unless concentrations for all
four locations were not detectable for ^particular day, in which
case a "highest" value could not be determined, ;. .
Of the eleven days there were a total of fpur days wherein
there were matches between predicted and,, observed maximums. As
shown in Table 9-1, these days • are 01/28/88 (SLAMS), 03/04/88
(Route 4), 05/27/88 (Watkins Ave.) and 06/20/88 (Watkins, Ave.),
The probability of a random match, between maximum observed and
maximum predicted concentrations on any particular day with four
sites is 0.25. From the binomial distribution, the probability
(p-value) of four or more matches out of eleven trials is 0.286.
Since this result was not statistically significant (p> 0.05), the
number of matches observed was not greater! than expected due to
chance variation alone, i.e., the maximum predicted and measured
concentrations of lead occurred at the.same site only by chance.
One reason for the small number ,pf matches was that SLAMS
consistently showed the highest levels of lead even though this
site was predicted to have, the maximjim concentration only once
during these eleven days. This suggests the possibility that the
primary source of lead at SLAMS was something other than the MWC.
To eliminate the possibility that, the results of the above
modified sign test might be biased by consistently high lead levels
9-6
-------
at SLAMS originating front an unidentified source, the SLAMS site
was excluded and the modified sign test repeated for the remaining
three sites. These results are shown in Table 9-3.
With the elimination of SLAMS from the analysis, the number
of days for which data was available was reduced to ten because on
one of the original eleven days (2/21/88) no lead was detected at
the remaining three sites (i.e., Watkins Avenue, River Street and
Route 4). The maximums for both measured concentrations and
predicted concentrations were compared for the three sites giving
a total of six matches out of ten (01/16/88 Watkins Ave., 01/28/88
Watkins Ave., 03/04/88 Route 4, 05/27/88 Watkins Ave., 06/20/88
Watkins Ave. and 07/26/88 Watkins Ave.).. The probability of a
random match on a particular day with three sites is 0.33. From
the binomial distribution, the probability of 6 or more matches out
of 10 trials is 0.073. This p-value of 0.073 suggests the
relationship between the modeled concentrations and the measured
concentrations was not significant at the 0.05 level, i.e., the
primary source of lead at these sites was not likely to be the MWC.
The modified sign test was repeated using the locations
predicted to have maximum concentrations from the dispersion
modeling with the River Street meteorologic data. Complete
information to perform the test was available for five days
(05/27/88, 06/20/88, 07/14/88, 08/07/88, 08/19/88) as shown on
Table 9-2. The probability of a random match between the location
of the maximum observed and maximum predicted concentrations on any
particular day with four sites is 0.25. There were two matches
between predicted and measured maximums (Watkins Ave. on 05/27/88
9-7
-------
TABLE 9-3
Ranks for Three Sampling Sites (SLAMS Excluded) Based on Both
Measured* and Modeledb Lead Concentrations
Date
01/16/88
01/28/88
02/21/88
03/16/88
04/21/88
05/27/88
06/20/88
07/14/88
07/26/88
08/07/88
08/19/88 „
Watkins Ave.
Mo
3
3
1.5
3
1.5
3(3)
3(3)
3(3)
3
3(2)
(3)
Me
3
3
2
1.5
1
3
3
1
3
2
2
River St.
Mo
2
2
1.5
1
3
2(1)
1.5(2)
2(2)
2
2(3)
(1)
Me
1
1
1
l."5
2
2
2
3
1
1
1
Route 4
Mo
1
1
3
2
1.5
1(2)
1.5(1)
1(1)
1
1(1)
(2)
Me
2
2
3
3
3
1
1
2
2
- 3
3
Me1: Ranks based on measured concentration data
Mob: Ranks based on dispersion model using SLAMS meteorologic data.
Ranks based on dispersion model using River St. meteorologic
data are in the parentheses.
9-8
-------
and 06/20/88). The probability of finding two or more matches
out of five independent trials (or days) as a random occurrence is
0.367, indicating there is no relationship between the location of
the modeled and measured maximum lead concentrations.
The analysis was again repeated excluding the SLAMS; the
results are shown in Table 9-3. Again, there were two matches (the
same two as when SLAMS was included) of the location of the maximum
predicted and modeled lead concentrations. The probability of a
random match between the location of the maximum observed and
maximum predicted concentrations on any particular day with three
sites is 0.33. From the binomial distribution, the probability of
two or more matches out of 5 trials is 0.532. Therefore, there was
no evidence for a correlation between the measured lead
concentrations and the lead concentrations predicted by the
dispersion model (using the River Street meteorologic data) at
these three monitoring sites, supporting the results of the
analysis using the SLAMS meteorologic data for the predicted
concentrations. It should be noted, however, that the power of the
test to detect a deviation from the hypothesis of random matching
of the predicted and measured maximums would be quite low with only
five trials in the experiment.
The findings of no relationship between the maximum measured
and modeled lead concentrations are consistent whether the SLAMS
or River Street meteorologic data are used, suggesting the quality
of the SLAMS data is not compromised. Furthermore, the
consistently higher lead concentrations of the SLAMS (relative to
the other three sites) does not appear to influence the finding of
9-9
-------
no relationship between the modeled and measured concentrations
since the results are the same whether the site is included or
excluded from the analysis.
9.1.2. Friedman Nonparametric ANOVA for Lead. In the preceding
modified sign tests, an attempt was made to establish a direct
relationship between the predicted and measured lead levels. : In
this analysis the pattern in the ranked levels of lead was
established for the two data sets (measured and predicted)
separately. These two patterns were then compared to evaluate the
concordance between them. This test was conducted using only the
concentrations predicted from the air dispersion model using the
SLAMS meteorologic data, since a pattern of relative rankings for
the four sites could not be ascertained using the limited
meteorologic data available for River Street. Additionally,
information gleaned from conducting the modified sign test with
these data showed the results were similar using both meteorologic
data sets.
The daily ranks of the .four sampling sites, based on both
measured and modeled lead .concentrations are shown in Table 9-2.
Only days for which both the measured data were available for all
four sites and the meteorologic data were available for estimating
concentrations by the dispersion model were analyzed; eleven days
were used (08/19/88 in Table 9-2 was excluded).
The Friedman test statistic based on the ranks of the measured
concentrations was 13.4, which has a p-value of 0.0038. This
indicated a statistically significant difference between the sites
9-10
-------
for the measured concentrations of lead. The Friedman test
statistic for the ranks of the modeled concentrations was 11.5 with
a p-value of 0.0095, also indicating a significant difference
between the sites.
The average ranks for the eleven days associated with the four
sites, shown in Table 9-4, indicate that the measured and the
modeled concentrations did not follow the same pattern. The
dispersion model predicted the highest rank (i.e., lead
concentration) to occur at Watkins Avenue and the lowest at Route
4. The actual measured lead concentration ranked highest at SLAMS
and lowest at River Street.
Because of the possibility that SLAMS was receiving lead from
an unidentified source as discussed in Sections 9.1...1, the analysis
was repeated without that site. Table 9-3 shows the ranks of the
measured and predicted concentrations for the ten days for the
remaining three sites. The Friedman test statistic based on
measured concentrations is 3.13 with a p-value of 0.209. The test
statistic based on the modeled ranks was 9.56 with a p-value of
0.008. The average ranks associated with the three sites are shown
in Table 9-5. The average ranks for the modeled concentrations
suggest there should be a difference in lead concentration due to
the MWC, while the ranks of the measured concentrations do•not show
this difference.
The Friedman ANOVA test for the rank of the modeled and
measured concentrations indicated the sites differed in
concentrations. However, the pattern of lead concentrations
(highest to lowest concentration) differs between the modeled and
9-11
-------
TABLE 9-4
Average Ranks of Lead Concentrations for Four Sampling Sites
Site
Sample Size
Average Rank
Measured
Modeled
SLAMS 11
Watkins Ave. 11
River St. 11
Route 4 11
3.55
2.50
1.59
2.36
2.73
3.32
2.41
1.55
9-12
-------
TABLE 9-5
Average Ranks of Lead Concentrations for Three Sampling Sites
(Excluding SLAMS)
Site
Sample Size
Average Rank
Measured
Modeled
Watkins Ave. 11
River St. 11
Route 4 11
2.22
1.59
2.18
2.64
2.00
1.36
9-13
-------
measured concentrations. This finding indicates that the MWC is
not the primary contributor of lead to the monitoring sites.. Had
the MWC been the primary contributor, the pattern should have' been
the same. The results of the Friedman test excluding SLAMS differ
from those including the SLAMS (showing a statistically significant
difference between the sites), reaffirming the observation-that the
higher lead concentrations at SLAMS may be due to additional
sources of lead.
The results of both the modified sign test and the. Friedman
ANOVA suggest there are other sources contributing to the measured
lead levels and that the MWC is not the primary source responsible
for the measured lead levels.
9.2. COMPARISON OP MODELED AND MEASURED PCDD/PCDF .
The statistical comparison of the measured and modeled
concentrations of PCDD/PCDFs involved the .conversion of .the
PCDD/PCDF isomer concentrations to 2,3,7,8-TCDD equivalents. The
actual measured concentrations of individual isomers or congeners
would have been the most appropriate variable ..for comparison with
the modeled concentrations. However, lack of adequate isomer-
specific detectable concentrations for days on, which the
incinerator was operational and lack of corresponding meteorologic
data needed for air dispersion modeling preclude such a comparison.
For example, 2,3,7,8-TCDF was detectable at one or more monitoring
sites on only 6 days, and 2,3,4,7,8-PeCDF was detectable at one or
more monitoring sites on only 4 days, for which there are
meteorologic data and the incinerator was operating. Detectable
9-14
-------
concentrations of 2,3,7,8-TCDD, 2,3,7,8-HxCDD, and 2,3,7,8-PeCDD
occurred primarily during late 1988 and early 1989 when the
incinerator was not operating. OCDD was measured in ambient air
on nine days at concentrations greater than that detected in the
field blanks and method blanks. Since the OCDD concentrations for
the nine days reflected concentrations present in ambient air and
not just contamination from reagents and the analyticcil procedures,
they could be compared to the modeled concentrations possible for
these days. ,
There is uncertainty attendant in using the 2,3,7,8-TCDD
equivalent concentration in this context. As discussed in Section
3.1v4., the 2,3,7,8-TCDD equivalent concentration in ambient air
is calculated by applying both assumptions of proportionality of
isomers and equivalence of concentration to the detection limit if
i
the isomer-specific concentration was not detectable, and the TEF
approach (the specifics of these calculations are delineated in
Section 3.1.4.). The resultant concentration represents a
concentration "weighted" by the toxicity of the. isomers and has
been used for the determination of human health risks (U.S. EPA,
1989). The use of the 2,3,7,8-TCDD equivalent concentration
introduces uncertainty since PCDD/PCDF congener profiles (described
in Chapter 8) may be altered during transport arid deposition
(Eitzer and Kites, 1989). However, since the processes by which
these profiles are altered are not fully understood, possible
changes in congener profiles have not been accounted for here.
9-15
-------
9.2.1. Modified Sign Test Analysis for PCDD/PCDF. The modified
sign test was performed twice, once using the 2,3,7,8-TCDD
equivalent concentrations and once using OCDD concentrations.
These measured concentrations were compared with the concentrations
predicted by the dispersion model using meteorologic data collected
from SLAMS; all four monitoring sites were compared. The modified
sign test was not repeated for the 2,3,7,8-TCDD equivalent or OCDD
concentrations predicted by dispersion modeling using the River
Street meteorologic data, as there were only three days for which
complete information was available and the power of this test for
detecting a correlation is very low with only three trials.
Data for the calculated 2,3,7,8-TCDD equivalent concentrations
(henceforth referred to as "measured") and modeled concentrations
were available for nine days when the combustor was operating. The
relative rankings of the four sites for these dates are listed in
Table 9-6.
Results of the modified sign test indicate that the modeled
maximum coincided with the measured maximum concentration on six
of the nine days (01/16/88 Watkins Ave., 03/16/88 Watkins Ave. and
04/21/88 River Street, 05/27/88 Watkins Ave., 06/20/88 Watkins
Ave., 08/07/88 Watkins Ave.). The probability that this was the
result of a random matching is 0.010, showing. the number of
observed matches was greater than expected due to chance alone.
This statistically significant finding suggests there is a
relationship between the measured 2,3,7,8-TCDD equivalent
concentrations and those concentrations predicted to occur from
the MWC emissions.
' 9-16
-------
TABLE 9-6
Ranks for Four Sampling Sites Based on Both Measured* and
Modeled15 2,3,7,8-TCDD Equivalent Concentrations
Date
01/16/88
02/21/88
03/04/88
03/16/88
04/21/88
05/27/88
06/20/88
07/26/88
08/07/88
SLAMS
Mo
3
3
2
3
3
1
3
3
2
Me
2
1
1
2
2
1
3
4
3
Watkins Ave.
Mo
4
2
2
4 ,
1.5
4
4
4
4
Me
4
3
2
4
1
4
4
1
4
River St .
Mo
2
4
2
1
4
3
1.5
2
3
Me
3
2
4
3
4
2
2
3
1
Route 4
Mo
1
1
4
2
1.5
2
1.5
1
1
Me
1
4
3
1
3
3
1
2
2
Me": Ranks based on measured concentration data
Mob: Ranks based on dispersion model using SLAMS meteorologic data,
9-17
-------
The modified sign test was conducted for OCDD, the only
PCDD/PCDF congener for which adequate data were available for
statistical analysis, as discussed above. A comparison of the
"\
ranks of the measured concentrations and the ranks predicted from
the dispersion model is displayed in Table 9-7. The modified sign
test applied to these data showed only one match (01/16/88 Watkins
Ave.) of the maximum predicted and maximum measured concentrations
out of nine days. The p-value for this test was 0.925, indicating
there is no correlation between the measured and predicted OGDD
concentrations. This was in contrast to the 2,3,7,8-TCDD
equivalent concentration data that suggested a correlation between
measured and predicted concentrations.
