vvEPA
PROPEK
DIV; •
OF
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-82-013r
October 1982
Air
METEOROLOGY
NORTHEAST CORRIDOR
REGIONAL MODELING
PROJECT
ANNUAL EMISSION
INVENTORY COMPILATION
AND
FORMATTING
Volume XVIII:
Inventory Review And Evaluatior
-------
EPA-450/4-82-013r
NORTHEAST CORRIDOR REGIONAL
MODELING PROJECT
ANNUAL EMISSION INVENTORY
COMPILATION AND FORMATTING
Volume XVIII:
Inventory Review And Evaluation
by
GCA Corporation
Bedford, MA
Contract No. 68-02-3510
EPA Project Officers: James H. Southerland
Thomas F. Lahre
Prepared For
U.S. Environmental Protection Agency
Office of Air, Noise and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
October 1982
-------
This report has been reviewed by the Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, and approved for publication as received from GCA Corporation, Bedford, MA. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S. Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation for use. Copies of this report are
available from the Air Management Technology Branch, Monitoring and Data Analysis Division, Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711.
-------
CONTENTS
Figures iv
Tables v
Executive Summary viii
1. Introduction 1
Background 1
Objectives 1
Agency contacts 3
Report organization 4
2. Emissions Summary 7
Introduction .... 7
Data procurement 7
Emissions summary 7
3. Evaluation Techniques 19
Introduction 19
SIP compatibility review 19
NEDS compatibility review 20
Emission densities 26
Per capita emissions from selected area source
categories 28
Point source inventory comprehensiveness 50
Area source inventory comprehensiveness 52
4. Data Quality Review 57
Introduction 57
Connecticut 57
Delaware 61
Maine 63
Maryland 65
Massachusetts 69
New Hampshire 73
New Jersey 75
New York 79
Ohio 83
Pennsylvania 91
Rhode Island 95
Vermont 98
Virginia 100
Washington, D.C 103
West Virginia 109
Summary 112
5. References 113
111
-------
FIGURES
Number Page
1 Boundaries of regional model grid system ... 2
2 VOC point and area source emissions densities . 30
3 NOX point and area source emission densities ... 31
4 VOC point and area source emissions per area 32
5 NOx point and area source emissions per area 33
6 VOC area source emission densities 34
7 NOx area source emission densities 35
8 VOC point source emission densities 36
9 NOX point source emission densities 37
10 VOC area source emissions per area 38
II NOx area source emissions per area 39
12 VOC point source emissions per area 40
13 NOx point source emissions per area 41
14 Massachusetts air pollution control districts and counties . 55
-------
TABLES
Number Page
1 Agency Contacts for the NECRMP Annual Emission Inventory . . 5
2 Directory of the NECRMP Annual Regional Emission Inventory
Reports 6
3 Summary of Point Source Data Procurement . 8
4 Summary of Area Source Inventory Data Sources 9
5 Statewide Totals of Area and Point Source Emissions of
VOC and NOX for Northeast Corridor Study Region 10
6 NECRMP Point Source Emissions by Reporting Category .... 12
7 Major Categories of Oxidant Precursor Emissions from
Point Sources in the NECRMP Region 15
8 NECRMP Area Source Emissions by Category 16
9 Major Categories of Oxidant Precursor Emissions from
Area Sources in the NECRMP Region 18
iO Annual Emission (Tons) of Volatile Organic Compounds (VOC)
for Selected SIP Areas 22
11 Annual Emissions (Tons) of Nitrogen Oxides (NOx) for
Selected SIP Areas 23
12 Comparison of NEDS and NECRMP Emissions Estimates for
Volatile Organic Compounds, by State 24
13 Comparison of NEDS and NECRMP Emissions Estimates for
Nitrogen Oxides, by State 25
14 Electric Utility Emissions in New York 26
15 NECRMP Study Area VOC and NOX Emissions, by State, as a
Function of Population (Within Study Area Only) 27
-------
TABLES (continued)
Number Page
16 NECRMP Study Area VOC and NOx Emissions as a Function
of Land Area (Within Study Area Only) .......... 29
17 Per-Capita Emissions from Residential Fuel Combustion ... 43
18 NEDS vs. NECRMP Residential Emissions for Massachusetts
and Vermont ....................... 42
19 Per Capita Emissions for Highway Vehicles ......... 45
20 Average Light-Duty Vehicle Emission Factors, by State ... 47
21 VOC/NOX Ratios for Highway Vehicle Emissions ........ 48
22 Per Capita Emissions for Gasoline Handling (Point and
Area Sources) ...................... 49
23 Numbers of Facilities and Individual Point Sources
Covered in the NECRMP Inventories ............ 51
24 Coverage of "RACT" Sources in the NECRMP Data Base ..... 53
25 Area Source Categories Inventoried for the NECRMP Study . . 56
26 Comparison of NEDS, SIP and NECRMP Emissions for
Connecticut ....................... 60
27 Comparison of NEDS, SIP and NECRMP Emissions for
Delaware ......................... 62
28 Comparison of NEDS and NECRMP Emissions for the NECRMP
Study Area Portion of Maine ............... 64
29 Comparison of NEDS, SIP, and NECRMP Emissions for
Maryland ......................... 68
30 Comparison of NEDS, SIP, and NECRMP Emissions Estimates
for Massachusetts .................... 72
3i Comparison of NEDS and NECRMP Emissions for
New Hampshire ...................... 74
32 Comparison of the NEDS and NECRMP Inventories for
New Jersey ........................ 78
-------
TABLES (continued)
Number Page
33 Comparison of the NEDS, SIP, and NECRMP Inventories for
New York 82
34 Comparison of the NEDS, SIP, and NECRMP Inventories for
the Ohio Portion of the NECRMP Study Area 85
35 Comparison of GCA and NOACA Area Source Emission
Calculations for the Four Cleveland Area Counties .... 86
36 Data Sources for Commercial/Institutional and Industrial
Fuel Categories Used in the NECRMP and NOACA Annual
Emission Inventories 90
37 Comparison of NEDS, SIP, and NECRMP Inventories for
Pennsylvania 94
38 Comparison of NEDS, SIP, and NECRMP Inventories for
Rhode Island 97
39 Comparison of NEDS and NECRMP Inventories for Vermont ... 99
40 Comparison of the NEDS and NECRMP Inventories for
Virginia 102
41 Comparison of the NEDS, SIP, and NECRMP Inventories for
Washington, D.C 106
42 Major VOC Discrepancies in D.C. Area, SIP Versus NECRMP . . 107
43 Major NOx Discrepancies in D.C. Area, SIP Versus NECRMP . . 108
44 Comparison of NEDS and NECRMP Inventories for the
West Virignia Portion of the NECRMP Study Area Ill
VII
-------
EXECUTIVE SUMMARY
INTRODUCTION
On a nationwide basis, nonattainraent of the National Ambient Air Quality
Standards (NAAQS) for ozone is one of the most serious and widespread air
pollution problems facing the air quality management community. The Northeast
Corridor, a megalopolis of urban and suburban areas extending from Washington,
D.C. to Boston, bears an especially large portion of this ozone nonattainraent
problem. The United States Environmental Protection Agency (USEPA), in
cooperation with the northeastern states, local agencies, and Metropolitan
Planning Organizations (MPOs), has undertaken the Northeast Corridor Regional
Modeling Project (NECRMP) to develop regional and urban ozone control
strategies through the use of photochemical air quality simulation models.
To employ a regional model, an inventory of point and area source
emissions covering the entire NECRMP study area had to be assembled and placed
into a common format. Unfortunately, exiting data bases were inadequate to
either properly test or validate a regional model. To this end, USEPA1s
Office of Air Quality Planning and Standards retained GCA/Technology Division
to complete an annual inventory for use in NECRMP. The study area, shown in
Figure 1, includes the entire northeast quadrant of the United States from
longitude 69 degrees to 82 degrees West and latitude 38 degrees to 45 degrees
North, plus the Ohio Counties of Franklin, Licking, Perry, and Fairfield. To
accommodate these additional counties, the NECRMP grid system extends West to
84 degrees longitude.
OBJECTIVES
The objective of the effort reported in this volume and summarized here
was to compile and review the entire NECRMP emission inventory data base,
evaluate its quality and comprehensiveness, and develop corrective measures in
coordination with each state. Particular emphasis was placed on the
suitability of the data for further temporal spatial, and species resolution
for subsequent use in a photochemical model of the entire 15 state* area.
This inventory review volume provides a summary of point and area source
emission data from each state, including a discussion of their coverage and
completeness and their agreement with 1982 ozone State Implementation
Plan (SIP) data bases, and other inventory totals in use. Also included
*For the purposes of this report, the District of Columbia is treated as a
state equivalent.
viii
-------
i.s a description of ongoing data collection efforts in each state, and the
current and future availability of additional data from individual state
agenc ies.
The major pollutants of interest in the NECRMP study were Volatile
Organic Compounds (VOC) and Oxides of Nitrogen (NOX) with carbon monoxide
(CO) being included as a potentially important tracer. Particulate (TSP) and
Oxides of Sulfur (SOX) emissions were also included as available, but
quality assurance efforts focused on VOC and NOX. Regarding the
completeness and accuracy of data, the greatest emphasis was placed on point
sources with emissions of VOC greater than 500 tons/year and NOX above 750
tons/year, as these are generally found to be responsible for a large portion
of the total point source emissions and are likely to have a distinct and
identifiable impact on photochemical modeling results. Sources exceeding
these cutoffs were considered "major" and will be handled as individual point
sources by the Regional Model Data Handling System (RMDHS), although the
NECRMP point source inventory generally includes all available data on point
sources regardless of size. Point sources with emissions below the cutoff
levels will be treated as area sources by the RMDHS system, their spatial
allocation to grid cells based on UTM coordinates.
EMISSIONS SUMMARY
A summary of the NECRMP emissions data is provided in Section 2.
Separate tables summarize emission totals for point and area sources by state
and by type of emission category. Emissions for the entire NECRMP study area
are as follows.
VOC NOX
(tons/year) (tons/year)
Area 2,981,759 2,795,288
Point 1,041,043 3.480,344
Total 4,022,802 6,275,632
EVALUATION TECHNIQUES AND DATA QUALITY REVIEW
Section 3 describes the techniques used to evaluate the quality of the
NECRMP emissions data. It includes a discussion of how the NECRMP data were
compared with inventories filed by each state in support of State
Implementation Plan submittals and as well as with data contained in the
National Emission Data System (NEDS). Also provided in Section 3 are tabular
and graphic presentations of emission densities of VOC and NOX expressed
both in terms of population and land area. VOC/NOX ratios are likewise
presented for a number of emission categories by state to identify possible
anomalies.
ix
-------
SIP Compatibility Review
All available 1982 ozone State Implementation Plans for ozone
nonattainment areas within the NECRMP region were reviewed. In most
instances, the NECRMP emission data were obtained directly from the same state
agencies who were simultaneously developing the SIP inventories. Each state
received special funding for NECRMP liaison activities, in addition to their
regular program grants, and were required to keep GCA informed of any
significant changes made to their emission inventories. In theory, the SIP
and NECRMP inventories should agree for all ozone nonattainment areas within
the study region; however, this is not always true. In some instances,
certain states made significant changes to their SIP data either without
notifying GCA or after the respective NECRMP inventory had been completed. In
fact, as of this writing, some states are still actively modifying SIP
inventory submittals. In other cases, entire area source categories were
omitted by some states, and to maintain uniformity and completeness across the
entire NECRMP region GCA developed its own emission estimates in these missing
categories. Some states supplied GCA with computerized point source
inventories which bore little or no resemblance to the point source data
reported in the hard-copy SIP inventories.
While there is only one "correct" inventory for a given area, use of
different methods, assumptions, and data sources can yield greatly differing
results. To avoid inter-state biases introduced as a result of different
inventory methods, the NECRMP inventories were compiled and verified using a
single standard set of methods. Deviations from the standard methods or
emission factors were brought to the attention of the appropriate state agency
for reconsideration. While one objective of NECRMP was to ensure consistency
from state to state, it was also an objective to achieve consistency between
the SIP and NECRMP inventories for individual states. There were, of course,
instances when these two priorities came into conflict. These situations were
resolved case-by-case individually, with deviations from the standard methods
documented in the individual state volumes.2-16
Because of their differing objectives, the SIP and NECRMP inventories are
reported differently. The SIP requirements entailed reporting (hard copy)
reactive VOC and NOX emissions on a daily basis for a typical summer weekday
to support SIP-required attainment demonstration (e.g., using EKMA). The
NECRMP inventory required annual emissions reflecting total VOC and NOX in
computerized form appropriate for further temporal, spatial, and reactivity
resolution in conjunction with EPA's Eulerian Regional Oxidant Model (ROM).
Thus, direct comparison between the SIP and NECRMP inventories was not
possible within the time and resource limitations of this project. However, a
limitted comparison was attempted by converting the SIP-reported emissions
from kg/day to tons/year using a conversion factor based on 365 days/year. An
exact match would not be expected since the SIP inventory totals reflect
emissions on a typical summer weekday rather than an average annual basis, and
because they include only reactive VOC emissions rather than total VOC.
However, if adjusted totals were within approximately 10 percent of each
-------
other, the SIP and NECRMP inventories were considered to "match." All known
differences between the two are identified and evaluated in state-by-state
review of data quality found in Section 4.
NEDS Compatibility Review
States are required annually to supply updated emissions data to EPA's
National Emission Data System (NEDS), which is the national data repository
upon which most air quality analyses are based. For a number of reasons, not
all states have kept their NEDS files sufficiently detailed and up-to-date for
use in modeling analyses as complex as NECRMP. Still, since the NECRMP, NEDS,
and SIP inventories are all based on state supplied data, theoretically all
three data bases should agree. However, since at EPA's urging each of the
states in the study area provided assistance and assigned high priority to
GCA's questions regarding compilation and quality assurance of the data base,
it ia resonable to assume that overall the NECRMP inventory represents the
most suitable emissions data for the intended application. Unfortunately, the
degree of cooperation extended by the various states in the NECRMP region was
not uniformly high. Thus, for certain individual sources the data contained
in NEDS nuiy still be more representative of the actual level of emissions. It
ia useful, then, to compare data contained in NEDS with the SIP and NECRMP
data to identify areas of disagreement for future resolution.
Section 3 presents a comparison, by SIP area and by state, of the NEDS,
SIP, and NECRMP inventories. Regionwide totals are compared below for those
L3 SIP areas which fall within the NECRMP study boundaries:
VOC (tons/year)
NOX (tons/year)
i _> air
area totals
Point
Area
NEDS
850,548
2,563,302
SIP
505,741
1,763,874
NECRMP
590,672
1,675,771
NEDS
789,584
1,229,828
SIP
974,144
1,340,214
NECRMP
1,339,483
1,520,048
Total
3,413,850 2,269,615 2,266,443 2,019,412 2,314,358 2,859,567
The NECRMP VOC emission estimates generally agree with the SIP totals for
the 13 SIP areas in the study region. The totals for all 13 SIP areas for
NECRMP are within about 0.1 percent of the reported SIP values. The NEDS
emission estimates for the same areas are 47 percent higher than those
reported in the SIPs.
The NECRMP emission estimate for NOX, across all 13 SIP areas, is
approximately 25 percent higher than that reported in the SIPs. On the other
hand the NEDS NOX totals are approximately 13 percent lower than the
corresponding figure compiled from the SIPs. The bulk of this latter
difference can be attributed to the NECRMP point source inventory for New
xi
-------
York. As discussed in greater detail below, and again in Section 4, the
reported NECRMP electric utility emissions for New York are higher than those
reported in NEDS by a factor of 7. The utility emissions reported in NEDS are
believed to be more accurate than those reported in NECRMP. Further, GCA
recommends that the NEDS estimates for utilities be substituted for the NECRMP
estimates in New York. With this adjustment, the NECRMP NOX total for the
13 SIP areas would agree with the SIP totals within less than 3 percent.
VOC (tons/year)
NOX (tons/year)
Entire NECRMP Region
NEDS
NECRMP
NEDS
NECRMP
Point
Area
1,
4,
072
367
,643
,962
1
2
,041
,981
,043
,759
2,
2,
258,
299,
689
300
3
2
,480,344
,795,288
Total
5,440,605 4,022,802 4,557,989 6,275,632
For area sources, the NEDS VOC estimates are consistently much higher
than those provided in NECRMP, although there are a couple of exceptions.
Nearly all of the difference in area source VOC emissions can be attributed to
evaporative VOC losses from organic solvent use. This difference results
primarily from the fact that GCA and the states utilized methodologies
consistent with EPA Guidelines,19 which do not consider all of the
identified industrial end uses of organic solvent (given in Table 4 of End Use
of Solvents Containing VOC (EPA-450/3-79-032))20 as potentially being area
sources, while NEDS does. In addition, when end use cannot be identified the
organic solvent is all treated as area source emissions in NEDS. The
assumption implicit in the Guidelines is that all industry end users
classified as "other" or "unidentified" will be completely accounted for in
the point source inventory. Thus, on a regionwide basis, it would be expected
that NECRMP point source VOC emissions would exceed NEDS point source totals.
As is shown in Section 3, this is clearly not the case. Either the
methodology inherent in the Guidelines underestimates emissions from organic
solvent use, or NEDS includes emissions estimates from sources that do not
really exist. Because of the magnitude of this discrepancy, GCA recommends
that the organic solvent data contained in End Use of Solvents Containing VOC
be reexamined with respect to how the EPA Guidelines interpret these data.
Regionwide the NEDS and NECRMP NOX emissions differ by over
1.7 million tons/year, with most of this difference atributable to point
sources. Specifically, over 1.1 million tons can be accounted for by electric
generation sources in New York. Since the NEDS data on electric generation
xii
-------
are updated annually baaed on DOE data (confirmed by E. H. Pechan estimates of
S02 from electric utilities), independent of the states, it is likely that
the NEDS estimates in this case are correct and should be substituted for
those in NECRMP. The NEDS and NECRMP estimates for electric utility emission
in New York are discussed further in Sections 3 and 4.
Emission Densities
Emission densities for VOC and NOX were calculated for each state or
region within the NECRMP study area, both in terms of emissions per capita and
emissions per square mile. This analysis was performed primarily to identify
potential anomalies in the emissions data. To identify "suspicious" values,
the mean and standard deviation were calculated for VOC and NOX for point,
area, and total emissions. States whose per-capita emissions fell more than
two standard deviations from the mean were singled out for further
investigation. All state area source per-capita emissions estimates were
found to fall within two standard deviations for the mean, indicating that no
individual state values fall outside the range of values expected for this
type of aource.
One atate value, for Delaware, falls more than two standard deviations
(2.57) from the per-capita point source VOC mean. For point source NOX
emissions, only one value (for West Virginia) falls outside of the expected
range. These state values are discussed in greater detail in Section 4.
Emission densities for each state were calculated as a function of land
area, as were population densities. It would be expected that population
den.siti.es and emission densities would correlate well. Regression analyses
for VOC and NOX emission densities versus population densities were
performed, and as predicted, population and emission densities showed a very
strong correlation (r2 >0.98) for both VOC and NOX emissions.
Some area source categories (e.g., commercial/consumer solvent use, dry
cleaning, etc.) are estimated based on population and a per-capita emission
factor. Others are derived as a function of an activity rate (e.g., gasoline
saLen, vehicle miles of travel, etc.). One evaluation measure applied to the
NECRMP area source inventory data was to calculate emissions on a per-capita
basis for certain area source categories that were not originally derived
using per-capita emission factors. The results of this analysis are presented
in Section 3.
PointSource Inventory Comprehensiveness
Comprehensiveness of the point source emissions data was evaluted for
each state by comparing facilities included in each state's inventory against
chose listed in the Directory of Volatile Organic Compound Sources Subject to
Reasonably Available Control Technology Requirements (RACT Directory). This
comparison was actually performed as a quality assurance measure during the
point source compilation phase of NECRMP. The intention of this check was to
! '.'uriry .md subsequently include VOC sources that had been omitted in the
xiii
-------
individual state data bases. For the entire NECRMP study 548 sources were
identified as being included in the RACT Directory but not in the
state-supplied data. This exercise resulted in the addition of six facilities
to the NECRMP data base (one in Rhode Island and five in Massachusetts)
totaling over 9,000 tons of VOC. A state-by-state comparison of RACT coverage
is presented in Section 3.
Area Source Inventory Comprehensiveness
Area source inventory comprehensiveness was evaluated in terms of
geographic coverage, area source category coverage, and resolution of area
source categories. Area source emission inventories were compiled for all
areas within the NECRMP study boundaries. Some variations were found to exist
between individual state components in terms of all three dimensions as
discussed below in Section 3.
In certain instances, state agencies determined that some of the less
significant categories did not contribute sufficient VOC and NO,, emissions
f A
to warrant their inclusion. In other cases emissions representing a number of
the NECRMP categories were aggregated to form composite categories. For this
latter situation, one of two courses of action were taken by GCA. When
sufficient data were available from the states, or when national data were
adequate to allow their independent disaggregation, GCA calculated
category-specific estimates from the composite data. When such information
was unavailable, GCA included the composite data in that state's EIS/AS master
file and properly "commented" the composite category so as to identify its
component constituents. These adjustments, unique to each state, are
discussed individually in Section 4.
Point Source Inventory Quality
In general the NECRMP point source files are suitable for use in
regional modeling, although some significant problems remain. Ongoing efforts
in two states, New Jersey and West Virginia, should eventually result in
significant improvements to the existing data. In New Jersey, the DEP is
currently updating their APEDS file to reflect important changes in emissions
estimates for a number of sources which resulted from field office reviews
conducted in preparation for SIP revision. The DEP also recently addressed
"bugs" in the conversion routine which had resulted in numerous conversion
errors. When completed, the revised New Jersey data should be substituted for
the point source file currently contained in the NECRMP annual inventory.
West Virginia's APCC is developing a new comprehensive point source
inventory. Since the data currently in the NECRMP point source file for West
Virginia are at least 7 years old, it should be used only as an interim file
until the new state data are made available.
Three states, (Massachusetts, New York, Pennsylvania) and the District
of Columbia, did not directly respond to GCA's specific QA question/problem
xiv
-------
lists. Massachusetts and New York did supply revised NEDS tapes and
additional data which addressed some of the more serious point source
problems, although difficulties remain in both files, as described in
Section 4. Except for "cleaning up" data errors that prohibited conversion to
EIS/PS format, the Pennsylvania and District of Columbia files reflect data
"as received."
The Maryland BAQNC has reinventoried two major facilities since
completion of the NECRMP inventory for that state, namely Eastern Stainless
and Bethlehem Steel. These revised data should replace the current NECRMP
information for the two facilities.
By far the largest source of problems in the NECRMP inventory relates to
data conversion. Since a regional scale inventory necessitates use of a
common data format, all of the inventories had to undergo one or more data
conversions. Most inventories had to be converted twice, once from the
state's own unique systems to a common system and again to the system being
used in the project. This double conversion was required since states with
their own systems usually only had software to convert to one "common" system
and this varied between NEDS, EIS/P&R, and EIS/PS. These conversions were
expensive, often involving hundreds of dollars in computer time. Because the
inventories had been inadequately checked for transcription and keypunching
errors and the AEROS instructions were so often misinterpreted, the error
Listings resulting from the conversions were massive. Errors had to be
prioritized before resolution despite the allocation of significant resources
toward this end. For many states, the allocated resources were completely
expended in identifying and resolving data "glitches" which prohibited
conversion into EIS/PS. As a result, it was not always possible to verify
that all of the states' emissions estimates were accurate.
It is unfortunate that this situation exists. A properly coded NEDS
file will translate into EIS/PS format with very few if any problems. The
NEDS system has been in existence long enough so that state agencies should be
able to produce properly formatted "NEDS" data. However, many of the
state-specific inventory systems are apparently not sufficiently compatible to
allow complete and accurate conversion into NEDS format. Other systems which
should be compatible appear to have a number of "bugs" in their conversion
software. In still other instances, it appears that certain states have
either misinterpreted or intentionally ignored the AEROS manual instructions.
The above problems have been addressed and the NECRMP point source
inventory contains commonly formatted data for the entire study region. In
most instances, these data were obtained directly from state air agencies and
presumably represent the individual states' most current and accurate data.
Thus, as a collective set, the NECRMP point source inventory should represent
the most suitable data base available for the intended application. However,
aa described in this volume, there are some residual problems with the data.
Since the data exists in a single commonly formatted, easily updatable data
xv
-------
base, these residual deficiencies should not be irreparable. Rather, the
NECRMP point source inventory can be utilized as a continuously maintainable
"working" file of point source data independent from NEDS and, thus, not bound
by the same requirements. However, if the states follow through and update
NEDS to reflect the data supplied for NECRMP, the two data bases should
eventually be compatible.
Area Source Inventory Quality
In general, the NECRMP area source inventories are suitable for further
use in NECRMP modeling, although some problems exist. The most significant
residual problem with the area source inventories is the use of composite
categories. Since the temporal and organic species allocations are
category-specific, these will adversely affect the NECRMP modeling.
Disaggregated data were sought from the appropriate agencies in all
instances. Should this information be made available in the future, the
NECRMP area source files should be appropriately updated. Similarly, the
interim file for West Virginia should be superseded by West Virginia's 1980
area source inventory when it becomes available.
A number of differences between NECRMP and NEDS have been noted above.
The largest single source of difference appears to relate to differing
interpretations of solvent sales data used by OAQPS to develop the emission
factors in Procedures for the Preparation of Emissions Inventories for
Volatila Organic Compounds—Volume I, Second Edition1^ and by NADB to
estimate emissions in NEDS. These differences should be resolved internally
within EPA and the NEDS and/or NECRMP VOC estimates adjusted accordingly.
Many other differences between NEDS and NECRMP will be resolved if the states
file updates to NEDS, to reflect the data supplied for the SIPs and NECRMP.
SUMMARY
An emission inventory is only a compilation of estimates and as such
does not allow one to compare actual versus predicted values. Nevertheless,
if proven and standard procedures are followed and the resulting data are
subjected to logical quality assurance and validation checks, one can develop
reasonable assurance that an inventory data base is reliable for the purposes
intended.
In general, GCA believes that the NECRMP emission inventory is complete
as is practicable given the resources available and the current distribution
of responsibilities between EPA and the individual states. Known problems
with data from specific states are discussed in Section 4. Undoubtedly
additional errors remain and some sources may be omitted. Since the data base
is commonly formatted and capable of continuous updates using existing EPA
software, it is likely that some additional corrections may be warranted in
the future. For example, through the modeling process itself, peculiar
results will often surface which result in the identification of a missing
source, or erroneous information on an included source. Additionally, some
states are still working on revisions to their ozone SIP inventories,
presumably reflecting newer, better data. These updates should also
eventually be reflected in NECRMP.
xvi
-------
It should be reiterated that an emission inventory can never be proven
precise and accurate. The NECSMP inventory however, has undergone rigorous
checks and validation efforts and we believe it to be capable of driving the
ROM. Any errors or inconsistence identified can be subsequently corrected and
available new data incorporated.
xvix
-------
SECTION 1
INTRODUCTION
BACKGROUND
On a nationwide basis, nonattainment of the National Ambient Air Quality
Standards (NAAQS) for ozone is one of the most serious and widespread air
pollution problems facing the air quality management community. The Northeast
Corridor, a megalopolis of urban and suburban areas extending from Washington,
D.C. to Boston, bears a large extent of the ozone problem. The United States
Environmental Protection Agency (USEPA), in cooperation with the northeastern
states, local agencies, and Metropolitan Planning Organizations (MPOs), has
undertaken the Northeast Corridor Regional Modeling Project (NECRMP) to
develop regional and urban ozone control strategies through the use of
photochemical air quality simulation models.
