-------
20
9 10 11 12 1 23
Time of Day, By Hourly Interval
.-0— Obsar/ed
—<=? Predicted
(c) 3 August 1976
FIGURE 1-4. (Concluded)
-------
10 h
OBSERVED
PREDICTED
MEAN OF
3 DAYS
J 1°
Q.
ro
S-
It)—
0
o
c:
0 s
<^> s
CD
C
o
£ o
3 AUGUST 1976
28 JULY 1976
10
29 JULY 1975
Time of Day By Hourly Averaging Period
FIGURE 1-5. VARIATIONS OVER ALL STATIONS OF OBSERVED AND PREDICTED
AVERAGE OZONE CONCENTRATIONS
-------
22
Averaging differences does not reveal any significant bias during
hours of peak ozone concentrations. This finding is also confirmed by
the results shown in Figure 1-6. In this figure, the fractional mean
deviation from perfect agreement between predictions and observations
is shown to vary randomly at the higher ozone concentrations. The mean
deviations are not negligible, but they are not excessive and both
positive and negative differences occur.
Predicted and observed concentrations of carbon monoxide were also
compared. Carbon monoxide is effectively an inert pollutant. Its major
source is motor vehicle exhaust. Thus it is primarily emitted along
roadways. The comparisons showed outstanding agreement at some stations
and very poor agreement at others. This performance is consistent with
the well-established microscale variability of roadway CO. The varia-
tion of the CO concentration from one side of a downtown city street to
the other may well be from one to several hundred percent. The extremely
good CO correlation at some stations is taken to represent the typical
performance of the DAQM on a grid-cell-average basis (see the discussion
in Chapter II).
DAQM validation statistics were compared with statistics published
by the EPA (Turner et al., 1973) for annual average computations with models
EPA recommends for use with nonreactive pollutants (Table 1-1). Although
averaging over a year should improve statistics, measures for DAQM's
one-hour-average results exceeded those for the EPA models.
No validation was carried out for the application of the COM to
estimate annual average concentrations of NOp and particulates, since
it was in use and had been calibrated by Colorado state agencies. As
discussed in Chapters III and IV, neither calibration nor validation
are appropriate concepts for these N02 studies since comparison data
are only available from a single station in the central business district.
At that station we judged that the uncalibrated predictions were in error
by substantially more than they were in outlying regions, where high NO
concentrations are more likely.
-------
23
Root Mean Square Ozone Concentration (pphm)
..1/2
(Observed) + (Predicte
-2.0
FIGURE 1-6. ESTIMATE OF BIAS IN MODEL PREDICTIONS AS A
FUNCTION OF OZONE CONCENTRATION
-------
S09 Annual Averages*
AQDM
COM
Gifford (box)
Hanna (box)
Mean
Oq Hourly Averages
SAI Denver Model
TABLE 1-1. VALIDATION MEASURES OF VARIOUS MODELS
[differences = (predicted-observed)/observed]
Number
of
Comparisons
75
75
75
75
RMS
Difference
0.90
0.39
0.61
2.44
Mean
Absolute
Difference
0.68
0.27
0.53
1.32
Lower
Limit
Difference
0.64
0.87
1 .30
1.07
Upper
Limit
Difference
2.30
1.23
0.22
9;13
Difference
Range
2.94
2.10
1.52
10.20
75
279
1.08
0.31
0.70
0.35
0.97
0.90
3.22
1.16
4.19
2.06
ro
* S0? model measures were computed from data in Turner et al. (1973).
-------
Calibration constants recommended by the APCD were used for partic-
ulates. The constants were such that 51 ppm background was to be added
to 1.96 times the predicted concentration. This large a correction to COM
model results suggests substantial observational and/or emissions inven-
tory problems.
F. BASE CASE STUDIES
Estimates of present and future air quality in the Denver region
were made from model computations using SAI's Denver Air Quality Model
(DAQM) and EPA's Climatological Dispersion Model (COM). Computations
were made for comparison with each NAAQS. The DAQM, a "real time" model,
was used to estimate the highest one-hour-average concentrations of ozone
and carbon monoxide. COM was used to estimate annual average concentra-
tions of particulates. Annual average concentrations of nitrogen dioxide
were inferred from both COM and DAQM results.
1. Ozone
As described above, two of the three summer days for which data were
available were reasonably representative of worst-case ozone episodes.