9.2.2. Friedman Nonparametric ANOVA for PCDD/PCDF. The results
of the modified sign test suggested a correlation between the
measured and modeled maximum concentrations of 2,3,7,8-TCDD
equivalents, but the results of the Friedman analyses examining
the pattern in the ranked levels of 2,3,7,8-TCDD equivalent
concentrations for the two data sets (measured and predicted) did
not provide strong support for that conclusion.
The Friedman test was conducted using only the concentrations
predicted from the air dispersion model with the SLAMS meteorologic
data. The test statistic for comparing the measured 2,3,7,8-TCDD
equivalent concentrations over the four sites was 2.73, which has
a p-value of 0.43. This indicated that the hypothesis of equality
of the four sites based on the measured concentrations cannot be
rejected; the 2,3,7,8-TCDD equivalent concentrations at the four
9-18
-------
TABLE 9-7 ' "' •'•"
Ranks for Four Sampling Sites Based on Both Measured8 arid
Modeled1" OCDD Concentrations
Date
01/16/88
02/21/88
03/04/88
03/16/88
04/21/88
O5/27/88
06/20/88
07/26/88
08/07/88
SLAMS
Mo
3
3
2
3
3
1
3
3
2
Me
3
1
1
2.5
4
4
4
4
2.5
Watkins Ave.
Mo
4
2 ' '
2
4
1.5
4
4
4
4
Me
4
2.5
3
2,5,
3
2
3
3
2.5
River St.
Mo
2 ' .
4
2
• 1 , •
4
3
1.5
2
3
Me
1
2.5
4
4
1.5
2
1.5
2
1
Route 4
Mo
1
1
4
2
1.5
2
1.5
1
1
Me
2
4
2
1
1.5
2
1.5
1
4
Me": Ranks based on measured concentration data
Mob: Ranks based on dispersion model using SLAMS meteorologic data,
9-19
-------
sites are similar. The Friedman analysis of the rankings of the
modeled concentration for the same nine days gave a value of the
test statistic of 7.54 with a p-value of 0.06. While not
significant at the 0.05 level, this p-value indicates that there
is more difference in the relative rankings of the four sites for
modeled concentrations than with the measured concentrations. The
average ranks for both measured and modeled concentrations of the
four sites are shown in Table 9-8.
The Friedman tests were repeated with the OCDD concentrations.
The daily ranks of the four sampling sites based on both measured
and modeled OCDD concentrations are shown in Table 9-7. The same
nine days as used for 2,3,7,8-TCDD equivalent concentrations were
analyzed. The Fsiedman analysis gave a test statistic of 3.11 (p-
value = 0.38) for the measured concentrations of OCDD and 6.76 (p-
value « 0.08) for the predicted OCDD values. This is similar to
the result obtained for the 2,3,7,8-TCDD concentrations; that is,
there is no statistically significant difference in the measured
or modeled concentrations between the four ambient air monitoring
sites. The results of the Friedman test for OCDD support the
findings of the modified sign test, suggesting the MWC is not the
primary contributor of OCDD to the monitoring sites. The average
ranks for the nine days associated with the four sites are shown
in Table 9-9.
For both the 2,3,7,8-TCDD and OCDD, the average ranks of the
modeled concentrations suggest that the concentrations should
differ but the actual concentrations are very similar as shown by
the average ranks of the measured concentrations. While the
9-20
-------
TABLE 9-8
Average Ranks of 2,3,7,8-TCDD Equivalent Concentrations for
Four Sampling Sites
Site
Sample Size
Average Rank
Measured
Modeled
SLAMS 9
Watkins Ave. 9
River St. 9 .
Route 4 9
2.1
3.0
2.7
2.2
2.6
3.3
2.5
1.7
9-21
-------
TABLE 9-9
Average Ranks of OCDD Concentrations for Four Sampling Sites
Site
Sample Size
Average Rank
Measured
Modeled
SLAMS 9
Watkins Ave. 9
River St. 9
Route 4 9
2.8
2.9
2.2
2.1
2.6
3.0
2.8
1.6
9-22
-------
modified sign test was statistically significant for 2,3,7,8-TCDD
equivalent.concentrations, the results of the Friedman analyses do
not support the findings. No relationship between the modeled
andmeasured concentrations of OCDD were found. However, a direct
relationship between the results of the OCDD analyses and those of
2,3,7,8-TCDD equivalent would not necessarily be expected. The
2,3,7,8-TCDD equivalents include OCDD in the determination, albeit
OCDD has a small TEF value and would not be expected to contribute
substantially to the 2,3,7,8-TCDD equivalent value even if present
at high concentrations. Instead, the other more "toxic" isomers
(those with higher TEF values) present at levels close or equal to
the detection limit most likely influence the overall 2,3,7,8-TCDD
equivalent concentration as calculated in this report. This
information, then, adds uncertainty to the meaning of a significant
statistical finding, particularly if not supported by subsequent
statistical analyses or by other congeners. Taken together, these
results suggest the MWC is not the primary contributor to PCDD/PCDF
concentrations at the ambient air monitoring sites and that there
are other sources for these pollutants.
9.3. CONCLUSION
The statistical analyses of the measured and predicted lead
and PCDD/PCDF data suggest that there are other sources
contributing to these measured levels and that the MWC was not the
primary source of the pollutants. This finding is supported by the
observation that other pollutants, which only occasionally were
found at detectable concentrations, were often located at different
9-23
-------
sites on the same day. Table 9-1 shows the location of the maximum
detectable concentrations for the pollutants. When two or more
pollutants that rarely show up at levels above their detection
limits occur on the same day but at different sites, such as on
03/04/88, it seems unlikely that they would have originated from
the same source unless there were changes in the meteorologic
conditions coinciding with changes in composition of the stack
output.
9-24
-------
10. LONG-TERM AIR DISPERSION MODELING
Additional modeling of the MWC stack emissions was performed
to determine the magnitude of the long-term ambient air
concentrations of pollutants in Rutland. The Industrial Source
Complex Long-Term (ISCLT) model utilized one year of Rutland
meteorologic data collected at the meteorologic monitoring sites
once in operation. The ISCST model as discussed in Chapter 4
predicted daily concentrations based on the meteorologic data of
Rutland for the sampling days when the MWC was in operation. This
chapter describes both the ISCLT modeling methodology and the
modeling results.
10.1. MODELING METHODOLOGY
The ISCLT model was run using some information that was also
incorporated into the ISCST and the initial ISCLT modeling for the
placement of the monitoring sites (as discussed in Chapter 2) . The
source characteristics of the MWC and meteorologic data were input
parameters for the ISCLT model. The source parameters, described
in Section 1.3, consisted of the same general information about the
MWC as was used in the ISCST and previous ISCLT modeling. Exhaust
from the incinerator was vented from a single stack which was 1.040
m in diameter and 50.3 m high. The exhaust gas exited at a
temperature of 327.60 K and a velocity of 15.24 m/s. „ Unit emission
rate (1 g/s) was assumed so that the predicted concentrations from
10-1
-------
the ISCLT could later be converted to pollutant-specific
concentrations using the stack emission rate for each pollutant.
The meteorologic data input consisted of Glens Falls, New York
cloud cover information and Rutland, Vermont wind speed and wind
direction. Glens Falls cloud cover information was .used in the
ISCLT as in the ISCST because no such information was available for
Rutland. Glens Falls has the closest National Weather Service
Station and has similar topography to Rutland (see Section 4.2.)
Wind speed .and wind direction data were collected at 3
monitoring sites in Rutland (as discussed in Chapter 2) : SLAMS,
River Street and Watkins Avenue. The ISCLT was run 3 separate
times using the available data collected at each site during 1988.
The data of Watkins-Avenue were modeled even though the wind speeds
observed during the summer months were much lower than those
observed during the other months.
The ISCLT required meteorologic data in the STability ARray
(STAR) format. A STAR summary is a statistical tabulation of joint
frequency of occurrence of wind speed and wind direction
categories, classified according to the Pasquill stability
categories (U.S. EPA, 1986a). STAR summaries combining wind speed,
wind direction and cloud cover were based on the available 1988
data. A separate STAR summary was developed for each site. Each
STAR summary had six stability classes and a wind-speed category
consisting of various combinations of wind speed and Pasquill
stability categories. The wind speed categories used for modeling
10-2
-------
were 0-0.89 m/s> 0.90-2.46 m/s, 2.47-4.47 m/s, 4.48-6.93 m/s, 6.94-
9.61 m/s, and 9.62-12.5 m/s.
The ISCLT was run using the STAR summary and anemometer
height for each monitoring site and the Urban 3 Mode. The Urban
3 Mode was used because the incinerator was located in a rural area
with complex terrain. The ISCLT was run with the same polar and
discrete receptors for each of the data sets (i.e., Watkins Avenue,
SLAMS, and River Street) as used for the initial long-term
modeling. A total of 160 polar receptors and 59 discreite receptors
were used with each modeling run. The polar receptors were located
at radial distances of 0.2, 0.5, 1.0, 2.0, 5.0, 10, 20, 30, 40, and
50 km from the MWC for 16 wind directions. The discrete receptors
were used to better define the point ,of maximum deposition.
The output from each ISCLT modeling run was a prediction of
long-term ground-level ambient air concentrations at each of the
receptors based on an emission rate of 1.0 g/s. To determine the
maximum pollutant-specific ground-level ambient air concentrations,
the predicted concentrations at each receptor were multiplied by
the maximum measured stack emission rate of the pollutant. The
stack emission rates used were from the stack testing in March 1988
(see Section 4.2.1).
10.2. ISCLT RESULTS
The five highest predicted concentrations and the respective
receptor location using the metebrologic data from the 3 sites are
10-3
-------
TABLE 10-1
Results of Site-Specific ISCLT Modeling
UTM Coordinate
Direction
Relative
to MWC
Predicted Annual Ground-
Level 'Concentration of
Pollutant (fj,g m)*
River Street
661700/4829950
661700/4830050
661700/4829900
661700/4830200
661623/4829885
SLAMS
661700/4829950
661700/4830050
661700/4830200
661776/4829885
661700/4829900
Watkins Avenue
North, 250 m
North, 350 m
North, 200 m
North, 500 m
NNW,
200 m
North, 250 m
North, 350 m
North, 500 m
NNE, 200 m
North, 200 m
1.4
1.2
1.1
0.97
0.94
1.3
1.2
1*0
0.99
0.89
661700/4829950
661700/4830050
661700/4829900
661700/4830200
661776/4829885
North, 250 m
North, 350 m
North, 200 m
North, 500 m
NNE, 200 m
1.8
1.5
1.4
1.1
0.97
Based on unit emission (1 g/s) (See text.)
10-4
-------
shown in Table 10-1. The receptors having the highest ground-
level ambient air concentrations were all within 500 m of the
\
incinerator and were all north of the incinerator. Receptors
located south to southwest of the MWC were consistently the
receptors with the lowest ground-level ambient air concentrations
within any particular radius or distance from the incinerator.
Assuming unit emission (1 g/s), the five highest concentrations
predicted using the SLAMS data ranged from 0.89 to 1.3 /Ltg/m3.
Those predicted using the Watkins Avenue data ranged from 0.97 to
1.8 ng/ic?, and those predicted using the River Street data ranged
from 0.94 to 1.4 jug/m3.
All three data sets predicted the same receptor as having the
highest ground-level ambient air concentrations. This is" a
discrete receptor located 250 m north of the MWC. This discrete
receptor (661700/4829950) is the site predicted by the initial
modeling using Albany, New York "data as having the highest ground-
level ambient air concentrations. All three Rutland data sets
predicted the same five receptors as having the five highest
ground-level concentrations, except for the River Street data which
predicted a receptor located to the northwest rather than the
northeast as one of the five highest points. These results support
the initial modeling using the Albany, New York meteorologic data.
The ISCLT modeling results could not be directly compared to
the ISCST modeling results because both the loceitions of the
discrete receptors and the meteorologic information used in the
10-5
-------
modeling differed. In general, however, the maximum concentrations
predicted at the polar receptors by the ISCST using both the SLAMS
and River Street (Tables 4-5 and 4-6) are of the same magnitude as
the maximum predicted annual ground-level concentrations listed in
Table 10-1.
10.2.1. Pollutant-Specific Concentrations
The five-highest ground-level concentrations from the three
data sets were used 'to estimate .the concentrations of specific
pollutants. These predicted concentrations were converted to
pollutant-specific concentrations by multiplying the model-output
predicted concentration by the pollutant-specific emission rate.
The pollutant-specific emission rates were derived from stack
emission testing (see Section 4.2.1).
The five-highest predicted concentrations for the 3 data sets
for the pollutants for which an emission rate was available are
listed in Table 10-2. Beryllium was not detected during the stack
emission testing (Lodi, 1988), so the emission rate was assumed to
equal the detection limit. The range of the five-highest ground-
level ambient air concentrations for each ISCLT run are summarized
with the maximum emission rate measured during the stack testing
in Table 10-3.
10-6
-------
S
O".
03 W
H «j
S'H
O co
•**
§i!
o o
•H «
*>-j
? 0
US
7So
o o •"••
-<§v?
S°o
m-a %
•H S o
P -p «
O-P
•H «)
•o x
8 «
a, g
•P #•
CO t1
4*
55
?S
fl) O
S*1
fa
•0
1
«^
O
•H
4J
l"
II
1
ij
£
5
9
»H
iH
S
O 10
£t!
1 O
00 rdinate)
8-S8
8-8*
R
Meteorologic L<
Site (UH
o o o o o
«H rt t-) t-l cH
• x x x x x
CO O H CO fH
... co to
CM H O • •
H H H CO CO
O O O O O
X X X X X
H o «* r> >o
O M Ot <• PI
in i" to <*) n
o o o o o
X X X X X
r- M «H o\ o
•* co in o o
•* o n n n
o o o o o
X X X X X
<-4 if •* r-l t~
• in r- t- «*
rH . . . .
