vvEPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-85-003
June 1985
Air
Evaluation And
Application Of The
Urban Airshed
Model In The
Philadelphia Air
Quality Control
Region
-------
EPA-450/4-85-003
Evaluation And Application Of The Urban
Airshed Model In The Philadelphia Air
Quality Control Region
Prepared By
Jay L Haney
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, California 94903
and
Thomas N. Braverman
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Office of Air and Radiation
Research Triangle Park, North Carolina 27711
Prepared For
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office Of Air And Radiation
Office Of Air Quality Planning And Standards
Research Triangle Park, NC 27711
June 1985 U.S. Environmental P~-otrctl-
Region 5, L.'^/ary ' :: - -'
230 S. 1\J '-• '. •-! '•- ^ •
Chicago, IL C'j.•-.>.
-------
This report has been reviewed by The Office Of Air Quality Planning And Standards, U.S. Environmental
Protection Agency, and has been approved for publication. Mention of trade names or commercial products
is not intended to constitute endorsement or recommendation for use.
EPA-450/4-85-003
-------
ACKNOWLEDGMENTS
A number of people made significant contributions to the work reported
herein. The authors thank the many individuals at Systems Applications
and EPA-OAQPS who participated in the various technical tasks associated
with these modeling efforts. Particular recognition is due Drs. Tom
Tesche and Christian Seigneur and Mr. James Kill us of Systems Applica-
tions, and Messrs. Norman Posseil and David Layland of EPA-OAQPS. Special
thanks are also extended to the Systems Applications' production staff and
to the project editor, Ms. Carol Wade.
m
-------
CONTENTS
Notice 11
Acknowl edgements ill
Figures vii
Tab! es xi i i
1 INTRODUCTION 1
Overview of Study 1
Technical Approach 2
2 CHARACTERIZATION OF THE 13 JULY 1979 OZONE EPISODE 7
3 DESCRIPTION OF MODEL INPUTS FOR 13 JULY 1979 11
Modeling Region Specifications 11
Mixing Heights 12
Wi nd Fi el d 17
Background Concentrations 37
Initial Conditions 37
Boundary Conditions 40
Emission Inventory 42
Metscal ars 49
Terrai n 52
4 CHARACTERIZATION OF THE 19 JULY 1979 EPISODE 55
5 DESCRIPTION OF MODEL INPUTS FOR 19 JULY 1979 57
Mi xi ng Hei ghts 57
Wind Field 59
Background Concentrations 69
Initial Conditions 75
Boundary Conditions 75
Metscal ars 78
6 ANALYSIS OF URBAN AIRSHED MODEL PERFORMANCE FOR THE
PHILADELPHIA SIMULATIONS OF 13 AND 19 July 1979 81
Model Performance Evaluation Measures 81
Model Evaluation Results for the 13 July 1979 Simulation.... 87
Model Evaluation Results for the 19 July 1979 Simulation.... 96
Comparison of Philadelphia Results with
UAM Performance in Other Cities 107
-------
7 OZONE SENSITIVITY ANALYSIS Ill
Met hodol ogy Ill
Results of Ozone Sensitivity Simulations for 13 July 116
Results of Ozone Sensitivity Simulations for 19 July 130
8 SUMMARY AND CONCLUSIONS 175
References 183
Appendix A: COMPILATION OF AIRSHED RESULTS FOR 13 JULY 1979
Appendix B: COMPILATION OF AIRSHED RESULTS FOR 19 JULY 1979
VI
-------
LIST OF FIGURES
1-1 Geographical location of the Philadelphia airshed modeling region 4
1-2 Philadelphia airshed modeling region 5
2-1 Synoptic situation, 0700 EST, 13 July 1979 8
3-1 Temperature sounding for JFK Airport on
13 July 1979—0700 EST 13
3-2 Temperature sounding for JFK Airport on
13 July—1900 EST 14
3-3 Temperature sounding for Dulles airport on
13 July 1979—0700 EST 15
3-4 Temperature sounding for Dulles Airport
on 13 July 1979—1900 EST 16
3-5 Mixing height profile for urban and rural cells for
the 13 July 1979 simulation 19
3-6 Observed surface wind vectors and wind vectors used in
preparing the interpolated surface wind field for
13 July 1979 20
3-7 Upper level winds aloft at JFK Airport on
13 July 1979—0700 EST 27
3-8 Upper level winds at JFK Airport on
13 July 1979—1900 EST 28
3-9 Upper level winds at Dulles Airport on
13 July 1979—0700 EST 29
3-10 Upper level winds at Dulles Airport on
13 July 1979—1900 EST 30
vn
-------
3-11 Airshed Model surface winds for 13 July 1979 33
3-12 Physical boundaries used in the 13 July 1979 Airshed
Model simulation 41
3-13 Surface layer total hydrocarbon emissions for the
Philadelphia airshed region for 1979 47
3-14 Surface layer total NOX emissions for the
Philadelphia airshed region for 1979 48
3-15 Land use classification for the Philadelphia airshed
modeling region 54
4-1 Synoptic situation, 0700 EST, 19 July 1979 56
5-1 Temperature sounding for Philadelphia,
0410 EST, 19 July 1979 58
5-2 Mixing height profiles for urban and rural
cells on 19 July 1979 61
5-3 Observed surface wind vectors and vectors used in
preparing the interpolated surface wind field for
three time periods for 19 July 1979 62
5-4 Pibal sounding at Wilmington, Delaware on
19 July 1979, 1200 EST 66
5-5 Pibal sounding at Philadelphia on 19 July, 1979, 1150 EST 67
5-6 Pibal sounding at Trenton, New Jersey on 19 July 1979,
1200 EST 68
5-7 Airshed Model surface winds for 19 July 1979 71
5-8 Boundary specifications for 19 July 1979 simulations 77
6-1 Scatter plot of predicted and observed station peak
ozone concentrations for 13 July 1979 91
6-2 Mean normalized bias, mean normalized error as
function of measured 03 concentrations and distri-
bution of residuals for 13 July 1979 93
vm
-------
6-3 Time series (24 hours) of predicted versus observed
ozone, and spatially predicted ozone (pphm), for the
13 July 1979 simulation, 1600-1700 EST 94
6-4 Scatter plot of predicted and observed ozone
concentrations for 13 July 1979 97
6-5 Scatter plot of predicted and observed station peak
ozone concentrations for 19 July 1979 102
6-6 Mean normalized bias, mean normalized error as
function of measured 03 concentrations and
distributions of residuals for 19 July 1979 104
6-7 Time series (24 hours) of predicted versus observed
ozone and spatially predicted ozone for the simulation
of 19 July 1979, 1400-1500 EST 105
6-8 Scatter plot of predicted and observed ozone concen-
trations for 19 July 1979 108
7-1 Predicted ozone response to hydrocarbon emission
reductions for peak regional ozone in the Philadelphia
urban plume for 13 July 119
7-2 Predicted ozone response to hydrocarbon emission
reductions for 13 July at the Roxy Water, PA monitor 121
7-3 Predicted ozone response to hydrocarbon emission
reductions for 13 July at the Norristown, PA monitor 122
7-4 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the
Philadelphia urban plume for 13 July 123
7-5 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Roxy Water, PA monitor 124
7-6 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Norristown, PA
monitor 125
7-7 Total ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the
Philadelphia urban plume for 13 July 126
-------
7-8 Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Roxy Water, PA
monitor 127
7-9 Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Norristown, PA
monitor 128
7-10 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.25HC minus D.BASE) 132
7-11 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (0.50HC minus D.BASE) 133
7-12 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.75HC minus D.BASE) 134
7-13 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.BK03 minus O.BASE) 135
7-14 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.BKHC minus D.BASE) 136
7-15 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.BKHC.03 minus D.BASE) 137
7-16 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BK03 minus D.BASE) 138
7-17 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BKHC minus D.BASE) 139
7-18 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BKHC.03 minus D.BASE) 140
7-19 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BK03 minus D.BK03) 141
7-20 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BKHC minus D.BKHC) 142
7-21 Maximum deficit/enhancement for ozone (pphm) for all hours
for 13 July (D.50HC.BKHC.03 minus D.BKHC.03) 143
-------
7-22 Predicted ozone response to hydrocarbon emission reductions
for peak regional ozone in the Philadelphia urban plume
for 19 July 146
7-23 Predicted ozone response to hydrocarbon emission reductions
for 19 July at the Downington, PA monitor 147
7-24 Predicted ozone response to hydrocarbon emission reductions
for 19 July at the Roxy Water, PA monitor 148
7-25 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the
Philadelphia urban plume for 19 July 149
7-26 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 19 July at the Downington, PA monitor 150
7-27 Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 19 July at the Roxy Water, PA monitor 151
7-28 Total ozone reduction (%) versus percent hydrocarbon emission
reduction of peak regional ozone in the Philadelphia urban
plume for 19 July 152
7-29 Total ozone reduction (%) versus percent hydrocarbon emission
reduction for 19 July at the Downington, PA monitor 153
7-30 Total ozone reduction (%) versus percent hydrocarbon emission
reduction for 19 July at the Roxy Water, PA monitor 154
7-31 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.25HC minus 6.BASE) 158
7-32 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.50HC minus 6.BASE) 159
7-33 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.75HC minus 6.BASE) 160
7-34 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B25HC minus 6.BASE) 162
7-35 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC minus 6.BASE) 163
XI
-------
7-36 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B75HC minus 6.BASE) 164
7-37 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.BK03 minus 6.BASE) 165
7-38 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.BKHC minus 6.BASE) 166
7-39 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.BKHC.03 minus 6.BASE) 167
7-40 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BK03 minus 6.BASE) 168
7-41 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BKHC minus 6.BASE) 169
7-42 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BKHC.03 minus 6.BASE) 170
7-43 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BK03 minus 6.BK03) 171
7.44 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BKHC minus 6.BKHC) 172
7-45 Maximum deficit/enhancement for ozone (pphm) for all hours
on 19 July (6.B50HC.BKHC.03 minus 6.BKHC.03) 173
-------
LIST OF TABLES
3-1 Urban and rural mixing height values used in the DIFFBREAK
file for 13 July 1979 18
3-2 Spatially constant wind vectors for level 4 winds for
13 July 1979 31
3-3 Background concentration values for 13 July at the top of the
modeling region (TOPCONC), as initial concentrations above the
mixing height, and for all levels of all boundaries except
the levels below the mixing height on the Southeast boundary 38
3-4 Initial conditions for 13 July 1979 39
3-5 Southeast boundary conditions for cells below the
mixing height for the simulation of 13 July 1979 43
3-6 Hourly emissions of NOX and hydrocarbon (tons/hr) used
for the 1979 Philadelphia emission inventory 45
3-7 Total daily emissions by source type (g«mole) in the
1979 Philadelphia inventory 46
3-8 METSCALAR file for the 13 July 1979 airshed simulation 51
3-9 Surface roughness and vegetation factor values 53
5-1 Urban and rural mixing height values used in the DIFFBREAK
file for 19 July 1979 60
5-2 Estimated spatially constant, temporally varying wind input
for levels 2, 3, and 4 for the 19 July 1979 wind file 70
5-3 Background concentration values for 19 July at the top of
the modeling region (TOPCONC), as initial concentrations
above the mixing height, and for all levels of all boundaries
except the levels below the mixing height on the Northeast
and East boundaries 74
xi
-------
5-4 Initial conditions at station monitors for 19 July 76
5-5 Boundary conditions for the Northeast and East boundaries
below the mixing height estimated from data collected at
the Van Hi Seville, New Jersey monitor 79
5-6 METSCALAR inputs for the 19 July 1979 simulation 80
6-1 Maximum predictions/observations for 13 July 1979
for ozone 88
6-2 Performance measures for the DAM simulation of the
13 July 1979 episode in Philadelphia (for peak station
predictions/observations and ozone measurements
above 5 pphm) 89
6-3 Maximum predictions/observations for 19 July 1979
for ozone (pphm) 99
6-4 Performance measures for the UAM simulation of the
19 July 1979 episode in Philadelphia 100
6-5 Comparison of UAM performance evaluation for ozone in
Los Angeles, Tulsa, Sacramento, Denver, and Philadelphia 109
7-1 Simulation designations for Philadelphia Airshed
sensitivity analysis for 13 and 19 July 1979 112
7-2 Total RHC concentrations used as initial conditions for
base case and hydrocarbon reduction simulations for 13 July 113
7-3 Total RHC concentrations used as initial conditions for
base case and hydrocarbon reduction simulations for 19 July 115
7-4 Reduced background concentrations used in the ozone
sensitivity simulations of 13 and 19 July 117
7-5 Hourly predicted maximum ozone (pphm) and ozone
reductions for the ozone sensitivity simulations
of 13 July 118
7-6 Hydrocarbon emission reductions required to meet the NAAQS
for ozone from the sensitivity simulations of 13 July 129
xiv
-------
7-7 Deficit/enhancement figures for ozone for the
13 July sensitivity simulations 131
7-8 Hourly predicted maximum ozone (pphm) and
ozone reductions for the ozone sensitivity
simulations of 19 July 145
7-9 Hydrocarbon emission reductions required to meet the NAAQS
for ozone from the sensitivity simulations of 19 July 156
7-10 Deficit/enhancement figures for ozone for the
19 July sensitivity simulations 157
xv
-------
1 INTRODUCTION
OVERVIEW OF STUDY
The U.S. Environmental Protection Agency's (EPA) Office of Air Quality
Planning and Standards (OAQPS) contracted with Systems Applications, Inc.
to carry out a photochemical air quality modeling evaluation study in the
Philadelphia Air Quality Control Region (AQCR). The goal of the study was
to assess the potential utility of the Urban Airshed Model (UAM) as an air
quality planning tool in a large metropolitan area that often receives a
considerable influx of ozone and precursors from other parts of the North-
east urban corridor. To achieve this goal, it was necessary to evaluate
the model's performance to ensure that it provides an adequate representa-
tion of the physical and chemical processes that influence ozone formation
in the Philadelphia atmosphere. Model performance evaluation is typically
carried out through comparison of predictions of hourly averaged ozone
concentrations with corresponding ozone measurements on one or more
historical episode days. This report discusses the results of simulations
of two summer ozone episodes—13 and 19 July 1979—in Philadelphia. The
13 July episode is characterized as a stagnation period, whereas the
19 July episode resulted in part from the transport of regional precursor
emissions from the New Jersey/New York urban area located northeast of
Philadelphia.
The requirements for the evaluation study were to
Acquire, format, and install on computer the 1979 Philadelphia air
quality, emissions, and meteorological data base;
Prepare UAM input files for the 13 and 19 July 1979 ozone episodes;
Evaluate the model's performance in estimating the magnitude of ozone
concentrations and temporal and spatial distributions;
Using base cases for both days, carry out a number of ozone sensi-
tivity simulations; and
Prepare a final report describing model inputs, model base-case
results, performance evaluation findings, and results of the ozone
sensitivity simulations
-------
This report is organized as follows: Section 2 provides a brief overview
of the characteristics of the 13 July 1979 ozone episode in Philadelphia;
Section 3 describes the procedures used to prepare each of the model input
files for this day. The characteristics of the 19 July 1979 episode are
described in Section 4; a description of the model input preparation for
this day is presented in Section 5. Section 6 contains an analysis of
model performance on 13 and 19 July 1979. Section 7 discusses several
hydrocarbon emission reduction scenarios and input sensitivity simula-
tions. Summary and conclusions are presented in Section 8.
Two appendixes are included with this report. Appendixes A and B present
predicted hourly average ozone isopleths and time series plots of ozone
predictions and observations at monitoring stations for the 13 and 19 July
1979 simulations, respectively.
TECHNICAL APPROACH
The photochemical modeling approach adopted in the Philadelphia study is
similar to, and draws heavily from, past UAM application studies in other
urban areas in the United States and abroad, including Los Angeles (Tesche
et al., 1982a,b), St. Louis (EPA, 1983a), Denver (EPA, 1983b), and Tulsa
(Reynolds, 1982). Although the Philadelphia application can be considered
straightforward in many respects, certain unique characteristics of the
region require special consideration (e.g., treatment of upwind boundary
conditions to reflect precursor emissions transported from the New
Jersey/New York urban area). The steps followed in the Philadelphia study
are summarized next.
The evaluation of the UAM in the Philadelphia metropolitan area comprised
the following technical steps:
(1) Selection, through consultation with the EPA project officer, of
the ozone modeling episodes to be simulated;
(2) Identification, receipt, and installation of meteorological, air
quality, emissions, and other geographic data;
(3) Performance of a quality assurance audit of the data base to
identify and subsequently correct errors and biases found in the
meteorological and air quality data bases;
(4) Specification of the modeling region;
(5) Preparation and review of UAM input files;
-------
(6) Initial UAM simulations and evaluations;
(7) Diagnostic analyses of simulation results in order to identify
and subsequently remove biases or errors in the model input
files;
(8) Final simulation of the ozone episodes to obtain suitable base
cases; and
(9) Evaluation of UAM performance following prescribed statistical
procedures.
After preparation and evaluation of the base-case simulations, a series of
hydrocarbon emission reduction simulations involving various assumptions
of background concentration for hydrocarbons and ozone were performed to
(1) assess the sensitivity of predicted ozone concentrations, and (2)
demonstrate the value of using the UAM in the formulation of emission con-
trol requirements for improving air quality in large metropolitan areas
such as Philadelphia.
Figure 1-1 presents the general geographical setting of the Philadelphia
airshed modeling region. Figure 1-2 shows a more resolved view of this
region. (Further details of the modeling region are given in Section 3.)
-------
PH1LBDELPMIR
RIRSHED
FIGURE 1-1. Geographical location of the Philadelphia airshed
modeling region.
-------
NORTH
0
10
20
30
20
10
0
I I I I i I I
-H—i—I—'—I—'—I—'—I
10 20 30 40 50
KILOMETERS
Philadolphial pp
13
BURLINGTON
22-
i i i i i i i i i I i i
i I i i i
30
20
CO
a:
LLJ
10
10
DOWNTOWN PHlLflDELPHIR RREfl
20
30
0
SOUTH
* STflTION
1
17
I-
0 5 10
KILOMETERS
1 RMS Lab.
2 Rncora
3 Bristol
4 Bnpantine
5 Cemden
6 Chester
7 Claymont
8 Conshohocken
9 Defense Suppcrt
10 Dowmngton
11 Franklin Inst itute
12 Island Rd. Ri-port Ci-c'e
13 Lunberton
14 Norristown Rrrcory
15 Northeast Rirprrt
16 Roxy Water Purap
17 SE Sewage Plant
18 South Broad
19 Sunmit Bridge
20 SW Corner Breed/Butler
21 Trenton
22 Van Hisevi I le
23 Vmeland
FIGURE 1-2. Philadelphia airshed modeling region. Philadelphia AQCR for which
emissions are available is shown by bold lines. Lighter lines denote county
boundaries. Stippling represents urban areas.
-------
CHARACTERIZATION OF THE 13 JULY 1979 OZONE EPISODE
The highest and most widespread ozone concentrations measured during the
summer of 1979 occurred on Friday, 13 July. These high concentrations
were the result of a buildup of precursors from near-stagnant conditions
on the previous day (12 July), and generally weak and variable winds pre-
vailing until noon of the 13th, when a stronger southerly flow was
established. The synoptic pattern showed high pressure throughout the day
dominating the surface and upper-level (2000 m) flow fields (see Figure
2-1). This high pressure was part of the seasonally semipermanent Bermuda
high-pressure cell centered east of Florida. The high pressure weakened
through the course of the day as a trough (the remnants of Hurricane Bob)
over the Ohio Valley moved slowly eastward bringing precipitation to
western Pennsylvania.
Light west-northwesterly surface winds were present throughout the region
from noon on 12 July until the early morning hours of 13 July when the
winds became calm or very light with a northerly direction. This wind
flow pattern was responsible for transporting ozone and precursors to the
southeast of the region during this period. Surface wind measurements
showed these calm-to-very-light winds during the early morning hours of 13
July, with northerly directions increasing slightly in speed and shifting
to a general southerly direction by late morning (1000 EST). This shift
in surface wind direction resulted in a recirculation of material that had
been transported to the southeast during the previous 24 hours. Influx of
ozone and precursor from the southeast, a buildup of regional ozone and
precursors from near-stagnant conditions on the previous day (12 July),
the day's emissions, generally light winds, high region-wide temperatures,
and mostly clear skies were the primary conditions leading to high ozone
concentrations.
Evidence of this recirculation of ozone and precursor material and the
possible existence of a reservoir of ozone aloft from the previous day is
shown in the observed ozone concentration data at upwind monitors. For
the morning of 13 July, stations to the south and east were upwind.
Examination of the ozone concentration during the initial onset of mixing
is one way of providing information on ozone levels aloft. At 1000 EST,
observed ozone concentrations were 8.4 pphm at Brigantine, New Jersey; 9.7
-------
i»
FIGURE 2-1. Synoptic situation, 0700 EST, July 13, 1979.
(Source: AT lard et al., 1981)
-------
pphm at Lumberton, New Jersey; and 12.3 pphm at Ancora, New Jersey. At
another upwind monitor (Vine!and, New Jersey), ozone data were missing for
1000 EST; however, an hour earlier (900 EST) this monitor recorded an
ozone concentration of 10.2 pphm. For this time period (onset of mixing),
these ozone concentrations were among the highest observed during the
summer of 1979. This indicates that a large reservoir of ozone existed
aloft and was available for mixing down to the surface.
No upper-air measurements were available in Philadelphia on this day;
however, radiosonde data from New York City (JFK) and Washington D.C.
(Dulles International Airport) showed very light westerly winds at 2000 m
throughout the day.
Because surface wind flow patterns established a southerly direction by
late morning, peak ozone concentrations occurred north of the high urban
emission source region. The highest ozone concentration recorded on this
day was a value of 20.5 pphm at Conshohocken at 1600 EST. The second
highest value of 20 pphm occurred earlier in the afternoon at the Roxy
Water Pump monitor at 1400 EST. Thirteen monitors recorded ozone concen-
trations greater than 12 pphm.
-------
DESCRIPTION OF MODEL INPUTS FOR 13 JULY 1979
This section describes the preparation of the UAM input files for the
13 July 1979 episode. The majority of the files were prepared after
examination of the air quality and meteorological data from the
Philadelphia Oxidant Data Enhancement Study carried out during the summer
of 1979 (Allard et al., 1981). The emissions and terrain files (estimates
of surface roughness and uptake) were prepared after examination of the
study detailing the preparation of the emission inventory for the
Philadelphia AQCR (EPA, 1982). All input files times were set to Eastern
Standard Time (EST). Input specifications for some of the files were
prepared either wholly or in part by members of the EPA project team at
OAQPS, whereas the computer simulations were all performed at Systems
Applications, Inc.
MODELING REGION SPECIFICATIONS
The modeling region of the Philadelphia airshed specified in this study is
180 x 170 km, or a total area of 30,600 km . This region covers parts
of Pennsylvania, Delaware, Maryland, and New Jersey, and includes the
metropolitan areas of Philadelphia, Pennsylvania; Wilmington, Delaware;
and Trenton, New Jersey (see Figures 1-1 and 1-2).
The modeling region specifications (e.g., grid origin, grid size, number
of horizontal cells, etc.) are contained in what is known as the REGION
packet. For the Philadelphia application, the modeling grid contains 36
by 34 horizontal cells that are fixed at 5000 m by 5000 m, along with
4 vertical cells (layers) that vary in thickness depending on the hourly
mixing height and the height of the top of the modeling region. Two
vertical cells are specified below the mixing height, with two cells
above. In the simulations, the cells, or layers, below the mixing height
may attain the minimum cell thickness specified (for this application,
50 m) during the night when the mixing height is at a minimum. As the
mixed layer thickens during the day, the vertical thickness of layers
1 and 2 grows, while the thickness of the upper layers (3 and 4)
decreases. The specified height for the top of the modeling region was
fixed for each hour and was used in preparing the REGIONTOP file. The
Airshed Model REGION packet in all input files was specified as follows:
11
-------
UTM ZONE: 18
X ORIGIN: 387000. m EASTING
Y ORIGIN: 4340000. ra NORTHING
GRID SIZE: 5000.0 m
TOP OF REGION: 1630.0 m
NX: 36 cells
NY: 34 cells
NZ: 4 cells
MINIMUM CELL THICKNESS-LOWER LAYERS: 50 m
MINIMUM CELL THICKNESS-UPPER LAYERS: 50 m
MIXING HEIGHTS
Mixing heights for 13 July 1979 were estimated every half hour for urban
and rural areas for daytime, transitional, and nighttime regimes. Urban
and rural mixing height cell designations are included in the discussion
of the terrain file. Since no soundings were available for Philadelphia
on this day, the daytime mixing heights between 0600 and 1400 EST were
estimated using the available morning (0700 EST) and evening (1900 EST)
radiosonde soundings in New York (JFK) and Washington, DC (IAD). These
temperature soundings are shown in Figures 3-1 through 3-4. The procedure
used to estimate daytime mixing heights is similar to the methodology used
by Holzworth (1972). First, spatially averaged surface temperatures for
the Philadelphia region were computed for each hour using all available
meteorological measurement sites. The resultant average hourly surface
temperatures were plotted on temperature-versus-height graphs of each
sounding. The height of the intersection between the dry adiabat of the
surface temperature and the temperature sounding was then taken as an
estimate of the mixing height.
Between 0600 and 1200 EST, the mixing height was computed as the average
of the values estimated from the JFK and Dulles airport morning sound-
ings. An average of the JFK and Dulles morning and afternoon soundings
was used to compute mixing heights between 1200 and 1400 EST. An excep-
tion to this procedure was made for estimating rural mixing heights
between 0600 and 0900 EST, before the dissipation of the rural surface
stable layer. During this period rural mixing heights were subjectively
estimated to increase from the overnight value of 100 m to the 0900 EST
value. Because of the general nature of the "adiabatic" procedure used
for estimating daytime mixing heights, no other distinction was made
between urban and rural mixing heights during the 0600-1400 EST time
period.
12
-------
JP
28ES-
18ES--
16BE-
15
35
13Z0-
see-
600
4120
20Z
I I I I I I I
15 20 25 3Z
TEMPERPTU*E
Figure 3-1. Temperature sounding for JFK airport on
13 July 1979 — 0700 EST.
13
-------
15 23 25 33 35
2400
2000 \
1800\
1600\
Sf 1400
1200
1000
800
600
200
15 23 25 30
TEMPERATURE 1C)
35
Figure 3-2. Temperature sounding for JFK airport on
13 July 1979 — 1900 EST.
14
-------
20
25 30 35
2200^
2000%
1800
1600
^ 1400
^ 1200
J£ 1000
600
600
400
200
0.
IS 20
25 30 35
TEMPERPTURE !C)
45
Figure 3-3. Temperature sounding for Dulles airport on
13 July 1979 — 0700 EST.
-------
10 15 2? 25 30 35
15
20 25 3C
TEMPERPTUPE fCJ
Figure 3-4. Temperature sounding for Dulles airport on
13 July 1979 -- 1900 EST.
16
-------
The overnight (0000 through 0600 EST) mechanically dominated mixing
heights were set to a constant value of 250 m in urban areas and 100 m
in rural areas. These "default" values were obtained from analyses of
overnight mixing heights in Philadelphia, St. Louis, and other cities
(Godowitch, 1984; Godowitch, et al., 1984b; Bornstein, 1968; Clarke,
1969).
Mixing heights during the evening transition between convectively domina-
ted and mechanically dominated mixing are difficult to estimate, particu-
larly in the absence of local measurements. To approximate this transi-
tion in a manner consistent with the capabilities of the Airshed Model
(and at the same time avoid sharp discontinuities in mixing height),
mixing heights were decreased at a rate of -2-3 m/min after the time of
maximum average surface temperature at 1400 EST (Noonkester, 1976; Kaimal,
et al., 1982). This rate was used for urban mixing heights until 2300 EST
when the overnight value of 250 m was assumed to be applicable. In rural
areas, this rate was applied until 1800 EST, at which time the mixing
height was decreased more rapidly to reach the overnight value of 100 m by
2000 EST (Godowitch, 1984b). The resulting half-hour mixing height values
for urban and rural grid cells are presented in Table 3-1 and graphically
illustrated in Figure 3-5.
WIND FIELD
Preparation of the wind field for a specified simulation day is one of the
more critical inputs in achieving acceptable model performance. There is
no "correct" way of specifying a three-dimensional flow field using only a
limited number of surface observations and even fewer upper air observa-
tions covering an area as large as the Philadelphia airshed region. Yet
the "modeled" flow field is the crucial element in the final spatial
alignment of the urban plume. In any Urban Airshed Model application,
certain areas of the airshed grid will lack surface data. These data
"gaps" must be filled in with realistic estimates so that a smooth con-
tinuity is maintained for the mass flow both in time and space. Because
of the light and variable nature of the observed surface winds on 13 July,
it was apparent that it might be difficult to prescribe a three-dimen-
sional flow field free from directional and speed biases. The following
procedure was used in preparing the hourly flow fields for the 13 July
1979 simulation.
Surface wind data collected during the 1979 Philadelphia Oxidant Data
Enhancement Study (Allard et al., 1981) were available for 16 stations
for 13 July. The first step in the preparation of the wind field was to
graphically plot the surface vectors on the airshed grid. Figure 3-6
depicts these plotted vectors for four separate hours of the day. This
17
-------
TABLE 3-1. Urban and rural
mixing height values used
in the DIFFBREAK file for
13 July 1979.
Time
(EST)
0000
0030
0100
0130
0200
0230
0300
0330
0400
0430
0500
0530
0600
0630
0700
0730
0800
0830
0900
0930
1000
1030
1100
1130
1200
1230
1300
1330
1400
1430
1500
1530
1600
1630
1700
1730
1800
1830
1900
1930
2000
2030
2100
2130
2200
2230
2300
2330
2400
Urban
(m)
250
250
250
250
250
250
250
250
250
250
250
250
250
270
295
375
450
680
925
1160
1200
1330
1480
1480
1480
1500
1530
1530
1530
1475
1420
1365
1310
1220
1130
1035
950
855
770
675
590
500
410
370
330
290
250
250
250
Rural
(m)
100
100
100
100
100
100
100
100
100
100
100
100
100
135
150
250
350
620
925
1160
1200
1330
1480
1480
1480
1500
1530
1530
1530
1475
1420
1365
1310
1220
1130
1035
950
730
525
320
100
100
100
100
100
100
100
100
100
18
-------
Legend:
^_ Urban
... Rura I
12
Time (hours)
FIGURE 3- 5. Mixing height profile for urban and
rural cells for the 13 July 1979 simulation.