Simulations with the meteorological conditions on the two days, 28 July
and 3 August 1976, were made with current emissions inventories for the
validation study. For the base case study, additional simulations were
made with the same meteorological inputs, but with emissions data esti-
mated to represent conditions in the years 1985 and 2000 (see Appendix C
for details).
The simulations of air quality in future years compared with 1976
(see Figure 1-7) show that in spite of population growth and growth
in transportation, commerce, and industry, peak ozone concentrations
are predicted to drop from 24 pphm in 1976 to 13 pphm in 1985 and
9 pphm in 2000, assuming strict compliance with the FVECP.
-------
26
Time of Day, By Hourly Interval fstart hourl ES3 Reduction 1976-1985
LstOP h°Ur J m Reduction 198S-2UOO
!a) Meteorology for 28 July 1976 Assumed
FIGURE 1-7.
REDUCTION IN PREDICTED OZONE CONCENTRATIONS
(pphm) AT DENVER STATIONS DUE TO PREDICTED
FUTURE EMISSIONS CHANGES
-------
27
15
i JL 12 II
9 ia n 12
, .start hourj ESI Reduction !976-198b
Time of Day, By Hourly Interval [stop hour J ^ Reduction 1985-2000
(b) Meteorology for 3 August 1976 Assumed
FIGURE 1-7. (Concluded)
-------
28
The emissions per vehicle mile traveled (VMT) of ozone precursors--
hydrocarbons and nitrogen oxides--drop drastically by the year 2000
because of Federally mandated reductions in emissions from motor vehicles.
The most significant reduction in peak ozone concentrations, however,
occurs during the first period--1976 to 1985--when only hydrocarbon emis-
sions are reduced. Reductions in NO emissions per vehicle-mile traveled
X
are not expected to decrease the total NO emissions because of increases
A
in VMT; but reductions in total HC emissions are expected from the Fed-
erally mandated HC reductions.
The results of the Denver model simulations also show that fractional
decreases in the area! extent of NAAQS violations exceed the fractional
decreases in peak ozone concentrations. In 1976, computed NAAQS viola-
tions (> 8 pphm taken as > 9 pphm) occurred over a maximum of 300 to 400
square miles. By 2000 the area is expected to be reduced to 0 to 40
miles, although the predicted exceedance is so small that a violation
cannot be confidently assumed (Figure 1-8).
2. Carbon Mcmoxide
Carbon monoxide concentrations were computed with DAQM for a winter
day, 15 November 1975. Winter days tend to have the highest concentra-
tions of nonreactive pollutants because the low solar intensity permits
the shallow morning mixing layer to last longer, and sunlight is not
required for production of nonreactive pollutants. In addition, the
late sunrise assures that the thinnest mixing layer will exist during
much of the morning rush hour period.
Concentration gradients can be much larger for nonreactive pollu-
tants than for ozone. This is because peak concentrations occur very
close to sources (primarily traffic for CO) before dispersion dilutes
the pollutant. As a result, the definition of peak concentrations of
CO is very difficult. CO monitors on opposite sides of a heavily
traveled street can easily show concentrations differing by a factor
of two (or, in some cases, by an order of magnitude) (see Chapter II).
-------
260 |-
O) J
C 290
O)
QJ I
> ;?oh
Year 1976 Emissions
Year 1985 Emissions
Year 2000 Emissions
12
T
OlUNt CUILtHIHAI IOH. 11
0
U p|>liii
ID pH.ii
I/ jiplui
Time of Day By Hourly Interval
_l I I L-
a y lu it
T IB TT T?
T T~
ro
10
(a) Meteorology for 28 July 1976 Assumed
FIGURE 1-8. SIZE OF AREA IN WHICH PREDICTED OZONE CONCENTRATIONS EXCEED
GIVEN VALUES FOR YEARS 1976, 1985, AND 2000
-------
CD
S^ rti
ra
3
O" ;ui
ICi
O
Year_ 1976 E_mi^s_sions
IWUNb UJfttlNIHAI lUN Q b |i|J
0 It pi
0 111 M
J L __
Year 1985 Emissions
,60|-
QZONE CONCEHTHATION: 2 3 ppm
A '0 aof
6 8 $ 1} ' 1 12 1
Year 2000 Emissions
LONUHIUAT10N u u j
TTTTT ^iTTT
f I ?
fo If tt Y I 4 I I f f
CO
O
Time of Day By Hourly Interval [||—-^'J"'
(b) Meteorology for 3 August 1976 Assumed
FIGURE 1-8 (Concluded)
-------
31
The DAQM computation has a two-mile resolution and therefore microscale
street effects are suppressed.