To employ a regional model, an inventory of point and area source
emissions covering the entire NECRMP study area had to be assembled and placed
into a common format. Unfortunately, existing data bases were inadequate to
cither properly test or validate a regional model. To this end, USEPA1s
Office of Air Quality Planning and Standards retained GCA/Technology Division
to complete an annual inventory for use in NECRMP. The study area, shown in
Figure 1, includes the entire northeast quadrant of the United States from
longitude 69 degrees to 82 degrees West and latitude 38 degrees to 45 degrees
North, plus the Ohio Counties of Franklin, Licking, Perry, and Fairfield. To
accommodate these additional counties, the NECRMP grid system extends West to
84 degrees longitude.
OBJECTIVES
The objective of the effort reported in this volume was to compile and
review the entire NECRMP emission inventory data base, evaluate its quality
and comprehensiveness, and develop corrective measures in coordination with
each state. Particular emphasis is placed on the suitability of the data for
further temporal, spatial, and species resolution for subsequent use in a
photochemical model of the entire 15 state* area. This inventory review
document includes a summarization of point and area source emissions from each
state and includes discussion of coverage and completeness of area sources,
agreement with 1982 ozone State Implementation Plan (SIP) data bases,
*For the purposes of this report, the District of Columbia is treated as a
state equivalent.
-------
§
u
CO
>*
(0
•o
•H
(J
00
0)
•a
o
e
o
CO
HI
CO
OJ
T3
c
o
CO
00
-------
comprehensiveness and adequacy of point source data, and agreement with or
deviation from other inventory totals in use. Included is a description of
residual or ongoing data collection efforts occuring in each state, and the
current and future availability of additional data from individual state
agencies.
The major pollutants of interest were Volatile Organic Compounds (VOC)
and Oxides of Nitrogen (NOX) with carbon monoxide (CO) being included as a
potentially important tracer. Particulate (TSP) and Oxides of Sulfur (SOX)
emissions were also included as available, but quality assurance efforts
focused on VOC and NOX. Thus, the quality of TSP, CO and SOX emission
data is generally undetermined. A higher emphasis for completeness and
accuracy of data was placed on point sources with emissions of VOC greater
than 500 tons/year and NOX above 750 tons/year, as they are generally found
to be responsible for a large portion of the total point source emissions and
are likely to have a distinct and isolatable impact on photochemical modeling
results. Sources exceeding the above cutoffs are considered "major" and will
be handled as individual point sources by the Regional Model Data Handling
System (RMDHS), although the NECRMP point source inventory generally includes
all available data on point sources regardless of size. Point sources with
emissions below the above cutoffs will be treated as area sources by the RMDHS
system. Their spatial allocation to grid cells will be based on UTM
coordinates.
AGENCY CONTACTS
Collection of raw emission inventory data is generally the responsibility
of the individual states, who are required to annually supply data to EPA's
National Emission Data System (NEDS), which is the national repository and
data base upon which air quality analyses are performed. Due to conflicts in
resources, priorities and differences in emphasis, not all states have kept
their NEDS data files in a sufficiently detailed and up to date form for use
in highly complex modeling exercises such as NECRMP. When it was determined
that a data base of a more current and complete nature than available was
required, it was necessary to identify the NECRMP effort as a state emphasis
program. In addition to their regular grants, each state received funding for
NECRMP emission inventory work and for responding to GCA's questions and
comments relating to the inventory compilation and quality assurance efforts.
In return, the states agreed to play a major role in the collection and
correction of the data base.
Under previous work assignments,* EPA and GCA staff visited each of the
involved EPA Regional Offices and the state agency offices. The objectives of
the state meetings were to:
• determine state and local contacts who were to be actively involved
in the regional emission inventory,
*EPA Contract Number 68-02-3168, Work Assignments 1-3.
-------
• determine if the needed data were available or the needed manpower,
time, and resources for the states to provide EPA with a current,
annual point and area source emission inventory in NEDS or EIS
format had been committed, and
• determine if the manpower and costs associated with the liaison and
assistance needed to support the NECRMP study were available and
committed.
At the outset of this study, the EPA Project Officer provided agency
contacts in the four EPA Regional Offices, 14 state agencies, and the District
of Columbia, which resulted from the earlier meetings. These contacts, listed
in Table I, served as the source of initial inventory data in most instances,
and assisted in a detailed review of the data for errors, inconsistencies, and
missing elements.
REPORT ORGANIZATION
'Hie results of the NECRMP annual emission inventory compilation effort
are reported in an 18 volume set of documents. A directory of the NECRMP
annual emission inventory reports is presented in Table 2. Volume I in this
series describes the background of the program and discusses the methodologies
used to compile and verify the annual emission inventory. Volumes II through
XVI present a more detailed discussion of each state's inventory effort.
Volume XVII describes the spatial, temporal, and species allocation factors
developed to allow for the creation of modeler's tapes from the completed
inventory. This volume, number XVIII, presents a summary of the point and
area* source inventories for the entire study area. Also included is a
detailed analysis of the overall quality of the data base and an assessment of
the data's utility for photochemical modeling.
This volume consists of five sections plus an executive summary which
preceeds this introduction. Section 2 summarizes point and area source
emissions, by state, for the entire NECRMP study area. Section 3 discusses
che techniques used to evaluate the quality of the emission data, and presents
relative emission densities as a function of population and land area for VOC
and NOX. Section 4 presents a state by state review of the quality of the
inventory data including identification of uncertainties, deficiencies, and
recommendations for future improvements. All references cited in this volume
presented in Section 5.
*For the purposes of this report, area sources include highway vehicles
-------
1~M
Jj^
CM
C5
£•4
J^
M
pj
^>
^»
S
r^
z
0
CO
CO
M
S3
1
PLJ
j^
s*
u
u
55
.
**J
53
P
I
CO
H
CJ
H
Z
8
^H
0
s
o
•-
i
H
x-N
M
iJ
O
jj
a
o
CJ
^
u>
CO
3
Ui
cu
CJ
C
01
a
> &u
co C
D o
ca a:
c
o
•H
JJ
CJ
01
o
Cu
CO
^j
C
o
>
e
'w
>-i o
0 C
.-1 01
oo e
X U
~~- CO
«£ o.
CU 0)
u o
u
3
CJ
1 -^
1 u
1 O
u
c
c
Q
u
•-1
01
e ^
o 01
co co ~^
Ul -H 01
U 0 O
•a en z
s
^ Ul CO
Ui n)
s 01 e
t J3 O
O O J3
Z 0! H
Ui
cu
01
c
.^4
00
c c
o w
••4
4J >\
o *•»
01 -H
ij •— '
O co
ui 3
cu cr
— 1 —1 CJ
10 fl C
U 4-* 01
e c oo
0) 01 <
s e
§ c —
0 O O
• •* ^4 u
> > C
ceo
W UJ CJ
UJ IJ-* C
0 O O
CCS
flj <1) ^-1
a B -.
4~J LJ O
U Ul CU
flj (fl
o. a. u
D v --4
CO 0)
U l-i
u ••*
, Ui
co co
> O
CJ
S T3
90 O
3 U,
6 CO
C
4)
0
CU
e
0» —i
a n
C u
O C
-.•* g
> c
c o
o-l >
o c
u
C '-i-i
01 O
a
u >.
Ui cj
CO C
O. 0)
01 00
0 <
•o
c
cO
~^
en
k-l XJ
C
11 O
-0 a
o C
.n u
as >
01
F-«4
CO
-o
01
cu
o
Q
Qd
«• co
Ui X
01 O
01 C
00 O
ui S
0 ?<
cu co
O «
e
o
4j
CJ
0>
U
O
Cu
CO
4J
C
0)
o
.J
>
c
Ul
1— I I4_l
>-t o
0 C
•-* 11
oo e
O- Ui
-~- CO
< s.
CXt 01
^
01
1 CO
1 Ul
1 01
!-«}
3
u
_i
•-*
•o
c
0
oo
01
>,
CO
U
C
o
tg
>
Ul
CO
c
0
CJ
u
nmen
0
,J^
>
c
Ul
U-l
o
c
e
4J
Ul
<0
a.
^
Ui
o
3
1)
Z
Ul
CU
e
r— t
36
0)
CO
Ul
CO
M
C
O
a
O
CO
3
C/)
r^
H
CU
Z
«
Ul
CU
oo
Q
CO C
-i O
D tJ
03 CO
S
CO ,-H
•rf CU
>•< 3
_, S
.-< co
3 c/3
^^
O
Ul
e
o
^*t
CO
4J
C
01
S
o
Ul
1-1
>-! U3
^^
c o
o
• -< e
00 O
0) "J
OJ n
— -^
CU .H
1 11
1 Ul
1 CO
3
CO
01
o
>— i
*-*
co a
Ul cp
01 -a
3 ui
cr 0
UJ
o c:
-1 <0
o
•C 8
££
S «
u IV
C 0
O Ul
U 3
O
01 CO
CO
•O u
B B
cfl 01
u o
'J^ .^
CO >
S-5
U, t^j
•- o
<
MH C
O 01
s
3 u
CO Ui
CU CO
uj a.
03 Q
CO
• *4
C
CO
•o >
C — i
CO >s
~-t CO
^ G
u* c
CO J)
S Cu
01
c
H
.^
en
td
S
CO
p4
^
3
T)
Ul
CO
O
Ul
a
o
o
3
^
O
CU
Ul
• f4
<
01
u
CO
CO
CO
• »4
C
Of
±4
•i-H
>
0
01 -H
-j a.
.rf CO
CO J3
CQ Crt
>, C
£ 01
CO ~-i
^
O
CO
en
S
o
CJ
yl
aj —4
e o
i "
c c
Ui 0
D CJ
0 C
u o
O 3
*-* —j
'J CU
CJ
3 U
CJ
Q flj
^- C
o ao
u Ui
Tjl --^
c >
j: u
CP CO
n u
3 3
01
• f4
Ul
•3
-1
^ 2
U CO
O --I
a -<
-a ^
a 3
^
CJ
(3
01
e
o
CJ
0)
o
Ul
cu
i-4
> co
c c
O (V
-* s
^0 C
•v o
OS U
"*v. "H
0, G
W CtJ
I
1
1
o
_£
0
>
-------
TABLE 2. DIRECTORY OF THE NECRMP ANNUAL REGIONAL
EMISSION INVENTORY REPORTS
Volume Contents
I Project Approach^
II Connecticut Emission Inventory^
III Delaware Emission Inventory3
IV Maine Emission Inventory^
V Maryland Emission Inventory^
VI Massachusetts Emission Inventory^
VII New Hampshire Emission Inventory?
VIII New Jersey Emission Inventory^
IX New York Emission Inventory^
X Ohio Emission Inventory^-0
XI Pennsylvania Emission Inventory^
XII Rhode Island Emission Inventory^2
XIII Vermont Emission Inventory^
XIV Virginia Emission Inventory^
XV Washington, D.C. Emission Inventory
XVI West Virginia Emission Inventory^
XVII Development of Allocation Factors^-'
XVIII Inventory Review and Evaluation
-------
SECTION 2
EMISSIONS SUMMARY
INTRODUCTION
This section summarizes the Northeast Corridor Regional Modeling Project
Annual Emission Inventory. Included are tables summarizing the original
source of each state's emissions data and emission totals for point and area
sources. Also included are summarizations of regionwide emissions by category.
DATA PROCUREMENT
Each state agency contact (see Table 1) was requested to provide that
state's most current point source emission inventory in computer readable
form. Data were accepted in NEDS, EIS/P&R, or EIS/PS format. A summary of
the original procurements is presented in Table 3.
Two different approaches were used in compiling area source inventories,
depending on whether the state had a current area source inventory (typically
the case in ozone nonattainment areas) or not, in which case, GCA had to
develop the inventory. Table 4 shows which states had a current area source
inventory and for which states GCA developed the area source data.
EMISSIONS SUMMARY
Table 5 presents a summary of VOC and NOX emissions, by state, for the
entire NECRMP study area. Included in Table 5 are statewide summaries of
point, area, and total emissions of VOC and NOX. For Maine, Ohio, Virginia,
and West Virginia, only the NECRMP study area portion of each state is
reflected in the Table 5 emission levels.
For reporting purposes, regionwide point source emissions were aggregated
into 70 emission categories. Aggregation of individual point sources was
accomplished using a PL/1 computer routine which is described in Volume I.I
The program screened each point source and allocated emissions into reporting
categories based on the Source Classification Code (SCC) and Standard Industry
Classification (SIC) associated with each point. The emission reporting
categories include categories for which reasonably available control
technology (RACT) requirements have been defined. In addition to "RACT"
categories, the reporting categories include other source categories which
were selected to facilitate reporting of emissions.
-------
H
It3
j-t
S
Z3
CJ
O
as
CU
<£
C— 1
«£
a
UJ
CJ
OS
ID
O
t/1
H
Z
(—1
o
j^
a
S
^
5;
X
VI
•
CO
£j4
J
"~3
^
H
^-s
C
0
•-4
"U a)
01 «)
> •-!
•* a
cj 3
cfl
»J C
cfl -H
a ao
• H
Ui
O
^"
jj
cfl
O
Cu
01
O
Vd
3
O
01
to
4-1
tfl
Q
01
4J
cfl
GO
00
"•»«,
0
, 1
en
a
c
0
. -1
4J
O
Ol
4j
o
lj
CU
1-1
cfl
4J
G
01
e
G
O
u
• -4
C
td
'4-1
0
•
•U
O»
•
c
o
CJ
4J
3
U
CJ
0)
G
O
CJ
CO 00
~\ "•»»*
i/*i m
OS
'•O
04
"*-• CO
CO O
I-l JJ
e
r-4 0
0 -r4
U 4J
4J O
d 01
O 4J
CJ 0
Vw
<— < dl
?0
u 1— 1
C cQ
D iJ
S 4?
O S
^4 C
•^ o
> u
a •*
Id >
e
o
U-J
O
•H •
Q ' '
a.
Ol 0)
u a
3 01
co C
—> -iH
5 S
01
u
3 0)
CO C
— « -H
Ol CB
Q X,
oo 00
***^ ""^
00 CJV
co co
a a
W M
z z
cfl
r—l 00
o cT
U -r-l
4-1 r-l
C 0)
O 0)
CJ C
• H
o> 60
03 C
•-I W
o
z >^
4^
•O •>•»
C —1
CO !B
3
>> O'
4J
•r^ r«^
—4 CO
CO 4->
cD- o>
•^ O
< tl
VM '>
0 C
3
co lu
01 O
^
3
CQ 4J
a.
-o 01
c o
CO
(— 1 «
^*l *A
U 01
*9 J9
03
4J
4^
01
03
-O 3
C J=
CO U
r-t co
U CO
(fl (tj
QO 00
""-s. "****
>O CN
P^
as
cu
CO ^*s.
Q CO
Q
C
o
• f^
4^
o
01
4J
o
u
CU
v-4
CO
4J
G
0)
S
e
o
• H
>
C
'-M
O
•
4J
a.
01
Q
>m
0)
CQ U3
q u
< 01
z >-)
< 3
cu oi
Ed Z
01
U
•*J
-C >>
en 01
a. 01
S u
_cfl 01
3 3
O! 01
Z Z
oo oo
"•"^ *"*^
to t-4
co co
Q Q
Cd Cd
Z Z
C
O
•U
CO
)-i
0)
01
C
0
CJ
f-4
CO
4J
C
0)
|
o
• I-l
>
c
^
0
•
I 1
a.
01
a
u
o
u
3 <:
01 CJ
Z 0
•^
o
o
3 •-<
Ol J=
z o
00 00 00
"*^» "*^ c
C 3
Ed O
CJ
y-j
0 >,
G
• 0)
a. oo
0) 01
C5 -4
f-4
cfl •<
• H •— '
C
CO •<
> CJ ffl CQ
^° § 3
01 -O Z Z
C C -~ ^,
c to •< <;
01 CU CU
CU td Cd
co -a
•^ (2
C CO
CO r^
> CO
-* M 4-1
PS C
03 01 O
G -ap
C C C
oi j: 01
cu aS >
co
"^s»
CN
r-4
cd
Cu
•"•s,^
CO
Cd
•o
M
CO
o
w
,-4
0
4J
e
o
CJ
£
o
4J
3
f— *
O
CU
I-l
• H
^
01
CO
co
=s
CO
C
•H
V?
>
0)
00 00
• — -^
0
•^ 4J
cue:
o en 01
•H e
30 30 C
01 C 00
as —i -H
-^ U CO
•< ai co
Cu cu cfl
Cd C
• ^ cu
j3 ai 4J
so e co
3 Cd M
O ca
ui co a,
J3 •<-< 0)
H > co
co .a o
0]
^
C
01
a
*^
M S
•H B
•< 0
CJ
CO
•H U
O 01
•i-l -H
00 -H
M I-l
•i-l cfl
> 01
4-1 01
03 4J
0) -H
rs Q.
09
O) 01
A "O
A
O CQ
4J 4-1
cfl
2^
•H >>
3 1-
0 O
Jj
T3 C
• 0) 01
CO M >
4-1 • r-4 C
CO •-> -ft
^3 'H
4J CJ
00 3 -H
l-» CO
C^ CO CO
1—4 CO Q
3
oo oi
-4 JC "O
•« CJ --4
C -4 >
O JS O
•H 3 u
00 CLt
Ol 00
OS -i 0
Cd >\
Ori H ^\
3 O 4J
CO 4J --4
G •*
01 Ol --4
4J e , CO G
u o
O 01 -^
4-1 J3 CD
C 4J 01
01 -^
^ U^ 3
col
M 0
fl 0
01 O
C '^ ^^ v^
O 4J O r-4
•H CU k4 .
01 — 1 4J 1— 1
co a. G >
•H SoX
S o cj
Cd O 0)
C S
01 03 o 3
.C 4J .H --H
4J O 4J O
01 3 >
S -4 r-4
O '4_ --I D
Vj Ol O Ol
Cu OS Cu CO
•3 01
-------
TABLE 4. SUMMARY OF AREA SOURCE INVENTORY DATA SOURCES
State
Source of Inventory
Connecticut
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
Washington, D.C.
West Virginia
Connecticut Department of Environmental Protection
GCA developed
Maine Department of Environmental Protection/GCA
developed (category dependent)
Maryland Bureau of Air Quality and Noise Control
(nonattainment areaa)/NUS Corp. (attainment areas)8
Massachusetts Department of Environmental Quality
Engineering
GCA developed
New Jersey Department of Environmental Protection
New York Department of Environmental Conservation
GCA developed/Northeast Ohio Areawide Coordinating Agency
(Cleveland area)
Pennsylvania Department of Environmental Resources
Rhode Island Department of Environmental Management
GCA developed
Virginia State Air Pollution Control Board
Washington Council of Governments3
GCA developed
aThe area source data for the Maryland Counties of Montgomery and Prince
George's were compiled by the Washington Council of Governments.
-------
TABLE 5. STATEWIDE TOTALS OF AREA AND POINT SOURCE EMISSIONS OF VOC
AND NO* FOR THE NORTHEAST CORRIDOR STUDY REGION*
Emissions (tons/year)*5
State
Connecticut
D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
Area
VOC
167,730
21,309
35,743
66,403
197,699
235,862
59,187
415,035
651,570
379,588
465,776
49,264
33,551
115,842
87,200
Point
N0x
113,160
15,684
40,649
42,783
156,845
266,046
45,306
348,568
587,109
432,098
491,114
48,830
25,791
89,343
91,962
VOC
34,295
1,952
21,222
13,651
77,557
113,328
21,041
158,228
191,633
85,452
264,054
13,375
4,921
24,443
15,841
NO*
62,292
6,345
42,730
16,646
138,095
151,196
49,206
204,844
1,516,743
310,576
538,019
5,037
1,595
35,868
401,152
Total
VOC
202,025
23,261
56,965
80,054
275,256
349,190
80,228
573,263
843,253
465,040
729,830
62,639
38,472
140,285
103,041
NOx
175,452
22,029
83,379
59,429
294,940
417,242
94,512
553,412
2,103,852
742,674
1,029,133
53,867
27,386
125,211
493,114
REGIONAL
TOTALS
2,981,759 2,795,288 1,041,043 3,480,344 4,022,802 6,275,632
a Emission totals for Maine, Ohio, Virginia, and West Virginia reflect only
the study area portion of each state.
emissions reflect totals reported by the respective states or MPOs.
GCA made no adjustments to reported data to specifically reflect either
cotal or reactive VOC. Generally evaporative sources reflect reactive VOC
per EPA" 3 policy summarized in Referenda 19. Highway vehicle emissions
generally consist of nonmethane VOC as generated by MOBILE1 or 2. Other
combustion sources most often reflect total VOC emissions. State-specific
deviations from the above are discussed in Section 4.
10
-------
Table 6 presents a summary of regionwide point source emissions by
reporting category. Summarization of each state's point source emissions into
the same reporting categories as shown in Table 6 can be found in the
individual state volumes of the NECRMP report.2-16 pOr points with multiple
SCCs, no attempt was made to split emissions into more than one category. In
these instances, the primary SCC was used to account for all emissions at that
point. It was further assumed that the first SCC listed in the point source
record was the "primary" SCC.
The emission totals in Table 6 reflect 55,474 individual emission points
representing more than 1,400 different source classifications (SCCs) at 12,657
facilities. However, a large portion of ozone precursor emissions occur
within a few major categories as indicated in Table 7. These categories
account for approximately 98 percent of the NOX and 84 percent of the VOC
emitted from point sources in the NECRMP region.
Table 8 presents area source emissions, by category, for the entire study
region. Summaries of each state's area source emissions by category can be
found in the individual state volumes of the NECRMP report.2-16 ^g can be
derived from Table 8, ten major classifications of area source emissions
account for over 82 percent of the total area source VOC and over 86 percent
of the NOX in the NECRMP Study Region. These area source classifications
are summarized in Table 9.
11
-------
efl
O
• l-l
M
M
CO *->
u-|
ON
m
ox —^
co
CO
a
o
O
O
•H d
co td
en w
4) H 4) U
U 00 O
CU 4) O «-J
OO -u O
•^ co O w
U • 4) «9
3 S -O U
•O 4) 3 H
o si C
x-s O
M O
M >
M 4) H-t
60 O
60 ,
*^S T3 ^0
33 c "° o
4) eg
•^ « V( 4)
Hen OM pM>eucxiu3OMHPKiHi-JO-iC/:
12
-------
CO
0
o
CO
0
O
•H
09
CO
O
cn
0
0
O
CJ
w
J
o
o
u
o
CM
O
CO
0,
B
0
O CQ
'r4 U
CO 0
U) -H
•H 0
I a
o
00
4)
U
iOCMCMl^OOOOOOOOOOOOOvOi-lOOOOOOOcMr>.iOvooO
lOr-li—IIO O ^ CO CO r-4 i—I
r>»CMCM r-l P>.Or-IPOP^
^t -^ vo CM CO ON CO
CM vO $• r-i o
00 CO CO r-l
CM
CO
O
• H
O
0)
Cd
CO
r-t 00
Cfl AJ
jj u
o
££
CO
O ^8
I*
CMCM
CO
0)
0)
CO
01
CJ
O
09
4)
0
0)
3 -H
CD to S
00 C O 35
ij 50
CO
O
CM
m
u
3
X f
o oc
>rl -i-l Ol JJ
OO
OO
ooooo
tJ°
co a.
d, .H
JJ 01
CO 00
cfl O
M O
O
l-i •
en
CO
• i-l
M
CO
3
1
CO r-l
^ M CO
03 CO rH
M (U U
01 M
I
0) 01
U M
J3 K -H
a
-------
^^V
inuec
u
C
0
y
^•^
O
tt
_3
>— 4
3
H
8
/^
u
nj
0)
>•»
-.. 0
01 O
a >
o
u
0]
a
o
•^
M x
a) O
•* Z
£
0
0
C/3
>> CM
u o
m en
J
'|j
(X
a.
CO
H
C
O CO
•^ U
« a
03 •*
•H O
fl *
•X
>>
cr
o
00
41
CO
O
OOP«.CN«NOOOOO u"i
in m oo ^o o fi
vr o ~d- oo CM r»
co o^ co co cy* cr>
—* CM t
o
m
o
CO
m eo vo — <
oo
CO
o>
01
4)
9
CO W
-• 4)
« « e
•t-i 01 o
h 4J -1-4
•u y u
c o o c o a
o o u M a. >, o
• ^ I I I CO M O
jj ^ _l ,—l .^ 4)
« « «3 « Q > -"
3 01 « GO o a
jQ O O O Q) y C
a a. a. a. u 4) w
503 0] 0) 03 ai 4)
.^ >rj .,-1 0) JJ
O O O 3 u C
^ CM
4) 4) 4) 4) -O 4)
co
00
CM
0 OO •«
co c • y-j
4) 4) -i-l >» C O O
y y cu 4) •,-( BH AJ
s wi a to 3
JJ -*>% 4)
OS 3 3 3 3 Q C 1-
yCOO4)OO4J'-^3
uju-i-l -OZoB-OCiJO
O -1-4 O 4) 4) 01
ya.'T3oioo>^
•a co-^coocc4)AJ
AJCUca"'<8jS>' «4)>^O
CO-O 5 _ U 04 CU U h O.
•« 4) -U O
3 >
•<-l 3 u JJ
4) -o y i-i i-
yflfi ca«8v4)wCaO
W 3 •(-( lu 4J u C -i-( -H
3 y-iotose-H js
O OO4J'^O03O4)
co . u jj js tJ --i Cu y 4)
r-» 3 4) u M-I 03 u X2
0) O >, U U ^0)3-U
4) i y: ti o as : jZ 0
Wl^f-H 3-UC U 33 OJ
aj -o i—4 4) o o«a y
QycoCi'^ C u 4)
01-OO.Ul O03U.i-(C>-4
-OJ->O.O 4)34)>>
«moc X4)W-4j:4)
4)1 .^ 41 4) T3 CJWCO
4) - 4) i-l Op C I-i 4) C
l-l ^ OJ j: olCf'H O U4) 3
oj -^ M JJ C -H 4J nj 41 4)
I -H 03 O C U Z
u CN 3m u a) co o
3 ONQ0041 >4)01W
^aco^-jajjr^c-H o
4). Hi^.^ijH u
co oj co i—i •»•
4)
O -^ i-4
O 4)
JM u y
04 e I-
H 4) 3
> o
** «w »v •—' *w
<8 3 W 00 3 Vj
cu -o"C4|Hw-^cja.
cu i-i M 4) y
y £
u
4)
^ 4) J-l C
3 /-» • 4)
j oi e o >
a •* p e
93
a
01
3
O
co
y
•
>» M
a a)
a y
8 ^1
TJ u so-^ -H H
f-* 04 4) g
3 u 4) O.