19
-------
FIGURE 3-6. Observed surface wind vectors and
wind vectors used in preparing the interpolated
surface wind field for 13 July 1979.
20
-------
I I
WIND SPEED (M/S!
448B —
- 4448 -
i 4428 -
4480 —
4360 -
4368-
1 1 i 1 1 1 1 1 1 1 1 1 1
[
-
-
DOWN
1 1 | 1 1 1 | i i 1 | 1 1 1 | 1 1 ! | i 1
TR^N
H&L
NORR
*°RTBR*1S VRN~"I
' MCGJ LHKE ~
jt RLLE '
LUMB
CHES
HILM
SU1M
HILL
I I I 1 I I I I I I I I I I I I I 1 I I I I 1 I
418 432 450 478 496 518 530 550
ERSTINC IKM1
Observed wind vectors
448C
44BE
442
44ZZ
438B -
f
TREN
f
HILL
NORR
DOHN
NORT
' RLLE
MCC-U LflrE
LUM3
CHES
HILM
SUM1
SE
SH
KILL
407 427 447 467 46? 507 5£7 547
ERSTIKG (KM)
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 3-6a. 0000-0100 EST
21
-------
I I I 1 I I
0 s
WIND SPEED (M/S)
4460
4468
- 4440
4420
4400
4380
4362
NORR
-At
NORT
DOMN
TREN
BRIS
RLLE
LUH8
O
MCGU
VflN
LBKE\
PHIL
CMES
SUWW
HILL
410 430 450 470 490 SIB 530 552
ERSTING (KM)
Observed wind vectors
448Z -
4468 -
= 4440 -
i
4420 -
44BB
4380
4360
1 ' ' I
Hri
i
NE
I
TREN
NORR
HILL
DOKN
/ VPN
HOST - »- \
KCG'J LRK5,
RLLE »-
PHIL
CHES
HILM
SUKH
SE
SH
V ^
MILL
i I i i i i i t i I i i t I i i i I i i i I i i i 1 i i t I i i i
4B7 427 447 467 467 537 527 5<7
EBSTIHG (KM)
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 3-6b. 0600-0700 EST
22
-------
B 5
HIND SPEED (M/SJ
4482
4460
- 4440
4420
4400
4360
4360
I ' I i ] i i i I
7REN
HILL \
NORR \ BRIS
DOWN
BLLE
LUMS
GU /
PHIL
CMES
H1LM
SUMM
MILL
410 430 450 470 490 SIB S30 550
EflSTING IK«)
Observed wind vectors
446!
445!
: 4*28
4380
43E2
\
\
NH
OCK'N
\
5". I.
RT \
-
Lfi-.i -
LUHB
*\ PHIL
CHES
NILM
MCCU
\
%.
SUMM
SE
SM
HILL
4B7 427 447 467 467 507 527 547
EB5TING (Ml
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 3-6c. 1200-1300 EST
23
-------
HIND SPEED IM/SI
4486
4460
- 4440
x 4420
44B0
4380
4360
TREN
WILL
NORR BRIS
/ JiDSJ—
DOWN
BLLE
MCGU
V*N
LPKE
LUHB
CHES
HILM
f
SUM*
MILL
41B 438 458 47a 490 510 530
ERSTING (KM)
Observed wind vectors
550
44SC —
446C —
4'.22
44ZI
436?
DCWN
-I
NORR
J HLl
///
CHES
HLLE KCC-U
LUK.3
SUMH
SE
MILL
I I 1 I I I ! I I I I I I I I I I >!! I I I I I t I 1 I
4Z7 427 447 467 467 507 527 547
ERSTINS 1KMI
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 3-6d. 1800-1900 EST
24
-------
graphical plotting is performed as a data screening procedure. By examin-
ing the hourly observed vectors plotted on the airshed grid, it is pos-
sible to gain insight into the actual flow field one is attempting to
model.
The plotting of surface vectors can uncover questionable data (e.g., a
surface vector that is 180° out of phase with the other nearby sta-
tions). It can also be used to identify areas of the airshed grid for
which observational data are sparse; these areas may need to be supple-
mented with estimated data to complete the spatial coverage of the mass
flow field.
Observed surface wind data were available at the following locations:
Station
Northeast Philadelphia Airport, Pennsylvania
Philadelphia International Airport, Pennsylvania
McGuire Air Force Base, New Jersey
Willow Grove Naval Air Station, Pennsylvania
Lakehurst Naval Air Station, New Jersey
Trenton-Mercer County Airport, New Jersey
Millville Airport, New Jersey
Greater Wilmington Airport, Delaware
Chester, Pennsylvania
Bristol, Pennsylvania
Al1egheny-Phi1adel phi a, Pennsylvani a
Summit Bridge, Delaware
Downington, Pennsylvania
Lumberton, New Jersey
Van Hiseville, New Jersey
Norristown, Pennsylvania
NW
NE
SE
SW
UTM Easting
(km)
499.0
478.6
534.1
488.7
556.9
515.6
494.3
448.6
467.8
510.0
491.7
441.0
436.0
518.0
559.0
473.5
430.0
550.0
535.0
410.0
UTM Northing
(km)
4436.0
4414.6
4431.3
4449.8
4431.5
4459.0
4357.0
4390.7
4409.3
4439.5
4425.2
4376.0
4426.0
4423.0
4439.5
4440.0
4480.0
4490.0
4380.0
4360.0
The last four stations in this list were the designated "pseudo" stations
used to fill the gaps in the observed surface field before the interpola-
tion procedure. Each hourly plot of observed vectors was subjectively
reviewed, and certain inconsistent or questionable data points were
removed from the data file before proceeding. Overall, the wind data
collected during the field program (for those days reviewed) looked
25
-------
reasonable, and only minor corrections or deletions were needed before
wind field preparation could commence. Figure 3-6 presents observed wind
vectors for 13 July and the set of vectors used in the wind interpolation
scheme for the corresponding time period. Comparison of the sets of
vectors indicates those vectors deemed unreasonable and data used to fill
the gaps in the interpolation.
The next step in the wind field preparation involved an interpolation
program to obtain a complete surface wind field. The program WINTER
employed a 1/R interpolation scheme and used the screened observation
file to create a gridded surface wind field. Preparation of the winds
in the upper layers was the next step in the process.
Although they were performed on other days during the 1979 Philadelphia
Oxidant Study, upper-level atmospheric measurements were not performed on
13 July in central Philadelphia. Consequently, the nearest upper-level
radiosonde data for this day correspond to the twice-daily releases at JFK
and Dulles airports. These data were carefully analyzed to establish
their utility for prescribing upper-level wind and temperature fields over
Philadelphia. Figures 3-7 through 3-10 present vector plots of the
morning and afternoon radiosonde wind profiles at JFK and Dulles airports,
respectively. These data confirm the existence of a southwesterly through
northwesterly synoptic flow, which is apparent from an examination of the
NWS daily weather map for this day (see Figure 2-1). On the basis of
these four soundings, the 1900 EST sounding of 12 July, and the 0700 EST
sounding of 14 July, an attempt was made to estimate average synoptic-
level (i.e., =1500 m) wind velocities at JFK and Dulles for the entire
day. This was achieved by averaging the radiosonde values over a range of
1200 to 1800 m for each sounding. Hourly estimates of winds aloft were
derived using vector interpolation between soundings. This review of both
sets of soundings resulted in the selection of a constant vector to be
used for each hour to describe the level-4 winds. The spatially constant,
temporally varying wind vectors used for level 4 are presented in Table
3-2.
To obtain wind vectors for levels 2 and 3, the program WINDCHANG was
used. This program performed a linear vector interpolation to obtain the
vectors for levels 2 and 3 using the surface vector and the level-4 vector
for each grid cell. The program read the REGIONTOP and DIFFBREAK mixing
height files to determine the height of the node for each level for each
hour. These files were read because the mixing height changes hourly and
determines the thickness of each vertical cell. After the node heights
were determined for levels 2 and 3, a linear interpolation was performed
to compute the mid-level direction. The speed of the level 2 or level 3
vector was then increased by employing a power law equation as follows:
26
-------
Delta H = 250 Meters
Delta S * 2 M/S
Figure 3-7. Upper-level winds at JFK airport on
13 July 1979 -- 0700 EST.
27
-------
Delta H = 250 Meters
Delta S = 2 M/S
Figure 3-8. Upper-level winds at JFK airport on
13 July 1979 — 1900 EST.
28
-------
Delta H = 250 Meters
Delta S = 2 M/S
Figure 3-9. Upper-level winds at Dulles airport on
13 July 1979 « 0700 EST.
29
-------
Delta H = 250 Meters
Delta S = 2 M/S
MOUTH
Figure 3-10. Upper-level winds at Dulles airport on
13 July 1979 -- 1900 EST.
30
-------
TABLE 3-2. Spatially constant
wind vectors for level 4 winds
for 13 July 1979.
Hour
(EST)
0000-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
Di rection
(Degrees)
284
282
280
278
276
274
272
270
270
270
270
268
268
268
268
268
265
265
265
250
250
240
240
240
Speed
(m/s)
5.5
5.5
5.2
5.0
4.8
4.4
4.2
4.0
4.0
4.0
4.0
4.0
4.0
3.5
3.5
3.0
3.0
3.0
3.0
4.0
4.0
4.0
4.0
4.0
31
-------
V2 = (V4 " v1)*(NODE2/150°)a + vi
V3 = (V4 - V1)*(NODE3/1500)a + Vj
where
NODE2 = the height of the node of level 2,
NODE3 = the height of the node of level 3,
V* = the speed of the vector for level 4,
Vj = the speed of the vector of the surface wind, and
a = 0.2
This method is only one means of interpolating for the winds in levels
2 and 3. In the equation, a is the surface friction coefficient that
varies as underlying features on the surface vary. If the a coefficient
had been dropped, a straightforward linear interpolation would have
resulted from the equation; however, the magnitude of the winds in levels
2 and 3 would have been underestimated with this procedure. Only one
value of a was used in this application to reflect an average value of
surface friction throughout the region.
A full three-dimensional flow field is thus created for each hour. Either
because of noise in the wind sampling network, or because of the nature of
the vectors sampled at two distinct points, the resulting flow field will
contain a great deal of divergence. If the field is employed by the Urban
Airshed Model with the divergence left in, spurious artificial vertical
motions will result in unrealistic vertical transport. The last phase of
the wind field preparation involves the elimination of nearly all diver-
gence.
The program DIVFREE was written to read the three-dimensional file and
eliminate most of the divergence in the flow field. This program was
adopted from an EPA algorithm developed by Clark and Eskridge (1977) and
is the final step in the creation of the wind field. As a quality control
check, the wind vectors are then plotted on the airshed grid and examined
for reasonableness before use in the UAM simulation. Trajectory analysis
can be performed using this final file to determine, for example, the
origin both in space and time of the parcels affecting the peak ozone
concentrations. Figure 3-11 presents the gridded, smoothed surface vec-
tors for the same hours presented in Figure 3-6. The gridded plots of the
32
-------
j 1
HIND SPEED (M/S)
15 20
(a) 0 - 100 EST
35
FIGURE 3-11. Airshed model surface winds for 13 July 1979.
33
-------
I
i I
0
10
15
HIND SPEED (H/S)
20 25
30
30
i t t 4 4 i 4 i i i
15
10
*-* 4 i i
m M « * M M * * * * M i H V ' ' * * * *\ \
^*;***^iVV^x^x\jM4(v * * * * * x \ \
J44***^5i\\\s*^>»\\\V4^V\^ * « * * * x x ^>
4 i 4 t * * > \ x x v^
-------
-30
0
15
(c) 1200 -
20
1300 EST
25
30
35
FIGURE 3-ll(continued)
35
-------
30
0
10 15 20 25
(d) 1800 - 1900 EST
30
35
FIGURE 3-11(concluded)
36
-------
upper-level winds are not presented here, but are routinely examined for
consistency before a UAM simulation is carried out.
BACKGROUND CONCENTRATIONS
Estimates of background concentrations at the top of the modeling region
(for TOPCONC file), initial concentrations aloft (above mixing height-
levels 3 and 4), and concentrations for all boundaries except those below
the mixing height (levels 1 and 2 along the Southeast boundary) are
presented in Table 3-3. Concentration values of NO, N02, CO, and
hydrocarbons were specified on the basis of work by Killus (1982) to
reflect concentrations of a typical urban atmosphere. Hydrocarbon
speciation is that used in the Carbon-Bond Mechanism (Killus and Whitten,
1981). To represent an aged polluted atmosphere with a reservoir of ozone
aloft, a value of 0.08 ppm for ozone was specified. This value was
obtained from the ozone measurement at upwind monitoring sites during mid-
morning when the surface-based inversion was dissipating. These stations
included Brigantine, Lumberton, and Ancora, New Jersey. A background
concentration of 0.000025 ppm peroxyacetylnitrate (PAN) was specified
following the work performed by Singh and Hanst (1981).
INITIAL CONDITIONS
Initial conditions for all species were specified using all available
monitoring data in the Philadelphia region. The simulation of 13 July
commenced at midnight, requiring concentration values corresponding to
this time. Rather than using the hourly averaged observed value for each
species for the midnight hour, initial condition concentrations were
calculated for each monitor by averaging the observed hourly average
pollutant concentrations from 2300-2400 EST on 12 July with the 0000-0100
EST concentration on 13 July to obtain a two-hour average value for mid-
night. Initial condition concentrations used in the preparation of the
AIRQUALITY file for 13 July are presented in Table 3-4 for 23 monitoring
sites.
Because no carbon monoxide (CO) data were available at rural monitors
on 13 July, background CO concentrations of 0.20 ppm were input at the
Downington, Summit Bridge, and Van Hiseville monitors. Also, a CO concen-
tration of 1.6 ppm at Ancora was eliminated because it was not considered
representative of an outlying rural area, having possibly been the result
of some local pollutant source. Also included in the table are the emis-
sion inventory reactive hydrocarbon (RHC) split factors used to compute
initial concentrations of single-bond carbons (PAR), double-bond carbon
groups (OLE), ethylene (ETH), aromatic rings (ARO), and carbonyl groups
37
-------
TABLE 3-3. Background concentration values
for 13 July at the top of the modeling region
(TOPCONC), as initial concentrations above the
mixing height, and for all levels of all
boundaries except the levels below the mixing
height on the Southeast boundary.
Concentration
Species
NO 0.001
N02 0.002
03 0.08
CO 0.2
ETH 0.001
OLE 0.0004
PAR 0.040
CARB 0.010
ARO 0.0008
PAN 0.000025
BZA 0.00001
38
-------
TABLE 3-4. Initial conditions for 13 July 1979 (concentrations in ppm),
Station
AMS Lab.
Ancora
Brigantine
Bristol
Camden
Chester
Claymont
Conshohocken
Defense Support
Downington
Franklin Inst.
Island Rd. Airp. Cir.
Lumberton
Norristown Armory
Northeast Airp.
Roxy Water Pump
SE Sewage Plant
South Broad
Summit Bridge
SW Corner Broad/Butler
Trenton
Van Hi Seville
Vine! and
Easting
(m)
491600
511800
546000
511000
491700
469000
461500
474500
483800
436000
485200
480300
518000
473500
499000
479500
487200
486100
441000
487000
520000
559000
498200
Northing
(m)
4428500
4392400
4377506
4440000
4419000
4410000
4406400
4435600
4418300
4426000
4422800
4414800
4423000
4440000
4436000
4433100
4417300
4421600
4376000
4428000
4452000
4439500
4371200
NO
0.
—
—
--
0.
—
—
—
—
0.
0.
—
0.
—
0.
—
—
0.
0.
—
—
0.
—
005
022
002
010
006
035
020
000
001
N02 CO
0.
—
—
—
0.
0.
0.
—
—
0.
0.
—
0.
0.
0.
—
—
0.
0.
—
—
0.
0.
075 1.5
—
—
—
105 2.4
030 --
045 —
—
--
012 0.2
065 —
—
031 1.2
075 —
060 —
—
--
080 3.5
003 0.2
--
—
006 0.2
024 —
0
0
0
0
0
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
°3
.035
.051
.037
.000
.004
.058
-
.000
.020
.073
.035
.010
.010
.000
.000
.045
.010
.035
.046
.035
.009
.005
.054
RHC
1.05
—
—
--
—
2.10
—
—
—
0.00
—
—
0.20
0.39
—
--
—
0.30
0.05
—
—
—
—
Carbon-Bond Fraction
RHC Component (% as Carbon)
PAR
OLE
ETH
ARO
CARB
74.
2.
4.
13.
5.
0
8
1
2
9
39
-------
(CARB) as ppmC according to the Carbon-Bond Mechanism formulation. The
observed total RHC observation value for each hour is multiplied by each
splitting factor to obtain the speciated reactive hydrocarbon component as
ppmC.
Initial conditions in surface grid cells without monitors were obtained by
employing a Poisson interpolation. After this surface field was computed,
a vertical interpolation method was chosen such that the background con-
centration at the top of the modeling region (TOPCONC) was used in levels
3 and 4, and the level 2 value was obtained by a linear interpolation
between the surface (level 1) value and the level-3 value. Using this
method, all grid cells in all levels were initialized with appropriate
concentrations for all species before commencement of the simulation.
BOUNDARY CONDITIONS
The physical boundaries used in the 13 July simulation are presented in
Figure 3-12. Because of the stagnation characteristics of this episode,
much of the large airshed region was "blocked off" and not included in the
simulation. No grid cells containing major emissions sources were
excluded by this procedure. The winds were light and variable throughout
much of the simulation day, so it was determined that a large portion of
the grid need not be modeled. This procedure cut simulation costs, but
did not hinder the analysis of the core of the urban ozone plume located
to the north of the urban center.
Background values for all species were designated for all boundaries
except the Southeast boundary below the mixing height. Because the Phila-
delphia region was dominated by a large, slow-moving high-pressure system
on the days preceding the 13 July episode, stagnation and recirculation of
urban oxidant precursors occurred. The wind flow on 12 July was light and
westerly, with northerly flow established early on the 13th. By mid-
morning (1000 EST), the flow had veered from a northerly to southeasterly
direction, and continued south-southeasterly throughout the rest of the
day. It is believed that the urban plume from Philadelphia was trans-
ported to the east late on 12 July and recirculated back toward the Phila-
delphia area by the moderate southeasterly flow on the 13th. This
hypothesis is substantiated by an unusally high ozone reading of 13.9 pphm
at 1200 EST at the Vine!and monitor located 50 km to the southeast of
Philadelphia. This advection of oxidant precursors from the previous
day's emissions, along with the daily emissions and meteorological condi-
tions conducive to ozone formation (e.g., lack of ventilation), led to
high ozone levels on 13 July. Because the flow field for most of the
daylight hours exhibited a southerly component, a different set of concen-
trations for the Southeast boundary grid cells below the mixing height was
40
-------
NORTH
10
30
30
0
NORR
VfiM
JXY
SHE&S
LUM8
ISL
cues
CLRY
RHCO
SUMH
;ii'-;X# x:it * :jf'x':j':-:^i :'':::jj:-:-:j:':-'::i:'':'::t': '::<:''::::i:: -':>: ' ^ '••:'''+ -':|: '':':tx ':i::':::Y'- '•>"'' '> '.- ''•[' ' t ' t; -:i '• ± }
30
SOUTH
FIGURE 3-12. Physical boundaries used in the 13 July 1979 Airshed Model
simulation (shaded area is not included in the simulation).
41
-------
made to duplicate the inflow of aged air parcels containing ozone and
precursors from the previous day's emissions.
For the Southeast boundary, the ozone concentrations on 13 July at the
Vineland monitor, located on the designated Southeast boundary of the
modeling region (see Figure 3-12), were used to obtain temporally varying
ozone concentrations for the two vertical cells below the mixing height.
Because of recirculation through the Southeast boundary, early morning
(0000-0700 EST) NO, N02, and PAN concentrations for grid cells below the
mixing height were obtained from late-afternoon measurements of these
species at the Downington, Pennsylvania and Van Hiseville, New Jersey
monitors when these sites were being influenced by the urban plumes. An
early morning (0000-0700 EST) nonmethane hydrocarbon (NMHC) concentration
of 0.15 ppmC was specified for cells below the mixing height, based on an
analysis of data from other Northeast urban corridor cities under similar
meteorological conditions (stagnation and carryover). The hydrocarbon
species split of NMHC was chosen to represent an aged urban air mass as
recommended by Kill us (1982.).
The early morning (0000-0700) NO, N02» PAN, and other speciated NMHC con-
centrations were input to the CBM-III mechanism of the EKMA trajectory
model, OZIPM (Whitten and Hogo, 1978) to simulate daytime (0700-1700)
variation in these pollutant species due to chemistry. The resultant
values of these pollutant species were used for temporally varying NO,
N02, PAN, and other NMHC species concentrations for grid cells below the
mixing height along the Southeast boundary, except that background
concentrations were substituted for resultant concentrations below
background values. The OZIPM kinetics model simulation produced a similar
amount of ozone as recorded at the Vineland monitor, thus giving credence
to these boundary condition species. The 1700 EST pollutant species
concentrations were extended until 2400 EST. The concentration values
used for the Southeast boundary below the mixing height for all species
are given in Table 3-5.
EMISSION INVENTORY
The gridded minor point, area source, mobile source, and elevated point
source emission inventory was prepared for the EPA in 1981 by Engineering
Science, Inc. (EPA, 1982). (The boundary of the gridded inventory is
shown in Figure 1-2.) Hourly (local daylight time) emission values for
total NOX and total hydrocarbon are presented in Table 3-6. Total daily
emissions by source type are presented in Table 3-7 for NO, N02, and
hydrocarbons. This table also gives the average hydrocarbon split for the
entire emission inventory. The gridded spatial distribution of surface-
layer hydrocarbon emissions is presented in Figure 3-13. The gridded
spatial distribution of surface-layer NOX emissions is presented in Figure
3-14; the bold line defines the area for which gridded emission estimates
42
-------
en
c •
X E
•«- Q.
E CL
(J^ ^
JC *^
0 0
'a) Ij
.a ITS
10 4->
.— C
i— ai
QJ O
0 C
o
t- 0
O »•— '
C CT^
O r-H
•a "3
C "3
o
o ro
«-H
r—
10 3
(0 E
JC 10
o -c
c_
• 0
LT> 14-
1
LU CT>
_J •!-
CQ ai
tfsiOLOOooooooooo
oooooooooooooooooo
oooooooooooooooooo
OOOOOOOOOOOOOOOOOO
oooooooooooooooooo
COr— ir~VOOOO<— 'lOCVJCSJi— ii-HCT^COLOCOCTiCO
LO ^f CO CO CNJ t— * CvJ CO f*^ ^D »™^ CVJ CO t-H i_? O^ ^O tO
OOOOOOOOO^Ht-lr-lt-H^Hr-lOOO
oooooooooooooooooo
oooooooooooooooooo
oooooooooooooooooo
oooooooooooooooooo
rH t— 1 f-H f— 1 t-H ff"^ r— t f»H r-H f— ( t— 1 t— 4 t— 1 t— 1 f— 1 f-H t-H t-H
oooooooooooooooooo
oooooooooooooooooo
OOOOOOOOOOOOOOOOOO
oooooooooooooooooo
oooooooooooooooooo
OOOOOOOOOt— 1 i— It— It— It— 1 i— 1 ,— 1 r-l r- 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
oooooooooooooooooo
oooooooooooooooooo
OOOOOOOOOOt-ii— it— i,_it— it— it— it— i
43
-------
•
•o
O)
•o
3
r—
U
0
o
*—*
ID
CO
1 1 1
i
CO
OJ (.
E
o
o
tr>
vo
o
o
CO
c
o
o
r— 1
o
o
o
o
o
o
CM
1
o
o
en
i— i
e\j
co
o
o
o
0
o
o
CO
o
0
o
c
o
0
0
0
o
c
o
l£>
o
ID
o
o
cr>
ID
O
o
«M
O
0
o
I-H
o
o
o
0
o
I-H
CM
1
O
o
o
«VJ
C\J
co
o
o
0
o
o
o
00
o
c
c
o
o
o
o
o
o
o
o
VO
o
ID
o
o
vo
ID
o
c
CsJ
o
o
o
1— 1
o
o
c
o
o
evj
CVJ
1
o
o
1— 1
CM
CM
co
o
o
o
o
o
0
CO
o
o
o
o
o
o
o
o
o
o
o
VO
o
ID
o
o
t— 1
ID
o
o
CM
o
o
0
t— 1
o
o
o
o
o
CO
CM
1
o
o
CM
CM
CM
CO
o
o
o
o
o
o
CO
o
o
o
0
o
0
0
o
o
o
o
VO
o
ID
o
0
o
ID
0
c
CM
o
o
o
1— 1
o
o
o
o
o
«3-
CM
1
o
o
co
CM
44
-------
TABLE 3-6. Hourly emissions of NOX and hydrocarbon (tons/hr)
used in the 1979 Philadelphia emission inventory.
Hour
0000
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
Total
Total
(LOT)
- 0100
- 0200
- 0300
- 0400
- 0500
- 0600
- 0700
- 0800
- 0900
- 1000
- 1100
- 1200
- 1300
- 1400
- 1500
- 1600
- 1700
- 1800
- 1900
- 2000
- 2100
- 2200
- 2300
- 2400
(tons/day)
(tons/year)
Total
NOX %
23.
21.
19.
19.
20.
24.
33.
53.
53.
49.
50.
51.
46.
50.
51.
54.
58.
52.
42.
35.
32.
30.
29.
26.
934.
341,172.
413
070
886
712
461
625
658
234
623
794
629
720
660
451
662
883
815
584
650
866
448
728
480
664
718
000
2.
2.
2.
2.
2.
2.
3.
5.
5.
5.
5.
5.
4.
5.
5.
5.
6.
5.
4.
3.
3.
3.
3.
2.
100.
50
25
13
11
19
63
60
70
74
33
42
53
99
40
53
87
29
63
56
84
47
29
15
85
0
Total
HC %
17.
15.
12.
11.
11.
14.
25.
81.
94.
89.
89.
84.
77.
75.
75.
81.
74.
67.
56.
39.
35.
33.
31.
22.
1217.
444,532.
255
090
413
777
444
422
677
214
468
978
809
016
34
132
917
869
434
097
362
228
537
102
363
956
897
000
1.
1.
1.
0.
0.
1.
2.
6.
7.
7.
7.
6.
6.
6.
6.
6.
6.
5.
4.
3.
2.
2.
2.
1.
100.
42
24
02
97
94
18
11
67
76
39
37
90
35
17
23
72
11
51
63
22
92
72
58
88
0
45
-------
0
f
>.
t.
o
-*->
c
eu
c
•f—
(O
•r—
x:
o.
F^
Ol
•o
£
•fBI*
f*
Cv
cn
cn
^H
^
-C
c
•r-
^— %
co
i
o>
^^
Ol
o.
5^
^j
Ol
U
3
O
VI
^.
J2
VI
C
0
°«/j
CO
(r_
&
dl
>^
_„
•S
*
O
z
OJ
u
3
O
CO
vo
A
cn
co
1-H
LO
vo
A
o
f*^
in
CO
CM
CO
A
1-H
o
CM
Cn
t*^
A
f*^.
^J.
^^
1— I
A
CM
i— l
VO
CO
o
A
f*^
in
CM
m
in
cn
A
00
CO
f**,.
A
in
c
•P—
O
Q.
•o
tu
(O
>
0)
LU
cn
O^
p*^.
A
O
t— i
LO
o
CO
CO
A
CO
1-H
CO
cn
CM
r^~
A
VO
^f
00
A
in
i— i
CM
CO
VO
A
CO
00
i— l
^r
cc
CM
A
in
oo
co
f-
co
00
A
^*
^H
i-H
0»
r*.
UD
A
r-4
^*
c
A
CM
*->
C
•t"
O
CL
t.
O
c
•^
S.
CO
CO
00
A
1-H
cn
LO
in
in
A
cn
vo
vo
A
t— 1
co
cn
vo
A
1-H
*f
00
A
vo
CM
1—4
r*^
^j-
A
^
cn
1-H
00
in
CM
A
in
cn
CM
cc
o
CO
A
1 —
o
CM
,^
er>
o
A
in
o
A
CM
1-H
A
vo
CM
CM
A
m
in
CO
in
in
A
^
VO
VO
LO
in
CO
A
f1^
CM
CO
CO
^>
A
0
in
o
A
1-H
f— <
*oh
o
A
0
CM
LO
A
1-H
VO
en
^«
A
CO
CO
r1^
f^
CM
CM
A
CO
1-H
CO
A
1-H
00
CO
\G
A
*-O
<3-
VO
A
vo
CD
*— i
«-j-
A
CM
CO
r- 1
A
5
o>
u
^
3
O
co
Ol
^~ I* ™
*r™ fQ
-Q j->
0 . 0
vo
vo
A
O
LO
^J-
A
OO
1-H
II
X
0
ID
O
1—
f-H
O
f-H
A
CM
CO
CO
A
CO
vo
II
0
^d
r"—
IT3
^j
£
Q.
UO
&
m
O)
T3
u c
TO O
(- -Q
Li_ (-
(D
•O O
C
O co
CO
co m
u
o.
to
CQ
c on
cz. <:
<: o
46
-------
S S3
m ,_
UJ
o
CM ^
3C
SI
is
(M
IS!
IS
ts
(3
CS
!Sb3
•a
ai
to
0}
CO
C\J
I I i I I i T I I I I I I
I 1 I I I I I I j I l 1 I I I
Q.