Under present conditions, CO concentrations in violation of the
NAAQS can surely be found in any major urban area at particularly criti-
cal spots such as industrial, shopping center, or stadium parking lots
or at busy intersections. This is certainly true in Denver, even though
the DAQM predicts peak CO concentrations of 26 ppm for one hour and 15
ppm for eight hours (NAAQS are 35 ppm and 9 ppm, respectively). Since
CO emissions are expected to decrease in the future [under the FVECP),
one-hour violations are not predicted for 1985 and 2000 with DAQM. The
predicted maximum one-hour-average CO concentrations for those years are
17 and 4 ppm, respectively. DAQM does predict a small eight-hour-average
violation (10 ppm) in 1985, but none (3 ppm maximum) in 2000. The
occurrence of violations in those years at "hot spots" will depend more
on future vehicle emissions than on regional growth. That is, the
number of vehicles in a parking lot should not change much, but the
emissions per vehicle should decrease. The CO emissions at low speed
(5 mph) are projected by the Federal Vehicle Emissions Control Program
(FVECP) to drop by two-thirds from 1976 to 1985 and by an additional 50
percent from 1985 to 2000 (see Appendix B). From this it could be antic-
ipated that many of the "hot spots" in the Denver region in
1976 will not produce NAAQS violations in those future years. Since the
present hot spots have not been identified or monitored, this anticipa-
tion cannot be confirmed.
3. Nitrogen Dioxide
Estimates of N0? concentrations are very difficult to make. The
primary emitted species is NO. Only 15 percent or less of the NOV emis-
A
sions are N0? (see Appendix B). There is no NAAQS for NO, however, and
the only standard for N02 is on an annual average basis. After entering
the atmosphere, NO is typically oxidized to N02 in the same set of
photochemical reactions that generates ozone. During daylight hours,
N02 photolyzes to regenerate NO. Because most NO eventually becomes NO-,
-------
32
the typical NCL analysis involves not N02 but N0x> under the assumption
that all NO emissions are N0? (CDH, 1976). Since an annual average is
X ^
desired, NO modeling was done in this project with the COM.
X
While most NO is eventually oxidized to N02, all emitted NOX is
clearly not in the form of N02 at every position downwind of the NO
source. The portion of NO that is not N0? is particularly high near
s\ £—
NO sources. COM can therefore be expected to seriously overestimate
N02 concentrations in the densest source areas of a region. Raw COM
estimates should be more accurate near the periphery of the modeling
region.
COM is a calibrated model; that is, its results are multiplied by
a factor to make them correspond with observations. The Denver region
had just one N02 monitoring station, which was in the center of the
densest downtown NO source area. COM results were calibrated with
.A
data from this station; the COM results for that station should be
equivalent to the annual average concentration at that site for the
calibration year. After calibration, the results from COM should be
progressively more in error (underpredicting) at greater distances from
the calibration site. Since the emissions of all interacting species
may change differently in the future, the calibration may not be valid
for future years.
The problem of the complex photochemical reactions is resolvable by
using DAQM; in fact, NO and N0? concentrations are computed by DAQM.
This model does not, however, give annual averages. Peak one-hour-
average N02 concentrations computed with DAQM for worst case summer days
show decreases in N02 from 1976 to 1985 and again to 2000. This is in
contrast to predictions of steady increases of N02 using the erroneous
assumptions of COM. It thus seems fair to conclude that:
> Observations show current compliance with the NAAQS standard
of 5 pphm N02 on an annual average basis.
-------
33
> N02 concentrations may decrease from 1976 to 1985 and again
to 2000 by the ratio of DAQM results for those years. They
should not increase as much as was predicted by COM.
> Current modeling techniques are inadequate to produce accurate
annual average N02 predictions.
4. Particulates
Annual average particulate concentrations were computed with COM.
The model results were calibrated with observational data. Since the
data are collected with monitors usually located near major source areas
(e.g., roads) the calibration tends to overestimate the regional average
concentrations. This is not to say the observations are wrong, merely
that they are not directly comparable with model results. It is probable
that the particulate concentration estimates contained in Chapter III
are overestimates. Nevertheless, the predicted concentrations are so
high (four to five times the state standard) that current and future
violations may be expected.
Particulate standards are violated in many areas throughout the
United States, but monitoring, source identification, and size analysis
problems are so great as to make appropriate response recommendations
difficult. Common monitoring problems are sample flow variation and
chemical reactions on the filter. Much of the particulate material
comes from natural sources that are difficult to control, and often a
large fraction of the material is in particles so large that they cannot
penetrate deep into the respiratory system. Thus they have little
physiological effect.