O 4) «9 S _•
CO O
4) U
O Z
0-HVrf
4)
~* G - —
aj u C u
3>O4J WOUJ->>4.T*
OtawC CMOUH
CO 3 04 34)^
14
-------
0]
a
o
CO 0)
ta co
CJ -rJ
o f3
o
CO M
O
z u o
r-l 3 >
O 0
Oc 01
M
0
fV* it
PM (Zl
05
Z -H
O >H O CU d
3 CO O 03 -H
U M 03 U
O 3 "O , d M-t >, co OH CU
00 M JO 3 CO O O
d CU X d r-4 r-4 efl
•rH d r^
o o a\ CM I-H
r«- r-»
i— i
eg
d
o
OJ -iH
M d u
(DO 3
r-4 -.H U
•iH U -i-l
O CO U
PQ r4 0)
cu d
d d M
O CU "••»
•fH O I— 1 r-l
u eg co
CO O -iH -rl
3 -iH IH u
J3 k JJ IH
to S u °> v
AJ O O 3 S
00 M O CU *^ a
C < r-l d O
•H I-H W r-l O
co o eg
eg -H d 1 I 1
-------
TABLE 8. NECRMP AREA SOURCE EMISSIONS BY CATEGORY
EIS/AS
t''4Un$vJVy
Number
001
002
003
004
005
006
007
008
009
010
Oil
012
013
014
015
016
017
018
019
020
021
022
023
024
025
026
027
028
Annual Emissions
(tons/year)
Category Description
Stage I Gasoline Evaporation
Stage II Gasoline Evaporation
Storage Tank Breathing
Gasoline Loading/Transit
Small Indus trial /Commercial De greasing
Dry Cleaning
Architectural Surface Coating
Auto Body Refinishing
Small Industrial Surface Coating
Graphic Arts
Commercial/Consumer Solvent Use
Cutback Asphalt Application
Pesticides
On-Highway Light Duty Vehicles
On-Highway Light Duty Trucks - Class I
On-Highway Light Duty Trucks - Class II
On-Highway Heavy Duty Gas Trucks
On-Highway Heavy Duty Diesel Trucks
On-Highway Motorcycles
Residential Anthracite Coal
Residential Bituminous Coal
Residential Residual Oil
Residential Distillate Oil
Residential Natural Gas
Residential LPG
Residential Wood
Ccramercia I/ Institutional Anthracite
Commercial/ Institutional Bituminous
(continued)
16
VOC
79,774
121,240
12,511
44,561
105,047
40,848
145,651
57,235
139,944
26,231
198,844
54,811
24,153
1,063,448
112,544
69,721
124,903
57,103
11,905
6,284
678
246
4,781
4,761
134
2,976
300
513
NOx
0
0
0
0
0
0
0
0
0
0
0
0
0
1,005,785
102,675
64,614
137,504
421,473
6,472
7,138
498
11,817
82,848
60,144
1,437
769
8,678
1,629
-------
TABLE 8 (continued)
EIS/AS
Category
Number
029
030
031
032
035
036
037
038
039
040
041
042
043
044
045
046
047
048
049
050
051
052
053
054
055
STUDY
Annual Emissions
(tons/year)
Category Description
Commercial/ Institutional Residual Oil
Commercial/Institutional Distillate Oil
Commercial/Institutional Natural Gas
Commercial/Institutional LPG
Industrial Bituminous
Industrial Residual Oil
Industrial Distillate Oil
Industrial Natural Gas
Industrial LPG
Industrial Wood/Other
Military Aircraft
Civil Aircraft
Commercial Aircraft
Railroad Locomotives
Gasoline Powered Vessels
Distillate Oil Powered Vessels
Residual Oil Powered Vessels
Off-Highway Vehicles - Gas
Off-Highway Vehicles - Diesel
On-Site Incineration
Open Burning
Structural Fires
Field/Slash Burning
Forest Fires
On-Highway Vehicles (composite)
AREA TOTALS
VOC
977
1,167
1,630
32
8,403
368
560
11,885
12
17
12,117
4,529
15,599
25,492
63,372
1,987
3,460
47,012
20,582
19,386
98,220
15,855
8,308
6,538
103,104
2,981,759
NOx
54,213
40,861
29,266
549
126,763
22,253
14,951
62,305
440
500
6,123
632
17,213
100,320
16,994
9,651
3,726
34,710
219,166
2,691
17,668
3,274
1,380
925
95,233
2,795,288
aDuring development of the Connecticut Area Source inventory, vehicle class-
specific emissions data were unavailable. Therefore, a composite category
was utilized.
17
-------
TABLE 9. HAJOR CATEGORIES OF OXIDANT PRECURSOR EMISSIONS FROM
AREA SOURCES IN THE NECRMP REGION
Percent of Area Source
Oxidant Precursor Emissions
Emission Category
NO,
VOC
Gasoline Marketing
Degreasing
Architectural Surface Coating
Commercial/Consumer Solvent Use
Highway Vehicles
Residential Distillate Oil
Industrial Bituminous
Railroad Locomotive
Off Highway Vehicles
Open Burning
0.0
0.0
0.0
0.0
65.6
3.0
4.5
3.6
9.1
0.6
8.7
3.5
4.9
6.7
51.8
0.2
0.3
0.9
2.3
3.3
18
-------
SECTION 3
EVALUATION TECHNIQUES
INTRODUCTION
This section describes the techniques used to evaluate the quality of the
NECRMP emissions data. Included are discussions of how the NECRMP data were
compared with inventories filed by each state in support of State
Implementation Plan (SIP) submittals and data contained in the National
Emission Data System (NEDS). Also provided in this section are tabular and
graphic portrayals of emission densities of volatile organic compounds (VOC)
and oxides of nitrogen (NOX) expressed both in terms of population and land
area. Ratios of emissions (VOC/NOX) are also provided for a number of
emission categories, by state, to identify possible anomalies.
SIP COMPATIBILITY REVIEW
A review was made of all 1982 ozone State Implementation Plans (SIP)
available for ozone nonattainment areas within the NECRMP region. In most
instances, the NECRMP emission data were obtained directly from the same state
agencies who were simultaneously developing the SIP inventories. GCA reviewed
the state-supplied data and, to the maximum extent possible, worked directly
with each state to resolve problems. Further, each state received funding for
NECRMP liaison activities, in addition to their regular grants, and were
obligated to inform GCA of any changes made to their emission inventories
after they were supplied to GCA. In theory, then, the SIP and NECRMP
inventories should agree for all ozone nonattainment areas within the study
region, however, this is not always true. In some instances, certain states
made significant changes to their SIP data either without notifying GCA or
after the respective NECRMP inventory had been completed. In fact, as of this
writing, some states are still actively modifying SIP inventory submittals.
In other cases, entire area source categories had been omitted by some states,
and to maintain uniformity and completeness across the entire NECRMP region,
GCA estimated emissions from these apparently omitted categories. Some states
supplied GCA with computerized point source inventories which bore little or
no resemblance to the point source data reported in the hard-copy SIP
inventories.
While there is only one "correct" inventory for a given area, use of
different methods, assumptions, and data sources can yield greatly differing
results. To avoid inter-state biases introduced as a result of the employment
of different inventory methods, the NECRMP inventories were compiled and
verified using a single standard set of methods. Deviations from the standard
19
-------
methods or emission factors were brought to the attention of the appropriate
state agency for reconsideration. While one objective of NECRMP was to ensure
consistency from state to state, it was also the objective to ensure
consistency between the SIP and NECRMP inventories. There were, of course,
instances when these priorities conflicted. These cases were resolved
individually. Deviations from the standard set of methods are documented in
the individual state volumes.2-16
The SIP requirements entailed reporting (hard copy) reactive VOC and
NOX emissions on a daily basis for a typical summer weekday to support
SIP-required attainment demonstration (e.g., using EKMA). The NECRMP
inventory required annual emissions reflecting total VOC and NOX in
computerized form for further temporal, spatial, and reactivity resolution to
be used in EPA's Eulerian Regional Oxidant Model (ROM). Because of these
differing requirements and issues of timing and availability, the SIP
inventories were not directly used in NECRMP. However, the NECRMP inventories
were compared with the SIP inventories to identify major differences for
possible future resolution.
Since the SIP inventories seldom reported annual emissions and/or the
temporal factors from which daily emissions were derived, direct comparisons
with the NECRMP inventories were not possible within the time and resource
limitations of this project. To attempt some comparison, the approach taken
was to convert the SIP-reported emissions from kg/day to tons/year using a
conversion factor based on 365 days/year. An exact match would not be
expected since the SIP inventory totals were adjusted to reflect emissions on
a typical summer weekday rather then an average annual basis. Also, no
attempt was made to adjust the SIP-reported reactive VOC totals to reflect
total VOC as is represented by the NECRMP totals. However, if the adjusted
totals were within approximately 10 percent of each other, they were
considered to "match". Regardless of how close the inventory totals are, all
known differences are identified and evaluated in the state by state review of
data quality that can be found in Section 4.
NKDS COMPATIBILITY REVIEW
States are required to annually supply updated emissions data to EPA's
National Emission Data System (NEDS) which is the national repository and data
base upon which many air quality analyses are based. Due to conflicts in
resources, priorities, and differences in emphasis, not all states have kept
their NEDS files sufficiently detailed and up-to-date for use in modeling
analyses as complex as NECRMP. Since the NECRMP, NEDS, and SIP inventories
are all based on state supplied data, in theory, all three data bases should
agree. However, since EPA negotiated with each of the states in the study
area to provide increased assistance and priority to GCA1s questions and
comments relating to the compilation and quality assurance of the data base,
it is believed that overall the NECRMP inventory represents the most suitable
emissions data for the intended application.
Unfortunately, the degree of cooperation of some of the states in the
NECRMP region was less than ideal. Thus, for certain individual sources or
area source categories, the data contained in NEDS may actually be more
20
-------
representative of the actual level of emissions. It is useful, then, to
compare data contained in NEDS with the SIP and NECRMP data to identify areas
of disagreement for future resolution. Tables 10 and 11 present comparisons
for each major SIP area of the NEDS, SIP, and NECRMP data bases for VOC and
NOX, respectively. Statewide or study area totals from NEDS are compared to
the NECRMP data in Tables 12 and 13 for VOC and NOX, respectively.
As shown in Table 10, the NECRMP VOC emission estimates generally agree
with the SIP totals for the 13 SIP areas in the study region. The totals for
all 13 SIP areas for NECRMP are within <1 percent of the reported SIP values.
The NEDS emission estimates for the same areas are 47 percent higher than
those reported in the SIPs.
As shown in Table 11, the NECRMP emission estimate for NOX, across all
13 SIP areas, ia approximately 25 percent higher than those reported in the
SIPs, while the NEDS data are approximately 13 percent lower. The majority of
this difference can be attributed to the NECRMP point source inventory for New
York. As is discussed in greater detail in Section 4, the reported NECRMP
electric utility emissions for New York are higher than those reported in NEDS
by almost a factor of 7. For reasons discussed in Section 4, the utility
emissions reported in NEDS are believed to be more accurate than those
reported in NECRMP. Further, it is recommended that the NEDS estimates for
utilities be substituted for the NECRMP estimates in New York. If so
adjusted, the NECRMP NOX total for the 13 SIP areas would agree within less
than 3 percent.
As shown in Tables 12 and 13, a number of large differences exist between
the NEDS and NECRMP estimates of emissions by state. Major differences, or
differences that apply to all states in the study region are discussed below.
Differences that apply to individual states are discussed in Section 4.
As shown in Table 12 on a regionwide basis, point source VOC emissions
reported in NEDS and NECRMP are reasonably consistent (NECRMP point source VOC
is 5.5 percent lower). However, on an individual state basis, some large
differences occur; these are discussed in Section 4.
For area sources, the NEDS VOC estimates are consistently much higher
than those provided for NECRMP, although there are a couple of exceptions.
Nearly all of the difference in area source VOC emissions can be attributed to
evaporative VOC emissions from organic solvent use. This occurs primarily
because GCA and the states utilize methodologies consistent with the EPA
Guidelines, which do not consider all of the identified industrial end uses of
organic solvent given in Table 4 of End Use of Solvents Containing VOC
(EPA-450/3-79-032)20 as potentially being area sources, while NEDS does. In
addition, the organic solvent when end use cannot be identified is all treated
as area source emissions in NEDS. The assumption inherent in the
Guidelines^ is chat all industry end users classified as "other" or
"unidentified" will be completely accounted for in the point source
inventory. Thus, on a regionwide basis, it would be expected that NECRMP
point source VOC emissions would exceed NEDS point source totals. As shown in
21
-------
2
^^
CJ
O
>
^^
Vj
9
f.
p
5i
'-'
^ ^
^
5
O
cd
o
UJ
*— *
Z-^
i
i—4
O
^>
C/3
-»
o *
•r-4 GO
*•<
"H
y: ai
a -3
t/5 OH
C/3 >— t
—l c/5
Ui Q
U
3 U
£2 t)J
•z. -J
xS -^J
<^ C*O
0
— "
u
P-J
SO.
•^
H
CO CM O -J O
Cu r*** ^^ 'A — «4 r^
aj -— i
c/5 GO *J ^^ ^3* tn
^j NO ON O -3" CO
Z f» CM CM
ID
41 "S
U — I
< 14 U C
4) •- -3 4)
3* u 14-1 U >
-H ej <0 ;fl O !fl
•n o 4; C '-" s
U U --H tJ
•j) u u V. 3
0 o o- * 41
a 3 crt = Z
ON
en
NO
CM
NO
^J1
ON
in
p^
NO
r-*
0
CM
CM
0
^H
x3*
NO
CM
O>
-3-
0
0
m
—4
m
-3-
m
00
>n
r-.
en
en
—4
ON
O
NO
_^
en
NO1
^
P.H
4)
o
C
41
-o
-^
^
O
3-
^.
ON
r-4
00
O
00
NO
r^
PN^
m
m
00
^*-
NO
V-*
CM
>j-
tN
_
CJN
CM
m
CM
^3"
NO
NO
CM
ON
ON
NO
00
en
0
m
ON
in
NO
m
NO
—4
O
in
NO
m
—4.
NO
r«*
,_,
ON
CM
V
)H
O
3
V
_i
^3"
en
NO
CM
00
en
f-s
en
CM
O
0
NO
^1
r*»
ON
00
00
«4
en
o
r~.
P^
p— 4
f^4
O
CM
en
CJN
O
in
m
en
en
NO
CO
^
in
en
in
H^
-*
in
_H~
m
m
O
CM
m
^H
CM
p^
p^
vO
ON
^H
CM
O
en
NO
CM
en
O^N
CO
CO
ON
en
—4
CO
m
en
~H
m
CO
O
ON
r^
-H
in
co
en
ON
NO
^H
pi*
r**
ON
en
p*.
CM
CM
-3-
CM
CM
•^
en
en
CM
en
^
-H
ON
CM
0
en
— H
00
p*.
ON
CM
m
NO
CM
oo
-T
O
,-H
^H
•c
c
4>
J>
CJ
r~4
U
cn
>^
^
o
^o
(N
A
CM
m
•H
NO
ON
NO
CM
CM
O
m
00
en
rH
^
en
_H
p*
r*
m
r*^
NO
J"
^f
p».
00
en
NO
Pv
A
^H
CM
O
en
en
m
*
CM
CM
P-.
NO
O
ON
m
^
l~-
m
o
m
00
in
o
m
CO
3
rTi
r?
^
a.
CO
•J
J
<;
•
4)
00
C
2
o
u
u
41
jQ
3
03
41
l^
CQ
tn
tH
41
.0
B
3
a.
HH
trt
•
Ui •
oj en
0) 00
>> o>
u
41 -H
Q.--I
U
N
91 "°
41 -o
Id d)
3 >
• H 41
0)
03
3 e
C 0
e "^
to u
0 I
U M
O
•0 u-l
4) C
U -H
u
41 -U
> w
e 41
O *J
—4
9)
u U
(C 41
S -i
• fH U-l
U 41
41
aj
flrt ^J
H- * (B
V) Q
•K
22
-------
u
w
35
CO
3
i o-
•* P-, M
M yj
rl «
10 &
Q
H
O c/5
W O
,-j S
W SB
to
OS
~ * ,
x 14.1 2
o o 3
z ^ w
rft O
r? o
U4 r i
Q C
.j m M t* .
S^ W »y Ew
O a M w
o £ ws
•*4 «
z «
w «
CJ .p-
1
M '"'
25 3 Q
* 1 %
o •*
to
6 a,
H 5
N— ' 2
to g
O
I 1
to
to u
M a a.
M O US
fc*4 CU
pj
25
C
g Q
5 M
35
^
HJ
3
^^ *9
r~^ QJ
^
•<
a,
M
to
ON
CO
ps.
o
CO
00
in
CM
in
00
O
m
ao
O
CM
CM
CM
CO -*
00 CM
O r»
oo \o
in M?
CO •»
—4 ^
CO vO
00 P»
in o
co
O
IX
ao
m
_,
ON
-1
p-l
2
"*
NO
1— 1
P*.
co
O
,3.
NO
o
p-l
p*.
rf
o
m
CO
00
m
ON
P-.
CM
p-i oo
in r^
vO ON
%O p-l
•^
^O NO
00 CM
co m
5 °
CO O
00 OO
•-4 00
00 ON
O CM
ON ON
CM r-»
CM CO
pH
CM ">
CO P-I
p-l p-l
rv PH
in ON
p_j
CM m
CM c~
f- CM
CM CO
CO 00
1-
4) 00
M P<
O -3
a -Q
.H
o
NO
^^
p^
^
NO
oo
m
O
1— 1
05
r-
0.
i-l
ID
•o
pH
•i-l
j:
a,
p-4
NO
in
m
m
P-I
ON
P*.
ON
CO
CO
rH
ps.
P-.
CM
CO
CO
ON
p-
O
in
O
p-l
00
o
CO
00
00
0
o
rs.
CO
CM
00
O
p-l
vO
NO
^f
ON
CM
NO
sj-
.
Q
^
a
o
4-1
9
If4
A
a
to
5
in r—
O NO
p*. in
NO ON
ON m
-H 00
CM
co 00
co m
CM CO
CM -31
00 r-l
p"4 CO
CM
CO CM
r*s p^
co ~3-
ON ON
m P-I
-4 O
CM
P-. 00
-4 O
p-l O
m CM
p* in
M
""*
PS. ^
m P-I
O CM
ON O
O -3-
— 1 CO
A
p-l
S31 00
in CM
r-. oo
1-4 O>
ON CM
CM
p-4
•* CO
co oo
m
-------
a
H
O
oi
O
Cb
w
H
t-1
CO
Ed
2
O
t—'
CO
CO
a,
X
Qd UJ
Z H
>n
10
CO «
O CO
LJ O
Z Z
Cb O
O 3-
§3
CO
1-1 a
8
•jj
J
39
c
o
H
en
a
o
•H
CO
09
's
fd
3
C
O
H
co
01
o
a,
o
z
en
a
g
o
M
Z
w
o
•jj
z
o
z
fO
jj
en
o
CN
oo m
en so
O -4
en
vO
00
tN
O
00
CN
O
O
•o
*
o
sr
-*
M
ul
EN
00
CN
ao
CN
SO
m ON
oo
es
m ON ^ r» co vo
CN
r^
O
09
%
1 "I I" s
03 cq cu cu
H
8
z
3
a
jj
CO
J^
1-4
O
0)
CD
U
24
-------
in"
u
a
X
o
z
W
a
M
Z
oei
O
c*.
W3
W
2
3?
rH
H
V3
W
CO
z
2
Cfl
M
9
fl-
ee
u
*
o
5
05
a
z
o
z
o
S 0) rH
* « d fs
0) rH -rt IH
5 CH 5 S
CM CM
St rH
CM m
r-» st
ft Os
•*
rH St
m r-»
in oo
CO rH
CM
^O \Q
St 0
O CO
so m
sO St
CM
o r-
co r^
OS rH
en o
oo en
rH
i^ sO
OS O
rH CM
rH OS
m st
rH
CM Os
sO so
OS rH
00 St
CO CD
4-1 rl
CM
CM
st
*sf
00
^.
o
CM
en
O
co
iv.
m
i— i
^
0)
V
3
£
CM
m
oo
en
O
rH
CM
O
en
rH
en
sO
p-
os
o
rH
r«.
oo
m
p^
sO
m
sO
CM
st
en
^^
P*
sO
rH
m
r-i
en
in
sO
en
o
^
01
z
st en
p^ en
SO rH
CM Os
St CM
p^ C3
rH
p~ o
en 00
CM 00
CM Os
Os CM
ITS rH
**
rH
00 St
Ov rH
O rH
CM rH
en Os
St St
CM OS
CM sO
sO 00
O CO
OS O
CM in
sO Os
^^ ^"4
m O
O 00
rH en
co m
in i— i
rH rH
sO O
rH SO
0 CM
fO \O
ctt
'd
rt
rH
* CO
o d
•H d
SZ 0)
O £<
P-* sO
sO 00
oo en
en P*
m CM
00 OS
O 00
m oo
sO CO
CO St
O rH
en os
oo r^
oo m
St CM
00 CO
0 CM
m o
rH CO
en st
P- m
en o>
o m
in rH
O sO
C5 \&
O CO
m
•o
C
cd
rH
to
r-l i-i
a
rH st
rH rH
CM ft
m co
CM OS
rH St
CM CM
i— 1 Px
CM rH
O sO
rH CM
I-H m
en CM
st so
en ON
OS rH
00 OS
00 rH
m m
00 Os
Os rH
P-. Os
00 CM
00 4J
IH co
> S
CN
co
sO
m
r>«
CM
sO
OS
00
O\
m
m
M
st
00
oo
CM
m
OS
p-^
M
CM
0
O
CO
Os
Os
CM
CM
St
st
CO
o"
00
st
CO
Os
oo
SO
CO
m
CM
•«
CM
3
^
H
8
3
z
o
R
o
g
•
01
0}
«
01
r*
rW
U
U-l
o
a
o
•rt
U
Q
a.
g)
0)
rl
a)
S
u
CO
>,
rH
d
o
00
3
rH
C
M
*
25
-------
Table 12, this is clearly not the case. This result implies that either the
methodology inherent in the Guidelines underestimates emissions from organic
solvent use, or that NEDS includes emissions estimates from sources that do
not really exist. Because of the magnitude of this discrepancy, it is
recommended that the organic solvent data contained in End Use of Solvents
Containing VOC^O be reexamined with respect to the Guidelinesiy
interpretation of these data.
As shown in Table 13, the NEDS and NECRMP NOX emissions differ by over
1.7 million tons/year. Most of this difference is attributed to point
sources. Specifically, over 1.1 million tons can be accounted for by electric
generation sources in New York. Since the NEDS data on electric generation
are annually updated based on DOE data (confirmed by E. H. Pechan estimates of
SC>2 from electric utilities), independent of the states, it is most likely
that the NEDS estimates, in this case, are correct and should be substituted
for the NECRMP estimates for electric utility emissions. The NEDS and NECRMP
estimates for electric utility emissions in New York are compared in Table 14.
TABLE 14. ELECTRIC UTILITY EMISSIONS IN NEW YORK
Annual Emissions (1,000 tons) from Electric Utilities
Pollutant NECRMP NEDS
ISP
S02
NOx
voc
CO
81
1,405
1,340
21
97
112
495
188
1
12
EMISSION DENSITIES
Emission densities for VOC and NOX were calculated for each state or
region within the NECRMP study area, both in terms of emissions per capita and
emissions per square mile. Table 15 presents a summary of VOC and NOX
emissions per-capita for each state. Included are area source, point source,
and total emissions expressed in tons per capita. Area source emissions would
be expected to have little variability on a per-capita basis. As shown in
Table 15, per-capita emissions for area sources of VOC range from
0.0334 tons/capita-year in the District of Columbia to 0.0744 tons/capita-year
in Virginia, while per-capita emissions for area sources of NOX range from
0.0246 tons/capita-year in the District of Columbia to 0.0738 tons/capita-year
in Ohio.
26
-------
TABLE 15. NECRMP STUDY AREA VOC AND NOX EMISSIONS, BY STATE, AS A
FUNCTION OF POPULATION (WITHIN STUDY AREA ONLY)
Annual Emissions Per Capita (tons/capita-year)
Ar««
State
Connecticut
Washington, D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
MEAN
STD. DEVIATION
VOC
.0540
.0334
.0601
.0707
.0471
.0386
.0644
.0564
.0355
.0648
.0393
.0520
.0656
.0744
.0588
.0543
+.0131
NOx
.0364
.0246
.0683
.0456
.0374
.0435
.0493
.0473
.0320
.0738
.0414
.0516
.0504
.0574
.0620
.0481
+.0134
Point
VOC
.0110
.0031
.0357
.0145
.0185
.0185
.0229
.0215
.0104
.0146
.0223
.0141
.0096
.0157
.0107
.0162
+.0076
NOx
.0200
.0100
.0718
.0177
.0329
.0247
.0535
.0278
.0827
.0530
.0453
.0053
.0031
.0230
.2703
.0494
+.0655
Total
VOC
.0650
,0365
.0958
.0853
.0656
.0572
.0873
.0778
.0460
.0794
.0615
.0661
.0752
.0901
.0694
.0705
+.0164
NOx
.0564
.0345
.1402
.0633
.0703
.0683
.1028
.0751
.1147
.1268
.0867
.0569
.0536
.0804
.3323
.0975
+.0712
27
-------
As would be expected, when evaluated on a per-capita basis, point source
emissions are somewhat more variable than area source emissions. Point source
VOC omissions range from 0.0031 tons/capita-year in the District of Columbia
to 0.0357 tons/capita-year in Delaware, while point source NOX emissions
range from 0.0031 tons/capita-year in Vermont to 0.2703 tons/capita-year in
West Virginia.
The above analysis was performed primarily to identify potential
anomolies in the emissions data. To identify "suspicious" values, the mean
and standard deviation were calculated for VOC and NOX for point, area, and
total emissions. States whose per-capita emissions fell more than two
standard deviations from the mean were identified for further investigation.
All state area source per-capita emissions estimates were found to fall within
two standard deviations from the mean indicating that no individual state
values fall outside the range of values expected.
One state value, for Delaware, falls more than two standard deviations
(2.57) from the per-capita point source VOC mean. For point source NOX
emissions, only one value for West Virginia falls outside of the expected
range. These state values are discussed in greater detail in Section 4.
Emission densities were calculated as a function of land area, as
presented in Table 16. Also shown in Table 16 are population densities for
each state. It would be expected that population densities and emission
densities would correlate well on a linear basis. Regression analyses for VOC
and NOX emission densities versus population densities were performed. As
expected, population and emission densities showed a very strong correlation
(r* >0.98) for both VOC and NOX emissions.
Emission densities, both as a function of population and land area are
portrayed graphically in Figures 2 through 13.
PER CAPITA EMISSIONS FROM SELECTED AREA SOURCE CATEGORIES
Some area source categories (e.g., commercial/consumer solvent use, dry
cleaning, etc.) are estimated based on population and a per-capita emission
factor. Others are derived as a function of an activity rate (e.g., gasoline
sales, vehicle miles of travel, etc). One evaluation measure applied to the
NECRMP area source inventory data was to calculate emissions on a per-capita
basis for certain area source categories that were not originally derived
using per-capita emission factors. Three categories were selected for
evaluation:
• residential fuel combustion,
• highway vehicles, and
• gasoline marketing.
28
-------
TABLE 16. NECRMP STUDY AREA VOC AND NOX EMISSIONS AS A FUNCTION
OF LAND AREA (WITHIN STUDY AREA ONLY)
Annual Emissions (tons/square mile-year)
Area
State
Connecticut
Washington, D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
VOC
34.49
349.33
18.03
4.11
19.99
30.13
6.56
55.18
13.62
24.06
10.36
46.96
3.62
8.93
4.37
NOX
23.27
257.11
20.51
2.65
15.86
33.99
5.02
46.35
12.27
27.39
10.92
46.55
2.78
6.89
4.60
Point
VOC
7.05
32.00
10.71
.85
7.84
14.48
2.33
21.04
4.01
5.42
5.87
12.75
.53
1.88
.79
NOX
12.81
104.02
21.56
1.03
13.96
19.32
5.45
27.24
31.71
19.69
11.97
4.80
.17
2.77
20.08
Total
VOC
41.54
381.33
28.74
4.96
27.83
44.62
8.89
76.22
17.63
29.48
16.23
59.71
4.15
10.81
5.16
NOx
36.08
361.13
42.07
3.68
29.82
53.31
10.47
73.58
43.99
47.08
22.89
51.35
2.95
9.65
24.69
Population
Density
Capita/
sq. mi.
