United States Office of Pollution Prevention
Environmental Protection and Toxics
Agency Washington, DC 20460 December 1998
&EPA Ground-Truthing of the Air
Pathway Component of
OPPT's Risk-Screening
Environmental Indicators
Model
DRAFT: Do not cite or quote December 1998
-------
Ground-Truthing of the Air Pathway Component of
OPPT's Risk-Screening Environmental Indicators
Model
Nicolaas W. Bouwes, Sr., Ph.D.
Steven M. Hassur, Ph.D.
Economics, Exposure, and Technology Division
Office of Pollution Prevention and Toxics
United States Environmental Protection Agency
December 1998
DRAFT: Do not cite or quote December 1998
-------
ACKNOWLEDGMENTS
This report evaluates the air pathway component of the Office of Pollution Prevention and
Toxics' (OPPT's) Risk-Screening Environmental Indicators Model. This report is one of many
products of the OPPT's Risk-Screening Environmental Indicators Model Project. The project,
initiated in 1991, has resulted in the Risk-Screening Environmental Indicators Model, a unique and
powerful analytical tool for risk communication. The Indicators Model has the potential to make a
significant contribution to environmental improvement. We wish to thank our contractor, Abt
Associates Inc., for their support and creativity throughout the development of this project.
We also want to thank several persons at State agencies who were very helpful in providing
data and information for the analyses described in this report. These include Mr. Eric Wade and Mr.
Tom Gentile of the New York State Department of Environmental Conservation; Mr. Christopher
Nguyen of the California Environmental Protection Agency's Air Resources Board; Mr. Orlando
Cabrera of the Wisconsin Department of Natural Resources; and, Mr. Greg Stella of EPA's Office
of Air Quality Planning and Standards.
Project Managers: Nicolaas W. Bouwes, Sr., Ph.D.
Steven M. Hassur, Ph.D.
Abt Associates
Project Staff:
Elizabeth Fechner Levy, Project Manager
Brad Firlie, Deputy Project Manager
Hernan Quinodoz, Senior Analyst
Peter Eglinton, Senior Analyst
Jeremy Castle, Research Assistant
Cindy Gould, Research Assistant
DRAFT: Do not cite or quote
December 1998
-------
EXECUTIVE SUMMARY
EPA's Science Advisory Board (SAB) advised the Office of Pollution Prevention and Toxics
(OPPT) to conduct a "ground-truthing" analysis of the exposure model components of OPPT's Risk-
Screening Environmental Indicators Model. The objective of the Indicators Model is the analysis of
Toxics Release Inventory (TRI) releases and their relative risk-related impacts, which can be used for
relative ranking purposes.
In this ground-truthing analysis, the air model component of the Indicators Model was
evaluated. Air pollutant concentrations estimated by the Indicators Model were compared to
concentrations obtained from Air Guide-1 (AG-1), an air dispersion model used by the New York
State Department of Environmental Conservation for regulatory purposes. The air pollutant
concentrations calculated by the Indicators Model are based on a combination of median data (e.g.,
stack height and exit gas velocity) and generic assumptions, whereas the AG-1 model relies on a
greater variety of facility- and stack-specific data. The differences in pollutant concentrations
predicted by both models were analyzed for 24 test cases in New York. This representative sample
was designed to capture the variability observed in three input variables. Four metropolitan areas
were selected to sample different meteorological conditions, and two types of pollutants, with and
without decay rates, were modeled in each metropolitan area. The distribution of stack heights was
represented by three discrete bins, each containing about a third of the stack heights reported by all
TRI facilities in New York. Two test cases (one for a pollutant with a decay rate and one for a
pollutant without a decay rate) were selected from each stack height bin for each metropolitan area.
The Indicators Model estimates air pollutant concentrations for each 1 km2 cell in a 21-km
by 21-km grid surrounding a TRI facility. Each TRI facility is represented with a single stack located
at the center of the central cell in the grid. Cell by cell concentrations predicted by the Indicators
Model and AG-1 were compared by calculating a concentration ratio for each cell (a ratio of one
indicates perfect agreement between the models). Two sets of tests were conducted: in the first, the
Indicators Model used facility-specific median stack heights and exit gas velocities; in the second, the
Indicators Model used stack heights and exit gas velocities corresponding to the median values for
the facility's 3-digit Standard Industrial Classification (SIC) code. These SIC code-based values were
nationally derived, based on available data.
Concentration ratios for individual cells ranged from 0.23 to 3.1 when using facility-specific
parameters, and from 0.25 to 3.4 when using SIC code-based parameters. Average concentration
ratios computed over all 440 cells surrounding a single facility differed by 48 percent or less when
using facility-specific parameters, and by 35 percent or less when using SIC code-based parameters.
Average ratios computed over the 24 test cases were within two percent of unity (with a standard
deviation of 13 percent) when using facility-specific parameters, and within six percent of unity (with
DRAFT: Do not cite or quote ii December 1998
-------
a standard deviation of 13 percent) when using SIC code-based parameters. Thus, the Indicators
Model does not seem to consistently overpredict or underpredict pollutant concentrations.
Average concentration ratios were also computed over concentric square rings around the
central cell. These averages show a pattern consistent across most facilities: concentration ratios
converge to within a narrow band around one as distance from the stack increases. Average
concentration ratios in the innermost ring, where air pollutant concentrations are highest, ranged
from 0.6 to 1.7 when using facility-specific parameters, and from 0.5 to 1.8 when using SIC code-
based parameters. Average ratios at the outermost ring ranged from 0.8 to 1.5 when using facility-
specific parameters, and from 0.6 to 1.2 when using SIC code-based parameters. Overall, the results
obtained demonstrate that predictions of pollutant concentrations are not only comparable, but are
extremely close, even though key input data to the two models are not the same. Although the
Indicators Model is not designed as a substitute for more comprehensive, site-specific risk
assessments, the results of this ground-truthing analysis indicate that the air exposure pathway of the
Indicators Model provides very good estimates of air pollutant concentrations at the facility-specific
level.
Pollutant concentration is one component in the calculation of an Indicator Element, which
can be used to rank facilities. An Indicator Element is the product of three components: the surrogate
dose, which is based on pollutant concentration and exposure assumptions; the toxicity weight for
the chemical of interest; and, the exposed population. Besides pollutant concentration, for a given
chemical with one toxicity weight and one set of exposure assumptions, it is only the variation in
population which influences the value of the Indicator Element. To ascertain the possible impact of
population on the Indicator Element, the relative contribution of each ring to the Indicator Element
was examined. Results indicate that population around a TRI facility can have a significant impact
on Indicator Element values, depending on the population size and distribution relative to the
predicted pollutant concentrations. The accuracy of the Indicator Elements, however, is directly
dependent on the accuracy of the pollutant concentration estimates.
As done in the Indicators Model, Indicator Elements were used to rank facilities. Facilities
corresponding to the 24 test cases were ranked using each set of available concentration estimates:
AG-1, ISCLT3 with facility-specific median stack heights and exit gas velocities, and ISCLT3 with
SIC code-based median stack heights and exit gas velocities. Separate rankings were obtained for
facilities emitting chemicals that decay and those emitting chemicals which do not decay. With only
one exception, the rankings corresponding to different input parameters were identical for both
categories of chemicals, for all three sets of input parameters. This result lends further support to the
use of the Indicators Model to develop relative rankings of TRI facilities.
DRAFT: Do not cite or quote iii December 1998
-------
TABLE OF CONTENTS
ACKNOWLEDGMENTS i
EXECUTIVE SUMMARY ii
TABLE OF CONTENTS iv
LIST OF TABLES v
LIST OF FIGURES vii
1. INTRODUCTION 1
2. DESIGN OF GROUND-TRUTHING ANALYSIS FOR NEW YORK 2
2.1 SCOPE OF THE ANALYSIS 3
2.2 SAMPLING FRAMEWORK 3
2.3 TESTING STRATEGY 5
3. PRELIMINARY TESTS 7
4. MODEL COMPARISON: AG-1 VERSUS ISCLT3 9
4.1 INPUT DATA 9
4.2 RESULTS 10
4.2.1 Impact of Exit Gas Velocity Assumptions 13
4.2.2 Impact of SIC Code-based Stack Height
and Exit Gas Velocity Assumptions 13
4.3 FUGITIVE EMISSIONS ANALYSIS 15
5. PERSPECTIVE ON FINDINGS 17
5.1 CALCULATION OF INDICATOR ELEMENTS 18
5.1.1 Toxicity 18
5.1.2 Surrogate Dose 18
5.1.3 Population 19
5.2 COMPARISON OF INDICATOR SUB-ELEMENTS' CONTRIBUTIONS
BY RING 20
5.3 FACILITY RANKINGS BASED ON INDICATOR ELEMENTS 21
6. CONCLUSION 23
7. REFERENCES 25
DRAFT: Do not cite or quote iv December 1998
-------
LIST OF TABLES
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9
Table 10
Table 11
Table 12
Table 13
Table 14
Ground-Truthing Test Cases 28
Parameter Values Used by Each Model in the Ground-Truthing Exercise 29
Location and Stack Coordinates of TRI Facilities in New York Selected for
the Model Comparison Exercise 30
Facility-Specific Stack Heights (m) 31
Facility Specific Exit Gas Velocities (m/s) 32
Facility-Specific Stack Diameters (m) 33
Facility-Specific Stack Exit Temperatures (K) 34
Facility-Specific Chemical Emission Rates (g/sec) 35
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area,
Chemical Characteristic, and Stack Height. Scenario: Facility-Specific
Median Stack Height and Median Exit Gas Velocity 36
Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations
and by Metropolitan Area. Scenario: Facility-Specific Median Stack
Height and Median Exit Gas Velocity 37
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area,
Chemical Characteristic, and Stack Height. Scenario: Facility-Specific
Median Stack Height and Exit Gas Velocity of 0.01 m/sec 38
Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations
and by Metropolitan Area. Scenario: Facility-Specific Median Stack
Height and Exit Gas Velocity of 0.01 m/sec 39
Comparison of AG-1, Indicators Model and 3-digit SIC Code Parameters
40
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area,
Chemical Characteristic, and Stack Height. Scenario: SIC Code-Based
Median Stack Height and Median Exit Gas Velocity 41
DRAFT: Do not cite or quote
December 1998
-------
Table 15 Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations
and by Metropolitan Area. Scenario: SIC Code-Based Median Stack
Height and Median Exit Gas Velocity 42
Table 16 Exposure Event Counts Surrounding TRI Facilities 43
Table 17 Facility Rankings Based on Indicator Elements for
Chemical with Decay Rate (Toluene) 44
Table 18 Facility Rankings Based on Indicator Elements for
Chemicals without Decay Rate 45
DRAFT: Do not cite or quote
VI
December 1998
-------
LIST OF FIGURES
Figure 1A Example Concentrations (ug/m3) Predicted by AG1. Scenario:
Facility-Specific Median Stack Height and Constant Exit Gas Velocity
of 0.01 m/sec 47
Figure IB Example Concentrations (ug/m3) Predicted by ISCLT3. Scenario:
Facility-Specific Median Stack Height and Constant Exit Gas Velocity
of 0.01 m/sec 48
Figure 1C Example Concentration Ratios (ISCLT3/AG1). Scenario: Facility-Specific
Median Stack Height and Constant Exit Gas Velocity of 0.01 m/sec 49
Figure 2 Example Contour Plots of Concentrations Predicted By Each Model and
Example Contour Plot of the Concentration Ratios. Scenario:
Facility-Specific Median Stack Height and Constant Exit Gas Velocity
of 0.01 m/sec 50
Figure 3 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Albany. Scenario: Facility-Specific Median Stack
Height and Median Exit Gas Velocity 51
Figure 4 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Buffalo. Scenario: Facility-Specific Median Stack
Height and Median Exit Gas Velocity 52
Figure 5 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Rochester. Scenario: Facility-Specific Median Stack
Height and Median Exit Gas Velocity 53
Figure 6 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Syracuse. Scenario: Facility-Specific Median Stack
Height and Median Exit Gas Velocity 54
Figure 7 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Albany. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 55
Figure 8 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Buffalo. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 56
Figure 9 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Rochester. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 57
DRAFT: Do not cite or quote vii December 1998
-------
Figure 10 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Syracuse. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 58
Figure 11 Average (ISCLT3/AG1) by Ring and Stack Height Bin. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 59
Figure 12 Average (ISCLT3/AG1) by Ring and Chemical Characteristic. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 60
Figure 13 Average (ISCLT3/AG1) by Ring and Metropolitan Area. Scenario:
Facility-Specific Median Stack Height and Median Exit Gas Velocity 61
Figure 14 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Albany. Scenario: SIC Code-Based Median Stack
Height and Median Exit Gas Velocity 62
Figure 15 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Buffalo. Scenario: SIC Code-Based Median Stack
Height and Median Exit Gas Velocity 63
Figure 16 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Rochester. Scenario: SIC Code-Based Median Stack
Height and Median Exit Gas Velocity 64
Figure 17 Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case
and For All Cases: Syracuse. Scenario: SIC Code-Based Median Stack
Height and Median Exit Gas Velocity 65
Figure 18 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Albany. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 66
Figure 19 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Buffalo. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 67
Figure 20 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Rochester. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 68
Figure 21 Average (ISCLT3/AG1) by Ring, Chemical, and Case: Syracuse. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 69
Figure 22 Average (ISCLT3/AG1) by Ring and Stack Height Bin. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 70
DRAFT: Do not cite or quote viii December 1998
-------
Figure 23 Average (ISCLT3/AG1) by Ring and Chemical Characteristic. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 71
Figure 24 Average (ISCLT3/AG1) by Ring and Metropolitan Area. Scenario:
SIC Code-Based Median Stack Height and Median Exit Gas Velocity 72
Figure 25 Difference in Median Stack Height (SIC Code-Based Stack Height
Minus Facility-Specific Stack Height) 73
Figure 26 Difference in Median Exit Gas Velocity (SIC Code-Based Exit Gas
Velocity Minus Facility-Specific Exit Gas Velocity) 74
Figure 27 Indicator Sub-element Contributions and Concentration
Ratios (ISCLT3/AG1) by Ring and Case: Albany 75
Figure 28 Indicator Sub-element Contributions and Concentration
Ratios (ISCLT3/AG1) by Ring and Case: Buffalo 76
Figure 29 Indicator Sub-element Contributions and Concentration
Ratios (ISCLT3/AG1) by Ring and Case: Rochester 77
Figure 30 Indicator Sub-element Contributions and Concentration
Ratios (ISCLT3/AG1) by Ring and Case: Syracuse 78
DRAFT: Do not cite or quote ix December 1998
-------
1. INTRODUCTION
The Science Advisory Board (SAB) of the U.S. Environmental Protection Agency (EPA)
advised the Office of Pollution Prevention and Toxics (OPPT) to conduct a "ground-truthing"
analysis of the exposure model components of OPPT's Risk-Screening Environmental Indicators
Model (the Indicators Model). The Indicators Model is intended for analysis of trends in Toxics
Release Inventory (TRI) releases and their relative risk-related impacts. The Indicators Model is not
the equivalent of site-specific risk assessment, in part because a number of simplifying assumptions
have been made to limit the data requirements of the model. These assumptions do not inhibit the
use of the Indicators Model at the national level, but may have the potential to restrict the usefulness
of the model at a site-specific level. To explore the use of the model for more site-specific analyses,
OPPT requested a ground-truthing analysis of the air model component of the Indicators Model. The
purpose of this ground-truthing analysis was to compare air pollutant concentrations predicted using
a combination of median data (e.g., stack height and exit gas velocity) and generic assumptions in the
Indicators Model to pollutant concentrations predicted using facility- and stack-specific data in a
model used for regulatory purposes.
For this analysis, pollutant concentrations estimated by the Indicators Model were compared
to concentrations obtained from an air dispersion model used by the New York State Department of
Environmental Conservation. Section 2 of this memo describes the design of the ground-truthing
analysis. Section 3 presents preliminary model comparisons which were conducted to assess the
default assumptions built into each model. Sections 4 and 5 then present the results of the ground-
truthing analysis and discuss them, respectively.
DRAFT: Do not cite or quote 1 December 1998
-------
2. DESIGN OF GROUND-TRUTHING ANALYSIS FOR NEW YORK
Personnel from the New York State Department of Environmental Conservation (NY DEC)
indicated an interest in providing assistance to EPA in this ground-truthing exercise. The NY DEC
provided EPA with a copy of the model Air Guide 1 (AG-1), and assisted in making the model
operational. AG-1 contains facility-specific data, such as stack heights, for New York facilities,
including TRI reporting facilities. AG-1 is used by NY DEC to verify facility compliance with air
quality standards (NY DEC, 1991; 1995). AG-1 is composed of two models: a simple model for
screening analyses, and a more complex model for refined analyses. The screening analysis produces
a single worst-case concentration for the facility, while the refined analysis can predict concentrations
at multiple locations chosen by the user. The refined analysis is far more comparable to the air model
component of the Indicators Model, and therefore was chosen for the ground-truthing analysis.
Both the Indicators Model and the more complex model in AG-1 use the same analytical
algorithm to predict air concentrations of pollutants emitted from industrial point sources. Both
models implement the long-term Gaussian plume algorithm included in EPA's Industrial Source
Complex (ISC) models (U.S. EPA, 1992a; 1995a, b). Because the two models were developed at
different times, they use different versions of ISCLT (AG-1 uses ISCLT2, while the Indicators Model
uses ISCLT3). However, the same algorithm is used to model dispersion from point sources in both
versions of ISCLT. Thus, identical results should be obtained when both models are used with the
same input data set. The major difference between ISCLT2 and ISCLT3 lies in the treatment of area
sources, for fugitive emissions. The algorithm for area sources was significantly improved in ISC3.
In this ground-truthing exercise, the results obtained from the Indicators Model are compared
to results obtained from a model which uses more facility-specific data. The results from the
Indicators Model are not being compared to air monitoring data because the ISC series of models
(versions 1,2, and 3) have already been validated. The EPA and others (e.g., Bowers and Anderson,
1981; Bowers et al., 1982; Heron et al. 1984; Moore et al., 1982) have repeatedly tested separate
components and features of the ISC models. Tests have included comparisons with experimental
(wind tunnel) and site-specific (air quality monitoring) data. These studies have validated
improvements in model algorithms and confirmed that the ISC models can adequately reproduce field
observations of pollutant concentrations. Currently, ISC3 is one of nine models recommended by
EPA for refined air quality analyses (U.S. EPA, 1995c). Recently, ISC3 was used as a benchmark
to which the performances of other models were compared (U.S. EPA, 1995d).
DRAFT: Do not cite or quote 2 December 1998
-------
2.1 SCOPE OF THE ANALYSIS
The overall objective of the ground-truthing exercise was to assess the degree to which results
from the Indicators Model differ from those of another state-of-the-art air model currently used for
regulatory purposes. Given that the Indicators Model uses a combination of facility-specific median
data, where available, and generic assumptions, while the AG-1 model uses almost all facility-specific
data, different air pollutant concentrations are predicted for emissions from the same facility. By
analyzing the differences in pollutant concentrations for a number of facilities, the degree to which
predictions differ between the two models was quantified.