G. SENSITIVITY STUDIES
The base case projections described in the section above are esti-
mations of the air quality that would result from predicted emissions.
The air quality thus reflects growth projections for each community,
-------
34
plans developed for a transportation system to serve the communities'
needs, and Federally mandated vehicle emissions reductions.
It is important to estimate not only the air quality consistent with
expected development but also the changes in air quality that would be
caused by major growth modifications. Such modifications might be chosen
to achieve some goal or might occur because communities develop contrary
to projections. To determine the sensitivity of air quality predictions
to growth projections, model simulations were run with a series of hypo-
thetical changes in projected emissions patterns as described in Chapter
IV. The changes chosen were 25 percent emissions reductions in arbitrar-
ily selected suburban communities or clusters of communities, or a 17
percent reduction (i.e., no growth) in future emissions in the city of
Denver itself. In each case, emissions in the remainder of the region
were raised uniformly to provide a constant regional emission rate.
These changes are intended to represent larger, more isolated effects
than are likely to be obtainable by deliberate policy decisions or are
likely to occur in an uncontrolled course of events.
In spite of the drastic changes in emission patterns, the change in
predicted ozone distributions was almost undetectable. Thus it does not
seem likely that deliberate or uncontrolled redistribution of growth
patterns could greatly affect peak ozone concentrations or ozone patterns
The sensitivity of the ozone concentrations predicted by the Denver
Model to the assumed degree of atmospheric dilution was also studied.
Changes in meteorological input were made to represent the uncertainty
in specifying worst-case meteorology. Lower average wind speeds or
shallower mixing layers than originally estimated from data were used.
Both these changes would produce less dilution and thus higher ozone
concentrations than were computed in base case runs. Decreases of one-
third in wind speed, mixing layer thickness, and in both wind speed
and mixing layer thickness were assumed at each grid point. (Note that
the proportional decreases preserve wind patterns.) The results of
these simulations showed that a one-third decrease in wind speed had
-------
35
a negligible effect on the peak ozone concentration, but substantially
increased the area in which the NAAQS was violated. In contrast, peak
concentrations increased but areas did not when changes were assumed in
mixing layer thickness. When both parameters were changed, the results
changed by less than the product of the parameter ratios.
Simple "box model" analyses, which are often recommended for regional
planning applications, would predict a 50 percent ozone increase for a
33 percent reduction in either wind speed or mixing depth. Ozone con-
centrations predicted by the Denver Model increased 4 percent for the
one-third decrease in wind speed and 16 percent for a one-third decrease
in mixing depth. If both wind speed and mixing depth are decreased by
one-third, the box model predicts a 125 percent increase but the Denver
Model shows only a 33 percent increase.
These results show that worst-case meteorology estimates would have
to be in error by implausible amounts to invalidate the conclusions of
the base case studies.
H. MITIGATION STUDIES
The final studies using model simulations were to determine the
effectiveness of potential air quality mitigation measures. Although
improved air quality in Denver was predicted by the base case studies,
present air quality in Denver is seriously in violation of NAAQS, espe-
cially for ozone. Predictions in these studies show continued violations
of particulate standards through 2000 and violations of ozone and eight-
hour CO standards through 1985. Estimates of NOo concentrations vary
from compliance to violation through the entire period. Although no
projections were made past the year 2000, all projected vehicle emission
factor reductions should be achieved by 1993. Thereafter increases in
VMT may again increase ozone concentrations.
One major mitigation measure is strict compliance with FVECP
standards, which was assumed for base case studies. Data from EPA
-------
36
Region VIII indicate that compliance with the FVECP is unlikely without
controls more strict than presently recommended inspection and main-
tenance programs. The assumptions of growth, transportation planning,
and an effective vehicle emissions control program may not be realized.
In such circumstances, the institution of further air pollution mitiga-
tion measures may become necessary. If so, it would be desirable to know
how effective such measures might be and how much control might be
required to achieve an air quality goal.
Simulations were made to address these questions. First, a simula-
tion was made with a 30 percent reduction in emissions (from base case)
in every grid cell. This reduction could represent one or a set of
mitigation measures that apply proportionately to all sources. It
would seem that 30 percent regionwide emissions reductions would be
quite severe, and that larger reductions would be difficult to impose on
a community. Simulation results show a reduction in peak ozone concentra-
tions of 15 percent or less. Thus, if local mitigation measures need to
be applied to achieve air quality goals, they will have to be severe to
have a significant effect on ozone concentrations.