639
9,788
300
58
298
781
102
979
384
371
264
903
55
120
74
29
-------
J-l
as
tU
>%
(3
u
•H
P.
CO
y
en
a
o
CD
0)
w
c
0)
CO
c
o
1-1
en
en
•H
a
-------
M
fl
CU
>>
Jl
8-
o
co
c
o
u
CO
0)
co
cu
T)
c
o
•H
CO
CO
•H
§
(U
O
O
CO
CO
(U
M
CO
O
O.
CO
00
•H
fa
31
-------
o
LA LA
-------
cd
cu
CN
CO
o
fl
0)
0)
p.
co
fi
o
•H
co
co
•H
s
cu
cu
o
O
co
tfl
0)
T3
cd
c
•H
O
in
-------
u
«]
0)
C8
CU
fl
O
w
o
co
0)
en
c
CU
T3
fl
O
•H
CO
cn
•-*
a
01
0)
a
u
3
O
M
at
ai
u
CJ
00
34
-------
td
cu
a.
cfl
a
w
cs
o
co
-------
>^
g
_
0
o
V
D
«M
0
o"
1
,_
o
o
Q
f^i
O
O
1
>
O.
nj
U
en
C
o
w
C
o
•H
M
CO
i-t
s
01
HI
y
u
o
co
O
0.
o
o
oo
0)
1-1
00
36
-------
cfl
01
ex
(0
CJ
•-v.
en
c
o
CO
o;
c
OJ
o
•H
co
CO
•H
a
CU
0)
o
o
CO
o
a
o
2
(1)
1-4
60
•H
Pn
37
-------
I
cs
•H
CD
c
o
fl
0)
u
to
14
-------
I
UA
I
O
LTV
-T
D
O
vO
A
M
CO
a)
co
c
O
OJ
a)
M
0)
CO
C
O
•H
tn
co
•H
a
-------
u
ta
OJ
;*>
I
CO
e
o
OJ
a,
w
o
•H
co
en
•H
s
cu
OJ
o
o
co
o
a.
§
00
40
-------
s
v|
•H
CO
a
o
nJ
cu
M
al
QJ
CU
to
c
o
•H
03
co
0)
o
M
O
co
O
a.
m
i—i
-------
The above categories were selected because they: (1) were not derived
using per-capita emission factors; and, (2) would be expected to correlate
with population. Emissions for the above three categories were expressed in
terms of population and compared from state to state. The purpose of this
comparison was to determine the degree of variability from state to state and
identify potential anomolies with the data. Values which deviated
significantly from the regionwide mean were examined in more detail to
determine if the deviation was explainable, a result of differing methodology,
or a probable error. Analyses for the above categories are discussed
individually, as follows.
Residential Fuel Combustion
Residential fuel emissions were calculated on a per-capita basis for each
state, as was the ratio of VOC/NOX emissions, which are shown in Table 17.
The per-capita emissions for VOC range from 0.25 Ib/capita in Massachusetts to
2.48 Ib/capita in Vermont while NOX emissions range from 3.32 Ib/capita in
Virginia to 10.64 Ib/capita in Vermont. The per capita emissions for Vermont
are more than two standard deviations greater than the mean for VOC and nearly
two standard deviations above the NOX mean. Thus, the Vermont data were
selected for further review. Since the Massachusetts VOC emissions, on a per
capita basis, are one order of magnitude lower than Vermont, residential
emissions from these two states were investigated in greater detail. The NEDS
system residential emission totals for these two states are compared with the
NECRMP estimates for these two states in Table 18.
TABLE 18. NEDS VS. NECRMP RESIDENTIAL EMISSIONS FOR MASSACHUSETTS AND VERMONT
Annual Emissions from Residential Combustion (Tons)
VOC NOX VOC/NOX
State NEDS NECRMP NEDS NECRMP NEDS NECRMP
Massachusetts
Vermont
783
306
111
633
13,628
1,414
15,561
2,721
0.057
0.216
0.050
0.233
As shown in Table 18, the NECRMP and NEDS data are in reasonable
agreement for residential emissions in Massachusetts, while the NECRMP
estimates for Vermont are approximately twice as great as the NEDS estimates.
This difference can be attributed to differing methodologies. The NEDS
residential emissions are estimated using annual Department of Energy (DOE)
£ucL use statistics while the NECRMP estimates were calculated by GCA by
employing published housing and degree day information in the fuel use
estimation equations contained in the EPA Guidelines.^ The GCA
42
-------
TABLE 17. PER-CAPITA EMISSIONS FROM RESIDENTIAL FUEL COMBUSTION
Annual Emissions (Ib/capita)
State
Connecticut
Washington, D.C.
Delaware
Maine*
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
REGIONWIDE AVERAGE
I-EAN OF STATE AVERAGES
STD. DEVIATION
VOC
0.35
0.75
0.83
2.06
0.86
0.25
1.51
0.40
0.36
0.60
0.95
0.48
2.48
1.49
0.66
0.62
0.93
±0.66
HO,
4.59
5.98
6.63
8.35
6.56
5.09
8.93
5.39
4.83
4.12
4.66
9.63
10.64
3.32
3.76
4.. 6 8
6.17
+2.26
Ratio VOC/NOX
0.076
0.125
0.125
0.247
0.131
0.049
0.169
0.074
0.075
0.146
0.204
0.050
0.233
0.449
0.176
0.132
0.151
*Residential fuel combustion emissions for Maine reflect a composite of
residential, commercial/institutional, and industrial area source categories.
See Volume IV.4
43
-------
calculations were rechecked and verified to be accurate. Thus, in this
instance, either all of the fuel used in Vermont was not reported by DOE or
the Guidelines19 algorithms overestimate fuel used. Since the Vermont data,
on a per-capita basis, fall approximately two standard deviations from the
mean, the later case is more likely true. (However, it is known that many
Vermont residences that utilize a non-wood fuel as the primary heat source
suppliment with wood that is either self-cut or purchased in such a manner
that DOE will not have a record of its consumption).
Utilizing the NEDS data for Vermont in place of NECRMP still results in
Vermont having five times the per-capita residential fuel combustion emissions
experienced in Massachusetts. Some of the difference can be attributed to
heating degree-days, which are compared below:
City Heating Degree Days
Boston, MA 5,634
Worcester, MA 6,969
Burlington, VT 8,269
The remaining difference can be attributed to: (1) significantly different
mixes of fuels used in the residential sector; and, (2) use of an "adjusted"
VOC emission factor for natural gas combustion in Massachusetts.
A significantly greater percentage of homes in Massachusetts utilize
natural gas for heating than in Vermont:
Fuel Massachusetts Vermont
Fuel Oil 65% 80%
Natural Gas 28% 5%
Electricity 4% 67,
Wood/Coal 0.5% 3.1%
LPG 1% 5%
All Others 1.5% 0.9%
The VOC emission factor utilized for residential natural gas combustion
Ln Massachusetts had been adjusted by the Massachusetts DEQE to reflect
reactive VOC only. A VOC emission factor of 3.2 Ib/lO^ cu. ft. was employed
in Massachusetts, whereas a factor of 8 lb/10^ cu. ft. was utilized in
Vermont and elsewhere in the NECRMP study area. The VOC emission factor used
for all other residential fuels in Massachusetts are consistent with the
remaining NECRMP states. The use of a reactive VOC emission factor in
Massachusetts can be accounted for when the NECRMP inventory is speciated for
input to the Regional Oxidant Model (ROM).
Highway Vehicles
Highway vehicle emissions were calculated for each state in the study
region, as shown in Table 19. VOC emissions range from 99.96 Ib/capita in
Virginia to 39.32 Ib/capita in Pennsylvania. The regionwide averages are
-------
CO
M
33
S!
<
ai
£
CO
ISSION
H
M
Si
u
(A
u
CU
£
3
H
g
j<
i
o
-5
CL Q
(i 1 X
U 1 O
*"* 9
^ u
u §
o
u
a Q
o *
** U
t9
i ®
a
3
j<
i
P
3
u
s
M
i
o
J<
i
g
Os 00 NO
CM M3 00
-- ^r r-
m CM CM
NO »n os
OS
o
o
O en
CM CM
IA m
~
*> S
CM m
O 00
CM
CM
Os CM
00 P"»
a>
00
r* o — i
O vO O
r*> tn u^
O -» N
so tr> o>
2 ^ ^
cn NO -
i-4 ~H CM
r^ 00 r*
r*» CN CM
O O ~*
** O A -* r»*
-< «A Os
S S 2
O 00 oo
1A fl CM
V
l<
• H
-C X
CD V
a. to jrf
Ju u
J? 5
J > »
CO W 41
Z 2 Z
O Os Ps| 00
0 00 ~i O
os os os
.* Os -^ -*
lA
NO %O ^ ^*
-* IA *A r-
S S 3 «
IA os CM cn
CM i-» ^
f^ lA 00 •*
w3 rx CM cn
en CM ao cn
>A O O f-
~* ^f iA 00
cn CM CM
cn CM CN
o o O
r-. CM cn ^
CM OS -* <3
ao NO cn IA
•H
NO oo LA ^.
NO (^ CO f^
o
lA
r-«
°. Ill
CM
00
NO
Os
m
00
lA
ao
— * «^ CM
lA lA 00
CM 1 CM **
West Virginia 28
REGIONWIDE
AVERAGE
WAN 42
STD. DEVIATION »12
T3
•a
13
X
a
V
9
U
(0
c
o
.,4
8
V
*
9
V
i
3
Q.
C
1
U
u
Dh
X
M
0
00
o
H
J
•o
a.
£
s
/chiclet
:ategori
39
« 49
is
•5 2
5.2
45
-------
47.82 Ib VOC/capita and 56.92 Lb N0x/capita. Means of the individual state
values are 58.57 Ib VOC/capita and 63.43 Ib N0x/capita. Since the majority
of highway vehicle emissions are generated by li^ht-duty vehicles (LDV), means
and standard deviations were examined separately t'or LDVs.
For both light-duty vehicles and total vehicles, emissions for Virginia
an; generally two standard deviations higher than the means. All other state
vnlues fell within two standard deviations of the mean. To determine why the
values for Virginia are so high, light-duty vehicle emission factors are
compared Ln Table 20. As shown in Table 20, the Virginia light-duty vehicle
emission factors fall within the expected range. Thus, the anomolies in the
Virginia per capita highway vehicle emissions occur as a result of the implied
VMT per capita.
Emissions reported in NECRMP were compared with those published by the
Virginia State Air Pollution Control Board21 and were found to be
consistent. Since the relationship of VMT per capita in Virginia is much
higher than observed elsewhere, the VMT estimates for Virginia should be
further investigated.
Ratios of VOC/NOX emissions are compared from state to state and by
vehicle class in Table 21. Results in Table 21 for New Jersey appear
anomoLous because New Jerseys' VOC and NOX emissions were forcefully
disaggregated into vehicle classes soley on the basis of VMT distribution,
thus the same VOC/NOX ratio applies to all vehicle classes.
Highway vehicle emissions were calculated by the individual states using
various versions of EPA's mobile source emission factor model. Some states
utilized MOBILE1 while others used MOBILE2. Following completion of most of
the NECRMP inventory, serious errors were detected in the MOBILE2 emission
model prompting the release of MOBILE2.5. Additional corrections are
currently underway and MOBILES is expected to be released in the fall of 1983.
Since highway vehicles account for a significant percentage of total
emissions in the NECRMP region, the inconsistency of emission models used and
the uncertainty introduced by the known errors in the version of MOBILE2 used
in several states raises serious questions about both the accuracy of mobile
source emission in the NECRMP inventory as well as the likely impact on model
results. Recalculation of highway vehicle emissions following the release of
MOBILS3 should be considered.
Ga90line Handling
VOC emissions from various gasoline handling activities are shown in
Table 22 on a per-capita basis. State to state variations in gasoline
handling emissions occur as a result of differing gasoline consumption rates
(on a per-capita basis), and the use of differing emission factors owing to
different percentages of splash, submerged and balance loading and the
implementation of Stage II RACT controls in some areas.
-------
TABLE 20. AVERAGE LIGHT-DUTY VEHICLE EMISSION FACTORS, BY STATE
State
Connecticut
Washington, D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
MEAN
STD. DEVIATION
VOC
11.00
7.99
9.70
10.17
6.69
9.27
7.38
9.02
7.57
7.57
10.56
9.72
10.54
7.63
8.92
+ 1.42
NOX
7.09
7.49
7.08
8.71
7.60
8.68
10.14
6.07
7.68
7.59
9.90
10.35
8.07
6.49
8.07
+ 1.33
47
-------
TABLE 21. VOC/NOX RATIOS FOR HIGHWAY VEHICLE EMISSIONS
State
Connecticut
Washington, D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
LDV
—
1.55
1.07
1.37
1.17
.88
1.07
.73
1.49
.99
1.00
1.07
.94
1.31
1.18
LDT1
—
1.59
1.03
1.44
1.27
1.04
1.04
.73
1.50
1.04
1.09
1.02
.92
1.46
.99
LDT2
—
1.59
.99
1.44
1.11
1.04
1.06
.73
1.50
1.01
.98
.98
.87
1.05
— —
HDG
—
.27
.84
.98
.59
.75
.91
.73
1.51
.76
.77
.84
.75
1.11
.95
HDD
—
.268
.110
.985
.163
.094
.114
.73
.159
.090
.091
.111
.093
.123
.115
MC
—
—
10.32
—
14.95
12.51
14.78
0.73
11.63
10.03
0.66
11.15
9.19
—
—
Total
1.08
1.30
0.87
1.30
1.02
0.77
0.91
0.73
1.03
0.66
0.66
0.90
0.94
1.10
0.80
-------
TABLE 22. PER CAPITA EMISSIONS FOR GASOLINE HANDLING (POINT AND AREA SOURCES)
Annual Emissions (Ib/capita)
State
Connecticut
Washington, D.C.
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New York
Ohio
Pennsylvania
Rhode Island
Vermont
Virginia
West Virginia
REGIONWIDE
AVERAGE
MEAN
STD. DEVIATION
Stage I
gasoline
evaporation
3.20
.08
4.12
5.55
1.20
1.28
3.63
.58
2.22
4.45
3.42
2.85
3.99
2.86
3.63
..__
Stage II
gasoline
evaporation
4.25
2.65
4.66
4.68
2.76
3.67
4.18
4.58
2.95
5.13
3.67
3.79
4.63
5.78
4.22
___
Storage
tank
breathing
.44
.27
.49
.49
.29
.38
.44
.47
.30
.53
.41
.39
.48
.35
.43
— —
Gasoline
loading/
transit
0.33
1.87
3.38
7.25
1.95
0.73
2.98
3.71
1.12
3.74
1.79
5.85
3.28
3.75
2.99
__-.
Total
8.22
4.88
12.65
17.95
6.23
6.06
11.23
9.34
6.58
13.86
9.29
12.88
12.38
12.74
11.26
8.73
10.37
+3.58
49
-------
As shown in Table 22, per capita emissions from stage I evaporation
appear quite Low for the District of Columbia and New Jersey and quite high
for Maine. These differences can be attributed to different assumed mixes of
splash, submerged, and balance filling. The District of Columbia Council of
Governments assumed that all service station loading was performed using
balance filling (0.3 Ib/thousand gallons). Similiarly, a considerable portion
of New Jersey (AQCRs 043 and 045) also reflects balance filling only.
Conversely, the Maine Department of Environmental Protection assumed all
service station loading was performed using splash filling (11.5 Ib/thousand
gallons). These factors account for most of the differences in Stage I
emissions. The remaining differences can be attributed to varying gasoline
consumption rates.
Gasoline loading/transit losses were examined on a per-capita basis. As
shown in Table 22, values for Connecticut and Massachusetts appear quite low,
while values for Maine and Rhode Island appear high. The Maine and Rhode
Island loading losses come primarily from large point sources. Both states
house large gasoline terminals which serve as major transfer points for areas
that go beyond the state boundaries. This could, in part, also account for
the low values in Connecticut and Massachusetts. Further, gasoline loading
losses that were supplied for the metropolitan Boston area were calculated
assuming 100 percent balance filling. The Connecticut area source inventory
covered only transit losses for the category, with loading losses a'ssumed to
be adequately covered in the point source inventory. With the exception of
Maine, discussed above, total emissions from gasoline marketing for all
states, on a per-capita basis, fall within two standard deviations from the
mean.
POINT SOURCE INVENTORY COMPREHENSIVENESS
Numbers of facilities and individual emission points covered in each
state component of the NECRMP inventory are provided in Table 23.
Comprehensiveness of the point source emissions data was evaluated for each
state by comparing facilities included in each state's inventory against those
listed in the Directory of Volatile Organic Compound Sources Subject to
Reasonably Available Control Technology Requirements (RACT Directory). This
comparison was actually performed as a quality assurance measure during the
point source compilation phase of NECRMP. The intention of this check was to
identify and subsequently include VOC sources that had been omitted in the
individual state data bases. For the entire NECRMP study, over 548 sources
were identified as being included in the RACT Directory, but not in the
state-supplied data. This exercise resulted in the addition of six facilities
to the NECRMP data base (one in Rhode Island and five in Massachusetts)
totaling over 9,000 tons of VOC.
GCA fowarded, to each state agency contact, a list of potential omissions
based on a comparison of the state-supplied data with the RACT Directory.
State agency contacts who responded to GCA's lists of potential emissions
usually indicated that the apparent omissions were for one or more of the
following reasons:
50
-------
TABLE 23. NUMBERS OF FACILITIES AND INDIVIDUAL POINT SOURCES COVERED
IN THE NECRMP2~16 INVENTORIES
State
Connecticut
Delaware
Maine3
Maryland
Maaaachusetts
New Hampshire
New Jersey
New York
Ohioa
Pennsylvania
Rhode Island
Vermont
Virginia*
Washington, D.C.
West Virginia8
REGION TOTALS
NECRMP
Facilities
3,342
472
487
131
1,274
316
331b
1,145
731
2,137
90
162
1,628
102
309
12,657
(1979/1980)2-16
Points Points/facility
10,239
1,277
713
1,477
4,363
459
13,329
2,564
2,980
13,590
319
331
2,641
207
985
55,474
3.1
2.7
1.5
11.3
3.4
1.5
17.6
2.2
4.1
9.7
3.5
2.0
1.6
2.0
3.2
5.4
aThe NECRMP data include only a portion of this state.
bNEDS "Card 1" data were provided for 14,643 facilities while complete
records were provided for 1,749 facilities, reflecting 30,742 emission points.
Further, to properly "balance" with the New Jersey area source inventory,
only facilities emitting more than 100 TPY are included in the NECRMP point
source inventory.
51
-------
• che facility has ceased operation or transferred ownership,
• the facility was listed in the inventory under a different name,
• the facility emitted less VOC than the state's minimum source size
cutoff,
• the source was incorrectly listed in the RACT Directory.
Nevertheless, since some states had nearly perfect "RACT coverage", it is
believed that relative coverage of "RACT" sources is a reasonable measure of
the comprehensiveness of the point source inventories.
The RACT Directory22 attempts to list all VOC sources subject to RACT
regardless of size. Many of the point source files supplied by the NECRMP
states reflected a minimum source size cutoff and therefore would not have
included small VOC emitters listed in the RACT Directory.22 This explains
the apparant low coverage of sources in New York, for example. Also, many of
the facilities listed in the RACT Directory22 as industrial Perchloro-
ethylene Drycleaners are questionable. This suspicion was confirmed at
least for Massachusetts, by GCA staff members familiar with industrial
perchloroethylene drycleaners in Massachusetts. On the other hand, some
sources listed in the RACT Directory22 but not in individual state point
source files are suspected to be significant emitters. Thus, while a very low
percentage of RACT facility coverage does not guarantee that major VOC sources
have been omitted, it does cast some uncertainty on the comprehensiveness of
the states' point source data. Conversely, a very high percentage of RACT
sources in a state inventory does provide some assurance of the
comprehensiveness of the data for these states.
Table 24 provides a comparison of the relative coverage of RACT sources
in the NECRMP inventory. As shown in Table 24, two states, Virginia and
Delaware had virtually 100 percent coverage of these sources in their
inventories. It is the author's opinion that these states' inventories were
outstanding in terms of comprehensiveness.
ARKA SOURCE INVENTORY COMPREHENSIVENESS
Area source inventory comprehensiveness can be defined in terms of
geographic coverage, area source category coverage, and resolution of area
source categories. Area source emission inventories were compiled for all
areas within the entire NECRMP study area. However, some variations exist
between individual state components in terms of category coverage and
resolution, and in one instance, geographic resolution.
The primary objectives of the area source inventory development were:
• to ensure all VOC and NOX emissions had been accounted for, and
were adequately resolved into individual source categories to enable
proper temporal and organic species disaggregation to support
photochemical modeling, and
52
-------
TABLE 24. COVERAGE OF "RACT" SOURCES IN THE NECRMP DATA BASEa
Facilities listed
in the 2?
"RACT" directories
Connecticut
Delaware
Maine
Maryland
Massachusetts
New Hampshire
New Jersey
New Yorkb
Ohio
Pennsylvania'3
Rhode Island
Vermont
Virginia
Washington, D.C.b
West Virginia^3
Region Totals
79
30
35
69
123
11
336
232
233
245
43
10
90
3
49
1,588
"RACT" directory
sources omitted
in NECRMP
(after resolution)
14
1
4
31
44
4
44
159
62
138
19
4
1
2
15
542
Percent
coverage
82
97
89
55
64
64
87
31
73
44
56
60
99
33
69
66
aDoes not include sources identified as having been closed or included under
a different name. Individual facilities are identified in specific state
volumes.
states did not respond to lists of potential omissions forwarded by GCA
as part of the quality assurance review and therefore may not actually have
as poor a coverage as indicated above since facilities that may have been
closed or had their ownership transferred were not identified.
53
-------
• to ensure complete geographical coverage and adequate spatial
resolution to enable further disaggregation to grid cells.
Spatial Resolution
The area source inventories for NECRMP were compiled on a county level
for all but one state within the NECRMP study area. The single deviation
involves the Commonwealth of Massachusetts, whose Department of Environmental
Quality Engineering utilizes Air Pollution Control Districts (APCD) for
virtually all aspects of air quality management, including emission
inventories. The Massachusetts APCDs, and the relationship to counties is
portrayed in Figure 14. As shown in Figure 14, the Massachusetts APCDs do not
adhere to country boundaries. To accomodate Massachusetts' deviation from the
other NECRMP states, the following steps were taken: (1) the Massachusetts
point and area source inventories were compiled on an APCD basis, and
(2) spatial allocation factors^? for Massachusetts were altered to enable
spatial resolution to NECRMP grids from APCDs rather than counties. Since the
Massachusetts spatial allocation factors have been based on APCDs, this
deviation should not adversely impact the use of the Massachusetts data in
photochemical modeling.
Area Source Category Coverage and Category Resolution
One objective of the area source inventory development was to ensure
sufficient resolution of area source emissions into categories to enable
detailed temporal and spatial resolution for photochemical modeling. A total
of 54 unique area source categories were defined for use in NECRMP. These
categories are shown in Table 25. Not all of the Table 25 categories were
covered in all 15 state inventories. In some instances, certain state
agencies determined that some of the less significant categories did not
contribute sufficient VOC and NOX emissions to warrant their inclusion. In
other instances, emissions representing a number of the Table 25 categories
were aggregated to form composite categories. For these latter cases, two
courses of action were taken by GCA. First, when sufficient data were
available from the states or when national data were adequate to allow their
independent disaggregation, GCA calculated category-specific estimates from
the composite data. When such information was unavailable, GCA included the
composite data in that state's EIS/AS masterfile and properly "commented" the
composite category so as to identify its component constituents. Since these
adjustments were unique to each state, they are discussed individually, by
state, in Section 4.
54
-------
C
o
o
c
to
CO
•H
T3
c
o
u
c
o
o
a
M
•H
cfl
CO
-------
TABLE 25. AREA SOURCE CATEGORIES INVENTORIED
FOR THE NECRMP STUDY
Category number Category description
001 Stage I Gasoline Evaporation
002 Stage II Gasoline Evaporation
003 Storage Tank Breathing
004 Gasoline Loading/Transit
005 Small Industrial/Commercial Degreasing
006 Dry Cleaning
007 Architectual Surface Coating
008 Auto Body Refinishing
009 Small Industrial Surface Coating
010 Graphic Arts
Oil Commercial/Consumer Solvent Use
012 Cutback Asphalt
013 Pesticides
014 On-Highway Light Duty Vehicles
015 On-Highway Light Duty Trucks - Class I
016 On-Highway Light Duty Trucks - Class II
017 On-Highway Heavy Duty Gas Trucks
018 On-Highway Heavy Duty Diesel Trucks
019 On-Highway Motorcycles
020 Residential Anthracite Coal
021 Residential Bituminous Coal
022 Residential Residual Oil
023 Residential Distillate Oil
024 Residential Natural Gas
025 Residential LPG
026 Residential Wood
027 Commercial/Institutional Anthracite
028 Commercial/Institutional Bituminous
029 ConanerciaI/Institutional Residual Oil
030 Conanercial/Institutional Distillate Oil
031 Commercial/Institutional Natural Gas
032 Commercial/Institutional LPG
033 Commercial/Institutional Wood
034 Industrial Anthracite
035 Industrial Bituminous
036 Industrial Residual Oil
037 Industrial Distillate Oil
038 Industrial Natural Gas
039 Industrial LPG
040 Industrial Wood
041 Military Aircraft
042 Civil Aircraft
043 Commercial Aircraft
044 Railroad Locomotives
045 Gasoline Powered Vessels
046 Distillate Oil Powered Vessels
047 Residual Oil Pownred Vessels
048 Off-Highway Vehicles - Gas
04'> Off-Highway Vehicles - Diesel
0)0 On-Site Incineration
n^i Open Burning
052 Structural Fires
053 Field/Slash Burning
054 Forest Fires
56
-------
SECTION 4
DATA QUALITY REVIEW
INTRODUCTION
This section presents a state-by-state review of the overall quality of
the NECRMP inventory data. It includes a discussion of the results of the
evaluation techniques presented in Section 3 as they relate to each state.
Also included are lists of uncertainties, deficiencies, and recommendations
for future corrective actions. Commentary on the overall suitability of the
NECRMP data base for use in photochemical modeling is offered.
The quality of the NECRMP point source inventories were evaluated in
terms of currency (year represented by the data), missing data elements,
residual problems, and apparent completeness for other pollutants (TSP, SO^
and CO). Although the point source inventories were subjected to considerable
quality assurance checks and subsequent corrections, not all problems
identified were fully rectified due to time and resource limitations. Error
corrections were prioritized based on source size, severity of the error, and
likely impact photochemical modeling results.
The quality of the NECRMP area source inventories were evaluated in terms
of consistency with the EPA-prescribed methodologies,^ age of the data, and
deviations from generally accepted emission factors. Area source inventories
were also examined in terms of comprehensiveness, which addresses both
geographical coverage and area source category coverage and resolution.
The NECRMP point and area source emissions were previously compared to
NEDS and the 1982 Ozone SIPs in Section 3. Differences identified in that
comparison are examined on a state-by-state basis in the remainder of this
section.
CONNECTICUT
The Connecticut Department of Environmental Protection (DEP) provided the
input data used in the NECRMP point and area source inventories for
Connecticut and assisted in the quality assurance review of the data by
supplying required corrections and additional data as was necessary.