H en co f- t~-
rt M M M M
o o o o o
X X X X X
c>) t^
V V V V V
o o o o o
*
<-l 03 r-
Kl M ro M M
O O O O O
X X X X X
t to o t- en
vo PI co r- •*
PI PI M cj eg
o o o o o
X X X X X
to •«• CO t- •*
VO in P) (N cH
0 O O 0 0
X X X X X
co n o ro to
co CO P-
K1 ro ro ro Kl
O O O O O
X X X X X
•*r o PJ co oi
o CM en o r-
Ul •* PI PI PJ
0 0 O O O
X X X X X
o CM en H •*
PI O\ p- xf CM
o o o o o
X X X X X
r> •* vo vo r»
... pi PI
rH H H CO p~
V V V V V
0 O 0 0 O
rH rH H H H
X X X X X
pi in CM PI H
• •* co en tH
iH en co vo vo
o o o o in
in in o o oo
en o en CM co
en o en o en
CM CO M CO CM
o o o o vo
o o o o f>
Watkins
Avenue 661
661
661
661
661
V
a
•H
r-(
cj
t!
li
t)
•«
o
• is
5c
2°
V
K (II
o in
•H «J
10 J3
0)
•H 10
§td
jj
x
-------
TABLE 10-3
The Highest Modeled Ambient Air Concentrations
for the Three Rutland Sites
Pollutant
Emission Rate (g/s)
Air Concentration (jiig/m )
Arsenic
Beryllium*
Cadmium
Chromium
Lead
Mercury
Nickel
2,3,7,8-TCDD
Equivalents
6.30x10,
-6
7.60X10
-6
1.28x10"
2.80X10"3
7.95X10"4
3.19x10
-4
3.58X10
-3
9.16x10
-8
5.61X10"6 to 11.3X10"6
<6.76X10"6 to <13.7xlO"6
1.14X10"4 to 2.30X10"4
2.49X10"3 to 5.04X10"3
7.07X10"4 to 1.43X10"4
2.84X10"4 to 5.74X10"4
3.19X10'3 to 6.44X10"3
8.15X10"8 to 16.5X10"8
Beryllium was not detectable during stack emission testing, so the emission
rate was based on the detection limit.
10-8
-------
10.3. CONCLUSION
The modeling using site-specific Rutland data confirmed the
initial modeling efforts using long-term air dispersion modeling
to locate the meteorologic and ambient air monitoring sites.
However, there is uncertainty associated with the air dispersion
modeling ,as a result of the lack of long-term Rutland meteorologic
data as input into the ISCLT model, and the use of limited MWC
stack monitoring data. The air dispersion modeling was performed
using limited site-specific data; the modeling was performed using
< 1 year of Rutland wind speed and wind direction data. Ideally,
long-term modeling should incorporate 5 years of meteorologic data.
The stack emission data were also limited; only the maximum stack
emission rate of the 3 runs were used to estimate the maximum
annual average concentration. Variation in stack emissions may
have occurred as a result of varying operating conditions of the
incinerator, and these possible variations were not incorporated
into the modeling.
The modeling results, with the exception of PCDDs/PCDFs
indicate that the majority of the pollutant levels attributable to
the MWC may not be measurable using the current analytical
techniques. The predicted concentrations of some of the chemicals
modeled were orders of magnitude less than the analytical limit of
detection. Table 10-4 lists the maximum predicted ground-level
concentration and the detection limit for each chemical.
Consequently, the pollutant ambient air concentrations emitted by
10-9
-------
TABLE 10-4
Maximum Predicted Annual-Average Concentration
and Analytical Limit of Detection for Each Pollutant
Predicted Concentration
Pollutant (fig/m3)
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
Nickel
2,3,7,8-TCDD
Equivalents
1.13X10"5
1.37X10"5
2.30X10"4
5.04X10"3
1.43X10"3
5.74X10"4
6.44X10"3
1.65X10"7
Limit of
Detection (jug/m )
4.6X10"3
2.4xlO"4
1.4xlO"3
6.9X10"3
S.lxlO"3
ND
7.7xlO"3
6.4X10"9
ND: Not Determined
10-10
-------
the MWC generally could not have b.een measured. Since the minimum
limits of detection varied for each PCDD and PCDF isomer, the value
in the table is the lowest 2,3,7,8-TCDD equivalent concentration
estimated from the measured ambient air samples... Assuming this
estimate is reflective of what could be measured, the 2,3,7,8-TCDD
equivalent concentrations attributable to the MWC could have been
measurable.
10-11
-------
-------
11. ENVIRONMENTAL MEDIA RESULTS
Environmental media were sampled in areas surrounding the
Rutland MWC during the project in October and November 1987, and
June 1988. Water, sediment, soil and milk were sampled twice
before, and once after, the combustdr was operational, while
agricultural crops (carrot and potato) and forage (grass hay) were
sampled only before commencement of combustor operation. The
sampling and analytical procedures have been described in Section
2.2.4 and 2.2.5. All environmental samples were analyzed for
metals; soil, sediment, milk, produce and forage were analyzed for
PCBs and PCDD/PCDFs.
Samples collected in 1987 prior to operation of the Rutland
MWC represent background levels of pollutants in the environment
for comparison with those samples taken after the initiation of
incinerator operations. The primary objective of sampling during
both pre-operational and operational periods of the combustor is
to provide an indication of the incremental increase of pollutant
concentrations in these media, if any, caused by emissions from
the MWC. While several sites were sampled (e.g., for metals, five
sites were sampled for water and sediment, and seven sites were
sampled for soil) , each site was sampled only once during each
sampling round producing a limited number of samples. Thus, a
quantitative risk assessment, such as determination of human
exposure via the food chain (U.S. EPA, 1990), was precluded by the
small sample sizes. Therefore, a qualitative risk assessment was
performed in which samples of each pollutant in the same media
11-1
-------
(e.g., soil) were pooled across the various sites for each sampling
round and then compared statistically. For example, the mean
concentration for each metal for October 1987, November 1987 and
June 1988 was calculated for each media and the three sampling
rounds then compared. Additionally, the metal concentrations for
the sampling rounds prior to operation (i.e., background) have been
pooled and the mean compared with the mean from the sampling round
during operation of the Rutland MWC. Statistical analyses have
been discussed in Section 5.2.
Milk samples collected at Route 100 (Westfield, VT) have been
excluded from statistical analyses because samples were collected
only during one sampling period (November 1987). Similarly, soil
samples collected at Creek Road in June 1988 have been excluded
from statistical analysis of PCDD/PCDF and PCB concentrations since
no corresponding samples were collected in either October or
November 1987. Thus, no comparison of, pollutant concentrations
before and after incinerator operation could be made for these
sites-. Only background concentrations of pollutants for produce
and forage are presented since sampling only occurred during 1987.
Results for the carrot and, potato have been pooled to estimate
average produce concentration.
Concentrations that were reported by the analytical laboratory
as being non-detectable were conservatively assumed to equal the
reported detection limit (i.e., thus giving an upper limit estimate
of concentration) . Data are expressed for each chemical in the
same units as received from the analytical laboratory. Replicate
analyses of the same chemical in the same sample are averaged to
11-2
------- An error occurred while trying to OCR this image.
-------
D)
*x
D)
O.
co co S
o> o> °>
T- T- T^
CD
o o §
O z -^
o
c
o
O
^
^^™
i i
j- O
-------
o
E
o
01
*•»
(0
+*
CO
w"
-------
§8§
c
O
O
QC
QC
geo
JC
O
co
a
u
3
O
CO
T~
O
CO
—r~
co
T~
IO
CO ' CM T-
11-6
-------
Q)
•I-'
CO
c»
c:
v.
c>
c:
O
(
•f- >-
3 I
I w -
r- CT3
,- o c
c
o
£
I
••—
o
0)
•*-»
-------
c
CD
••-•
CO
(U
E
•4-»
k.
rt
a.
0)
o
3
O
CO
11-8
-------
c
o
^
3
O
co
a)
QC
i»
0)
4-*
I
0)
E
Q.
0)
Q
0)
O
3
O
CO
11-9
-------
Cl)
o>
c
c
o
c
o
E
._ ^_,
•o c
O o
h.»>
i.l?
to —
OC-ts
3
£<
O
^
O
c
o
O
H-
O
0)
•*-•
03
•*-*
CO
w"
a
o
co
-------
1
0)
O)
c
c
o
c
•t-*
a
CO
a
o
3
O
CO
11-11
-------
Ic
l-i
CO
O)
£=
o
^- WO
^ c 5
o
-------
o
E
•- I
OJ
.& cc
D O "•
<5 §.E
u- O w
o ®
TO
+••
CO
w
0)
o
fc_
3
O
m
0)
OC
k.
0)
4-1
i
c
0)
E
o.
0)
Q
0
o
3
O
CO
CM i-
11-13
-------
CO
c
o
LLJ
CC
2
UL
c
o c
O ~
.= Q.
CO
o
E
o
0)
-l-»
ro
•t-t
CO
05"
eg
o
3
O
05
eg
CC
k.
o
•*-*
I.
c
0)
E
Q.
OJ
Q
0)
o
3
O
CO
11-14
-------
cv,
O 4_>
co c,
-El
» £
§>
LJJ C *-
QC
-------
.si
"I
CO O ^
T- c
^
E 0)
.2 a
E E
O Ki
-CO
O
cd
•*-•
CO
w
o
3
O
CO
11-16
-------
o
£
o
0
**
(0
*•
CO
CO
o
o
3
O
(0
0)
DC
t»
<0
Q)
E
•»*
i_
RJ
Q.
4)
2
'3
O
CO
11-17
-------
C
o
£
I
•*-
O
0)
•*-*
OS
**
CO
to
d)
o
O
co
-------
V O -a
*- +: c
»- «J (0
uu i: ^
a
E
c
o
E
W
0)
M—
O
-------
TABLE 11-1
Metal Concentrations in Milk, Produce and Forage
October and November 1987 and June 1988
Metal
As
Be
Cd
Cr
Pb
Hg
Ni
Sample*
Date
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
Milk
fua/Ll
X ± SD
ND
ND
ND
125°
1.0C
1.0C
1.0C
1.0C
NQ
NQ
NQ
5C
9 . 0+5 . 2
4 . 0±3 . 5
6 . 5+4 . 8
5°
118±25 '
43.0+37.4
80.3+49.7
25C
ND
ND
ND
0.2-1.0°
NQ
NQ
NQ
50°
High
Value
ND
ND
ND
125
1.0
1.0
1.0
1.0
NQ
NQ
NQ
5
15.0
8.0
15.0
5
145
111
145
25
ND
ND
ND
1.0
NQ
NQ
NQ
50
Produce*3
(mcr/kcf)
X ± SD
0.5C
0.5C
0.5C
NS
0.30-0.10°
0.03°
0.30-0.10°
NS
0.2
0.3
0.2+0.1
NS
1.0C
1.0C
1.0C
NS
2.5C
2.5C
2.5C
NS
0.05C
•0.05C
0.05C
NS
2.5C
2.5C
2.5C
NS
High
Value
0.5
0.5
0.5
NS
0.1
0.03
0.1
NS ,
0.3
0.3
0.3
NS
1.0
1.0
1.0
NS
2.5
2.5
2.5
NS
0.05
0.05
0.05
NS
2.5
2.5
2.5
NS
Forageb
(ma/kcr}
X + SD
0.5C
0.5°
0.5C
NS
0.03C
0.03C
0.03C
NS
0.1°
0.1
0.1
NS
1.0C
1.0C
1.0C
NS
2.5C
2.5C
2.5C
NS
0.05C
0.05C
0.05C
NS
2.5C
2.5C
2.5C
NS
High
Value
0.5
0.5
0.5
NS
0.03
0.03
0.03
NS
0.1
0.1
0.1
NS
1.0
1.0
i.o
NS
2.5
2.5
2.5
NS
0.05
0.05
0.05
NS
2.5
2.5
2.5
NS
For October 1987, Milk n=3; Produce n=2; Forage n=2;
For November 1987, Milk n=3; Produce n=l; Forage n=2;
For June 1988, Milk n=3
S.D. not calculated for n<3
c No value exceeded analytical detection limits
ND = Concentration not determined due to analytical problems, e.g.,
interference
NQ » Determined present but not quantified
NS « Not sampled
11-20
-------
Arsenic values found in produce and forage in this study were
>',
non-detectable. The lower detection limit was greater than the
value reported by Johnson and Manske (1976) for potatoes (<0.1
M'g/g) but within the range reported by Pyles and Woolson (1982) for
potato flesh (0.02-2.4 ppm) . Chromium concentrations measured in
this study are below the detection limit (<1.0 mg/kg). This
detection limit is greater than that reported for chromium
concentrations in potato (0.15 mg/kg) by U.S. EPA (1978b) . Gerdes
et al. (1974) reported mercury concentrations of 1-123 jttg/kg in
vegetable samples from Texas. Concentrations of mercxiry in produce
and forage in this study were below the detection limit (0.05
mg/kg) . Data for background concentrations of the other metals
(beryllium, lead, nickel) in produce and forage in other
geographical areas were not immediately available in the
literature.
11.1.2. Milk. Mean concentrations of the metals in milk are
reported in Table 11-1 and in Figures 11-1 and 11-2. The milk was
collected from bulk storage tanks at the sampling sites.
Arsenic, cadmium, mercury and nickel were not determined in
milk due to analytical problems (e.g., interference) during the
October and November sampling rounds, and were : not found at
concentrations exceeding the detection limit in June 1988.
Concentrations of beryllium in milk did not exceed the
detection limit of 1.0 Mg/L for all sampling periods and sites,
including Route 100 (Westf ield, VT) . .
11-21
-------
Chromium and lead concentrations were found in milk in
measurable quantities at several sites in October and November
1987, but were below the detection limit in June 1988. The
detection limit for these metals increased between the 1987 and
1988 sampling rounds. There was, however, no statistically
significant difference in chromium concentrations between the three
sampling periods when analyzed by a one-way ANOVA or by the
Kruskal-Wallis test. Samples collected prior to MWC operations
were pooled and compared with those collected during operation by
a two-sample pooled t-test. The average chromium concentrations
between the pooled pre-operation period and the operational period
are similar, but could not be analyzed by t-test since the variance
of the operational period was zero (i.e., all values are the same) .
Lead concentrations showed a statistically significant difference
(ANOVA, p=0.010) between the three sampling periods, with the
samples collected in October 1987 being greater than the other
sampling periods (Scheffe test, p<0.05). However, since all
concentrations of lead during the operational period (June 1988)
were non-detectable and were set equal to the detection limit (the
variance was zero), and because the mean concentrations in October
and November 1987 were statistically significantly different, a
pooled t-test could not be conducted.