O)
re
IS
CO
X
I—
o
SI
CVJ
a.
a)
IS
CO
S3
Cxi
O
CO
c
o
I/)
to
d)
c
o
(J
O
T3
O
+•>
i.
•u
re
01
' o.
CO M-
I I I I I I
i i i
I I I I I
I I I I I
SI
CO
IS
CNJ
SI
OL O
C3 a>
>-i 0)
u. s_
IS
1S3M
47
-------
cn
z
LU CO
O CD
O LJ
cn
x
CE O
UJ Z
ct:
CE _1
CE
>~
CE
Q
-------
were available for the 1979 inventory (the Philadelphia AQCR). Emission
estimates contained in the inventory and used in the simulation represent
a typical summer weekday in 1979.
METSCALARS
The term "metscalars" refers to those model inputs that are considered
spatially invariant across the modeling region. Metscalars include the
region-wide average temperature gradient below the mixing height, the
average temperature gradient above the mixing height, the exposure class
(related to the degree of boundary-layer thermal stratification), and the
radiation factor (related to the NC^ photolysis rate).
The hourly temperature gradient above and below the inversion was speci-
fied after examining the available temperature soundings for 13 July
1979. Radiosonde data from JFK and Dulles airports at 1900 EST, 12 July;
0700 and 1900 EST, 13 July; and 0700, 14 July were used. To obtain the
temperature gradients, a simple procedure was followed. The vertical
temperature profile was plotted graphically (temperature versus height)
for each sounding. Using the hourly specified mixing height, an average
temperature gradient was subjectively drawn from the surface to the top of
the mixed layer. This average temperature gradient was also obtained for
the layer above the mixing height to the top of the modeling region using
the same procedure. The temperature gradients thus obtained were the
result of examining data from multiple soundings near the region.
Specific hourly values for the temperature gradients above and below the
mixing height were obtained by averaging the two sets of sounding data.
For the temperature gradient above, the data showed a nearly neutral atmo-
sphere for the entire day with a value of -0.006°K/m specified for all
hours. The data showed a nearly adiabatic atmosphere below the mixing
height for the entire day, with a temperature gradient of -0.01°K/m
specified for the daylight hours. For nighttime hours, a value of 0.0°K/m
was specified to reflect a slightly stable rather than adiabatic atmo-
sphere.
As detailed in Ames et al. (1978), the exposure class is a measure of near
ground-level stability due to surface heating or cooling and can be esti-
mated from insolation as follows:
exposure
class
3
2
1
0
-1
-2
strong \
moderate)
slight )
heavy overcast
> -K-
cloud cover
, < -5 cloud cover
- o
daytime insolation
day or night ,
nighttime cloudiness.
49
-------
The exposure class categories chosen for each hour were specified after
reviewing the available meteorological data including surface tempera-
tures, solar radiation data, and synoptic weather summaries for 13 July.
A computer program developed by Schere and Demerjian (1977) was used to
calculate layer-averaged N02 photolysis rate constants based on month,
day, year, latitude, longitude, time zone, time of day, mixing height, and
measured solar radiation data. The first seven parameters were used to
calculate zenith angles and corresponding clear sky theoretical surface
N02 photolysis rate constants. The mixing heights, along with the
parameters used to calculate clear sky theoretical surface N02 photolysis
rate constants, were used in the calculation of clear sky theoretical
layer-averaged N02 photolysis rate constants. The methodology used to
determine the clear sky theoretical surface and layer-averaged N02
photolysis rate constants is described by Demerjian et al. (1980). The
total measured solar radiation data in langley/min was multiplied by the
constant 0.40 min"1 min/langley derived by Jeffries et al. (1982) to
convert the solar radiation data to empirical surface N02 photolysis rate
constants (min"1). The empirical layer-averaged N02 photolysis rate
constant was calculated by the following equation:
Empirical surface
p • • i i N°, Photolysis , ,
Empirical layer- 2 / theoretical layer-
., ..« u ,. -, • pate constant . ...
averaged NO- photolysis averaged NO
photolysi s
rate constant Clear sky theoretical surface __ ^ ^
._ , . rate constant
N00 photolysis
rate constant
Because dewpoint values were in the upper 60s to lower 70s throughout the
region for the entire day, a constant of 24000 ppm was designated for the
concentration of water. Atmospheric pressure throughout the region ranged
from 1010 to 1013 mb as measured by the surface stations of Philadelphia
(PHL) and radiosonde data from JFK and Dulles airports. On the basis of
these measurements, a constant value of 1.0 atm was specified for atmo-
spheric pressure for all hours of the simulation day. The complete set of
inputs used in the METSCALARS file for 13 July 1979 is contained in Table
3-8.
50
-------
c
o
-u
(O
*3
E
(/)
TJ
_c
CO
t_
tr-
IO
cn
1—4
>^
r—
3
"3
CO
1— 1
co
5
<4-
0.
IE
C£
o
[^
H^
UJ
00
1
CO
UJ
_l
OQ
^£
1—
1="
0 *->
•i- (O
OJ
s: cu
Q. L.
to 3
O (/>
E en
-t-> CO
«C t.
CL
c 'i'
CO 0.
O D.
O
CJ C
o
t- .f
(11 4-)
•M V
«} t-
«/> o;
•r- O
t^ ^™
5>> O
i— O
0 <
-C OL
Q. •
CVJ O)
0 *J
== <0
01
Ol
t. to
3 CO
0 ^
Q. C_)
X
UJ
0 O
"oi
4-> OQ
C
•P"
T3
C- CU
C5 >
0
d
-------
TERRAIN
The terrain of the modeling region was classified according to land-use
values estimated while compiling the emission inventory (EPA, 1982). For
grid squares not contained in the emission inventory region, land-use
classifications were obtained from a United States Geological Survey
(USGS) map with a 1:250,000 scale. Each land-use category was assigned
a surface roughness value and an estimate of surface uptake velocity
(vegetation factor) according to Wesely (1983) (Table 3-9). Land-use
classification by grid cells for the Philadelphia modeling region is pre-
sented in Figure 3-15.
52
-------
TABLE 3-9. Surface roughness and vegetation factor values.
Land-Use Category
Surface
Roughness
(m)
Surface
Resistance
(s/cm)
Uptake
Velocity
(cm/s)
Vegetation
Factor
Mixed Grassland and Cropland 0.10
Deciduous Forest 1.0
Coniferous Forest 1.0
Urban Area 1.0
Ocean Water 0.001
1.0
0.6
1.5
3.0
20
1.0
1.7
0.7
0.3
0.05
1.0
1.7
0.7
0.3
0.05
53
-------
Legend:
m G = Mixed Grassland
ana cropland
ID D = Deciduous Forest
10 C = Coniferous Forest
H U = Urban Area
Q N = Ocean Water
4500 -
4480 -
446B -
4440 -
0>
c
j= 4420 - SK
-u
c.
O
4400 -
4380 -
4360 -
*34!
i:.;::u:: ......,. t:«
.87
407
427
447
467 487
Easting (km)
507
527
547
FIGURE 3-15. Land-use classifications for the Philadelphia airshed
modeling region.
54
-------
CHARACTERIZATION OF THE 19 JULY 1979 OZONE EPISODE
The second-highest peak ozone readings in the Philadelphia AQCR during the
1979 summer oxidant data enhancement study occurred on Thursday, 19 July
(Allard et al., 1981). On this day, the Philadelphia urban plume was
transported to the west of the central city and precursor transport from
the New York/New Jersey urban area was the apparent cause of a substantial
portion of the ozone concentrations measured in the Philadelphia area.
A broad high-pressure area extending from the midwest through the north-
east to Nova Scotia influenced surface wind patterns. The eastern core of
this high-pressure area was situated to the north of Philadelphia in the
early morning hours, bringing light northerly winds through the airshed
(see Figure 4-1). This high-pressure ridge moved eastward throughout the
day. As a result, the surface winds veered to an easterly direction
around noon and then to a southeasterly flow that persisted the rest of
the day. Winds aloft showed a persistent west-southwesterly flow during
the entire day. Surface trajectory analysis indicated that air parcels
arriving in the urban center of Philadelphia at 1200 EST originated to the
northeast in the New Jersey/New York urban area.
The peak ozone concentration recorded on 19 July 1979 was 17.0 pphm at the
Roxy Water Pump monitoring station, 10 km northeast of central Philadel-
phia. An ozone reading of 16.0 pphm was also recorded at Downington, 45
km due west of downtown Philadelphia. Six monitors recorded ozone concen-
trations equal to, or greater than, 12 pphm; an additional four monitors
recorded ozone concentrations greater than, or equal to, 10 pphm.
55
-------
VtT
FIGURE 4-1. Synoptic situation, 0700 EST, 19 July 1979.
(Source: Allard et al., 1981).
56
-------
DESCRIPTION OF MODEL INPUTS FOR 19 JULY 1979
This section describes the preparation of the UAM input files for the
19 July 1979 episode. The area source emissions, elevated point source,
top of region, and terrain files prepared for the 13 July 1979 simulation
were not day-specific. Only the internal date was changed in these files
before they were used in the UAM simulation for 19 July; therefore, they
are not described in this section. The modeling region specified for 19
July 1979 was that used in the 13 July 1979 simulation (see Section 3).
However, different boundaries were defined for the 19 July 1979 simulation
(see Figure 5-8).
MIXING HEIGHTS
The hourly averaged mixing height values were estimated using available
radiosonde observations and sodar data for the Philadelphia area. The
radiosondes were released in downtown Philadelphia; the sodar was located
at Summit Bridge, Delaware. An example of temperature sounding data for
19 July 1979 is given in Figure 5-1. Both urban and rural mixing heights
were specified on the half hour beginning at midnight. The following
procedures were used in specifying the mixing heights used in the simu-
lation:
Urban mixing heights from 0500 through 1500 EST were estimated using
vertical potential temperature profiles obtained from Philadelphia
upper-air soundings. For each hour, the mixing neight was considered
to be at the base of the layer in which the potential temperature
increased rapidly with height. Upper-air sounding data were avail-
able from the urban site for 0500, 0714, 0937, 1150, and 1500 EST.
For hours between soundings, potential temperature profiles were
obtained by interpolation;
Urban mixing heights between 0000 and 0500 EST were set at the value
estimated from the 0500 EST sounding;
The data indicate that the maximum mixing height was reached at 1200
EST. Urban mixing heights following the time of the 1500 EST sound-
ing were determined by decreasing the mixing height at a rate of -2-3
m/min (Noonkester, 1976; Kaimal, et al., 1982) to reach the 250-m
overnight default value by 2100 EST.
57
-------
IS
20
25
30
35
1200
1000
800-
600
400
200
" 1 1 L 1 I I 1 I i I t 1 I J I i 1 i J I I __t i l_.._L.-_i 1 i L"
IB IS 20 25
TEMPERATURE Id
3B
35
FIGURE 5-1. Temperature sounding for Philadelphia, 0410 EST,
19 July 1979.
58
-------
Rural mixing heights between 0700 and 1100 EST were estimated from
sodar data measured at Summit Bridge, Delaware. Until 0700 and after
1800 EST, the rural mixing heights were held at the 100-m overnight
default value used for 13 July. Between 1100 and 1700 EST, when the
mixing height exceeded the vertical range of the sodar, rural mixing
heights were set at 100 m below the mixing heights estimated from
urban soundings (Spangler et al., 1974).
The resulting mixing height values for both urban and rural grid cells
for 19 July are presented in Table 5-1. A graphical representation of the
temporal increase and subsequent decrease in the mixing depth layer for
both urban and rural cells is presented in Figure 5-2.
WIND FIELD
For the 19 July 1979 episode, surface wind measurements were available for
16 stations. Gridded surface wind fields used with the Urban Airshed
Model were prepared following the same procedure as that for the 13 July
1979 simulation day (see Section 3). The observed surface wind pattern
for three hours during the day is presented in Figure 5-3. Pseudostations
were placed in the same locations as for the 13 July application, and data
were specified to make the interpolation produce a spatially and tem-
porally consistent flow field. The surface winds veered from a light
northeasterly direction in the morning to a consistent southeasterly
direction with moderate speeds by late afternoon. Also included in Figure
5-3 are the actual vectors used in the interpolation of the surface wind
field. This figure shows those vectors that were changed and also data
used to fill gaps for the interpolated field.
Because of the availability of upper-air data within the airshed region
on this day, the procedure used for specifying winds in levels 2, 3, and 4
differed from that used for the upper-level winds for 13 July. Radiosonde
or pibal wind measurements were available for three sites: Wilmington,
Delaware, Downtown Philadelphia, and Trenton, New Jersey. These data were
plotted graphically for review purposes. Figures 5-4, 5-5, and 5-6 pre-
sent measured winds aloft up to 3250 m above Wilmington, Downtown Phila-
delphia, and Trenton, respectively for 1200 EST on 19 July. The data
plotted in these figures show a light surface flow (2 to 4 m/s) in a
general southeasterly direction, while the wind speeds aloft (1250 to
3000 m) were moderate with a general westerly direction. All available
upper-level wind data were plotted in this manner and utilized to estimate
specific hourly inputs for each level. After the data were examined, a
constant average wind vector was obtained for each hour for levels 2, 3,
and 4.
59
-------
TABLE 5-1. Urban and rural
mixing height values used
in the DIFFBREAK file for
19 July 1979.
Time
(EST)
0000
0030
0100
0130
0200
0230
0300
0330
0400
0430
0500
0530
0600
0630
0700
0730
0800
0830
0900
0930
1000
1030
1100
1130
1200
1230
1300
1330
1400
1430
1500
1530
1600
1630
1700
1730
1800
1830
1900
1930
2000
2030
2100
2130
2200
2230
2300
2330
2400
Urban
(m)
190
190
190
190
190
190
190
190
190
190
190
190 .
190
190
190
240
350
460
600
740
890
1060
1240
1420
1530
1530
1480
1410
1340
1270
1200
1120
1020
940
850
750
660
570
480
410
340
300
250
250
250
250
250
250
250
Rural
(m)
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
260
440
490
540
560
580
630
880
1320
1430
1430
1380
1310
1240
1170
1100
1020
920
840
750
480
160
100
100
100
100
100
100
100
100
100
100
100
100
60
-------
0,
0
Legend:
_ Urban
... Rura I
12
Time (hours]
18
FIGURE 5-2. Mixing height profiles for urban and rural
cells on 19 July 1979.
61
-------
FIGURE 5-3. Observed surface wind vectors and
wind vectors used in preparing the interpolated
surface wind field for 19 July 1979.
62
-------
I 5
HIND SPEED IN/SI
4488
4468
- 4448
tc
2
z 4428
QC
s
4488
4388
4368
_' i i | i i i | i i i | i i i I i i i I i i i I i i i 1 i
TREN
Hill
NO^R N0.RT BR°IS
- D°/N fiULE/ LUMB 1
/ PHIL I \|
/
t CH*S//
~ NILM
1
-
~ SUMM 1
i i i 1 i i { 1 i i i 1 i i I 1 i i i 1 i i i 1 i i i 1 i i
487 427 447 467 487 587 527
ERSTINC (KM)
i 1 i i i
-
•
L«& "
1 :
-
_
-
1 1 1 1 1 ~
547
Observed wind vectors
448C
44EC
£
i
4422
4400
4388
NM
TREN
HILL
NORR I EPIS
N3=T
PjN
PHIL
CHES
I -]
GU LRKE -|
I ^
HIL1
sum
/
SH
M:.L
487 427 447 467 467 • SB7 527 547
ERSTINC (KK)
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 5-3a. 0400-0500 EST
63
-------
I I I I I I
0 5
HIND SPEED l«/S>
II I I I I II I I I i ( ! i I
4460
4460
- 4440
4420
4480
4360
4360
HILL
TREN
HILL
I I i I i i i I I I i 1 i I i I i i i I i i i I i i i I i i
t I i i i
407 427
447 467 467 507 527
ERST ING (KM)
547
Observed wind vectors
436Z -
4C7 427 447 457 4E7 5B7 527 557
EP;TINS (KM)
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 5-3b. 1200-1300 EST
64
-------
0 5
H:«.: SPEED (M/S)
44SC
= 4422
44B3 -
43S3I-
43Se>-
DOKN
FrIL
CMES
i i i I
\ KILM
SUKM
KILL
I I I I 1 t I I I i I I 1 I I I I I t 1 I T I ,1
4Z7
427 447 457 4S7 SD7 52' 547
ERSTING (KM I
Observed wind vectors
445C
- 444?
442B
44CC
435:
4363
* JFEN
HILL
X
NOR?
OOHN
NORT
L--.E
RLLE
LUM3
PKIL
CHES
U1LH
\
SW
SUMM
SE
KILL
407 427 447 457 467 507 527 54'
EBST1NC IK'U
Wind vectors used in preparing model input wind fields
for surface layer.
FIGURE 5-3c. 1800-1900 EST
65
-------
DELTfiH = 250 METERS
DELTfiS = 2 M/S
FIGURE 5-4. Pibal sounding at Wilmington, Delaware on 19 Ju1,y
1979, 1200 EST.
66
-------
> H
DEL TRH = Z50 METERS
DEL TRS = 2 M/S
FIGURE 5-5. Pibal sounding at Philadelphia, Pennsylvania on 19 July
1979, 1150 EST.
67
-------
DELTffH = 250 METERS
DELTfiS = Z M/S
FIGURE 5-6. Pibal sounding at Trenton, New Jersey on 19 July
1979, 1200 EST-
68
-------
The preparation of inputs for upper-level winds involved the following
steps:
For the hours of 0500 through 1600 EST, layer-average wind speed
(arithmetic average) and wind direction (predominant direction)
values were obtained for each of the three layers from the rawinsonde
sounding and pibal observations. Linear interpolation was used for
estimating winds for hours without upper-air measurements during this
period.
For the hours of 0000 through 0400 EST, layer-average wind speeds and
wind directions were obtained by interpolation between upper-air
winds from the Washington, D.C. NWS sounding at 1900 EST on July 18
and the Philadelphia sounding at 0500 EST on July 19.
For hours after 1600 EST, winds in layers 3 and 4 were held constant
using the 1600 EST values. In layer 2, winds were held constant
through 1800 EST using the 1600 EST values. After 1800 EST, layer 1
winds were used for layer 2.
A complete list of spatially constant, temporally varying wind inputs for
layers 2, 3, and 4 is presented in Table 5-2. The intermediate step for
completing the three-dimensional wind field involved the use of the inter-
polated surface field, and the program WINDCHAN6. This program reads the
gridded surface field and input data presented in Table 5-2 to create the
three-dimensional wind file. As was done for the 13 July wind file, this
three-dimensional field was rendered nearly free of divergence by running
it through the DIVFREE program. The completed, divergence-free wind file
was then plotted on the airshed grid and inspected for consistency. The
surface flow field used in the simulation is plotted for three hours in
Figure 5-7, and clearly depicts the northeasterly-to-southeasterly shift
in wind direction through the day.
BACKGROUND CONCENTRATIONS
Values used for 19 July background concentrations are the same as those
used in the 13 July simulation except for the ozone background value. On
the basis of examination of upwind monitoring data at the time of mixing,
the value specified for ozone above the mixing height was 0.06 ppm. This
represents a decrease from the 0.08 ppm used in the 13 July stagnation
simulation. Table 5-3 lists the background concentrations used at the top
of the modeling region, as initial concentrations above the mixing height
and for all boundaries above the mixing height. Below the mixing height,
for all boundaries except the Northeast and East boundaries, an ozone
69
-------
TABLE 5-2. Estimated spatially constant, temporally varying wind
input for levels 2, 3, and 4 for the 19 July 1979 wind file.
Level 2
Hour (EST)
0000-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0800-0900
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
Wind
Speed
(m/s)
2.0
2.0
2.0
2.5
3.0
3.0
2.5
2.5
5.0
4.5
4.0
3.5
3.0
4.0
3.0
3.0
3.0
3.0
3.0
Surface
H i
H i
,i
H i
Wind
Di rect i on
(Degrees)
30
40
50
50
55
55
50
40
55
75
90
85
75
90
105
120
165
165
165
Winds Used
,
11 "
,
ii ii
Level 3
Wind
Speed
(m/s)
5.5
5.5
5.5
5.5
5.5
5.5
6.0
6.0
7.4
6.0
5.5
4.5
4.5
2.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
Wind
Direction
(Degrees)
30
40
50
50
50
50
55
55
60
75
85
70
70
95
50
155
155
155
155
155
155
155
155
155
Level 4
Wind
Speed
(m/s)
5.0
4.5
4.5
4.0
4.0
4.0
4.5
5.0
6.0
4.0
3.0
2.0
4.0
6.0
6.0
4.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
Wind
Di rection
(Degrees)
350
350
345
345
345
345
10
30
30
35
25
150
265
250
275
240
280
280
280
280
280
280
280
280
70
-------
10
15
20
0 5
HIND SPEED (M/S)
25 ' 30
35
30
25
I i I I | I I i . I . | . I . I IT | II I Ij TTI 17] . TT1 II \
(a) 400 - 500 EST
FIGURE 5-7. Airshed model surface winds for 19 July 1979.
-30
I
71
-------
I
J
0
10 .
• 5
HIND SPEED (M/S)
20 25
30 35
— 10
-5
0
10
15 20
(b) 1200 - 1300 EST
25 30
35
0
FIGURE 5-7 (continued).
72
-------
1111
it
0.
0
FIGURE 5-7 (concluded)
15 20 25
(c) 1800 - 1900 EST
30 35
73
-------
TABLE 5-3. Background concentration values
for 19 July at the top of the modeling region
(TOPCONC), as initial concentrations above the
mixing height, and for all levels of all boundaries
except the levels below the mixing height on the
Northeast and East boundaries.
Species
NO
N02
°3
CO
ETH
OLE
PAR
CARB
ARO
PAN
BZA
Concentration
(ppm)
0.001
0.002
0.06*
0.2
0.001
0.0004
0.040
0.010
0.0008
0.000025
0.00001
A value of 0.05 ppm was used below
the mixing height for all boundaries
except the Northeast and East
boundaries.
74
-------
value of 0.05 ppm was specified. The value should have been 0.06 ppm;
however, the difference is not significant because these are outflow
boundaries.
INITIAL CONDITIONS
The initial-condition field for 19 July was created following the same
procedure as that used for 13 July 1979 (see Section 3). Because the
simulation began at midnight, the initial-condition values were obtained
by computing a two-hour average starting at 2300 on 18 July. Stations
without data for a particular species were left out of the Poisson inter-
polation scheme. Background concentrations for all species were used in
levels 3 and 4 above the mixing height, and a linear interpolation was
used between the background values aloft and the surface (layer 1) value
for each cell to obtain the concentration value in level 2. The initial-
condition input values used in the preparation of the AIRQUALITY file for
19 July are presented in Table 5-4.
BOUNDARY CONDITIONS
Boundary conditions were specified after examination of all air quality
and meteorological data collected during this episode. Because of wind
flow through the airshed on this day and the need to limit simulation
costs, certain unnecessary grid cells were eliminated from the simulation
by the boundary specifications. No grid cells containing major emissions
sources were excluded by this procedure. The boundaries used in the simu-
lation are shown in Figure 5-8. After the hourly three-dimensional wind
fields were prepared, a number of parcel trajectories were released at
various locations within the grid. The trajectory analysis revealed that
air parcels arriving in central downtown Philadelphia near noontime
orginated in the New Jersey/New York urban area directly to the northeast
in the early morning hours. Because of this inflow from an adjacent air-
shed region, it was important to estimate the levels of migratory regional
pollutants being advected from the neighboring "dirty" air mass. There-
fore, two sets of boundary conditions were specified: one set of condi-
tions for the East and Northeast boundaries below the mixing height to
estimate the inflow of pollutants from the New Jersey/New York urban area,
and another set for all other boundary conditions as an estimate of back-
ground conditions presented earlier (see Table 5-3). The estimates of the
Northeast and East boundaries below the mixing height were the result of
further analysis.
Precursor transport from the New Jersey/New York urban area into the
Philadelphia AQCR occurred on 19 July 1979. This occurrence is supported
75
-------
TABLE 5-4. Initial conditions for 19 July 1979 (concentrations in ppm),
Station
AMS Lab.
Ancora
Brigantine
Bristol
Camden
Chester
Cl aymont
Conshohocken
Defense Support
Downington
Franklin Inst.
Island Rd. Airp. Cir.
Lumberton
Norristown Armory
Northeast Airp.
Roxy Water Pump
SE Sewage Plant
South Broad
Summit Bridge
SVi Corner Broad/Butler
Trenton
Van Hi Seville
Vineland
Easting
(m)
491600
511800
546000
511000
491700
469000
461500
474500
483800
436000
485200
480300
518000
473500
499000
479500
487200
486100
441000
487000
520000
559000
498200
RHC
Component
PAR
OLE
ETH
ARO
CARB
Northing
(m)
4428500
4392400
4337506
4440000
4419000
4410000
4406400
4435600
4418300
4426000
4422800
4414800
4423000
4440000
4436000
4433100
4417300
4421600
4376000
4418000
4452000
4439500
4371200
NO
0.000
—
—
—
0.011
—
—
—
—
0.006
0.030
—
0.000
—
0.000
—
—
0.020
0.000
—
—
0.000
—
N02 CO
0.025 0.5
—
—
0.039 —
0.051 2.5
0.024 —
0.045 —
—
—
0.015 0.2
0.050 0.5
—
0.028 0.0
—
0.015 —
—
—
0.045 1.5
0.009 0.2
—
—
0.006 0.2
0.025 —
°3
0.015
0.015
0.017
0.012
0.004
0.026
0.020
0.000
0.000
0.024
0.010
0.015
0.008
0.030
—
0.010
0.005
0.010
0.009
—
0.002
0.021
0.010
RHC
0.35
—
—
—
—
0.95
—
--
—
0.05
—
—
0.40
0.29
—
—
—
0.25
0.10
—
—
—
—
Carbon-Bond Fraction
(% as Carbon)
74.0
2.8
4.1
13.2
5.9
76
-------
NORTH
10
20
ortheast Boundary:-
East
Boundary
FIGURE 5-8. Boundary specifications for 19 July 1979 simulation,
(Bold line is emissions region.)
77
-------
by the following observations: (1) overnight winds were northerly in the
New York and Philadelphia areas, (2) high NOX concentrations -were recorded
in the early morning hours upwind of Philadelphia at the Van Hiseville,
New Jersey monitor, and (3) a surface layer back trajectory from the high
ozone concentration of 16.0 pphm recorded at Claymont, Delaware, using
interpolated surface wind fields generated from observed data, indicates a
pathway back towards the New Jersey/New York metropolitan area that does
not traverse the high-emission-density area of Philadelphia. Therefore,
for the East and Northeast boundaries, the Van Hiseville monitor, which is
located near the intersection of these two boundaries, was used to define
boundary conditions. On the basis of the expected behavior of the New
York plume derived from field studies in New York and several other
cities, overnight precursor transport is likely to have occurred just
above the surface (Alkezweeny et al., 1981; Clarke, 1969; Godowitch et
al., 1984b, Possiel et al., 1984). Thus, the Van Hiseville monitor is
likely to have been just below the urban precursors transported throughout
the night. Therefore, the 0500-0600 EST surface NOX measurements at Van
Hiseville were extrapolated back in time to midnight to account for over-
night transport of precursors just above the surface monitor at Van Hi Se-
ville, but within the mixed layer.
Estimates of boundary conditions for the Northeast and East boundaries
below the mixing height are presented in Table 5-5. Observed NO and NOo
data for the hour of 0500-0600 EST were used as input boundary values for
the hours of 0000-0600. Since hydrocarbons were not measured at the Van
Hiseville monitor, estimates of the influx of total reactive hydrocarbons
across the Northeast and East boundaries below the mixing height were
specified by multiplying the hourly NOX concentrations at the Van Hise-
ville monitor by the Philadelphia surface layer emission inventory hydro-
carbon/NOx ratio of 6. The total reactive hydrocarbon value was then
speciated into carbon-bond components using the carbon-bond fractions of
the emission inventory.
MESTSCALARS
The inputs for the METSCALARS file were obtained for 19 July using the
same procedures as those employed for the 13 July simulation (see Section
3). Radiosonde data from central Philadelphia were used to specify the
temperature gradients above and below the mixing height. A value of 24000
ppm for the concentration of water vapor was specified because dewpoints
for a number of stations in the region were in the upper 60s and low
70s. Surface pressure at 0700 EST was 1021.7 mb, so a value of 1.0 atm
was specified for the atmospheric pressure model input. The N02 photoly-
sis rate constants were obtained by the procedure used in the 13 July
simulation. The values used for all parameters in the METSCALARS file for
19 July are presented in Table 5-6.
78
-------
C
X
•f—
E
.c
O
"ol
J3
•r—
c
U
•a
c
3
CC
•
in
1
UJ
_l
co
1—
monitor
I/I
t-
OJ
ai
a>
(/I
c
C
*o o
?1ght esti
:oncentra1
^ — •
D.
CO
CC
^^ f~t f*^ c3 <^^ <*^ »~H if) QO p**» r^^ r^^ p*^ Lt~* LO Lfy ^^
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOO
OOOOOOOOt— li— IT— (1— Ir—
-------
c
o
03
3
>>
's
CTi
4J
S_
O
3
Q.
C
OO
I—
LU
LU
co
t_ IT]
in
a;
di CL
O Q.
C
O
+J (O
(/> Qi
•i- O
O 0.