Because many input variables affect model predictions, the tests conducted for this ground-
truthing analysis assessed the combined impact of those variables used in the air exposure pathway
of the Indicators Model. Uncertainty and sensitivity analyses would be needed to obtain a complete
perspective on the range of variability in model concentrations that occurs for alternative
combinations of input parameters. Such analyses were not included in this ground-truthing
comparison. Instead, results from a preliminary sensitivity analysis conducted using ISCLT3 were
reviewed to identify the relative impact of different input variables. In that analysis, a single input
variable was varied over a range of values while holding all other variables constant; the process was
repeated for all stack-specific variables (stack height, stack diameter, exit gas velocity, and exit gas
temperature). Relative impacts were measured in terms of the average air concentration over a grid
identical to that used by the Indicators Model. The results indicated that the pollutant concentrations
predicted by ISCLT3 are most sensitive to the stack height value used; exit gas velocity also has a
measurable, although smaller, impact on predicted concentrations. Both stack height and exit gas
velocity are negatively correlated with the average air concentration; that is, larger values of these
parameters will yield smaller concentrations, and vice-versa. More extensive tests conducted by the
NY DEC have reached similar conclusions (NY DEC, 1991).1
2.2 SAMPLING FRAMEWORK
This ground-truthing analysis compares air pollutant concentrations estimated by using a
combination of facility-specific (e.g., median stack height and median exit gas velocity) and generic
(e.g., stack diameter and exit gas temperature) air modeling parameters in the Indicators Model to
concentrations estimated using facility-specific data. Specifically, 24 test cases were constructed to
evaluate the impact of Indicators Model parameters for facilities with different stack heights,
geographic location, and chemical characteristics of emissions (see Table 1).
1 NY DEC quantified the impact of stack height on pollutant concentrations under different conditions,
including a range of downwind distances, varying building dimensions, and differing numbers of stacks (NY DEC,
1991).
DRAFT: Do not cite or quote 3 December 1998
-------
Test cases were designed to capture the variability in stack heights, because this input variable
has the largest impact on predicted air concentrations. The Indicators Model uses either the median
stack height of all stacks (regardless of the chemical emitted) for TRI facilities with this information
or an SIC code-based median stack height for facilities without stack data (Bouwes and Hassur,
1998). The latter is based on the median of stack heights for facilities in a particular 3-digit SIC code
(or in the 2-digit SIC code if the 3-digit SIC code is invalid. If no valid 2-digit SIC code is available,
the median of all stack heights in SIC codes 20 through 39 is used). Stack height data were obtained
from the AIRS Facility Subsystem (AFS) within the Aerometric Information Retrieval System
(AIRS), the National Emission Trends Database, and databases from three individual states
(California, New York, and Wisconsin). In the calculation of median stack height for facilities with
a particular SIC code, statistical analyses were conducted to determine whether heights for stacks not
emitting any TRI chemicals should be included. For some SIC codes, significant height differences
did not exist between stacks emitting TRI chemicals and stacks not emitting TRI chemicals. Thus,
in those test cases, all stack heights for all facilities in that SIC code were used to estimate the
median stack height for that SIC code. For other SIC codes, a significant height difference between
the two groups of stacks did exist, and only those stacks emitting TRI chemicals were used in the
calculation of a median stack height for that SIC code.
When running AG-1, NY DEC uses actual stack height data for those individual stacks
emitting chemicals of concern at a selected facility. The sampling framework for the ground-truthing
analysis was designed to evaluate in part the impact of using a facility-specific median stack height
in the Indicators Model versus using multiple stack-specific heights in the AG-1 model. Three
categories of facilities were represented: (1) TRI facilities with median stack heights less than seven
meters, (2) TRI facilities with median stack heights between seven meters and ten meters, and (3) TRI
facilities with median stack heights greater than ten meters. These categories reflect the distribution
of facility-specific median stack heights for TRI facilities in New York: approximately one-third of
these facilities are found in each of the stack height bins. Once the test cases were chosen for
analysis, the facility-specific median stack height was used in the Indicators Model runs and the actual
stack-specific heights were used in the AG-1 model runs. To evaluate the impact of using stack
heights based on SIC codes, a further comparison was made, using the stack heights based on each
facility's SIC code in the Indicators Model.
As previously indicated, the preliminary sensitivity analysis showed that exit gas velocity also
has a measurable impact on predicted concentrations. The Indicators Model uses either the median
exit gas velocity of all stacks (regardless of the chemical emitted) for TRI facilities with this
information or an SIC code-based median exit gas velocity for facilities without exit gas velocity data
(Bouwes and Hassur, 1998). The latter is based on the median of exit gas velocities for facilities in
a particular 3-digit SIC code (or in the 2-digit SIC code if the 3-digit SIC code is invalid. If no valid
2-digit SIC code is available, the median of all exit gas velocities in SIC codes 20 through 39 is used).
Exit gas velocity data were obtained from AFS within AIRS, the National Emission Trends Database,
DRAFT: Do not cite or quote 4 December 1998
-------
and databases from two individual states, New York and Wisconsin. The same statistical analyses
as described above for stack heights were conducted before a median exit gas velocity was calculated
for each SIC code. Again, the facility-specific median exit gas velocity was used in the Indicators
Model runs and the actual stack-specific exit gas velocities were used in the AG-1 model runs for one
comparison; a second comparison was made using exit gas velocities based on SIC codes.
Specific TRI facilities were selected from urban and rural areas covered by meteorological
stations in Albany, Buffalo, Rochester, and Syracuse.2 These four metropolitan areas were chosen
to determine if particular air modeling parameters have greater impacts in certain areas due to
possible interactive effects with different meteorological conditions. For each metropolitan area and
stack height bin, two facilities were selected: one to represent stacks emitting chemicals with decay
rates and the other to represent stacks emitting chemicals without decay rates. The distinction was
intended to reflect another difference between the Indicators Model and AG-1: the Indicators Model
incorporates chemical decay rates (based on photo-oxidation), while AG-1 does not. These decay
rates reduce the resultant air concentrations predicted by the Indicators Model.
An attempt was made to construct the sample of test cases by selecting one chemical with a
decay rate and one without a decay rate, as well as facilities that emitted both chemicals, to minimize
the variability across sites. However, these restrictions yielded an insufficient number of facilities for
analysis. The final set of 24 test cases reflects a compromise: a single chemical (toluene) with a
decay rate and four of the most commonly released chemicals without decay rates (mercury,
aluminum, lead, and nickel) for New York TRI facilities in the four locations. Four of the facilities
represented in the sample discharge both types of chemicals: Facility A (Albany), Facility G
(Syracuse), Facility Q (Rochester), and Facility S (Rochester). Although the information on these
facilities was used for the analysis of both chemicals with decay rates and those without decay rates,
each facility is considered to be two separate test cases because different sets of stacks are evaluated
by AG-1 and, therefore, results do not represent the effect of changing only chemical characteristics.
2.3 TESTING STRATEGY
To conduct this ground-truthing analysis, the ISCLT3 model (U.S. EPA, 1995a, b) was used
directly, rather than as implemented in the Indicators Model. Because of this choice, a three-way
model comparison was necessary. First, the Indicators Model and ISCLT3 were compared to verify
that the ISCLT3 algorithm was successfully incorporated into the Indicators Model. Second, AG-1
and ISCLT3 were compared to verify that they yielded the same results with identical inputs for point
sources. Although both models implement the same ISCLT point-source algorithm, this comparison
was necessary to test whether other assumptions were built into AG-1. Third, AG-1 and ISCLT3
2 "Urban" areas are defined in the Indicators Model as having populations greater than 119,070 people. In this
ground-truthing analysis, fifteen facilities are located in urban areas and five are in rural areas.
DRAFT: Do not cite or quote 5 December 1998
-------
were compared, with AG-1 using all available facility-specific data and ISCLT3 using the combination
of facility-specific data and generic assumptions used in the Indicators Model. This third test
evaluated how model predictions of pollutant air concentrations from point sources differ when
facility-specific data (e.g., building parameters, such as height and area dimensions, and stack
parameters, such as height, exit gas velocity, and temperature) are used as compared to median stack
height and exit gas velocity data and generic assumptions.
DRAFT: Do not cite or quote 6 December 1998
-------
3. PRELIMINARY TESTS
This section describes the first two model comparisons conducted prior to the actual
comparison of results from the Indicators Model and AG-1 model. First, EPA already conducted
several tests in the past that verified that the Indicators Model yielded results identical to those of the
ISCLT3 model when predicting air concentrations from point sources.
Second, tests were conducted to compare results from AG-1 and ISCLT3. These tests were
conducted with Facility A in Albany, for which all facility-specific data were available in the AG-1
database. A single chemical (mercury) was selected from all the TRI compounds emitted by this
facility. All input data from AG-1 were used as input to ISCLT3, and two tests were run, one for the
urban mode and one for the rural mode. In both tests perfect agreement was obtained between the
two models' predictions for all nodes in a 21-km by 21-km grid. In the Indicators Model, each node
is centered in a 1-km by 1-km cell, and the concentration at the node is assigned to that cell. The
facility is located in the center cell of the 441 cells, and no concentration is attributed to that cell. The
grid size is not finer because the Indicators Model assesses general population exposures, not risk to
a Most Exposed Individual (MEI).
Although one facility was used to test both the urban and rural modes, only one mode is used
for a given facility in the Indicators Model. If the total population in a 21 -km by 21 -km grid centered
at the facility is larger than 119,070, the urban mode is used. Different dispersion algorithms are used
for the rural and urban modes (U.S. EPA, 1995a, b), but for a given mode, the same algorithms are
used in both AG-1 and ISCLT3. The two models, however, make different assumptions about
building dimensions. When site-specific data are available, AG-1 calculates individual stack heights
as the sum of two variables: building height and stack height above structure. When site-specific data
are not available, AG-1 assumes that all building dimensions (height, width, and length) are equal to
the stack height; this assumption is intended to make the model more conservative. ISCLT3 makes
no specific dimension assumptions, and adopts zero building dimensions. By forcing ISCLT3 to
make the same assumptions about building dimensions as AG-1, perfect agreement was obtained
under both rural and urban modes. However, in the actual ground-truthing tests reported in the next
section, no such correction was made. Therefore, this difference in assumptions accounts for a
fraction of the total difference in air concentrations observed at each facility. Different concentrations
are predicted because the presence of a building produces higher concentrations near the source due
to building downwash. After downwash, there is less pollutant mass to be distributed further away
from the building, because the total pollutant mass being emitted into the air is the same regardless
of building dimensions. Thus, when all other inputs are the same, the Indicators Model will produce
slightly higher air pollutant concentrations further away from the source than AG-1 and lower
concentrations nearer the source. However, the differences in predicted concentrations are small for
the range of distances sampled by the computational grid used in the Indicators Model (1 to 14.8 km,
where 14.8 km is the diagonal distance from the source to the corner of the 21-km by 21-km grid).
DRAFT: Do not cite or quote 7 December 1998
-------
Typical maximum differences are on the order of one to two percent, and decrease to insignificant
levels with increasing distance from the source.
DRAFT: Do not cite or quote 8 December 1998
-------
4. MODEL COMPARISON: AG-1 VERSUS ISCLT3
As indicated in Section 2, ISCLT3 was used directly for this ground-truthing exercise. All
facility-specific median data and generic assumptions used in the Indicators Model were also used in
ISCLT3, to obtain the same model predictions that would be produced by the Indicators Model. In
the remainder of this section these results are referred to as the "Indicators Model results" for
convenience.
4.1 INPUT DATA
AG-1 and ISCLT3 share the same input parameters, but assign different values to them, as
summarized in Table 2. For stack diameter, exit temperature, and building dimensions, the Indicators
Model uses constant, generic values, whereas AG-1 uses facility-specific data (if available). In
addition, AG-1 computes concentrations from all individual stacks that emit a particular chemical,
while the Indicators Model treats all such emissions as emanating from a single stack at a central
location, with stack height equal to the median height of all stacks at the facility and exit gas velocity
equal to the median exit gas velocity from all stacks at the facility. For chemicals which may decay
through photodegradation, the Indicators Model uses a decay rate, whereas AG-1 assumes no
chemical decay occurs. Both models use comparable meteorological data, i.e., STability ARray
(STAR) data from local meteorological stations.3 For a given meteorological station, the Indicators
Model uses average conditions computed over many years (typically 25 years or more), while AG-1
uses one year's worth of data corresponding to the most recent year with valid STAR data. For
purposes of this ground-truthing exercise, both models used STAR data from AG-1.
The stack coordinates of the TRI facilities selected for the model comparison are listed in
Table 3. All coordinates are in meters, with values corresponding to the Universal Transverse
Mercator (UTM) coordinate system. Two sets of coordinates are listed, corresponding to the NY
DEC and national TRI databases. The national TRI database contains a single pair of coordinates
for each facility, while the NY DEC database contains stack-specific coordinates. The values listed
for the latter in Table 3 are the coordinates of the point located in the middle of all stacks that emit
the particular chemical selected for the model comparison. AG-1 centers the computational grid at
this middle point. Note that some of the TRI database and NY DEC coordinates included in Table 3
differ by hundreds or thousands of meters, which would cause the contaminant plumes to be mapped
3ISCLT uses as input meteorological data that have been summarized into joint frequencies of occurrence for
particular wind speed classes, wind direction sectors, and atmospheric stability categories. These STAR summaries
may include frequency distributions over a monthly, seasonal, or annual basis.
DRAFT: Do not cite or quote 9 December 1998
-------
in non-overlapping locations. Therefore, the single stack for the ISCLT3 runs was placed at the same
middle point that AG-1 uses to center the grid.4
Tables 4 to 8 display the input data used by each model for the following parameters: stack
height, exit gas velocity, stack diameter, exit temperature, and chemical emission rate. For stack
diameter and exit temperature, the Indicators Model has single default values (Table 2), while AG-1
uses stack-specific values. Because the AG-1 emissions data are from different years for different
stacks, reported releases from the TRI database could not be used. Instead, as indicated in Table 2,
for a given facility the sum of the emission rates of a particular chemical from all relevant stacks in
AG-1 was used as the chemical emission rate for that facility in the Indicators Model (ISCLT3).
Although AG-1 uses unique chemical emission-stack combinations, the mean and median stack
heights and exit gas velocities are presented in Tables 4 and 5 for purposes of comparison to ISCLT3
inputs. As shown in Tables 4 and 5, the number of stacks used in the calculation differ, as AG-1
mean and median values are based only on those stacks which emit the chemical being analyzed,
whereas mean and median values in ISCLT3 are based upon all stacks at the facility.
4.2 RESULTS
Three sets of Indicators Model runs were conducted to explore the impact of having facility-
specific median data or relying on assumptions when such data are not available. The first set uses
facility-specific median stacks heights and exit gas velocities, representing the case with most stack-
specific data. The second set uses facility-specific median stacks heights and a constant exit gas
velocity of 0.01 m/sec. The third set uses median stacks heights and exit gas velocities corresponding
to the 3-digit or 2-digit SIC code of the facility, representing the case with the least stack-specific
data. Results from the three sets of tests are described below.
Both the Indicators Model and AG-1 report pollutant concentrations on a discrete grid. The
Indicators Model uses a 21-cell by 21-cell grid composed of 1 km2 cells, with a total of 441 cells.
The same grid dimensions were chosen for the AG-1 model runs to compare results at the same
locations. Figure 1A displays the pollutant concentrations in each cell predicted by AG-1 for an
example facility, while Figure IB displays the concentrations predicted by the Indicators Model.
Figure 1C displays the ratio of concentrations predicted by each model for each cell (i.e., ISCLT3
concentration/AG-1 concentration); a ratio of one indicates perfect agreement between the Indicators
Model and AG-1. The arrays of results shown in these figures provide a wealth of information, but
they are not the most convenient means to analyze spatial patterns. Instead, concentrations can be
displayed as a pollutant concentration plume with the aid of a contour plot. Figures 2A and 2B
display contour plots of the pollutant plumes predicted by each model for the example facility. Figure
4 In the Indicators Model, the facility stack is centered in the model cell that contains the facility coordinates
from the national TRI database.
DRAFT: Do not cite or quote 10 December 1998
-------
2C displays a contour plot of the concentration ratios shown in Figure 1C. Figure 2C reveals that
concentration ratios in about 20 cells around the stack range in value from 0.6 to 0.9; concentration
ratios in all other cells located further away from the stack are between 0.9 and 1.0.
Without reference to the location of individual cells, a histogram of all cell ratios provides a
more compact way of comparing plumes and illustrates the variability within and among test cases.
Figures 3 to 6 display such histograms for all 24 test cases, individually and averaged by metropolitan
area. While some of the histograms (e.g., test case 3 in Albany) are narrowly clustered around a
single value (usually one), others display more dispersion (e.g., test case 1 in Rochester), with the
maximum value for any single cell ratio being 3.1 (for test case 4 in Rochester). The histograms in
Figures 3 to 6 show that the average concentrations calculated by the Indicators Model for an
individual facility may differ from those calculated by AG-1 by up to 48 percent, with the largest
deviation corresponding to test case 4 in Albany (average concentrations are calculated over the 440
cells surrounding each facility).
In addition to the contour plots and histograms, another type of plot was developed to
examine the variability of model results with distance from the source. Because the computational
grid used by the Indicators Model is made up of square cells surrounding the source, a surrogate
measure was used to approximate the radial distance from the source. The grid can be visualized as
being made up of concentric square rings located around the central cell containing the source; in
a 21-km by 21-km grid, there are ten such rings, with ring one being closest to the source and ring
ten being the outermost ring. The ring number serves as a surrogate measure of distance in
kilometers from the source. For each of the ten concentric square rings, an average concentration
ratio was calculated; because of averaging effects, these concentration ratios display a narrower range
of values than the variations depicted by the histograms in Figures 3 to 6. Figures 7 to 10 display the
average concentration ratios over concentric square rings for individual test cases, grouped by
metropolitan area. The shapes of the plots for test cases in the same metropolitan area are somewhat
similar, but not enough to define distinct patterns for each metropolitan area. Instead, two patterns
are apparent for individual test cases: concentration ratios decrease with distance when there is a
maximum at ring one, or increase with distance when there is a minimum at ring one. For the second
ring and further, ratios for individual test cases are within ten percent of unity for Albany, and within
about 20 percent of unity for Buffalo, Rochester, and Syracuse, except for two test cases discussed
below. Within the first ring, ratios for individual test cases are within 35 percent of unity, except for
the two test cases discussed below.