The lead concentration measured in milk from Route 100 is in
the range of the lead concentrations of the milk samples collected
in Rutland during November 1987 and June 1988. Assuming the water
content of milk is 87% (Baes et al., ,1984), the concentrations in
fresh milk collected from bulk storage tanks in Rutland in June
11-22
-------
1988 (<0.19 jug/g) is within the range of that reported for fresh
milk by Murthy etal. (1967) (see Table 11-2) . The average of the
lead concentrations (again corrected for water content) of the
samples collected before the incinerator was operational in October
and November 1987 (0.91 and 0.33 Atg/g, respectively) and the sample
collected from Route 100 for background comparison (0.25 Mg/g) /
however , are greater than the concentrations found by Murthy et al .
(1967) . It -appears, then, that the lead concentrations measured
in milk in Rutland are most likely representative of background
variability of lead concentrations for this area. This conclusion
is further supported by the fact that the highest lead
concentrations in milk were found before the incinerator was
operational, and that there are no significant increases in ambient
air (see Chapter 9) , soil or forage lead concentrations. It would
be expected that the air, soil and food chain would have increased
lead levels that would coincide with, or precede, contamination in
cows milk. Data for background concentrations of chromium in milk
in other geographical areas were not immediately available in the
literature.
11.1.3. Water, Sediment and Soil. Average water, sediment and
soil concentrations of metals are presented in Table 11-3 and in
Figures 11-3 through 11-16. Water concentrations of arsenic,
beryllium, and nickel were below their respective detection limits
at all sites for all three sampling periods. Cadmium and mercury
concentrations each were detectable at one site during one sampling
period, but the measured concentration was equal to the detection
11-23
-------
^
s:
c
•**
W
7 3n
3 of
03 c -S
£ .2
.Jj
(0
. 1
c
0)
o
o
CJ
(U
o
c
Q)
0)
U-4
0)
•H
10
0)
J-l
Cu
.— 4
*s
4J
0)
(0
c
0
—I
4J
tin ivo
o o o " o o o
vo- o- tr- r-i- I-H' . vo-
r^O no c-io ojo ^i-O rHO
O — • O ••-' . O — O — • O-— O' — •
o o 0,0 o o
pr" p~* ' t t r i
O 2 - ' K ?>
_r r • - - s
—t i-i ff Q) C '
•!-> 0) S , o O
« -P C 4-> T3
C UJ ~ (U tj> C
C Q) C -0 C (0
•••* J^ O -H ' -H »-(
0 U *J > ^ , . 4J
C C U) O M U
*t-t fft Q ^-[ *J Q
O S CQ (X CQ . . (V
•o
(0
(U
-
• • <•
U)
3
O
o
VO
M
°tJ'
' § 1
a) . ?j
CP T3 tj
-H 0) ^
0) 4-> •Z
3 co. »
>• (0 -H „
>-i
-------
TABLE 11-3
Metal Concentrations in Water, Sediment and Soil
October and November 1987 and June 1988
Met a]
As
Be
Cd
Cr
Pb
Hg
Ni
Sample8
L Date X
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
10/87
11/87
10-11/87
06/88
Water
(u.a/'D
± SD High
Value
5b
s
b
1.0b
1.0b
lb
1
ib
2.8+1.1
2b
2.4+0.8
2b ~
9.4+2.6
7.6±5.8
8.5±4.4
0.2+ 0
0.2^
0.2
0.2b
5b '
5b
5b
5b
5
5
5
5
1.0
1.0
1.0
1.0
1
1.
1
1
4
2
2
2
13
18
18
5
0.2
0.2
0.2
0.2
5
5
5
5
Sediment
(mq/kq}
X ± SD
3.3 ±1.7
2.9 ±1.6
3.1 +1.6
2.3 ±1.2
0. 12±0.08
0.20±0.07
0.16±0.08
0.12±0.04
0.3b
0.74+0.54
0.52+0.43
0.5b
3.1 ±2.0
4.3 ±2.1
3 . 7 ±2 . 0
3.6 ±1.6
10.5+2.1
13.8 ±6.6
12.2 ±4.9
10.8 ±4.0
0.10b
0.10b
0.10b
0.02b
4.4 ±2.3
3.6 ±1. '5
4.0 ±1.9
5.7 ±2.0
High
Value
4.4
5.0
5.0
3.5
0.2
0.3
0.3
0.2
0.3
1.7
1.7
0.5
0.3
5.8
5.8
4.7
13.8
25.1
25.1
15.2
0.10
0.10
o.io
0.02
7.3
4.5
7.3
7.7
Soil
fmq/kcrt
X ± SD
5.9 ±1.5
4.0 ±1.9
5.0 ±1.9
4.4 ±1.2
0,, 16+0 .=07
0,,17±0.05
0,,17±0.06
0,2 ±0
0.56±0.67
0.56±0.11
0.56+0.48
0.8 ±0.67
14.8 ±28.28
7.7 ±6.1
11.4 ±20.6
16.0 ±27.1
57.5 ±72.9
44.2 +48.0
51.3 ±60.8
79.3 ±93.9
0.18+0.22
0 . 10r
0.14±0.16
0 . 11±0 . 19
23.5 ±48.4
9.4 +10.2
16.9 ±35.6
19.4 ±30.0
High
Value
7.8
7.8
7.8
5.9
0.3
0.2
0.3
. 2
2.2
0.8
2.2
2.3
4.4
21.5
84.4
77.4
216.0
143.0
216.0
246.0
0.71
0.10
0.71
0.53
143.0
32.4
143.0
87.4
a For October 1987, Water n=5; Sediment n=5; Soil n=7;
For November 1987, Water n=5; Sediment n=5; Soil n=7;
For June 1988, Water n=5; Sediment n=5; Soil n=7 for each,metal
b No value exceeded analytical detection limits
11-25
-------
limit. Therefore, since the average concentrations for thesemetals
were equal for the three sampling periods, no statistical analyses
could be performed. The concentrations of arsenic, beryllium,
*i
cadmium and nickel, at or equal to the detection limit, are less
than, or within the range of, the respective metal concentrations
found in other surface waters as presented in Table 11-4.
Concentrations of mercury in surface waters were not readily
available in the literature.
Chromium and lead water concentrations exceeded the detection
limit (Figures 11-3 and 11-4) in several samples collected in the
pre-operational sampling periods (October and November 1987), and
these data were therefore statistically analyzed. A one-way ANOVA
of chromium or lead concentrations over the sampling periods showed
no statistically significant difference in mean concentrations,
When the pre-operational sampling intervals were pooled, a two-
sample pooled t-test could not be conducted since all values of
chromium or lead were the same (below the detection limit) for the
June 1988 collection (variance was 0) . The non-parametric analysis
of variance (Kruskal-Wallis test) showed a statistically
significant (p=0.02) difference in the mean lead concentration for
the different sampling periods. This difference was due to the
large difference between the non-detectable concentrations observed
in June 1988 and the relatively high concentrations observed in
October and November 1987. As discussed in the methodology
section, the fact that the parametric and non-parametric analyses
did not give the same results suggests that the assumptions made
11-26
-------
1
H
rH
"
CO
rtj
EH
^
0)
-P
(0
c
•H
CO
i-H
(0
4J
0)
4H
0
W
o
-rH ^
-P J
0)
T!
C
3
O
&•
O
CQ
Reference
.e
0)
ij
CO
0)
-p
o
g
Q)
a
eshwater
JH
fc
M
0)
-P
(0
0)
0
O rH fT . 2
co r- co r- cn
cn cn cn cn f\j ,H
• rH rH rH
O . . o
CO rH r-H ffi
cn to (0 ^
rH
-P -PC
0) Q) -H
Ul
CM 0) 73
•HO) 43 o
.* -P (0 4J
CX O-H
0) (0 O O
H) CQ S v_
ct; cj 05
2 § 55
OS QJ Qi
22 g
vo
n --^
Cd ^r pd
2 rvo^ 2
in'
0 0
0 0
rH 0
i « o"
2 rH
0 V
rH
O
-H . '
C
0) "
CQ
£_i
rH
(0
CO
cn *£
rH VO 55
VO O
cn t- C
^ rH Cn '»H
Ok rH
W -o
C - 0)
CO S O-H
0 O O
ID cQ >*£ x— '
Oi OH OH
2 22
CD
rH
O
0
v 02
V
/
CM
\M o]
* """•
.« K rH
2 21*
o 2.
•
O :
0)
01 r-~
• co • rH
cn o -^
1 rH PS rH
2 1 •
rH O O) O
O O
• ' •
0 O
e
3
•H
r-H
i-H
^,
0)
< . CQ
CO.
cn ..
rH "O •
C 0'
- (0 C
M
^ C
(0 r^ o (0 -
-P cn cn >H o
Q) rH H < 4J
C ^ '"•«. N -H
a to g o g
>t < 0) 0 (0
OH B5 W CQ K
« OJ « OH
2 222
J
rH
0 V
« *0 r^
2 rH J^rH rH
r-H (0
-------
*
ij
o
u
^-*
T
f-H
H
a
CO
i5
0)
o
Q)
rH
W W rH
• " *• ^1
W W g V4
• Q) (0
D D K U
pr* p£ Q^ fV^
2 2 22
«,.
K pi PJ 1 H"
2 2 2 —
•
O
to
OJo OJaj
rH -— * rH -— ^
rH t*^ rH C"*
• • cd P^
i a\ . i o\ 2; 2:
O rH
o
O^ ^-^ C0«j
1 S" -«^
| (Sj t~t • p^
^- «r — o 2
0
o
e
3
•H
O
5-1
°
- , - . rn
vo o
CO CO O)
^ O% rij
' . i-H rH 2
O
o • , ^ - r- C
f^ rft rf* O\ 'H
rH .'• W - W TJ
': - Q)
«. » * Oi -P
g W W O -H
Q) . . O U
X D D « —
in
o
o
i p£ pj, ps
2 2 2
vo
o
o
»
o
Pi PH PH PH
'•
O
0 ^* O
in ^ I-H --^
in cr> n
Pi • Pi 1 O-4
2 1 r^ 2 CM ^
H —
Q
rH •
f- o O CO
• Pi °,° 1 **>
co 2 in o co •— *
rH
8
r0
(0
0)
^1
O 0
co en
en H
H
M"
ft, 1
JH
O
£_l
CD
S .
11-28
-------
0)
o
c
0)
0)
Cft to"
f~l f™» |rf
o 02
o • r- in ^ r-
rtj r*» o\ c— ^ CTI C
di &i ' t-H ••->
W rH ^1 W 73
• - U - • ft ij '
10 g K 10 CO O -H
• d) rtj rtj * Of)
C3 ffi n 2 D « ^
« O5 tf OS . « •'#
2 2 2222
i-T .
A OS DH OS OS OS
•— 2 2222
in • o
o
CNJ VO" -O
rH A f^> 03
OS • - rH ^-~
OS OS 121 CK
22 1 •-•
O IT) rH ^
V
CMC- V •* 'T •* rH^--
— — > — •
i-H
0)
X
O
•H
2
'
T!
(U
0) •£*
-H °
-*-1 m
(— • Q/
Q) ^
.S-g
3 a c z
e « „
C -H J3 -rH "
(0 X -P T3 Q,
0) (0 O> Q) g
11-29
-------
for the parametric ANOVA (i.e., equal variances, normally
distributed data) were not met. In fact, the variance for the
June, 1988 sampling period was zero.
Surface water concentrations of chromium found in Rutland are
at the lower end of the range of chromium concentrations (0-112
Mg/L) reported by U.S. EPA (1978b; 1980a). Similarly, lead
concentrations found in this study are within the ranges for other
surface waters (3-1000 jug/L; Koop, 1970; U.S. EPA, 1986d), but
greater than those found in remote streams (mean concentration 3.7
/zg/L; Hem, 1970) . ,
The majority of metals, with the exception of cadmium and
mercury, were found to be present in sediment in concentrations
above the detection limit (Figures 11-5 through 11-9). Only one
sample each of cadmium and mercury were detectable. Except for
these two metals, statistical analyses did not show any
significant differences in mean concentrations of any metals when
compared across sampling periods, nor when the pre-operation period
(October and November 1987) was compared with the operational
period (June 1988). Mercury concentrations in sediment were
statistically significantly .lower in June 1988 than both of the
1987 sampling periods (Kruskal-Wallis test, p = 0.00091). This,
however, is attributable to the lower detection limit for the 1988
analysis. Similarly, cadmium concentrations in sediment were
statistically significantly lower in October 1987 than in November
1987 or June 1988 (Kruskal-Wallis test, p = 0.0018) due to the
11-30
-------
lower detection limit during that sampling period. The November
1987 and June 1988 sediment cadmium concentrations were not
statistically significantly different. Mercury, and lead
concentrations in sediments were not readily available in the
literature. Arsenic sediment, concentrations found in this study
are in the range (<10 M9/9) of those reported by Cerelius (1974).
The concentration of the majority of metals in soil exceeded
the detection limit. Mean soil concentrations of metals are
reported in Table 11-3 and Figures 11-11 through 11-16. Soil
concentration of metals, particularly chromium, lead and nickel,
appeared to be much higher at the MWC/Rte.4 sampling site than the
other sampling sites. However, this pattern (MWC/Rte.4
consistently the highest metal concentrations) was observed at all
three sampling periods. This resulted in a statistical design that
was balanced,'and, thus, parametric statistical analyses showed no
difference in the means between sampling periods for any of the
metals. Non-parametric analyses (Kruskal-Wallis test) showed
statistically significant differences between sampling periods for
'' '. V . ' I -
cadmium and mercury soil concentrations. This was attributable to
differences in the detection limits of the analytical methods at
the different sampling periods and also to the large number of tied
ranks in these rank-transformed analyses. The soil metal
concentrations in Rutland were generally within the lower range of
values reported for background and/or farm soil concentrations.
Table 11-5 lists concentrations of metals in soil.