Si
£ OL
£%
Of.
o o o c o c
oocooocooooocooocococooo
ocococoooooocoococoocccc
oooooi—t T-I i—levjcsjcMPOfncofOPoevjcvjr-ii—icooo
OOOOOOOOOOOOOOOOOOOOOOOO
ooooooooooooo
I I I I I I I I I I I I I
ooooooooooo
I I I I I I I I I I I
§^-«a-*±«d-»a-oLOuoi«tni*-i>^p^r».oocOLOtr)uf)LO^i-^-^r
oooootDp*-cocriO'—iCMco^a-mvo
OOOOOOOOO«-t«-i«-^«-««—'«—<^H
I I I I I I I I I I I I I I I I I I I I I I I I
oooooooooooooooooooooooo
oooooooooooooooooooooooo
O«—iCMco^-to^or^cocnO'—'CMCo<3-ir>vor».cocriO'—'cviro
OOOOOOOOOO'-f'—'i—">—
-------
ANALYSIS OF URBAN AIRSHED MODEL PERFORMANCE FOR THE PHILADELPHIA
SIMULATIONS OF 13 and 19 JULY 1979
This section presents a brief overview of the results and the analysis of
model performance for the UAM simulations conducted for the Philadelphia
region for 13 and 19 July 1979. These simulations originated at midnight
and continued for 24 hours. For a more complete presentation of the
results of the 13 and 19 July 1979 simulations, refer to Appendixes A and
B. The statistical measures used in the analysis of model performance are
first defined; then the simulation results and the performance evaluations
for 13 and 19 July 1979 are presented.
MODEL PERFORMANCE EVALUATION MEASURES
The primary objective of the model performance evaluation effort is to
ascertain how well the UAM computes the peak concentrations and spatial
and temporal distribution of 03 throughout the Philadelphia study area.
This computation can be accomplished by comparison of available hourly
averaged measurements of Og with corresponding predictions for surface-
level grid cells containing the monitoring sites. Several means of quan-
tifying model performance have been developed (e.g., Bencala and Seinfeld,
1979; Hayes, 1979; Fox, 1981; Layland and Cole, 1983). Hayes (1979)
reported a detailed examination of candidate evaluation measures for air
quality dispersion models that identified five attributes of desirable
model performance:
(1) Accuracy of the calculated peak concentration
(2) Absence of systematic bias
(3) Lack of gross error
(4) Temporal correlation
(5) Spatial alignment
The evaluation of UAM results for Philadelphia focused on quantitative
assessments of each of these performance attributes. Some modifications
to the performance measures suggested by Hayes (1979) were made to take
into account the recommendations of the AMS workshop on air quality model
performance (Fox, 1981).
81
-------
Accuracy of the Calculated Peak Concentration
The accuracy of calculated 1-hour 03 peak concentrations can be evaluated
by several methods. The measured 1-hour 03 peak concentrations can be
compared with (1) the highest 1-hour 03 concentrations calculated at the
same monitoring station, (2) the highest 1-hour 03 concentration calcula-
ted at any monitoring station (i.e., the calculated 1-hour 03 peak station
concentration), or (3) the highest 1-hour 03 concentration calculated in
any surface-level grid cell (i.e., the calculated 1-hour 0^ peak grid con-
centration).
Lay!and and Cole (1983) propose that of these three approaches, the com-
parison of the measured peak concentration with the calculated peak con-
centration at any monitoring station is the most useful method. They
believe that the comparison of the measured peak concentration with the
calculated peak concentration in any surface-level grid cell will bias the
evaluation toward overestimation because the station monitoring network is
less likely to report the highest 03 concentrations occurring in the grid-
ded area. They note also that a comparison of the peak predicted and
observed concentrations for the same monitoring station would bias the
evaluation toward underprediction because the model may not exactly repro-
duce the spatial pattern of the concentration field.
Consequently, in this study, the accuracy of the 1-hour 03 peak concentra-
tion is evaluated by comparing the measured and calculated peak 03 concen-
tration at any of the monitoring stations. It should be noted that this
model evaluation procedure is consistent with the regulatory aspect of the
03 NAAQS since the concern in this case is with the peak 1-hour 03 concen-
tration occurring in the monitoring network regardless of time or loca-
tion. The 03 concentration at a monitoring station is calculated as the
distance-weighted average of the surrounding four grid cells.
For completeness, the highest 1-hour 03 concentrations calculated and
measured at the same monitoring station are compared. These pairs of
station peak 0)3 concentrations are plotted on a scatter diagram of calcu-
lated versus measured values, and the correlation coefficient is calcula-
ted. The highest 1-hour 03 concentration calculated in any surface-level
grid cell is also reported.
Absence of Systematic Bias
Absence of systematic bias refers to the ability of a model to avoid
either underestimating or overestimating pollutant concentrations. Esti-
mates of the mean bias were calculated in the following manner:
82
-------
Mean Deviation
1
¥
N
E
1=1
(CB
Mean Normalized Deviation =
N
N
E
(C
m,i
- C
0.5 x re
c,i
c,'
where C and
Cm are
is
the calculated and measured concentrations, respec-
tively, and N is the total number of comparisons. These measures are
consistent with the recommendations made by the AMS workshop (Fox,
1981). A negative bias corresponds to model overprediction and a positive
bias corresponds to model underprediction. The mean deviation is normali-
zed with respect to the arithmetic mean of the measured and calculated
concentrations according to Hayes (1979). The standard deviation for each
of these measures of bias is also calculated to provide
the variability of the quantities (Cm ^ - Cc ^ ) and (C
[0.5 x (C . +
deviations'(a)
C .)].
ar£ as
m
c
m
Mathematically, the expressions
follows:
an indication
- C
of
for
c
the
standard
o (Deviation)
(C
m.i
o (Normalized Deviation) =
'*.<
1/2
1
N
1
N
N
E
1-1
N
E
[/ C
10.5 x
/ c»
10.5 x
m,i
(Cm,i *
,.1 - Cc
-------
Absence of Gross Error
The absence of gross error can be determined through the calculation of
the mean absolute deviation and the mean normalized absolute deviation.
These measures are given by the following expressions:
Mean Absolute Deviation =
C . - C
.
»
Mean Normalized Absolute Deviation = TT
' m.i " c,i
The standard deviation for each of these measures is calculated as
follows:
o (Absolute Deviation)
- C
Ni=i
- r
L
c,i
1/2
a (Normalized
Absolute Deviation)
" C
0.5
c,i
0.5
r - c
Vi c.i
--
x (Cm,i + Cc,i>J
2
1/2
If the mean absolute deviation of the pairs of calculated and measured
values is small, then the model is said to exhibit "skill" as a predic-
tor. If the mean absolute deviation is large, the model suffers from the
presence of large gross errors. As with the bias determination, these
measures were calculated for 03 for all cases in which the measured values
exceeded a 5 pphm threshold. The mean normalized absolute deviation for
6 was plotted as a function of measured concentration to complement the
84
-------
corresponding bias displays discussed previously. The mean absolute
deviation and mean normalized absolute deviation were also calculated for
the pairs of measured and calculated peak 1-hour 63 concentrations at
monitoring stations. These pairs of peak station concentrations were also
plotted on scatter diagrams.
Temporal Correlation
Temporal correlation refers to the "timing" or "phasing" of the measured
and calculated ozone levels at a specified station. The temporal correla-
tion for a given station can be determined by using the hourly pairs of
measured and calculated concentrations to define the appropriate mean
values. A correlation coefficient can then be calculated according to the
following expression (Hoel, 1962):
Correlation =
Coefficient
1_
N
E
i = l
l- £ /c -
N Aj \S:,i
/ i N \ / ,
r 1 ^ J / 1
\ c,i " ¥ r-* 0,1 J \ m,i ~ ¥
i N \2
1 V r \
"N ^ °c i
N i=1 c,y
1/2
IE(C .
N & ^,,
N \
Scm1
i=l m»V
1 f c Y
~ "N 1=1 m'V
1/2
where N is the number of comparisons for a particular station. Lack of
temporal correlation can be ascribed to one or more causes, including
inadequate characterization of emissions, wind, or mixing depth inputs.
To calculate an average temporal correlation coefficient on a particular
day, the following change of variable is performed (Hoel, 1962):
where r- is the computed correlation coefficient for station j. Next, the
mean value of j is estimated:
M
,E (n.j - 3) ^
"* =~M
E (nj - 3)
85
-------
where M is the number of monitoring stations and nj is the number of com-
parisons made for station j. This somewhat complicated transformation is
used because the variance of an estimated correlation coefficient is a
function of both sample size and the population correlation coefficient,
p. The transformation to $j converts rj to an approximately normally
distributed random variable with a variance of I/(HJ - 3) that is not
dependent on the population value, p. The average of all $,• is computed
by weighting each value by its variance. Then p can be determined from
the following equation:
p = exp(2T) - 1
exp(2$~) + 1
Spatial Alignment
The spatial alignment of measured and calculated concentration fields is
another useful measure of model performance. For a given hour, imagine
two concentration isopleths, one constructed from measured pollutant con-
centrations, and the other from the corresponding model calculations. If
one isopleth were placed over the other, the degree of spatial misalign-
ment would be easy to discern. Spatial alignment can be quantified by
considering a sequence of "time slices." The calculated and measured
concentrations for a particular hour are employed to calculate a correla-
tion coefficient using the formula given in the previous section. When
applying the expression for the correlation coefficient for this example,
N is equal to the number of comparisons available for a specific hour.
Spatial correlation coefficients can thus be computed for each hour, and
estimates of the average spatial correlation coefficient for each simula-
tion day can also be made following the general procedure described in the
previous section.
A high correlation coefficient means that the calculated spatial distri-
bution of pollutants over the modeling region for a specific hour corre-
sponds closely to that indicated by the measurements. Poor correlation is
sometimes to be expected during hours when the measured values do not
exhibit significant spatial variability. Although the absolute magnitudes
of the calculated and measured values might be in reasonably close agree-
ment, the small variations in pollutant levels from station to station
might not be replicated in the variations of the calculated values, thus
86
-------
yielding a relatively low spatial correlation coefficient. Common sources
of spatial misalignment include discrepancies between modeled and observed
wind velocities and directions, inaccuracies in the emission inventory,
and improper treatment of photochemistry.
The results of the application of these performance measures to the UAM
simulations of the 13 and 19 July 1979 ozone episodes in Philadelphia are
presented next.
MODEL EVALUATION RESULTS FOR THE 13 JULY 1979 SIMULATION
Following the procedures outlined in the previous section, measures of
overall model accuracy, bias, error, and temporal and spatial correlations
were calculated. Bias and error were calculated for all pairs of ozone
concentrations for which the measured value exceeded 5 pphm, as well as
for pairs of station peak ozone concentrations. These results are sum-
marized in Tables 6-1 and 6-2. Table 6-1 presents a detailed comparison
of maximum measured and calculated ozone concentrations for each of the 22
monitoring stations. Table 6-2 presents the measures of model performance
for this simulation.
Accuracy of Calculated Peak Concentrations
From a regulatory point of view, the ability of a model to reproduce mea-
sured peak concentrations is a major attribute of model performance.
Table 6-1 presents the measured and calculated peak ozone concentrations
at each monitoring station, the normalized residuals, the times of occur-
rence and the lag time. A scatter plot of these pairs of station peak
ozone concentrations is presented in Figure 6-1; bias and gross error are
given in Table 6-2.
The maximum measured 1-hour ozone concentration on 13 July was 20.5 pphm
at the Conshohocken station at 1600-1700 EST. The peak calculated concen-
tration at this same station was 18.5 pphm. The calculated ozone maximum
concentration occurred two hours prior to the measured ozone peak value.
The model-predicted maximum station value was 19.3 pphm, at the nearby
Roxy Water Pump station, at 1300-1400 EST. This value is close to the
measured peak concentration of 20 pphm at this same station and to the
measured airshed-wide peak concentration of 20.5 pphm. The model under-
predicts the station peak ozone concentration by 6 percent. The largest
calculated 1-hour ozone concentration on the entire grid was 26.6 pphm at
1600-1700 EST.
87
-------
E
Q.
O.
o;
c
o
M
O
s_
o
en
en
r—I
>)
3
S_
O)
.a
o
to
c
o
-t->
u
OC
a: OZH-
s> •- — z
O Q X ul
X UJ < O
QCZ2
a. C
U
o Z <
u. uj r> oi
O > Z t-
oi — z
of uj X ul
3 (/> < U
o tor z
X O O
U
CD
LU UJ O
UJ O ^N
H-
O U Qi Z
» i_i LU LJ
i- a in o
< u co z
a: a: O O
a. u
t-
<
O 2 I—
1-1 -H Z
O X LU
U < U
I* Z Z
a- O
u
O H
LUZ<
(A X ul
« < U
O Z Z
O
u
in
Z
o
^ai — CM
z o.
LU a-
to -i < a. _i -
U t/J O O Z 3 ui LU O o£ >
I— O l-zoiOOt-<.nctai LU
co _i a: z i LU ui — H t- < < no z c/i
< < O Z ui O O to z _i O <* V) ;* 5 I- O "-
-jail-ujl-Siz — ^zuj—' uj i >-• I— x
<->— J
88
-------
TABLE 6-2. Performance measures for the UAM simulation of the
13 July 1979 episode in Philadelphia.
Performance Attribute
Performance Measure
Values
Accuracy of the airshed Ratio of predicted to
peak prediction measured station peaks
Time difference between
predicted and measured
station peaks
Station Peaks
Predicted
Measured
Station peaks
Systematic bias
Gross error
Mean Deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Bounds at the 90
percent confidence
level
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
0.940
-2 hours
19.3 pphm
(Roxy Water)
20.5 pphm
(Conshohocken)
-0.086
0.180
•0.998 pphm
2.491 pphm
•2.790 and 0.793 pphm
0.148
0.130
2.021 pphm
1.718 pphm
89
-------
TABLE 6-2. (Concluded)
Performance Attribute
Performance Measure
Values
All 03 Concentrations > 5 pphm
Systematic bias Mean deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Gross error
Temporal correlation
Spatial alignment
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Temporal correlation
coefficients
Each station
All-station average
Spatial correlation
coefficients
Each hour
All-hour average
•0.150
0.361
•1.660 pphm
2.885 pphm
0.295
0.255
2.670 pphm
1.981 pphm
0.273 to 0.956
0.737
0.041 to 0.735
0.456
90
-------
25.00
20.00
15.00
UJ
o
M4
O
UJ
10.00
LU
O£
0.
£ 5.00
0.1
I I I I I I I
B
I I I I I I I I I I I I I I I I
r. 00
5.00 10.00 15.00
MRXIMUM OBSERVED OZONE CPPHM)
20.00
25.00
FIGURE 6-1. Scatter plot of predicted and observed station peak
ozone concentrations for 13 July 1979. The correlation coefficient
for the 19 pairs of station values is 0.731. The solid line repre-
sents perfect agreement between observed and predicted concentrations;
the dotted lines represent the envelope of the predicted values that
are within a factor of two of the corresponding observed values.
91
-------
The scatter plot of station peak concentrations (Figure 6-1) shows a cor-
relation of 0.731. The normalized bias and error are -0.086 and 0.148,
respectively. The model appears to overestimate, on average, the station
peak ozone concentrations by about 1 pphm. All the calculated peak sta-
tion values are within a factor of 2 of the measured peak station values.
Estimates of Overall Systematic Bias
Measures of potential systematic bias were calculated as both nonnormali-
zed and normalized quantities. Table 6-2 presents these biases for the 13
July 1979 simulation with the constraint that the measured ozone concen-
trations equaled or exceeded 5 pphm. Figure 6-2a shows a plot of the
normalized bias as a function of the measured ozone concentrations. For
measured ozone concentrations ranging from 5 to 16 pphm, the model tends
to overestimate concentrations. The model tends to underestimate only at
very low or very high ozone concentrations. The model also appears to
overestimate ozone concentrations over the entire range of concentrations
measured on 13 July 1979. For concentrations greater than 5 pphm, the
model tends to overestimate measured values by 15 percent. The magnitude
of the average overestimation is 1.66 pphm.
Estimates of Gross Error
The mean absolute deviation as an indication of gross error is estimated
by averaging the absolute (unsigned) differences between the pairs of
calculated and measured concentrations. This gross error is presented as
both nonnormalized and normalized values in Table 6-2. Figure 6-2b pre-
sents the normalized gross error as a function of measured ozone concen-
trations. The average error for measured ozone concentrations greater
than 5 pphm is 29.5 percent. The magnitude of the average error is 2.67
pphm. At elevated measured concentrations (i.e., above 12 pphm), the
error is nearly independent of the measured ozone concentrations (see
Figure 6-2b).
Temporal Correlation
The temporal evolution of observed and predicted ozone concentrations at
16 selected monitoring stations is presented in Figure 6-3.
The temporal correlation coefficients for ozone indicate a broad range of
values for the individual monitoring stations (see Table 6-2). For ozone
concentrations above 5 pphm, the average coefficient is 0.737. An
examination of Figure 6-3 indicates that most of the observed temporal
92
-------
2.0
2 i-B
cc
UJ
0 0.0
X
o
* -1.01-
UJ
-2.8
12 15 18 21
24
I 1 1 1 1 1 1 1 1 1
_ 03
DflTfl POINTS 307
\
%
V
>
*.
X _ j- _ _ B- <- -EJ- *" "
v _.Bjj™ — ^f ^u
=- „•»"
~«r
i I i i i I i i i I
1 i ' i ' i
^
«
_-,~B
^
—
-
i i i i i i
2.0
- 1.0
0.0
- -1.0
2.0
1.5
6 9 12 15 18 21 24
HERSURED CONCENTRRTION (PPHM)
(a)
6 9 12 15 16 21 24
UJ
1.0
0.5
«E
UJ
1.0
' I 'I ' I
03
DRTH POINTS 307
I
T ' I
-•or
i I i I
""--*a
•3*»^ 4
I
--*«.—a—«--^
6 9 12 15 18
HERSURED CONCENTRRTION (PPHM)
-2.0
21 24
2.0
1.5
1.0
0.5
0.0
(b)
0.5
UJ
o
O£
0 , _
o i>. 3
o
u.
o
>_ 0.2
h""
_i
S0.i
CD
O
20 -16 -12 -8-40 4 8 12 16 2
i 1 i 1 i 1 i | i
03
DRTfl POINTS 307
-
-
PI
. . OffS-0 .
1 1 1 1 1 1 1 1 1
-
_
s ~
0.4
3.3
6.2
RESIDURL (PPHM)
(c)
FIGURE 6-2. Mean normalized bias, mean normalized error as function
of measured 03 concentrations and distribution of residuals for 13
July 1979.
93
-------
g .'
Z
3 O
O QK
uo oo.
i I i I i I • ! ! I i I ' I i I i
o
1 o
" S«
t—. Ott,
i t t I i ! i ! i 1 < I i I i It I i*
m m •!->
S- f3
01 T—
00 3
-C E
o ••-
uo
1/1
01
-a T—
0) =5
QJ
S_ QJ
a. -c
o s-
o
.. — . 4-
1/1
1- '
o 'B
OJ
on c
O) O
•r- M
S- O
ai
(/) ~O
a>
cu -u
e u
I— T3
a>
co
LU 03
a: ••-
=5 4-»
O T:
i— i Q.
LJ_ 1/1
(N « «• IB 10 ft*
94
-------
c ""^
> §£
I I i I . I I I 1 I I I I I I
1S«3
in ^_
en CM
HS
i
! fNJ
o
x
1S3M
i *
= tnuj
K taoe
U 00.
i I i I i ! i I i I i I i I i I i I
m
-------
trends in the ozone measurements are represented in the model predictions;
however, the model overestimates ozone levels after the time at which the
peak observed value occurs.
Spatial Alignment
Spatial isopleths of calculated hourly averaged ozone concentrations at
1600-1700 EST are presented in Figure 6-3. Because of the consistent
southerly wind flow established by late morning of 13 July 1979 in the
Philadelphia region, the center of the ozone cloud depicted in Figure 6-3
is located north of the urban emission area. Spatial correlation coef-
ficients are typically smaller than the corresponding temporal coef-
ficients (see Table 6-2). For instance, for ozone levels greater than or
equal to 5 pphm, the average spatial coefficient is 0.456. The spatial
correlation is expected to be lower than the temporal correlation. The
temporal correlation of calculated and measured concentrations will
generally be satisfactory since the diurnal pattern of the ozone concen-
trations depends strongly on solar irradiation and will therefore be
fairly well reproduced by a model such as the Urban Airshed Model that
takes into account the effect of solar irradiation on ozone chemistry. On
the other hand, uncertainties in meteorological inputs (e.g., wind speeds,
wind directions, mixing heights, and mixing rates) can greatly affect the
numerical value of the spatial correlation coefficient. For example,
reduction of surface and upper-air meteorological data in Urban Airshed
Model simulations of the Los Angeles basin had a large effect on the tra-
jectories of air parcels and on the calculated ozone concentrations (Seig-
neur et al., 1981).
Overall Correlation
The distribution of the residuals (measured ozone concentrations minus
calculated ozone concentration) is shown in Figure 6-2c. Most of the
residuals are within + 4 pphm, with a mean value of -1.7 pphm. A scatter
plot of all ozone prediction-observation pairs is shown in Figure 6-4.
Seventy-three percent of the predictions are within a factor of 2 of the
observations. The correlation coefficient is 0.828.
MODEL EVALUATION RESULTS FOR THE 19 JULY 1979 SIMULATION
Following the procedures outlined in the first section, measures of over-
all model accuracy, bias, error, and temporal and spatial conditions were
calculated. Bias and error were calculated for all pairs of ozone con-
centrations in which the measured value exceeded 5 pphm, as well as for
96
-------
25.00
O *
an n ,O0 / a
n •' D -' / n
m m H
— ™ nD
I rfi I I I I I I I I I I I I I
5.00 10.00 15.00 20.00
OBSERVED OZONE CONCENTRATION (PPHM)
25.00
FIGURE 6-4. Scatter plot of predicted and observed ozone
concentrations for 13 July 1979. The correlation coefficient
for the 377 pairs of ozone concentrations is 0.828. The solid
line represents perfect agreement between observed and predicted
concentrations; the small dotted lines represent the envelope of
the predicted values that are within a factor of two of the
corresponding observed values, and the heavy dashed line repre-
sents the regression line forced through the origin.
97
-------
pairs of station peak ozone concentrations. The results are summarized in
Tables 6-3 and 6-4. Table 6-3 presents a detailed comparison of maximum
measured and calculated ozone concentrations for each of the 22 monitoring
stations. Table 6-4 presents the measures of model performance for this
simulation.
Accuracy of Calculated Peak Concentrations
Table 6-3 compares the measured and calculated peak ozone concentrations
at each monitoring station, the normalized residuals, the times of occur-
rence of the peak, and the lag times. A scatter plot of these pairs of
peak 03 concentrations is presented in Figure 6-5; bias and gross error
are given in Table 6-4.
The maximum 1-hour concentration measured on 19 July was 17.0 pphm at the
Roxy Water Pump station at 1400-1500 EST. The calculated peak concentra-
tion at this same station was 13.8 pphm. The largest station peak ozone
concentration calculated by the model was 14.2 pphm at Downington at 1500-
1600 EST. The model underpredicted the station peak ozone concentration
by 16 percent. The largest 1-hour ozone concentration calculated by the
model on the entire grid was 17.7 pphm.
The scatter plot of station peak concentrations (Figure 6-5) shows a cor-
relation of 0.478. The normalized bias and error are -0.021 and 0.205,
respectively. The model does not show any particular trend toward either
over- or underestimation of the peak station ozone concentration. All
calculated station peak ozone concentrations except one are within a fac-
tor of two of the measured station peak ozone concentrations.
Estimates of Systematic Bias
Normalized and nonnormalized measures of model bias are presented in Table
6-4 for ozone concentrations above 5 pphm. The normalized bias of 0.5
percent does not show a significant trend toward underestimation. The
magnitude of the underestimation is only 0.024 pphm. The lack of under-
or overestimation is exemplified in Figure 6-6a, which represents the
normalized bias as a function of the measured ozone concentrations. It
appears that the model overestimates ozone concentrations between 3 and 9
pphm and underestimates ozone concentrations below 3 pphm and above 9
pphm.
98
-------
Q.
Q.
OJ
c
o
IM
O
CT1
CTi
i—i
>>
p—
3
CTN
S_
O
01
c
o
OJ
CO
.0
o
-*»
1/1
c
O
•o
0)
OL
X
(O
CO
i
10
<0
U. O h-
O UJ X <
h-soi
oc ori-
3 -• — z
O o x ui
X UJ < <_)
Of £Z
o. O
U
or <
u- LU z> at
O > Z 1-
at — z
a: uj x uu
3 v> < o
O ca z: z
I O O
U
09
UJ LU O
U) O >*
at
a. o a.
u. o a <
O ui uj oi
K> l-
o o a z
«•" "" LU LU
k- Q O
< LU CO Z
a: of o O
a. u
a h-
LUZ <
I- => us
OZH-
i-, _ Z
o x LU
LU < U
oc z: z
Q. O
u
z
o
Q H
LUX <
> = at
tn x LU
co < o
02Z
O
U
— <-»ci i
ill
^^1 — (M^CVJlOVD —
I-
to
z a. co ~ < •— a. _i u>
LU OL z< < a. a a
u ii z> z « z o: < — •
Z ULOQ QZ3H-UIUJOQ;
«-• h-O h- ZoiOOc/JI— Ua£Q3
CO -IK Qi Z I LU t9 — t- h- < < < CO Z
< < O z z ui o O to z _i a os co ui3 3 I- o
-i at l- < LU t- z i z •— ^ z LU — x LU x •-• (-
O LO u a co >- co LU z z < ca o£ i- >- vi i- z z
MO — — Z LU < ZU. 3 < _l Z OS OiX ZSZUU
>
LUO
CO Z
— <
x _j
99
-------
TABLE 6-4. Performance measures for the UAM simulation of the
19 July 1979 episode in Philadelphia.
Performance Attribute
Performance Measure
Values
Accuracy of the airshed Ratio of predicted to
peak prediction measured station peaks
Time difference between
predicted and measured
station peaks
Station Peaks
Predicted
Measured
Station peaks
Systematic bias
Gross error
Mean deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
Bounds at the 90
percent confidence
level
Mean absolute deviation
Normalized
Average
Std. dev.
Nonnormalized
Average
Std. dev.
0.836
1 hour
14.2 pphm
(Downington)
17.0 pphm
(Roxy Water)
-0.021
0.275
0.047 pphm
3.038 pphm
-1.437 and 1.531 pphm
0.205
0.179
2.293 pphm
1.930 pphm
100
-------
TABLE 6-4 (Concluded)
Performance Attribute
Performance Measure
Values
All Oo concentrations > 5 pphm
Systematic
bias
Mean Deviation
Normalized
Average
Std. dev.
0.005
0.398
Gross error
Temporal correlation
Spatial alignment
Nonnormalized
Average 0.024 pphm
Std. dev. 3.163 pphm
Mean absolute deviation
Normalized
Average 0.288
Std. dev. 0.273
Nonnormalized
Average 2.400 pphm
Std. dev. 2.051 pphm
Temporal correlation
coefficients
Each station
All-station average
Spatial correlation
coefficients
Each hour
All-hour average
-0.914 to 0.970
0.525
-0.421 to 0.643
0.202
101
-------
25.80
'•'\ i i i I i i i i I i i i i I i i i i 1 i i i t
,00
5.00 10.00 15.00
MRXIMUM OBSERVED OZONE CPPHM)
20.00
25.00
FIGURE 6-5. Scatter plot of predicted and observed station
peak ozone concentrations for 19 July 1979. The correlation
coefficient for the 22 pairs of station values is 0.478.
The solid line represents perfect agreement between observed
and predicted concentrations; the dotted lines represent the
envelope of the predicted values that are within a factor of
two of the corresponding observed values.
102
-------
Estimates of Gross Error
The normalized and nonnormalized gross errors are presented in Table
6-4. The average error for measured ozone concentrations greater than
5 pphm is 28.8 percent. The magnitude of the error is 2.4 pphm. The nor-
malized gross error is presented as function of the measured ozone concen-
trations in Figure 6-6b. The normalized gross error appears to decrease
as the measured ozone concentration increases.
Temporal Correlation
The temporal evolution of observed and predicted ozone concentrations at
16 selected monitoring stations is presented in Figure 6-7. The temporal
correlation coefficients are presented in Table 6-4. They indicate a
large variation among stations. The overall temporal correlation is
0.525, which is lower than that obtained for the 13 July 1979 simula-
tion. This lower temporal correlation may result from uncertainties in
the boundary concentrations for the 19 July 1979 simulation since the
ozone inflow through the Northeast and East boundaries strongly affects
the airshed ozone field. This pattern is characteristic of the 19 July
1979 simulation and is discussed in the next paragraph.
Spatial Correlation
Spatial isopleths of calculated hourly averaged ozone concentrations at
1400-1500 EST are presented in Figure 6-7. The ozone isopleths show three
centers of predicted high ozone buildup for that hour: one southwest of
Downington, Pennsylvania, another just southwest of Norristown, Pennsyl-
vania; and the third just west of Ancora, New Jersey. Because of the wind
flow through the airshed on this particular day, high ozone concentrations
were generally anticipated west of the Philadelphia urban center. The
ozone isopleths for 1400-1500 EST clearly show this pattern. The large
mass of ozone aligned in a north-south direction north of Vine!and, New
Jersey results from oxidant precursor material advected from the New York
City area into the Philadelphia region on 19 July. This inflow is speci-
fied in the model by the boundary concentrations for the Northeast and
East boundaries (see Section 5).
The average spatial correlation for ozone levels greater than or equal to
5 pphm is 0.202. This value is lower than the temporal correlation coef-
ficient. This feature has been discussed with the results of the 13 July
1979 simulation. The spatial correlation is lower than that obtained for
the 13 July 1979 simulation. This lower spatial correlation may result
from uncertainties in the Northeast and East boundary concentrations for
103
-------
12 15
18
21 24
2 1-0
tr
UJ
0 0.0
X
0
i -1.0
LJ
-?. a
i 1 i I i I i I i I i I i 1 i
_ 03
DflTfl POINTS 298
_ >R
V
KJ d **
^ ^_ ^
"-~-B_-B ET
—
—
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1
-
—
-
1
12
MEfiSURED CONCENTRRTION (PPHM)
(a)
2.0
1.0
-1.0
-2.0
2.0
o
1-4
£1.5
LlJ
I 1.0
O
12 15 18 21 24
to
m
cc
0.5
0.0
i T
03
DflTR POINTS 298
--o—
i I i I i
'13--.