A second simulation indicates the much greater improvement in air
quality possible with selective vehicle emission controls. Such controls
would probably only be feasible by Federal mandate, since auto manufac-
ture is not locally regulated except in California. The selective control
simulation used relaxed NO emissions standards roughly corresponding to
X
a proposal now before Congress to permit 1 gram per mile of NO emissions
A
rather than the 1985 goal of 0.4 grams per mile assumed in the base case
run. Results of this simulation for conditions otherwise the same as
the 1985 base case run showed no violations of the NAAQS for ozone in
the Denver modeling region. Note that this result is restricted to the
modeling region. Relaxation of NOV emission standards would increase
X
N02 concentrations in downwind areas, possibly exacerbating air quality
problems in those areas.
-------
37
The reason for an improvement in air quality with increased NO
X
emissions is that the ratio of the two ozone precursors (hydrocarbons
and N0x) becomes less favorable for ozone production. Not enough hydro-
carbons would be present to oxidize the NO to N02 and thus stimulate
ozone production before atmospheric mixing and dilution lowered all
concentrations. Note that this result applies to summer days in Denver
with worst-case meteorology (as do many other results in this report,
to some degree). Regions with other emissions mixes or patterns and
other climatology might experience other effects.
A final simulation involved studying the effects of noncompliance
with FVECP (compliance was assumed in the base case). In this simula-
tion, light-duty vehicle emission factors were provided by EPA Region
VIII to represent vehicle emissions expected assuming the present limited
effectiveness of controls and without the implementation of an inspection
and maintenance program. Simulations of 1985 and 2000 base case runs with
0.4 and 0.9 grams per mile of NO emissions and with the new, degraded
A
emission factors were made. Qualitatively, the results of these runs
were similar to those of the base case, but the decrease in ozone con-
centrations was not as great. Figure 1-9 shows predicted ozone peaks
in each year for the base case and for noncompliance with the FVECP.
Also shown is the predicted ozone peak for a uniform 30 percent reduc-
tion in emissions and for a 5 percent reduction in hydrocarbon emissions
(estimated to be obtainable from the installation of vapor recovery
devices on gasoline station delivery nozzles).
The lines of equal predicted ozone concentrations shown in Figure
1-9 put predicted ozone peaks in perspective. They show graphically
the complexity of the dependence of ozone concentrations on precur-
sors. Note that compliance with the ozone NAAQS is not predicted
by 2000 if FVECP is not followed. Peak ozone concentrations about 2
pphm higher than base case results are predicted. This is not a large
absolute effect, but it is a substantial percentage change from the
small concentrations predicted if FVECP is enforced. Vapor recovery
at gasoline stations is shown to be an effective measure for achieving
a small decrease in ozone concentrations.
-------
38
EMISSIONS INVENTORY
FOR YEAR -1976
0.6
0.5
E <
O-O
-2>,0.
c at
(J O_
XI-
O "J
z cu
0.3
E -O
• tu
0.2
0.1
PREDICTED MAXIMUM
ONE-HOUR-AVERAGE
OZONE CONCENTRA-
TION (ppm)
(0.24)
0.9 GRAMS/MILE NOX EMISSIONS STANDARD
EPA REGION VIII EMISSION FACTORS
AP42 EMISSION FACTORS
ISOPLETHS OF MAXIMUM OZONE CONCENTRA-
TIONS COMPUTED FROM SIMULATIONS OF
SMOG CHAMBER RUNS WITH THE CARBON-
BOND MECHANISM
POINTS COMPUTED WITH CARBON-BOND
MECHANISM IN OAQM SIMULATION RUNS
FOR 3 AUGUST 1976 METEOROLOGY
0.05
MAXIMUM OZONE CONCENTRATION
(ppm)
0.10 0.15 0.20 0.25
2000
(0.07)
(0.24)
1976 STAGE II HC CONTROLS
30% REDUCTION
IN 1976 HC AND
(0.
0.5 1.0 1.5 2.0 2.5
6-9 a.m. Reactive Hydrocarbon Concentration (ppmC)
(computed near CAMP station by DAQM)
FIGURE 1-9. MAXIMUM OZONE COMPUTED BY CARBON-BOND SIMULATIONS
OF SMOG CHAMBER EXPERIMENTS AND BY DENVER MODEL
FOR VARIOUS ASSUMED FUTURE EMISSIONS
-------
39
In winter simulations, CO concentrations were found to be increased
over that assumed in the base case analysis by 60 to 70 percent in 1985 and
by 50 percent in 2000 by the deqraded emissions factors. Thus, the eight-hour
CO standard would not be met on even a grid-cell-average basis in 1985.