11-31
-------
rH
*0
CO
C
•H
UJ
rH
? S
-L <»
f" f *g ^^*
f»^ *^ ^yi
^_
ttj ^»
M W Cn
TJ O «*
M a.
2 «~
s .s
4J
10
H
C
Q)
O
C
o
0
Reference
rl 10
03 (0
J3 CO
•P rl
O "C
c
1
(0
Cn
•a
c
o
tn
JM*
o
(0
rH 0)
•H D.
O >i
CO EH
rH
(0
-P
Q)
E
•*-->..
•rl VO CO rH '
C r^ 5 co
O O^ ' *• O^
•rl rH . rH
f^
co •% w - ^ -r _-
•* vo 5 P- aj r> 2
t~ a> J en f-n ai S
CftrH -O H nJ-PrH^
rH C Q> ~C
(0 ^ 4J e W -H
^ & 3 & • •§ r5«
OS 0 W XSrHQ)
WnJ C^^ CSOl-P
4JrH &CO C <0 tQ rl-H
(0 rH -H ff> -H >H -H Q) O
'
_Q
^ O
ryj fV^ VO P^ p4 Pi rH
22 ^ 222V
B B B B B • fi S
PJ Pi K Pi K PS PS
22 2 2 22 2
O
• o
•* O O ^-»
rH ^" *o . in {3 t^ VO
^-«. ^-v O3
I I vo K 1 in • 1 c£
*-* 2 v_ vo 2
in rH rH in
• • ^-^
o o
Ul
0)
&
22 2 2 ' *J 2 2
i-H
rH
u
•rl
C
n)
CO
r** o^
^ r**
rH G\
c-H
w id
Ul
co "c
-rl
Pi Pi
2 2
Pi Pi
2 2
2 2
F
0
' t"^ ID ^* (O
*-*• * — *.
1 rH 1 \0
•* — f - — '
rH rH
•
O
*
OH PH
2 2
e
3
i~H
rH
J_l
0)
a> CQ
w
«£
11-32
-------
-p
c
0
o
in
I
rH
[ 1
w
(_3
03
<*£
EH
—
(0
O
C
0)
£_J
0)
*W
0)
p*
0) 01
X3 ,
W EH
rH
It!
-P
Q)
S
CM co
co r*j CTI
CTl CO rH
rH CTl • .
rH CO *•
CO « CTl
• ^ CTl rH t^
i-H • rH (0 . CTl
(tf r— 1 rH
10 ^ 4J
-p - c a) ; -
Q) -P Q) >i
0) M 1 (0 rH 0) , • C .
0) >1 O -H ' ' -rH
CQ PH S Pi »_3
P^ On PH PH Pi
22 2 2 2
p^ OS p^ pS PS
22 2 2 2
• CD
CM - —
PH f¥j O^ CNJ P^
22 2 1-2
o
rH
O
V
CO. ,
CO ^^
rH in co r-~co
• CM • ^^o • • — *
o M o r- o o vo
OJ CNJ O
1 1 1 . . OS 1 •
002 0
CTl rH VO >-- rH ^
o o o o
• • • •
000 0
OS OS OS OS OS
22 2 2 2
B
,-rH
B
•o
(C
^*
CO O\ CO
CTl t~ VO
rH CTl CTl
rH rH CM
•oo
•H >i >i rH
y) (Tj rt
xt w >
X! T3 (0 >.
W C i-H J-|
•H -H r-H «J
OS PS PS OS
2 • 2 ' 2 2
OS OS PS OS
2222
OS OS OS OS
2 2 22
000
O co O to O O
in ^ o --^ Oo in
tM O rH O fO O rH
o o o
1 rH 1 rH 1 rH 1
N
0) rH m in
o
CO
OS OS OS OS
2 22 2
S
3
•rH
B
0
S-4
x:
0 .0
XI O
co n .. •
r- co
CTl ffi
rH i-H - • .
•* •» ^< *
< < t^
04 fW CTl
W ' W rH
• • •*
CO' CO CO '
'
C<£ /yf r^
222
OS OS OS
222
OS OS OS
222
o o
o o o
O O On
n rH n -^
o
1 1 1 <*
in rH in
V
OS PS OS
222
11-33
-------
~
c
o
u
in
i
i-H
rH
w
iJ
SO
Q)
U
Referen
M U)
0) BJ
JS 0)
1 1 j_j
o •<
c
1
g
TJ
C
o
>H
X
o
Q
CO
•H a
O >i
W E-<
^
flj
4J
(U
s
CM CO VO
co co tn vo r-»
en aj co t-» en
«-i en en rH
rH H
*" JH **1"
co • 3 *. en co,-"*
^0 i-H rH • f^. rft 4. •
cn > C 0) >i -H 0)
i .- c w ft A; •
* g 3 g TJ x; c
iH E tji O C 0) (d
•-H ,
z z z z • z z -P-
r-H
rH
T3 ,
(0
Q)
rH
«• (0
O en
rH -rH O ^ r-
* (d ^> r^ t^* en
O en en I-H
4_) rH rH rH
0) O
0) - >i
C O >i x: id
tn TS
T5 W M 4-> C
5-1 -3 (d ~ -rH
' W D W CX J
CX CX IX CX
2 ,Z Z Z
ex « c*; ex
z z z z
*
ex cx ex cx
z z z z
0
>£>
rH in rira
o '— o -— o n
rH rH O
1 • 1 • 1
O O O
in ^ rH ^-- rH
•* o . o
o
0 0
0
cx ex tx cx
Z Z Z Z
D
O
J-i
0)
*~* ' ^*
VD
-^^
en
rH
CO
en
rH rH
(d
Fishbein,
Frank et
cx ex
z z
ex cx
z z
cx tx
z z
^ ^
in co
o o
o o
U)
0)
a
cx >.
Z -P
rH
r-H
11-34
-------
Q)
U
C
(U
Q)
Q)
a
S-i (fl
0) (0
.C Q)
-M J-J
o <
-p
c
o
U
m
i
rH
rH
W
CQ
C
(0
i!
J-l
(0
"O
c
a
o
U
It)
CQ
rH 0)
•H a
O >!
co E-i
(0
-p
0)
s
CO
CO
CTi
rH
CO
VD
CTl
(0
s
CO
CO
c
0)
n
tp
rH
o
X
cr>
t^-
CTi
to
M
CTl
in
o
t)
to
r-l
CP
o
c
in
t-~
Ol
CO
•n
VO
CO
w.
CO
a
•z
a
a
•z,
a
OS
a
o
o
0,0
o
C0
-------
11.1.4. Conclusion. Overall, these results indicate that there
were no apparent increases in metal concentrations in the
environmental media during the period the Rutland MWC was
operational relative to the period prior to combustor operation..
However, because many metal concentrations were non-detectable and
assumed equal to the limit of detection and because method
detection limits often changed between sampling periods, this
conclusion contains some uncertainty. It is still possible that,
had lower concentrations of these metals been quantifiable,
differences between sampling periods (operational ,vs. non-
operational) might have been observed.
11.2. PCB
The analytical results 'for PCB were reported as congener-
specific concentrations for both the field ' samples and method
blanks. As discussed in Chapter 2, congener concentrations for
each sample were analyzed by HRGC-HRMS, corrected by the respective
detected method blank and then summed to ,estimate the total PCB
concentration present in each sample. Total PCB concentrations in
the environmental media are reported in Table 11-6 and Figures 11-
17 through 11-19.
11.2.1. .Produce and Forage. The concentrations of PCB in the
produce and forage range from 1.86xl03 (carrot) to 6.18xl03 (potato)
pg/g. The produce collected during October 1987 had an average PCB
11-36
-------
u>
1
rH
rH
TABLE
(0
T3
Q)
s
rH
(0
-p
c
Q)
E
C
O
.^J
C '--
W Cr>
C CP
--i a
(C Q
-P CO
•£+•'••
X
c
o
-rH
•P
(0
-p
c
CO
u
c
o
u
.ca
o
cu
•a
o
-^
0)
a*
c
-rH
rH
a
e
(O
CO
CO
CO
av
rH
CO
c
3
I—
co
cr\
r-H
JH
0)
n
Novem
ro
co
CT>
SH
Q)
.Q
O
-P
U
0
'(0
TS
0)
"o "a " "O-
i i _j
<~H (^ -^"
XXX
vo C-H v "CTJ;
O r. .
+1 4-1 4-1
'b "o :So" ' :
rH rH ' rH
XXX
n r- vD.
r~- CN) th
co co
2 Z co co •*
o -o o . •
rH rH rH
XXX
rH ••••}• VD '
• ^j* CT\. • ' . r*" . .. .,
O rH VD
-H 4-1 +1 '
fi ro fNJ ^ -*y
o o o o o
rH rH rH rH rH
X X X X X
n CNJ CM H in
in co rH co
0 O 0 ...
' ' " H rH . rH
XXX
rH {^ VD
. VO O; O
o ' '•••*' fo
4-1 4-1 4-1
ro K> fN) t^» u-»
O O O O O
rH rH rH rH rH .. •
X X X X - X
CN **O CT^ "^ O^
O CN CO P- C~-J
^r in (N r- rH
-, *-. ......
-p
(U -D C
o. • a> co ...
D • Cr* E
TU (B ^"4 '"-H rH
O J-l r-< T3 -H
S-I O -r-l Q) O
. . ~ . ...
in
in, II
11 '• C •
c
-p
• . -.4-) C
•CO)
Q) E
S.-i
* TJ Q)
0) CO
w
• «. vD
vo 11
-,-.|.«v :.,
rH
•• '• -H O - -
o co"
co
n II
II C
'X rH in
r-H -H ||
-^;sc
••• 4-J
•- 00 C
" ' ' (N II Q)
' II C E
C "H
Q) T3
a) i CT> a)
CT1 <0 CO
(0 S-i
li Q *^
O fT I VD'
CM' II
•••- C
•- r— I
. (N II VH
II C -H
C O co
0) O C
O 3 •-
,3 T3 co M
'-'•'. T3 O II 0
O >H C t-i
rH Ot
&< Ai TU
-rH 0)
^r-. -H 4->
r~ co S 10
CO O*i ' — '
CT> rH -3
rH CO U
r-l CO r-H
JH- CO O> (0
0) £t rH U
Q ^
O CO CO -P
4J >' C 0'
U O 3 C
0 Z ID
r-l r-l r-l Q
O O O •
Oid-SCOCO
-------
c
o
E
I
»*-
O
0)
CO
m
Q)
o
3
O
09
0)
QC
k.
0)
•*•*
i
0)
E
•»-»
L.
n
a
0)
a
!o
o
3
O
CO
11-38
-------
CO
c
CD
0) o
COE
•*- CO
CO
o>
c
c
o
111
^ ««
LU iZ^-
O o c
u- o co
O o>
Q.
CD
o
03
•CO
o
E
O
0)
•+-'
CO
^-»
CO
w"
V
o
3
O
(0
(U
IT
k.
0)
-*-•
-------
o
CO
.E
O) tfi
Ill l_ *-
°- c
-------
concentration of 4.02xl03 pg/g and the potato collected in November
1987 had a concentration of 2.53xl03 pg/g. .Replicate analyses of
the same potato were averaged to determine the value for that
potato (i.e., the duplicate analyses of the potato collected at
Quarterline on 10/09/87 were averaged to determine the value for
the potato at that collection period). The average PCB
concentrations in forage were 5.26xl03 and 3.82xl03 pg/g during
October and November 1987, respectively.
The PCB concentrations in produce in Rutland are similar to
those concentrations found elsewhere. Carey et al. (1979) did not
1 '
detect any PCBs in crop samples collected from 1483 sites in 37
states.
11.2.2. Milk, Sediment and Soil. The results of the analyses of
milk, sediment and soil samples do not indicate that PCB
concentrations in these environmental media increased due to
deposition of PCBs from the stack emissions, but indicate the
concentrations are similar to those found elsewhere.
A one-way ANOVA was performed to compare the total
concentrations of PCB in milk for each sampling round (i.e.,
October 1987, November 1987 and June 1988). The average PCB
concentration in milk for the samples collected after the
commencement of MWC operations (8.73xl01 pg/g) was statistically
significantly less than the average concentrations in samples
collected in October (2.39xl02 pg/g), but not significantly
11-41
-------
different from that for November 1987 (1.12x10* ,pg/g). The
Kruskal-Wallis nonparametric ANOVA showed no, statistically
significant differences between mean milk PCB concentrations ,f9r
any of these sampling periods. , ,.
Since the October and November 1987 milk PCB,concentrations
were statistically significantly different, they could not be
pooled for comparison of pre-operational and operational
concentrations. Yet, it can be concluded that operation of the iprc
is not the likely source of the milk PCB concentrations since June
1988 levels were below both pre-operational sampling period
concentrations. Due to the small number of milk samples .analyzed,
however, this conclusion contains a degree of uncertainty that
cannot be estimated precisely.
A milk sample collected at Route 100 (Westfield, VT.) during
November 1987 was used for background comparison. This sample had
a PCB concentration of 1.32xl02 pg/g. The concentration of this
single background sample is similar to the concentration range of
the samples collected during November 1987 and June 1988,but is
less than the concentrations in samples collected during October
1987. No statistical tests were performed to compare the Rutland
concentrations to that of Westfield since only one sample was
collected in Westfield. .
The average PCB soil concentration for June 1988 was 4.56x10
pg/g. While this value is less than the average concentration in
'October 1987 samples (1.29xl05 pg/g), and slightly greater than the
11-42
-------
average concentration in November 1987 (3.25xl04 pg/g), the means
for these sampling periods are not statistically significantly
different.
The average PCB concentrations detected in Rutland soil
samples are within the PCB concentration ranges found in other
areas. For example, Carey et al. (1979)^sampled soils from five
U.S. urban areas (43-156 samples per site) in 1971; concentrations
were detected in three areas with PCB levels ranging"from 2.0xl04
to 1.19xl07 pg/g. Greaser and Ferriandes (1986) analyzed 99 soil
samples to estimate background concentrations in British soils.
PCBs were identified in all samples within the range of 2.3xl03 to
4.44xl05 pg/g.