I I I
j I
6 9 12 15 18
MEfiSURED CONCENTRflTION tPPHM)
21 24
2.0
1.5
1.0
0.5
0.0
(b)
-15
.9
...
LJ
cr
o
u_
o
>. c.i
CO
o
°" 0.S-
~i
C3__
» r* ' ** 7_4« I ^-.j
-*
•^
k
rj
1
-: J . 3
i
~i " • 2
'^ 1 t
V -, i
d3 -IS -12" -8-40 4 B 12 16 2cf "
(b)
FIGURE 6-6. Mean normalized bias, mean normalized error as function
of measured 03 concentrations and distribution of residuals for 19
July 1979.
104
-------
1S3M
OV7UJ
_ etc
Q on-
u
— eo-
in oa-
"3
CL
to
T5
C
OJ
C
O •
M i—
O 1/5
UJ
-a
ai o
> o
i. LD
01 ^t
01
r-
-a en
OJ ^H
-M
u >>
T3 ^
ai --3
10 C
i- O
3 •!-
O -M
J= (O
QJ ai
•t- JC
S_ 4->
CD
to s_
o
<1) 14-
•r- 0)
(— C
O
Nl
• O
ID O)
UJ U
or -i-
o a>
i— i S_
U. Q.
105
-------
1S3M
1 • I ' I ' I ' I
i
LJ 00. «
1 I I t ' I ' I I I I I I I I I i '
ILJ
40
•a
a>
T3
3
o
u
r-«.
i
UJ
a:
106
-------
the 19 July 1979 simulation since the ozone field appears to be sensitive
to these boundary concentrations and air quality data were insufficient to
determine these concentrations with accuracy.
Overall Correlation
The distribution of the residuals (measured ozone concentration minus
calculated ozone concentrations) is shown in Figure 6-6c. Most of the
residuals are within + 3 pphm, with a mean value of -0.6 pphm. A scatter
plot of the pairs of calculated and measured ozone concentrations is pre-
sented in Figure 6-8. Seventy-five percent of the predictions are within
a factor of 2 of the observations. The correlation coefficient is 0.769.
COMPARISON OF PHILADELPHIA RESULTS WITH
UAM PERFORMANCE IN OTHER CITIES
To provide perspective on how the Philadelphia simulation compares with
other recent model applications, Table 6-5 summarizes the results of simu-
lations for Los Angeles, Tulsa, Sacramento, and Denver. These data indi-
cate that the model generally tends to underpredict by from 1 to 15 per-
cent. The cases of overprediction are the Tulsa simulation of 2 Septem-
ber, the Los Angeles simulation of 26 June, and the Philadelphia simula-
tions of both 13 and 19 July 1979.
The Philadelphia simulations show gross errors of 34 and 31 percent for
the 13 and 19 July 1979 simulations, respectively. Other Airshed Model
simulations show gross errors ranging from 21 to 57 percent. Thus, model
performance for the Philadelphia simulations appears to be commensurate
with performance obtained in previous model applications.
In a study of the expected accuracy of photochemical air quality simula-
tion models (AQSM) arising from likely uncertainties in model inputs,
Seinfeld (1977) stated that "one is inclined to place an overall uncer-
tainty on oxidant level predictions from current AQSM of ± 50 percent."
The results of the application of the UAM to Philadelphia appear to be
consistent with this expected level of performance, which is the only
available stated "standard" of photochemical grid model performance.
107
-------
25.00
I I I I I I I I I I I. I I I I I I I I I I I I /
"5.130 10.00 15.00 20.00
OBSERVED OZONE CONCENTRRTION (PPHM)
25.00
FIGURE 6-8. Scatter plot of predicted and observed ozone
concentrations for 19 July 1979. The correlation coefficient
for the 445 pairs of 03 concentrations is 0.769. The solid
line represents perfect agreement between observed and pre-
dicted concentrations; the dotted lines represent the envelope
of the predicted values that are within a factor of two of the
corresponding observed values, and the heavy dashed line rep-
resents the regression line forced through the origin.
108
-------
«t
1C
._
3
H-
A
I/)
CU
p_
CU
en
c
^£
O
_J
c
cu
c
o
M
o
c_
o
<*-
c
o
••—
[ *
1C
3
.—
03
L. t_
(C CU
CL >
E C
O CU
0 O
*
• o
LO 4->
1 C
to cu
E
LU 03
__l t_
CO U
E
3
•o
M
••_
._*
*c
E
o
c
03
cu
s:
•o
ai
N
.»
'C
E
t_
o
c
*c
cu
s:
to
c
o
to
•t"-
t^
«o
("t
E
o
o
03
c
o
'£
1C
•f—
>
cu
a
cu
4-1
3
"o
oo
f*»
•^
CO
*~*l
c
o
•^
-M
*o
r—
3
j=
(.0
>*J
J^J
•^
CJ>
VO O Cn . r*^ to en
r*^ p-- P*» r** en c. p-^ •— i
en en en en ^^ cu en
<— 1 i— 1 t— 1 t— 1 JD i— 1 4J
4-> E W
eueu >i >v aj eu3
CC r- i— 3 4-J CO*
33 33C7>Q. 33
•OO 1^>'O3CU "~S *£
*f. to
vo r>» <-H en oo ^~
CMCM CMCMCOCM CMCM
^*>
a>
cu g
^C 03 03
to c_
10 r— O
O 3 1C
_l 1— OO
CM i— < I — LO
to en o co
CO CO
o r^ «s- •-*
LO LO CO CO
o o o o
en LO oo co
O «-H CM O
• • • •
o o o o
1 1
^O t«O C^ ^^
^^ ^^ ^^ ^^
o^ o^ o^ o^
>fc >fc >_ >fc
^.P. ^_ fWB ^_
33 3 3
^O '"^ '^ ^"D
^n oo ^^ ^^
CM CM *-* t-H
fQ 'O
•p~ *^
JC JC
CL CL
fB^ •—
CU CU CU
t- -o -a
CU 1C fC
^ •_ •_—
C 1- T~
cu jc jc
o Q- a.
to
c
o
•^
^_J
en
•o
c
(C
CJ
E
•o
c
1C
•t
T3
O
4^
cu
E
i
to
cu
4_>
C7^
C
•r—
10
3
T3
CU
1 *
3
0.
Q
O
CU
£_
cu
oo
eu
03
O
._
•_
C
eu
o
c
0
o
f
o
•o
cu
t_
3
to
03
^
eu
JC
4->
o
4_>
4-J
O
ai
r^
to
eu
t_
JC
4->
S
-o
a>
N
.fM
1—
fl
O
c
cu
t.
cu
s
cu
r*
4_)
O
4->
O
cu
a.
to
eu
t_
j^
i *
••«•>
5
eu
IM
••—
1C
E
L.
O
c
cu
(_
cu
>s
to
c
Q
.._
4->
(O
•r—
>
eu
-a
c
cu
^
cu
JC
4->
CU
^
cu
JC
91
CO
1
to
•a
c
t— •
^™
^c
•
c
o
•f—
4->
03
4->
C
CU
0
c
o
o
C*1
0
"O
cu
4-)
fC
0
••••
4->
cu
^
JC
4->
"E
>c
eu
_g"
^J
.d
4->
*r—
^
«•>
E
^~
CL
a.
CM
A
OO
eu
3
'(C
>
TJ
CU •
t- E
3 JC
00 CL
1 03 CL
CU
E to
JC A
4->
•r— 00
5 -t-
t/> *^-
C <4-
o o
CO 4->
•i- 3
t- 0
ra
CL CU
E J=
O 4->
O
cu
^ t-
o eu
*(- JC
3
eu
t_ «
t^_
C 0
c
•o o
CU t-
4^ 4^
t- CL
O CU
r^
CM
00
T-H *~^
CM
OO
• en
r— •— '
03 *•— ^
4-> •
CU •—
03
eu
JC 4->
u eu
to
Cu oo
h- T3
>> "o
J=> C
>^
-o cu
cu oi
00
to >,
3 J3
U
oo "^3
•r— CU
"O oo
I/)
CU 3
(_ O
1C OO
.•_
OO "O
4->
i— CU
3 t-
oo fc ^^
cu •—
c
>1 >»
C£
^5
CU >,
••—
t- T3
o et/
OO <^i
Cu T—
TJ «-
o
CU to
L_ CU
03 ^3
to CU
4J S-
3
CU 4->
*3
O oo
4-> CU
CU
e t-
U C
oo o
cu cu
JC JC
1— (—
D CU
109
-------
OZONE SENSITIVITY ANALYSIS
METHODOLOGY
After the base case simulations for 13 and 19 July were completed, a num-
ber of ozone sensitivity simulations were performed. These simulations
were carried out to examine the UAM's sensitivity to the principal factors
that might affect National Ambient Air Quality Standard (NAAQS) ozone
attainment control strategy development. In these sensitivity simulations
only hydrocarbon emissions and initial conditions were reduced; emissions
and initial conditions of NOX were left unchanged. Along with the
emission and initial condition reductions, simulations were performed to
test the sensitivity of predicted ozone levels to key inputs for the
simulation (e.g., boundary conditions and assumed background
concentrations). A total of 34 sensitivity simulations were performed
combining hydrocarbon emission reductions with various assumptions of
boundary conditions and background concentrations specified for
hydrocarbons and ozone. The list of simulations is presented in Table 7-
1. All simulations cover'the time period 0000-2000 EST.
Hydrocarbons were reduced in the low-level area source and elevated point
source files by multiplying all source emission rates by appropriate fac-
tors, resulting in a total reduction of hydrocarbon emissions by 25, 50,
or 75 percent. Initial condition concentrations were reduced in a differ-
ent manner. Reduced initial concentrations at stations with available
data were obtained by first subtracting a background total RHC value of
0.06 ppmC (Killus, 1982) from the base case initial concentrations, then
reducing the resulting value by the appropriate factor (25, 50, or 75
percent), and finally, by adding the background value of 0.06 ppmC to this
value to obtain a new initial value. Table 7-2 presents the initial con-
ditions for RHC used in the base case and the hydrocarbon reduction simu-
lations for 13 July. The observed value for Summit Bridge (0.05 ppmC) was
close to background initially and was left unchanged in the hydrocarbon
reduction simulations. Initial conditions in surface grid cells without
monitors were obtained by employing a Poisson interpolation. After this
surface field was computed, a vertical interpolation method was chosen
such that the background concentration at the top of the modeling region
(TOPCONC) was used in levels 3 and 4, and the level 2 value was obtained
by a linear interpolation between the surface (level 1) value and the
111
-------
Table 7-1. Simulation Designations for Philadelphia Airshed Sensitivity
Analysis for 13 and 19 July 1979
Simulation Name
Date
Description
D.BASE
D.25HC
D. 50HC
D.75HC
D.B50HC
D.BKHC
D.25HC.BKHC
D.50HC.BKHC
D.75HC.BKHC
D.BK03
D.25HC.BK03
D.50HC.BK03
D.75HC.BK03
D.BKHC. 03 .
D.25HC.BKHC.03
D.50HC.BKHC.03
D.75HC.BKHC.03
6. BASE
6.25HC
6.50HC
6.75HC
6.B25HC
6.B50HC
6.B75HC
6.BKHC
6.B25HC.BKHC
6.B50HC.BKHC
6.B75HC.BKHC
6.BK03
6.B25HC.BK03
6.B50HC.BK03
6.B75HC.BK03
6.BKHC.03
6.B25HC.BKHC.03
6.B50HC.BKHC.03
6.B75HC.BKHC.03
13 July
n
n
n
n
n
n
n
it
n
n
n
n
n
n
n
n
19 July
n
n
n
n
n
H
n
n
H
n
n
n
n
n
n
n
n
n
Base Case
Reduce HC's 25% in Emiss and Initial conditions
Reduce HC's 50% " " n " n
Reduce HC's 75% " " " " n
Reduce HC's 50% in Emiss, Initial, SE Boundary
Reduce Background HC's in Run D.BASE
Reduce Background HC's in Run D.25HC
" " " in Run D.50HC
" " "in Run D.75HC
Reduce Background 03 in Run D.BASE
Reduce Background 03 in Run D.25HC
" " " in Run D.50HC
" " "in Run D.75HC
Reduce Background HC's, 03 in Run D.BASE
Reduce Background HC's, 03 in Run D.25HC
" " " " in Run D.50HC
" " in Run D.75HC
Base case
Reduce HC's 25% in Emiss and Initial conditions
Reduce HC's 50% " n
Reduce HC's 75% n n "
Reduce HC's 25% in Emiss, Initial, NE Boundary
Reduce HC's 50% "" " "
Reduce HC's 75% "" " "
Reduce Background HC's in Run 6. BASE
Reduce Background HC's in Run 6.B25HC
" " " in Run 6.B50HC
" " " in Run 6.B75HC
Reduce Background 03 in Run 6. BASE
Reduce Background 03 in Run 6.B25HC
" " " in Run 6.B50HC
" in Run 6.B75HC
Reduce Background HC's and 03 in Run 6. BASE
Reduce Background HC's, 03 in Run 6.B25HC
" " " " in Run 6.B50HC
" " " " in Run 6.B75HC
112
-------
TABLE 7-2. Total RHC concentrations used as initial condi-
tions for base case and hydrocarbon reduction simulations for
13 July.
Station
AMS Lab
Chester
Lumberton
Norristown
South Broad
Summit Bridge
Base Case
RHC (ppmC)
1.05
2.10
0.2
0.39
0.30
0.05
25 Percent
Reduction
RHC (ppmC)
0.803
1.59
0.165
0.308
0.24
0.05
50 Percent
Reduction
RHC (ppmC)
0.555
1.08
0.13
0.225
0.180
0.050
75 Percent
Reduction
RHC (ppmC)
0.308
0.570
0.095
O.M25
0.120
0.050
113
-------
level-3 value. This method was used to initialize all grid cells in all
levels with appropriate concentrations for all species before commencement
of the ozone sensitivity simulation. Initial concentrations for 19 July
were obtained using the same procedure as that for 13 July. These values
are presented in Table 7-3.
For 13 July, the base-case emissions and initial conditions were reduced
by 25, 50, and 75 percent, respectively, for three simulations (D.25HC,
D.50HC, and D.75HC). The analysis was extended by examining the sensi-
tivity of predicted ozone to reductions in background hydrocarbons and
background ozone in separate simulations and by reducing both background
hydrocarbon and ozone in the same simulations along with the reduced
hydrocarbon emissions and initial conditions. Only one simulation
involving a reduction of the Southeast boundary hydrocarbon concentrations
was performed for 13 July. Differences of less than 1 ppb for predicted
ozone were found for this simulation (D.B50HC). Because of this result
and the fact that the designated boundary hydrocarbon concentrations along
the Southeast boundary were already close to background when inflow from
this boundary occurred, all subsequent sensitivity simulations for 13 July
used combinations of reduced emissions, initial conditions, and back-
ground, but did not reduce concentrations along the Southeast boundary.
For 19 July, ozone precursors from the New Jersey/New York urban area that
were transported through the Northeast and East boundaries mixed with the
local emissions to produce high levels of ozone. Hydrocarbon emissions
and initial conditions were reduced in three simulations (6.25HC, 6.50HC,
and 6.75HC). In all subsequent hydrocarbon reduction simulations for 19
July, boundary conditions on the Northeast and East boundaries were also
reduced by 25, 50, or 75 percent. These simulations were performed to
test the sensitivity of ozone by control of hydrocarbons in the local
airshed (Philadelphia) and a neighboring airshed (New Jersey/New York
urban area). Values of reduced hydrocarbons for the Northeast and East
boundaries were obtained in a manner similar to that used to obtain the
reduced initial conditions. Background concentrations of the speciated
hydrocarbons (see Table 5-3) were subtracted from the base-case hourly
hydrocarbon boundary value (Table 5-5). This intermediate value was then
reduced by 25, 50, or 75 percent, and the background was added, resulting
in a new hourly value for the boundary concentration. Boundary conditions
for hydrocarbons already below background were left unchanged.
Background values for hydrocarbons and ozone were reduced in a number of
the sensitivity simulations. These background values were designated for
the top of the modeling region, as initial conditions above the mixing
height, and for all boundaries, except the Southeast boundary for 13
July and the Northeast and East boundaries for 19 July below the mixing
114
-------
TABLE 7-3. Total RHC concentrations used as initial condi-
tions for base case and hydrocarbon reduction simulations for
19 July.
25 Percent 50 Percent 75 Percent
Base Case Reduction Reduction Reduction
Station RHC (ppmC) RHC (ppmC) RHC (ppmC) RHC (ppmC)
AMS Lab
Chester
Downington
Lumberton
Norristown
South Broad
Summit Bridge
0.35
0.95
0.05
0.4
0.29
0.25
0.1
0.2775
0.7275
0.05
0.315
0.2325
0.2025
0.09
0.205
0.505
0.05
0.23
0.175
0.155
0.08
0.1325
0.2825
0.05
0.145
0.1175
0.1075
0.07
115
-------
height. These values are presented in Table 7-4. The reduced RHC back-
ground values for both days represent a reduction of nearly 50 percent
from the base-case values. For 13 July, the background ozone value was
reduced 50 percent, and for 19 July, the background ozone was reduced 33
percent.
RESULTS OF OZONE SENSITIVITY SIMULATIONS FOR 13 JULY
A total of 16 ozone sensitivity simulations were performed for 13 July
(Table 7-1). Ozone response in all simulations, compared to base-case
predictions, was obtained by examining both the predicted maximum over the
entire grid for the simulation day and the predicted distance-weighted
average maximum value at the two highest monitors. Both peak grid and
peak station maxima were examined because previous evaluations of the UAM
(EPA, 1983b) have shown that when hydrocarbon emissions are reduced, the
daily maximum ozone predictions may migrate and "... emission control
requirements should not be based on ozone predictions constrained to a
particular monitoring site" (Layland and Cole, 1983). The peak predicted
grid value is the actual peak value obtained in a particular grid cell,
whereas the maximum at a station monitor is a distance-weighted average
value of the four closest grid cell values. Table 7-5 presents the hourly
predicted maximum ozone for the base case and all sensitivity simulations
for the Philadelphia urban plume peak (located north of urban center), the
Roxy Water monitor, and the Norristown monitor. Also included in the
table are the relative and total ozone reductions for each sensitivity
case. The relative reductions were obtained by comparing the emission
reduction simulation with the corresponding simulation without reduced
emissions (e.g., compare D.BKHC with D.50HC.BKHC). The total ozone reduc-
tion value was obtained by comparing sensitivity simulations to the base
case (D.BASE).
The values in Table 7-5 are presented graphically in two ways. The first
type of graph compares the percent hydrocarbon emission reduction with the
peak predicted hourly average ozone value, while the second type compares
the percent of hydrocarbon emission reduction with either the total or
relative percent of ozone reduction. Also depicted on the former graph is
the level of the NAAQS (12 pphm) for ozone.
Figure 7-1 presents a comparison of the peak predicted hourly average
ozone of the Philadelphia urban plume for all sensitivity simulations
The distance weighted average concentration is obtained by computing the
distance weighted average of the concentrations at the centroids of the
four grid cells closest to the station monitor in question.
116
-------
TABLE 7-4. Reduced background
concentrations used in the
ozone sensitivity simula-
tions of 13 and 19 July.
Concentration
Species (ppm)
ETH 0.0
OLE 0.0
PAR 0.03
CAR8 0.003
ARO 0.0
CO 0.1
03 0.04
117
-------
<+-
o
w>
C
o
•F-
^>
HJ
r«_
0)
>^
-I--*
^>
•^
t/)
c
^
O
<4-
in
C
o
'£
o
3
•o
2i
cu
c
o
IM
o
TJ
C
,
"71
3
O
•
in
i •
f** >•>
LU 3
1 ^K
~™ «J
CO
^C co
•^
O-
«t
c
o
in
• fB»
<_
t_
o
TS-
^
Q.
A
J^
CU
4-i
*o
^*g
X
0
"no
CU
'°"
cu
E
3
f^
C
(0
t_
™*^
(0
•f*
.c
Q.
ID
"w
^»
•rv
t~
Q-
c •->
O •(->
cu -i- c
C 4-> CU
o o o
M 3 «-
o -a cu
CU Q.
fV* ^— *
c
o
•f"
CU •!->
0 i-
tsl T3
O CU
t_
0.
C ^""^
o +->
CU T- C
C j-> 01
o o u
N 3 t-
o -c cu
CU CL
o: — •
c
o
*r»
CU -M
^— t*
O f
N -a
0 CU
t_
o.
c '-—
O 4->
CU i- C
C •(-> CU
O 0 0
N 3 (.
O T3 CU
cu a.
0£ — •
c
o
CU 4->
c o
o ••-
NJ -a
o cu
t,^
Q.
*
0
h-
cu
*f~
4^>
rd
fa>
O£
^"^.
e
a.
a.
^ -
,_
(O
o
1—
cu
>
•1—
4_5
(Q
,_
CL
p^
^— m.
E
a.
a.
r—
(B
O
1—
CU
•f™
-[--I
m
^_
cu
fy
*£"
a.
CL
o
i *
1 •••! •••{ •••) •••
CTll^CTl ft Id QO r-HCOCO CX3^fr-(
t— i oo ^ CM<-UO CM^-UO CMir>r^
covoor^-o«-ioor«^criOCMir)^-CMV£><— <
c^LncMO>vocMcovor^<3-oi>^«3-ovc<4-
»-* T-1* i-^ 1—4 1-^ i— 4 f— * •— 1 T— 4 f— 4
co«3-i^vo«a-if>cncoc\ji^.i»~cy>cocTi'— "
CM ^^ LO ***4 ^^ Lf) ^O CO ^^ VO CVJ LO VO ^"^
oo ^f f^^ *"^ r** to ^™* t~~* co f™** i-H *"H
t •••! •••! •••! •••
LO CO CO CTt CTi CNJ 00 VO CT* *»1" CO r*^«
CM ^J^ LO CM ^f VO CM ^J^ LO CO LO *^
cocoioLocyicooocyicy>r«."!rLOLoeooo
Cn*S-^-iCTiLOi— iOOVOr-^CMCT>l~-*l-CT>LO*3-
«— 4 f— 1 i-H «— 1 f— 4 1—4 i— 1 i-H
CM ^J^ LO f™4 CO LO LO CM LO LO •••* ^" LO ^*
LO r1^* ^" CM ^^ 9** * LO C\J ^tfr CT^ CM ^"
1 •••) •••! •••! •••
CM ^f" rO CTi LO CO ^^" C7^ O^ CNJ CM CO
CM^LO CVJLOLO CM^-LO COLOf^-
voior~.^-coiO'a-LjOLocoo^-CMcri^rcT>
LOO'S-CMCOLOOOOlOCncOOCM'tJ-COLO
CM CM .— It— 1 CM t— 4 «— 4 CM >— 1 i— 1 i— 4 CM »— '
CO CO CO
o o o
CJOO COCOCO OOO
0:0:0: ooo 0:0:0:
CO CQ CD COOQOOOCQCOCO
UJOOOOOOOCOOOOCJOOO
000:0:0:0:0:0:0:00:0:0:0:0:0:0:
cocMLor^coeMLOi-.cQCMLOr-~coCMior->
aoooaooooaoooooa
118
-------
I I
BRSE CflSE BRCKGROUND
REDUCED BflCKGROUND 03
REDUCED BRCKGROUND HC
REDUCED BflCKGROUND HC/03
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-1. Predicted ozone response to hydrocarbon emission reductions
for peak regional ozone in the Philadelphia urban plume for 13 July.
119
-------
with the percent of hydrocarbon emission reduction. Similar graphs are
presented in Figures 7-2 and 7-3 for the station-monitors of Roxy Water
and Norristown (both located in Pennsylvania, northwest of the urban cen-
ter of Philadelphia), respectively. Figures 7-4, 7-5, and 7-6 present
response curves of relative ozone reductions for the peak of the urban
plume, the Roxy Water, and Norristown monitors, respectively. Figures
7-7, 7-8, and 7-9 present response curves of total ozone reduction for the
urban plume, Roxy Water, and Norristown monitors, respectively.
All of the figures contain information that might guide the possible
formulation of broad control requirements for attaining the ozone NAAQS;
however, since specific source categories were not addressed (e.g., mobile
and stationary sources), the results cannot be used directly to formulate
an ozone NAAQS attainment strategy. Figure 7-1, for example, shows that
the peak predicted value in the base case (26.6 pphm) would not meet
ambient standards for ozone even with a 75 percent reduction in hydro-
carbon emissions, assuming no change in background levels for hydrocarbons
and ozone. If, however, background ozone levels are reduced by 50 per-
cent, the standard can be met with only a 58 percent reduction of hydro-
carbon emissions. If hydrocarbon background levels are reduced by 50
percent, with ozone background unchanged, then modeling analysis shows
that the ozone standard will be met with only a 43 percent hydrocarbon
emission reduction. If background hydrocarbon and ozone levels are
reduced by 50 percent, Figure 7-1 reveals that ambient ozone levels will
meet the national standard with only a 35 percent reduction in hydrocarbon
emissions. The percent of hydrocarbon emission reductions needed to reach
the standard at the Roxy Water and Norristown monitors is lower for the
four cases as shown in Figures 7-2 and 7-3, respectively. All emission
reduction values required to meet the ozone standard for all sensitivity
simulations of 13 July are summarized in Table T-6. This table shows a
wide spectrum of control requirements both in comparison of the regional
peak with the highest station predictions, and in comparisons of the vari-
ous assumptions for background hydrocarbons or ozone. Using the values of
the regional peak is a conservative approach for estimating control
requirements because the peak predicted ozone (26.6 pphm) is more than
30 percent greater than that observed during the entire summer field pro-
gram. Control requirements at the two highest monitors are very similar
for the four sets of sensitivity simulations, with nearly 50 percent
hydrocarbon control needed, assuming no change in background concentra-
tions, and only 14 percent control needed if background for both ozone
and hydrocarbon is decreased by 50 percent.
Spatial patterns of predicted ozone can be illustrated with the use of
Deficit/Enhancement (D/E) plots which show the area! extent of the
decrease or increase for ozone in the Philadelphia airshed for each sen-
sitivity simulation compared to the base case simulation. Ozone decreases
120
-------
BRSE CRSE BflCKGROUND
REDUCED BflCKGROUND 03
REDUCED BflCKGROUND HC
REDUCED BflCKGROUND HC/03
25 50 75
PERCENT HYDROCARBON EMISSION REDUCTION
100
FIGURE 7-2. Predicted ozone response to hydrocarbon emission reductions
for 13 July at the Roxy Water, PA monitor.
121
-------
i | i
BflSE CflSE BRCKGROUND
REDUCED BRCKGROUND 03
REDUCED BRCKGROUND HC
REDUCED BRCKGROUND HC/03
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
FIGURE 7-3. Predicted ozone response to hydrocarbon emission reductions
for 13 July at the Norristown, PA monitor.
122
-------
100
u
Z)
Ci
LU
UJ
z
D
M
a
u
ctr
Ld
BflSE CflSE BflCKGROUND
REDUCED BflCKGROUND 03
REDUCED BRCKGROUND HC
REDUCED BflCKGROUND HC/03
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
100
FIGURE 7-4. Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the Philadelphia.urban
plume for 13 July.
123
-------
100
o
t—I
t-
u
UJ
LJ
Z
o
D
UJ
LJ
'
BflSE CflSE BRCKGROUND
REDUCED BRCKGROUND 03
REDUCED BRCKGROUND HC
REDUCED BRCKGROUND HC/03
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
100
FIGURE 7-5. Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Roxy Water, PA monitor.
124
-------
100
u
ZD
O
UJ
UJ
•z.
o
M
O
LJ
U
(£
UJ
Q_
BRSE CRSE BflCKGROUND
REDUCED BRCKGROUND 03
REDUCED BflCKGROUND HC
REDUCED BflCKGROUND HC/03
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
100
FIGURE 7-6. Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Norristown, PA monitor.
125
-------
1021
LJ
3
O
UJ
ce
Ld
Z
o
M
O
UJ
u
EC
LJ
D_
BflSE CflSE BRCKGROUND
REDUCED BRCKGROUND 03
REDUCED BRCKGROUND HC
REDUCED BRCKGROUND HC/03
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-7. Total ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the Philadelphia urban
plume for 13 July.
126
-------
100
o
I—1
I—
u
Z3
LU
•z.
O
o
UJ
U
01
LJ
0_
I | I I I I
BflSE CRSE BRCKGROUND
REDUCED BRCKGROUND 03
REDUCED BflCKGROUND HC
REDUCED BRCKGROUND HC/03
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-8. Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Roxy Water, PA monitor.
127
-------
100
o
I—f
H-
LJ
LJ
LJ
Z
O
tvj
O
UJ
LJ
cc:
u
D_
BRSE CRSE BflCKGROUND
REDUCED BflCKGROUND 03
REDUCED BflCKGROUND HC
REDUCED BflCKGROUND HC/03
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-9. Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 13 July at the Norristown, PA monitor.
128
-------
TABLE 7-6. Hydrocarbon emission reductions required to meet the NAAQS
for ozone from the sensitivity simulations of 13 July.
Percent Hydrocarbon Emission Reduction
Sensitivity Simulation Regional Peak Roxy Water Norristown
Base case background > 75 45 50
Reduced background ozone 58 32 38
Reduced background hydrocarbons 43 21 27
Reduced background ozone and 35 13 14
hydrocarbons
129
-------
or increases can be depicted hourly, or the maximum D/E can be shown with
one plot summarizing the maximum hourly increase-or decrease for each grid
cell over the entire simulation day. In this analysis, the maximum D/E
patterns are presented for a number of the sensitivity simulations in
Figures 7-10 through 7-21, as summarized in Table 7-7. Some of the
Deficit/Enhancement plots show comparisons with the overall base case
simulation (D.BASE), whereas others show comparisons with the correspond-
ing simulation without hydrocarbon emission reduction (e.g., compare
D.50HC.BKHC with D.BKHC). All the Deficit/Enhancement plots presented
show decreases in predicted ozone. This is due to the fact that in the
sensitivity simulations, combinations of urban emissions, background
concentrations, and boundary conditions were reduced.