In two of the cities there is a single curve that displays consistently higher concentrations for
all rings: test case 4 in Albany (mercury) and test case 4 in Rochester (nickel). These same test cases
can be identified using the histograms in Figures 3 and 5. Inspection of Table 4 reveals that test case
4 in Albany and test case 4 in Rochester share a common characteristic: the facility-specific median
stack height used in the Indicators Model i s significantly shorter than the corresponding median height
DRAFT: Do not cite or quote 11 December 1998
-------
of the stacks that actually emit the given chemical (although AG-1 uses individual stack heights, their
median was computed to allow a simple comparison; other measures, such as the emission-weighted
mean or median, could be used as well). The differences are 26 meters (m) and 6 m for the Albany
and Rochester test cases, respectively. Calculations using the shorter stack height from the Indicators
Model result in higher concentrations predicted by the Indicators Model, and therefore, higher
concentration ratios. Test case 4 in Albany, which has the largest discrepancy between median stack
heights, produces the largest ratios over the entire grid in the 24 test cases. These results are
consistent with previous sensitivity analyses of the influence of stack heights on pollutant
concentrations. However, the tests conducted for this ground-truthing analysis were not designed
to isolate the influence of a single variable. Hence, the range of variability in calculated pollutant
concentrations reflects the combined effect of all input variables that take different values in each
model (this includes not only all stack parameter data, but also building dimensions and treatment of
chemical decay).
In interpreting the average concentration ratios over concentric rings, it is important to note
that the inner rings have fewer cells (e.g., 8 cells for ring 1 of an individual test case), as compared
to outer rings (e.g., 80 cells for ring 10 of an individual test case). Therefore, the statistics for the
inner rings are more sensitive to single high values. In contrast, the ratio statistics for the outer rings
are more stable and seem to approach a constant value, typically very close to unity. In subsequent
figures similar "ring" curves are used to examine the variability of concentration ratios by stack height
bin, chemical, and metropolitan area.
Figure 11 displays the average concentration ratio computed for each ring for the three stack
height bins. Agreement between the Indicators Model and AG-1 seems to be independent of stack
height bin, because most ratios are within five percent of unity; even within the two innermost rings,
ratios are within fifteen percent of unity.
Figure 12 compares the ring statistics grouped by chemical type (each group has twelve test
cases). The ratios for the chemical with a decay rate are consistently lower than those for chemicals
without a decay rate, which is expected, given that the Indicators Model accounts for decay rates,
while AG-1 does not. Figure 12 indicates that ratios for the chemical with a decay rate are about five
percent lower than unity on average, while those for the chemical without a decay rate are about two
percent higher than unity. However, this figure should be taken as indicative only. Evaluating the
effect of this individual variable would require running each test case with both chemical types,
holding all other parameters constant.
Figure 13 shows the average ring statistics for each metropolitan area (six test cases each,
averaged over both chemical types). Except for Syracuse, the ratios for all rings in the four curves
shown in Figure 13 are within ten percent of unity. The concentration ratios in the first ring of
Syracuse are within 17 percent of unity.
DRAFT: Do not cite or quote 12 December 1998
-------
Table 9 contains similar information, but also provides the standard deviations, minimum
values, and maximum values of the concentration ratio for each metropolitan area, by chemical
characteristic and by stack height bin. The mean concentration ratio for the entire sample is 0.984,
indicating that on average, the predictions of the Indicators Model are virtually the same as those of
AG-1. Subsample average ratios (e.g., by metropolitan area, chemical characteristic, and stack height
bin), shown in Table 9, vary between 0.93 5 and 1.05, again representing very good agreement. Table
10 contains the statistics corresponding to the concentration ratios by ring for all locations together
and by metropolitan area. A complementary view is provided by the histograms in Figures 3 to 6.
These figures show that the average histograms of concentration ratios for each metropolitan area
have most cells clustered around one, with the highest frequency corresponding to ratios between
0.95 and 1.05.
4.2.1 Impact of Exit Gas Velocity Assumptions
When this ground-truthing exercise was initiated, the corresponding version of the Indicators
Model assumed a constant exit gas velocity (0.01 m/s) for all stacks. Given that the preliminary
sensitivity analysis indicated that exit gas velocity had a measurable impact on predicted
concentrations, and that the default value of 0.01 m/s was three orders-of-magnitude smaller than
most available data on exit gas velocities, the way in which exit gas velocities are treated in the
Indicators Model was changed (Bouwes and Hassur, 1998). Tables 11 and 12 contain a summary
of results for the constant exit gas velocity case, in the same format as Tables 9 and 10. Although
each single statistic in Tables 11 and 12 can be compared to its counterpart in Tables 9 and 10, only
the mean concentration ratio calculated over the whole sample (all rings, all metropolitan areas) is
analyzed here. The mean ratio in Tables 11 and 12 equals 0.980, approximately equivalent to the
mean ratio (0.984) shown in Tables 9 and 10; the corresponding standard deviations are virtually the
same (0.136 and 0.134, respectively). Although these statistics are very similar, EPA believes that
it is more defensible to use available data on exit gas velocities and to treat the data in the same
manner that stack height data are treated than to use a default value that is three orders-of-magnitude
smaller than most available data.
4.2.2 Impact of SIC Code-based Stack Height and Exit Gas Velocity Assumptions
The results presented so far correspond to the case in which facility-specific data are available
to calculate median stack heights and exit gas velocities. However, only a small fraction of facilities
nationwide (about ten percent) have such data in the Indicators Model database. For the vast
majority of the facilities, the Indicators Model uses the median stack height and exit gas velocity
corresponding to the 3-digit SIC code of the facility. Table 13 contains the median stack heights and
exit gas velocities corresponding to the 3-digit SIC codes of the 24 facilities in the sample, along with
the facility-specific median values (used in the previous comparison) and the chemical-specific median
values (which summarize the stack by stack emissions calculated by AG-1). A brief inspection of
DRAFT: Do not cite or quote 13 December 1998
-------
Table 14 reveals that stack heights for individual facilities may differ by as much as a factor of seven.
To test the performance of the Indicators Model when data based on SIC codes are used, the
3-digit SIC code median values in Table 13 were used in ISCLT3 and the results were compared to
AG-1. Results are displayed in Figures 14 through 24 and Tables 14 and 15. Because the figures
and tables contain results parallel to those previously discussed, a side-by-side comparison is possible.
For example, the histograms in Figures 14 to 17 show a summary of cell-by-cell concentration ratios
similar to those in Figures 3 to 6. Overall, the histograms in Figures 14 to 17 show more scatter than
those in Figures 3 to 6. This scatter is consistent with the larger differences in input parameters
(stack heights) for some facilities, as shown in Table 13. An inspection of the histograms in Figures
14 to 17 shows that the average concentrations calculated by the Indicators Model for an individual
facility may differ from those calculated by AG-1 by less than 35 percent (the largest average
deviations correspond to test case 1 in Albany and test case 4 in Rochester). The maximum value for
any single cell ratio is 3.4 (for test case 4 in Rochester).
The summary statistics in Tables 14 and 15 can be readily compared to those in Tables 11
and 12 (and Tables 9 and 10). The mean concentration ratio calculated over the entire sample (all
rings, all facilities) equals 0.936 (Tables 14 and 15), somewhat lower than the mean ratio (0.984)
obtained when using facility-specific median stack heights and exit gas velocities (Tables 9 and 10).
This result is consistent with the inputs shown in Table 13: given that a majority of 3-digit SIC
median stack heights are larger than the corresponding facility-specific median values, the Indicators
Model predicts smaller concentrations and therefore the concentration ratios are lower on average.
(This result in turn is consistent with the findings from sensitivity analyses already discussed.) The
standard deviation of the concentration ratio (0.131) is approximately equivalent to the previous one
(0.134).
A majority of the 24 test cases have 3-digit SIC code median values significantly higher than
the corresponding facility-specific median values. On a nationwide basis, the Indicators Model could
be expected to sometimes overpredict and sometimes underpredict, depending on the discrepancies
between actual and assumed parameter values. To assess the range of discrepancies on a larger
sample, parameter values for all facilities with site-specific data were compared to SIC code based
values. The comparison was performed by subtracting facility-specific median values from SIC code
based median values, for stack heights (1504 facilities) and exit gas velocities (1063 facilities). The
results are displayed in Figures 25 and 26 for stack heights and exit gas velocities, respectively. SIC
code based median stack heights range from 69 m less to 29 m more than the facility-specific median
stack heights. The 95th and 5th percentiles are 18m less and 7.0 m more, respectively. SIC code
based median exit gas velocities range from 295 m/s less to 17 m/s more than the facility-specific
median exit gas velocities. The 95th and 5th percentiles are 49 m/s less and 7.1 m/s more,
respectively. Ground-truthing analyses were not repeated for these additional facilities, although
DRAFT: Do not cite or quote 14 December 1998
-------
previous results show that using median values based on SIC codes yields a wider range of
concentration ratios (subsample statistics in Table 14 vary between 0.871 and 1.00, a range only
slightly wider than the corresponding ranges in Tables 9 and 11). Because the concentration ratio
statistics (overall average and standard deviation) are reasonably close to the values obtained when
using facility-specific median values, it is concluded that the Indicators Model performs very well
when using 3-digit SIC code median values for stack heights and exit gas velocities.
4.3 FUGITIVE EMISSIONS ANALYSIS
Fugitive releases, which are modeled as area sources, are a significant fraction of the total
reported air emissions of TRI chemicals. The ISCLT model used by AG-1 and the Indicators Model
can predict fugitive emissions from area sources as well as stack emissions from point sources. Thus,
it is theoretically possible to conduct a ground-truthing exercise for fugitive emissions to test the area
source component of the Indicators Model.
A ground-truthing exercise for fugitive emissions using AG-1, however, would not be very
useful. Recall that AG-1 uses ISCLT2, and the Indicators Model uses ISCLT3; the area source
algorithm in ISCLT3 has been improved over that used in ISCLT2 to calculate pollutant
concentrations from fugitive emissions (U.S. EPA, 1992a, 1995b). Therefore, predictions made by
the two models will differ even when identical input data are used. In addition, AG-1 and the
Indicators Model use different data to characterize the dimensions of area sources. While AG-1 uses
site-specific data for the surface area and height of an area source, the Indicators Model uses default
values. Hence, comparing the fugitive emission component of AG-1 and the Indicators Model would
require separate evaluations of the differences due to model algorithms and due to input data.
The essential difference in the area source algorithms used in ISC2 and ISC3 can be
summarized as follows. Both algorithms are based on integrations of the Gaussian plume formula
used for point sources, but the integration is carried out over different area geometries to describe
the shape of an actual area source. In ISC2 the integration is carried out over a crosswind line, and
calculations assume square area sources. Actual area sources may have irregular shapes; they can
be represented with many small squares that approximately overlay the actual area. In ISC3 the
integration is carried out over a rectangular area, and calculations allow arbitrary dimensions for each
rectangle. By using rectangles of variable dimensions (aspect ratios can be as high as ten to one), area
sources of irregular shape can be represented more accurately than in ISC2. (Note that these
integrations cover the area source itself and therefore are independent of the computational grid used
in the Indicators Model to estimate pollutant concentrations in square cells.) The revised area source
algorithm included in ISC3 has been thoroughly evaluated and its predictions compared to wind
tunnel data (U.S. EPA, 1992b, c, d). Because the computational algorithms are different, ISC2 and
ISC3 will predict different concentrations for an identical area source, square or otherwise. However,
the differences between predictions of ISC2 and ISC3 are more significant close to the source. ISC2
DRAFT: Do not cite or quote 15 December 1998
-------
(and therefore AG-1) can underestimate concentrations close to the source by as much as a factor
of three (NY DEC, 1995).
If the area source algorithms were identical in ISCLT2 and ISCLT3, as the point source
algorithms are, a ground-truthing analysis would compare the results obtained from site-specific data
on area source sizes with results obtained using default assumptions. The Indicators Model uses
default values for the dimensions of all area sources: a surface area of 10 m2 and a height of 3 m. The
AG-1 Guidelines (NY DEC, 1991) recommend using a surface area of 84 m2 in the absence of site-
specific data; no default value is recommended for the height of the area source.
Sensitivity analyses conducted on ISCLT2 demonstrate that for an arbitrary area source size,
there is a distance from the source at which the concentrations approach those of a point source (NY
DEC, 1991). As would be intuitively expected, this distance decreases for smaller area sources. For
an area source of the size used in the Indicators Model (10 m2), this distance is about 50 m; for an
area source of the size recommended in the AG-1 Guidelines (84 m2), this distance is about 400 m
(NY DEC, 1991). Therefore, at the distances sampled by the Indicators Model grid (one kilometer
and larger), both models yield practically identical results (NY DEC, 1991). These results from
ISCLT2 only reflect the impact due to different area sizes, not the impact of different area source
heights. A similar sensitivity analysis was conducted using the ISCLT3 model to evaluate the impact
of both area source size (10 m2 and 84 m2) and height (3 m and 0 m). From this analysis it was
determined that the distances from the source at which the concentrations approach those of a point
source are also less than one kilometer. Thus, a separate ground-truthing exercise for area sources
would be redundant with the analysis of point sources already conducted.
DRAFT: Do not cite or quote 16 December 1998
-------
5. PERSPECTIVE ON FINDINGS
This ground-truthing analysis shows that pollutant concentrations predicted by the Indicators
Model are in excellent agreement with those predicted by AG-1, even though the models use different
input data (median and generic values versus stack-specific data) and assumptions (e.g., building
dimensions and treatment of chemical decay). Although the range of concentration ratios for
individual cells is 0.23 to 3.4, the vast majority of individual cells in all 24 test cases have
concentration ratios that are close to unity (within five percent of unity when facility-specific median
parameters are used, and within ten percent of unity when SIC code based parameters are used).
Because any one individual cell contributes very little to the impact of the facility as a whole, average
concentration ratios over concentric rings around the stack were analyzed. For the majority of the
test cases in the sample, average concentrations within each ring predicted by the two models are
within 20 percent of each other. In the rings closest to the source, in which the largest discrepancies
occur, average concentrations within each ring predicted by the two models are within a factor of 0.5
to two of each other, even when SIC code based parameters are used. Thus, although the Indicators
Model is not designed as a substitute for more comprehensive, site-specific risk assessments, the
results of this ground-truthing analysis indicate that the air exposure pathway of the Indicators Model
provides very good estimates of air pollutant concentrations at the facility-specific level.
Not surprisingly, this ground-truthing analysis showed that the Indicators Model performs
best when facility-specific median stack heights and exit gas velocities are available, rather than when
median stack heights and exit gas velocities based on SIC codes are used. When facility-specific
median values were used, results indicated a very close agreement between the Indicators Model and
AG-1: average concentrations calculated over the approximately 10,560 cell concentrations estimated
by each model for all 24 test cases differ by less than two percent, with a standard deviation of
approximately 13 percent. Even when parameters based on SIC codes are used, the results of the
Indicators Model compare very well to those of AG-1: average concentrations computed by both
models for the 24 test cases differ by approximately six percent, with a standard deviation of
approximately 13 percent.
Average ring concentrations predicted by the two models are within a factor of 0.5 to two of
each other near the facility; these concentration ratios become smaller and often converge within a
narrow band around unity with increasing distance from the source. Only two of the 24 test cases
departed from this general pattern when using facility-specific median parameter values. As
previously mentioned, such disagreements are probably due to the markedly different stack heights
used by each model in these two test cases. Similar discrepancies are expected to occur in a fraction
of the cases nationwide, because the facility-specific stack statistics (e.g., median) may not always
accurately approximate the corresponding statistics for the subset of stacks that emit a particular
chemical. This may happen regardless of whether facility-specific or SIC code based parameters are
used. The sample is too small to allow precise inferences of how often this may occur, but the fact
DRAFT: Do not cite or quote 17 December 1998
-------
that such discrepancies occurred only twice in the 24-case sample gives some indication that this
situation may occur in only a small fraction of cases on a nationwide basis as well.
5.1 CALCULATION OF INDICATOR ELEMENTS
Although the ground-truthing exercise has affirmed the accuracy of the pollutant
concentrations predicted by the Indicators Model, pollutant concentration is only part of the
calculation of an Indicator Element, which can be used to rank facilities. Therefore, it is imperative
to ascertain the contribution of pollutant concentration, as well as other components, to the
estimation of Indicator Elements. An Indicator Element is the product of three components: the
surrogate dose, which is based on pollutant concentration and exposure assumptions; the toxicity
weight for the chemical of interest; and, the exposed population. For each of the 440 cells
surrounding a TRI facility, cell-level products, called Indicator Sub-Elements, are calculated and then
added to yield the Indicator Element. Consideration of these other Indicator Element components
while taking into account the increased predictive accuracy of the ISCLT3 model at greater distances
from a facility will aid the analyst when interpreting Indicators Model results at the facility-level.
5.1.1 Toxicity
Toxicity weights are chemical and pathway-specific; each facility emitting a given chemical
will receive that same pathway-specific weighting factor for that chemical release. Weights range
from 0.1 to 1,000,000 for carcinogens and from 0.001 to 100,000 for non-carcinogens. The impact
of toxicity weights on Indicator Elements will be irrelevant only when comparing facilities emitting
the same chemical. In all other cases they may account for a significant fraction of the total Indicator
Elements value calculated for a facility.
5.1.2 Surrogate Dose
The air pollutant concentration estimated by the Indicators Model is converted to a surrogate
dose using standard assumptions for body weight and inhalation rate. These exposure assumptions
are the same from facility to facility and will not influence the ranking of facilities. Thus, the
surrogate dose can be viewed as the ISCLT3 concentration multiplied by a constant. As discussed
above, the results of this ground-truthing exercise demonstrated that the methods employed by the
Indicators Model to estimate facility stack heights and exit gas velocities result in pollutant
concentrations that compare very favorably to those of the AG-1 model, which uses much more
facility-specific data. Generally, the results of the two models converged at approximately 2
kilometers from the facility, resulting in only a small percentage of the 1-km by 1-km cells being
prone to over or underestimation of pollutant concentrations by an appreciable amount. These cells
with an appreciable amount of over or underestimation are usually located in the immediate vicinity
of the source. While pollutant concentrations are also highest near the source, one cannot conclude
DRAFT: Do not cite or quote 18 December 1998
-------
that these cells have the greatest impact on Indicator Elements without considering the impact of
population distribution.
5.1.3 Population
In addition to pollutant concentration, population is the other component of the Indicator
Element that is of interest for this ground-truthing exercise. Unlike exposure assumptions and
toxicity weight, which are applied consistently across all cells surrounding a facility, population is not
distributed evenly around a facility. Generally speaking, it would be ideal if population was
distributed at distances from the facility where the correspondence between ISCLT3 and AG-1
concentration estimates was nearly identical. Then the resulting facility rankings would be a fair
representation of facilities' relative risk. If the population was concentrated primarily within 2 km
of a facility, the resultant relative-risk rankings would be subj ect to greater error because the potential
for discrepancies in estimated pollutant concentrations is higher nearer to the facility.