I. RECOMMENDATIONS
The results of this project show the critical importance of main-
taining vehicle emission standards if predicted improvements in Denver's
air quality are to occur. Although emission standards must be set by
Federal mandate, realization of the benefits will also depend on a strong
local vehicle inspection and maintenance program. If additional local
mitigation measures are required, substantial across-the-board reduc-
tions of regionwide emissions will be necessary to have a significant
effect on ozone concentrations.
Continued attention to collecting and improving emissions, air
quality, and meteorological data bases is vitally important for further
studies. Particularly useful additions to the data base would be data
on hydrocarbon species emitted and more extensive observations of wind
throughout the mixing layer and above and across the entire region.
It is suggested that further use of DAQM in the Denver region by
contractor or by local agencies would be quite valuable. Attention
should be given to shortcomings in modeling techniques for all pollu-
tants,ybut particularly for NO,,. No suitable models are now available
for estimating compliance with the annual average NAAQS for N02-
-------
41
II DENVER AIR QUALITY MODEL VALIDATION STUDY
The present study is the first application of SAI's Denver Model.
Before relying on model predicti-ons in simulations of future conditions
for which no observations exist, SAI believes it would be very desirable
to have at least some measure of the model's performance for situations
where observations do exist.
The results reported here show that the Denver Model performs as
well or better for one-hour-average ozone predictions as EPA models do
for annual average SOp predictions. However, measurement of the Denver
Model's performance is limited by uncertainty in available air quality
observations. A true validation of an air pollution model would require
comparison of its output concentrations against actually occurring pol-
lutant concentrations. The Denver Model output is in terms of averages
over grid cell volumes. Pollutant concentration averages at this scale
are not directly observable, since instruments sense the concentration
only within a very small sample of air. Values of "actually occurring"
concentrations, averaged over appropriate volumes, cannot be perfectly
determined: they can only be inferred by instrument observations of
limited accuracy representative of small samples from a grid cell volume.
Validation comparisons of model results versus observations must be
considered, nevertheless, in the absence of direct comparisons of
"observed versus true." "Model (predicted) versus true" comparisons
must depend on generalized statistical information as to the "observed
versus true" relationship.
Another concern in a validation exercise is to reveal individual
aspects of the model's performance. For example, the ability of the model
to simulate concentration maxima, concentration patterns, and the location
and orientation of those patterns may be considered individually.
-------
42
This is particularly important in a model to be used for predictions
of future air quality. In such an application, the driving conditions--
emissions inventory, meteorology, etc.—are hypothetical. The results
can be characteristic, at most, of the air quality on some specific
future occasion.
Given the uncertainty of the hypothetical future conditions, it
is not essential that a model be able to perfectly simulate observed
past conditions. If errors are due only to errors in wind patterns,
a model that could be depended on to predict concentration patterns
equivalent in magnitude and shape but displaced somewhat might be very
useful although validation comparisons at points in time and space
might not be impressive.
In the validation program reported here, every attempt was made to
examine all relevant issues—the standards of model performance, the
reliability of observational data and the individual and statistical
comparability of model output and atmospheric measurements.
A. SIMULATION CASES
The model simulations run for validation purposes were also used
to evaluate critical Denver air quality situations. These purposes
are consistent, since high ozone concentrations present the best test
of the model's ability to simulate the production of ozone from emitted
pollutants. Lower concentrations of ozone are not only less critical
in terms of air quality but are also more apt to be dependent on subtle
variations in background and to be less functionally related to reactions
between emitted materials.
The data files provided by the Colorado Division of Highways and
the Air Pollution Control Division of the Colorado Health Department
are described in Appendix B. They include meteorological data for
three high ozone summer days. The first of the three days, 29 July 1975,
-------
43
had been preselected and extra data were taken in anticipation of
stagnation conditions and high ozone on that day. Although the ozone
standard (8 pphm) was exceeded that day (11 pphm was observed), a
frontal system moved through the Denver region during the day and
air quality conditions did not become as severe as expected.