Point Source Inventory Quality
The NECRMP point source inventory for Connecticut reflects 1979 and 1980
data and was supplied in NEDS format as converted from Connecticut's STARTER
57
-------
ays Cera. A number of data formatting problems were addressed by GCA prior to
incorporation of the data into the NECRMP inventory. These included
elimination of repetitive NEDS Card 1 records for each point, correction of
erroneous hexideciraal bit patterns (internal computer formatting) for
alphabetic point IDs, and duplicate point ID assignments. Also, a large
number of incomplete and/or incorrect SCC codes were corrected with the
assistance of the DEP. Since the inventory, as received, was virtually
incapable of conversion into EIS format, the corrections in this regard
represent a considerable improvement. However, some problems remain. One
"generic" correction undertaken involved seasonal emissions distribution
percents. Many of these fields, as received, were blank or zero filled.
These fields were changed to reflect an equal seasonal distribution, i.e.,
four 25 percent values. This may impact the temporal allocation of the
emissions data since temporal factors for most point source categories are
derived from the operating rate data contained on the point source records.
However, for power plants, the largest point source NOX emission category,
the percent annual throughput fields will be overridden by the temporal
allocation factors reported in Volume XVII.^
Area Source Inventory Quality
The Connecticut area source inventory, representing a 1979* base year,
was supplied in hard copy format by the Connecticut DEP. The inventory
reflected county summaries of emissions of all five criteria pollutants from
52 area source categories. However, these categories did not correspond well
with the 54 NECRMP area source categories. Two major groups of area source
emissions, gasoline handling and highway vehicles, had been aggregated, and
six categories had been omitted. Most of these problems were resolved with
the assistance of the DEP. The major residual problem with the Connecticut
data is the aggregation of on-highway vehicle emissions. Vehicle type
specific emissions were reportedly unavailable. All highway vehicle emissions
were reported in a single composite category. For the remaining categories,
the DEP data conformed resonable well with the EPA-prescribed
methodologies.19
One deviation from the EPA-prescribed methods relates to VOC emissions
from dry cleaning. The DEP estimated dry cleaning solvent usage using a 3 lb
solvent/capita factor. Emissions were derived with an emission factor of
210 lb VOC/ton of solvent. The EPA-recommended factor of 1.5 lb VOC/capita
would have resulted in an increase in the VOC emissions estimate of
approximately 1,900 tons/year. Since the DEP believed their methodology to be
more consistent with actual conditions in Connecticut, their estimates were
maintained in NECRMP.
The Connecticut DEP indicated that pesticide usage data were unavailable
and, therefore, did not supply an estimate of VOC emissions from this
*Except per-capita based emissions, which were recalculated by GCA using data
from the 1980 census.
58
-------
category. For the purposes of NECKMP, GCA calculated emissions from the
application of pesticides assuming an emission rate of 3.5 Ibs/harvested
acre. Also, the DEP emissions estimate from off-highway vehicles accounts for
diesel fuel only.
Comparisons with the NEDS and SIP Inventories
The NECRMP emissions data for Connecticut are compared to NEDS and SIP
inventory estimates in Table 26.
As shown in Table 26, NECRMP VOC estimates are somewhat higher than those
reported in the SIP but lower than that reported in NEDS. NECRMP NOx
emissions compare reasonably well with the SIP and NEDS estimates for the SIP
areas, but are about 13 percent higher than the statewide estimate in NEDS.
Virtually all of the VOC difference is related to area sources of organic
solvent evaporation. For area sources, NEDS shows 130,628 tons VOC/year,
while NECRMP shows 31,690 tons VOC/year from solvent evaporation. The reason
for this difference relates to different interpretations of organic solvent
sales data20 between the EPA Guidelines19 and those employed in NEDS, as
discussed earlier in Section 3.
Most of the NOX difference between the reported NEDS and NECRMP
estimates relates to electric generation emissions, NEDS shows
25,745 tons N0x/year while NECRMP shows 44,157 tons NOx/year. Some of
this difference may relate to different years of record between NEDS (1981/82)
and NECRMP (1979/80). Electric generation emissions are updated in NEDS using
DOE data, independent of the states, whereas the NECRMP estimates are believed
to be consistent with the DEP estimates for 1979.
Apparent Completeness for Other Pollutants
The Connecticut point source inventory also contains data on TSP, S02
and CO emissions, which are compared to the NEDS system totals below. NECRMP
data correction efforts did not focus on these pollutants, but rather,
emphasized VOC and NOX.
NECRMP point
NEDS point
(41,746) 3,146 (51,013)
A complete discussion of the Connecticut inventory is presented in
Volume II.2
TSP (tons/year)
16,231
57,977
SO? (tons/year)
51,492
48,346
CO (tons/year)
14,268
71,281
59
-------
TABLE 26. COMPARISON OF NEDS, SIP AND NECRMP EMISSIONS FOR
CONNECTICUT
S [ V urea
Hartford
PoinC
Area
Total
New Haven
Point
Area
Total
Statewide
Point
Area
VOC
NEDS
4,410
61,740
66,150
28,317
57,530
85,847
26,386
259,652
(tons/year)
SIP NECRMP
7,242 2,901
36,211 46,243
43,453 49,144
10,058 22,934
30,176 37,826
40,234 60,760
34,295
167,730
NOx
NEDS
7,843
27,988
35,831
16,710
24,054
40,764
49,519
105,793
(tons/year)
SIP NECRMP
8,047 5,219
27,761 31,483
35,808 36,702
20,922 15,629
25,750 25,905
46,672 41,534
62,292
113,160
Total
286,038
202,025 155,312
175,452
60
-------
DELAWARE
The Delaware Division of Environmental Control (DEC) provided the input
data used in the Delaware point source inventory. The area source inventory
for Delaware was developed primarily by GCA using the EPA Guideline^
methodologies. The DEC assisted in this process by supplying much of the
input data utilized, reviewing the GCA-developed inventory, and suggesting
some changes which were subsequently carried out.
Point Source Inventory Quality
The Delaware point source inventory was obtained from the Delaware DEC in
EIS/P&R format and reflected 1980 data. In general, the Delaware inventory
was found to be complete and exceptionally well prepared. Problems that were
detected primarily related to duplicate SCC's at single points and UTM
coordinate errors. The former errors were completely corrected as EIS/PS will
reject duplicate SCCs at single points, thus assuring their identification.
All UTM coordinates that fell "out of range" were addressed since these were
identified in the EIS edit checking program. All suspicious UTMa at large
facilities were also addressed since these sources were subjected to a manual
review. There is little assurance, however, that UTM coordinates on minor
facilities that did not fall "out of range" were all identified. However,
since 95 percent of all point source NOx emissions and 97 percent of all VOC
emissions are represented by "major" facilities (these emitting more than
100 tons per year of VOC and/or NOx), resulting errors in photochemical
modeling are expected to be insignificant.
Area Source Inventory Quality
When contacted in December 1980, the Delaware DEC had indicated that they
intended to utilize projections of a 1976 base year inventory for their 1982
State Implementation Plan submittal.3 Since one of the program objectives
was to avoid data older than 1978,3 GCA developed a 1980 area source
inventory of VOC and NOX emissions using state-supplied and published
activity data, AP-42 and other EPA-approved emission factors, and
methodologies generally consistent with the Procedures for the Preparation of
Emission Inventories for Volatile Organic Compounds—Volume I.iy
For some area source categories, the methodologies described in
Volume 1^-9 had to be slightly modified to accommodate the source activity
data available for Delaware. Additionally, the quality of existing activity
data necessitated making certain assumptions or adjusting older data to
reflect 1980. These assumptions, adjustments, and deviations from the
Volume 1^9 methodologies are discussed in detail in Volume III.'
Comparisons with NEDS and SIP Inventories
The NECRMP emissions data for Delaware are compared to the NEDS and SIP
inventory estimates in Table 27.
61
-------
TABLE 27. COMPARISON OF NEDS, SIP AND NECRMP EMISSIONS FOR DELAWARE
SIP area
VOC (tons/year)
NEDS SIP NECRMP
NEDS
(tons /year)
SIP NECRMP
New Castle Co.a
Point
Area
Total
16,664
30,585
21,026
20,576
47,249 41,602
20,041
19.612
39,653
16,338
18,793
35,131
27,117
29.421
56,538
24,709
26,809
51,518
Statewide
Point
Area
Total
17,429
47,081
64,510
21,222
35,743
56,965
50,921
32.690
83,611
42,730
40,649
83,379
*Part of Philadelphia SIP.
As shown in Table 27, the NECRMP and NEDS area source inventories differ
by approximately 11,000 tons VOC/yr. Essentially all of this deviation
results from a single NEDS area source category, solvent evaporation loss.
The reasons for this difference, as previously discussed, involve differing
interpretations of solvent use data between those used to estimate emissions
in NEDS and those used to develop the per-capita factors reported in the the
EPA Guidelines.19 The SIP and NECRMP inventories appear to be in
reasonable agreement for New Castle County while the NEDS inventory is
somewhac higher for VOC and lower for NOx in that County. Statewide NOx
totals between NEDS and NECRMP also appear to be in reasonable agreement
overall although the distribution between point and area source emissions
differ.
Apparent Completeness for Other Pollutants
The Delaware point source inventory also contains data on TSP, S02 and
CO emissions, which are compared to the NEDS system totals below.
NECRMP point
NEDS point
Deviation
TSP (tons/year)
42,066
38,973
SO? (tons/year)
132,463
125,198
CO (tons/year)
9,126
2,605
3,093
7,265
6,521
62
-------
A complete discussion of the Delaware inventory is presented in Volume
MAINE
The Maine Department of Environmental Protection (DEP) provided the input
data used in the NECRMP point and area source inventories for Maine and
assisted in the quality assurance review of the data by supplying required
corrections and additional data as necessary*
Point Source Inventory Quality
The Maine point source inventory reflecting 1979, was obtained from the
Maine DEP in NEDS format. The most frequently encountered problems detected
were missing UTM coordinates and exhaust flow rates. The Maine DEP provided
all missing UTM coordinates, and flow rates for all major sources were either
supplied by the DEP or calculated by GCA using agreed upon default equations,
which are presented in Volume I.I Known remaining problems with the Maine
point sources include missing flow rates for minor sources. Since these will
be modeled as area sources in NECRMP, and not as individual point sources,
their omission is of little consequence.
Area Source Inventory Quality
The Maine 1979 area source inventory was supplied by the Maine DEP. A
portion of the inventory was supplied in hard copy format several months
before the complete inventory was supplied in EIS/AS format. The Maine EIS
file did not conform to the category numbering scheme selected for NECRMP.
Rather, the Maine DEP utilized the NEDS numbering system. The Maine EIS file
was renumbered to match the NECRMP category codes and was merged with the
previously supplied data which had been coded into EIS/AS format by GCA. For
a number of categories, the Maine EIS/AS files reported only emission totals.
The emission factors, in these instances, were obtained from the hard copy
documentation of the area source inventory. In turn, the process rates for
these categories were derived by GCA from the DEP reported emissions and
emission factors. The Maine area source inventory did not address
architectural surface coating, autobody refinishing, and small industrial
surface coating. Emissions from these categories were inventoried by GCA
using the methods presented in Volume I.
For several area source categories, the Maine DEP"a inventory reflected
methodologies that differed from those recommended in Volume I.1 These are
highlighted as follows.
Degreasing—
Rather than employ the per-capita method recommended in Volume I, ^ the
Maine DEP utilized data obtained from an inventory questionnaire. All
degreasers emitting less than 10 tons/yr were considered area sources. As a
result, the Maine DEP estimate of VOC emissions from degreasing operations is
approximately 3,000 tons/yr less than would have been estimated had the
per-capita method been employed. Apparently, the Maine DEP felt that the
questionnaire survey was comprehensive enough to justify the deviation.
63
-------
Dry Cleaning—
Rather than utilize the per-capita approach suggested in Volume I,19
the Maine DEP relied on survey data to develop the VOC emissions estimate for
dry cleaning. Use of the per-capita method would have resulted in a
260 ton/yr increase (approximately 60 percent higher) in emissions from this
source category. The DEP believes the survey to be more accurate for Maine
than the national average per-capita factor.
Commercial/Consumer Solvent Use—
The Maine DEP utilized solvent sales data obtained from three companies
to estimate VOC emissions from this category. The DEP estimate may not have
adequately accounted for household use of cleaners, etc. Had the per-capita
approach been employed, the VOC emissions estimated for this category would be
approximately 2,300 tons/yr higher. However, the DEC believes the national
per-capita approach to be less accurate for Maine.
Fuel Combustion—
The Maine area source inventory reported emissions from fuel combustion by
fuel type, but not by user sector. As a result, all area source fuel
combustion emissions were included under residential use, except LPG, for
which use-specific data were available. This will impact temporal
distribution of the emissions, but not speciation, which is based primarily on
fuel type.
Civil Aircraft—
The Maine DEP concluded that emissions from LTOs of civil aircraft were
insignificant, based on survey data. Therefore, emissions from this category
were not included.
Comparison with the NEDS Inventory
The NECRMP emission data for the study area portion of Maine are compared
to the NEDS inventory on Table 28.
TABLE 28. COMPARISON OF NEDS AND NECRMP EMISSIONS
FOR THE NECRMP STUDY AREA PORTION OF MAINE
VOC (tons/year) NOx (tons/year)
NEDS NECRMP NEDS NECRMP
Point
Area
Total
12,718
54,172
66,890
13,651
66,403
80,054
11,462
31,010
42,472
16,646
42,783
59,429
64
-------
TSP (tons/year)
10,811
11,733
SO? (tons/year)
60,578
75,692
CO (tons/year)
22,530
29,026
As shown in Table 28, point source emissions from VOC are in reasonable
agreement, while the NECRMP NO* estimate is somewhat higher than NEDS. The
bulk of this difference can be attributed to industrial boiler emissions.
Much of the area source difference can be attributed to highway vehicles.
Differences in the year of record between NEDS (1981/82) and NECRMP (1979) may
account for this difference, the NECRMP highway vehicle estimates were
obtained directly from the Maine DEP and are believed to accurately represent
1979 highway vehicle emissions.
Apparent Completeness for Other Pollutants
The Maine point source inventory includes data on TSP, SC>2 and CO
emissions, which are compared to the NEDS system totals, below, for the NECRMP
study area portion of Maine.
NECRMP point
NEDS point
Deviation (922) (15,114) (6,496)
A complete discussion of the Maine inventory is presented in Volume IV,^
MARYLAND
The Maryland Bureau of Air Quality and Noise Control (BAQNC) provided the
input data used in the NECRMP point and area source inventories for Maryland
and assisted in the quality assurance review of the data by supplying required
corrections and additional data as necessary.
Point Source Inventory Quality
The Maryland point source data were supplied by the Maryland BAQNC in
NEDS format, as converted from the Maryland Premise Files. The data reflect
1980.
A number of "generic problems" resulted from the conversion of the
Premise File data into NEDS format. These included redundant NEDS card 1
records and dropped trailing zeros in NEDS fields with implied decimal
points. These generic problems were addressed using GCA-developed computer
routines.5 in addition to these problems, the most frequent errors
encountered were incomplete or missing SCC codes, stack parameters, and UTM
coordinates. A number of "out of range" UTM-Y coordinates were detected as
well.
All of the incomplete or missing SCC codes were addressed. Due to the
EIS system's handling of these problems (card rejection), their complete
correction is reasonably ensured. However, missing stack data were not sought
65
-------
for minor point sources and, therefore, remain incomplete. Since these point
sources will not be modeled as individual points in NECRMP, these remaining
problems will be of little consequence in that regard. All missing UTMs were
obtained from the BAQNC and are incorporated in the final data base.
Similarly, "out of range" Y coordinates were corrected with the assistance of
the BAQNC. However, since Maryland crosses a UTM zone boundary, virtually any
UTM-X coordinate would pass the EIS edit check, therefore no UTM-X "out of
range" errors were detectable by the EIS system edit checks. During the
manual review of major sources, inconsistencies in UTMs between points within
any facility were identified. However, if all points within a facility had
the same set of incorrect UTM-X coordinates they were not likely noticed.
Since there were a number of Y-coordinate errors, it must be assumed that some
incorrect X-coordinates could exist, particularly among small sources. Since
the overwhelming majority of point source VOC and NOX emissions in Maryland
are attributable to major sources (over 99 percent), this potential problem
will be of minor consequence in photochemical modeling.
The BAQNC, when responding to GCA's error list, indicated that two major
facilities, Eastern Stainless and Bethlehem Steel, were being reinventoried.
These new data were not received by GCA in time for their inclusion in the
Maryland point source inventory. When completed, these data should replace
the current information for these two sources.
Area Source Inventory Quality
The 1980 Maryland area source inventory was supplied in hard copy format
by the Maryland BAQNC. The inventory for portions of Maryland that are in
attainment of the National Ambient Air Quality Standard (NAAQS) for ozone was
completed by NUS Corporation, under contract to the BAQNC. For nonattainment
areas, the BAQNC developed the inventory.
Although the methodologies reported by NUS and the BAQNC were generally
consistent with those recommended in Volume I^, annual emission totals,
activity rates, and emission factors on a county-specific basis had not been
provided for all categories. Also, a number of categories had been omitted.
The BAQNC supplied most, but not all, of the data requested by GCA as a result
of the data review. Some deviations from the EPA Guidelines^ are
identified below.
Small Industrial/Commercial Degreasing—
According to the BAQNC, all degreasing emissions for the nonurban
portions of Maryland are accounted for in Maryland's point source inventory.
Therefore, area source emissions from this category were assumed to be zero in
the nonurban areas. This resulted in a VOC estimate that is approximately
2,800 tons/yr lower than would have been estimated by the Volume I*-* method.
Small Industrial Surface Coating—
According to the BAQNC, all surface coating emissions are covered in the
point source inventory. Area source emissions from this category were assumed
to be zero.
66
-------
Fuel Combustion—
The BAQNC was unable to disaggregate fuel combustion emissions by fuel
type. Composite categories were utilized for residential and commercial/
institutional fuel combustion. This will have a significant impact on VOC
speciation since species profiles are fuel-specific.
According to the BAQNC, all industrial fuel combustion emissions are
covered by the Maryland point source inventory. Therefore, area source
emissions from these categories were assumed to be zero. The Maryland
emission totals for all fuel combustion categories exceed these reported in
NEDS, which gives some weight to this assumption, even though NEDS shows over
5,000 tons/yr of NOx and close to 1,200 tons/yr of VOC from area source
industrial fuel combustion.
Other Composite Categories—
The Maryland area source inventory also fails to disaggregate emissions
from the following composite categories: aircraft (civil, commercial,
military), oil-powered vessels* (distillate, residual), off-highway vehicles
(gasoline, diesel), and on-highway light-duty trucks (classes I and II). Use
of composite categories will adversely affect VOC and NOx speciation.
Temporal factors will not seriously be affected by these aggregations, except
aircraft, for which separate hourly patterns were derived for civil,
commercial, and military aircraft operations.
Comparison with NEDS and SIP Inventories
The NECRMP emissions data for Maryland are compared to the NEDS and SIP
inventory estimates in Table 29.
For the Baltimore SIP area, the NEDS, SIP, and NECRMP estimates for VOC
are quite variable, while agreement for NOX is significantly better. In
both instances, the NECRMP estimates fall almost midway between the SIP and
NEDS estimates.
Area source emissions reported in the (December, 1981 version) Maryland
SIP do not include four categories that are included in the NECRMP inventory;
Degreasing, Dry Cleaning, Autobody Refinishing, and Graphic Arts, which
account for much of the VOC difference. The SIP VOC point source data are
more than 30 percent lower than NECRMP, and should be further investigated.
On a statewide basis, the NEDS and NECRMP estimates for NC^ are within
less than 10 percent of each other. For VOC, however, the NEDS system totals
are significantly higher. Much of the difference occurs as a result of the
organic solvent evaporation differences between NEDS and the EPA Guidelines,
as previously discussed in Section 3. For point sources, much of the
difference between NEDS and NECRMP can be attributed to two categories: in-
process fuel use and unclassified fuel combustion.
*In nonurban areas, the composite category also includes pleasure craft
(gasoline-powered).
67
-------
TABLE 29. COMPARISON OF NEDS, SIP, AND NECRMP EMISSIONS FOR MARYLAND
SIP area
Baltimore
Point
Area
Total
Statewide
Point
Area
VOC
NEDS
62,231
152,793
215,024
60,092
263,589
(tons/year)
SIP NECRMP
49,890 72,691
78,054 90,806
127,944 163,497
77,557
197,699
NO;
NEDS
32,722
81,983
114,705
116,673
153,882
x (tons/year)
SIP NECRMP
57,132 52,290
66,386 66,516
123,518 118,806
138,095
156,845
Total
323,681
275,256 270,555
294,940
68
-------
Apparent Completeness for Other Pollutants
The NECRMP point source inventory for Maryland includes estimates for TSP
SC-2 and CO emissions, which are compared to the NEDS system totals, below.
TSP (tons/yr) SOy (tons/year) CO (tons/year)
NECRMP point 87,313 447,997 39,139
NEDS point 29,137 297,743 52,176
Deviation 58,176 150,254 (13,037)
A complete description of the Maryland inventory is presented in
Volume V.5
MASSACHUSETTS
The Massachusetts Department of Environmental Quality Engineering (DEQE)
provided the input data used in the NECRMP point and area source inventories
for Massachusetts and assisted in the quality assurance review of the data by
supplying a revised point source tape reflecting corrections to problems
identified by GCA and supplying required corrections for area source data as
necessary.
Point Source Inventory Quality
The Massachusetts point source inventory was obtained from the
Massachusetts DEQE, through EPA. The inventory reflected 1979 data and was
supplied in NEDS format as converted from the DEQE's, EIS/P&R file. A number
of generic problems existed throughout the Massachusetts NEDS file which were
addressed using GCA-developed computer routines.6 This included
inconsistent coding of the emission control device and efficiency fields,
incorrect common stack fields, and incorrectly entered "action" codes. In
addition, the following problems were frequently encountered: emissions
estimation methods reported as "3" (i.e., computer calculated), but fields
needed for calculation were missing; missing stack data; illegal SCC codes;
and duplicate SCCs at single points. A comprehensive list of specific
problems was forwarded to the Agency. However, the DEQE chose not to respond
to the error list directly, rather, the DEQE updated the Massachusetts NEDS
file, reportedly using the GCA error list as a guide. Updates were reportedly
based on source size and error severity. Although the revised data reflected
considerable improvements, numerous problems persisted. These included
missing stack data and duplicate SCCs at single points. The former problems
were addressed using a computer routine which inserted data reflecting ground
level plumes at ambient temperature (77°F) for the missing stack data. The
duplicate SCC problems had to be manually corrected by GCA, since they
frequently occurred when the emissions estimation method of "3" was used. In
these instances process rates were added and sulfur and ash contents were
averaged, weighed by relative process rates.
69
-------
In addition to the stack data problem, there were a significant number of
incorrect UTM coordinates that were identified during the QA review. While
many of these were addressed by the DEQE in the revised inventory, all UTM-X
coordinate errors may not have been identified since Massachusetts falls
within two UTM zones making virtually all X coordinates "in range." Based on
the number of "out of range" Y coordinates it would not be surprising if a
number of X coordinate errors exist within the Massachusetts data.
Upon review of the initial NECRMP report for Massachusetts, 6 the DEQE
noted discrepancies in VOC emissions for a number of facilities. The problem
was traced back to conversion of the Massachusetts data from EIS/P&R format to
NEDS format. Apparently, for a number of major facilities, DEQE had stored
reactive VOC data in a separate, "sixth" pollutant field in their EIS/P&R
file. These data were subsequently "lost" in the conversion to NEDS format.
A follow-up meeting was held at EPA/Region I offices to determine the most
efficient approach to restore the missing VOC data.
The DEQE subsequently supplied listings of VOC emissions data, which
included the reactive VOC data stored in the above mentioned separate
pollutant field, to EPA/Region I. Update transactions to enable the
appropriate corrections were coded and keypunched by EPA/Region I and
forwarded to GCA, who updated the EIS/PS master file, accordingly. GCA also
took this opportunity to correct the SCC codes of several lime manufacturing
point sources which had previously been miscoded as a result of a
typographical error in the 11/78 version of the NEDS SCC listing (on page C-59
of A
Area Source Inventory Quality
The Massachusetts DEQE supplied area source data, reflecting 1979, in
hard copy format. The area source inventory was reviewed for
comprehensiveness and consistency with the Volume 1^- area source
methodologies. The documentation supplied provided only the VOC and NOX
emissions in kg/day. Activity or process rates, emission factors, and annual
emissions in tons/year were not provided. Additionally, a few categories had
not been addressed, and a few others required further disaggregation to
maintain conformity with the NECRMP requirements.
GCA was also unable to duplicate DEQE's emission calculations for
gasoline handling, architectural surface coating, and autobody refinishing.
Additional information was requested. The DEQE estimate for small
industrial/commercial degreasing included only cold cleaning. GCA asked DEQE
to verify that all open-top vapor and conveyorized degreasing had been covered
In the point source inventory's 8665 tons VOC/year (covering 112 emission
points). The DEQE estimate for structural fires included VOC emissions only;
GCA requested NOX data as well. All of the above were subsequently resolved
except that the DEQE did not disaggregate emissions for three categories:
vessels (gasoline, distillate, residual), aircraft (civil, commercial,
military), and off-highway vehicles (gasoline, diesel). Composite categories
were used in these instances. As is the case for other states which used
composite categories, speciation and temporal resolution will be adversely
affected.
70
-------
Comparison with NEDS and SIP Inventories
The NECRMP emissions data for Massachusetts are compared to the NEDS and
SIP inventory estimates in Table 30.
With the exception of the Worcester SIP area, NECRMP point source VOC
totals are somewhat higher than the SIP estimates. Some of this difference is
related to the DEQE's handling of nonreactive VOC for point sources. As
discussed above, the DEQE had stored reactive VOC emissions in "sixth"
pollutant field in their EIS/P&R file which were not supplied in Massachusetts
original NECRMP submittal. Adding these emissions into the NECRMP data base
resulted in a statewide increase in VOC point source emissions of
approximately 51,000 tons/year. For the Springfield area the NEDS and NECRMP
estimates for point source VOC are in good agreement. The NEDS point source
VOC estimates are approximately 10 percent higher than NECRMP for the Boston
area and 10 percent lower for the Worcester area.
The Massachusetts SIP and NECRMP area source inventory totals deviate
significantly from NEDS. The bulk of the VOC difference can be attributed to
solvent evaporation. Reasons for deviations in estimates for this category
have previously been addressed.
For NOX, the NECRMP inventory shows over 80,000 tons/yr more than NEDS
(approximately 30 percent higher). Nearly half of this deviation is
attributable to highway vehicle emissions (32,000 tons/yr). The NECRMP annual
totals for highway vehicles were derived from the Massachusetts supplied
kg/day values using the temporal adjustment factors from Volume XVII.^
This will ensure that when temporally resolved to a typical summer weekday,
the NECRMP emissions will match the Massachusetts SIP values. Thus, a
deviation in annual totals may not necessarily be a major problem in this
instance. The remaining difference can be attributed to commercial/
institutional oil combustion. For these categories (distillate, residual),
GCA utilized the DEQE supplied fuel use estimates and emission factors to
calculate emissions. Although the distribution to sectors varies, the total
area source distillate oil use was found to be in good agreement (less than
one percent difference) with the 1978 NEDS Fuel Use Report,32 yet the NOX
emissions estimates from commercial/institutional distillate oil combustion
differ by an order of magnitude. Some of this difference appears to be
related to emission factors utilized. The DEQE used a factor of
60 lb/10^ gallons, which is somewhat higher than the emission factor used by
CCA in other NECRMP states (22 lb/103/gal).