To consider this issue, revisit Figures 18 through 21, which show the concentration ratios
using SIC code based parameters for the 24 test cases for the four metropolitan areas in New York
State. Generally, concentration ratios become relatively constant at approximately 2 km. Within 1
km the ring-average estimates of the concentration ratios for the 24 test cases range from 50 percent
below unity to almost 80 percent above unity. As seen in Table 15, the largest concentration ratio
for a single cell of the 192 cells composing the 1 km rings of these 24 test cases (8 cells x 24 sites)
was 3.4; the average of these 192 concentration ratios was 0.89.
To calculate an Indicator Element, it is necessary to multiply pollutant concentration in each
cell by the number of people living in each cell. Therefore, population distribution in concentric rings
around each facility was examined to see whether higher pollutant concentrations closer to the facility
were counterbalanced by lower populations closer to the facility. The number of people living in each
of the 440 cells surrounding the 24 facilities was obtained from the Indicators Model (AG-1 does not
have a population database); these numbers were then added over all cells in a given ring for a given
facility. The resulting population distributions do not display a consistent pattern, but rather vary
significantly from facility to facility. While some facilities have the majority of the population living
in rings 1 to 3, many facilities have increasing numbers of people living at greater distances. There
is also significant variability among metropolitan areas: in Albany, most people live relatively far
away from TRI facilities, while in Buffalo a high percentage of people live close to TRI facilities. In
an attempt to obtain a national perspective of this, a nationwide distribution of exposure events, i.e.,
persons impacted by multiple TRI facilities with non-zero air releases, was also analyzed. Table 16
presents the exposure events within specific "distance rings" of TRI facilities reporting air releases.
The values shown in this table are derived by assigning each person in the U.S. to each TRI facility
located within a specified distance; this procedure allows a person to be counted multiple times, as
is done in the Indicators Model, depending on how many TRI facilities potentially impact them.
DRAFT: Do not cite or quote 19 December 1998
-------
Thus, the total exceeds the U.S. population, because of individuals experiencing multiple exposures.
Although approximately 28 percent of the U.S. population resides within 2 km of TRI reporting
facilities, Table 16 shows that only five percent of all exposure events occur within 2 km.
When a large percentage of the population lives close to a TRI facility and when significant
discrepancies exist between the AG-1 and ISCLT3 predictions of pollutant concentrations near that
facility, the generated Indicator Elements could conceivably influence relative rankings of facilities.
In those instances where significant discrepancies exist between the A G-l andlSCL T3 concentration
predictions close to the facility but only a small percentage of the population live close to the
facility, the impacts on the Indicator Elements and the associated facility rankings will be negligible.
5.2 COMPARISON OF INDICATOR SUB-ELEMENTS' CONTRIBUTIONS BY RING
As described above, Indicator Elements are the sum of Indicator Sub-Elements calculated for
each of the 440 cells surrounding a TRI facility. To investigate the relative contribution of cell rings
to the total Indicator Element value, Indicator Sub-Elements were calculated for each ring around
each facility by multiplying just the population and the pollutant concentration in each cell, and adding
the products over all cells in a ring. (These results were not multiplied by toxicity because the focus
was only on analyzing a single pollutant in a given case.) The percent contributions of each ring to
a facility's Indicator Element are displayed in Figures 27 to 30 (one figure per metropolitan area),
along with the corresponding concentration ratio (ISCLT3/AG1) distributions by ring (these
distributions are identical to those shown in Figures 7 to 10).
Inspection of Figures 27 to 30 reveals the absence of a typical profile. In fact, the distribution
of the percent contribution by ring varies widely, as a consequence of the cell-by-cell combination of
population and pollutant concentrations. While there are test cases where the largest contribution
to a facility's Indicator Element comes from the first few rings (e.g., test case 1 in Syracuse), the
converse is true in other test cases (e.g., test case 1 in Rochester). These two test cases illustrate the
correlation between the distributions of population and Indicator Sub-Elements, and help visualize
the impact that discrepancies in concentration estimates (measured by concentration ratios) may have
on Indicator Elements. When there is a high population density near the facility, discrepancies in
concentration estimates can translate into discrepancies of similar magnitude in Indicator Elements.
In the worst case, the same factor of 0.5 to two that bounds discrepancies in pollutant concentrations
will apply to Indicator Elements as well. This case is exemplified by case 4 in Albany, where
concentration discrepancies in excess of 40 percent occur for all rings, and therefore the Indicator
Element value is also 40 percent overestimated. This case was previously identified as unique,
because of significant differences in median stack height input parameters. When a small percentage
of the population lives near the facility, discrepancies in concentration estimates in the first few rings
will have a much smaller impact on the total Indicator Element value. An extreme case is exemplified
by case 4 in Rochester (Figure 29); although the concentration ratio indicates discrepancies between
DRAFT: Do not cite or quote 20 December 1998
-------
30 and 60 percent for the first two rings, these discrepancies do not impact the Indicator Element
because there is no population living in the first two rings. Correspondingly, in those instances where
concentrations are correctly estimated, so will be the Indicator Elements, regardless of population
distribution.
As with pollutant concentration analyses, these conclusions cannot necessarily be extrapolated
to the U.S. as a whole. This sample reveals the wide variability in the distributions of Indicator Sub-
Elements and the significant impact on Indicator Sub-Elements that results from the particular
population distribution around a facility (although higher concentrations occur close to the source,
their impact on the Indicator Sub-Elements is greatly dependent on the size of the population living
in that area). Because of the wide variability observed from test case to test case, the Indicators
Model needs to be employed to capture the unique population distribution around each modeled
facility to ensure proper treatment of population and exposure.
5.3 FACILITY RANKINGS BASED ON INDICATOR ELEMENTS
The objective of the Indicators Model is to perform relative rankings of risk-related impacts.
To evaluate the use of different assumptions concerning stack heights and exit gas velocities, a
ranking exercise was performed on the 24 New York test cases. Facilities were ranked by each set
of available concentration estimates, generated by AG-1, by ISCLT3 with facility-specific median
stack heights and exit gas velocities, and by ISCLT3 with SIC code-based median stack heights and
exit gas velocities. Using the Indicator Elements calculated above, facilities were ranked in two
groups, those emitting chemicals that decay (toluene) and those emitting chemicals which do not
decay (aluminum, mercury, nickel, or lead). Note that because toxicity weights for individual
chemicals are not included in the above Indicator Elements, it is possible to group and rank all
facilities emitting chemicals which do not have decay rates, because the dispersion of inorganic
chemicals is modeled without any chemical-specific data (i.e., for a given facility, a pound of lead
released to the air is predicted to undergo the exact same dispersion as a pound of aluminum). The
two sets of rankings are listed in Tables 17 and 18, one for the chemical with decay and one for the
chemicals without, respectively.
Inspection of Tables 17 and 18 reveals that the rankings corresponding to different input
parameters are virtually identical for both categories of chemicals. The only exception is the rankings
of facilities F and Q. Facilities F and Q were assigned the same rankings (3 and 2, respectively) when
using ISCLT3 with both sets of input parameters, but were assigned slightly different rankings (2 and
3, respectively) when using AG-1. Indicator Element values for facility F are 2633 when facility-
specific parameters are used, 2729 when SIC code-based parameters are used, and 3226 when using
AG-1. Indicator Element values for facility Q are 2736 when facility-specific parameters are used,
2919 when SIC code-based parameters are used, and 3097 when using AG-1. In all three cases,
Indicator Elements values for facility Q are very close (within four percent, seven percent, and four
DRAFT: Do not cite or quote 21 December 1998
-------
percent, respectively) of the values corresponding to facility F. This suggests that relative rankings
depend not only on the Indicator Element values of a given facility, but also upon the corresponding
values of facilities with similar Indicator Element values. Differences in rankings may not be
meaningful when the corresponding Indicator Elements are very close in magnitude.
DRAFT: Do not cite or quote 22 December 1998
-------
6. CONCLUSION
This comparison of the Indicators Model to the AG-1 model was designed to measure
whether the Indicators Model yields air pollutant concentrations comparable to an air dispersion
model (AG-1) currently in use by a state agency, and to give an indication of the discrepancies in
predictions. The air pollutant concentrations calculated by the Indicators Model are based on a
combination of median and generic data and assumptions, whereas the AG-1 model relies on a greater
variety of facility- and stack-specific data. The differences in pollutant concentrations predicted by
both models were analyzed for 24 test cases in New York. The results obtained demonstrate that
predictions of pollutant concentrations are not only comparable, but are extremely close, even though
key input data to the two models are not the same. Average ratios computed over the 24 test cases
were within two percent of unity (with a standard deviation of 13 percent) when using facility-specific
parameters, and within six percent of unity (with a standard deviation of 13 percent) when using SIC
code-based parameters. The accuracy of concentration estimates close to a facility is usually less than
the accuracy observed further away from the facility, but the Indicators Model does not seem to
consistently overpredict or underpredict pollutant concentrations.
The impact of population distributions around TRI facilities on the Indicator Element was also
examined. Population around a TRI facility can have a significant impact on Indicator Element
values, depending on the population size and distribution relative to the predicted pollutant
concentrations and on the accuracy of the pollutant concentration estimates. The impact of
population on the accuracy of the Indicator Element depends on the cell-by-cell combination of
population and pollutant concentrations. Indicator Element values of lesser accuracy result from a
combination of less accurate concentration estimates near the facility and a majority of the population
living near the facility. When the concentration estimates are accurate, so are the Indicator Elements,
regardless of population distribution. When a small percentage of the population lives near the
facility, discrepancies in concentration estimates near the facility will have only a small impact on the
Indicator Element value. Thus, the Indicators Model needs to be employed to capture the unique
population distribution around each modeled facility to ensure proper treatment of population and
exposure.
Indicator Elements were used to rank the facilities that correspond to the 24 test cases in New
York. Facilities were ranked using each set of available concentration estimates: AG-1, ISCLT3 with
facility-specific median stack heights and exit gas velocities, and ISCLT3 with SIC code-based
median stack heights and exit gas velocities. Separate rankings were obtained for facilities emitting
chemicals that decay and those emitting chemicals which do not decay. With the exception of one
facility, the rankings corresponding to different input parameters were identical for both categories
DRAFT: Do not cite or quote 23 December 1998
-------
of chemicals, for all three sets of input parameters. This finding supports the use of the Indicators
Model to develop relative rankings of TRI facilities based on their risk-related impacts.
DRAFT: Do not cite or quote 24 December 1998
-------
7. REFERENCES
Bouwes, Sr., N.W., and S.M. Hassur. 1998. Estimates of Stack Heights and Exit Gas Velocities for
TRI Facilities. U.S. EPA, Office of Pollution Prevention and Toxics, Economics, Exposure,
and Technology Division. October.
Bowers, J. F., and A. J. Anderson, 1981. An Evaluation Study for the Industrial Source Complex
(ISC) Dispersion Model. EPA-450/4-81-002.
Bowers, J. F., A. J. Anderson, and W. R. Hargraves, 1982. Tests of the Industrial Source Complex
(ISC) Dispersion Model at the Armco, Middletown, Ohio Steel Mill. EPA-450/4-82-006.
Heron, T. M., J. F. Kelly, and P. G. Haataja, 1984. Validation of the Industrial Source Complex
Dispersion Model in a Rural Setting. Journal of the Air Pollution Control Association, 31 (4),
365-369.
Moore, G. E., T. E. Stoeckenius, and D. A. Stewart, 1982. A Survey of Statistical Measures of
Model Performance and Accuracy for Several Air Quality Models. EPA-450/4-83-001.
N. Y. DEC, 1991. Draft. New York State Air Guide-1. Guidelines for the Control of Toxic Ambient
Air Contaminants. New York Dept. of Environmental Conservation.
N.Y. DEC, 1995. User's Guides for the Air Guide-1 Software program and USEPA's SCREENS
model. Guidelines for the Control of Toxic Ambient Air Contaminants. New York Dept. of
Environmental Conservation.
U.S. EPA, 1992a. User's Guide for the Industrial Source Complex (ISC2) Dispersion Models.
Volume II, Description of Model Algorithms. EPA-450/4-92-008b.
U.S. EPA, 1992b. Comparison of a Revised Area Source Algorithm for the Industrial Source
Complex Short Term Model and Wind Tunnel Data. EPA-454/R-92-014.
U. S. EPA, 1992c. Sensitivity Analysis of a Revised Area Source Algorithm for the Industrial Source
Complex Short Term Model. EPA-454/R-92-015.
U.S. EPA, 1992d. Development and Evaluation of a Revised Area Source Algorithm for the
Industrial Source Complex Long Term Model. EPA-454/R-92-016.
U.S. EPA, 1995a. User's Guide for the Industrial Source Complex (ISC3) Dispersion Models.
Volume I, User Instructions. EPA-454/B-95-003a.
U.S. EPA, 1995b. User's Guide for the Industrial Source Complex (ISC3) Dispersion Models.
Volume II, Description of Model Algorithms. EPA-454/B-95-003b.
DRAFT: Do not cite or quote 25 December 1998
-------
U.S. EPA, 1995c. Guideline on Air Quality Modeling (Revised) and Supplements. EPA-450/2-78-
027R et seq., published as Appendix W to 40 CFR Part 51.
U. S. EPA, 1995d. Testing of Meteorological and Dispersion Models for Use in Regional Air Quality
Modeling. EPA-454/R-95-005.
DRAFT: Do not cite or quote 26 December 1998
-------
TABLES
DRAFT: Do not cite or quote 27 December 1998
-------
TABLE 1
Ground-Truthing Test Cases
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
0
P
Q
R
S
Q
T
S
Indicators
Model
Median
Stack
Height (m)
10.06
9.45
1.22
10.06
8.08
4.88
11.43
9.14
3.96
28.35
9.14
5.49
14.63
9.14
6.10
11.73
8.53
3.66
15.24
7.92
6.10
15.24
7.92
6.10
Chemical
With
Decay
Rate
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
Chemical
Without
Decay
Rate
Mercury
Aluminum
Mercury
Lead
Lead
Lead
Nickel
Nickel
Nickel
Nickel
Nickel
Nickel
Land Use
Mode
Urban
Urban
Urban
Urban
Urban
Urban
Rural
Rural
Urban
Rural
Rural
Urban
Urban
Urban
Urban
Urban
Rural
Urban
Urban
Urban
Urban
Urban
Rural
Urban
-------
TABLE 2
Parameter Values Used by Each Model in the Ground-Truthing Exercise 1
Parameter
stack height (SH)
stack diameter
exit gas velocity
exit temperature
decay rate
emission rate
wind speed and direction
building height (BH)
building width (BW)
building length (BL)
location coordinates
(latitude, longitude)
Indicators Model (ISCLT3)
single value; median stack height for each
facility; calculation based on all stacks at the
facility
1 m(d)
single value; median exit gas velocity for each
facility; calculation based on all stacks at the
facility
293 K(d)
chemical-specific
total of all stack emissions for the selected
chemical, from AG-1 database
same as AG-1 (both models use the same type
of meteorological data)
assume BH=0 (d)
assume BW=0 (d)
assume BL=0 (d)
single value for each facility (TRI database)
AG-1
single or multiple values; actual height for each
stack-chemical combination
actual stack-specific value
actual stack-specific value
actual stack-specific value
no decay (d)
actual stack-specific value
AG-1 STAR database
actual stack-specific value; in the absence of
stack-specific data, assume BH=SH (d)
actual stack-specific value; in the absence of
stack-specific data, assume BW=SH (d)
actual stack-specific value; in the absence of
stack-specific data, assume BL=SH (d)
single or multiple; stack-specific, as reported in
AG-1 database
Default values are indicated with (d).
-------
TABLE 3
Location and Stack Coordinates of TRI Facilities in New York Selected for the Model Comparison Exercise 1
Urban Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
O
P
Q
R
S
Q
T
S
UTME from
TRI2'3
606266
605871
605972
606266
604574
597218
419367
407979
409308
371672
407979
602462
188265
192038
179187
187367
171697
182600
286491
284606
290572
286491
269772
290572
UTMN from
TRI2'3
734199
732227
729363
734199
729742
726925
761384
770435
767507
756557
770435
773533
759084
755007
766125
753204
782845
765699
781069
784275
783821
781069
764903
783821
Central UTME
fromAG-12'4
606300
605800
606100
606200
604600
597100
419400
403500
409400
371600
403500
402500
188300
192100
179800
187400
171600
182500
285250
284600
291000
285250
291000
291000
Central UTMN
fromAG-12'4
734200
732200
730200
734050
729400
727000
761500
767200
767500
756500
767200
773700
758750
755300
766300
753300
785000
765600
786200
784200
784100
786200
784100
784100
1 Note that certain facilities are used for the evaluation of chemicals both with and without decay rates.
However, these two types of chemicals may be emitted from different stacks within the facility.
2 All coordinates are in meters, with values corresponding to the Universal Transverse Mercator (UTM)
coordinate system.
3 TRI coordinates are a single pair for each facility, contained in the TRI database.
4 Although each stack is provided with its own coordinates in AG-1, for the purposes of comparison to the single
pair of coordinates used in the Indicators Model, a single pair of coordinates was calculated for AG-1.
Coordinates listed for AG-1 are the arithmetic average of the individual coordinates of the set of stacks that emit
the particular chemical elected for the model comparison.
-------
TABLE 4
Facility-Specific Stack Heights (m)
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
O
P
Q
R
S
Q
T
S
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
# Stacks
Emitting
Selected
Chemical
2
3
1
2
6
2
7
7
1
1
3
1
2
1
12
1
1
8
121
1
4
3
1
1
AG-1 Parameters1 [indicators Model Parameters
Mean
Stack
Height
5.49
9.45
1.83
36.58
7.37
4.88
12.63
6.57
2.44
28.35
7.47
5.49
10.97
14.94
4.75
8.23
3.66
3.39
12.51
7.92
8.31
20.93
9.14
6.10
Median
Stack
Height
5.49
9.45
1.83
36.58
9.14
4.88
12.80
7.01
2.44
28.35
8.23
5.49
10.97
14.94
3.35
8.23
3.66
2.44
12.19
7.92
8.84
21.34
9.14
6.10
Minimum
3.66
9.45
1.83
36.58
3.05
3.05
11.58
3.96
2.44
28.35
7.92
5.49
10.36
14.94
1.83
8.23
3.66
2.44
1.83
7.92
3.96
17.68
9.14
6.10
Maximum
7.32
9.45
1.83
36.58
9.14
6.71
12.80
8.84
2.44
28.35
9.75
5.49
11.58
14.94
9.14
8.23
3.66
7.62
35.05
7.92
11.58
23.77
9.14
6.10
Stack #
(Total)
19
3
3
19
24
2
12
17
5
3
17
3
40
7
21
24
99
14
859
11
47
859
31
47
Median
Stack
Height
10.06
9.45
1.22
10.06
8.08
4.88
11.43
9.14
3.96
28.35
9.14
5.49
14.63
9.14
6.10
11.73
8.53
3.66
15.24
7.92
6.10
15.24
7.92
6.10
Mean
Stack
Height
12.48
9.04
5.28
12.48
11.96
4.88
10.19
8.53
3.35
24.38
8.53
5.49
14.67
10.32
11.57
15.19
8.57
4.68
17.97
8.40
6.94
17.97
9.48
6.94
1Although AG-1 uses unique chemical emission-stack combinations, the mean and median heights are presented for model input comparison
purposes. The number of stack heights used in the calculation differ, as AG-1 averages are based only on those stacks which emit chemicals being
analyzed, whereas average stack heights in ISCLT are based upon all stacks at the test case site.