The remaining two days represented in the data set were selected
by Colorado Division of Highways as high ozone days out of an extended
period of intensive monitoring during the summer of 1976. The two days
were 28 July 1976 and 3 August 1976. The peak ozone concentration
measured on the 28th was 16 pphm (twice the standard) and on the
3rd it was 17 pphm. The peak ozone measured during the 1976 summer
study was 18 pphm on 30 June, which was not a day selected for meteor-
ological data collection.
These three days were the only days for which data were available
to SAI for air quality modeling. The favorable validation results re-
ported below may or may not be fortuitous, but they are based on the
entire available data set.
B. OBSERVATIONS AND COMPUTED RESULTS
The map of the Denver modeling region in Figure II-l shows a
15 X 15 array of two-mile-square grid cells. All of the simulation
results discussed here were obtained using this horizontal grid. The
upper limit of the analysis region was the inversion base or 2500 feet
above ground level, whichever was the lesser. The region was divided
into three equal layers, but all results discussed here pertain to the
layer adjacent to the ground.
Atmospheric ozone measurements were obtained at the stations indicated
in Figure II-l. Generally, 24 one-hour-average ozone measurements were
available for each station, but some data were missing. For the simulated
hours, beginning at 5 a.m., six stations had complete data for 29 July 1975
and eight stations for 28 July 1976 and 3 August 1976. Similar data
-------
44
KEY
NG - Northglenn NJ
WE - Wei by GM
AR - Arvada 0V
CR - C.A.R.I.H. PR
CM - Continuous Air Moni-
toring Program [CAMP]
National Jewish Hospital
Green Mountain
Overland
Parket Road
10
NORTH
20
CO
LU
3:
.-3'
SW^v
•$$
*?*'-*''•
if.:
jrt-'i'-""
• «i¥v.
:**~!KZ.
••f^*^\.
s
10
SOUTH
20
30
CO
CE
LU
30
FIGURE II-.l. MAP OF DENVER AIR QUALITY MODELING REGION
SHOWING 15 x 15 MODELING GRID OF 2-MILE
SQUARES AND AIR QUALITY MONITORING STATIONS
-------
45
sets were available for carbon monoxide but only one or two stations
reported NO data.
A
Computer results were printed out as arrays of grid cell concentra-
tions (for the ground layer only) at each hour and also as one-hour
averages. , The one-hour averages are the only ones discussed here since
they compare most directly with the atmospheric measurements. The
result predicted by the model and used to compare with a station obser-
vation is the grid cell value for the cell in which the station is
located. Use of this convention presumes that the predicted grid value
represents the value everywhere in the cell to the degree of accuracy
of which the model is capable. It does not represent a point value
(e.g., at the cell center) of a continuous distribution. Thus, no
interpolations were used to extract results at the coordinates of the
observation stations.
C. VALIDATION RESULTS
1. Comparisons of Ozone Predictions and Observations
Predicted and observed ozone concentrations for each of the three
days modeled are plotted versus time in Figure II-2. These comparison
plots show good general agreement although small differences between
predictions and observations occur on many occasions and substantial
differences occur on some occasions. Specific phenomenological expla-
nations cannot be identified for particular discrepancies, but some
circumstances that could produce the differences shown are known for
early morning, midday, and afternoon periods. These are described below.
Differences between predictions and observations near the start of
the simulation period are to be expected. At this early time of day,
measured values of ozone are always small and are in the range where
instrument errors can easily exceed 100 percent [See the full discussion
of ozone monitoring problems in Burton et al. (1976)]. Ozone trapped
in a lowering inversion on the previous afternoon could be present in
-------
46
5673
10 11 12 1 23 4
6 7 8 9 10 11 12 1 2 I ? 5
- 5
Tin* of Day, By Hourly Interval
Observed
Predicted
(a) 29 July 1975
FIGURE II-2. OBSERVED AND PREDICTED HOURLY OZONE CONCENTRATIONS
(pphm) AT VARIOUS STATIONS
-------
47
-0 0 0 PARKER RD
67.
78
89 10 11 12 i
9 10 TT T2 T 2
Time of Day, By Hourly Interval
fstart hour! —0
(.stop hour
Observed
Predicted
(b) 28 July 1976
FIGURE II-2. (Continued)
-------
48
c
o
o
OJ
§
9 lOTTTzTI
of Day. By Hourly Interval
...0 —
Gbsar/ed
Predicted
(c) 3 August 1976
FIGURE II-2. (Concluded)
-------
49
"pools" in areas protected from ventilation. Thus in the stable early
morning atmosphere, particularly strong ozone concentration gradients
can be expected, with the highest concentrations expected in protected
areas very near the ground. Since the predicted concentrations represent
averages over four square miles horizontally and at least 50 feet vertically,
it is likely that these values would be quite low, much lower than some
possible measurements. At any rate, subsequent ozone peaks will be
mainly determined by morning emissions of hydrocarbons and NO rather than
A
by morning ozone, which is low enough according to the measurements to
be of secondary importance.