Figures 7-10 through 7-12 depict areal changes in ozone that resulted from
reducing hydrocarbon emissions without changing background hydrocarbon or
ozone levels. The largest decrease in ozone was 16 pphm (located in the
center of the peak in the base case) with a 75 percent decrease in hydro-
carbon emissions.
If no controls were specified for urban hydrocarbon emissions, the largest
reduction in ozone would be 3 pphm if background ozone levels were halved
(Figure 7-13). Similarly, a reduction of 4 pphm ozone over limited areas
was obtained in the simulation by decreasing only background hydrocarbons
(Figure 7-14). By decreasing both background ozone and hydrocarbons, the
simulation revealed large areas to the north and west of the urban center
where decreases of 5 pphm were calculated to occur (Figure 7-15). Figures
7-16 through 7-18 show comparisons with the base case of combinations of
50 percent hydrocarbon emission reduction and various assumptions for
background hydrocarbon and ozone. Figure 7-18 compares the sensitivity
simulation of 50 percent emission reduction and 50 percent decrease in
both background ozone and hydrocarbon with -.he base case. This figure
shows a decrease in predicted ozone of 18 ppnm in an area north of the
urban center. Figures 7-19 through 7-21 show relative differences between
simulations with 50 percent reduction in hydrocarbon emissions and those
with no emission reductions. The magnitude and areal extent of these
patterns are, as expected, similar to the pattern depicted in Figure 7-11
where emissions were also reduced 50 percent.
RESULTS OF OZONE SENSITIVITY SIMULATIONS FOR 19 JULY
For the meteorological conditions of 19 July, a total of 18 ozone sensi-
tivity simulations were performed (Table 7-1). As was done for 13 July,
both the peak regional predicted ozone and the peak station ozone are
examined. Because 19 July was a transport day with inflow of fresh pre-
cursors from the New Jersey/New York urban area, the peak predicted ozone
130
-------
TABLE 7-7. Deficit/enhancement Figures for ozone for the 13 July
sensitivity simulations.
Figure Sensitivity Simulation minus Base Case Simulation
7-10
7-11
7-12
7-13
7-14
7-15
7-16
7-17
7-18
7-19
7-20
7-21
D.25HC
D.50HC
D.75HC
D.BK03
D.BKHC
D.BKHC.03
D.50HC.BK03
D.50HC.BKHC
D.50HC.BKHC.03
D.50HC.BK03
D.50HC.BKHC
D.50HC.BKHC.03
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
D.BASE
D.BASE
D.BASE
D.BASE
D.BASE
D.BASE
D.BASE
D.BASE
D.BASE
D.BK03
D.BKHC
D.BKHC.03
131
-------
NORTH
0
10
20
30
30
10
0,
0> :
ROBB
%&8mx&•mZ-Atm;. DOWN ••
.-•'..-•••• -LUMB
:.../ ; criks ••' V.
RNCO
SUMM
VflN--'
30
20
tn
a:
10
0
10
20
30
SOUTH
FIGURE 7-10. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.25HC minus D.BASE].
132
-------
NORTH
0
10
20
30
•w*7
30
T::::::
i£:::
rr;:::
•i'-"'
20
• : : • • : :
•: I NORR^..
'
'- .•'.'•••'••'•''••''•Bk3''
'•- CHES'
•CLftT .-'
'••['A"--"
flNCO
10
SUHM
?:':
K<-
0,
>:jj:;:':;!|:::':::y: I'-trxY^^ijyiSifiy^xii::::::-!;;:;:;-);;^ : -:!t: ' :> :;:-:|:-:;!-:t:: :;>•:' ::-i: :'i '-f'-: ::
30
20
tn
a:
LLJ
10
10
20
30
SOUTH
FIGURE 7-11. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC minus D.BASE].
133
-------
NORTH
0
30 i
20^
10*:
20
30
pXi:j:|x|:jxjx|:jxjxj:x>X;::X;yxX^;X:X:X|:j:|x|:v:x:jx: |x:
S:';::::::x-:.:::X:-:X:X::x:::x:::::::x-:::::::::::;::x::;:;:;::::':'':-x::- ••:•'
lllli;llll;li:lll-|H|::|:':!:i;'; ••' /
•^:::::j:jxj:'x|:jxjx':::-:jxx:;xWx-X;'f .' .•• — .•• ""
SxlS;S::;sS|ii|Sl;SiCi:::;:x; E>OHN ;
( '''••-"
iili|si:i§;i|?:m:P;:i::l s U MM
0 10
'.•.•X'.'X*x-"'"'''X'.'; ;•;• '* * •"';•'• •••••-••;• ••••••• "•" •'•''* '•I'X-.-.'X-x-'v'XvXvX'X'' '•'. -.•.•x- •
/ T; / s. <•'''•''.•'' /. -is "'• '''^i •'.-•y/l ^'H:"i-
; ^;-ii-:i:^ ;,•'' /i '•.v-.;-.'.:T:REN'i ROBE ;-
/jlWM j-l:^:^::^:U {
r^r^|S5§£:::::::::'
^ISSf';;;X
fiNCO .....,.,,,,;:;;::$:
:.x,.x.:vll|-||l^illlll::ll:-?
20 30
SOUTH
: VflN
wj&M:' :'
._
~
.'.-
,:r
;r
'-
~
-30
:-20
to
cr
^10
FIGURE 7-12. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.75HC minus D.BASE].
134
-------
10
NORTH
20
30
1J.J II »>q tT T^HT I 3 I 1 J t »
20
UJ
10
0,
> t y. i <
0
DOHN
CHES
CLflY
.. SUHM
•-3-
> t < < ± t .1 .t i < t
30
TREN ROBB
NORR BRIS
COIt&XY
SHEOHS
VfiN
20
tn
cr
RNCO
10
BRIG
i ± i t ..t t t j i
30
SOUTH
FIGURE 7-13. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.BK03 minus D.BASE].
135
-------
NORTH
0
10
20
30
30
20
10
PI
•C-Xv ''•'•'.•'•. •' .•'. X'X'.' ' '.•'.•'.''.•'.•'.
jT;Xv'*X '.•'..' • X'X-X'"' .'.',', '•'
'*.•' .''. '.•'•'.•'.•'•'. '•'• •'• *XvX •'''.•'.•'•'.•
^. '.•.•, . .•.• • .' • .-.•, . . .• •.• • .-.
Iv.'X X"X-". ' '.''.•', '. '.•'.•',•'.''.''.''.'.'',''.''.''.•'
*!*'.''.•'.• '.•'.•'.• X*.v'.'* '•'•'•'• '.• .''.-'"'. '. ' '.•',•'.•
IjX'X'X'X'* ,'.'.'. •'•'•',•'.''.''.''•'''.''''.''. '
ifrxvx'x'x-l '. ', '. X'X'X'X'X'X-X
vX'X'X'X' **','''' '.'•''. •'•!•! '',''.''. X*X'X
|§p;;ip||||;;l|
Pliljxi^iiipif-ii
i*||lli:'y:::;;;||i;!:;;:i'|
fe-:-:?:-:?:^'':^'':^:::^:^:1^^:?:
iiiS:i:;:.i|:;xi. ... "'::.::>:|:i:|:: :• :Vf
feSSvi'S- '. -': $:£y -i'"'-':'-::-
Si-SSxX;?:- :•!• ::$S5:i'5:S
Illlllll !|i |:|i |; |j. j: | p! illlll ||::-:;:|;:||||: : ; % ?j$j£ j- !'.i J' '! ^ i'^
x-:* ix^l? |;|. t. W^:^ ;i| jiS ?' - • ^W':':'i;i;i; |;||s;x:-;:;p:;:;i::i -^ M •• :x:;i ;-::i: 4;'v-:S-' : :v;--
ijix^iiriii'SiSHv:- ::;:::;:X'::^:":::::-': ' .-• '''•••^^&?ffi-$tMi 'M ,:?;; "^Vv-x '':::x
Slil>! !F /• '::;> v---/ / C'F^^|f^;f : f ;
:V'Y;:'::'>;;;:S::;-': . •''' \ / i TREN ROBB " 'V: "'.: ' '• ': •
•'.•'•'.•'.•'.•I' . Kl'nPD DD T G • ' ' \/ Ql
••:••• .-:•: .• NUKK BRI.5 •• VHN .
" X'X'.'I " 'l^CN ^- ^ "'
1 ;'DoNf/ f '-.-s/-..-^..,-;;' C^ j -••• ' ;
ii /'""' '••' /' -ollPW-''' '.-••••-v'" :
JvJ: ; .-' ISfcB .-•': :
H; N '•••••-.. CHES ••' '••••'' / '•
;:S : "CLRY .-' /"•••••••• " .;...:.;:
i; / .-' ..••'' RNCO .;.. :.x,.*;;i; v' ::::$;-:':: •
:i:i:i ."' ••'' .:•.•.•.•.•.:•. =S : x&'S: :^: ^: '. •: X :
•S: '•••••SUMH :::::::;::x:::|||:|5£:^ :;' '£ £
M
-30
•*
-;
-za
-
'*•
~
*
•±-
- 10
-
:
JPI
cn
cr
SOUTH
FIGURE 7-14. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.BKHC minus D.BASE].
136
-------
NORTH
0 10 20 30
xxfe :'•• f:•:•>:•.-••: •X.--:-&:*.'yf-&'\-\-T:&f.-s:-:f :• .*-:':':!:.-: .:y.'-::::»:::::-i:'::::*.-:: ::J x-:;!::'.- -1
30
TREN .--'.ROBB "***&
... ....{R '':. : BR.IS
20 : CO
NORR .'v, : BRIS *"-.
SHBfflMS5''
/ -ISU
CHES/ ..--3-''
CLftY .' .- '
10
RNCO
.
:;pP*i(S.::;;::::::::;: x^
pi t:v::t:v::t:;:;^TX;TS-:-:iv:;:±:v::t:v:;f;.:.:t:!:!.;Ax:!:>!:v:^^ V-'
30
20
CD
cr
LU
10
0 10 20 30
SOUTH
FIGURE 7-15. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.BKHC.03 minus D.BASE].
137
-------
NORTH
0
10
20
30
20
in
UJ
10
?
*•
0.
0
DOWN:
-»•...
'• w
-.;-..CLOY
-5UMH
--3-
30
t t i i .. r.. j... H
< < I- * I I r t i
flNCO
10
30
SOUTH
30
VRN-:
20
to
CE
LU
10
^J0
FIGURE 7-16. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BK03 minus D.BASE].
138
-------
0
T-X;
•fe;
20
10
10
•...•SUHM
NORTH
20
30
NORR
flNCO
a
^ f !:::::ii:>>::i:$xif!:;::!
30
20
in
cr
[0
0
10
20
30
: I ri
SOUTH
FIGURE 7-17. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BKHC minus D.BASE].
139
-------
NORTH
0
10
20
30
^
30
20
en
10
0
0
:::::;:::::::SSSO:''"''',."••' ••' =::\ j \5...'.
30
'VON
>
: ..>'
SilMM j
flNCO
20
en
cr
UJ
10
t . j. j
10
30
SOUTH
FIGURE 7-18. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BKHC.03 minus D.BASE].
140
-------
0
301
201
101
10
SUHH
I
NORTH
20
30
:j g . 'Sx ';:•: * '•<•'• ''•"" / •'.-• k '.-'.' / . V . •.•••':::.:-::.:::::::.: x ::
11; i^:::;|?^-;// ////^/^-Cl .^-,'^V.-. :"i>SHl^
.:; ••:.\ TN- . ROBB
-LUMB
.
'••-••."•.'••... CH£S-
' •
RNCO
•x^x^Sf^'J'^-:^-^-^^^:-^^-''^' -: :t::::::-i: :::-i'- >::!''-
30
20
en
cr
10
"0
10
20
30
SOUTH
FIGURE 7-19. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BK03 minus D.BK03].
141
-------
NORTH
0
20
30
-f^ .i.V-i
30 m
30
10
POHN \
SUMM
v=
. •. \ : 7REN-.
RNCO
0
3 I { t •» t t 1 t t ...jr.. I t 1 1 •xfyxi-x-'lix
ySife'lT!:!:-]:::-:^^ I
20
to
cc
10
0
10
20
30
SOUTH
FIGURE 7-20. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BKHC minus D.BKHC].
142
-------
0
NORTH
20
30
11 i i*I » »
30
30
20
10
i
en
-DOHN
SUHH
£ 'TREN ROBB
LUMB
flNCO
20
to
cc
10
BRIG
> ...I...T...I J I t 1 t t. .I....1...*...*..J lit 1 > > 1 1 f___± I
I I t f... j . I
0
10 20
SOUTH
30
FIGURE 7-21. Maximum deficit/enhancement for ozone (pphm) for all
hours for 13 July [D.50HC.BKHC.03 minus D.BKHC.03].
143
-------
at a station affected by this interurban transport is also examined. As
is shown in the base case for this day (Figure 6-7), the simulated urban
ozone plume from Philadelphia was transported nearly 90 km west of the
urban center. As a result, the station of Downington, Pennsylvania
received the highest predicted ozone. In the simulation, this station was
affected almost exclusively by urban Philadelphia emissions, whereas all
other stations were affected to a large degree by the New Jersey/New York
plume. Of those stations affected by this plume, the one for which maxi-
mum ozone was predicted was the Roxy Water, Pennsylvania monitor. Table
7-8 presents the hourly predicted maximum for the base case and all sensi-
tivity simulations for the Philadelphia urban plume (located west of Down-
ington, Pennsylvania), the Downington monitor, and the Roxy Water moni-
tor. Also included in the table are the relative and total ozone reduc-
tions for each sensitivity simulation. Because 19 July experienced inter-
urban transport of ozone precursors, all but three of the sensitivity
simulations (6.25HC, 6.50HC, and 6.75HC) resulted in reductions in
boundary condition hydrocarbons on the Northeast and East boundaries by
25, 50, and 75 percent, along with the corresponding reduction in hydro-
carbon emissions and initial conditions.
For 13 July, the information contained in Table 7-8 is presented graphi-
cally for ease of interpretation.
Figure 7-22 compares, for all sensitivity simulations, peak predicted
hourly averaged ozone from the Philadelphia urban plume with the percent
of hydrocarbon emissions reduction. Similar figures are presented for the
Downington and Roxy Water monitors in Figures 7-23 and 7-24, respec-
tively. Figures 7-25 through 7-27 present relative ozone response, and
Figures 7-28 through 7-30 present total ozone response for the Philadel-
phia urban olume, the Downington monitor, a.id the Roxy Water monitor,
respectively.
Figure 7-22 shows the effects on ozone of possible hydrocarbon and ozone
reduction scenarios, both from the Philadelphia emission region and from
the neighboring emission area of New Jersey and New York. This figure
shows that the peak regional value located west of Philadelphia is also
influenced by transport through the boundaries to the northeast. This is
evident in a comparison of the curve representing the no-change boundary
conditions with the curve for the simulation in which the Northeast and
East boundary hydrocarbons are decreased by 25, 50, or 75 percent concur-
rently with emissions and initial conditions. In these simulations, back-
ground hydrocarbon and ozone were left unchanged. In the no-change
boundary condition emission reduction simulations, a value of 57 percent
hydrocarbon reduction is needed to meet the standard for ozone, but with
Northeast and East boundary conditions reduced, only 44 percent hydrocar-
bon reduction is required. This difference is an important consideration
144
-------
c
o
nj
3
•r—
10
C
at
at
o
N
o
at
t_
o
(A
C
o
o
3
-o
a>
c
o
N
O
-a
c
•a
o
N
o
x
T3
TJ
at
o
•?—
•a
3
O
CO
UJ 3
_J O
CO
^ c^
ai
c -—-
O -P
>f- C
c 4-> at
ooo
Nl 3 S_
O T3 OI
at o.
at
o
at Z> *£*
o •»— ex
Nl "O CX
o at *—
c *-»
O 4->
•r- C
4J Ot
O O
o -a
at
a> ex
a
c
o
at *J E
o •<- o.
N -a ex
o at —
c.
CL.
C <-*
O 4->
O) •— C
CT -f.-* ^
o o o at
M 3 t-
o -o at
at a.
at
at 4-> E
c o .c
O ••- Q.
N ~a cx
o at •—
o
ID
3
E
to
vo vo 10
CM CO
o •— i ro
«M
o
CSJ
c*
CM
V£>
CO
O
CO
LO
CVJ
IT)
CM
CM
*±
CO
CM
•*
Lf)
COLOVO
<— i evi PO T-H co
cvt
cvj
CM CO «-* CO
oooocr>
ooo
000
• • •
000
co co ro
ooo
CO CO CO
ooo
ooo
re =c a:
CO CO CO ^O CO CO CO
• • • o • • •
CJ O O • O O O
GOcvjinr^cQcococococacQcocacocQcacococo
145
-------
28
g-20
u
Ci
UJ
cr
UJ
D_
8
'*., *X
'-.X
""-., X
'"-
No change 1n background;
Mo Chinee 1n HE and E boundary
hydrocarbons
No change 1n background;
Reduced HI and t boundary
hydrocarbons
Reduced background ozone;
Reduced K and E boundary
hydrocarbons
Reduced background hydrocarbons;
Reduced HE and E boundary
hydrocarbons
Reduced background hydrocarbons
and ozone. Reduced N£ and E
boundary hydrocarbons
1 f A
—-^
NRRQ5
I
J
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
IBS
FIGURE 7-22. Predicted ozone response to hydrocarbon emission
reductions for peak regional ozone in the Philadelphia urban
plume for 19 July.
146
-------
28
S-20
o
M
O
16
0,
1 i T
I T
I I 1 I I
to change 1n background;
No change 1n K and E boundary
hydrocarbons
No change In background;
Reduced HE and E boundary
hydrocarbons
Reduced background ozone;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons
•nd ozone; Reduced NE and E
boundary hydrocarbons
I I
NRflQS
I I
J I
I
0
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
100
FIGURE 7-23. Predicted ozone response to hydrocarbon emission
reductions for 19 July at the Downington, PA monitor.
147
-------
TO _
£.0 i i i i | i ~i ' i i j i i r " r
No change 1n backgroun
No change 1n NE and E boundary
hydrocarbons
No change in background;
Reduced HE and E boundary
21 — • hydrocarbons
Reduced background ozone;
Reduced Hi and E boundary
hydrocarbons
^ Reduced background hydrocarbons;
Q_ _ Reduced HE and I
Q_ 20 hydrocarbons
Reduced background hydrocarbons
LlJ and ozone; Deduced K. ane E
boundary hydrocarbons
O
161-
^"
O
0 *>' NflflQS
UJ 1 ,
i—
u
o
UJ
Q_
CE
UJ _
T.1,
.,..
4)_ ", A
I I
I
I
I
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-24. Predicted ozone response to hydrocarbon emission
reductions for 19 July at the Roxy Water, PA monitor.
148
-------
100
LJ
13
O
LJ
DC:
LJ
Z
D
r*j
o
LJ
U
Q£
LJ
0.
75
I I I I ] I
No change 1n background;
Ho change In ME and E boundary
hydrocarbons
Mo change 1n background;
Reduced HE *nd E boundiry
hydrocarbons
Reduced backgroufid ozont;
Reduced HE and E boundary
hydrocarbons
Reduced background hydrocarbons;
Reduced HE and E boundary
hydrocarbons
Reduced background hydrocarbons
and ozone; Reduced NE and E
boundary hydrocarbons
I I
I T
.--'" X ..••
.••"" / ,, •'''.,.. "'"
-" s •• ..-•''
/ / ^;>-
- X - A'1' ^——
25
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
FIGURE 7-25. Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for peak regional ozone in the Philadelphia urban
plume for 19 July.
149
-------
Mo Clunoe In b*ck ground ;
No change 1n NE i"d E bound**-}
hydrocarbons
No Change 1n background;
Reduced HE ind t boundary
Xydrocirbons
Reduced background ozone;
Reduced HI mo I bounairy
hydroc»roons
Reduced background hydrocirbons
Reduces HE ind E
hydrocirbons
Reduced btckground hydrocirbonj
and ozone; Reduced NE ma E
boundiry nydrocjroons
25 50 75
PERCENT HYDROCfiRBON EMISSION REDUCTION
100
FIGURE 7-26. Relative ozone reduction (%) versus percent hydrocarbon
emission reduction for 19 July at the Downington, PA monitor.
150
-------
100
75
a
LJ
LJ
(£
z 50
a
u
LJ
C£
u
Q_
25
I I I I I i I I I
No change 1n background;
No change In NE and E boundary
hydrocarbons
No change 1n background;
Reduced NE and E boundary
hydrocarbons
Reduced background ozone;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons,
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons
and ozone; Reduced NE and E
boundary hydrocarbons
I 1
.<>" .,--
// .,-•:;; *
-" •/• . ' ..•'•''
// . ,;;..'•"
-~f •&'"''
.-' ..V
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
103
FIGURE 7-27. Relative ozone reduction (%} versus percent hydrocarbon
emission reduction for 19 July at the Roxy Water, PA monitor.
151
-------
100
75
u
3
O
UJ
a:
UJ
z
D
M
O
UJ
D_
•o Change 1n background;
No change 1n N£ and E boundary
hydrocarbons
No change in background;
Reduced HE and E boundary
hydrocarbons
Reduced background ozone;
Reduced Nt and E boundary
hydrocarbons
Reduced background hydrocarbons;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons
— and ozone. Reduced Nt and E
boundary hydrocarbons
50
.' f m '
.-•* x
,-'- / ,. •'
(_.- / ., • 1(1.,.
25 50 75
PERCENT HYDROCflRBON EMISSION REDUCTION
100
FIGURE 7-28. Total ozone reduction (%) versus percent hydrocarbon
emission reduction of peak regional ozone in the Philadelphia urban
plume for 19 July.
152
-------
100
LJ
r)
o
UJ
Oi
UJ
z
o
M
D
UJ
LJ
01
UJ
QL.
Ho change 1n background;
No change 1n N£ and E bound*
kydrocirbons
No Ctwn9« In background;
Reduced HE and E boundary
hydrocarbons
Reduced background ozone ;
Reduced Nt and E boundary
hydrocarbons
Reduced background hydrocarbons
Reduced NE and I boundary
hydrocarbons
Reduced background hydrocarbons
and ozone. Reduced NE and I
boundary hydrocarbons
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
100
FIGURE 7-29. Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 19 July at the Downington, PA monitor.
153
-------
100
75
o
Z)
o
z 50
D
1 I I I j
Ho change 1n background;
Ho clunge 1n HE and E boundary
hydrocarbons
Ho change in background;
Reduced NE and E boundary
hydrocarbons
Reduced background ozone;
Reduced HE and E boundirjr
Hydrocarbons
Reduced background hydrocarbons;
Reduced HE and E boundary
Hydrocarbons
Reduced background nydrocirbons
and ozone, Reduced NE and E
boundary hydrocarbons
i r
...-*
LJ
LJ
C£
UJ
0_
25
.-••" / ,*•••"' ..
' ,--•/ ,.-•-
' //
— •" j> A' «••"
-- / • .."'
-• s
1 I
25 50 75
PERCENT HYDROCRRBON EMISSION REDUCTION
IB;
FIGURE 7-30. Total ozone reduction (%) versus percent hydrocarbon
emission reduction for 19 July at the Roxy Water, PA monitor.
154
-------
for air quality planners in Northeast Corridor cities. This finding sug-
gests that reduced ozone levels in one airshed may not be maintained if
neighboring cities do not also make comparable emission reductions.
A summary of hydrocarbon reductions required to meet the ambient standard
for ozone for all sensitivity simulations of 19 July is presented in Table
7-9. In the no-change background simulations with no decreases of
boundary hydrocarbons, the peak predicted ozone value at the Roxy Water
monitor decreases very slightly with increased hydrocarbon control. This
shows that the peak predicted ozone value at the Roxy Water monitor is the
result of inflow of ozone precursors from the Northeast and East
boundaries. Peak predicted ozone, however, is decreased as boundary con-
centrations for hydrocarbons are decreased, as illustrated in Figure 7-24.
At the Downington and Roxy Water monitors, no hydrocarbon controls are
needed to meet the ozone standard if either background hydrocarbons alone
or both background hydrocarbons and ozone are decreased, because the peak
predicted values in the simulation without emission control are already
below 12 pphm for ozone (Table 7-9). Hydrocarbon controls needed to meet
the standard for the regional peak predicted ozone drop from 57 percent to
16 percent if background hydrocarbon is reduced by 50 percent, background
ozone is reduced by 33 percent, and the inflow of hydrocarbons from the
New Jersey/New York urban area is reduced by the same amount as the reduc-
tion in the Philadelphia emissions.
Relative and total ozone response curves for the regional peak in the
Philadelphia urban plume, Downington, and Roxy Water monitors are pre-
sented in Figures 7-25 through 7-27 and Figures 7-28 through 7-30, respec-
tively. The decrease in predicted maximum ozone for the regional peak (71
percent) as shown in Figure 7-28 is nearly equal to the reduction in
hydrocarbon emissions and boundary conditions (75 percent) if the back-
ground ozone is decreased by 33 percent and the background hydrocarbons
are decreased by 50 percent. As is evident in Figure 7-30, the calculated
percentage reduction in ozone concentration at the Roxy Water monitor is
only slightly affected by hydrocarbon emission reductions from the Phila-
delphia area alone. Substantial ozone reductions can be calculated only
if DAM input changes that correspond to emission reductions from the New
York City/New Jersey airshed are achieved.
For 13 July, spatial patterns of changes in predicted ozone for the sensi-
tivity simulations are displayed with the use of Deficit/Enhancement (D/E)
plots. The list of D/E plots for the sensitivity simulations of 19 July
is presented in Table 7-10. Figures 7-31 through 7-33 present D/Es for
the sensitivity simulation in which only hydrocarbon emissions from Phila-
delphia were reduced 25, 50, and 75 percent, respectively. No noticeable
decrease in ozone is found in central Philadelphia for any of these three
155
-------
TABLE 7-9. Hydrocarbon emission reductions required to meet the NAAQS for ozone
from the sensitivity simulations of 19 July.
Sensitivity Simulation
Percent Hydrocarbon Emission Reduction
Philadelphia
Urban Plume Downington Roxy Water
No change in background;
No change in NE and E boundary
hydrocarbons
No change in background;
Reduced NE and E boundary
hydrocarbons
Reduced background ozone;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons;
Reduced NE and E boundary
hydrocarbons
Reduced background hydrocarbons
and ozone; Reduced NE and E
boundary hydrocarbons
57
44
40
20
16
27
23
15
> 75
15
12
156
-------
TABLE 7-10. Deficit/enhancement figures for ozone for 19 July sensitivity
simulations.
Figure
Sensitivity Simulation minus Base Case Simulation
7-31
7-32
7-33
7-34
7-35
7-36
7-37
7-38
7-39
7-40
7-41
7-42
7-43
7-44
7-45
6.25HC
6.50HC
6.75HC
6.B25HC
6.B50HC
6.B75HC
6.BK03
6.BKHC
6.BKHC.03
6.B50HC.BK03
6.B50HC.BKHC
6.B50HC.BKHC.03
6.B50HC.BK03
6.B50HC.BKHC
6.B50HC.BKHC.03
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
minus
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6. BASE
6.BK03
6.BKHC
6.BKHC.03
157
-------
0
30
20
tn
LJ
10
0
NORTH
10
20
30
vX'-Jx'-^.vX1: :'x':::'::.J'.-. .;l''- ::! :::'?':::':*:::vlt'-'-.::'.::: :$-::>: !:'':- ^x-'^x.v'1:':':':'^: ;:*.: '. *•: -.:l :: :::* :::::*::x::'' • ':''x-xJ.;:': :•!-.': ::!-x:':tx-. J:;.''::l:''
1 "^^11111
#:••' •.•.-.-.-.•.• .;;;;:;: ..:;:;.. :::;.Xx:
*x; ';': :':':'
i£ '' /•'•' . ,••-'' \ TREN
SK '. ,-\ ••" '••••"' •••• '-.
S?- ''.,. '" •::-.-. '•-- ~i ...
i!&: '"'••-. ,--•. ''NjORR BRIS
1 c.-'....-- v ,. *»ftxY NORT
I.''.-. .. -::::::::-. DOH«.° \ flMS
ii;s ' .-: "•"' '••-. '•-....•' .... ''-• ^flN- LUMB
:S-:. ••' •:"'••• '': ' < ''.'• ':• 0£fgC|MD
*y -; ••'.•• -. ''-. ..••'' CHES
gH; . -.. •••-3- oi .-' CLflY
&•'. • •' .. :•
m '•••" ': \ .-. '-. .... RNCO
'(K:-:-:-:-:-.-.::-
S:;:*:?::::i:::x::',-,., VINE
Si^KiS-x^xXx::-;:;,,,,.
j^x :•:•:• •:•:•:•:•.•:•.•:•:•.•:-:•: ......
SS::::::::::::::.:.:':v:.:::'::::::: :;::::: ::::: • : ': x:i:::: .•:•••:
*c:x:x::':: ::":':.:':-:':::::::.:: :x::; :x- •. -:: x:::': -::x:.- • •.•.•.•.•.•.•.•...-.•.•.•. ...•.;.:.;.:.••:•:.:•:.:•:. x.x.:.:.:-:-:.:.:-x.:.X'X-:.:.:.:.:.'.:.:-:.:-:-:.x-x.:.:-:.:.:-:-:.:-:.:-:.:.>:-: : :•: :-:•
y8®w^^&?^^
• ;f .;v.f •: x:*- ' •'-' *' :
: '^ :; ;:': • :+::.•.•'• 'ir
VflN;:':~
1?
jx.-r
BRIG £•
:K:~
iS~
•x-ii
X:ij-iv »; . :i:x: i : '.i' ;:. •
30
20
tn
10
10
20
30
SOUTH
FIGURE 7-31. Maximum deficit/enhancement for ozone (pphm) for all
hours on 19 July [6.25HC minus 6.BASE].