-------
TABLE 5
FaciIity-Specific Exit Gas Velocities (mis)
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
0
P
Q
R
S
Q
T
S
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
# Stacks
Emitting
Selected
Chemical
2
3
1
2
6
2
7
7
1
1
3
1
2
1
12
1
1
8
121
1
4
3
1
1
AG-1 Parameters1 (indicators Model Parameters
Mean Exit
Gas
Velocity
12.21
15.79
23.32
24.54
20.26
20.13
8.26
19.19
9.14
2.77
6.28
4.57
15.03
10.79
0.17
8.23
10.51
15.57
11.01
3.96
14.32
13.72
30.48
11.58
Median
Exit Gas
Velocity
12.21
15.79
23.32
24.54
19.51
20.13
8.63
10.88
9.14
2.77
8.05
4.57
15.03
10.79
0.07
8.23
10.51
16.73
10.67
3.96
16.57
18.90
30.48
11.58
Minimum
4.36
15.79
23.32
11.89
17.01
20.13
6.10
1.19
9.14
2.77
0.70
4.57
13.11
10.79
0.00
8.23
10.51
7.44
0.00
3.96
2.59
2.44
30.48
11.58
Maximum
20.06
15.79
23.32
37.19
26.52
20.13
8.63
80.77
9.14
2.77
10.09
4.57
16.95
10.79
0.61
8.23
10.51
16.73
39.32
3.96
21.55
19.81
30.48
11.58
Stack #
(Total)
19
4
1
19
25
2
13
32
5
3
32
3
40
7
21
27
99
14
873
11
48
873
32
48
Median
Exit Gas
Velocity
4.36
15.79
23.16
4.36
14.72
20.13
8.63
5.82
20.42
7.50
5.82
3.57
15.76
10.79
0.076
8.23
12.80
16.73
11.67
10.06
8.18
11.67
12.12
8.18
Mean Exit
Gas
Velocity
8.64
12.44
23.16
8.64
13.56
20.13
11.66
7.85
15.95
95.37
7.85
3.90
15.21
11.12
1.07
10.68
14.42
15.18
14.69
12.91
8.20
14.69
27.01
8.20
Although AG-1 uses unique chemical emission
comparison purposes. The number of exit gas
which emit chemicals being analyzed, whereas
-stack combinations, the mean and median exit
velocities used in the calculation differ, as AG-1
average exit gas velocities in ISCLT are based
gas velocities are presented for model input
averages are based only on those stacks
upon all stacks at the test case site.
-------
TABLE 6
Facility-Specific Stack Diameters (m)
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
O
P
Q
R
S
Q
T
S
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
# Stacks
Emitting
Selected
Chemical
2
3
1
2
6
2
7
7
1
1
3
1
2
1
12
1
1
8
121
1
4
3
1
1
Mean
Stack
Diameter
0.18
0.91
0.05
1.30
0.49
0.15
1.05
0.26
0.10
0.10
0.66
0.15
0.43
0.91
0.21
0.61
0.33
0.22
0.43
0.10
0.86
0.59
0.10
0.20
Median
Stack
Diameter
0.18
0.91
0.05
1.30
0.61
0.15
1.07
0.36
0.10
0.10
0.61
0.15
0.43
0.91
0.15
0.61
0.33
0.20
0.23
0.10
0.91
0.36
0.10
0.20
Minimum
0.10
0.91
0.05
1.07
0.20
0.15
0.97
0.10
0.10
0.10
0.51
0.15
0.25
0.91
0.10
0.61
0.33
0.20
0.03
0.10
0.20
0.10
0.10
0.20
Maximum
0.25
0.91
0.05
1.52
0.61
0.15
1.07
0.36
0.10
0.10
0.86
0.15
0.61
0.91
0.48
0.61
0.33
0.30
2.69
0.10
1.42
1.32
0.10
0.20
Note: The default value for stack diameter in the Indicators Model is 1 m.
-------
TABLE 7
Facility-Specific Stack Exit Temperatures (K)
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
0
P
Q
R
S
Q
T
S
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
# Stacks
Emitting
Selected
Chemical
2
3
1
2
6
2
7
7
1
1
3
1
2
1
12
1
1
8
121
1
4
3
1
1
Mean Stack
Exit
Temperature
302
311
294
333
293
294
294
303
294
408
371
366
296
294
325
294
294
293
299
450
296
383
366
300
Median Stack
Exit
Temperature
302
311
294
333
293
294
294
297
294
408
326
366
296
294
311
294
294
293
294
450
295
295
366
300
Minimum
294
311
294
333
293
294
294
293
294
408
297
366
294
294
284
294
294
293
284
450
295
294
366
300
Maximum
311
311
294
333
294
294
294
315
294
408
489
366
297
294
363
294
294
293
394
450
300
561
366
300
Note: The default value for stack exit temperature in the Indicators Model is 293 K.
-------
TABLE 8
Facility-Specific Chemical Emission Rates (g/sec)
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
O
P
Q
R
S
Q
T
S
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
# Stacks
Emitting
Selected
Chemical
2
3
1
2
6
2
7
7
1
1
3
1
2
1
12
1
1
8
121
1
4
3
1
1
Mean
Chemical
Emission
Rate
2.20E-05
1.97E+00
3.79E-04
1.19E-04
3.44E-04
7.03E-06
5.22E-02
1 .76E-02
1 .08E-06
3.39E-02
7.85E-03
5.76E-05
3.31 E-02
9.07E-04
1 .39E-03
1.15E-06
7.20E-07
1 .44E-05
2.04E-02
1.16E-01
8.15E-05
2.16E-05
1.18E-04
1 .44E-08
Median
Chemical
Emission
Rate
2.20E-05
1.97E+00
3.79E-04
1.19E-04
4.73E-04
7.03E-06
4.44E-02
1.18E-02
1 .08E-06
3.39E-02
4.60E-03
5.76E-05
3. 31 E-02
9.07E-04
1 .86E-04
1.15E-06
7.20E-07
1 .44E-05
1 .27E-03
1.16E-01
1.90E-05
1.15E-07
1.18E-04
1 .44E-08
Minimum
1.41E-05
1.97E+00
3.79E-04
1.19E-04
4.32E-05
7.03E-06
7.20E-03
1 .02E-03
1 .08E-06
3.39E-02
6.10E-04
5.76E-05
9.50E-03
9.07E-04
5.26E-06
1.15E-06
7.20E-07
1 .44E-05
4.32E-08
1.16E-01
1 .44E-08
1 .44E-08
1.18E-04
1 .44E-08
Maximum
3.00E-05
1.97E+00
3.79E-04
1.19E-04
4.73E-04
7.03E-06
8.88E-02
4.43E-02
1 .08E-06
3.39E-02
1 .83E-02
5.76E-05
5.67E-02
9.07E-04
1 .36E-02
1.15E-06
7.20E-07
1 .44E-05
5.88E-01
1.16E-01
2.88E-04
6.48E-05
1.18E-04
1 .44E-08
Note: These values were used in both AG-1 and ISCLT3 for this analysis. The Indicators Model uses annual
emissions reported to TRI.
-------
TABLE 9
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area, Chemical Characteristic, and Stack Height
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
All Cases
By Metropolitan Area:
Albany
Syracuse
Buffalo
Rochester
By Chemical Characteristic:
Chemical with Decay Rate
Chemical without Decay Rate
By Stack Height:
Om10m
Average
0.984
1.049
0.935
0.962
0.989
0.948
1.020
0.972
0.958
1.021
Standard
Deviation
0.134
0.196
0.067
0.071
0.135
0.066
0.171
0.023
0.076
0.214
Minimum
0.231
0.810
0.527
0.518
0.231
0.231
0.347
0.841
0.518
0.231
Maximum
3.101
1.731
1.097
1.097
3.101
1.417
3.101
1.008
1.097
3.101
Number of
Cells
10539
2640
2640
2640
2619
5259
5280
3520
3520
3499
-------
TABLE 10
Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations and by Metropolitan Area
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
OVERALL Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.955
0.973
0.981
0.984
0.986
0.986
0.986
0.985
0.984
0.984
0.984
Rochester Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
1.053
1.007
0.996
0.991
0.989
0.988
0.987
0.985
0.983
0.986
0.989
Albany Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
1.041
1.056
1.057
1.057
1.055
1.053
1.050
1.048
1.045
1.042
1.049
Buffalo Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.899
0.940
0.954
0.960
0.963
0.965
0.965
0.966
0.965
0.965
0.962
Syracuse Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.828
0.891
0.915
0.928
0.936
0.940
0.942
0.943
0.944
0.943
0.935
Average
Standard
Deviation
0.258
0.180
0.142
0.125
0.113
0.106
0.101
0.098
0.095
0.094
0.134
Average
Standard
Deviation
0.450
0.280
0.189
0.153
0.124
0.106
0.095
0.086
0.081
0.075
0.135
Average
Standard
Deviation
0.281
0.227
0.207
0.199
0.194
0.192
0.190
0.190
0.190
0.190
0.196
Average
Standard
Deviation
0.137
0.101
0.084
0.076
0.071
0.067
0.065
0.063
0.061
0.060
0.071
Average
Standard
Deviation
0.162
0.113
0.087
0.073
0.064
0.057
0.054
0.051
0.050
0.050
0.067
Minimum
0.347
0.231
0.472
0.348
0.590
0.701
0.754
0.790
0.810
0.845
0.231
Minimum
0.347
0.231
0.472
0.348
0.590
0.701
0.754
0.790
0.810
0.887
0.231
Minimum
0.810
0.904
0.928
0.936
0.935
0.931
0.925
0.919
0.912
0.906
0.810
Minimum
0.518
0.680
0.759
0.805
0.833
0.855
0.859
0.862
0.860
0.857
0.518
Minimum
0.527
0.603
0.709
0.754
0.787
0.819
0.833
0.840
0.843
0.845
0.527
Maximum
3.101
2.182
1.879
1.672
1.546
1.497
1.491
1.485
1.482
1.478
3.101
Maximum
3.101
2.182
1.879
1.672
1.546
1.462
1.402
1.356
1.322
1.295
3.101
Maximum
1.731
1.595
1.547
1.521
1.505
1.497
1.491
1.485
1.482
1.478
1.731
Maximum
1.091
1.097
1.097
1.096
1.094
1.092
1.089
1.087
1.084
1.081
1.097
Maximum
1.097
1.076
1.056
1.045
1.039
1.033
1.030
1.027
1.024
1.023
1.097
Number of Cells
192
384
576
768
960
1152
1344
1536
1728
1899
10539
Number of Cells
48
96
144
192
240
288
336
384
432
459
2619
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
-------
TABLE 11
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area, Chemical Characteristic, and Stack Height
Scenario: Facility-Specific Median Stack Height and Exit Gas Velocity of 0.01 m/sec
All Cases
By Metropolitan Area:
Albany
Syracuse
Buffalo
Rochester
By Chemical Characteristic:
Chemical with Decay Rate
Chemical without Decay Rate
By Stack Height:
Om10m
Average
0.980
1.047
0.935
0.964
0.976
0.946
1.015
0.973
0.942
1.027
Standard
Deviation
0.136
0.191
0.069
0.069
0.147
0.072
0.170
0.022
0.093
0.206
Minimum
0.232
0.829
0.459
0.549
0.232
0.232
0.336
0.840
0.406
0.232
Maximum
3.032
1.658
1.001
1.097
3.032
1.434
3.032
1.008
1.097
3.032
Number of
Cells
10539
2640
2640
2640
2619
5259
5280
3520
3520
3499
-------
TABLE 12
Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations and by Metropolitan Area
Scenario: Facility-Specific Median Stack Height and Exit Gas Velocity of 0.01 m/sec
OVERALL Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.944
0.966
0.975
0.979
0.982
0.983
0.983
0.983
0.982
0.982
0.980
Rochester Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
1.017
0.982
0.975
0.974
0.974
0.975
0.975
0.974
0.973
0.976
0.976
Albany Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
1.028
1.050
1.054
1.054
1.053
1.051
1.049
1.046
1.044
1.041
1.047
Buffalo Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.907
0.945
0.957
0.963
0.965
0.967
0.967
0.967
0.967
0.966
0.964
Syracuse Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.822
0.888
0.914
0.927
0.935
0.940
0.942
0.943
0.944
0.943
0.935
Average
Standard
Deviation
0.252
0.181
0.144
0.128
0.116
0.109
0.104
0.100
0.098
0.096
0.136
Average
Standard
Deviation
0.467
0.299
0.208
0.169
0.139
0.120
0.107
0.097
0.091
0.084
0.147
Average
Standard
Deviation
0.254
0.215
0.200
0.193
0.190
0.188
0.187
0.187
0.187
0.188
0.191
Average
Standard
Deviation
0.131
0.096
0.081
0.074
0.069
0.066
0.063
0.062
0.061
0.059
0.069
Average
Standard
Deviation
0.157
0.112
0.088
0.074
0.066
0.060
0.057
0.055
0.054
0.053
0.069
Minimum
0.336
0.232
0.473
0.348
0.591
0.702
0.755
0.790
0.800
0.805
0.232
Minimum
0.336
0.232
0.473
0.348
0.591
0.702
0.755
0.790
0.811
0.838
0.232
Minimum
0.829
0.913
0.934
0.941
0.938
0.933
0.927
0.920
0.914
0.907
0.829
Minimum
0.549
0.706
0.780
0.823
0.848
0.860
0.859
0.862
0.860
0.857
0.549
Minimum
0.459
0.570
0.665
0.715
0.744
0.771
0.782
0.795
0.800
0.805
0.459
Maximum
3.032
2.160
1.866
1.663
1.540
1.487
1.482
1.478
1.475
1.472
3.032
Maximum
3.032
2.160
1.866
1.663
1.540
1.457
1.398
1.353
1.319
1.292
3.032
Maximum
1.658
1.555
1.521
1.504
1.492
1.487
1.482
1.478
1.475
1.472
1.658
Maximum
1.090
1.097
1.097
1.096
1.094
1.092
1.089
1.087
1.084
1.081
1.097
Maximum
0.994
1.001
1.001
1.001
1.001
1.001
1.001
1.001
1.001
1.001
1.001
Number of Cells
192
384
576
768
960
1152
1344
1536
1728
1899
10539
Number of Cells
48
96
144
192
240
288
336
384
432
459
2619
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
-------
TABLE 13
Comparison of AG-1, Indicators Model, and 3-digit SIC Code Parameters
Urban
Area
Albany
Syracuse
Buffalo
Rochester
Case
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
Facility
A
B
C
A
D
E
F
G
H
I
G
J
K
L
M
N
O
P
Q
R
S
Q
T
S
SIC Code
324
329
295
324
331
281
251
326
356
331
326
367
371
344
331
326
329
344
386
267
383 3
386
334
383 J
Chemical
Toluene
Toluene
Toluene
Mercury
Aluminum
Mercury
Toluene
Toluene
Toluene
Lead
Lead
Lead
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
Toluene
Toluene
Toluene
Nickel
Nickel
Nickel
AG-1
Median
Stack
Height1
5.49
9.45
1.83
36.58
9.14
4.88
12.80
7.01
2.44
28.35
8.23
5.49
10.97
14.94
3.35
8.23
3.66
2.44
12.19
7.92
8.84
21.34
9.14
6.10
Indicators
Median
Stack
Height1
10.06
9.45
1.22
10.06
8.08
4.88
11.43
9.14
3.96
28.35
9.14
5.49
14.63
9.14
6.10
11.73
8.53
3.66
15.24
7.92
6.10
15.24
7.92
6.10
3-Digit SIC
Median
Stack
Height1
32.00
12.19
9.14
32.00
24.38
13.11
9.14
9.45
9.14
24.38
9.45
9.14
12.19
9.14
24.38
9.45
12.19
9.14
12.19
9.14
9.14
12.19
12.19
9.14
Ratio of 3-
Digit SIC to
Indicators
Stack
Height
3.18
1.29
7.49
3.18
3.02
2.69
0.80
1.03
2.31
0.86
1.03
1.66
0.83
1.00
4.00
0.81
1.43
2.50
0.80
1.15
1.50
0.80
1.54
1.50
AG-1
Median
Exit Gas
Velocity2
12.21
15.79
23.32
24.54
19.51
20.13
8.63
10.88
9.14
2.77
8.05
4.57
15.03
10.79
0.07
8.23
10.51
16.73
10.67
3.96
16.57
18.90
30.48
11.58
Indicators
Median
Exit Gas
Velocity2
4.36
15.79
23.16
4.36
14.72
20.13
8.63
5.82
20.42
7.50
5.82
3.57
15.76
10.79
0.076
8.23
12.80
16.73
11.67
10.06
8.18
11.67
12.12
8.18
3-Digit
SIC
Median
Exit Gas
Velocity2
12.19
12.10
14.01
12.19
8.96
9.08
10.72
9.28
8.37
8.96
9.28
8.10
10.76
8.63
8.96
9.28
12.10
8.63
9.71
10.79
8.00
9.71
9.30
8.00
Ratio of 3-
Digit SIC to
Indicators
Exit Gas
Velocity
2.80
0.77
0.60
2.80
0.61
0.45
1.24
1.59
0.41
1.19
1.59
2.27
0.68
0.80
117.89
1.13
0.95
0.52
0.83
1.07
0.98
0.83
0.77
0.98
1 Stack height in meters.
2Exit gas velocity in meters per second.
3Facility S reported an incorrect SIC code (there is no code 383). The median stack height and exit gas velocity used are those of SIC code 38.