The average PCB sediment concentration of the samples
collected during June 1988 (8.27xl03 pg/g) is similar to the
average concentration of the October 1987 samples (7. 74xl03 pg/g) ,
but is approximately one-half the average concentration of the
November 1987 samples (l.SlxlO4 pg/g). The average concentration
in the November samples is high due to the high concentration
measured at Rocky Pond (5.08xl04 pg/g). The mean concentrations of
t
the samples collected during these three periods, however, are not
statistically different.
The PCB levels found in the sediment in Rutland are less than
those found elsewhere in the United States. PCB levels of 9.8xl04
to 5.4xl05 pg/g have been detected in the sediments from four
remote high-altitude lakes in the Rocky Mountain National Park
11-43
-------
(Heit et al., 1984). Sediment from the Milwaukee harbor has been
found to contain PCB le
(Christensen and Lo, 1986).
found to contain PCB levels of l.OSxlO6 to 1.34xl07 pg/g
11.2.3. Conclusion. The effect of incinerator emissions on total
PCB concentrations in forage and produce could not be determined,
since these media were only sampled prior to MWC operations. No
difference in total PCB concentrations was found in milk, sediment
or soil sampled both before and during incinerator operations.
11.3. PCDD/PCDP
The analytical results for the PCDD/PCDFs in environmental
media were reported as follows. Concentrations were blank-
corrected and converted to 2,3,7,8-TCDD equivalent concentrations
as explained in Section 3 and presented in Table 11-7 and in
Figures 11-20 through 11-22.. Means presented refer to 2,3,7,8-
TCDD equivalent concentrations. Since only the octachlorodibenzo-
p-dioxin (OCDD) congener was. consistently detected in- the
environmental media, mean concentrations of this congener (as
reported by the analytical laboratory) were also compared for the
various sampling periods.' These data are presented in Table 11-
8.
11-44
-------
f^
i
rH
rH
1
"
(0
•o
0)
s
"iti
-p
c
0)
g
o
5-1
-rH
C
w
c
-H
to c:
J-i O*
, o*
Concentratj
[X ± SD) (p
1^
c
0)
i-H
m
-H
3
w
Q
Q
CJ
EH
CO
t^
O
CM
"O
O
•H
fl)
Sampling Pi
CO
CO
en
a)
c
.0
CO
Cn
_j
November ;
et
CO
en
rH
^
Q)
O
-P
O
O
a
-H
T5
a)
S
en
CM
rH CM
O Cn
H
+1
+1
en ^
co
CO CO • • CM
2 20 rH
t"~
rH •*
• co
o
in
+1
+1
en
^r co o ON
•* co CM en
en -«3* o n
*d*
rH
O Ol
O rH
OJ
+1
+1
rH O CM f-
• rH CO •
rH • • rH
rH VD O rH
^
^0) -D
O 0)
D tr>
T3 (TJ ^J i — 1
O >H rH -rt
M O -H O
o< fa S w
en
*
in
VO '
in
in
in
in
en
CM
a\
VO
in
CO
^
+1
CO
en
r-
-P
C
0)
e
T3
0)
CO
in
II
^
-p
C
0)
e
•rH
•0
Q)
CO
.^
in
II
C
, — |
•H
0
CO
II
rH
-H
S
• •.
CM
11
C
0)
tr>
(0
O
fa
•^
CM
II
C
a>
o
3
0
i-i
cu
j^-
CO
en
rH
i_i
Q)
0
4->
O
O
0
fa
in
II
c
-p
c
0)
g
-rH
'O
0)
CO
in
II
C
rH
-rH
O
CO
.^
ro
II
C
rH
-rH
S
CM
II
C
CO
CT
(0
M
0
fa
• -.
c-H
II
C
Q)
U
3
TJ
0
CU
CO
en
rH
0)
.Q
Q)
O
2
in
O
fa
m J3
in
II
-P
c
0)
e
-H
•o
0)
co
•«.
in
II
c
rH
•H
0
W
.^
CO
II
C
_y
r-]
-H
00
CO
en
rH
0)
c
3
O
fa
o 1
CO
V
C
O
-------
O>
V,
D)
Q.
CO
*-•
c/5
O)
c
c
o
o
c
o
o
Ml^MnMiMM
10
N.
•
1
to
C)
1
in
CM
c
O
r\i
i
UJ
CC
£>
•t-* .
«= T3
0) C
O «
c ^:
O 3
O DC
c
o
E
k.
0)
0)
^-*
03
•4*
(0
w
a)
O
•s
o
05
4)
CC
k_
03
•*- c
CD —
5
> Q.
0
o
•*-•
c
E
o.
0)
Q
tt)
o
2
o
CO
11-46
-------
N. r-
00 00
co
o> o> 2
0>
o 5 5
O Z -3
0)
+••
CO
O)
c
o
OJ O H~-
I O 3
T-. c OC
wO.E
DC*- ro
UL| *-
Q
-------
40
c
I
|
o
•*•*
CO
D)
«>.
O)
c
o
c
o
o
o
CD
I
o
IO
o
CO
o
CM
CM C T3
CM 0) C
i O CO
^ 5^
r- O 3
UJ O QC
<3 — w
— ca JD
"• •- Q.
SI
"-S
Q —
9 o
o
£
O
at
+*
.#••
n)
c
4)
E
4-1
k.
ca
Q.
03
Q
O
CO
11-48
-------
CO
1
rH
rH
CQ
«S
•H
"8
0)
S
lental
s
c
o
•H
t^
w
c
•H
05
c
o
Concentrati
so) (pg/g)
Q +'
Q ....
O C.
Q
f-t
•rH
0
•H
13
. |
a
o
N '
a\
Lorodib<
Octach!
•o
o
•H
rH
0)
O4
D>
C
•H
rH
a
B
re
w
CO
co
CTi
rH
Q)
C
^"3
.0
r-
co
cr>
rH
M
' ' ':
CO
in
O CT>
VO
4-1 •*
r~ +i
. r- o • cr>
VD no M
CM
r-
rH
O O
CTl
+1
rH CM
CO "d* tH
vo in -CM
VO •* CM in
^0) TJ
O 0)
3 O>
T! (0 X rH
O Vj rH -H
(H O --H O
PJ P-. S W
CM
VD
rH
+1
rH
CM
' "•'
CM
CM
rH
+1
O
CM
a\
CO
+1
CM
CO
rH
4J
C
Q)
B
Tf
0)
in
II
•P
0)
B
•H
•o
0)
w
in
II
C
•H
o
w
CO
II
rH
s
• fc.
CM
II
c
0)
JJI
R3
^-j
0
CM
• *.
CM
II
q
reduce
cu
CO
r-
^_
0)
Octob
S-i
o
fe
<0
in
II
C
-j
C
a)
B
•H
•d
w
*"•
in
II
C
r— i
o
u
i— i
•H
s
CM
II
a)
(0
^H
O
^LI
•^
H
II
q
Produce
CO
rH
j^i
Q)
O
a
o
.0
in
II
q
Q)
B
•rH
T3
0)
W
•-
in
II
q
rH
-rH
o
w
u
rH
•H
S
CO
CO
CT\
rH
Q)
C
0
u
r">
V
q
o
culated
rH
(0
O
•P
O
Q
TJ
0)
rH
CU
B
ro
Uj
-P
0
"Z,
II
2
11-49
-------
11.3.1. Produce and Forage. Most of the 2,3,7,8-TCDD equivalent
average concentrations were derived from values that were non-
detectable but were conservatively set equal to the detection
limit. The average concentrations in the forage and produce ranged
from 4.88 to 11.1 pg/g, as shown in Table 11-7. The 2,3,7,8-TCDD
equivalent concentrations are lowest in forage samples, with
averages of 6.10 and 4.88 pg/g for samples taken in October and
November 1987, respectively. The carrot sample had the highest
2,3,7,8-TCDD equivalent concentration of 11.2 pg/g. Potato samples
collected in October and November 1987, had average concentrations
of 10.9 and 9.44 pg/g, respectively.
Although TCDD contamination of fruits, vegetables or grains
has not been reported in the United States (all congeners of
PCDD/PCDF were not considered), 2,3,7,8-TCDD was found in locally
grown garden fruits and vegetables (concentration not reported)
following an industrial accident in Seveso, Italy in 1976 (U.S.
EPA, 1985).
11.3.2. Milk, Sediment and Soil. Table 11-7 lists average
concentrations and corresponding standard deviations by sampling
period for milk, sediment and soil. The majority of PCDD/PCDF
isomer concentrations in these samples were nondetectable, and were
set equal to the detection limit for the purposes of calculating
average 2,3,7,8-TCDD equivalent concentrations. Statistical
analyses of the milk, sediment and soil samples indicate that there
11-50
-------
were no statistically significant differences (ANOVA and Kruskal-
• ' i • , -..,".' ' .
Wallis tests) between the concentrations of PCDD/PCDFs (as 2,3,7,8-
TCDD equivalents) detected while the MWG was in operation and the
concentrations found before the MWC was operational. Similar
results were observed when the OCDD congener data were analyzed.
No statistically significant differences (ANOVA and Kruskal-Wallis
tests) were observed between pre-operational and operational OCDD
concentrations in soil or sediment. However, both the ANOVA and
Kruskal-Wallis tests indicated that the OCDD concentration in milk
was statistically significantly higher in October 1987 than in
November 1987 or June 1988.
2,3,7,8-TCDD equivalent concentrations of all Rutland milk
samples were within an order of magnitude of the concentration of
the Route 100 sample collected for background comparison (0.120
pg/g). The 2,3,7,8-TCDD equivalent concentrations detected in milk
from cows around the Rutland facility, both before and during
operation of the MWC, are also within an order of magnitude of
those reported in milk from cows located near incinerators in
Switzerland (0-2 ppt; Rappe et al., 1987).
The PCDD/PCDF concentrations detected in sediment samples in
this study are generally within the range of concentrations
measured in sediments exposed to combustor emissions in other
areas. Czuczwa et al. (1984) measured sediment concentrations at
several depths in Siskiwitt Lake on Isle Royale in Lake Superior,
and found similar levels. Comparable PCDD/PCDF concentrations were
11-51
-------
found in archipelago of Stockholm, Sweden (Rappe and Kjelier,
1987b), and at various locations in Japan (Yasuhara et al., 1987)<
The mean 2,3,7,8-TCDD equivalent concentration in soil
collected in June 1988 was 12.4 pg/g. This was similar to the mean
concentration of samples collected in October 1987 (11.7 pg/g), but
greater than the average concentration of samples taken in November
1987 (3.99 pg/g). There was high variability in concentrations,of
these samples. For example, the three sampling periods at the
Route 4 site had one sample that was at least fifteen times
greater, and one sample up to 42 times greater, than the other
(e.g., values of 2.32 and 96.6 pg/g for October). The average
total PCDD/PCDF concentrations in the Rutland area are greater than
concentrations found in soil samples taken from rural areas in
Europe (Rappe and Kjelier, 1987). However, the average values in
the Rutland area are generally within the range of soil
concentrations measured near stack emissions in Florence, Italy
(Berlincioni and di Domenico, 1987) and in various locations in
Japan (Yasuhara et al., 1987). For example, Berlincioni and di
Dimenico (1987) sampled topsoil from open meadows and farmland
within a 1 km radius of an incinerator, and found comparable
results (0-500 pg/g).
11.3.3. Conclusion. Since samples of forage and produce were only
collected prior to commencement of operations of the MWC, it was
not possible to determine whether concentrations of PCDD/PCDFs in
11-52
-------
these media were altered due to combustor emissions.. In samples
of milk, sediment and soil, there were no statistically significant
increases in 2,3,7,8-TCDD equivalent concentrations in samples
collected after commencement of operations of the MWC, when
compared with samples taken prior to operation. However, because
many PCDD/PCDF concentrations were non-detectable and assumed to
be equal to the limit of detection and because sample sizes were
small, this conclusion contains some uncertainty. For the one
congener for which concentrations were consistently measurable
(OCDD), no contribution of MWC operation to milk, sediment or soil
OCDD concentrations was observed.
11.4. SUMMARY
Small sample sizes resulting from single samples being taken
at - each field monitoring site, large numbers of samples with
concentrations at or close to the limit of detection of the
analytical methodology and large variability; of detectable sample
concentrations precluded a quantitative risk assessment (such as
determination of human exposure via the foodchain using the
observed sample concentrations as input data). In the qualitative
analysis performed, there were no apparent differences in the
concentrations of metals, PCB or PCDD/PCDF (as 2,3,7,8 - TCDD
equivalents) in produce, forage, milk,.soil, sediments or water
(metals only) before or during the operation of the;Rutland MWC. ..
The measured concentrations are within the range of background
11-53
-------
concentrations found in other geographical areas. The sporadic
statistically significant findings are not supported by similar
altered concentrations in other media, such as ambient air or the
food chain, which would have been expected to have been altered
coincidently. The values found in Rutland do not suggest
alterations due to operation of the MWC, and are therefore
considered indicative of typical background concentrations.
11-54
-------
12. CONCLUSION
The objective of this multimedia, multipollutant field study
of the MWC in Rutland, Vermont was to determine human exposure
resulting from MWC emissions. With the exception of PCDD/PCDFs and
lead, the majority of pollutants in the ambient air and
environmental media were not present in concentrations that could
be detected by the analytical methods employed, a direct
determination of the contribution of the incinerator to the
measurable concentration of pollutants was not possible.
Therefore, an analysis of the likelihood that the incinerator was
a primary contributor to the measured pollutant concentrations was
assessed using several alternative approaches.
The conclusion reached by evaluation of the collected field
samples is that the measured concentrations of the pollutants in
the ambient air and environmental media cannot be correlated with
the emissions from or operation of the MWC. The MWC does not appear
to be the primary source of these pollutants. Evidence for this
conclusion comes from both qualitative and quantitative evaluation
of the measured pollutant concentrations in the ambient air and
environmental media, as well as comparison with predicted ambient
air concentrations of the pollutants using local meteorologic
information.