Apparent Completeness for Other Pollutants
The Massachusetts NECRMP point source inventory includes emission data on
TSP, S02, and CO, which are compared to the NEDS system totals below.
71
-------
TABLE 30. COMPARISON OF NEDS, SIP, AND NECRMP EMISSIONS ESTIMATES
FOR MASSACHUSETTS
SIP area
Boston
Point
Area
Total
Worcester
Point
Area
Total
Springfield
Point
Area
Total
Statewide
Point
Area
VOC
NEDS
76,883
312,231
389,114
9,417
63,732
73,149
20,981
56,163
77,144
101,681
445,041
I (tons/year)
SIP NECRMP
41,039 68,901
177,311 145,867
218,350 214,768
11,266 10,311
33,277 31,631
44,543 41,942
17,703 20,987
33,647 34,583
51,350 55,570
113,328
235,862
NOj,
NEDS
73,910
134,240
208,150
2,050
24,622
26,672
10,238
25,063
35,301
89,621
183,930
; (tons/year)
SIP NECRMP
105,815 133,135
179,443 174,224
285,258 307,359
80 2,662
26,957 28,972
27,037 31,634
25,750 11,895
14,484 36,415
40,234 48,310
151,196
266,046
Total
546,722
349,190 273,551
417,242
72
-------
TSP (tons/yr) SOV (tons/year) CO (tons/year)
NKCRMP point 13,368 320,261 18,541
NEUS point 11.975 301,229 10,057
Deviation 1,393 19,032 8,484
A complete description of the Massachusetts emission inventory can be found in
Volume VI.6
NEW HAMPSHIRE
Point source data for New Hampshire were obtained from the NEDS system
following an update in a previous contractual effort by GCA.23 jhe area
source data were developed by GCA using the procedures recommended by EPA in
the Guidelines^ document. The New Hampshire Air Pollution Control Agency
(NHAPCA) assisted in this effort by supplying required corrections and
additional data as needed for the point source inventory, and providing much
of the input data used in the area source inventory development.
Point Source Inventory Quality
The New Hampshire point source inventory, reflecting 1979 data, was
obtained from the NEDS system. The data had been updated in a previous
assignment by GCA.23 ^he majority of the problems corrected as a result of
the QA review entailed missing UTM coordinates, inconsistent use of blanks and
zeros in the control equipment and efficiency fields, and missing exhaust flow
rates. These were corrected through the cooperation of the New Hampshire
APCA. An erroenous emission total for one facility was detected by EPA and
subsequently addressed. No serious remaining problems are known.
Area Source Inventory Quality
GCA developed a 1980 area source inventory of emissions of VOC and NOX
using state supplied activity data and methodologies generally consistent with
the Procedures for the Preparation of Emission Inventory for Volatile Organic
Compounds—Volume 1.19Due to data availability,some minor deviations from
the Volume I19 methods were necessary. These are explained in Volume VII.1
One such modification to the Guidelines'^ methodologies involves
residential fuel combustion. Fuel used in the residential sector was
estimated using the equations presented in Volume I.*9 xhe number of
housing units using each individual fuel type was determined using the
following equation:
Hi 80 = Hi 70
73
-------
where H. Qri a Number of housing units using fuel type i in 1980
1 oU
H. * Number of housing units using fuel type i in 1970 (from the
1 1970 Census of Housing)
H a Number of total housing units in 1980 (from the 1980 Census)
T oU
H a Number of total housing units in 1970 (from the 1970 Census)
This method, used in lieu of fuel-specific data for 1980, assumes an
insignificant change in fuel mix from 1970 to 1980. The emission factor for
LPG was derived assuming a 1:1 ratio of propane to butane. Emission factors
for the remaining fuels were obtained directly from AP-42.2^
Comparison with the NEDS Inventory
The NECRMP emissions data for New Hampshire are compared to the NEDS
inventory estimates in Table 31.
TABLE 31. COMPARISON OF NEDS AND NECRMP EMISSIONS
FOR NEW HAMPSHIRE
VOC (tons/year)
Point
Area
Total
NEDS
11,192
68,683
79,875
NECRMP
21,041
59,187
80,228
NO* (tons/year)
NEDS
41,697
30,177
71,874
NECRMP
49,206
45,306
94,512
As shown in Table 31, NECRMP point source VOC emissions are nearly twice
those contained in NEDS. This is primarily due to recent updates in NEDS
(current as of January 1983). (The July 1982 version of NEDS data for
New Hampshire contained 18,461 tons/year from VOC point sources).
The NECRMP area source estimates of both VOC and NOX differ from the
NEDS system. As was the case in other states, the NEDS system shows higher
VOC emissions from solvent evaporation. The reasons for this deviation have
been previously discussed. For NOX, the NECRMP inventory shows over 30
percent more emissions from area sources than does NEDS. Most of this
deviation can be attributed to higher NOX emissions estimates for light-duty
highway vehicles and residential distillate oil combustion. The NECRMP
estimate of emissions from highway vehicles was calculated based on VMT data
from the NHDOT and MOBILE2. The NECRMP residential fuel combustion emissions
estimate was derived using the volume 1*9 methodology.
74
-------
Apparent Completeness for Other Pollutants
The NECKMP point source inventory includes data on TSP, SC>2 and CO
which are compared to the NEDS inventory, below.
NECRMP point
NEDS point
TSP (tons/year)
4,829
6,489
SO^ (tons/year)
68,015
102,088
CO (tons/year)
5,875
11,143
(1,660) (34,073) (5,268)
A complete discussion of the New Hampshire inventory is presented in
Volume VII.7
NEW JERSEY
The New Jersey Department of Environmental Protection (DEP) was the
original source of the point and area source emissions data used in NECRMP.
For the New Jersey counties included in the Philadelphia SIP area,
Engineering-Science, Inc. conducted an extensive update and quality assurance
effort prior to incorporation of the data into the NECRMP inventory. The
Now Jersey DEP also assisted in the quality assurance review of the NECRMP
data by supplying updates and corrections as required and additional data as
needed.
Point Source Inventory Quality
The NECRMP point source inventory data, reflecting 1980, were obtained in
EIS/P&R format from Engineering-Science, Inc., who had previously converted
the inventory data from NJDEP's Air Pollution Emission Inventory Data System
(APEDS). There were numerous significant problems with the data, which were
traced back to conversion of the APEDS data into EIS/P&R format. These
included dropped digits in stack diameter fields, occasional incorrect
conversion of emissions from Ib/hr to tons/year, invalid and/or missing
emission control equipment and efficiencies, blank UTM zones, "garbage" data
in various fields and asterisks in several process rate fields.
Due to resource limitations, the above problems were generically
addressed using a computer routine, which is further discussed in
Volume VIII.8
After completion of the above computerized corrections, a number of
discrepancies were noted between the NECRMP data and the New Jersey SIP
inventory. The EIS file for the involved facilities was compared to the
New Jersey APEDS file, the NJDEP field office reports, and the operating
permit files. Inconsistencies generally fell into two categories:
75
-------
• The APEDS to EIS conversion resulted in uncontrolled emissions being
reported for certain points, thus creating a discrepancy.
• The APEDS file matched the EIS file, but emissions reported in the
SIP reflected the more recent field office reports which superseded
the APEDS data.
The APEDS file reflects emissions by pollutant species rather than total
VOC, SOX, NOX, etc. It was noted when reviewing the APEDS data that in
some instances water vapor (H20) had been erroneously included as
hydrocarbon emissions in the APEDS to EIS conversion. Since at that point the
objective was primarily to reach agreement between the SIP and NECSMP
inventories, no attempt was made to correct these problems.
In many instances, the SIP facility emission totals resulted from the
NJDEP field offices' reports. For some field offices, only plant totals were
provided prohibiting matching of individual emission points. In these
instances, "new" emission totals were allocated to individual points based on
the distribution of the "old" emission totals. Since pollutant speciation in
preparation for photochemical modeling will be tied to point-specific SCO
codes, considerable error in speciation is inevitable.
Area Source Inventory Quality
The 1980 New Jersey area source emission inventory was supplied in hard
copy format by the New Jersey DEP. The EPA Project Officer also made
available an area source inventory covering the New Jersey portion of the
Philadelphia AQCR that was developed by Engineering-Science, Inc.
The above area source inventories were reviewed for comprehensiveness and
consistency with the area source methodologies presented in Volume I.I In
general, the NJDEP area source inventory was very well prepared and was
determined to be preferable to the Engineering-Science inventory. The results
of this review and comparison were documented and forwarded to the EPA Project
Officer. The primary problem with the Engineering-Science data reflected the
fact that the inventory had been developed prior to the release of the
Procedures for Preparation of Emission Inventories for Volatile Organic
Compounds, Volume I, Second Edition*9 and, therefore, reflected older
methodologies in many instances. The EPA Project Officer indicated
concurrence with GCA's recommendation to utilize the NJDEP inventory for all
of New Jersey in the NECSMP inventory. Some minor modifications were made to
the NJDEP data prior to computerization, primarily to ensure consistency with
the other NECRMP states' data. These involved those area source categories
whose emissions are based on population: degreasing, dry cleaning,
architectural surface coating, autobody refinishing, graphic arts, and
commercial/consumer solvent use. Since the DEP had completed calculation of
the area source inventory prior to release of the 1980 Census, they had relied
on population projections for estimating emissions using per-capita factors.
To maximize consistency with the other NECBMP states, GCA recalculated
emissions for the above categories using the 1980 Census of Population.
76
-------
Comparison with NEDS and SIP Inventories
GCA prepared its first draft of the New Jersey inventory reportS j.n
May, 1982. In August of 1982, the NJDEP submitted a revised ozone SIP
inventory. Differences between the August 1982 version and the December 1981
SIP inventory utilized by GCA are highlighted below.
• Eight major point sources of VOC were added to the August 1982 SIP
inventory and one was deleted, totaling an increase of 3,612 TPY of
VOC. One of the added sources was already contained in NECRMP, the
other seven are not; the deleted source is in the NECRMP inventory.
• Non-highway vehicle area source emissions were unchanged and are
generally in good agreement with NECRMP. Some minor differences
exist because GCA used 1980 Census data for per-capita area source
emissions calculation, while the NJDEP used earlier population
projections.
• Highway vehicle emissions have been modified to reflect an error in
MOBILE2, ostensibly involving I/M credits. This results in VOC
emissions adjusted upward, across-the-board, by about 15.3 percent.
Total summer weekday VOC emissions, which were 454 Mg/day in the
December 1981 SIP, are now 517 Mg/day in the August 1982 revision.
Extrapolating to annual rates, the NECRMP VOC totals for highway
vehicles should be increased from about 182,458 TPY to about
208,000 TPY, an increase of about 25,000 TPY.
• A source-by-source comparison was made of facility totals between
the New Jersey SIP and NECRMP inventories. These are summarized
below:
77 facilities emitting >_ 100 TPY of VOC are included in the
NECRMP inventory but not in the SIP inventory, totaling
20,307 TPY. The NJDEP has acknowledged that some of these
sources do emit 100 TPY or more.25
37 facilities emitting >_ 100 TPY of VOC are in the New Jersey
SIP inventory but not in the NECRMP inventory, totaling
13,199 TPY.
33 sources are present in both the NECRMP and SIP inventories
but differ in VOC emissions by >_ 100 TPY of VOC. The SIP
sources falling in this category total 7105 TPY more than the
NECRMP sources.
35 sources emitting >_ 100 TPY of NOX are in the NECRMP
inventory, but not in the SIP inventory, totaling 19,322 TPY.
The NJDEP has acknowledged that some of these sources emit
100 TPY.25
77
-------
19 sources emitting >_ 100 TPY of NOX are in the New Jersey
SIP inventory but not in the NECRMP inventory totaling
6,412 TPY.
17 sources present in both the NECRMP and SIP inventories, but
differ in NOX emissions by >_ 100 TPY. The SIP sources
account for 13,249 TPY of NOX more than the same sources in
the NECRMP inventory.
A source by source review of the above differences can be found in
Appendix D of Volume VIII.8
The New Jersey NECRMP emissions estimates are compared to the NEDS system
totals in Table 32.
TABLE 32. COMPARISON OF THE NEDS AND NECRMP INVENTORIES
FOR NEW JERSEY
VOC (tons/year)
Point
Area
Total
NEDS
418,896
562,841
981,737
NECRMP
158,228
415,035
573,263
NOx (tons/year)
NEDS
157,803
272,738
430,541
NECRMP
204,844
348,568
553,412
As shown in Table 32, the NEDS system estimates for VOC emissions from
both point and area sources are considerably higher than NECRMP. Changes in
the NJDEP SIP inventory are not believed to be reflected in NEDS because,
despite the above differences between the NEDS and SIP inventories, it was
found that the cumulative errors, for both VOC and NOX, almost exactly (and
coincidentally) cancel each other.
The differences between the NEDS and NECRMP estimates for point source
VOC can primarily be attributed to chemical manufactuing, organic solvent
evaporation, and petroleum storage transport. For NOX, NECRMP point sources
in NECRMP total 47,000 TPY higher than NEDS. Virtually, all of this
difference can be attributed to electric generation.
Area source VOC totals in NEDS are nearly 150,000 TPY higher than
NECRMP. Much of the difference can be attributed to organic solvent
evaporation. Reasons for this deviation have previously been discussed. The
NECRM? NOX emissions estimate from area sources is approximately 21 percent
higher than NEDS. Most of this difference can be attributed to highway
78
-------
vehicles - LDVs. The NECRMP highway vehicles emissions, and most likely the
NEDS highway vehicle emissions as well, do not reflect the August 1982 update
made by the NJDEP.
Because of the large number of differences between NEDS, NECRMP, and the
revised ozone SIP inventories for New Jersey, it is recommended that, if not
already accomplished, the latest set of SIP estimates be incorporated into the
New Jersey APEDS file. Since the errors which hindered conversion from APEDS
to EIS format have reportedly been corrected, the revised APEDS file should be
converted into EIS format for replacement of the current NECRMP file for
New Jersey point sources. Similarly, highway vehicle emissions for New Jersey
(as well as other affected areas) should be updated to reflect corrections to
MOBILE2.
Apparent Completeness for Other Pollutants
The New Jersey point source inventory also includes data on TSP, SC>2,
and CO emissions, which are compared to the NEDS system totals below:
TSP (ton/year) S02 (ton/year) CO (ton/year)
NECRMP points 160,027 588,276 113,773
NEDS points 351,017 273,589 94,925
Deviation (190,990) 314,687 18,848
A complete discussion of the New Jersey emission inventory can be found
in Volume VIII.8
NEW YORK
The New York Department of Environmental Conservation (DEC) provided the
input data used in the NECRMP point and area source inventories for New York.
The DEC assisted in the quality assurance review of the data by supplying a
revised point source tape reflecting corrections to many errors identified by
GCA, and supplying additional data as needed, most notably a cross-reference
file which enabled GCA to "match up" New York's unique process codes with the
most closely corresponding EPA SCC codes.
Point Source Inventory Quality
The New York point source emission inventory was supplied by the New York
Department of Environmental Conservation (DEC). The data were supplied in
NEDS format and represented 1980 data. The first NEDS formatted tape sent by
New York had considerable problems apparently relating to the conversion of
the point source data from New York's data system into NEDS format. Most
notably, no NEDS Card 6 records were provided for nonboiler sources, thus, no
process rates, SCC codes, design rates, or process descriptions were
included. Other problems related to the emissions control device and
efficiency fields. Further, a large number of sources had reported stack
temperatures of 80°F.
79
-------
The moat significant problem entailed SCC codes. New York utilizes their
own unique process codes which are not directly compatible with EPA's SCC
codes. The DEC did send a cross reference list of New York process codes to
SCCa, matching the "closest" SCC to the appropriate New York process code.
This, combined with a revised NEDS file which contained the previously missing
Card 6's which now contained the New York process codes, enabled GCA to
"match" approximately half of the SCC codes for processes. Process codes that
could not be "matched" with an appropriate SCC were assigned an "unidentified"
code of 3-99-999-99. This SCC accounts for 38 percent of the VOC emission but
only 0.2 percent of the NOX emissions from point sources in New York.
Since the VOC species allocation factors are SCC-specific, 38 percent of
the New York point source VOC will not be able to be properly speciated,
seriously affecting the data's suitability for photochemical modeling.
Area Source Inventory Quality
The New York DEC supplied separate inventories for the Metropolitan New
York City area and for "upstate" New York. GCA reviewed the DEC-supplied data
and identified numerous data gaps and a few emission factor inconsistencies
which were brought to the attention of the DEC. A number of composite area
source categories needed to be disaggregated to conform to the list of area
source categories being used in NECRMP. Emission factors and activity rates
were not provided for the majority of categories. GCA also identified
discrepancies with some of the emission factors that were provided by the
NYDEC.
Some of the missing data for the "downstate" inventory was contained in
the DEC's ozone SIP inventory which was later supplied by EPA/Region II.
NYDEC supplied VOC composite category disaggregation data, through EPA, and
the New York Department of Transportation supplied VMT data enabling
disaggregation of highway vehicle emissions data into the appropriate vehicle
classes.
GCA eventually chose not to utilize the upstate NOX emissions summary
for the following reasons. First, the emissions were aggregated into
composite categories. Although the NYDEC later disaggregated the VOC
emissions summaries into the appropriate NECKMP categories, no information was
provided to enable proper disaggregation of the NOX emissions.
Additionally, the inventory was reportedly based on Part 208 Water Quality
Management Program population projections, which were determined to vary
significantly from the 1980 Census data which was used in all other NECRMP
states.
In some instances, GCA had to develop emission factors, activity rates,
and/or emissions for various categories to supplement the New York DEC
supplied data. This was necessary to meet the requirements of the EIS/AS
system and, in some cases, to be more consistent with Volume 1^
methodologies. The specific categories to which this applies are discussed in
detail in Volume IX.9
80
-------
Comparison with the NEDS and SIP Inventories
The NECRMP inventory for New York is compared to the SIP and NEDS
emission totals in Table 33.
As shown in Table 33, the VOC totals for the New York State portion of
the New York City (NYC) Metropolitan Area agree within 5 percent. Closer
examination, however, shows significant differences in individual point source
totals, which are outlined in detail in Appendix E of Volume IX.9 More than
99 percent of the difference in NOX emissions in the Metro NYC area can be
attributed to two power plants, whose NECRMP NOX emission totals are shown
below, that are not present in the New York SIP:
• E. F. Barrett (105,481 TPY)
• Bowline Point (325,045 TPY)
Despite the known VOC emission total differences for specific point source in
New York and New Jersey, VOC emission totals for the entire New York City SIP
area agree within less than 6 percent. Differences in NOX emissions are
roughly equal to that reported for the two NYC area power plants identified
above.
Statewide totals of point source VOC emissions are considerably higher in
NECRMP than in NEDS. The majority of NECRMP VOC emissions fall into two
categories, in-process fuel use and not-classified industrial point sources,
which are not readily comparable to the NEDS reporting categories. Area
source VOC emissions reported in NEDS are considerably higher than those shown
for NECRMP area sources. Most of this difference can be attributed to organic
solvent evaporation. Reasons for this difference have previously been
discussed.
Statewide NOX emissions in NECRMP are approximately 1.3 million
tons/year higher than shown in NEDS. About 1.1 million ton/year of this
difference relates to electric generation sources, which account for most of
the NOX discrepancy between NECRMP and NEDS in the entire 15-state NECRMP
study area. As discussed in Section 3, NEDS emissions for power plants are
independently updated by NADB annually using DOE published activity data.
Yet, much of the difference between NECRMP and the SIP relates to power plants
which are not included in the SIP inventory, and may be omitted in NEDS as
well. Thus, the NECRMP NOX emissions from power plants in New York should
be reinvestigated.
Apparent Completeness for Other Pollutants
The New York Point Source inventory includes point source emissions data
for TSP, S02, and CO, which are compared to the latest NEDS totals below:
81
-------
2
0
^c
£ij
z
06
0
tL,
co
M
oi
g
Z
w
1— 1
1
a
w
z
Q
^J
A
a*
h-^
to
M
CO
a
Ed
z
NJ
S
H
Cti
0
z
o
CO
1-4
Cri
5"
a,
S
8
r^
f^
Ed
J
^
g^
U
Z
,-v
(3
01
>^
*«^,
c a>
O M
U CO
>^x
O
co
a
f*3
5
£
2
o
[d
Z
U
nj
0>
>> CU
"v. M
C CO
0
4j
^^
U
o
CO
a
pj
Z
(Q
rjj
^
flj
CM
i— i
CO
O
^•4
Z
0
5
25
o
a
o
^
^4
o
a.
^
M
0
3
D
CN r-4
CN r->
*"1 ™
in m
in c in
\o co
P- OO
m in
n M
so cy>
en r**
CN ^
a\
CO
O
Vl
U
0)
•g
O
SM 4J
Z C eo
• f^ D
0) O ^
U CU <
• l-l
U
fi
a
00
o*\
in
!_!"
o\
r-l
r-4
m
— ' 4 sO
1 1
*s.T r^
sO O
CN i-l
M M
*^ CJN
en oo
o*
9) Li
•o e ea
•H -H 0)
S 0 ^
4J
eQ
4J
M
CN
m
00
en"
O
CN
1
O
en
-^
en
r^
en
m
CN
m
^^
00
1
m
sO
en
M
m
CN
O
M
1— 1
»-H
CQ
4J
O
H
^J
3
U
• H
O
01
c
0
,
^
U
c
3
O
o
•o
1— I
o>
•-*
U-l
^
*^
OJ
•o
C
OJ
>^
01
CO
M
0)
•-J
3
0)
Z
a
^4
4)
,y
U
O
c
CO
4}
-a
3
— l
U
c
M
O
82
-------
TSP (ton/year) SO? (ton/year) CO (ton/year)
NECRMP points 272,343 1,769,409 203,972
NEDS points 225,320 989,761 92.996
Deviation 47,023 779,648 110,976
A complete discussion of the New York emission inventory is presented in
Volume IX.9
OHIO
The Ohio Environmental Protection Agency (OEPA) worked with GCA* in
updating their point source inventory data for use in NECRMP. Since Ohio does
not maintain an area source inventory of the detail required for NECRMP, GCA
developed the Ohio area source inventory for NECRMP using methods generally as
prescribed by the EPA Guidelines^ document. The Ohio EPA assisted in this
process by supplying much of the input data used by GCA in developing the area
source inventory.
Point Source Inventory
The Ohio point source inventory was obtained from the Ohio EPA, following
the GCA update work, in NEDS format as converted from the Ohio EIS system, and
represents 1980 data. The most prevalent problems identified in the quality
assurance review entailed missing SCC codes, SIC codes, and stack parameters.
Alao a number of sources contained multiple discharge points^" for which stack
parameters needed for modeling were unavailable. The majority of the missing
data was supplied by the Ohio EPA in response to GCA's request for those
data. For multiple discharge points, the Ohio EIS to NEDS conversion routine
substitutes 77°F, 49 foot plumes for the multiple "stack" data. Since this
substitution would have adversely affected modeling, the following method was
used to determine the most appropriate parameters to include in the source
record. Where several discharge points had different stack parameters but met
the criteria of AEROS Volume II for combining into a single point, the
following method (suggested in AEROS Volume II) was used to determine the most
appropriate parameters to code. A "K" value was calculated for each discharge
point as:
K = HVT/Q
*Contract 68-02-3510, Work Assignment 10.
"f"The Ohio EIS system maintains parameters for each discharge point individually
even though they are contained in a single emission "point". NEDS and EIS/PS,
however, can contain only one set of stack parameters for each "point".
83
-------
where H » individual stack height
V • individual gas flow rate
T « individual stack temperature
Q - individual emission rate for VOC if present, NOX otherwise.
Stack parameters were coded for the discharge point with the lowest "K" value.
Some residual problems remain with the Ohio point source data. The most
important of which relates to 96 source records which have missing data
(primarily SCO codes) owing to confidentiality. Missing SCCs were coded as
"other chemical manufacture" or "other industrial processes." Although VOC
speciation for modeling is based on SCO codes and omission of these SCCs will
adversely affect the photochemical modeling, these categories account for only
7.5 percent of the total VOC emissions from Ohio point sources. Therefore the
impact is not expected to be severe.
Area Source Inventory Quality
Since Ohio did not maintain a VOC/NOX area source inventory of the
detail required for NECRMP, the 1980 area source inventory for the Ohio
portion of the NECRMP study area was developed by GCA using methods generally
consistent with the Procedures for the Preparation of Emission Inventories for
Volatile Organic Compounds—Volume I, Second Edition,19 and includes
estimates for all 54 NECRMP area source categories. The Northeast Ohio
Areawide Coordinating Agency (NOACA) was concurrently preparing a VOC/NOX
area source inventory for use in the 1982 ozone SIP for the Cleveland area.
With a few exceptions, the emissions estimates used in NECRMP for the Ohio
Counties of Cuyahoga, Lake, Lorain, and Medina reflect the NOACA estimates.
A comparison of the NOACA versus NECRMP estimates can be found below. A more
comprehensive discussion of the differing methods used is presented in
Volume X.10
Comparison with SIP and NEDS Inventories
The NECRMP point and area source emission totals are compared to the SIP
and NEDS totals for the Ohio portion of the NECRMP study area in Table 34.
The Northeast Ohio Areawide Coordinating Agency (NOACA) prepared a
VOC/NOX area source inventory for the Cleveland area, consisting of
Cuyahoga, Lake, Lorain, and Medina Counties, for the 1982 State Implementation
Plan (SI?) submittal. GCA also calculated emissions for the four NOACA
counties in order to compare the results of the different methodologies and
data sources. The results of this comparison appear in Table 35 which shows
the emissions calculated by both GCA and NOACA for each area source category.
A ratio of the sum of all four counties' emissions as calculated by GCA
divided by the sum of emissions as calculated by NOACA is included as an
indicator of comparison for each category. If this ratio is nearly equal to
1.0, the calculation differences are small. A sum over all counties was used
for this ratio so that differences in distributing emissions to the different
counties were smoothed out and the ratio indicates primarily differences in
methodology and data sources. The entire table shows calculations before any
84
-------
o
1 j
3B
O
H
£_4
py*
O
CO
w
M
£"*
C^
z
j>
Z
M
s
2
i
z <
s
*« H
CO Q
p
CO CO
Q
w rs
z 5
M CJ
2 Ex3
H Z
Ct. W
o 5
H
S5
O Cti
CO O
Q &
82
^
CO
W
M
H
M
Cfl
<0
«^
a
O
4J
i
~
eg
at
-^^
r*
o
^
CJ
o
"*
p .
3
CJ
z
0!
M
co
CO
Q
Z
p.
£
o
z
a-
M
CO
co
Q
§
CO
co"
f*t»
CM
,_(
(TJ
U
fi
vC 00
m o
O CM
r-l CO
CO -*
1 1
m CM
r-l CM
^O \Q
ft M
r^ O
O ON
CO CM
CM 00
m oo
-* m
m o>
oo r-.
CO
1 1
O r-l
^O ^^
>* -*
co" vf
\O r*^
r-l
-------
00
z
O
1-4
•£
J
a
CJ
55
0
00
00
_*
a
sa
as
3 00
Q U
OO 1-4
H
OS 0
•< U
u u3
< as
0 <
z
Q5?