-------
TABLE 14
Summary Statistics for (ISCLT3/AG1) Ratio by Metropolitan Area, Chemical Characteristic, and Stack Height
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
All Cases
By Metropolitan Area:
Albany
Syracuse
Buffalo
Rochester
By Chemical Characteristic:
Chemical with Decay Rate
Chemical without Decay Rate
By Stack Height:
Om10m
Average
0.936
0.871
0.940
0.930
1.001
0.912
0.959
0.934
0.898
0.974
Standard
Deviation
0.131
0.125
0.065
0.113
0.169
0.119
0.138
0.076
0.105
0.178
Minimum
0.248
0.479
0.484
0.439
0.248
0.248
0.383
0.639
0.439
0.248
Maximum
3.385
1.079
1.002
1.099
3.385
1.565
3.385
1.008
1.099
3.385
Number of
Cells
10539
2640
2640
2640
2619
5259
5280
3520
3520
3499
-------
TABLE 15
Summary Statistics for (ISCLT3/AG1) Ratio by Ring for All Locations and by Metropolitan Area
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
OVERALL Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.889
0.917
0.928
0.933
0.937
0.938
0.939
0.939
0.938
0.938
0.936
Rochester Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Albany Summar
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
1.081
1.025
1.011
1.005
1.003
1.001
0.999
0.997
0.995
0.995
1.001
y
Average
0.816
0.857
0.869
0.873
0.874
0.875
0.874
0.873
0.872
0.871
0.871
Buffalo Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.858
0.904
0.919
0.927
0.930
0.933
0.934
0.934
0.934
0.934
0.930
Syracuse Summary
1st ring:
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Overall
Average
0.801
0.880
0.912
0.929
0.939
0.945
0.949
0.951
0.952
0.952
0.940
Average
Standard
Deviation
0.252
0.177
0.141
0.125
0.114
0.107
0.103
0.100
0.097
0.096
0.131
Average
Standard
Deviation
0.520
0.324
0.227
0.187
0.158
0.140
0.129
0.120
0.114
0.111
0.169
Average
Standard
Deviation
0.170
0.144
0.133
0.128
0.125
0.123
0.122
0.121
0.120
0.119
0.125
Average
Standard
Deviation
0.179
0.142
0.126
0.119
0.114
0.111
0.108
0.107
0.105
0.104
0.113
Average
Standard
Deviation
0.141
0.100
0.077
0.066
0.058
0.054
0.052
0.051
0.050
0.050
0.065
Minimum
0.383
0.248
0.505
0.371
0.630
0.662
0.663
0.664
0.662
0.660
0.248
Minimum
0.383
0.248
0.505
0.371
0.630
0.748
0.803
0.822
0.830
0.836
0.248
Minimum
0.479
0.596
0.633
0.649
0.656
0.662
0.663
0.664
0.662
0.660
0.479
Minimum
0.439
0.601
0.676
0.722
0.736
0.736
0.736
0.735
0.733
0.732
0.439
Minimum
0.484
0.590
0.681
0.730
0.757
0.783
0.793
0.804
0.809
0.814
0.484
Maximum
3.385
2.354
2.016
1.790
1.653
1.561
1.496
1.447
1.409
1.380
3.385
Maximum
3.385
2.354
2.016
1.790
1.653
1.561
1.496
1.447
1.409
1.380
3.385
Maximum
1.079
1.061
1.054
1.050
1.048
1.046
1.045
1.044
1.043
1.043
1.079
Maximum
1.095
1.099
1.099
1.097
1.095
1.093
1.090
1.087
1.084
1.082
1.099
Maximum
0.948
0.976
0.984
0.988
0.992
0.998
0.999
1.001
1.002
1.001
1.002
Number of Cells
192
384
576
768
960
1152
1344
1536
1728
1899
10539
Number of Cells
48
96
144
192
240
288
336
384
432
459
2619
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
Number of Cells
48
96
144
192
240
288
336
384
432
480
2640
-------
TABLE 16
Exposure Event Counts Surrounding TRI Facilities
All persons
Race sub-populations
White
Black
Native American
Asian/Pacific Islander
Hispanic
Age sub-populations
Age<18
Age >65
count
%
count
%
count
%
count
%
count
%
count
%
count
%
count
%
Distance to Facility (+/- 500m)
<1 km 1-2 km 2-3 km 3-4 km 4-5 km 5-6 km 6-7 km 7-8 km 8-9 km 9-10 km
36,359 116,782 187,508 246,084 297,454 339,672 377,853 413,268 449,694 470,159
1.2% 4.0% 6.4% 8.4% 10.1% 11.6% 12.9% 14.1% 15.3% 16.0%
25,598 81,439 128,781 168,139 202,677 231,605 258,394 282,899 308,878 323,517
1.3% 4.0% 6.4% 8.4% 10.1% 11.5% 12.8% 14.1% 15.4% 16.1%
6,632 21,605 35,750 47,300 57,411 65,952 72,971 79,173 84,926 87,440
1.2% 3.9% 6.4% 8.5% 10.3% 11.8% 13.1% 14.2% 15.2% 15.6%
197 611 948 1,212 1,424 1,595 1,735 1,850 1,971 2,029
1.5% 4.5% 7.0% 8.9% 10.5% 11.8% 12.8% 13.6% 14.5% 14.9%
1,027 3,700 6,579 9,260 11,787 13,611 15,291 17,153 19,454 20,903
0.9% 3.1% 5.5% 7.8% 9.9% 11.5% 12.9% 14.4% 16.4% 17.6%
5,472 18,134 29,737 38,909 46,750 52,553 57,933 63,641 68,652 72,224
1.2% 4.0% 6.5% 8.6% 10.3% 11.6% 12.8% 14.0% 15.1% 15.9%
9,492 30,177 48,086 62,773 75,553 86,163 95,519 104,133 112,815 117,843
1.3% 4.1% 6.5% 8.5% 10.2% 11.6% 12.9% 14.0% 15.2% 15.9%
4,668 14,779 23,360 30,321 36,533 41,603 46,172 50,354 54,669 56,949
1.3% 4.1% 6.5% 8.4% 10.2% 11.6% 12.8% 14.0% 15.2% 15.8%
Total
(0-1 Okm)
2,934,834
100.0%
2,011,927
100.0%
559,159
100.0%
13,571
100.0%
118,765
100.0%
454,006
100.0%
742,554
100.0%
359,409
100.0%
Notes:
1. Data are from facilities reporting air releases in 1996.
2. Counts are in thousands. Percentages are of subpopulation totals.
3. Each person in the U.S. is assigned to each TRI facility within a specified distance ring of them, but is not removed from the Census database.
Therefore, due to multiple impacts on one person of facilities located at varying distances, the total number of exposure events exceeds the U.S. population.
-------
TABLE 17
Facility Rankings Based on Indicator Elements for Chemical with Decay Rate (Toluene)
AG-1
Facility
B
F
Q
G
R
K
M
L
C
S
A
H
Total
Indicator
Element1
16671
3226
3097
801
144
47
14
1.4
0.78
0.49
0.13
0.0019
24003
Percent
of Total
69.45%
13.44%
12.90%
3.34%
0.60%
0.20%
0.06%
0.01%
0.003%
0.002%
0.001%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
ISCLT 3-Facility-Specific Median Values
Facility
B
Q
F
G
R
K
M
L
C
S
A
H
Total
Indicator
Element1
16642
2736
2633
670
138
41
14
1.5
0.74
0.48
0.12
0.0018
22876
Percent
of Total
72.75%
1 1 .96%
11.51%
2.93%
0.60%
0.18%
0.06%
0.01%
0.003%
0.002%
0.001%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
ISCLT 3- SIC-Code Based Median Values
Facility
B
Q
F
G
R
K
M
L
C
S
A
H
Total
Indicator
Element1
15767
2919
2729
638
137
44
10
1.5
0.74
0.48
0.08
0.0018
22248
Percent
of Total
70.87%
13.12%
12.27%
2.87%
0.62%
0.20%
0.05%
0.01%
0.003%
0.002%
0.0004%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
Indicator Elements are the product of pollutant concentration and population in each cell, summed over all 440 cells surrounding a TRI facility.
-------
TABLE 18
Facility Rankings Based on Indicator Elements for Chemicals without Decay Rates
AG-1
Facility
G
L
D
A
J
T
P
Q
E
O
N
S
Total
Indicator
Element1
130
101
6.3
0.42
0.28
0.15
0.09
0.06
0.05
0.003
0.0015
0.000022
238
Percent
of Total
54.45%
42.45%
2.66%
0.18%
0.12%
0.06%
0.04%
0.03%
0.02%
0.001%
0.001%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
ISCLT 3-Facility-Specific Median Values
Facility
G
L
D
A
J
T
P
Q
E
O
N
S
Total
Indicator
Element1
133
76
6.1
0.66
0.27
0.13
0.08
0.07
0.04
0.002
0.0014
0.000021
216
Percent
of Total
61 .47%
35.11%
2.84%
0.30%
0.12%
0.06%
0.04%
0.03%
0.02%
0.00%
0.001%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
ISCLT 3- SIC-Code Based Median Values
Facility
G
L
D
A
J
T
P
Q
E
O
N
S
Total
Indicator
Element1
127
83
4.6
0.44
0.26
0.12
0.08
0.08
0.04
0.002
0.0014
0.000021
215
Percent
of Total
58.91%
38.47%
2.15%
0.20%
0.12%
0.05%
0.04%
0.04%
0.02%
0.001%
0.001%
0.00001%
100.00%
Rank
1
2
3
4
5
6
7
8
9
10
11
12
Indicator Elements are the product of pollutant concentration and population in each cell, summed over all 440 cells surrounding a TRI facility.
-------
FIGURES
-------
FIGURE 1A
Example Concentrations (ug/m3) Predicted byAG1
Scenario: Facility-Specific Median Stack Height and Constant Exit Gas Velocity of 0.01 m/sec
751500
752500
753500
754500
755500
756500
757500
758500
759500
760500
761500
762500
763500
764500
765500
766500
767500
768500
769500
770500
771500
409400
3.6E-03
4.4E-03
5.4E-03
6.5E-03
7.9E-03
9.4E-03
1.2E-02
1.8E-02
2.4E-02
3.1E-02
3.7E-02
3.5E-02
3.3E-02
3.0E-02
2.7E-02
2.4E-02
2.1E-02
1.8E-02
1.6E-02
1.4E-02
1.2E-02
410400
3.8E-03
4.2E-03
5.2E-03
6.5E-03
8.1E-03
9.9E-03
1.2E-02
1.8E-02
2.6E-02
3.5E-02
4.3E-02
4.1E-02
3.7E-02
3.4E-02
3.0E-02
2.6E-02
2.2E-02
1.9E-02
1.6E-02
1.4E-02
1.2E-02
411400
4.0E-03
4.5E-03
4.9E-03
6.4E-03
8.2E-03
1.0E-02
1.3E-02
1.8E-02
2.9E-02
4.0E-02
5.1E-02
4.8E-02
4.3E-02
3.8E-02
3.3E-02
2.8E-02
2.3E-02
1.9E-02
1.6E-02
1.4E-02
1.2E-02
412400
4.3E-03
4.8E-03
5.3E-03
5.9E-03
8.0E-03
1.1E-02
1.4E-02
1.8E-02
3.1E-02
4.6E-02
6.2E-02
5.7E-02
5.0E-02
4.3E-02
3.6E-02
2.9E-02
2.4E-02
1.9E-02
1.6E-02
1.3E-02
1.1E-02
413400
4.5E-03
5.0E-03
5.7E-03
6.5E-03
7.4E-03
1.1E-02
1.5E-02
2.0E-02
3.3E-02
5.5E-02
7.7E-02
6.9E-02
5.9E-02
4.9E-02
3.9E-02
3.0E-02
2.4E-02
1.9E-02
1.6E-02
1.3E-02
1.1E-02
414400
4.7E-03
5.3E-03
6.1E-03
7.1E-03
8.2E-03
9.6E-03
1.5E-02
2.2E-02
3.2E-02
6.5E-02
1.0E-01
8.7E-02
7.1E-02
5.4E-02
4.1E-02
3.1E-02
2.4E-02
1.9E-02
1.5E-02
1.2E-02
9.7E-03
415400
5.0E-03
5.5E-03
6.4E-03
7.6E-03
9.0E-03
1.1E-02
1.3E-02
2.2E-02
3.6E-02
7.8E-02
1.4E-01
1.1E-01
8.4E-02
6.0E-02
4.2E-02
3.0E-02
2.3E-02
1.7E-02
1.3E-02
1.0E-02
8.7E-03
416400
6.7E-03
7.2E-03
7.5E-03
7.9E-03
9.7E-03
1.2E-02
1.5E-02
2.0E-02
4.0E-02
8.8E-02
2.0E-01
1.5E-01
9.9E-02
6.2E-02
4.1E-02
2.8E-02
2.0E-02
1.5E-02
1.2E-02
1.0E-02
9.1E-03
417400
8.6E-03
9.5E-03
1.1E-02
1.2E-02
1.3E-02
1.4E-02
1.7E-02
2.4E-02
3.5E-02
9.6E-02
3.5E-01
2.1E-01
1.1E-01
5.9E-02
3.5E-02
2.3E-02
1.8E-02
1.5E-02
1.3E-02
1.1E-02
9.4E-03
418400
1.0E-02
1.2E-02
1.4E-02
1.6E-02
1.9E-02
2.3E-02
2.8E-02
3.4E-02
4.6E-02
9.0E-02
8.3E-01
2.5E-01
8.8E-02
4.8E-02
3.4E-02
2.5E-02
2.0E-02
1.6E-02
1.3E-02
1.1E-02
9.5E-03
419400
1.2E-02
1.4E-02
1.7E-02
2.0E-02
2.5E-02
3.3E-02
4.5E-02
6.8E-02
1.2E-01
2.9E-01
2.5E-01
9.9E-02
5.5E-02
3.7E-02
2.6E-02
2.0E-02
1.6E-02
1.3E-02
1.1E-02
9.5E-03
420400
1.4E-02
1.6E-02
1.9E-02
2.4E-02
3.0E-02
4.0E-02
5.7E-02
8.9E-02
1.5E-01
2.8E-01
7.2E-01
2.0E-01
8.2E-02
4.8E-02
3.3E-02
2.5E-02
1.9E-02
1.6E-02
1.3E-02
1.1E-02
9.4E-03
421400
1.5E-02
1.7E-02
2.1E-02
2.6E-02
3.3E-02
4.4E-02
5.7E-02
7.6E-02
1.0E-01
1.6E-01
2.9E-01
1.8E-01
7.9E-02
4.8E-02
3.1E-02
2.2E-02
1.8E-02
1.5E-02
1.2E-02
1.0E-02
9.0E-03
422400
1.6E-02
1.8E-02
2.2E-02
2.7E-02
3.2E-02
3.8E-02
4.7E-02
5.7E-02
7.8E-02
1.1E-01
1.7E-01
1.3E-01
8.0E-02
4.5E-02
3.2E-02
2.3E-02
1.7E-02
1.3E-02
1.1E-02
9.8E-03
8.6E-03
423400
1.6E-02
1.8E-02
2.1E-02
2.4E-02
2.8E-02
3.2E-02
3.7E-02
4.7E-02
5.9E-02
8.3E-02
1.1E-01
9.4E-02
7.0E-02
4.6E-02
3.0E-02
2.3E-02
1.8E-02
1.4E-02
1.1E-02
9.4E-03
8.0E-03
424400
1.5E-02
1.7E-02
1.9E-02
2.1E-02
2.4E-02
2.7E-02
3.2E-02
3.9E-02
4.7E-02
6.5E-02
8.1E-02
7.1E-02
5.9E-02
4.3E-02
3.1E-02
2.1E-02
1.7E-02
1.4E-02
1.2E-02
9.8E-03
8.3E-03
425400
1.4E-02
1.5E-02
1.7E-02
1.8E-02
2.0E-02
2.4E-02
2.8E-02
3.2E-02
4.0E-02
5.2E-02
6.2E-02
5.6E-02
4.9E-02
3.9E-02
3.0E-02
2.2E-02
1.6E-02
1.4E-02
1.2E-02
9.8E-03
8.4E-03
426400
1.3E-02
1.4E-02
1.5E-02
1.6E-02
1.9E-02
2.1E-02
2.4E-02
2.7E-02
3.5E-02
4.3E-02
5.0E-02
4.6E-02
4.1E-02
3.5E-02
2.8E-02
2.2E-02
1.7E-02
1.3E-02
1.1E-02
9.7E-03
8.5E-03
427400
1.2E-02
1.2E-02
1.3E-02
1.5E-02
1.7E-02
1.9E-02
2.1E-02
2.4E-02
3.0E-02
3.6E-02
4.1E-02
3.8E-02
3.5E-02
3.1E-02
2.6E-02
2.1E-02
1.7E-02
1.4E-02
1.1E-02
9.5E-03
8.3E-03
428400
1.0E-02
1.1E-02
1.2E-02
1.4E-02
1.5E-02
1.7E-02
1.9E-02
2.2E-02
2.7E-02
3.1E-02
3.5E-02
3.3E-02
3.0E-02
2.7E-02
2.4E-02
2.0E-02
1.7E-02
1.4E-02
1.1E-02
9.2E-03
8.1E-03
429400
9.5E-03
1.0E-02
1.2E-02
1.3E-02
1.4E-02
1.5E-02
1.7E-02
2.0E-02
2.4E-02
2.7E-02
3.0E-02
2.8E-02
2.6E-02
2.4E-02
2.2E-02
1.9E-02
1.6E-02
1.4E-02
1.1E-02
9.5E-03
7.9E-03
NOTE: Row and column headings represent Universal Transverse Mercator (UTM) coordinates in meters.