12-1
-------
Many af the pollutants were not detectable in the ambient air (
and, when they were, the' sites and days at which they were detected
varied. If the MWC had been 'the primary source of these
pollutants, the detectable concentrations would have been expected
to occur more consistently at a given location and diiring the, time
period when the incinerator was operating. Instead, detectable
concentrations of • several pollutants, .appeared to be randomly
observed at the different monitoring sites. Furthermore, very high
concentrations of some of the pollutants, particularly PCDD/PCDFs,
occurred in December 1988 and January 1989, when the MWC was not
operating.
The four alternative approaches employed to address source
apportionment all indicated other sources were likely to be
contributing to the measured concentrations. In one approach, the
possible correlation of particulate (PM-10 fraction) concentrations
for the period of November 5, 1987 through October 6, 1988 with the
amount of waste burned daily was investigated since many .pollutants
adhere to particulate matter and many of the pollutant
concentrations were not detectable. This analysis did not reveal
a significant correlation between these variables, suggesting that
the MWC was not the primary source of the particles in the Rutland
ambient air.
The comparison of the levels of mutagenic activity associated
with particles in the ambient air with both the PM-10 particle
concentration and the amount of waste burned per day further
12-2
-------
supports the conclusion that the incinerator is not a. significant
source of these pollutants. The analysis of the relationship
between the amount of waste burned daily and mutagenicity was
conducted because emissions of organic mutagens result from
i
incomplete combustion of municipal waste (Watts et al., 1989).
While there was a positive correlation between particle
concentration and mutagenic activity at all four monitoring sites,
there was no correlation between the number of tons of waste burned,
per day and the mutagenic activity at any of the sites (nor between
the amount of waste burned and particle concentrations as discussed
above).
The source contribution of the pollutants measured in the
ambient air was also analyzed by comparing PCDD/PCDF congener
profiles of ambient air with potential sources. Ballschmiter et
al. (1986) have suggested that the distribution patterns of the
various congeners may^indicate the nature of the PCDD/PCDFs. It
would be expected that if one source was the primary contributor
of these chemicals, then the congener patterns of the Rutland
ambient air and that source would resemble each other. The
PCDD/PCDF distribution patterns of homologues were found to differ
between the ambient air monitoring sites as well as between the
sampling days at the same site, thus indicating that there were
various local sources influencing the PCDD/PCDF profile. The
congener profiles of Rutland ambient air were also compared with
congener profiles of the stack emissions of the MWC and the
12-3
-------
emissions from wood burning .systems. Profiles of the ambient air
samples collected during the winter months did not resemble either
the emissions from wood burning systems nor those of the MWC stack'
emissions. Although there was uncertainty in the interpretation
of the profiles due to the lack of daily MWC emission data, it can
ft .
be concluded that the PCDD/PCDFs originated from a variety of
sources.
The potential contribution of the MWC to the measured
pollutants in the ambient air was also assessed by comparing the
measured ambient air concentrations with concentrations predicted
by air dispersion modeling with local meteorologic information
using two nonparametric statistical methods. Only lead and
PCDD/PCDFs (as 2,3,7,8-TCDD equivalent concentrations and OCDD)
were analyzed, since the other pollutants were not detectable at
frequencies sufficient for a statistical analysis. The analysis of
lead showed there was no correlation between the measured and
modeled concentrations, as would be expected if the incinerator was
not the primary source. Additionally, it was apparent that the
SLAMS had another significant source of lead contributing to the
measured air concentrations. While one statistical test of
2,3,7,8-TCDD equivalent concentrations suggested a possible
relationship between .the maximum concentration predicted to occur
from the MWC and that measured in the ambient air, this
relationship was not supported by the other statistical test nor
by results of the statistical analysis of OCDD. Furthermore, the
12-4
-------
use of 2,3,7,8-TCDD equivalent concentrations for analysis of
ambient air concentrations introduces uncertainty since it
represents a composite of both chemical concentration andtoxicity
information.
The concentrations measured in ambient air in this study were
compared with those of other rural areas. Arsenic and chromium
levels in ambient air in rural areas of the United States
(Fishbein, 1984) were below the analytical detection limits of
this study. Concentrations of 2,3,7,8-TCDD equivalents for rural
areas were not available. However, the total detected PCDD/PCDF
concentrations have been reported for ambient air in Ohio (Czuczwa
and Edgerton, 1986; Tiernan et al., 1988). The maximum
concentrations of PCDD/PCDFs detected in ,these studies were
similar to or slightly greater than (within an order of magnitude)
those detected in Rutland, Vermont in this study.
To assess the potential contribution of the MWC to pollutant
concentrations in water, sediment, soil, milk and food chain,
parametric and non-parametric statistical comparisons of data
pooled across the various sampling locations were conducted.- The
results of these analyses indicated that, even though many
pollutant concentrations were .non-detectable and conservatively
set equal to the methodologic limits of detection, there were no
apparent increases in metal, PCB or PCDD/PCDF concentrations in the
environmental media during the period the Rutland MWC was
operational relative to the period prior to combustor operation.
12-5
-------
These findings are supported by the lack of altered pollutant
concentrations in-the ambient air that would have been expected to
have been altered coincidentally with those of the environmental
media. In addition, the concentrations of pollutants in the
environmental media were similar to those found at other
geographical locations.
All of the foregoing approaches to assessing the contribution
of the MWC to pollutant concentrations in Rutland, Vermont contain
uncertainty related to design of the study and analytical methods,
as occur in any field study. Because of practical limitations
associated with the selection of sites, the monitoring sites could
not be located at the exact point where the initial air dispersion
modeling had predicted the maximum ground-level concentrations to
occur. Additionally, there were limitations with the air
dispersion modeling. Air dispersion modeling was performed using
limited site-specific data (such as wind speed and wind direction
data). Site-specific mixing height data and stability categories
were not available and had to be derived for the ISCST model. The
modeled ground-level concentrations of the metals, except chromium
and nickel on two days, were less than the detection limits used
for the measured concentrations on these pollutants, confirming the
results that they would not have been expected to have been
quantified.
12-6
-------
While this field study did not show that the MWC was a primary
contributor to the measured levels of pollutants, the results
contain information about the background levels of pollutants and
the contribution of other sources to the Rutland, Vermont area.
12-7
-------
-------
13. REFERENCES
Agency of Environmental Conservation, State of Vermont, 1984. Air
Pollution Permit for Vicon Recovery Systems, Rutland Resource
Recovery Facility.
Agency of Environmental Conservation, State of Vermont. 1986.
Amended Air Pollution Control Permit for Vicon Recovery Systems,
Rutland Resource Recovery Facility.
Allaway, W.H. 1968. Agronomic controls over the environmental
cycling of trace elements. In; A.G. Norman (ed.) Advances in
Agronomy. Volume 20. Academic Press, New York, NY.
Baes, C.F., R.D. Sharp, A.L. Sjoreen and R.W. Shor. 1984. A
review and 'analysis of parameters for assessing transport of
environmentally released radionuclides through agriculture.
Prepared for the U.S. Department of Energy under contract No. DE-
AC05-840R21400.
Ballschmiter, K. , H. Buchert, R. Niemczyk, A. Munder, and M.
Swerev. 1986. Automobile exhaust versus municipal-waste
incineration as source of the polychloro-dibenzodioxins (PCDD) and
-furans (PCDF) found in the environment- Chemosphere 15(7): 901-.
915
Baxter, J.C., D.E. Johnson, and E.W. Kienholz. 1983. Effects on
cattle from exposure to sewer sludge. U.S. Environmental
Protection Agency, Washington, DC. EPA-500/2/83-012.
Berlincioni, M. and A. di Domenico. 1987. Polychlorodibenzo-p-
dioxins and polychlorodibenzofurans in the soil near the municipal
incinerator of Florence, Italy. Environ. Sci. Technol. 21(11):
1063-1069.
Bernstein, L., J._ Kaldor, M. McCann and C. Pike. 1982. An
empirical approach to the statistical analysis of mutagenesis data
from the Salmonella test. Mutat. Res. 97: 267-281.
Beyer, K.W., R.L. Chariey and B.M. Mulkern. 1982. Heavy metal
concentrations in earthworms from soil amended with sewage sludge.
J. Environ. Qual. 11(3): 381-385.
Bignoli, G. and E. Sabbioni. 1984. Environmental assessment of
arsenic released from potential pollution sources!. Environ.
Monitoring and Assess. 4: 53-65. .
Booz Allen and Hamilton, Inc. 1983. A background document on
cadmium in municipal sewage sludge. Revised Draft. Prepared for
U.S. EPA Sludge Task Force. April 29, 1983.
13-1
-------
Bowen, H.J.M. 1966. rrace Elements in Biochemistry.
Press, New York. (Cited in Fishbein, 1981).
Academic
Brown, K.W., J.W. Mullins, E.P. Richitt, G.T. Flatman and S.C.
Black. 1985. Assessing soil lead contamination in Dallas, Texas;
Environ. Monit. and Assess. 5: 137-154.
Gary, A.E., J.A. Gowen, H. Tai, W.G. Mitchell and G.B. Wiersma.
1979. Pesticide residue levels in soils and crops from 37 states,
1972 - National Soils Monitoring Program (IV). Pestic. Monit. J.
12: 209-229.
Gary, A.E. 1982. Chromium in air, soil and natural waters. In;
Biological and Environmental Aspects of Chromium. S. Langard
(ed.), Elsevier Biomedical Press, New York, NY.
Cerelius. 1974. (Cited in WHO, 1981).
Christensen, E.R. and C.K. Lo. 1986. Polychlorinated biphenyls
in dated sediments of Milwaukee Harbor, Wisconsin. Environ.
Pollut. 12: 217-232.
Conover,
Edition.
W. J. 1971. Practical Nonparametric Statistics, 2nd
John Wiley & Son.
Greaser, C.S. and A.R. Fernandes. 1986. Background levels of
polychlorinated biphenyls in British soils. Chemosphere. 15: 499-
508. • •
Czuczwa, J.M. and S.A. Edgerton. 1986. Congener profiles of
polychlorinated dibenzo-p-dioxins and dibenzofurans in atmospheric
samples. APCA. 634-639.
Czuczwa, J.M., B.D. McVeety and R.A. Kites. 1984. Polychlorinated
dibenzo-p-dioxins and dibenzofurans in sediments from Siskiwit
Lake, Isle Royale. Science. 226: 568-569.
Demayo, A., M.C. Taylor, K.W. Taylor and P.V. Hodson. 1982. Toxic
effects of lead and lead compounds on human health, aquatic life,
wildlife, plants and livestock. CRC Grit. Rev. Environ. Control.
12(4): 257.
Eitzer, B.D. and R.A. Kites. 1989. Atmospheric transport and
deposition of polychlorinated dibenzo-p-dioxins and dibenzofurans.
Environ. Sci. Technol. 23(11): 1396-1401.
Erdman, J.A., H.T. Shaklette and J.R. Keith. 1976. Elemental
composition *of selected native plants and associated soils from
major vegetation-type areas in Missouri. Geological Survey
Professional Paper, 954-C. Washington, DC.
13-2
-------
Fishbein, L. 1981. Sources, transport and alteration of metal
compounds: An overview. I. Arsenic, beryllium, cadmi,um, chromium
and nickel. Environ. Health Persp. 40: 43-64.
Fishbein, L. 1984. Overview of analysis of carcinogenic and/or
mutagenic metals in biological and environmental samples. 1.
Arsenic, beryllium, cadmium, chromium and selenium. Int. J.
Environ, and Anal. Chem. 17: 113-170.
Fitzgerald, B. 1990. Vermont Air Pollution Control Division,
Agency of Natural Resources. Letter to C. Sonich-Mullin, Systemic
Toxicants Assessment Branch, U.S. EPA, Cincinnati, OH, May 7,
1990.
Fleischer, M. 1970. Summary of the literature on the inorganic
geochemistry of mercury. In; Mercury in the Environment.
Geological Survey Professional Paper No. 17. Washington, DC. .
Frank, R., H. Braun, K. Ishida and P. Suda. 1976,. Persistent
organic and inorganic pesticide residues in orchard soils and
vineyards of southern Ontario. Can. J. Soil Sci. 56: 463-484.
Gerdes, R.A., J.E. Hardcastle and K.T. Stabenow. 1974. Mercury
content of fresh fruits and vegetables. Chemosphere. 1:13-18.
Harless, R.L. 1989, U.S. EPA, Research Triangle Park, NC. Letter
to M. Fragge, Syracuse Research Corporation, Cincinnati, OH, June
23, 1989. •'
Harless, R.L. and Lewis. 1987. Analytical procedures and quality
assurance plan for the determination of PCDDs and PCDFs in ambient
air near the Rutland, Vermont municipal incinerator.
Harless, R.L. and D. McDaniel. 1988. Method for determination of
polychlorinated dibenzo-p-dioxins and dibenzofurans., in stack gas
emissions and ambient air. Presented at 1988 EPA/APCA Symposium
on Measurements of Toxic and Related Air Pollutants, Raleigh, NC,
May 1-4, 1988. ,
Harper, S.L., J.F. Walling,
Simplex optimization of
atmospheric particulate.
D.M. Holland and L.J. Pranger. 1983.
multielement ultrasonic extraction of
Anal. Chem. 55(a): 1553-1557.
Heit, M. C. Klusek and J. Baron. . 1984. Eviidence of deposition
of anthropogenic pollutants in remote Rocky Mountain lakes. Water
Air Soil .Pollut. 22: 403-416.
Hem, J.D. 1970. Study and interpretation of the chemical
characteristics of natural water. Geological Survey Water Supply
Paper 1473. U.S. Governmental Printing Office, Washington,, DC.
13-3
-------
Holmgren, C. 1983. Personal Communication. National Soil Survey
Laboratory. Soil Conservation Service. USDA, Lincoln, NE.
IARC. 1973. Some inorganic and organometallic compounds,
Monographs on the Evaluation of Carcinogenic Risk of Chemicals to
Man, Vol. 2, International Agency for Research on Cancer, Lyon. .
Jenkins, 1980. Biological monitoring of toxic trace metals.
Volume 1.' Biological Monitoring and Surveillance. Office of
Research and Development, Las Vegas, NV. EPA 600/3-80-089.
Johnson, R.D. and D.D. Manske. 1976. Pesticide residues in total
diet samples (IX). Pest. Monit. J. 9: 157-159.