158
-------
0
20
in
UJ
10
0,
NORTH
10
20
30
3
••""' .-'" DOWN.
SUMM
TREN
"NORR.
BRIS
RMS
V. ;
CHES
CLflY
LUMB
RNCO
VINE
30
VRN-X
20
en
CE
10
BRIG
0
10
20
30
SOUTH
FIGURE 7-32. Maximum deficit/enhancement for ozone (pphm) for all
hours on 19 July [6.50HC minus 6.BASE].
159
-------
NORTH
10
20
30
30
20
TREN
BRIS
.-*•
RMS
LUHB
ISL
fe'
...: :' .-.-' '"' CHES
'''..--.. '••.. "x CLflY
.--.»..
flNCO
SUMM
VINE
VflN
BRIG
CO
cr
0
20
30
SOUTH
FIGURE 7-33. Maximum deficit/enhancement for ozone (pphm) for all
hours on 19 July [6.75HC minus 6.BASE].
160
-------
simulations, consistent with the hypothesis that ozone levels in the New
York plume must be decreased before substantial reductions in Philadelphia
ozone levels can be achieved on 19 July-type days. The decreases in ozone
occur, as expected, west of the urban center, in the area of the Philadel-
phia urban plume in the base case. If, however, hydrocarbon emissions
being transported through the Northeast and East boundaries are decreased
along with the urban emissions, then decreases for ozone occur throughout
the region as illustrated in Figures 7-34 through 7-36. Maximum changes
in predicted ozone for the regional peak were approximately 8 pphm when
hydrocarbon emissions were decreased 75 percent, as shown in Figures 7-33
and 7-36. Figures 7-37 through 7-39 were created to assess the changes in
ozone predictions with no change in hydrocarbon emissions and decreased
background estimates for ozone and hydrocarbons. For 19 July, background
estimates for ozone and hydrocarbons represent a decrease of 33 and 50
percent, respectively. When ozone background was decreased, changes of
less than 1 pphm occurred, as presented in Figure 7-37. By decreasing
hydrocarbon background, ozone decreases of nearly 4 pphm were calculated,
as shown in Figure 7-38. After decreasing both background hydrocarbons
and ozone, decreases of as much as 5 pphm were predicted, as shown in
Figure 7-39.
Maximum calculated changes in ozone from the base case simulation (6.BASE)
were evaluated for the simulations with a 50 percent decrease in hydro-
carbon emissions and boundary conditions, and various background assump-
tions for hydrocarbons and ozone. Figures 7-40 through 7-42 present the
spatial patterns of these changes. By decreasing urban hydrocarbon emis-
sions and boundary hydrocarbons 50 percent, and assuming a 33 percent
decrease in background ozone and a 50 percent decrease in background
hydrocarbons, predicted ozone was decreased by 11 pphm in a small area
west of Downington (Figure 7-42).
Patterns showing relative changes in predicted ozone for those simulations
with decreases of 50 percent in urban hydrocarbon emissions and boundary
concentrations are given in Figures 7-43 through 7-45. The patterns are
fairly similar to those in Figure 7-35, with minor differences due to the
nonlinearity of the atmospheric chemistry in the model.
161
-------
0
NORTH
10
20
30
30
20
10
r>
'•.. NORR ':
DOWN
..- •-'..."" .-•••"' \I3L
,-A '"' CHES.-/
CLfiY
SUMM
TREH
'.BklS
•-:. 1
VfiN ^
HNCQ
BRIG
'VI.N.E.
a
0
10
20
30
SOUTH
FIGURE 7-34. Maximum deficit/enhancement for ozone (pphm) for all
hours on 19 July [6.B25HC minus 6.BASE].
162
-------
0
30
20
LU
10
^
NORTH
20
30
DOWN
SUMM
TREN
RORR ''•• ': '.'• '.BRIS
'
• -V:HES
CLOY •.
flMS.
JSL
^
•-. VINE •-
.•: LUHB
flNCO
BRIG
10
20
30
SOUTH
30
20
en
CE
10
FIGURE 7-35. Maximum deficit/enhancement for ozone (pphm) for all
hours on 19 July [6.B50HC minus 6.BASE].
163
-------
NORTH
0
10
20
30
ora
•. •'.•' '• '• '. '•'. . •'. .-
fi-', '. • "'.^A* '*. '• ""'*'*. •••. f^ '.'.". "" .'.•.•.•. .'.• .'. •'.-,-',•';•'. .-,-' ',-'.-' . '•'.'. *|
i*C-' '. . .•••••' .. • • * • • . • ". . ' i '•*•.. i EfpCgw :. <=>
|;;; '•-.;•/' ^••'/ ......: i /•""-'''CHESN (i\ :.\ V^ /
/!•!•'• '•'.'*** * " •*.**•••*»*, * *"'•*•. • ' •
|yl •••••;•'"";: ^••'^•••' / ^ /' i. .':.-. ^v.:: "\\^RNCP
j£';: -A;:'../ .^\;-"-"-'»'.'.'.'.'- •.'.••'•.:;:'...•• '' •:!•.''•-.''••. '\--^ \. '"-.
~;- •' ' ; -x-x " « •• "."•'•' ***« •'
•'.•'.-*.: •' •'•.-:•.'• . •••• '. "^w • . ', - *•• • *•. ". •. DDir
|ij;:.;;;.:.:;:::::;::.:::::: '-. '•./ SUMM \ -.'•-.';••. : \ i '-. B G
s:'''V::;'S:'.:x ;SS; '••.'•••..^.^2-. .-••'. -•'
~..:,.:x,x:,x;v.:X:::x:: ............. _1 .••
^^^^^^^^j^^^^^^^^^^^^i^r^^i^^'&M^
3 10 20 30
:• '^
" 'i.
j"*
VP.rOf
X1**
i:l:~
1
S""
;:-f
j: ^
; 1 • i." ._ -•
30
20
en
cr
UJ
10
SOUTH
FIGURE 7-36. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.B75HC minus'6.BASE]
164
-------
20
en
rx:
•K*!
10
r i
0
NORTH
20
30
DOWN
CHES
CLRY
V
SUMM.
TREN
NORR BRI5
NORT
RMS
LUHB
RNCO
VINE
30
BRIG.
20
en
cc
10
> r c i. .t...j!.
.i. A^ 4 j i j-
i i 1
10
20
30
SOUTH
FIGURE 7-37. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.BK03 minus76:BASE]
165
-------
0
20
tn
10
NORTH
10
20
30
< j i i i j r i r j> ^ t t i r j i r t i T * » '
-1
30
v... !
NOW?
'"a
&*
ISL
, •-. TREN
t
BRr§.
NORT
.••' LUMB
VRN ;.
20
CO
cr
UJ
.
•--. CLflY -v ;
flNCO
....SUM*
VINE
/ BRIG
V.
tti i... f... j... .t.. .1.. .t
.J. . I... I. ..!....!.. t .A...T. I. i I *..»..! I t I I- t \ \ i i J '. .»•
0
10
30
SOUTH
FIGURE 7-38. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.BKHC minus 6.BASE]
166
-------
NORTH
0
10
20
30
» T at I f "i i i t 3 4 •
30
.-5"
20
DOWN-.'
10
X*
SUMH'
V
I I * > j f t r I } J ± J i
NORT
RHS
MD
CLRY
V. ':
t*
flNCO
V.
•' VINE
BRIG
. .1 . . .1 . . J. . J. .
. .1 >
-:;::T-:::::ii:::::^
30
20
en
cr
10
10
30
SOUTH
FIGURE 7-39. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.BKHC.03 minus76cBASE]
167
-------
NORTH
0
10
20
30
30
20
10
T -I i J 1 I I J^ > > i X 1 I 1 fjft T I r JI J tI [ ' *
C 'I •« «• «« „ _.
A;:|: . •' : <
00^
~3'
ii
> :
DOHN
.-••••CHES :.
•CLflY '-. ...
'
SUHM ...
f 1 I t I T t
30
: &'. '.
TREN
,
'*•. BR I S^
'
.VIME--3 •
.1..1 C i
VflNftx
20
10
.BRIG •
V i:fii:
t I >
10
20
SOUTH
30
FIGURE 7-40. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6B50HC.BK03 minus 6.BASE]
168
-------
NORTH
0
10
20
30
30
20
10
0
I I-1 J xl T I I » } 1 ± I 1 T JIJ I I + -r» I rI < ] } J 't I t
• * *•*.*•
1; ;-.".•:;::...;•, #'
-i
P . .1 . . .1 . . j. . j. . . i. . . ± . . i . . .1 . . .1 . . .t. . . I. . I. .
:' TREN
'. '•'•', \ :•• :
: .- : : >» -. •.
NORR " ! V,'"';. ''BR'IS <
• '••."YINE.--"
v,'••..Us
30
VflWx-x
20
in
en
10
.'.'J9RIG
!:::;ii:!g^^
'0
10
20
30
SOUTH
FIGURE 7-41. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.B50HC.BKHC minus 6.BASE]
169
-------
NORTH
20
30
F I ± T I I I if I I t T I
1 -I I T J t l £ I I ± J I
..••-?-.
-6
20
-•' w, 4—:NORR : \ •:''•-. ''-BIRIS
'" «WMI,~ • .uneT .•• N
tn
UJ
'«:'•-. .-: -•".-. "t fiRfliv/ U A-'bDHB
10
•
•..
«'•.;./._ ; "• ^ '-. •.. '•-..'"•-..' Sx
'•.. '•''':'•:. i \ "'\ \ 'BRIG "" ||
BL^a.
30
20
in
CE
UJ
10
10
30
SOUTH
FIGURE 7-42. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6B50HC.BKHC.03 minus 6.BASE]
170
-------
NORTH
0
10
20
30
20
10SS-
0,
£x
''•''•
$:•:
H
&:: • v
i!* ':':': ..- — . *-.. X \'
•X'-*- •" ". * * """"••
;J;X; . * " • "* n * "*"" "*•
£v:; /' '•-.'•-. •? '"-•• ....-5—
HS --V-- --\ v :• • ;
£; -- i» ..- i.., ';
V*V A '" '" **' *,
£:£ ""^3 = .._..•••""••••••••" .-•*' *•. -'"-
vX !"•••.,*.'* ff* ' ' • , ,' ' t
|| _.. v:.' •••"•...../ .... :-. V
K"X .••*"*•. .**'* '""
•X*X .'...*.*•.•* • ....... rlUKrf
•X*X ." " •.*•" ••** ••••" "^ *• .'* • iW
•frX* * . * " « ^ ft QUM* *'•*"• "•
x"x- •* •*•" ".* •* * . •" '• "• i
IjX> '•..••.. • • . . "" " •' ". i
gxi -V.'- J: ... .v .•'""•••*".•"...-" . • • . . \
^•'•" ' " • .-.. •"<>;"""•' ^ •*' ""• ]
X'X ;**.• .*.* ** / ..""t*.- t
|j:i: -V ....../ : ,.--' x' -'"CHES •
Ix'x: ,•' ,.-•" B ". ."*.-... '\ C?-"' '.
X*i> * ."»•*". •"**.»' * '
',•,•', • . . • .i^ • • *
X*X* '.*•»"* ••".". • f.
|:ii; '•/••'•'!•"' * /•.." \ '•-.. :'''--l..-'' ''-...
¥x* ".* •" *. "• *' ". .** . '. **•"" **. "*.
•£•"• •' " * • * * *••....' .../•' ^ _ *
r v.'
»• SUMM
•^
Vk
+
V
•^
h
K
' .1 . . .1 . . i .T J ± * 1 I t f. . . f . . ± . 1 . .1 T I {
.
'
. , ".
,. . •*'. • • ''X'X'X ':vXv/.;X;' y-'. vX;l; ^ X;X; X;.;. .%;.; ;.;! X;!; -l''-;':':
'. *, \ I'l'X-x';':" X;. x';'i'; ;• 'x-x-x';' '.'X :•;• 'x-i '.'xV.
-••"'. *'. \ '• 'XvX-X< Xs vXv ;t X;X;X;X >X; ;X X;> '...X;';
'x ''• '• ' •' ::i:::::::':.:: '::::: :::' :;::'- : :':::' :
.. '-.' ''. -, '•• : ":"X'X': ? $: S :|S. :':':•:>
v • • • : : '' ''""'''' '& . :•:'!:'
: ^ : \ :|
. '•..' -.-. • j:
f '. ':
:• .' • :=
• ' ,L j!
. *. RNCO ••
• • x •:
T '-. "*•• '• '•• =i
V'-..'>j, : \ '\ BRIG j;
'•. vmii:'" •' •' :~-
i 1 1.113 i -i. . d 111 f{ 1 I >
^
-
-.
—
ITT
;^,
•^
'.r
'*^
^
:7!
"iJ
^r
S
:^T!
"lij
*^M
i^
!•"
•s
".T
A^
JM
j_i
30
20
(__
to
cr
1 1 i
LLJ
10
d
0
10
20
30
SOUTH
FIGURE 7-43. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.B50HC.BK03 - 6.BK03]
171
-------
0
10
30
20
en
10
; V. •-.
NORTH
20
.-••;•-:.•>•'.r*--d«,!!
*' .-• .-•- CLflY '-. \
•. / •' •' •. '.
V
SUMM
•-!•
30
TREN
•. \ DEfi£uSMD •'•-•'
\ ISLf
BRIG
0
» I 3 I I ± i
ii:::^
30
20
tn
cr
10
0
10
20
30
SOUTH
FIGURE 7-44. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.B50HC.BKHC minus 6.BKHC]
172
-------
NORTH
0
10
20
30
20
tn
10
?:?:"*"
I i i I i j t » xi i -i * j j ^ t i""* I- J J. 3I I 3 »> * ' *
..v.
* X .
,;"N* XV^.'x
..-•-s-
•' '"*'
A: :
NORR '-' "•.••/:
'
TREN
-:flMS.;
CHES:. •
CLHY --. ••.'..
•:•:•:• ' • .• •
*SL>7V-
\ V. '•-
SUMM
LUMB
\flHCO
BRIG
r&$$$Z$%i^^
30
20
en
cc
10
10
20
30
SOUTH
FIGURE 7-45. Maximum deficit/enhancement for ozone (pphm)
for all hours on 19 July [6.B50HC.BKHC.03 minus 6.BKHC.03]
173
-------
8 SUMMARY AND CONCLUSIONS
Two summer days experiencing the highest and second-highest observed
region-wide ozone concentrations during the 1979 Philadelphia Oxidant Data
Enhancement Study were simulated with the Urban Airshed Model (UAM). This
study is one of a number of recent applications of the UAM to large metro-
politan areas sponsored by the EPA. Other applications have involved the
cities of Tulsa, Denver, and St. Louis; work is currently under way to
evaluate and apply the UAM to the New York metropolitan area.
The Philadelphia application is unique because the Philadelphia airshed is
nested in the expansive and emission-rich Northeast Urban Corridor. Under
certain meteorological conditions, interurban transport of fresh oxidant
precursors plays a key role in determining ozone levels in Philadelphia
due to the proximity of the Baltimore/Washington D.C. urban area to the
southwest, and the New Jersey/New York urban area to the northeast. In
this study, two distinct meteorological regimes were successfully simula-
ted. On 13 July 1979, high pressure, light winds, and carryover of urban
emissions from the previous day created conditions that led to the highest
and most widespread high ozone readings recorded during the summer of
1979. The second day simulated, 19 July, was characterized by an influx
of fresh migratory precursor emissions from the New Jersey/New York urban
area to the northeast into the Philadelphia region. The bulk of the
Philadelphia urban plume was advected to the west of the urban center on
this day, while the New Jersey/New York plume moved toward the Philadel-
phia urban center, affecting most station readings. The occurrence of
this transport was substantiated by the fact that a parcel trajectory
reaching the station recording a high concentration at the time of the
peak (Claymont, Delaware) did not travel over high-density emission areas,
but originated well to the northeast of the region in the New Jersey/New
York urban complex.
The first phase of the model evaluation study involved the preparation of
UAM inputs using day-specific air quality and meteorological data. These
inputs included estimates of initial conditions, boundary conditions,
modeling region specifications, surface land-use features, background
concentrations, meteorological scalars, mixing heights, and wind fields.
An emission inventory for a typical 1979 summer weekday was prepared and
evaluated by the EPA prior to the commencement of this study and was used
175
-------
in all model simulations for both days. Boundary conditions, "initial
conditions, and background concentrations were specified after examining
day-specific air quality and meteorological data. Meteorological scalars
(e.g., humidity, surface pressure, etc.), mixing heights, and wind fields
were prepared using meteorological data routinely collected in the region
and additional data such as radiosonde, pibal, solar radiation, and other
surface data from sites which operated specifically for the 1979 summer
oxidant data enhancement program in Philadelphia.
Base-case simulations were obtained for both days after a number of pre-
liminary partial-day simulations were performed. For the 13 July base
case, the regional peak predicted hourly average ozone concentration was
26.6 pphm, and the peak ozone concentration calculated at a station
monitor (Conshohocken, Pennsylvania) was 19.3 pphm; the peak concentration
observed at a station monitor was 20.5 pphm. The peak predicted value in
the region was 30 percent greater than the peak observed value. This fact
may indicate that the model is overestimating the region-wide peak ozone;
however, it is unlikely that a limited number of monitors will capture the
actual peak ozone that occurred in the region, especially when the urban
plume is advected downwind of the urban emission center to the fringe of
the metropolitan area where few monitors are located. Model evaluation
statistics for peak ozone prediction/observation values paired in space
but not paired in time show that the model slightly overestimates peak
concentrations by 8.6 percent, with an average overestimate of 1 pphm.
For all predictions/observations paired in space and time for observations
above 5 pphm, the model overestimates by 15 percent, with an average over-
estimate of 1.7 pphm. The gross error or mean absolute deviation was 14.8
percent for station peak pairs and 29.5 percent for all pairs. These
statistics reveal that for this simulation day, the model shows a tendency
toward overestimation of hourly ozone; however, overall model performance
is still fairly good.
The development of an appropriate base case simulation for 19 July was
more difficult because of the uncertainty in specifying the quantity of
precursor inflow from the New Jersey/New York urban area with boundary
concentrations on the Northeast and East boundaries. If data from an
extended network of monitors at the surface and aloft along a particular
boundary provided information on the quantity of inflow material, some
degree of uncertainty would still remain in the transport, mixing, and
photochemical transformation of this material because of the uncertainty
involved in specifying a suitably representative flow field on the surface
and aloft from a limited number of wind measurements. Not only is the
quantity of precursor material important for boundary specifications, but
the timing of the boundary inflow is also critical to the ozone-formation-
dependent factors of mixing, turbulent diffusion, and solar radiation as
the material is advected through the airshed. Trajectory analysis using
176
-------
the surface wind field prepared for the 19 July simulations was used to
verify parcel trajectory paths to the peak observed station ozone value.
Parcels originated in the early morning hours northeast of Philadelphia,
and data from an upwind monitor (Van Hiseville, New Jersey) were used to
estimate hourly precursor inflow along the boundary.
For this simulation, the predicted peak hourly ozone concentration was
17.7 pphm in the urban plume located west of the Downington, Pennsylvania
monitor compared to an observed peak value of 17.0 pphm recorded at the
Roxy Water, Pennsylvania monitor. The predicted maximum ozone value was
14.2 pphm for the Downington monitor, whereas a value of 15.7 pphm was
observed for the same hour. The peak regional value predicted for this
same hour was located 30 km west of Downington. This suggests that per-
haps spatial alignment errors were introduced because of the possible
overestimation of wind speeds due to the technique used in preparing the
wind field. In spite of possible minor spatial alignment errors, evalua-
tion statistics show good model performance for the 19 July base case.
Statistics on peak station predictions/observations paired in space but
not paired in time reveal that the model slightly overestimates peak ozone
by 2.1 percent, with an average overestimate of 0.05 pphm. Statistics for
predictions/observations paired in space and time show nearly zero bias.
The mean absolute deviation or gross error for all pairs, however, is 28.8
percent, with a nonnormalized average absolute deviation of 2.4 pphm. The
model underestimated ozone in the lower and upper ranges of ozone observa-
tions and overestimated ozone in the mid-range of observations. These
model performance evaluation statistics are important considerations for
both 13 and 19 July in the light of the next phase of the study, which
tested the sensitivity of ozone to changes in critical inputs (such as
boundary and background conditions) and used the model to simulate hypo-
thetical urban hydrocarbon emission-reduction scenarios.
A total of 16 20-hour ozone sensitivity simulations were carried out for
the 13 July episodes, and a total of 18 simulations were performed for the
19 July episodes. These simulations addressed various assumptions for
background ozone and hydrocarbons, inflow boundary hydrocarbons, and
across-the-board reductions in urban hydrocarbon emissions (reductions of
25, 50, or 75 percent). Background hydrocarbons were decreased for both
simulation days by approximately 50 percent, while background ozone was
decreased 50 percent for 13 July and 33 percent for 19 July. Only one
simulation was performed in which hydrocarbon boundary conditions on the
Southeast inflow boundary were decreased by 50 percent for 13 July.
Boundary value hydrocarbons on Northeast and East inflow boundaries for
19 July were decreased by factors corresponding to reductions in the
hydrocarbon inventory of 25, 50, and 75 percent. Ozone response in all 13
July sensitivity simulations was obtained by examining the peak predicted
hourly ozone in the Philadelphia urban plume and at the two monitors with
177
-------
the highest hourly predicted ozone. For 19 July, the peak predicted
regional ozone from Philadelphia urban emissions, as well as the highest
station affected by these emissions, was examined. The ozone response at
the station with the calculated highest ozone affected by the simulated
New Jersey/New York urban emission plume was also examined. Ozone
response in the sensitivity simulations revealed the model's sensitivity
to various model input assumptions such as background conditions. How-
ever, simulated emission reduction scenarios were performed to broadly
demonstrate the real purpose of the Urban Airshed Model, which is the
prediction of (1) future air quality and (2) controls on present-day emis-
sion sources needed to attain and maintain the ozone NAAQS. It is not the
intent of this study to dictate the actual control measures needed to
improve future air quality in the Philadelphia AQCR; however, the results
demonstrate the model's ability to identify critically important planning
variables (e.g., HC reductions planned in neighboring influential cities),
and thus its use in regulatory settings by air quality planners.
Results for the 13 July sensitivity simulations show that, for the peak
predicted ozone concentration in the entire grid, the hydrocarbon emission
controls needed to show attainment of the NAAQS for ozone range from 75
percent reduction, assuming no change in background, to 35 percent reduc-
tion if background ozone and hydrocarbons are decreased by 50 percent.
Background values for hydrocarbons and ozone in the Philadelphia AQCR are
the result of atmospheric loading in. the entire eastern United States,
encompassing the northeast urban emission corridor. Background values in
the future can decrease only if other AQCRs also control precursor emis-
sions throughout the region. It is beyond the scope of this study to
quantify the reduction in background levels under various urban emission
control scenarios throughout the East; however, the model results show the
response of ozone to a hypothetical decrease of 50 percent for backgrounds
of ozone and hydrocarbons using the meteorological conditions of 13
July. The effect of emission controls in upwind urban areas could be
quantitatively assessed if a regional oxidant model were used to define
the boundary conditions of the Airshed Model. Regional oxidant models
have been developed and applied to the northeastern United States (Liu et
al., 1984; Lamb, 1983; Schere and Possiel, 1984). By simulating the
effect of emission controls at the regional scale with a regional model,
it would be possible to determine the resulting reduction in background
hydrocarbons and ozone. These reduced background concentrations could
then be used as input to UAM simulations. Examination of the peak predic-
ted regional ozone for the sensitivity simulation of 13 July shows that
the difference in urban hydrocarbon emission control needed to meet the
standard for two background assumption cases (no-change background case
vs. 50 percent decrease in background hydrocarbons and ozone) is large--75
percent compared to 35 percent. This difference is significant for air
178
-------
quality planners and operators of urban emission sources because any
change in present-day emission levels has a large effect on the economic
factors involved in operating such sources. Similar large differences
were obtained from ozone response on 13 July at the peak station examined-
-45 percent urban hydrocarbon reduction with no change in background com-
pared to 13 percent reduction needed if background for hydrocarbons and
ozone is decreased 50 percent.
The 19 July sensitivity simulations were complicated by the transport
regime and influx of urban emissions from the New York city area. For the
regional peak predicted ozone concentration from the Philadelphia urban
plume, there were large differences in control requirements to attain the
ozone NAAQS (similar to those of 13 July)--57 percent control of urban
hydrocarbon emissions with no change in boundary conditions or background
hydrocarbon and ozone assumptions, compared to only 16 percent control
required if ozone background is decreased 33 percent, hydrocarbon back-
ground is decreased 50 percent, and Northeast and East boundary hydrocar-
bons are reduced by the same amount as are Philadelphia emissions.
Because overall peak ozone predictions in the simulation of 19 July were
substantially less than those for 13 July, lower control requirements for
urban hydrocarbon emissions were needed to show attainment. For 19 July,
all but three of the sensitivity simulations decreased boundary inflow
hydrocarbons by 25, 50, and 75 percent. The three simulations in which
boundary hydrocarbons were not reduced reveal that attainment may not be
maintained for ozone in the Philadelphia AQCR under certain meteorological
conditions if neighboring emission-rich areas do not als-o control emis-
sions. This is due to the close proximity of these emission areas and the
meteorological regimes under which interurban transport occurs.
Specific conclusions of this urban photochemical modeling study can be
summarized as follows:
(1) The UAM has been demonstrated to successfully simulate two dis-
tinct meteorological regimes in the Philadelphia AQCR--stagna-
tion and interurban transport—that led to widespread high ozone
concentrations. Model performance, using output statistical
measures recommended by EPA/AMS, was judged to be good for both
simulation days. For the 13 July base case, the model tended to
overpredict peak ozone by an average of 8.6 percent at all sta-
tion monitors. For 19 July, the model overpredicted peak ozone
by 2.1 percent. Because of the slight tendency to overpredict
peak ozone and because this tendency is conservative, the model
can be considered a satisfactory predictor of ozone in the
Philadelphia AQCR.
179
-------
(2) Using the simplified urban hydrocarbon emission/initial condi-
tion reduction scenarios in this study* the UAM-calculated urban
hydrocarbon emission control requirements for Philadelphia
needed to attain the NAAQS for ozone for 13 July were as fol-
lows: greater than 75 percent reduction using the calculated
peak regional ozone, or 45 percent reduction if the calculated
peak ozone at a station monitor is used. For 19 July the hydro-
carbon control requirements were as follows: 57 percent hydro-
carbon reduction using the calculated peak ozone of the Phila-
delphia plume, 27 percent reduction using the peak value calcu-
lated at a station monitor influenced by the Philadelphia urban
plume, and greater than 75 percent using the peak value calcula-
ted at a station monitor influenced by the New York plume. The
latter control requirement indicates that this monitor was
almost entirely influenced by the simulated New Jersey/New York
urban plume, because no change in peak ozone was found for those
simulations that used reduced Philadelphia urban hydrocarbons
coupled with no change in the Northeast and East inflow boundary
hydrocarbons.
(3) Urban hydrocarbon control requirements vary widely depending on
the background levels assumed for ozone and hydrocarbons.
For 13 July, if ozone background is reduced by 50 percent, the
hydrocarbon control requirement (using the peak regional value)
to meet the ozone NAAQS is decreased, from greater than 75 per-
cent to 58 percent. If background hydrocarbons are reduced by
50 percent, with ozone background unchanged, the hydrocarbon
control requirement is decreased to 43 percent. Reducing both
ozone and hydrocarbon by 50 percent decreases the hydrocarbon
control requirement to 35 percent. For 19 July, the hydrocarbon
control requirement for the Philadelphia urban plume is reduced
from 57 to 44 percent when Northeast and East inflow boundary
hydrocarbons are reduced by the same amount as are the Philadel-
phia emissions. In addition, when background ozone and hydro-
carbons are reduced, the following results are obtained: if
background ozone is reduced 33 percent, the hydrocarbon control
requirement is decreased from 44 to 40 percent; if background
hydrocarbon is reduced 50 percent, with ozone background
unchanged, the hydrocarbon control requirement is decreased to
20 percent; and if background ozone is reduced 33 percent and
background hydrocarbons are reduced 50 percent, the hydrocarbon
control requirement to meet the NAAQS for ozone drops to 16
percent. For both days modeled, reductions in background hydro-
carbons yield a much greater decrease in hydrocarbon control
requirements than do comparable reductions in background ozone.
180
-------
(4) Regional cooperation in reducing area emissions in the eastern
United States appears to be required ff background levels for
hydrocarbons and ozone are to be lowered. Reducing background
levels of ozone and hydrocarbons will reduce control require-
ments for individual sources to achieve attainment. Under cer-
tain meteorological conditions, interurban transport of fresh
hydrocarbon emissions may lead to violations of the ozone
standard in the Philadelphia AQCR regardless of the Philadelphia
AQCR emissions levels. It is, therefore, important for neigh-
boring emission source regions to also lower emissions.
181
-------
REFERENCES
Alkezweeny, A. T., K. M. Basness, R. C. Eastie, and T. S. Wetzel (1981),
"Northeast Corridor Regional Modeling Project: Aircraft Measurements
- New York and Vicinity," EPA-450/4-81-012, March 1981, pp. 235.
Allard, D., M. Chan, C. Marlia, and E. Stephens (1981), "Philadelphia
Oxidant Data Enhancement Study—Analysis and Interpretation of Mea-
sured Data," EPA-450/4-81-011, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina.
Ames, J., T. C. Meyers, L. E. Reid, D. C. Whitney, S. H. Golding, S. R.
Hayes, and S. D. Reynolds (1985), "SAI Airshed Model Operations
Manuals Volume I—User's Manual and Volume Il—Systems Manual," EPA-
600/8-85/007a,b, U.S. Environmental Protection Agency, Research Tri-
angle Park, North Carolina.
[AMS] American Meteorological Society (1981), "Air Quality Modeling and
the Clean Air Act: Recommendations to EPA on Dispersion Modeling for
Regulatory Applications," American Meteorology Society, Boston,
Massachusetts.