-------
FIGURE 1B
Example Concentrations (ug/m3) Predicted by ISCLT3
Scenario: Facility-Specific Median Stack Height and Constant Exit Gas Velocity of 0.01 m/sec
751500
752500
753500
754500
755500
756500
757500
758500
759500
760500
761500
762500
763500
764500
765500
766500
767500
768500
769500
770500
771500
409400
3.27E-03
4.04E-03
4.96E-03
6.03E-03
7.27E-03
8.68E-03
1.08E-02
1.63E-02
2.22E-02
2.82E-02
3.42E-02
3.23E-02
3.00E-02
2.75E-02
2.48E-02
2.19E-02
1.91E-02
1.66E-02
1.43E-02
1.23E-02
1.06E-02
410400
3.48E-03
3.81 E-03
4.82E-03
6.04E-03
7.49E-03
9.18E-03
1.11E-02
1.68E-02
2.43E-02
3.21 E-02
3.97E-02
3.72E-02
3.41 E-02
3.07E-02
2.71 E-02
2.34E-02
2.01 E-02
1.71 E-02
1.45E-02
1.23E-02
1.06E-02
411400
3.70E-03
4.08E-03
4. 51 E-03
5.88E-03
7.57E-03
9.62E-03
1.20E-02
1.67E-02
2.65E-02
3.68E-02
4.69E-02
4.34E-02
3.92E-02
3.45E-02
2.96E-02
2.50E-02
2.09E-02
1.75E-02
1.45E-02
1.23E-02
1.06E-02
412400
3.92E-03
4.37E-03
4.88E-03
5.45E-03
7.40E-03
9.86E-03
1.29E-02
1.65E-02
2.86E-02
4.26E-02
5.65E-02
5.14E-02
4.53E-02
3.87E-02
3.22E-02
2.64E-02
2.15E-02
1.74E-02
1.45E-02
1.22E-02
1.03E-02
413400
4. 11 E-03
4.63E-03
5.25E-03
5.97E-03
6.79E-03
9.70E-03
1.35E-02
1.83E-02
3.00E-02
4.99E-02
6.99E-02
6.22E-02
5.29E-02
4.33E-02
3.47E-02
2.74E-02
2.15E-02
1.74E-02
1.42E-02
1.17E-02
9.73E-03
414400
4.28E-03
4.87E-03
5.60E-03
6.47E-03
7.53E-03
8.79E-03
1.35E-02
1.99E-02
2.94E-02
5.88E-02
8.96E-02
7.68E-02
6.20E-02
4.81 E-02
3.66E-02
2.75E-02
2.15E-02
1.69E-02
1.35E-02
1.09E-02
8.88E-03
415400
4.60E-03
5.06E-03
5.89E-03
6.93E-03
8.25E-03
9.93E-03
1.20E-02
2.04E-02
3.25E-02
6.92E-02
1.20E-01
9.71 E-02
7.23E-02
5.22E-02
3.69E-02
2.72E-02
2.04E-02
1.56E-02
1.21 E-02
9.57E-03
7.94E-03
416400
6.21 E-03
6.59E-03
6.88E-03
7.28E-03
8.86E-03
1.10E-02
1.39E-02
1.78E-02
3.56E-02
7.74E-02
1.74E-01
1.25E-01
8.27E-02
5.32E-02
3.60E-02
2.50E-02
1.80E-02
1.32E-02
1.10E-02
9.55E-03
8.36E-03
417400
7.91 E-03
8.76E-03
9.71 E-03
1.07E-02
1.17E-02
1.23E-02
1.55E-02
2.13E-02
3.03E-02
8.30E-02
2.83E-01
1.57E-01
8.64E-02
4.99E-02
3.06E-02
2.06E-02
1.68E-02
1.39E-02
1.17E-02
1.00E-02
8.68E-03
418400
9.62E-03
1.10E-02
1.27E-02
1.48E-02
1.75E-02
2.10E-02
2.54E-02
2.99E-02
3.83E-02
6.95E-02
5.73E-01
1.71E-01
7.05E-02
4.18E-02
3.04E-02
2.31 E-02
1.81 E-02
1.47E-02
1.22E-02
1.03E-02
8.88E-03
419400
1.13E-02
1.31 E-02
1.55E-02
1.87E-02
2.32E-02
2.98E-02
4.03E-02
5.88E-02
9.75E-02
2.10E-01
2.00E-01
8.54E-02
4.96E-02
3.34E-02
2.44E-02
1.88E-02
1.50E-02
1.24E-02
1.04E-02
8.92E-03
420400
1.26E-02
1.49E-02
1.78E-02
2.19E-02
2.78E-02
3.67E-02
5.15E-02
7.91 E-02
1.33E-01
2.41 E-01
5.33E-01
1.58E-01
6.96E-02
4.22E-02
3.02E-02
2.28E-02
1.79E-02
1.45E-02
1.20E-02
1.02E-02
8.75E-03
421400
1.37E-02
1.62E-02
1.95E-02
2.40E-02
3.05E-02
4. 01 E-02
5.21 E-02
6.91 E-02
9.40E-02
1.44E-01
2.43E-01
1.53E-01
6.87E-02
4.25E-02
2.81 E-02
2.01 E-02
1.63E-02
1.35E-02
1.14E-02
9.73E-03
8.43E-03
422400
1.44E-02
1.70E-02
2.04E-02
2.47E-02
2.94E-02
3.54E-02
4.32E-02
5.28E-02
7.18E-02
9.96E-02
1.45E-01
1.13E-01
7.10E-02
4. 01 E-02
2.87E-02
2. 11 E-02
1.60E-02
1.24E-02
1.05E-02
9.12E-03
8.00E-03
423400
1.48E-02
1.70E-02
1.93E-02
2.22E-02
2.58E-02
3.00E-02
3.48E-02
4.42E-02
5.49E-02
7.56E-02
9.94E-02
8.40E-02
6.30E-02
4. 21 E-02
2.70E-02
2.09E-02
1.63E-02
1.30E-02
1.06E-02
8.74E-03
7.47E-03
424400
1.40E-02
1.56E-02
1.75E-02
1.97E-02
2.23E-02
2.50E-02
3.04E-02
3.66E-02
4.36E-02
5.93E-02
7.34E-02
6.48E-02
5.37E-02
3.97E-02
2.84E-02
1.97E-02
1.60E-02
1.31 E-02
1.08E-02
9.06E-03
7.68E-03
425400
1.29E-02
1.42E-02
1.57E-02
1.73E-02
1.91 E-02
2.25E-02
2.64E-02
3.05E-02
3.76E-02
4.79E-02
5.71 E-02
5.17E-02
4.47E-02
3.61 E-02
2.76E-02
2.07E-02
1.52E-02
1.28E-02
1.07E-02
9.14E-03
7.84E-03
426400
1.18E-02
1.29E-02
1.40E-02
1.52E-02
1.75E-02
2.01 E-02
2.29E-02
2.57E-02
3.25E-02
3.96E-02
4.59E-02
4.24E-02
3.77E-02
3.24E-02
2.60E-02
2.05E-02
1.59E-02
1.22E-02
1.05E-02
9.04E-03
7.84E-03
427400
1.08E-02
1.16E-02
1.24E-02
1.41 E-02
1.59E-02
1.79E-02
2.00E-02
2.30E-02
2.83E-02
3.34E-02
3.80E-02
3.55E-02
3.23E-02
2.86E-02
2. 41 E-02
1.97E-02
1.59E-02
1.27E-02
1.01 E-02
8.80E-03
7.72E-03
428400
9.80E-03
1.04E-02
1.17E-02
1.30E-02
1.45E-02
1.60E-02
1.75E-02
2.08E-02
2.48E-02
2.87E-02
3.21 E-02
3.03E-02
2.80E-02
2.53E-02
2.22E-02
1.87E-02
1.55E-02
1.28E-02
1.05E-02
8.49E-03
7.53E-03
429400
8.90E-03
9.85E-03
1.09E-02
1.20E-02
1.32E-02
1.44E-02
1.58E-02
1.89E-02
2.20E-02
2.50E-02
2.76E-02
2.63E-02
2.45E-02
2.25E-02
2.03E-02
1.75E-02
1.49E-02
1.26E-02
1.06E-02
8.80E-03
7.28E-03
NOTE: Row and column headings represent Universal Transverse Mercator (UTM) coordinates in meters.
-------
FIGURE 1C
Example Concentration Ratios (ISCLT3/AG1)
Scenario: Facility-Specific Median Stack Height and Constant Exit Gas Velocity of 0.01 m/sec
751500
752500
753500
754500
755500
756500
757500
758500
759500
760500
761500
762500
763500
764500
765500
766500
767500
768500
769500
770500
771500
409400
0.956
0.959
0.961
0.963
0.965
0.966
0.967
0.966
0.965
0.964
0.964
0.962
0.961
0.960
0.959
0.960
0.960
0.960
0.961
0.960
0.960
410400
0.958
0.959
0.962
0.964
0.965
0.967
0.968
0.966
0.965
0.964
0.963
0.961
0.960
0.959
0.958
0.959
0.960
0.961
0.962
0.962
0.961
411400
0.959
0.960
0.961
0.964
0.965
0.967
0.968
0.967
0.964
0.962
0.961
0.958
0.956
0.955
0.956
0.958
0.960
0.961
0.962
0.962
0.962
412400
0.960
0.961
0.962
0.962
0.964
0.965
0.966
0.967
0.961
0.959
0.957
0.954
0.951
0.950
0.953
0.956
0.959
0.962
0.962
0.963
0.963
413400
0.961
0.962
0.962
0.962
0.961
0.962
0.963
0.964
0.959
0.953
0.951
0.946
0.943
0.945
0.948
0.953
0.958
0.960
0.962
0.963
0.963
414400
0.961
0.962
0.962
0.961
0.959
0.957
0.958
0.957
0.956
0.944
0.940
0.932
0.928
0.935
0.944
0.951
0.956
0.959
0.962
0.963
0.964
415400
0.962
0.962
0.962
0.960
0.957
0.953
0.948
0.947
0.946
0.930
0.922
0.909
0.909
0.922
0.938
0.947
0.953
0.958
0.962
0.964
0.965
416400
0.964
0.963
0.962
0.959
0.956
0.949
0.940
0.929
0.927
0.912
0.894
0.872
0.885
0.912
0.929
0.942
0.953
0.959
0.963
0.966
0.967
417400
0.965
0.964
0.963
0.960
0.955
0.946
0.933
0.915
0.894
0.888
0.839
0.807
0.863
0.897
0.923
0.943
0.955
0.961
0.965
0.968
0.968
418400
0.965
0.965
0.963
0.960
0.954
0.944
0.929
0.906
0.866
0.795
0.704
0.717
0.840
0.901
0.931
0.948
0.959
0.965
0.969
0.970
0.971
419400
0.966
0.966
0.964
0.961
0.955
0.945
0.930
0.905
0.858
0.752
0.806
0.892
0.928
0.947
0.959
0.967
0.971
0.973
0.974
0.973
420400
0.967
0.967
0.965
0.963
0.958
0.949
0.937
0.918
0.893
0.880
0.766
0.802
0.878
0.920
0.942
0.956
0.965
0.969
0.972
0.973
0.973
421400
0.968
0.968
0.967
0.965
0.962
0.956
0.948
0.942
0.943
0.916
0.869
0.871
0.904
0.926
0.943
0.957
0.965
0.969
0.972
0.973
0.973
422400
0.968
0.969
0.968
0.967
0.967
0.965
0.962
0.963
0.951
0.934
0.914
0.911
0.922
0.938
0.949
0.959
0.966
0.969
0.972
0.973
0.973
423400
0.968
0.970
0.971
0.972
0.972
0.973
0.975
0.968
0.961
0.947
0.937
0.935
0.940
0.948
0.957
0.963
0.966
0.969
0.972
0.973
0.973
424400
0.970
0.972
0.974
0.976
0.977
0.980
0.977
0.973
0.968
0.958
0.953
0.951
0.954
0.957
0.962
0.965
0.968
0.970
0.972
0.972
0.972
425400
0.972
0.974
0.977
0.979
0.982
0.980
0.978
0.976
0.971
0.966
0.963
0.962
0.963
0.965
0.966
0.968
0.969
0.971
0.971
0.971
0.971
426400
0.973
0.976
0.978
0.982
0.981
0.980
0.979
0.978
0.973
0.970
0.968
0.967
0.967
0.969
0.969
0.970
0.971
0.971
0.971
0.971
0.970
427400
0.974
0.977
0.980
0.980
0.981
0.980
0.979
0.978
0.975
0.973
0.972
0.971
0.971
0.971
0.972
0.972
0.972
0.971
0.970
0.970
0.969
428400
0.975
0.978
0.978
0.979
0.979
0.979
0.979
0.978
0.976
0.975
0.974
0.973
0.973
0.973
0.973
0.973
0.972
0.972
0.970
0.968
0.968
429400
0.975
0.976
0.977
0.978
0.978
0.978
0.978
0.977
0.976
0.975
0.974
0.974
0.974
0.974
0.973
0.973
0.973
0.972
0.970
0.968
0.966
NOTE: Row and column headings represent Universal Transverse Mercator (UTM) coordinates in meters.
-------
FIGURE 2
Example Contour Plots of Concentrations Predicted By Each Model
and Example Contour Plot of the Concentration Ratios
Scenario: Facility-Specific Median Stack Height and Constant Exit Gas Velocity of 0.01 m/sec
FIGURE 2A
ISCLT-3 Concentration (ug/m3)
o
o
s
o
o o
o o
? a
I
o o
o o
IT) N-
IB
!
m
f
|
O O
o o
s ?
O
o
a
o o
o o
IT) N-
'b'bo no.
0
"759501 "
757501 D °"
5-1
3-0.8
t-0.6
2-0.4
0.2
O
O
S
FIGURE 2B
AG-1 Concentration (ug/m3)
1
(
c
FIGURE 2C
(ISCLT3/AG1) Ratio
§
§
a
i
§
s
-V
*
lv
is p"
§
s
T^
' i
§
3
§
*
§
a
§
s
§
D 0.9-1
0.8-0.9
0.7-0.8
D 0.6-0. 7
§
s
NOTE: All axes represent Universal Transverse Mercator (UTM) coordinates In meters.
-------
FIGURE 3
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Albany
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
Albany, Case 1
450 -
400 -
oi 350 -
" 300 -
° 250 -
n 200 -
| 150 -
z 100 -
50 -
OCM^CDCOT-CM^CD
OOOO T- T- T-
(ISCLT3/AG1)
CO CM
Albany, Case 2
Number of Cells
-* -* IO IO CO CO £* -
OIOOIOOIOOIOC
DOOOOOOOOC
OOOO T-T-T-T-
(ISCLT3/AG1)
Albany, Case 3
450 -
400 -
oi 350 -
" 300 -
° 250-
% 200 -
| 150 -
z 100 -
50 -
OOOO T-T-T-T-
(ISCLT3/AG1)
Number of Cells
-^ -^ rO h
O1 O O1 O C
O O O O C
D O O O O C
Albany, Case 4
Number of Cells
01 o cri o 01 o 01 i
DOOOOOOOC
o o
CD CD T- T- T- T-
(ISCLT3/AG1)
Albany, Case 5
Number of Cells
O1OO1OO1OO1OO1C
DOOOOOOOOOC
CD CD
CD CD T- T- T- T-
(ISCLT3/AG1)
Albany, Case 6
450 -
w 400-
oi 350 -
^ 300 -
f 250-
^ 200 -
| 150 -
2 100 -
50 -
CD CD
Albany, All Cases
OOOO T-T-T-T-
(ISCLT3/AG1)
CD CD T- T- T- T-
(ISCLT3/AG1)
-------
FIGURE 4
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Buffalo
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
Buffalo, Case 1
450 --
400 --
350 --
300 --
250 - -
200 - -
150 --
100 --
50 --
0
(ISCLT3/AG1)
Buffalo, Case 2
450 -
400 -
oi 350 -
" 300 -
° 250 -
£ 200 -
| 150 -
z 100 -
50 -
0 -
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 3
450 -
400 -
oi 350 -
" 300 -
° 250-
% 200 -
| 150 -
z 100 -
50 -
OCN'^CDCOT-CN'^-CDCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 4
450 -
400-
oi 350 -
" 300 -
° 250-
% 200 -
| 150 -
z 100 -
50 -
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 5
250 "
£ 200 -
° 150 -
E 100 -
50 -
0 -
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 6
450 -
400-
oi 350 -
" 300 -
° 250-
% 200 -
| 150 -
z 100 -
50 -
OOOO T-T-T-T-
ISCLT3/AG1
1800 -
1600 -
w 1 400 -
8 1200 -
5 1000 -
5 800 -
E 600 -
z 400 -
200 -
0 -
Buffalo, All Cases
OCN-3-COCOT-CN-3-COCOCN
OOOO T-T-T-T-
ISCLT3/AG1
-------
FIGURE 5
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Rochester
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
200
180
160
oi 140
120
100
80
60
40
20
0
2
°
500
450--
400-
oi 350 --
H 300 --
° 250-
£ 200 --
I 150--
z 100--
50--
Rochester, Case 1
(ISCLT3/AG1)
500 i
450-
400-
oi 350 -
2 300-
2 250-
£ 200-
| 150 -
z 100 -
50-
0-
Rochester, Case 2
O(M^l-CDCOi-(M^l-CDCO(M
OOOO 1-^1-^1-^1-^
(ISCLT3/AG1)
Rochester, Case 3
CD CO i-
(ISCLT3/AG1)
Rochester, Case 41
140
120-
I 100-
o
s
CD CO (M
(ISCLT3/AG1)
Rochester, Case 5
300
250-
$ 200-
2 150-
E 100-
Z 50-
0 -
.
1
1
OOOO *-^^^^
(ISCLT3/AG1)
Rochester, Case 6
450
400-
w 350 -
3 300 -
5 250-
S 200-
| 150-
z 100 -
50 -
0 -
,|
OOOO *-^^^^
(ISCLT3/AG1)
2000 -i
1800 -
1600-
oi 1400 -
H 1200 -
° 1000 -
n 800 -
| 600-
z 400-
200 -
0 -
Rochester, All Cases
' -f
1...
OOOO 1-^1-^1-^1-^
(ISCLT3/AG1)
1AII ratios greater than 2.1 are grouped in the last bar.
-------
FIGURE 6
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Syracuse
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
400 -
350-
^ 300-
0 250-
° 200 -
1 15°-
^ 100 -
50-
0 -
Syracuse, Case 1
|
OCN^TCDCOt-CN^rCDCOCN
oooo -^-^-^-^
(ISCLT3/AG1)
450 I
400-
w 350 -
8 SOD-
'S 250-
S 200-
E 150-
z 100-
50 -
0-
Syracuse, Case 2
" B|
OOOO T-T-T-T-
(ISCLT3/AG1)
Syracuse, Case 3
500 --
450 --
400 --
350 --
300 --
250 --
200-
150-
100-
50--
CD CO t-
(ISCLT3/AG1)
Syracuse, Case 4
250
200 --
150 --
100 --
50 --
CD CO CM
(ISCLT3/AG1)
Syracuse, Case 5
500
450 -
w 400-
H 300-
2 250-
5 200-
| 150 -
z 100-
50 -
0 -
OOOO T-T-T-T-
(ISCLT3/AG1)
Syracuse, Case 6
500
450 --
. 40° "
^ 350 --
300-
250-
200 --
150-
100-
50--
CD CO t-
(ISCLT3/AG1)
1600 -
1400 -
^ 1200 -
0 1000 -
° 800 -
| 600 -
z 400 -
200 -
0 -
Syracuse, All Cases
OCN-3-COCOT-CN-3-COCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
-------
FIGURE 7
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Albany
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
1.80 -I
1.60
1.40 -
5 120-
i-
0 1.00
CO
0.80 -
0.60 -
0.40 -
Ring
1 Case
Case
H _ _ ^ A Case
^^ ' * -*-Case
=>-'Case
1 23456789 10
Max. No. of 8 16 24 32 40 48 56 64 72 80
Cells for ea.
Case
1, Toluene
2, Toluene
3, Toluene
4, Mercury
5, Aluminum
6, Mercury
-------
1 RD -,
I .OU |
1 Rn i
1 Ad J
2 120^
CO
r^ 1 oo J
o ' -uu
c/3
0 80
n fin J
0 40 j
Ring
Max. No. of
Cells for ea.
Case
FIGURES
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Buffalo
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
1 Case
^m m m m | * Case
h ^*^^K * I ^^
Jtz*' i -""-""-Case
^
123456789 10
8 16 24 32 40 48 56 64 72 80
1, Toluene
2, Toluene
3, Toluene
4, Nickel
5, Nickel
6, Nickel
-------
1 Rn
1 4D
J< 1 9D
CO
Q 1.00-
c/2
Oon
0 60
0 40
Ring
Max. No. of
Cells for ea.