*•*
3 ^
< i-J
w
u a
O J
tn
0 OS
_j
a p
0 E
oo
1-1 Ul
os 3
•2 H
z as
o p
CJ &
.
n
>J
3
H
3'v u t
w ^ w
oo go
-
^
" 9
a 2
•o
1
d
-j
d
<
a i
•-
u
3 5
d
v O
^ 2
*7
d
*c
3
, I
90
^
X
-oN-ro- a^«>n 000000^-^=00 o <, o o o a» o - <-, ^,
(N CN CN otnt/^ en caaoom oaaaa i/\A^
^-OMOJCM^I^— I'j-itN^-. ^vr>r-.r~. ii .» ui IIP.
n" « ,r" - m" - W >n^) -
y y -4
* °
^ u *- o
e u 3 u .-, —
•-4 U >** « V fcj .W <«4
•O2fl--i-afl ^.o •-*
«C!(n3oOi«>3U -^ g) _» O
O "* C enscqaiu.^ oitaOVO) -^ »
— * -^ M m -^ .^ g -- -^ us jj M vnov<«
eveu — ^: i. o a a. •-* 3 — « us ^oo
j*v4>j*aOQ3C3U'-«a >* u g -a >*« u <« .*• 3 — • 9
*-— '-aw
-------
-a
OOOOOOfNf
CN (-1
fN n
— * O O O i** (N
—II ^ \£
1 1 -Ct
OOOOOO-* o
cr>
~* in i i ° ° ^- ^*
30
m co
CO
8c?^d<8o><^ox
01 M
u 3
-" 0
u e
2 'a
^ 3
U U
C -I
n a)
u o C t
3 3
-- -^ 0 C
3>-'-0>OCT'OOOO
— '--wO-3-OOiAJOr*eN— 4O--<
— t 1
-^JJJ(sl_0i00
o I
flfJi^^QOj'i-^OOO
CT» (N -J in (N
^o m .— * vo <• -o
— • fN m
^c?^^^^8c?g
H
— • Q
O U nj
U 00
3 — i ta
-O -H b
-• u 3 13
«) to u O O
41 -^ « Oi O
^ T3 C -3 9
o) cq aj eg «
M tn CA n w
33333
-d -O 13 73 -O
C C C C C
so r- 30 3> O
OOflCNf^-NltN^rfnor-.flUjv.OOOjOg
OOOOOOOO--—OOOOO-I
OO^rNOOrOinOOOOOO^r-tO^
1 ft (N •-* -^
OOO^-OOO^r-^OOOO^^J-oo
•— < •& cn
I fNOO'-H-^ st
1 ° ^•LAO'OtNeN^S^^^^r^ONOOfn
,M" ^" _r ^ _r v
* * Sr.S-« "- °.-3J
«?S«J'SJl8z>z>z>z>z>z:>z>z>z
O — i
•H
*J V O
u a) 4 <•* 41
M-t ^ ^-* cq u)
(0 U — « 3 Vt «
WWW -^ Tj t$ •-<
tl (fl CO W « OS
•^ U aj •-< a> >, >>
flj O i-^ 00 T) U fll CQ
U^H V Vj gj fllu^JZ
• ^ rt S —< w » g--48j«a»'wi4-i
3S O O 35 > > > OO
-------
£
0)
s
•H
0
u
s
w
H
d"» 3 "•»
4J 1 •« u
•JO 00
u 1 2 fcJ
* g
e *
Ml
d
d
a *
a
w
3 5
d
.* 2
J
d
d
a s
1
3
'
e
1
t
0
>O ^ •O «N
00 S* r<. >* 1 1 1 1 1 I
• • 1 • t 1 1 1 1
fl ("•• f*4 '"*
-^ -u r*. O 1 I 1 t 1 1
-T -N 1 i 1 1 1 1
f*» ^00?N'--33'OOOOr*dOC5 "^ O
^ ^ -»•
88
-------
balancing with point sources was done. Area source categories for which large
differences occur were reexamined. In some instances, the NOACA estimates
reflected survey or other local data believed to be more accurate than
national per-capita factors reflected in the EPA Guideline^ methods. In
these instances, the NOACA numbers were used in NECRMP. In other instances,
GCA concluded that the NOACA inventory had omitted or improperly estimated
emission from certain other area source categories. In these instances, GCA
estimates were used. A complete discussion of which estimates were used for
which category is presented in Volume X.10
A number of categories had not been completed by NOACA in time for
inclusion in the Ohio NECRMP inventory.10 These categories are:
• Small Industrial Surface Coating (09)
• On-Highway Light Duty Vehicles (14)
• On-Highway Light Duty Trucks—Class I (15)
• On-Highway Light Duty Trucks—Class II (16)
• On-Highway Heavy Duty Gas Trucks (17)
• On-Highway Heavy Duty Diesel Trucks (18)
• On-Highway Motorcycles (19)
• Structural Fires (52)
• Forest Fires (54)
GCA estimates, therefore, were used for the above categories. Other
categories for which NOACA and GCA estimates differ significantly, and
resolution of these differences are summarized below.
Tank Truck Load ing/Trans it—
The NOACA inventory did not include tank truck loading emissions. To
maintain consistency with the other NECRMP states, the emissions calculated by
GCA were used for this category.
Degrees ing—
The NOACA estimate for degreasing was based on a comprehensive survey in
the Cleveland area. Since the GCA method reflected a national average and the
NOACA survey appeared to be comprehensive and reasonable, the NOACA estimates
were used for this category.
Pesticide Application—
GCA estimated emissions from pesticide application using the per-
harvested acre emission factor prescribed in the EPA Guidelines,19 an(j
harvested acreage data from the Census Bureau's County and City Data Book.
NOACA used a similar procedure but utilized different acreage estimates
derived from the respective county conservation services. Since the acreage
data used by NOACA were more current than those reported by the Census Bureau,
the NOACA estimates were used for this category.
89
-------
Cutback Asphalt—
GCA utilized state totals of asphalt application, apportioned to counties
on the basis of population, to address this category. NOACA performed a
survey to determine asphalt application in the Cleveland area. Since the
NOACA survey was believed to be reasonably comprehensive, the NOACA estimates
were used for this category.
Stationary Source Fuel Combustion
The methods used by NOACA for these categories were similar to those used
for the NECRMP inventory, except for the data sources used. Table 36 compares
the NECRMP and NOACA data sources. A large discrepancy occurred in the
emissions computed by GCA and those used by NOACA for Industrial Bituminous
(GCA's calculations were greater by more than an order of magnitude). NOACA
used both the industrial point and area source bituminous use reported in NEDS
Fuel Use Report. Thus, electric generation bituminous use was excluded from
the NOACA calculations. However, when balancing this category with the Ohio
point source inventory, GCA subtracted all point source use, including
electric generation fuel use. Therefore, GCA's calculations were used for all
32 counties for Industrial Bituminous emissions.
TABLE 36. DATA SOURCES FOR COMMERCIAL/INSTITUTIONAL AND INDUSTRIAL
FUEL CATEGORIES USED IN THE NECRMP AND NOACA ANNUAL
EMISSION INVENTORIES
Fuel
NECRMP data source
NOACA data source
Anthracite coal EPA'a NEDS Fuel Use Report
Bituminous coal DOE's Bituminous and Sub-
bituminous Coal and Lignite
Distribution
Residual oil
API's Basic Petroleum Data Book
Distillate oil API's Basic petroleum Data Book
Natural gas AGA's Gas Facts
LPG
Wood
DOE's Sales of LPG and Ethane
EPA's NEDS Fuel Use Report
EPA'a NEDS Fuel Use Report
EPA's NEDS Fuel Use Report
Petroleum Encyclopedia
ODOE Energy Review
ODOE Reports
DOE's Sales of LPG and
Ethane
Not included in the NOACA
inventory
Further comparison of the NOACA and NECSMP estimates, on a category-by-
category basis can be found in Volume X.^0
90
-------
On a regionwide basis, the NEDS system totals for VOC are somewhat higher
than NECRMP. Moat of this difference is related to organic solvent
evaporation. The reasons for this discrepancy have previously been discussed.
For NOX, the NECRMP estimates are considerably higher than NEDS. Most of
this difference can be attributed to area sources of NOX from bituminous
coal combustion. The NECRMP estimate for this category reflects coal usage
reported by DOE when balanced with point source totals in this category.
Apparent Completeness for Other Pollutants
The NECRMP point source inventory for Ohio includes data on TSP, S02,
and CO, which we compared to the NEDS system totals for the Ohio portion of
the NECRMP study area below:
TSP (ton/year) SO? (ton/year) CO (ton/year)
NECRMP points 206,391 1,459,530 211,165
NEDS points 231.664 1,545,934 425.854
Deviation (25,273) (86,404) (214,689)
A complete discussion of the NECRMP inventory for Ohio is presented in
Volume X.10
PENNSYLVANIA
The Pennsylvania Department of Environmental Resources (DER) provided
most of the input data used in the NECRMP point and area source inventories
for Pennsylvania. The DER further assisted in the quality assurance review of
the inventory by supplying some additional area source data where required.
Unfortunately, the DER did not respond to GCA's questions concerning the point
source inventory. Thus, except for corrections to generic problems that were
needed to enable conversion into EIS/PS format, the point source data for most
of Pennsylvania reflect the data essentially as received from the DER.
Point source data, and subsequent corrections, for the Philadelphia area
(except Philadelphia County) were obtained from Engineering-Science, Inc.
(ESl), who had conducted an extensive QA review of those data. Data for
Philadelphia County were obtained from the Philadelphia Air Management
Services (PAMS), through ESI. Point source data for Allegheny County
(Pittsburgh) were developed by GCA under a separate assignment26 to fulfill
the requirements of both the 1982 ozone SIP for the Pittsburgh area, and the
NECRMP annual inventory.
Point Source Inventory Quality
The original data source and year of record for the various components of
the Pennsylvania point source emission data are summarized below:
91
-------
Area
Allegheny County (Pittsburgh)
Philadelphia County
Remaining Philadelphia area3
Remainder of Pennsylvania
Source
GCA26
PAMS
ESI
PADER
Year of data
1980
1979
1979/80
1978
aBucks, Chester, Delaware, and Montgomery Counties
The above data were merged into a single file and subjected to the
quality assurance procedures reported in Volume I.I The quality assurance
review of the Pennsylvania data revealed a number of duplicate SCCs at single
points, a considerable number of points with questionable stack data, and a
number of sources with suspiciously high SC^ and CO emission levels. A
detailed list of questions was forwarded to the Pennsylvania DER in January,
1982. No responses were received from the DER. Engineering-Science supplied
data on changes affecting four facilities in the Philadelphia area, which were
incorporated into the NECRMP inventory as appropriate. For the remaining
portion of the State, however, the NECRMP inventory reflects the point source
data, as received, from the PADER.
Aa discussed above, to enable conversion into EIS/PS format, some
"generic" problems had to be addressed using a computer routine which is
discussed in detail in Volume XI. ^
Area Source Inventory Quality
The 1980 Pennsylvania area source inventory was obtained in hard copy
format from the DER. GCA identified several area source categories which were
noc included in the Pennsylvania inventory. A number of composite categories
which had to be further disaggregated were identified. Subsequently, GCA
received the PADER responses to the GCA review, via EPA/Region III. Not all
of the issues identified by GCA were resolved, although the PADER did supply
responses to the questions concerning the aircraft and gasoline marketing
categories. GCA was able to disaggregate the remaining composite categories
based on information from neighboring states' area source emission
inventories, and utilized NEDS 1980 emissions data for the omitted
categories. In some instances, GCA had to back calculate activity rates from
reported emissions, using AP-42 emission factors, to enable coding in EIS/AS
format. Specific adjustments performed by GCA in this regard are discussed in
detail in Volume XI. *•!• More significant categories are summarized below.
Gasoline Marketing—
The PADER provided activity rates, emissions estimates and emission
factors for the Stage I and Stage II gasoline evaporation categories. The
PADER estimate for Stage I evaporation included storage tank breathing. GCA
disaggregated the emissions for Stage I and storage tank breathing into
92
-------
separate categories. An AP-42 emission factor of 0.12 Ib/lO^ gal was used
to cover evaporative losses during transport by tanker trucks. GCA reviewed
the Pennsylvania point source inventory and verified that emissions for tank
truck loading were completely covered.
Degreasing, Dry Cleaning, Architectural Surface Coating, Graphic Arts,
Commercial Consumer Solvent Use—
GCA calculated VOC emissions (ton/year) using the 1980 Census of
Population and the per capita emission factors provided by the PADER, which
agreed with those recommended in the EPA Guidelines.19
Highway Vehicles—
Combined emission totals (for both VOC and NOX) for the six on-highway
vehicle categories were available for each county in kg/day for a typical
summer weekday. GCA multiplied daily emissions by 365 to convert the
emissions to tons/year consistent with Volume XVII.1? The emission totals
were distributed to specific vehicle classes based on the ratios derived from
the Ohio Area Source inventory.10 Emission factors from the Ohio
InventorylO were used to back-calculate VMT.
Fuel Use—
The PADER did not supply data on area source fuel combustion. Therefore,
1980 emissions from the NEDS system were used. The statewide emissions
estimates were distributed to counties based on the 1980 Census of
Population. Activity rates were back calculated using the appropriate AP-42
emission factors.
Comparison with the NEDS and SIP Inventories
The NEDS, SIP and NECRMP emission totals for Pennsylvania are compared in
Table 37.
As shown in Table 37, the NEDS, SIP, and NECRMP estimates for NOX are
generally in good agreement. Similarly, the SIP and NECRMP VOC totals are in
reasonable agreement, while VOC totals in NEDS are consistently higher. As is
the case throughout the NECRMP region, the NEDS system estimates are higher
primarily as a result of organic solvent evaporation estimates. The reasons
for these differences have previously been discussed in Section 3. A
comprehensive comparison of the SIP and NECRMP inventories can be found in
Appendix D of Volume XI.H
Apparent Completeness for Other Pollutants
The NECRMP point source inventory for Pennsylvania includes data on TSP,
S02, and CO emission, which are compared to the latest NEDS system totals,
below.
93
-------
TABLE 37. COMPARISON OF NEDS, SIP, AND NECRMP INVENTORIES FOR
PENNSYLVANIA
VOC (tons /year)
SIP area
Allentown
Point
Area
Total
Pittsburgh
Point
Area
Total
Philadelphia3
Point
Area
Total
Statewide
Point
Area
Total
NEDS SIP
6,869 6,035
39,201 17,703
46,070 23,738
29,405 24,543
152,299 124,726
181,704 149,269
190,785 135,186
324,630 219,677
515,415 354,863
182,428
830,342
1,012,770
NECRMP
7,444
18,897
26,341
28,750
122,448
151,198
139,889
215,205
355,094
264,054
465,776
729,830
NOx (tons /year)
NEDS SIP
40,501 55,523
21,270 18,508
61,771 74,031
183,275 191,111
98,180 90,526
281,455 281,637
105,864 146,049
199,241 179,443
305,105 325,492
626,011
503,869
1,129,880
NECRMP
52,161
21,690
73,851
133,792
119,786
253,578
141,759
208,303
350,062
538,019
491,114
1,029,133
alncludes Southern New Jersey and New Castle County, Delaware.
94
-------
TSP (ton/year) S02 (ton/year) CO (ton/year)
NECRMP points 248,622 2,027,142 754,852
NEDS points 261,867 2,000,221 546,463
Deviation (13,245) 26,921 (208,389)
A complete discussion of the NECRMP inventory for Pennsylvania is
presented in Volume XI.H
RHODE ISLAND
The Rhode Island Department of Environmental Management (DEM) provided
the basic input data used for both the NECRMP point and area source
inventories for Rhode Island. The DEM further assisted in the quality
assurance checking of the inventory by working directly with GCA to resolve
point source problems and providing additional data as needed to complete the
area source inventory compilation effort.
Point Source Inventory Quality
The Rhode Island point source data were obtained from the Rhode Island
DEM (through EPA Region I). The inventory was provided in NEDS format, and
reflects 1979 data. When received, the Rhode Island data suffered from
missing UTM coordinates, incomplete stack data, missing maximum design rates,
and the omission of a major VOC emitting facility. All of these problems were
resolved through the cooperation of the Rhode Island DEM. There are no
significant residual problems with the Rhode Island point source inventory of
which GCA is aware.
Area Source Inventory Quality
The Rhode Island DEM provided a 1980 area source inventory in hard copy
format. The initial gubmittal provided only statewide VOC totals for 17
composite categories. GCA subsequently received some of the information
needed to disaggregate the composite categories and allocate them to
counties. Still, the lack of comprehensive information required GCA make
certain assumptions to enable computerization of the inventory in EIS/PS
format. These cases are discussed individually in Volume XII.12 Some of
the more significant categories are discussed below.
Highway Vehicles—
VMT in miles per day and VOC emissions were supplied by the Rhode Island
DEM. The NO^, emissons were estimated by GCA by utilizing the average ratios
of VOC emissions to NOX emissions from other NECRMP highway vehicle
inventories for each vehicle type and then applying those percentages to the
Rhode Island VOC emissions estimates.
Gasoline Marketing—
The statewide 1980 Gasoline Sales total of 370,355 x 103 gallons was
distributed to counties based on population. Emissions and emission factors
95
-------
were supplied by the Rhode Island DEM for Stage I and II Gasoline Evaporation
and Storage Tank Breathing. Emissions from Gasoline Loading/Transit had not
been accounted for, therefore, GCA applied an emission factor of 6.87 Ib
VOC/1C)3 gallons.* Point source emissions were subtracted from the area
source emissions as described in Volume I.I
Fuel Combustion—
VOC emission factors for all of the fuel combustion categories and
composite VOC emission totals for each county were provided by the Rhode
Island DEM. Based on the distribution of VOC emissions found in the 1978 NEDS
Fuel Use Report,23 the 333 tons per year of VOC emissions in Rhode Island
were distributed to the Residential, Commercial/Institutional and Industrial
Sectors, by fuel types. The activity rates were then back calculated using
the VOC emissions and emission factors. The appropriate AP-42 emission
factors were used with the back-calculated fuel usages to derive NOX
emiss ions.
Railroad Locomotives—
The Rhode Island DEM provided a statewide VOC emission estimate. Using
a railroad transportation map,28 QC^ calculated the percentage of track
mileage in each county in order to distribute the emissions. Using the
appropriate AP-42 VOC emission factor, the activity rates were back
calculated. By applying the corresponding AP-42 NOX emission factor, NOX
emissions were calculated.
Aircraft and Vessels—
The Rhode island DEM supplied statewide composite VOC emissions for both
aircraft and vessels. GCA distributed the emissions to counties based on
population. NOX emissions were determined by applying the corresponding
ratios of VOC emissions to NOX emissions from the Massachusetts Inventory^
to the Rhode Island VOC emissions. A 2,000 Ib/ton VOC emission factor was
employed Co enable coding in EIS/AS.
Comparison to the NEDS and SIP Inventories
The NECRMP emission inventory for Rhode Island is compared to .the NEDS
and SIP inventory totals in Table 38.
Aa shown in Table 38, the SIP and NECRMP estimates for VOC agree within
less than 10 percent, although the NEDS estimate is 50 percent higher. Most
of this difference is attributable to organic solvent evaporation, the reasons
for which have previously been addressed in Section 3.
The NECRMP estimate for NOX is more than 25 percent higher than the
SIP estimate. At the time of preparation of the Rhode Islandl2 inventory,
NOX data for many area source categories had not been provided, most
notably, highway vehicles and residential combustion of fossil fuels. NOX
*Derivation of thia emission factor is discused in Volume
96
-------
TABLE 38. COMPARISON OF NEDS, SIP, AND NECRMP INVENTORIES FOR
RHODE ISLAND
SIP area
Providence
Point
Area
Total
Statewide
Point
Area
VOC
NEDS
14,613
87,594
102,207
14,644
87,594
(tons/year)
SIP NECRMP
16,094 13,375
51,500 49,264
67,594 62,639
13,375
49,264
NOx (tons/year)
NEDS SIP NECRMP
964 7,242 5,037
11,341 35,003 48,830
12,305 42,245 53,867
5,000 - 5,037
31,508 - 48,830
Total
102,238
62,639
36,508
53,867
97
-------
emissions from these categories in NECRMP were calculated using the Rhode
Island supplied VOC estimates and ratios of the appropriate NOX versus VOC
emission factors.
Apparent Completeness for Other Pollutants
The Rhode Island point source inventory includes estimates of TSP, SOo
and CO emissions, which are compared to the latest NEDS system totals, below.
TSP (ton/year) S02 (ton/year)
NECRMP points 3,441 6,055
NEDS points 3,315 6.115
Deviation 126 (60)
A complete discussion of the Rhode Island emission inventory can be found
in Volume XI1.12
VERMONT
Point source data for Vermont were obtained from the NEDS system
following an update in a previous contractual effort by GCA.23 The area
source data were developed by GCA using the procedures recommended by EPA in
the Guidelines 19 document. The Vermont Agency of Environmental Protection
(VAEP) assisted in this effort by supplying required corrections and
additional data as necessary for the point source inventory, and providing
much of the input data used in the area source inventory development.
Point Source Inventory Quality
The Vermont point source inventory was obtained from EPA's NEDS system
and reflect 1979 data. Since an extensive update of the Vermont data had
previously been completed,23 relatively few problems were detected as a
result of the quality assurance review. Those problems that were encountered
entailed missing SIC codes, exhaust gas flow rates, and missing or incorrect
UTM coordinates, which were subsequently corrected through the cooperation of
the Vermont Agency of Environmental Protection. There are no known residual
problems with the NECRMP point source inventory for Vermont.
Area Source Inventor/ Quality
The 1980 Vermont area source inventory was developed by GCA using data
supplied by various Vermont state agencies and methods generally consistent
with the Procedures for the Preparation of Emission Inventories for Volatile
Organic Compounds—Volume I, Second Edition. ^-9 Estimates of emissions of
VOC and NOX were derived for all 54 NECRMP area source categories. The
Vermont area source inventory is believed to reflect the most current data
available at the time of its preparation, and follows the EPA-prescribed
methods^ as closely as the available data allowed.
98
-------
Comparison with the NEDS Inventory
The NECRMP and NEDS inventories for Vermont are compared in Table 39.
TABLE 39. COMPARISON OF NEDS AND NECRMP INVENTORIES
FOR VERMONT
VOC (tons/year) NOx (tons/year)
NEDS
Statewide
Point 4,171
Area 35,562
Total 39,733
NECRMP NEDS NECRMP
4,921 866 1,595
33,551 43,023 25,791
38,472 43,899 27,386
As shown in Table 39, VOC emissions from NEDS and NECRMP are, overall, in
good agreement. For NOX, however, the NECRMP point source totals are
approximately 700 tons/year higher than those shown in NEDS. Over 300 TPY of
this difference can be attributed to the Beld Moran Plant in Burlington. The
NEDS data for this facility shows 304 TPY of NOX, reflecting a 1975 year of
record, while the NECRMP inventory, reflecting the 1979 update,23 shows
621 tons/year. The bulk of the remaining point source difference can be
attributed to industrial external combustion, for which NECRMP shows
670 tons/year (across 119 individual emission points) while NEDS shows
334 tons/year. Some, but most likely not all, of this deviation could be
related to use of the AP-42 Supplement 13 emission factors in NEDS. The
emission factors in NECRMP reflect modifications made through Supplement 12,
as Supplement 13 was released (August 1982) after completion of the point
source inventory for Vermont.
The NEDS system area source NOX totals are 40 percent higher than those
reported in NECRMP. Virtually all of this difference can be attributed to
off-highway diesel vehicles. The NECRMP estimate was derived using the 1978
NEDS Fuel Use Report27 to obtain off-highway diesel usage. Emissions were
calculated by distributing off-highway diesel fuel to equipment categories
(e.g., agriculture, construction, recreational vehicle, etc.) using survey
data summarized in the EPA Guidelines'^ an
-------
estimates, however, it is doubtful that this alone could account for the
magnitude of emissions difference in this category. Evidently NEDS assumes a
different vehicle mix.
Across the entire NECRMP study region, off-highway diesels account for
approximately eight (8) percent of the total NOX emissions from area sources
in NECRMP.* The NEDS system totals for Vermont show more than 52 percent of
the area source NOX as being emitted from off-highway diesels while the
NECRMP data for Vermont show off-highway diesels as being responsible for
15 percent of the area source NOX emissions. In this instance, the NECRMP
NOX estimate is believed to be more accurate.
Apparent Completeness for Other Pollutants
The NECRMP point source inventory for Vermont includes data on TSP, SOo
and CO, which are compared to the latest NEDS system totals below.
NECRMP points
NEDS points
Deviation
A complete description of the Vermont emission inventory can be found in
Volume XIII.13
VIRGINIA
The Virginia State Air Pollution Control Board (VSAPCB) provided the
input data used in both the NECRMP point and area source inventories. The
VSAPCB further assisted in this process by working closely with GCA in a
review of the data and by supplying required corrections, as needed.
Point Source Inventory Quality
The Virginia point source emission inventory was supplied by the Virginia
State Air Pollution Control Board (VSAPCB). The inventory reflects 1979 data,
which were supplied in NEDS format, as converted from EIS/P&R by the VSAPCB.
Although the inventory was generally found to be complete and well prepared,
gome minor problems were encountered. All estimated emissions based on stack
testa (estimation Method =» 1) were apparently "lost" in the State's conversion
from EI3/P&R to NEDS. These data were subsequently supplied by the VSAPCB.
Other problems addressed included out of date SCC codes and missing UTM
coordinates. All deficiencies identified were addressed, and were corrected
in the final EIS/PS master file.
*Nationwide, NEDS also shows this category as contributing 8 percent of the
total area source NOX.
100
-------
Data Confidentiality—
Under the Air Pollution Control Law of Virginia, the VSAPCB cannot
release point-specific process information. In order to obtain Virginia's
point source inventory, GCA had to consent to a confidentiality agreement
prohibiting the release of the process data considered confidential by the
Commonwealth of Virginia.
Originally, the VSAPCB indicated that the following fields would have to
be deleted from the inventory data prior to release to EPA (and, hence, made
public information):29
SPACES NEDS CARD NO. 6 DATA ITEM CONTENTS
18-25 Source Classification Code (Process Description)
26-32 Annual Fuel or Process Volume (Process Volume)
33-39 Maximum Hourly Design Rate (Process Capacity)
51-70 Comments (Equipment Description)
Withholding of the SCC would have seriously affected the further use of
the Virginia data in the NECRMP study. Temporal and pollutant species
patterns will be determined by SCC in developing modeling files from the
inventory data. Further, application of various control scenarios in the
modeling portion of the study will likely be dependent on the SCC.
EPA and the Commonwealth of Virginia were able to work out a solution
that enabled release of the SCC without compromising any provision of the Air
Pollution Control Law of Virginia. This entailed deleting the following
fields from each EIS point source record:29
• boiler capacity (columns 65 to 69 on card 11)
• emission estimation method (column 64 on card 13)
• fuel, process, solid waste operating rate (columns 37 to 43 on
card 21)
• maximum design rate (columns 44 to 50 on card 21)
• emission factor origin (column 32 on card 23)
• emission factor (columns 38 to 46 and 54 to 62 on card 23)
The above fields were deleted prior to delivery of the final data base to
EPA. The above deletions should not seriously affect the data's utility in
photochemical modeling.
Area Source Inventory Quality
The 1979 Virginia area source inventory was supplied in hard copy format
by the Virginia State Air Pollution Control Board. The inventory reflected
county summaries of area source emissions for all five criteria pollutants
that covered most of the 54 NECRMP area source categories. However, for
101
-------
reporting purposes, many of the categories had been aggregated together and,
therefore, did not conform with the NECRMP categories. Also, five categories
were determined to have been excluded from the inventory.