Kneip, T.J., M. Eisenbud, C.D. Strehlow and P.C. Freudenthal.
1970. Airborne particulates in New York City. J. Air Pollut.
Control Assoc. 20(3): 144-149
Koop, J.F. 1970. The occurrence of trace elements in water. In:
Proceedings of the Third Annual Conference on Trace Substances in
Environmental Health, 1969,1 D.D. Hemphill, ed. .University of
Missouri, Columbia, pp. 59-73.
Lindsay, W.L. 1979. Chemical Equilibria in Soils. John Wiley &
Sons, New York.
Lodi, C. 1988. Compliance Emission Test Report: Vicon Recovery
Systems, Rutland Resource Recovery Facility. Prepared by Eastmount
Engineering, Inc. for Vicon Recovery Systems.
Logan, T.J. and R.H. Miller. 1983. Background levels of heavy
metals in Ohio farm soils. Research Circular 275. The Ohio State
Univ., Ohio Agric. Res. and Development Center, Wooster, OH.
Maron, D.M. and B. Ames. 1983. Revised methods for Salmonella
mutagenicity test. Mutat. Res. 113: 173-215.
McCabe, L.J.,'J.M. Symons> R.D. Lee and G.G. .Robeck. 1970. Survey
of community water supply systems. J. Am. Water Works Assoc. 62:
670-687.
Morita, M., A. Yasuhara and H. Ito. 1987. Isomer specific
determination' of polychlorinated dibenzo-p-dioxins and
dibenzofurans in incinerator 'related samples in Japan. Environ.
Sci. Technol. 21: 971-979.
National Academy of Sciences (NAS). 1974. Medical and Biological
Effects of Environmental Pollutants: Chromium. National Academy
of Sciences, Washington, DC. •
13-4
-------
National Academy of Sciences (NAS). 1975. Medical and Biological
Effects of Environmental Pollutants: Nickel. National Academy of
Sciences, Washington, DC.
National Academy of Sciences (NAS). 1977. Drinking Water and
Health, NAS Safe Drinking Water Committee Report.
Peters, J. and B. Seifert. 1980. Losses of benzo(a)pyrene under
the conditions of high-volume sampling. Atmos. Environ., 14: 117-
119.
Pierce, F.J., R.H. Dowdy and D.F. Grigal. 1982. Concentration of
six trace metals in some major Minnesota soils series,, J. Environ.
Qual. 11(3): 416-422.
Pyles, R.A. and E.A. Woolson. ' 1982. Quantitation and
characterization of arsenic compounds in vegetables grown in
arsenic acid treated soil. J. Agr. Food Chem. 30: 866-870.
Rappe, C., N. Nygren and G. Lindstrom. 1987. Polychlorinated
dibenzofurans and dibenzo-p-dioxins and other contaminants in cow
milk from various locations in Switzerland. Environ. Sci. Technol.
21: 964-970.
Rappe, C. and L. Kjeller. 1987. PCDDs and PCDFs in environmental
samples: air, particulates, sediments and soil. Chemosphere.
16(8/9): 1775-1780.
Ratsch, H.C. 1974. Heavy metal accumulation in soil and
vegetation from smelter emissions. ROAP/TUSK 21 BCI-01 U.S. EPA,
Office of Research and Development, Corvallis, OR.
Ryan, J.A., H.A. Pahren and J.B. Lucas. 1982. Controlling cadmium
in the human food chain: A review and rationale based on health
effects. Environ. Res. 28: 251-302.
Sander, T. 1989. PEI, Associates. Personal communication with
M. Eichelberger, March 24, 1989. ; .
Sandhu, S. and W. Lower. 1987. In situ monitoring of
environmental genotoxins. In: Short-Term Bioassays in the Analysis
of Complex Environmental Mixtures V. Environmental Science
Research, Volume 36, Plenum Press, NY, p. 145-160.
Spittler, T.M. 1973. A system for collection and measurement of
elemental and total mercury in ambient air over a concentration
range of .004 to 25 Mg/m • U.S. EPA Region I, Surveillance and
Analysis Division Laboratory. Needham:Heights, MA. •
Thoma, H. 1988. PCDD/PCDF Concentrations in chimney soot from
house heating systems. Chemosphere 17: 1369-1379.
13-5
-------
Tiernan, T.O., D.J. Wagel, G.F. VanNess, J.H. Garrett, J.G. Solch,
and A.A. Harden. 1988. Evaluation of the collection efficiency
of a high volume sample fitted with an organic sampling module for
collection of specific polyhalogenated dibenzodioxin and
dibenzofuran isomers present in ambient air. Proceedings of the
1988 EPA/APCA International Symposium: Measurement of Toxic and
Related Air Pollutants, Raleigh, N.C., May 1-4, 1988.
U.S. Department of Commerce, National Oceanic and Atmospheric
Administration. 1990. Mixing Height Inventory (Upper Air Albany
& Surface conditions Burlington, 1988). National Climatic Data
Center, Asheville, N.C.
U.S. EPA. 1978 (a) . Reviews of the Environmental Effects of
Pollutants: VI. Beryllium. Office of Research and Development,
Health Effects Research Laboratory, Cincinnati, OH. EPA-600/1-
78-028. NTIS PB-290966.
U.S. EPA. 1978 (b) . Reviews of the Environmental Effects of
Pollutants: III. Chromium. Office of Research and Development,
Health Effects Research Laboratory, Cincinnati, OH. EPA-600/1-
78-023. NTIS PB-282796. ,
U.S. EPA. 1979. Method of Chemical Analysis of Water and Wastes.
Office of Research and Development, Environmental Monitoring and
Support Laboratory, Cincinnati, OH. EPA-600/4-79-020.
U.S. EPA. 1980 (a). Ambient Water Quality Criteria for Chromium.
Prepared by the Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, Cincinnati, OH for
the Office of Water Regulations and Standards, Washington, DC. EPA
440/5-80-035. NTIS PB81-117467/AS.
U.S. EPA. 1980 (b) . Ambient Water Quality Criteria for Lead.
Prepared by the Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, Cincinnati, OH for
the Office of Water Regulations and Standards, Washington, DC.
EPA 440/5-80-057. NTIS PB81-117681/AS
U.S. EPA. 1980 (c). Ambient Water Quality Criteria for Nickel.
Preparedby the Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, Cincinnati, OH
fortheOffice of Water Regulations and Standards, Washington, DC.
EPA 440/5-80-060. NTIS PB81-117715/AS.
U.S. EPA. 1982 (a). User's Instructions for the SHORTZ and LONGZ
Computer Programs. Philadelphia, PA. NTIS PB-146 100.
U.S. EPA. 1982 (b). Methods for Organic Chemical Analysis of
Municipal and Industrial Wastewater. Office of Research and
Development, Environmental Monitoring and Support Laboratory,
Cincinnati, OH. EPA-600/4-82-057.
13-6
-------
U.S. EPA; 1983 (a). Standard Operating Procedure for the ICP-
AES Determination of Trape Elements in Suspended Particulate Matter
Collected on Glass-Fiber Filters. Office of Research and
Development, Environmental Monitoring System Laboratory, Research
Triangle Park, NC. SOP-EMD-OQ2. . ,
U.S. EPA. 1983.-(b). Methods for Chemical Analysis of Water and
Wastes. Office of Research and Development, Environmental
Monitoring and Support Laboratory, Cincinnati, OH.•-EPA-600/4-79.-
020. .; -. , • ...••• -. ; . .. v,' : .•• •• .-• •; • -.
U.S. EPA. 1983 (c) . Health Assessment Document for Chromium.
External Review Draft. Research Triangle Park, NC. EPA 600/8-
83-014A. NTIS PB83-252205/AS.
U.S. EPA. 1984 (a). Compendium of Methods for the Determination
of Toxic Organic Compounds in Ambient Air., Office of Research and
Development, Environmental Monitoring System Laboratory, Research
Triangle Park, NC. ,EPA-600/4-84-041. ,NTIS PB87-168688.
U.S. EPA. 1984 (b) . Standard Operating Procedure for 'NAA
Determination of Trace Elements in Suspended Particulate Matter
Collected on Glass-Fiber Filters. Office of Research and
Development, Environmental Monitoring System Laboratory, Research
Triangle Park, NC. SOP-EMD-017. . ;,
U.S. EPA. 1984 (c) . Mercury Health Effects Update. Office of
Health and Environmental Assessment, Environmental Criteria
Assessment Office, Research Tiangle Park, NC. EPA 600/8-84-019F.
NTIS PB85-123925/AS !
U.S. EPA. 1984(d). Review of National Standards for Mercury.
Office of Air Quality Planning and Standards, Research Triangle
Park, NC. EPA-450/3-r84-014.
U.S. EPA. 1985 Health Assessment Document for Polychlorinated
Dibenzo^p-Dioxins. Environmental Criteria and Assessment Office,
Cincinnati, OH. EPA-450/3-84-014
U.S. EPA. 1986 (a). U.S. Environmental Protection Agency.
Industrial Source Complex (ISC) Dispersion Model User's Guide.
2nd ed. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. EPA-450/4-86-005a. NTIS PB86-234259.
U.S. EPA. 1986 (b). Standard Operating Procedure for Ultrasonic
Extraction and Analysis of Residual Benzo(a)pyrene from Hi-Vol
filters via Thin-layer Chromatography. Environmental Monitoring
Systems Laboratory. Research Triangle Park, NC. December, 1986.
SOP-MDAD-015.
13-7
-------
U.S. EPA. 1986 (c). Hazardous Waste Treatment, Storage, Disposal
Facilities. Field Sampling and Analysis Protocol for Collecting
and Characterizing Soil Samples from TSDF's. Office of Air Quality
Planning and Standards, Emission Standards and Engineering
Division, Research Triangle Park, NC. EPA-450/3-86-014.
U.S. EPA. 1986 (d) . Air Quality Criteria for Lead. Office of
Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Research Triangle Park, NC. EPA 600/8-83-028F.
NTIS PB87-142378/AS
U.S. EPA. 1986 (e) . Health Assessment Document for Nickel and
Nickel Compounds. Office of Health and Environmental Assessment,
Environmental Criteria Assessment Office, Research Tiangle Park,NC.
EPA/600/8-83/012FF. NTIS PB86-232212/AS
U.S. EPA. 1987 (a). Municipal Waste Combustion Study: Assessment
of Health Risks• Associated With Municipal Waste Combustion
Emissions. Prepared by RADIAN Corporation. EPA/530-SW-87-021g.
U.S< EPA. 1987 (b) . Air Dispersion Modeling of a Municipal Waste
Combustor in Rutland, Vermont. Prepared by PEI Associates, Inc.
under Contract No. 69-02-4351. Office of Air Quality Planning and
Standards, Pollutant Assessment Branch, Research Triangle Park,
NC. ,
U.S. EPA. 1987 (c). Standard Operating Procedure No. MB^143/0.
Performing the Kado Assay. June 19, 1987. Prepared by
Environmental Health Research and Testing, Inc. under Contract No.
68-02-4456. Health Effects Research Laboratory, Research Triangle
Park, NC.
U.S. EPA. 1989 Interim Procedures for Estimating Risks Associated
With Exposures to Mixtures of Chlorinated Dibenzo-p-Dioxins and -
Dibenzofurans (CDDs & CDFs) and 1989 Update. Risk Assessment
Forum, Washington, DC. EPA/625/3-89/016.
U.S. EPA. 1990. Methodology for Assessing Health Risks Associated
with Indirect Exposure to Combustor Emissions. Interim Final.
Office of Health and Environmrntal Assessment, Environmental
Criteria Assessment Office, Cincinnati, OH. EPA/600/6-90/003.
NTIS PB 90-187055/AS
U.S. Geological Survey. 1970. Mercury in the Environment.
Geological Survey Professional Paper 713. Washington, DC.
Vermont Air Pollution Control Division, Agency of Natural
Resources. 1985. A Review of the Potential Health Effects of
Dioxin Emissions, Acid Gas Emissions and Disposal of Dioxin
Contaminated Ash from the Vicon Resource Recovery Facility Proposed
for Rutland, Vermont.
13-8
-------
Vermont Air Pollution Control Division, Agency of Natural
Resources. 1987 (a). Cooperative Agreement with the U.S. EPA.
NO. CX-814651-01-0.
Vermont Air Pollution Control Division, Agency of Natural
Resources. 1987 (b) . Rutland Resource Recovery Facility Site
Specific Environmental Assessment Quality Assurance Plan. ,
Vinogradov, A.P. 1959. The geochemistry of rare and dispersed
chemical elements in soils. Consultants Bureau.
Watts, R. , B. Fitzgerald, G. Heil, et. al. 1989. Use of Bioassay
to Evaluate Mutagenicity of Ambient Air Collected Near a Municipal
Waste Combustor. U.S. EPA, Office of Air Quality Planning and
Standards, Research Triangle Park, NC.
WHO (World Health Organization). 1981. Environmental Health
Criteria. World Health Organization Technical Report Series.
Geneva, Switzerland.
Williams, R., T. Pasley, S. Warren, et al. 1988. Selection of a
suitable extraction method for mutagenic activity from wood smoke
impacted air particles. Int. J. Environ. Anal. Chem. 34: 137.
Yasuhara, A., I. Hiroyasu and M. Morita. 1987. Isomer-specific
determination of polychlorinated dibenzo-p-dioxins and
dibenzofurans in incinerator-related samples. Environ. Sci.
Technol. 21(10): 971-979. '
*U.S. GOVERNMENT PRINTING OFFICE:! 991 -5 it 6 -187/20600
-------
-------
-------
c
cn
CD
00
O
O
rn
-o
I
(O
III
If* §3*
§—' ^*
W o
2 o -*
2 o m
r1 ^ o
5 =
5"
a O
05 X
- m
1
w m
= s
o-S
2S
&0
_ S
3 0)
C
T3
> me
CO
5) en
T3
s
CD
s
5
3
L< 3
3 £•
O
CD
§15.
Si -• cT
cf. o X
O 3 m
I 5
01 O
10 3
00 CD
3
5)
CD
cn
CD
Q>
—t
O
3"
13
O
T)
m
3D
Q
is
01
------- |
|