Case
FIGURE 9
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Rochester
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
"^ . . . ., _ _ -*-
^ ^__^ » V ? V M ^_
^^fK
*^
123456789 10
8 16 24 32 40 48 56 64 72 80
Case 1, Toluene
Case 2, Toluene
Case 3, Toluene
Case 4 Nickel
Case 5, Nickel
O^^P fi Nifkpl
-------
1 RD -,
I . OU
1 Rn
1 4n
^r- 1 9D
CO
I-
~~^ -i nn
O 1.00 -
CO
n Rn
n Rn
n 4n
Ring
Max. No. of
Cells for ea
Case
FIGURE 10
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Syracuse
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
I
|| 1 Case
ij ^^^^^_Qg|gg
A^^ $ tt» » % i 4 ^K^uase
i-^'^r* ' H as
r^
1 23456789 10
8 16 24 32 40 48 56 64 72 80
1, Toluene
2, Toluene
3, Toluene
4, Lead
5, Lead
6, Lead
-------
1.80
FIGURE 11
Average (ISCLT3/AG1) by Ring and Stack Height Bin
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
1.60
^ 1.40
5
1 Om
7m
-------
FIGURE 12
Average (ISCLT3/AG1) by Ring and Chemical Characteristic
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
1.80 -1
1.60
1.40
< 1.20
CO
o 1.00-
c/2
" 0.80
0.60
0.40
Ring
Max. No. of
Cells in ea.
m m - ~"~
1 1 1 t T T 1 1 I
1 23456789 10
96 192 288 384 480 576 672 768 864 960
Chemical
Class
Chemical with
Decay Rate
Chemical
without Decay
Rate
-------
FIGURE 13
Average (ISCLT3/AG1) by Ring and Metropolitan Area
Scenario: Facility-Specific Median Stack Height and Median Exit Gas Velocity
o o o o o
CO CD -sT CN O
(19V/8110S
""" 0.80
0.60
0.40 J
Rings
Max. No. of
,_
I =H=A
^^*** - i i i ^ j j j I 1 H ^^^^~
m~-^M* j 1 i -1 1 -A-R
1 S^
123456789 10
48 96 144 192 240 288 336 384 432 480
Cells1 forea.
Metropolitan Area
1 See Table 10 for actual number of cells per ring in each metropolitan area
bany
jffalo
ochester
/racuse
-------
FIGURE 14
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Albany
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
Albany, Case 1
500 -T-
450 --
400 --
oi 350 --
" 300 --
° 250 -
% 200 --
| 150 --
z 100 --
50 --
0 --
(ISCLT3/AG1)
400 -i
350 -
^ 300 -
0 250 -
° 200 -
| 150 -
£ 100 -
50 -
0 -
Albany, Case 2
OCN^CDCOT-CN^CDCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
Albany, Case 3
500 -i-
450 --
400 --
oi 350 --
" 300 --
° 250 -
% 200 --
| 150 --
z 100 --
50 --
0 --
(ISCLT3/AG1)
500 -i
450 -
400-
oi 350 -
if 300 -
° 250-
% 200 -
| 150 -
z 100 -
50 -
0 -
Albany, Case 4
OCN'^'CDCOT-CN'^rCDCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
500 -,
450 -
400 -
umber of Cell
c^ o 01 o 01
o o o o o
z 100 -
50 -
0 -
Albany, Case 5
OtN^rcDCOT-tN^TCDCOtN
OOOO T-T-T-T-
(ISCLT3/AG1)
Albany, Case 6
450 --
400 --
^ 350 --
$ 300-
5 250 --
200--
150 --
100 --
50 --
0 --
E
(ISCLT3/AG1)
Albany, All Cases
1 400 -y- ^^
1200 -
(/)
^ 1000-
0
5 800-
jj 600 -
E
= 400 -
200 -
o
II
M
l_ | | j j. _| |_H jB |H_|
OCN^CDCOT-CN^CDCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
-------
FIGURE 15
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Buffalo
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
500 -i
450 -
400 -
oi 350 -
" 300 -
2 250-
% 200 -
1 150 -
z 100 -
50 -
0 -
Buffalo, Case 1
OCM^CDCOT-CM^CDCOCM
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 2
500 -r
450 --
400 --
g" 350 - -
< 300 --
P 250 --
O 200 --
22_ 150 --
100 --
50 --
0 --
(ISCLT3/AG1)
Buffalo, Case 3
500 -r
450 --
400 --
5 35° "
g 300 --
P 250 - -
O 200 --
22. 150 --
100 --
50 --
0 --
(ISCLT3/AG1)
Buffalo, Case 4
500
450 --
400 --
oi 350 --
" 300 --
2 250-
^ 200 --
I 150 --
z 100 --
50 --
0
(ISCLT3/AG1)
350 -,
300 -
P 250 -
O
S 200 -
P
o
£. 100 -
50 -
0 -
Buffalo, Case 5
OCN^rcDCOT-CN^rcDCOCN
OOOO T-T-T-T-
(ISCLT3/AG1)
Buffalo, Case 6
500 j
450 --
400 --
5 35° "
g 300 --
P 250 --
O 200 --
22_ 150 --
100 --
50 --
0
(ISCLT3/AG1)
Buffalo, All Cases
1600
1400 --
1200 --
§ 1000 --
P 800 - -
O 600 - -
400 - -
200 --
0 --
Jl
CD CO T-
CD CO CM
(ISCLT3/AG1)
-------
FIGURE 16
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Rochester
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
250 -
« 200-
0)
£ 150-
o
| 100-
z 50-
0 -
Rochester, Case 1
J
\m
orsi^rcocot-rsi^-cocorsi
OOOO T-T-T-T-
(ISCLT3/AG1)
500 -,
450 -
400-
oi 350 -
H 300 -
° 250 -
| 200-
1 150-
z 100-
50-
0-
Rochester, Case 2
OOOO T-T-T-T-
(ISCLT3/AG1)
Rochester, Case 3
500T
450--
400-
oi 350 --
H 300 -
2 250-
£ 200-
I 150-
z 100-
50--
0-
CD CO T-
(ISCLT3/AG1)
Rochester, Case 41
140 -i
120 -
I 100-
0
5 80-
S 60
1 40 -
20 -
0 -
.
I
||
I
ll
Illi.. ..
OOOO -^-^-^-^
(ISCLT3/AG1)
400 -
350-
^ 300-
0)
0 250-
* 20°-
| 150 -
z 100 -
50 -
0-
Rochester, Case 5
OOOO T-T-T-T-
(ISCLT3/AG1)
450-1
400-
ui 350 -
3 300 -
S 250 -
S 200 -
E 150 -
z 100 -
50-
0-
Rochester, Case 6
orsi^rcDco-t-rsi^rcDcorsi
oooo -^-^-^-^
(ISCLT3/AG1)
1600-1
1400 -
^ 1200 -
0 1000 -
° 800-
| 600-
i 400-
200-
0 -
Rochester, All Cases
_
J
||1B
OOOO T-T-T-T-
(ISCLT3/AG1)
1AII ratios greater than 2.1 are grouped in last bar.
-------
FIGURE 17
Frequency Distributions of Concentration Ratios (ISCLT3/AG1) by Case and For All Cases: Syracuse
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
450-1
400-
w 350 -
$ 300 -
5 250-
S 200-
E 150 -
Z 100-
50-
0 -
Syracuse, Case 1
m
OOOO -^-^-^-^
(ISCLT3/AG1)
500
450--
400 --
t/i
= 350--
« 300-
° 250-
| 200 --
I 150--
Z 100-
50 --
400 -
350-
2 300-
0)
0 250-
*200-
| 150 -
Z 100 -
50 -
0-
Syracuse, Case 2
OOOO T-T-T-T-
(ISCLT3/AG1)
Syracuse, Case 3
CD co t-
CD CO CM
(ISCLT3/AG1)
Syracuse, Case 4
450 --
400-
350-
300--
250 --
200 --
150-
100-
50-
0
(ISCLT3/AG1)
450 -
400-
oi 350 -
H 300 -
2 250-
| 200-
1 150-
z 100-
50-
0-
Syracuse, Case 5
OOOO T-T-T-T-
(ISCLT3/AG1)
Syracuse, Case 6
500 --
450 --
400 --
i/i
= 350 --
2 300 --
200 --
150-
100 --
50 --
CD co t-
(ISCLT3/AG1)
2500 -
2000 -
2 1500 -
o
| 1000 -
Z 500-
0-
Syracuse, All Cases
OCN^rCDCOt-CN^J-CDCOCN
O'OOO T-T-T-T-
(ISCLT3/AG1)
-------
FIGURE 18
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Albany
1.80 -I
1.60
1.40 -J
S 1.20-1
co
i 1.00
0.80
0.60
0.40 J
Ring
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
* ~lB II 1*1 [* "J^r A * * '
£^^5 '-- «- - ;-j a | _,_Case
^^ ^ * * * * * * * | -B-Case
^ IL~ 1 1 ' ' ' ' ' | A Case
h Case
X Case
T T T T T i 1 1 1 1
123456789 10 =^ Case
8 16 24 32 40 48 56 64 72 80
Max. No. of
Cells for ea.
Case
1, Toluene
2, Toluene
3, Toluene
4, Mercury
5, Aluminum
6, Mercury
-------
FIGURE 19
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Buffalo
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
1.80 -,
1.60 -
1.40
5 1.20 -
;C
co
0 1.00-
0.80 -
0.60
0.40 -
Ring
1 Case
,: ' ' J i "7 i 1 7 1 1 ' CaSe
1 K ft * * * « | -*-Case
^ t^f JL A A A A A A ' ~ 'Case
*^^
123456789 10
Max. No. of 8 16 24 32 40 48 56 64 72 80
Cells for ea.
Case
1, Toluene
2, Toluene
3, Toluene
4, Nickel
5, Nickel
6, Nickel
-------
FIGURE 20
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Rochester
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
1.80 -|
1 fin
1 40
o 1.20 -
co
O 1 00
W
0.80
n fin
0.40
i 1
>y _
>.
N[-_ A
i t < * ' ' 1.
B-^*"1"" u - U ^ ^ u " x
-j.- JIQ )K iff ^ ^
)K **
>K^^
T T T T T T T T T T
Rin9 123456789 10
Max. No. of 8 16 24 32 40 48 56 64 72 80
Cells for ea.
Case
Case 1, Toluene
Case 2, Toluene
Case 3, Toluene
f^acp 4 Nirkpl
Case 5, Nickel
Case 6, Nickel
-------
FIGURE 21
Average (ISCLT3/AG1) by Ring, Chemical, and Case: Syracuse
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
1.80 -|
1.60
1.40
5 1.20
co
0 1.00
w
0.80
0.60
0.40
Ring
1 Case
A Case
^T ^ ^ jCase
_^"^
123456789 10
Max. No. of 8 16 24 32 40 48 56 64 72 80
Cells for ea.
Case
1, Toluene
2, Toluene
3, Toluene
4, Lead
5, Lead
6, Lead
-------
FIGURE 22
Average (ISCLT3/AG1) by Ring and Stack Height Bin
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
1.80 -i
1.60
CO
O
OD
1.40
1.20
1.00
0.80
0.60
0.40
Ring 1
Max. No. of 64
Cells in ea.
Stack Height Bin
Stack height > 10m
7m
-------
FIGURE 23
Average (ISCLT3/AG1) by Ring and Chemical Characteristic
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
1.80 1
1 RD
1 40
C5 -| on
<£ I -^U
o 1 00
n RD
0 RO
n dn J
Ring
Max. No. of
Cells in ea.
Chemical
Class
HI T T T T T T T T
r"~ n
I I I I I I I I I I
1 23456789 10
96 192 288 384 480 576 672 768 864 960
Chemical with
Decay Rate
Chemical without
Decay Rate
-------
FIGURE 24
Average (ISCLT3/AG1) by Ring and Metropolitan Area
1.80 -,
1.60
1.40
§ 1.20
co
0 1.00 -
w
0.80
0.60
0.40 J
Scenario: SIC Code Based Median Stack Height and Median Exit Gas Velocity
^r-^_. '< § A
^ ± A A A A 4 4 4 1 , A
^p4 f * * 1 * * t -^B
t^^r^^T -, . . , i i | _a_R
W^ |
1 23456789 10
Rin9s 48 96 144 192 240 288 336 384 432 480
Max. No. of
Cells1 forea.
Metropolitan Area
1 See Table 15 for actual number of cells per ring in each metropolitan area
bany
jffalo
ochester
/racuse
-------
FIGURE 25
Difference in Median Stack Height
(SIC Code Based Stack Height Minus Facility-Specific Stack Height)
350 i
300
250
o>
co
8 200
£ 150
|
z 100
50
0
n n n n H
,- E3 -,-13-,-.., ,
O
CO
CD
CN
CN
CN
CD
CD
CO
CN
CN
CD
CN
O
CO
Difference (m)
-------
160
140
120
$ 100
8
P 80
60
40
20
0
CD
.a
E
FIGURE 26
Difference in Median Exit Gas Velocity
(SIC Code Based Exit Gas Velocity Minus Facility-Specific Exit Gas Velocity)
nHO
CO
o
CO
IT)
CN
O
CN
IT)
IT)
O
CN
Difference (m/s)
-------
FIGURE 27
Indicator Sub-element1 Contributions and Concentration Ratios (ISCLT3/AG1)2
by Ring and Case: Albany
Albany, Case 1
40% -r
_ c 30% --
.£ o
er=
I 5 20% -
o c
a! <3 10% -
o% 4
-r 1.80
1.60
-1.40 o
1.20 fi .°
CO +-"
j , , i j 1 nn "-; $
^jf i i i i i i i i
illlll...
0.80 GO
0.60
0.40
12345678910
Ring
Albany, Case 2
40% j
°> c 30% -
.£ o
II 20%-
" §
I =§ 20% -
0 c
OJ °
a. 0 10% -
0% -
S^I__.1_LJ
illlll...
1.80
1.60
- 1 .40 5
- 1 20 ^ .°
CO <-
- 1 oo !m (2
0.80 w
0.60
0.40
12345678910
Ring
Albany, Case 5
40% j
°> c 30% -
.b O
c | 20% -
8 £
o3 o
o! o 10% -
o% 4
II
Ill 1
III 1
r 1.80
1.60
- 1 .40 5
- 1 .20 ^ o
- 1.0013 &
0.80 GO
0.60
0.40
12345678910
Ring
Albany, Case 6
40% -r
o> c 30% -
.£ o
_,_, ^
c ^ 20% -
Q! <3 10% -
0% -
lllllllll
1.80
1.60
- 1 .40 5
1 .20 ^ .2
j «« i
-------
FIGURE 28
Indicator Sub-element1 Contributions and Concentration Ratios (ISCLT3/AG1)2
by Ring and Case: Buffalo
Buffalo, Case 1
40% -r
ra ,- 30% -
^ ^
^ "-P
c 5 20% -
o c
^ ° 1 0% -
n%
ill, , ,
I '
him
1 2345678910
r 1.80
1.60
- 1.40 5
1 20 ^ o
CO '*-
-i nn *""! n5
O
0.80 W
0.60
n 4D
Ring
Buffalo, Case 2
40% -r
ra c 30% -
.c 0
3
c -° 20% -
-------
FIGURE 29
Indicator Sub-element1 Contributions and Concentration Ratios (ISCLT3/AG1)2
by Ring and Case: Rochester
40% Rochester, Case 1 -| 30
0%
12345678910
Ring
ra c
.£ o
Di'-g
c I
40% -r Rochester, Case 4 T 1 80
-I 1.60
30% 4- X
20% --
10% +
1 2345678910
Ring
.E o
40% Rochester, Case 2 -| QQ
1.60
1.40
ra ,- 30% --
5 20% --
10%
0%
llllJllll
O
1.20 S: o
CO i-'
1.00 j o:
o
0.80 w
0.60
0.40
23456789 10
Ring
40% -r Rochester, Case 5 T
12345678910
Ring
40% -r
Rochester, Case 3
2345678910
Ring
40%
D) c 30%
.g o
c | 20%
CD j^
o c
°- ° 10%
0%
Rochester, Case 6
Jlllllll
1.60
1 40 "-
O
1.20 ^
CO
1.00 j
O
0.80 52
0.60
0.40
123456789 10
Ring
Indicator Sub-elements (percent) shown as histogram and can be read on the left vertical axis. Indicator
Sub-elements are the product of pollutant concentration and population in each cell, summed over all cells
in a ring. They reflect percent contribution to Indicator Elements by ring (e.g., for case 1, ring 1
contributes 35% to the Indicator Element and ring 10 contributes 4%).
Concentration ratios (ISCLT3/AG1) are shown as a line and can be read on the right vertical axis (e.g.,
for case 1, ring 1, the ratio is 1.18, and for ring 10, the ratio is 0.97).
-------
FIGURE 30
Indicator Sub-element1 Contributions and Concentration Ratios (ISCLT3/AG1)2
by Ring and Case: Syracuse
40% -r
ra c 30% --
- .2
I == 20% -
" 0
^ ° 10% -
n%
Syracuse, Case 1
, , i i i i i
t '
1 2345678910
Ring
1.80
1.60
-1.405
1-20^.2
1.0013(2
O
0.80 w
0.60
0.40
40% -,
D) c 30% -
- .2
I == 20% -
" 1
^ ° 10% -
no/
UYO -f
40%
0) c 30%
.£ o
£ =
I =§ 20%
0 "c
OJ 0
°- ° 10%
0%
Syracuse, Case 2
i i ± i a
ihlll
12345678910
Ring
Syracuse, Case 3
lll.lih
1.80
1.60
-1.405
-1-20^.2
Hcc
ry
(o
0.80 w
0.60
n >i n
U.4U
T- 1.80
1.60
-1.405
1.20^.0
- 1 oo !m K
o
0.80 w
0.60
n 4n
12345678910
Ring
40% -r
ra c 30% --
.£ o
S?
I =§ 20% --
£ o
^ ° 10% --
n%
Syracuse, Case 4
tr.
2345678910
Ring
1.80
1.60
1.405
1 -2° ^ ~
1.00 j a!
O
0.80 w
0.60
0.40
40% -,
D) c 30% -
.£ o
Di'-g
g | 20% -
g §
^ ° 10% -
n%
40% -
D) c 30% -
- .2
c | 20% -
CD j^
fc o
°- U 10% -
U%
Syracuse, Case 5
' 1 1 |4i 1 1 1 l l l
..llllllll
1 2345678910
Ring
Syracuse, Case 6
r 1.80
1.60
-1.405
-1-20 §5
1 nn '""i rv
(o
0.80 co
0.60
n 4n
T- 1.80
1.60
1.405
"1-20g.5
Hcc
, , ,, , (V
o
0.80 w
0.60
1 2345678910
Ring
Indicator Sub-elements (percent) shown as histogram and can be read on the left vertical axis. Indicator
Sub-elements are the product of pollutant concentration and population in each cell, summed over all cells
in a ring. They reflect percent contribution to Indicator Elements by ring (e.g., for case 1, ring 1
contributes 35% to the Indicator Element and ring 10 contributes 4%).
Concentration ratios (ISCLT3/AG1) are shown as a line and can be read on the right vertical axis (e.g.,
for case 1, ring 1, the ration is 0.76, and for ring 10, the ratio is 0.92).
------- |