The documentation supplied with the inventory summary was reviewed for
consistency with the Volume l!9 recommended procedures and the EPA-approved
emission factors. In general, the methodologies used by Virginia were
consistent with those recommended as the standard in Volume 1.19 7^6 basic
problem with the inventory provided was the aggregation of categories for
reporting purposes. The information needed to disaggregate emissions into all
54 area source categories was subsequently supplied by the VSAPCB.
Comparison to the NEDS Inventory
The NECRMP inventory for Virginia is compared to the latest NEDS system
totals in Table 40.
TABLE 40. COMPARISON OF THE NEDS AND NECRMP INVENTORIES
FOR VIRGINIA
VOC (tons/year)
Regionwide
Point
Area
Total
NEDS
17,157
115,775
132,932
NECRMP
24,443
115,842
140,285
NOX (tons/year)
NEDS NECRMP
30,354 35,868
79,858 89,343
110,212 125,211
As shown in Table 40, area source VOC totals between NEDS and NECRMP are
in good agreement, while NECRMP point source totals are more than 7,000 TPY
higher than NEDS. Approximately half of this difference can be attributed to
three facilities (for which NEDS shows a 1977 year of record versus a 1979
year of record in NECRMP):
VOC (tons/year)
NECRMP NEDS
Allied Aviation - Arlington County 1535 Not listed
DuPont - Augusta County 2075 41
Reeves Brothers - Rockbridge County 2246 I860
5856 1901
102
-------
Similarly, for NOX point sources, NECRMP totals are more than 5,000 TPY
(18 percent) higher than NEDS. Over 4,400 TPY of this difference can be
attributed to the VEPCO Possum Pt. Power Plant, for which NEDS shows a 1978
year of record). The NECRMP inventory, reflecting 1979, is believed more
current for the above facilities.
Apparent Completenes for Other Pollutants
The NECRMP point source inventory for Virginia includes data on TSP, S02
and CO emissions, which are compared to the latest NEDS system totals, below.
TSP (ton/year) SO-? (ton/year)
NECRMP points 29,696 78,867
NEDS points 30.395 93.986
Deviation (699) (15,119)
A complete description of the Virginia emission inventory can be found in
Volume XIV.14
WASHINGTON, D.C.
The Washington, D.C. Council of Governments (WASHCOG) provide'* the basic
input data used for the NECRMP point and area source inventories. The Agency
also assisted in the area source inventory development by supplying more
detailed documentation of the emissions estimates supplied, when requested by
GCA. GCA forwarded a comprehensive list of suspected errors and questions
concerning the WASHCOG-supplied point source data, but a response was not
received. The implications of this lack of response are discussed below.
Point Source Inventory Quality
Point source data for the District of Columbia, reflecting 1980, were
provided in EIS/P&R format by the Metropolitan Washington Council of
Governments (WASHCOG). The point source inventory also contained data
covering five Virginia and two Maryland Counties. At the time of receipt of
these data, GCA had already obtained point source data covering those Virginia
and Maryland Counties from the Virginia SAPCB and the Maryland BAQNC,
respectively. The state supplied data were used in preference to WASHCOG1s
data for the following reasons:
• At the time of receipt of the WASHCOG data, work on the Virginia
point surce data had been completed and work on the Maryland data
was well under way;
• The Council of Governments did not formally respond to GCA'a lists
of questions/problems concerning the D.C. point source data; and
• GCA noted the obvious omission of major powerplants in the WASHCOG
data for Virginia and Maryland and, in general, was more confident
of the accuracy of the data supplied by the Virginia and Maryland
Agencies.
103
-------
GCA completed a QA audit of the WASHCOG-supplied data and forwarded a
list of suspected errors and questions to that Agency.
After awaiting a response from the agency for several months, GCA
corrected the obvious errors (dropped zeros) and determined that the remaining
problems were not sufficiently serious to warrant further delay. These were
primarily missing maximum design rates and in a few cases, seasonal throughput
percents.
Area Source Inventory Quality
The 1980 area source emission inventory was provided by the Washington
Council of Governments (WASHCOG). The data provided included emission
summaries for the District of Columbia, five surrounding Virgina counties and
two Maryland counties. Area source emissions for the Virginia counties had
already been inventoried by the Virginia State Air Pollution Control Board,
and had been reviewed by GCA and found to be comprehensive and reasonable.
Therefore, the WASHCOG data covering Virginia were not used in NECRMP.
However, although the Maryland Bureau of Air Quality and Noise Control
provided point source data for the entire State, the Maryland area source
inventory did not address the Washington, D.C. area counties of Montgomery and
Prince George's. Therefore, the District of Columbia area source inventory
provided by WASHCOG was utilized in NECRMP for those two Maryland counties.
The area source data were reviewed for comprehensiveness and consistency
with the prescribed procedures. The most notable deficiency was the lack of
process or activity data and emission factors. GCA requested this information
from WASHCOG and subsequently received sufficient information to enable GCA to
complete the District of Columbia area source inventory. For many categories,
only emission factors and emission levels were provided. In these instances,
GCA back calculated process rates from the WASHCOG supplied data.
Some deviations from EPA prescribed procedures and emission factors were
noted, and are discused below.
WASHCOG employed some emission factors which deviated from the factors
recommended by GCA^- or prescribed by EPA in Procedures for the Preparation
of Emission Inventories for Volatile Organic Compounds—Volume I, Second
Edition.19These primarily involved categories for which per-capita factors
are typically used. For degreasing, the WASHCOG inventory accounts for VOC
omissions from cold metal degreaaing only, and presumes all open top vapor and
conveyorized degreasing emissions have been accounted for in the point source
inventory.
For architectural surface coating, the WASHCOG inventory documentation
indicates that the EPA-prescribed VOC emission factor of 4.6 Ib/capita was
felt to overstate emissions in the Metropolitan Washington area. Based on
data derived from the National Paint and Coating Association's 1979 Data Bank
Program, a factor of 1.76 Ib/capita was determined to be a more accurate
estimate of VOC emissions from architectural surface coating in the District
of Columbia area.
104
-------
The WASHCOG inventory employed a 0.4 Ib/capita factor to account for VOC
emissions from graphic arts establishments. While the EPA prescribed value is
0.8 Ib VOC/capita, WASHCOG indicated that the lower value was employed as a
result of an extensive survey of local graphic arts facilities.
WASHCOG discarded both the EPA per-capita and per-eraployee factors as
resulting in unrealistically high emission estimates for autobody
refinishing. WASHCOG conducted a survey of 24 autopainting businesses to
determine average numbers of full and partial paint jobs performed on a
typical summer day by a single shop. WASHCOG derived a VOC emission estimate
using the above derived average shop workload; a 6 Ib VOC/total paint job
emission factor; a 3 Ib VOC/partial paint job factor; and the total number of
shops listed in the yellow pages. The above deviation from the recommended
methods were identified and justified in the WASHCOG documentation.
Comparison with the NEDS and SIP Inventories
The NECRMP inventory for the District of Columbia is compared with the
NEDS and SIP inventory totals in Table 41.
Overall the SIP and NECRMP inventory totals are in reasonable agreement,
within about 10 percent. There are, however, substantial differences for
individual source categories which are summarized in Tables 42 and 43 for VOC
and NOX, respectively.
Highway vehicles, most notably trucks, exhibit the greatest differences
in both VOC and NOX emissions. Part of this difference is due to using a
tons/day to tons/year conversion factor of 365 for SIP comparison. In NECRMP,
to ensure proper temporal allocation when creating modeler's tapes for use in
the Regional Oxidant Model (ROM), the NECRMP temporal allocation factors1?
were utilized to convert the WASHCOG tons/day data into tons per year. For
the District of Columbia and the Maryland counties of Montgomery and Prince
George, when the NECRMP temporal allocation factors are applied, on-highway
VOC and NOX emissions match exactly. The NECRMP highway vehicle emission
estimates for the Virginia counties were obtained from the Virginia State Air
Pollution Control Board (VSAPCB) for reasons discussed previously. Vehicle
emissions for these counties, particularly for trucks, differ significantly
even though WASHCOG reportedly obtained Virginia and Maryland data from the
respective state agencies.
VOC estimates for gasoline evapoation losses are significantly higher in
NECRMP than the SIP since NECRMP added an area source component (per the EPA
Guidelines methodology 1-9) to account for loading of tank trucks at gasoline
terminals and bulk plants.
The category having the largest NOX discrepancy—residential,
commercial, institutional fuel use—is actually of less consequence than
indicated by the large difference in emissions, since very little of this
total is allocated to the oxidant season. This category also accounts for a
significant difference in annual VOC emissions.
105
-------
as
O
Cn
en
U
i— i
on
p
H
Z
ril
>
z
— 1
(,
CcS
U
U
Z
o
z
<
a.
M
en
M
CO
a
u
z
U
as
H •
CJ
tL. •
O Q
Z -
O Z
en Q
^ H
5 Z
<4 Z
3,-j i-l
5 s
a en
r i <**
^ ^*«
,
C a*
O M
u en
^/
£
Z
en
a
u
z
p.
jjg*
2
0
^.^
U
«J
0)
>, a-
•». M
c en
o
u
s«^
CJ
o
en
a
w
§
«J
1)
Ui
(tj
Oi
^^
en
CN C*
CO r»
•* o
o in
m o
r-J
— 1 CO
r-. o
vO m
M M
\O r+
~f CO
O> CO
vO O
CN O
A M
^ P^
«* CO
vO O
in en
O fN
M A
r- O
CN
— i
tn en
CN cr>
O 0
M Ml
<• CT\
— 1
r-l
en CN
CN vO
C^ CT>
M M
en
M **
— ( 1-4
CN
CN CN
f- — i
vO «3-
CN
CN
-I -*
O CO
O> CN
A
en
>*
>J -u
r-l C (0
c -^ ca
O o ^
cu
CN
O
CN
CN
en
i— i
in
n
cn
»— i
m
-a-
0
M
CO
CN
,— 1
vO
CN
A
en
CN
•*
CO
o
en
CN
in
co
r— 1
A
<3-
c
•-H
W
0)
• H
JJ
e
3
o
o
5
!3
• H
i— 1
r— 1
• H
3 •
•o
0) C
U Q
C r-l
•"• f^
U U
&. m
S
•a
c c
a) .1-1
A CO
C -l
u O-i
oo
C T3
•-» C
r-l Cfl
JJ
^ £r
o} 0)
v s
ng o
r oo
i— 1 U
u c
C Q
M S
ca
106
-------
P-J
<-C
Cc
CJ
w
co
CO
w
£>
C-H
M
C/D
*
^«
i
u
Q
2^
(— "4
co
w
i — i
CJ
J
%
CJ
CO
Q
CJ
§
OS
O
"i
*y
•*•*
f
CN
W
H-J
CQ
"^4
fH
tn
u
C
es
a
o
CJ
a.
0? CO 5-
Qua.
f»1 O C— t
Z »-> — •
— < ^-N
ft. tfl >4
H-< 4-1 O«
yj o H
>,
4J
C
3
O
O
^-^
CU
to
u
t/5
>^
O
00
01
tfl
u
01
£
3
Q
CO
"O "O
01 0)
en en
3 3
en en
l-i P
0 0
U 4J
u a
to ttj
i i ( «i ]
, ,__(
efl tfl
0 0
a a
B 3
01 01
LJ L?
cu Cu
s ?
S Q?
O O
z S
U-4 M-l
• <-l »H
CN| — ^
00
3
en
O
u H
3 Q
* CM
ro ro
in ro
p^ O1^
^ 'vf
^
^^ >
< -
^
*x^
oO tn
oo c 01
cT •-< u
•^ C '-I
—4 >J U-l
T) 3
c .a — i
tfl ""-., tfl
j= -o si
4) 3
CO) •— * O
"-" tn o 3
<—* en v^ ^
00 4-1 >->
en r-i Cw
tfl o
O CJ
tn
a>
u
^4
•3
0
a>
u
pj
.H
O
CU
en
tfl
•o
0)
01
J>
0
O
^4
Cfl
•o
e
3
tn
en
tfl
CLj
*?;
ai
CJ
W
O m CN
CO CN
lA -J
^^
in o O
•«o
«s
r-4
^^
en
^^ r-4
Cfl 3
C Q
cfl C O ^
> O •*
4J 4J C
•-> QO cfl 3
^-i C Z O
a>
U -r-l
00 —(
01 -H
Q cu
tn >,
e 3 tfl
00) 3
op oo
C W •-!
•^ to J3
en c/3 u-i
tfl UJ
3 0
CO
u
tfl
-a
>^
o>
>
u
3
tn
i-^
u
o
13
0)
tn
3
8
CJ
CM
in
oo
i— i
u-l
O^
O
^
e-4
_4
«
X
M
tfl
Cu
cT
• H
u
tfl
0
CJ
01
CJ
tfl
U-l
14
tn
•
JS
o
ij
F4
>>,
1
CU
tfl
u
'^J
rH
^
v^
T3
01
en
3
Si
s
O
00
oo
•t
^
^J-
o
^
•«
CM
^
f%
X
CO
^
• H
tfl
\i
01
3
tn
c
o
CJ
^
tfl JJ
•-4 C
0) -4
§o
to
o
0
vD
CM
CM
i-4
>
e
o
u
,r4
^4
\4
«£
00
"O C
C -i-i
to
•H 00
4-1 .-1
t« V4
01 u-i
Si Ol
^
>
107
-------
Eu
2C
2
u
Z
CO
=3
Sri
>
tx
1— 1
CO
M
^«
i
•3
o
c
Z
1— 1
co
u
l_l
0
cu
W
U
W
h- 1
a
X
o
OS
O
5
^*
.
m
U
t i
eQ
H
^
U O H
Z *•> ^
••j ^**
a, eo >
M U CU
CO 0 H
>,
u
C
0
CJ
ct>
u
CO
x
o
00
0)
CO
U
0)
u
u
3
o
CO
C/J
W
o
0
fl
C (u
0
CO r— 4.
co -
o
^J
— 1
C •-<
03 -•
oO ^J *^<
c u
•rt C
— J •-!
U h
< a,
^,
4)
3
.
4J
V]
C
• r4
'"**x *^~*
• ^>
s c
S -1
C ^
u
-^ 01
^ '^ f>» [*«•
<3\ m ts r-»
^
> _ x
Q co
. Z uw
ca id ca
e « a •-* e
CQ 01 Q to CO
•M SO Cb -^
_l U „ ^4
^J 0 C --"
•^ o) o e •*
3 CJ u 03
00 J-*
OI O) C op 01
U CJ -^ C CJ
e c £ -H c
•H '^ CO *^ >r4
u u
-------
Several point sources, namely Allied Aviation at the two main airports,
and the two Washington newspapers, contribute significantly to the NECRMP VOC
totals, but are not represented at all in the SIP, even though NECRMP data for
these sources were obtained directly from WASHCOG.
No area source degreasing losses were included in the NECRMP inventory
for Virginia because all such activities were claimed to have been handled as
point sources, per Virginia's APCB. Nevertheless, WASHCOG included an area
source degreasing component, using the EPA^-9 suggested 3 Ib/capita-yr factor.
Auto painting and architectural surface coating emissions are
considerably lower in the SIP. WASHCOG developed lower factors for these two
source categories through contacts with the National Paint and Coating
Association (NPCA) and various local autobody shops. GCA used the
EPA-suggested factors in the NECRMP inventory.
Apparent Completeness for Other Pollutants
The NECRMP point source inventory for Washington, D.C. includes data on
TSP, S02 and CO emissions, which are compared to the latest NEDS system
totals below.
TSP (ton/year) S02 (ton/year) CO (ton/: aar)
NECRMP points 1,398 11,082 4,295
NEDS points 2,533 14,519 6,249
Deviation (1,135) (3,437) (1,954)
A complete discussion of the Washington, D.C. emission inventory can be
found in Volume XV. ^-5
WEST VIRGINIA
The EPA Project Officer provided an agency contact in the West Virginia
Air Pollution Control Commission (WVAPCC). The WVAPCC received EPA liaison
funding and were to have been responsible for confirming data, supplying
required corrections or additional data, interfacing with individual sources
as necessary, and concurring on the comprehensiveness and accuracy of the
final data base. However, during the period of performance of the project,
the WVAPCC were unable to provide the basic input data needed for the
inventory. Therefore, as an interim measure while the WVAPCC completes their
inventory, alternative data sources were agreed upon.
Point Source Inventory Quality
The West Virginia Air Pollution Control Commission (WVAPCC) did not
complete their point source inventory in sufficient time to be used in the
initial applications of the NECRMP inventory. As an interim measure, while
109
-------
the WVAPCC completes their in-house comprehensive inventory, GCA accessed
West Virginia emissions data from the Emissions Inventory for the SURE
Region. *-8 Since these data are to be used only as an interim measure, a
significant degree of data correction effort was not expended. Rather,
efforts focused on correcting only those data problems that prohibited
conversion of the NEDS formatted data into EIS/PS format. One such problem
related to data within the West Virginia file that had been claimed
confidential. For previous use in SURE,18 the boiler design capacities,
process rates, emission estimation methods, hourly design rates, and percent
space heat fields for certain facilities had been "masked out" with alphabetic
characters. Since EIS/PS would not permit nonnumeric or blank characters in
these fields, they had to be changed to numeric characters to enable
conversion. The emissions estimation fields were changed to "5" (special
emission factor) and the remaining fields were changed to zeros. The most
significant problem with the interim West Virginia point source data is its
age. Although the base year of the SURE1** inventory is 1978, most of the
data actually represent 1975.
Area Source Inventory Quality
The West Virginia APCC did not complete their area surce inventory in
sufficient time to be used in the initial applications of the NECRMP
inventory. As an interim measure, while the WVAPCC completes this inventory,
GCA assessed the West Virginia area source inventory from the Emissions
Inventory for the SURE Region. 18 4. few categories were not sufficiently
disaggregated in SURE to conform with the 54 NECRMP area source categories.
In these instances, data from NEDS were obtained and coded directly into
EIS/AS format. All population based categories were recomputed by GCA using
the per-capita factors from Volume I of the Procedures for the Preparation of
Emission Inventories for Volatile Organic Compounds, Second Edition-1-^ and
the 1980 Census of Population.
The interim area source inventory should be adequate for use in NECRMP
until West Virgina APCC completes their area source inventory.
Comparison with the NSDS Inventory
The NECRMP and NEDS inventory data for the NECRMP study area portion of
West Virginia is presented in Table 44.
Overall, the NEDS and NECRM? data for West Virginia are in good agreement
particularly for area sources. Some differences in point source totals exist,
however, even though SURE utilized NSDS point source data. Changes made to
the West Virginia point source data for SURE were not made in the NEDS
system. Similarly, NEDS annual updates of powerplant and Federal government-
owned facilities data are not reflected in NECRMP.
However, since total emissions match reasonable well, the NECRMP
inventory should be adequate for interim use until the WVAPCC completes their
emissions inventory update.
110
-------
TABLE 44. COMPARISON OF NEDS AND NECRMP INVENTORIES
FOR THE WEST VIRGINIA PORTION OF THE NECRMP
STUDY AREA
VOC (tons/year)
Point
Area
Total
NEDS
7,224
90,804
98,028
NECRMP
15,841
87,200
103,041
NOx (tons/year)
NEDS
434,221
91,951
526,172
NECRMP
401,152
91,962
493,114
111
-------
Apparent Completeness for Other Pollutants
The NECRMP point source inventory from West Virginia includes data on
TSP, SC>2, and CO emissions, which are compared to the latest NEDS system
totals below.
TSP (ton/year) SQ2 (ton/year) CO (ton/year)
NECRMP points 128,625 1,219,811 92,953
NEDS points 128,510 1,023,192 170,510
Deviation 115 196,619 (77,557)
A complete discussion of the West Virginia point source inventory can be
found in Volume XVI.^6
SUMMARY
A summary of the NECRMP inventory review and evaluation is presented in
che Executive Summary of this Volume.
112
-------
SECTION 5
REFERENCES
1. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy, S. V. Capone.
Northeast Corridor Regional Modeling Project Annual Emission Inventory
Compilation and Formatting - Volume I - Project Approach.
EPA-450/4-82-013a. August 1982.
2. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume II - Connecticut Emission Inventory.
EPA-450/4-82-013b. September 1982.
3. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy. Northeast
Corridor Regional Modeling Project Annual Emission Iw sntory Compilation
and Formatting - Volume III - Delaware Emission Inventory.
EPA-450/4-82-013c. September 1982.
4. Sellars, F. M., B. J. Bosy, M. J. Geraghty, A. M. Kiddie, S. V. Capone.
Northeast Corridor Regional Modeling Project Annual Emission Inventory
Compilation and Formatting - Volume IV - Maine Emission Inventory.
EPA-450/4-82-013d. September 1982.
5. Sellars, F. M., A. M. Kiddie, B. J. Bosy, S. V. Capone. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume V - Maryland Emission Inventory.
EPA-450/4-82-013e. September 1982.
6. Seilars, F. M., B. J. Bosy, A. M. Kiddie, S. V. Capone. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume VI - Massachusetts Emission Inventory.
EPA-450/4-82-013f. September 1982.
7. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume VII - New Hampshire Emission Inventory.
EPA-450/4-82-013g. September 1982.
8. Sellars, F. M., A. M. Kiddie, B. J. Bosy, M. J. Geraghty. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume VIII - New Jersey Emission Inventory.
EPA-450/4-82-013h. September 1982.
113
-------
9. Sellars, F. M., A. M. Kiddie, B. J. Bosy, M. J. Geraghty. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume IX - New York Emission Inventory.
EPA-450/4-82-013i. June 1983.
10. Haupt, S. E., F. M. Sellars, M. J. Geraghty, A. M. Kiddie, B. J. Bosy.
Northeast Corridor Regional Modeling Project Annual Emission Inventory
Compilation and Formatting - Volume X - Ohio Emission Inventory.
EPA-450/4-82-013j. September 1982-
11. Sellars, F. M., A. M. Kiddie, B. J. Bosy, M. J. Geraghty. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume XI - Pennsylvania Emission Inventory.
EPA-450/4-82-013k. June 1983.
12. Sellars, F. M., B. J. Bosy, A. M. Kiddie, S. V. Capone. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume XII - Rhode Island Emission Inventory.
EPA-450/4-82-0131. September 1982.
13. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume XIII - Vermont Emission Inventory.
EPA-450/4-82-013m. September 1982.
14. Sellars, F. M., M. J. Geraghty, A. M. Kiddie, B. J. Bosy, S. V. Capone.
Northeast Corridor Regional Modeling Project Annual Emission Inventory
Compilation and Formatting - Volume XIV - Virginia Emission Inventory.
EPA-450/4-82-013n. September 1982-
15, Sellars, F. M., A. M. Kiddie, B. J. Bosy. Northeast Corridor Regional
Modeling Project Annual Emission Inventory Compilation and Formatting -
Volume XV - Washington, D.C. Emission Inventory. EPA-450/4-82-013o.
September 1982.
Ib. Sellars, F. M., A. M. Kiddie, B. J. Bosy, S. V. Capone. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume XVI - West Virginia Emission Inventory.
E?A-450/4-82-013p. September 1982.
17. Geraghty, M. J., F. M. Sellars, A. M. Kiddie, B. J. Bosy. Northeast
Corridor Regional Modeling Project Annual Emission Inventory Compilation
and Formatting - Volume XVII - Development of Temporal, Spatial and
Species Allocation Factors. EPA-450/4-82-013q. September 1982.
18. Klemm, H. A. and R. J. Brennan. Emissions Inventory for the SURE
Region. GCA/Technology Division for the Electric Power Research
Institute. EPRI EA-1913. April 1981.
19. Lamason, W. H. II, and Thomas Lahre. Procedures for the Preparation of
Emission Inventories for Volatile Organic Compounds. EPA-450/2-77-028.
September 1980.
114
-------
20. End Use of Solvents Containing VOC. EPA-450/3-79-039. U.S.
Environmental Protection Agency, Research Triangle Park, NC. May 1979.
21. Virginia State Air Pollution Control Board. Emission Inventory for the
Year 1979. December 1980.
22. Mehta, D. H. and J. E. Spessard. Directory of Volatile Organic Compounds
Sources Covered by Reasonably Available Control Technology Requirements
(RACT). EPA-450/4-81-007. February 1981.
23. Shah, Mahesh C., S. E. Haupt and J. N. McHenry. New Hampshire and
Vermont Point Source Emission Inventory Update for NECRMP.
EPA-450/4-81-017. March 1981.
24. U.S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors, AP-42, Third Edition and Supplements. August 1977.
25. Dyba, Raymond, New Jersey Department of Environmental Protection.
Correspondence to George Kerr, EPA/Region II. 12 July 1982.
26. Green, R. D., B. J. Bosy, E. J. Limberakis, S. V. Capone. Allegheny
County VOC and NOX Point Source Emission Inventory. EPA-903/9-82-002.
January 1982.
27. U.S. Environmental Protection Agency. 1978 National Lmis&i n Data S; stem
(NEDS) Fuel Use Report. EPA-450/4-80-028. December 1980.
28. U.S. Department of Transportation, Federal Railroad Administration.
State Transportation Maps. GPO-050-005-00020-8, 1977.
29. Erskine, W. W. Virginia State Air Pollution Control Board. Written
Communication to F. M. Sellars, GCA/Technology Division.
December 27, 1980.
115
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
Hfcf'UHT NO 2.
KPA-450/4-B2-013r
4. nTLfT'ANo SUBTITLE Northeast~Corridor Regional Modeling
Project Annual Emission Inventory Compilation and For-
matting. Volume XVIII: Inventory Review and
Evaluation
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
July 1983
6. PERFORMING ORGANIZATION CODE
7 AUTHOfMS)
Frederick M. Sellars and Barbara J. Bosy
8. PERFORMING ORGANIZATION REPORT NO.
GCA-TR-82-17-G(18)
HCHHORMINCl ORGANIZATION NAME AND ADDRESS
CCA/Technology Division
213 Burlington Road
Bedford, Massachusetts 01730
12 SPONSORING AG1 NCY NAME AND ADDRESS
U.S. Environmental Protection Agency, Air Management
Technology Branch, Monitoring and Data Analysis Divi-
sion, Office of the Quality Planning and Standards,
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3510
Work Assignment 14
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
IB rut'Pl.t Ml N TARY NOT>S
KPA Project Officers: James H. Southerland and Thomas F. Lahre
This report discusses the development of the Northeast Corridor Regional
Modeling Project (NECRMP) annual regional emission inventory. The inventory
reflects 1979/1980 data and focuses on point, area and mobile source emissions of
Volatile Organic Compounds (VOC) and Nitrogen Oxides (NOX), although particulate,
sulfur oxides and carbon monoxide emissions were also compiled for point sources.
The study area includes the entire northeast quadrant of the United States from
longitude 69° to 82° West, and latitude 38° to 45° North. The Volume presents a
summary of emissions for the entire study area. Also included is a detailed analysis
of the overall quality of the data base and an assessment of the data's utility for
photochemical modeling.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Emission Inventory Nitrogen Oxides
inventory
Source Inventory
Area Sources
Ozone
Volatile Organic
Compounds
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATl Held/Group
18 OlSrHIUUI ION '.rATLMENT
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page)
j Unclassified
22. PRICE
form 2370-1 (R.v. 4-77)
PREVIOUS EDI TION IS OBSOL. £T£
-------
?is
~*
°?
tfl
P
Is
>mc
•832
3JQO
I
A
oo
ro
O
CO
vj 0) «
— = o
-•0.3
c/>
S
i
rnm 5 3
33- -1 r>
m o r* o
o o -i
DO -S
» I
m>T)rn-n-o
•DID £ 3 o o
------- |