Anderson, G. E., S. R. Hayes, M. J. Hillyer, J. P. Killus, and P. V.
Mundkur (1977), "Air Quality in the Denver Metropolitan Region--
1974-2000," 77-222, Systems Applications, Inc., San Rafael, Califor-
nia.
Bencala, K. E., and J. H. Seinfeld (1979), "An Air Quality Model Perform-
ance Assessment Package," Atmos. Environ., Vol. 13, pp. 1181-1185.
Boris, J. P., and D. L. Book (1973), "Flux Corrected Transport, I:
SHASTA, an Algorithm That Works," J. Comp. Phys., Vol. 11, pp. 38-69.
Bornstein, R. D. (1968), "Observations of the Urban Heat Island Effect in
New York City," J. Appl. Meteorol., Vol. 7, pp. 575-582.
Chock, D. P., and A. M. Dunker (1982), "A Comparison of Numerical Methods
for Solving the Advection Equation," Environmental Science Depart-
ment, General Motors Research Laboratories, Warren, Michigan.
183
-------
Clark, T. L., and R. E. Eskridge (1977), "Non-Di>ergent Wind Analysis
Algorithm for the St. Louis RAPS Network," Environmental Sciences
Research Laboratory Report EPA-600/4-77-049, U.S. Environmental Pro-
tection Agency, Research Triangle Park, North Carolina.
Clarke, J. F. (1969), "Nocturnal Urban Boundary Layer over Cincinnati,
Ohio," Hon. Wea. Rev., Vol. 97, No. 8, pp. 582-589.
Demerjian, K. L., K. L. Schere, and J. T. Peterson (1980), "Theoretical
Estimates of Actinic (Spherically Integrated) Flux and Photolytic
Rate Constants of Atmospheric Species in the Lower Troposphere," in
Advances in Environmental Science and Technology, Vol. 10, pp. 369-
459, J. Pitts and R. Metcalf, eds., John Wiley & Sons, New York, New
York.
[EPA] U.S. Environmental Protection Agency (1981), "Interim Procedures for
Evaluating Air Quality Models," Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.
[EPA] U.S. Environmental Protection Agency (1982), "Emissions Inventories
for Urban Airshed Model Application in the Philadelphia AQCR," 450/4-
82-005, Research Triangle Park, North Carolina.
[EPA] U.S. Environmental Protection Agency (1983a), "The St. Louis Ozone
Modeling Project," 450/4-83-019, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina.
[EPA] U.S. Environmental Protection Agency (1983b), "Evaluation of Per-
formance Measures for an Urban Photochemical Model," 450/4-83-021,
Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina.
Fox, D. G. (1981), "Judging Air Quality Model Performance," Bull. Am.
Meteorol. Soc.. Vol. 62, pp. 599-609.
Godowitch, J. M. (1984), personal communication.
Godowitch, J. M., J.K.S. Ching, and J. F. Clarke (1984a), "Spatial Varia-
tion of the Evolution and Structure of the Urban Boundary Layer,"
submitted for publication to Bound. Layer Meteorol.
Godowitch, J. M., J.K.S. Ching, and J. F. Clarke (1984b), "Evolution of
the Nocturnal Inversion Layer at an Urban and a Nonurban Location,"
submitted for publication to J. Climate and Appl. Meteorol.
184
-------
Hayes, S. R. (1979), "Performance Measures and Standards for Air Quality
Simulation Models," 78-93R, Systems Applications, Inc., San Rafael,
California.
Hillyer, M. J., S. D. Reynolds, and P. M. Roth (1979), "Procedures for
Evaluating the Performance of Air Quality Simulation Models," 79-25R,
Systems Applications, Inc., San Rafael, California.
Hoel, T. G. (1962), Introduction to Mathematical Statistics, John Wiley &
Sons, New York, New York.
Holzworth, 6. C. (1972), "Mixing Heights, Wind Speeds, and Potential for
Urban Air Pollution Throughout the Contiguous United States," Office
of Air Programs Publication No. AP-101, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina.
Jeffries, H. E., R. M. Kamens, K. G. Sexton, and A. A. Gerhardt (1982),
"Outdoor Smog Chamber Experiments to Test Photochemical Models," EPA
Cooperative Agreement No. 805843, University of North Carolina,
Chapel Hill, North Carolina.
Kaimal, J. C., N. L. Abshire, R. B. Chadwick, M. T. Decker, W. H. Hooke,
R. A. Kropfli, W. D. Neff, F. Pasqualucci, and P. H. Hildebrand
(1982), "Estimating the Depth of the Daytime Convective Boundary
Layer," J. Appl. Meteorol., Vol. 21, pp. 1123-1129.
Killus, J. P. (1982), "Background Reactivity Estimates for Atmospheric
Modeling Studies," presented at the XV Informal Conference on Photo-
chemistry, June 27-July 1, 1982, Stanford, California.
Kill us, J. P., and M. K. Liu (1981), "Development and Evaluation of a
Regional-Scale Photochemical Air Quality Simulation Model," paper
presented at the 62nd annual meeting, Pacific Division AAAS, Univer-
sity of Oregon, Eugene, Oregon, 14-19 June 1981.
Killus, J. P., and G. Z. Whitten (1981), "A New Carbon-Bond Mechanism for
Air Quality Simulation Modeling," SAI No. 81245, prepared for the
U.S. Environmental Protection Agency under contract #68-02-3281,
Systems Applications, Inc., San Rafael, California.
Killus, J. P., J. P. Meyer, D. R. Durran, G. E. Anderson, T. N. Jerskey,
S. D. Reynolds, and J. Ames (1977), "Continued Research in Mesoscale
Air Pollution Simulation Modeling: Volume V—Refinements in Numeri-
cal Analysis, Transport, Chemistry, and Pollutant Removal," 77-142R,
Systems Applications, Inc., San Rafael, California.
185
-------
Lamb, R. G. 1983. "A Regional Scale (1000 km) Model of Photochemical
Air Pollution, Part 1, Theoretical Formulatkm," EPA-600/3-83-035,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina.
Layland, D. E., and H. S. Cole (1983), "A Review of Recent Applications of
the SAI Urban Airshed Model," EPA-450/4-84-004, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
Liu, M. K., R. E. Morris, and J. P. Killus (1984), Development of a
Regional Oxidant Model and Application to the Northeastern United
States, Atmos. Environ., Vol. 18, No. 16, pp. 1145-1161.
•
McRae, 6. J., W. R. Goodin, and J. H. Seinfeld (1982), "Mathematical
Modeling of Photochemical Air Pollution," Environmental Quality
Laboratory, California Institute of Technology, Pasadena, California.
Noonkester, V. R. (1976), The Evolution of the Clear Air Convective Layer
Revealed by Surface Based Remote Sensors," J. Appl. Meteorol., Vol.
15, pp. 594-606.
Possiel, N. C., C. W. Spicer, P. R. Sticksel, G. M. Sverdrup, A. T.
Alkezweeny, and W. E. Davis (1984), "Northeast Corridor Regional
Modeling Project: Ozone and Precursor Transport in New York City and
Boston During the 1980 Field Program, EPA-450/4-84-011, U.S. Environ-
mental Protection Agency, Research Triangle Park, North Carolina.
Reynolds, S. D., H. Hogo, W. R. Oliver, and L. E. Reid (1982), "Applica-
tion of the SAI Airshed Model to the Tulsa Metropolitan Area," SAI
No. 82004, report to U.S. Environmental Protection Agency, Contract
No. 68-02-3370, Systems Applications, Inc., San Rafael, California.
Reynolds, S. D., L. Reid, M. Hillyer, J. P. Killus, T. W. Tesche, R. I.
Pollack, G. E. Anderson, and J. Ames (1979), "Photochemical Modeling
of Transportation Control Strategies. I. Model Development, Per-
formance Evaluation, and Strategy Assessment," Federal Highway
Administration, Office of Research, U.S. Department of Transporta-
tion, Washington, D.C.
Roth, P. M., C. Seigneur, S. D. Reynolds, and T. W. Tesche (1983),
"Assessment of NOX Emission Control Requirements in the California
South Coast Air Basin, Vol. Ill," SYSAPP-83/019, report to Western
Oil and Gas Association, Systems Applications, Inc., San Rafael,
California.
186
-------
Schere, K. L., and K. L. Demerjian (1977), "Calculation of Selected Photo-
lytic Rate Constants over a Diurnal Range: -A Computer Algorithm,"
EPA-600/4-77-015, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina.
Schere, K. L., and N. C. Possiel. 1984. "U.S. EPA Regional Oxidant
Model--Background and Overview." 77th Air Pollution Control Associa-
tion Annual Meeting, San Francisco, California, 25-29 June 1984.
Seigneur, C., T. W. Tesche, P. M. Roth, and M. K. Liu (1983), "On the
Treatment of Point Source Emissions in Urban Air Quality Modeling,"
Atmospheric Environment, Vol. 17, No. 9, pp. 1655-1676.
Seigneur, C., T. W. Tesche, P. M. Roth, and L. E. Reid (1981), "Sensi-
tivity of a Complex Urban Air Quality Model to Input Data," J. Appl.
Meteorol., Vol. 20, pp. 1020-1040.
Seinfeld, J. H. (1977), "Review and Analysis," in International Conference
on Qxidants, 1976—Analysis of Evidence and Viewpoints—Part VI. The
Issue of Air Quality Simulation Model Utility, EPA-600/3-77-118, U.S.
Environmental Protection Agency, Research Triangle Park, North Caro-
lina.
Singh, H. B., and P. L. Hanst (1981), "Peroxyacetyl Nitrate (PAN) in the
Unpolluted Atmosphere: An Important Reservoir for Nitrogen Oxides,"
Geophys. Res. Lett., Vol. 8, No. 8, pp. 941-944.
Spangler, T. C., and R. A. Dirks (1974), "Meso-Scale Variations of the
Urban Mixing Height." Bound. Layer Meteorol.. Vol. 6, pp. 423-441.
Tesche, T. W., W. R. Oliver, H. Hogo, P. Saxena, and J. L. Haney (1982a),
"Assessment of NOX Emission Control Measurements in the California
South Coast Air Basin—Appendix A: Performance Evaluation of the
Systems Applications Airshed Model for the 26-27 June 1974 03 Episode
in the South Coast Air Basin," SAI No. 82064, Systems Applications,
Inc., San Rafael, California.
Tesche, T. W., W. R. Oliver, H. Hogo, P. Saxena, and J. L. Haney (1982b),
"Assessment of NOX Emission Control Requirements in the California
South Coast Air Basin—Appendix B: Performance Evaluation of the
Systems Applications' Airshed Model for the 7-8 November 1978 N02
Episode in the South Coast Air Basin," SAI No. 82230, Systems Appli-
cations, Inc., San Rafael, California.
187
-------
Wesely, M. L. (1983), "Turbulent Transport of Ozone to Surfaces Common in
the Eastern Half of the United States," in Trace Atmospheric Con-
stituents: Properties, Transformations, and Rates, pp. 345-370,
Advances in Environmental Science and Technology, Vol. 12, S. E.
Schwartz, ed. (H. Wiley-Interscience Publication).
Westburg, H., and P. Sweany (1980), "Philadelphia Oxidant Enhancement
Study: Hydrocarbon Analysis," EPA-450/4-80-022, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
Whitten, G. Z., and H. Hogo (1978), "User's Manual for Kinetics Model and
Ozone Isopleth Plotting Package," EPA-600/8-78-014a, U.S. Environ-
mental Protection Agency, Research Triangle Park, North Carolina.
Yanenko, N. N. (1971), The Method of Fractional Steps, Springer Verlag,
New York.
Zalesak, S. T. (1979), "Fully Multidimensional Flux-Corrected Transport
Algorithms for Fluids," J. Computational Physics, Vol. 31, pp.
335-362.
188
-------
APPENDIX A
COMPILATION OF AIRSHED RESULTS FOR 13 JULY 1979
189
-------
APPENDIX A
COMPILATION OF AIRSHED RESULTS FOR 13 JULY 1979
This appendix compiles various sets of Airshed Model ozone results.
Figure A-l is a complete set of hourly average ozone isopleths in the
Philadelphia area throughout the day. Time-series plots comparing dis-
tance-weighted average and maximum/minimum ozone predictions within one
cell's distance with observations at 19 monitoring stations are presented
in Figure A-2. Model results presented in this manner are helpful in
developing a better qualitative understanding of the simulation results.
These grid model concentration estimates used in comparison with the
observed data are defined as follows:
Distance-Weighted Average. The solid lines presented in Figure A-2
represent an average model prediction (for comparison with observed
values), which is obtained by computing the distance-weighted average
concentration of the four grid cells nearest to the monitor where the
observed value was recorded.
Maximum/Minimum One Cell Away. The dashed lines presented in Figure
A-2 represent the maximum and minimum concentrations predicted in the
block of nine cells centered on the grid cell containing the monitor-
ing station. This envelope provides an indication of the spatial
variability of the predicted concentration values in the immediate
vicinity of the station.
These plots reveal the presence of steep spatial gradients in the concen-
tration field and the qualitative effect that wind field errors might have
on the performance results. In calculating the model evaluation sta-
tistics, the distance-weighted average estimates produced by the simula-
tion are used.
191
-------
NORTH
30-.
SOUTH
(a) BETHEEN THE HOURS OF 9 fiND 10
FIGURE A-l. Hourly variation in predicted ground-level ozone concentration
fields (pphm) for Philadelphia, 13 July 1979 (bold numbers correspond to
station observations).
192
-------
0
10
NORTH
20
30
-
Vf?t~, ,
§ ~ '
M
•.:-]
J ••> it: ::| •••••{• :;': Xjl'.'l'fr: ••^:.:' .1.. :. '}•'•' •(••• '-1 '•>. '•> "::.' ''' '':';'.-' •'•''.; ' '': > '" '•:
SOUTH
(b) BETWEEN THE HOURS OF 10 RND 11
FIGURE A-l continued.
193
-------
NORTH
SOUTH
(c) BE THEEN THE HOURS OF 11 RND 12
— c
FIGURE A-l continued.
194
-------
NORTH
0
10
20
30
30
20
t^
to
*
ie
•:;xl-:-xl--- • -i> v -I.::: J . .-.TTT- 17
'•'• ''"' i ::'-i''' -i:'.' i: v:|';!:!:t'' '{' '' -i -?;' ''?':'::::.:: '':-'?': <' ' '' :•:? >
SOUTH
(d) BET HE EN THE HOURS OF 12 ffND 13
FIGURE A-l continued.
195
-------
NORTH
0
10
20
SOUTH
(e) BE THE EN THE HOURS OF 13 RND
FIGURE A-l continued.
196
-------
NORTH
SOUTH
(f) BE THE EN THE HOURS OF 14 ffND 15
FIGURE A -1 continued.
197
-------
NORTH
0
10
SOUTH
(g) BETWEEN THE HOURS OF J5 fWD 16
A-! continued.
198
-------
NORTH
0
10
20
30
_ r-/?
rO'T -?,*
SOUTH
(h) BE THE EN THE HOURS OF 16 fJND 17
•••-i ^
3 s
FIGURE A-1 continued.
199
-------
UORTH
SOUTH
(i) BETWEEN THE HOURS OF 17 ffND 18
FIGURE A-l continued.
200
-------
NORTH
0
10
20
SOUTH
(j) BE THE EN THE HOURS OF 38 RND 19
FIGURE A-l concluded.
201
-------
RMS LRB
• 6
12
18
24
RNCORfl
0 6
12
18
24
20
16
I 12
o
6
i i iiiriiiiiiiiiir | iiiir^
~ AIRSHED RUN D ——
- 1 CELL MfiX/MIN —
_ OBSERVED _
6 12 18
TINE (HOURS)
24 24
20 20
16 16
12 i 12
8 6
4 4
I 1 I I 1 I i i i r i | i i i TI \ t
~ BIRSHED-RUN D
- 1 CELL HflX/MIN --•
_ OBSERVED .
20
16
12
I I I I I I I I I I I I I I I I I I I ! I I I
6 12 18
TIME (HOURS)
BRISTOL
6
12
18
24
20
16
x
t 12
o
B
AIRSHED RUN D —-
- 1 CELL MRX/MIN --.
_ OBSERVED m
24
CflMDEN
0 6
12
18
24
24 24
20 20
16 16
12 t 12
I T I I I i i t I I I I i 1 T I I
AIRSHED RUN D —
- 1 CELL MRX/MIN --•
_ OBSERVED .
I 1 t 1 I
20
IS
12
12
TIME (HOURS)
18
12
TIME (HOURS)
18
FIGURE A-2. Comparison of Airshed Model predictions
and ozone observations for the 13 July 1979 episode,
Philadelphia. Pennsylvania.
202
-------
CHESTER
0 6
12
24
CLflTMONT
0 6
12
16
24
20
16
i 12
o
i i i i I ] i i i I r
~ flIRSHED RUN 0 —
1 CELL MRX/MIN --.
. OBSERVED .
i i i i r
T i i i r
^•V^n"!
DEFENSE SU
0 6
12
18
24 24
20 20
16 16
12 £ 12
12
TIME (HOURS)
16
i i i i i | i i i i i i i i i i r
' flIRSHED RUN D
- 1 CELL MflX/MIN —
_ OBSERVED -
24
J? . . i
20
16
12
12
TIME (HOURS)
18
24
f
FIGURE A-2 continued.
203
-------
DOHNINGTON
6
12
18
24
FRflNKLIN I
6 12
16
24
AIRSHED RUN D
- 1 CELL HRX/HIN
_ OBSERVED _
- 24 24
- 20 20
16
i r i i i [ i i i r i
' BIRSHED'RUN D ——
- 1 CELL MftX/MIN —
_ OBSERVED B
I-B
I JL i
24
20
16
12
12
TIME (HOURS)
18
12
TIME (HOURS)
18
I5LRND RD
6
12
18
- 24 24
- 20 20
- 1 CELL HRX/MIN —
LUMBERTON
0 6
12
18
6 12 18
TIME (HOURS)
i i T i r I i i i r i | i
- AIRSHED RUN D —
- 1 CELL HflX/MIN —
__ OBSERVED _
20
16
12
1 a a
0
12
TIME (HOURS)
><" -
i i i i
18
FIGURE A-2 continued.
204
-------
NORR15TOHN
6 12
18
24
28
16
a.
i 12
o
8
I I I 1 1 I I I I 1 1 I
• AIRSHED RUN D —
1 CELL MflX/MIN -••
. OBSERVED .
24
RQXt HflTER
0 6
12
TIME (HOURS)
18
24 24
22 20
16 16
a.
12 i 12
12
18
24
AIRSHED
Wiiiii JIIT
N D —
T I I T
- 1 CELL HRX/MIN —
_ OBSERVED .
41 MS,
12
TIME (HOURS)
16
SE SEHRGE
8 6
12
18
24
SOUTH BROR
8 6 12
18
24
28
16
i 12
«n
AIRSHED
1 I I I II|lllllTlilII
RUN D
- 1 CELL HRX/MIN
_ OBSERVED B
i i i i i I i i i i i I i i i i i I i i i i >
~ AIRSHED RUN D —
- 1 CELL MflX/MIN —
_ OBSERVED .
TIME (HOURS)
TIME (HOURS)
FIGURE A-2 continued.
205
-------
SUMMIT BRI
B 6 12
16
24
TRENTON
0 6
12
18
20
16
o.
i 12
I I I I I i I I I I | I I
filRSHED RUN D —
1 CELL MflX/HIN - —
OBSERVED
24 24
20 20
16 16
12 i 12
6 12 16
TIME (HOURS)
i i i i i i r i r i i
filRSHED-RUN D —
- 1 CELL MflX/HIN —
_ OBSERVED
i i i i i i i i
20
16
12
i_qi
I I
12
TIME tHOURS)
18
VflN HISEVI
6 12
18
24
24
20
16
o.
t 12
«n
AIRSHED
IlJllflllllljIil 1 i
RUN D
- 1 CELL HflX/MIN --•
_ OBSERVED .
24
20
16
12
6 12 18
TIKE (HOURS)
FIGURE A-2 concluded.
206
-------
APPENDIX B
COMPILATION OF AIRSHED RESULTS FOR 19 JULY 1979
207
-------
APPENDIX B
COMPILATION OF AIRSHED RESULTS FOR 19 JULY 1979
This appendix compiles various sets of Airshed Model ozone results.
Figure B-l is a complete set of hourly average ozone isopleths in the
Philadelphia area throughout the day. Time-series plots comparing dis-
tance-weighted average and maximum/minimum ozone predictions within one
cell's distance with observations at 20 monitoring stations are presented
in Figure B-2. Model results presented in this manner are helpful in
developing a better qualitative understanding of the simulation results.
These grid model concentration estimates used in comparison with the
observed data are defined as follows:
Distance-Weighted Average. The solid lines presented in Figure B-2
represent an average model prediction (for comparison with observed
values), which is obtained by computing the distance-weighted average
concentration of the four grid cells nearest to the monitor where the
observed value.was recorded.
Maximum/Minimum One Cell Away. The dashed lines presented in Figure
B-2 represent the maximum and minimum concentrations predicted in the
block of nine cells centered on the grid cell containing the monitor-
ing station. This envelope provides an indication of the spatial
variability of the predicted concentration values in the immediate
vicinity of the station.
These plots reveal the presence of steep spatial gradients in the concen-
tration field and the qualitative effect that wind field errors might have
on the performance results. In calculating the model evaluation sta-
tistics, the distance-weighted average estimates produced by the simula-
tion are used.
209
-------
HQR7H
10
20
30
if { I H 1...J. 3 i. ..*....!....t..t I 1 { 1 4 t t lit I It
SOUTH
(a) BETWEEN THE HOURS OF 10 AND 11
FIGURE B-l. Hourly variation in predicted ground-level ozone concentration
fields (pphm) for Philadelphia, 19 July 1979. (Bold numbers correspond to
station observations.)
210
-------
NORTH
tfft-S-i*:-'?:! v^
2
SOUTH
(b) BE THE EN THE HOURS OF 11 AND 12
FIGURE B-l continued .
211
-------
HQRJH
10
20
30
•:::;:|:::::;
"''*>
—' U
SOUTH
(c) BETHEEN THE HOURS OF 12 AND 13
FIGURE B-l continued .
212
-------
NORTH
10
W&-tt:W^^$$sf$$3$* i J t { i i t i i tt I > t i i t
10
20
SOUTH
(d) BEJUEEN THE HOURS OF 13 AND 14
FIGURE B-l continued .
213
-------
NORTH
:• {• :::::i::::::;i::::: :i > I t < i I i I i { { > j i i { i i t
3O
0
SOUTH
(e) BE THE EN THE HOURS OF 14 AND 15
FIGURE B-l continued .
214
-------
NORTH
10
20
::::3!::'.-:f •:;:•:»•:•:• .1... .1... .1... I.. I 5 f 3 t .I...I. I
(f) BETWEEN THE HOURS OF 15 AND 16
FIGURE B-l continued .
215
-------
NORTH
(g)BETWEEN THE HOURS OF 16 AND 17
FIGURE B-l continued .
216
-------
NORTH
^x&i-i-jgx; : F; *!^
Si?
SOUTH
WBETUEEN THE HOURS OF 17 AND is
FIGURE B-l continued .
217
-------
KQRJH
20
30
•v.#vs:4-:.v:-?s^^^
30 fc^-'^ '• :«:«W»M*:«««-™->X.:.:.. . KWWA*: :«*. .-:•: i**,:*:* ™«:^:^^^^^^^::: .:!^SM^:V ? :WiW W.:^
20
VAN:':'.-
AftCO
BKIS
VINE
. .} . . .1 . . . . . . t. . . t. . . j. . . .1 > 1 I
t tit t 1 i:::::::Jt:::::::i:::::::i:'::::!if::x|;j::x::|:::::::t:-::::{-::::::i:::::.:i;'::-::t:' : :
5 1?
12
20
30
SOUTH
(\)BETHEEN THE HOURS OF 18 AND 19
FIGURE B-l continued .
218
-------
NORTH
IT J < 1 J .1 I 1
{ I •) > > f { I I >
(j) BETWEEN THE HOURS OF 19 AND 20
FIGURE B-l concluded .
219
-------
RMS LRB
0 6
12
RNCORfl
0 6
12
18
24 24
- 20 20
- 16 16
- 12 £ 12
en
o
IIIIIJI1II1^
~ RIRSHED RUN 6
- 1 CELL MRX/MIN —
_ OBSERVED .
i i i i i i i i i i i i * at
24
20
16
12
12
TIME (HOURS)
18
6 12 18
TIME (HOURS)
BRISTOL
0 6
12
18
24
BRIGRNTINE
0 6 12
18
24
24
20
16
AIRSHED RUN 6 ——
- 1 CELL HflX/MIN —
_ OBSERVED .
I I
1 I I I I I I I I I I I I I I I
24 24
20 20
16 16
12 i 12
8 6
i i i i i i i i i i i | i i i i i | i i r i r
RIRSHED RUN 6
- 1 CELL MflX/MIN —
_ OBSERVED " m
24
16
12
6 12 18
TIME (HOURS)
6 12
TIME (HOURS)
18
FIGURE B-2. Comparison of Airshed Model predictions and
ozone observations for the July 19, 1979 episode,
Philadelphia, Pennsylvania.
220
-------
CflMDEN
i 6
12
18
I I I I I I I I I I I I I I I I I t I
D RUN 6 -
- 1 CELL HfiX/MIN
CHESTER
0 6
12
16
24 24
- 20 20
- 18 16
- 12 i 12
m
o
i i i r
• RIRSHED
1I I I I I I I 1 I I I [I I
D RUN 6
- 1 CELL MRX/MIN —
_ OBSERVED .
I I Q
24
20
16
12
6 12 18
TIHE (HOURS)
12
TIHE (HOURS)
18
CLRYMONT
0 6
12
18
24
CON5HOHOCK
6 12
18
24
16
£: 12
en
o
6
i i i i i l i i r
flIRSHED
l i i r T i
RUN 6 —
- 1 CELL MfiX/HIN —
_ OBSERVED ri
D
D
Q D
- /**» D
*
B &
* . X. «a t
i i i iL^f ? i i i i i i i i i i Vj i i i
6 12 18
TIHE (HOURS)
24 24
20 20
16 16
12 t 12
m
o
4 4
i i r \ r r i i i i i T T i i i r j t i T i i
• flIRSHED RUN 6 —
- 1 CELL HflX/MIN --•
_ OBSERVED m
24
20
16
12
:**
12
TIME (HOURS)
18
FIGURE B-2 continued.
221
-------
DEFENSE SU
6
DOHNINGTON
6
~ flIRSHED RUN 6
flIRSHED RUN 6
- 1 CELL MflX/MIN
_ OBSERVED
- 1 CELL MRX/MIN -
I I I I I I I I ?>!_ ' I
6 12
TIME (HOURS)
18
12
TIME (HOURS)
FRRNKLIN I
0 6
12
18
24
20
16
£
n.
t 12
en
flIRSHED RUN 6
- 1 CELL MflX/MIN —
_ OBSERVED _
s-v.
24
12
TIME (HOURS)
18
ISLRND RD
0 6
12
18
24
24 24
20 20
16 16
12 t 12
tn
o
#
I I T- T 1 I I T I I I 171 f I T r T i 1 i i
' AIRSHED RUN 6 —
- 1 CELL MflX/MIN --•
_ OBSERVED m
IK da
IS
12
6 12 18
TIME (HOURS)
FIGURE B-2 continued.
222
-------
LUMBERTON
6 12
NORR3STOHN
0 6 12
~ AIRSHED RUN 6 —
T i i I i i i i i I i i i i i I i i i i
- 1 CELL HflX/MIN —
_ OBSERVED m
- 1 CELL MflX/MIN —
OBSERVED B
0 i i I i i i i i I i i i i i
6 12 18
TIME (HOURS)
6 12 18
TIME (HOURS)
NDRTHER5T
0 6 12
16
24
ROXT HRTER
0 6 12
18
24
20
16
c.
t 12
I I I 1 T 1 1 1 I 1 I 1 I i I I I i I 1 I I
U flIRSHED RUN 6 — ~*
- 1 CELL MBX/MIN
_ OBSERVED B
i i i i i i i i i r T i i i i i i i i
AIRSHED RUN 6
- 1 CELL HfiX/MIN —
_ OBSERVED m
I N-J I _ J«.t" I I I I I I I I
12
TINE (HOURS)
18
12
TIME (HOURS)
18
FIGURE B-2 concluded.
223
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/4-85-003
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation and Application of the Urban Airshed Model
in the Philadelphia Air Quality Control Region
5. REPORT DATF
June 1985
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
T. N. Braverman, EPA; and J. L. Haney, SAI
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Inc. (SAI)
101 Lucas Valley Road
San Rafael, California 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3870
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division (MD-14)
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The Urban Airshed Photochemical Grid Model has been applied to a data base assembled
for the Philadelphia metropolitan area consisting of meteorological and air quality
data collected during the EPA 1979 summer field study and a spatially, temporally,
and chemically resolved emissions inventory of hydrocarbons and nitrogen oxides.. The
report presents results of (1) the evaluation of Urban Airshed Model performance on
2 selected model test days, (2) application of the Urban Airshed Model for control
strategy planning, and (3) sensitivity tests to determine the response of Urban Air-
shed Model predictions and control requirements to uncertainty in background concen-
trations. The model tended to slightly overestimate calculated hourly ozone for both
simulation days; but performed well in replicating observed concentrations. Results
indicate that urban hydrocarbon emission reductions required to attain the National
Ambient Air Quality Standard (NAAQS) for ozone vary widely depending on assumed levels
of background hydrocarbons and ozone. Under certain meteorological conditions, fresh
precursor emissions may be transported between cities in the Northeast Urban Corridor
leading to violations of the ozone standard. Therefore, all large metropolitan areas
in the Northeast must cooperate to reduce emissions if region-wide attainment of the
ozone standard is to be maintained.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Photochemical models
Ozone
Hydrocarbon control strategies
Background concentrations
Interurban transport
18. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
------- |