The largest differences between model predictions and station obser-
vations are noted during peak ozone periods. During high ozone periods,
fresh emissions of NO can cause local reductions in ozone concentrations,
as discussed in Appendix A. This local effect would be most pronounced
at ground level along very heavily traveled thoroughfares. Studies of
microscale street canyon effects by Johnson (1974) show that concentrations
of auto emitted pollutants can vary quite significantly even from one
side of a street to the other.
Ozone suppression by NO emissions along heavy traffic routes would
cause some monitors to record less than the actual peak values of ozone
concentration that might occur very near their location. In Figure II-2(c),
for example, ozone peaks predicted for C.A.R.I.H. and National Jewish
Hospital on 3 August 1976 are similar, but the observed ozone peaks
at the two stations differ by 100 percent. Each of the stations is at
the edge of the central business district and might be expected to have
similar midday ozone concentrations. The existence of comparable
concentrations in the vicinity of both stations is not proved by the
simulation but, considering the foregoing discussion, neither is it
disproved by the observed values.
The series of two-dimensional isopleth maps of predicted ozone
concentrations in Figure II-3 show that on 29 July 1975 a few hours
"cooking" of the morning rush hour emissions were required to produce
-------
50
NORTH
SOUTH
(a) Hour 0800-0900 MST
FIGURE II-3. ISOPLETHS OF OZONE CONCENTRATIONS (pphm) ON
29 JULY 1975. Isopleth interval 1 pphm.
-------
51
NORTH
SOUTH
(b) Hour 1000-1100 MST
FIGURE II-3. (Continued)
-------
52
NORTH
SOUTH
(c) Hour 1200-1300 MST
FIGURE II-3. (Continued)
-------
53
NORTH
SO'
SOUTH
(d) Hour 1400-1500 MST
FIGURE II-3. (Continued)
-------
54
NORTH
SOUTH
(e) Hour 1600-1700 MST
FIGURE II-3. (Concluded)
-------
55
the maximum ozone concentration. (The highest ozone shown here is 23+
pphm at 12 a.m. to 1 p.m.) The ozone "cloud" subsequently drifted to
the south edge of the analysis region and then back northeastward
across the region. Any portion of the cloud that left the modeling
region before this flow reversal would be lost to the simulation.
It appears likely that in the real atmosphere, as in the model simu-
lation, ozone can be swept off the edge of the modeling region. If the
winds are such as to bring the material back, the model simulation results
will be in error; subsequent ozone concentrations will be underpredicted
along the path of the missing ozone. In the three simulation runs, this
effect appeared most pronounced on 29 July 1975 and could account for
the low afternoon concentration predictions on that day.
2. Analysis of Ozone Comparisons
Several statistics of the observations, the model predictions, and
the differences between them have been calculated. They are discussed
in the following sections to provide measures of the Denver Model's
performance and an understanding of what may be expected of it. In the
final section of this chapter, the Denver Model's performance is com-
pared with that of other air quality models.
The comparisons presented earlier (Figure II-2) clearly showed
predicted ozone variations during the course of a day that were of the
same magnitude and occurred in much the same time patterns as the
observed variations. If some part of the differences between predictions
and observations are regular then those differences might indicate
aspects of the model that could be improved; that is, consistent
biases may be indicative of functional shortcomings. To test this
possibility, the data were averaged over all stations for each day.
These averages are plotted in Figure II-4 along with the data averaged
over all three days.
-------
P 10 -
d.
Q.
C
o
O
c
o
o
Ol
c
o
-p— OBSERVED
A PREDICTED
Time of Day By Hourly Averaging Period stpp hour)
FIGURE II-4. TIME VARIATIONS OVER ALL STATIONS OF OBSERVED AND
PREDICTED AVERAGE OZONE CONCENTRATIONS
-------
57
It can be seen from this figure that the comparison between obser-
vation and prediction improves with each averaging process. The
correlation of observations and predictions when averaged over all days
and stations is indeed quite good. If the data are also averaged over
time of day, the average prediction is found to be just 0.4 pphm less
than the average observation.
Of course, individual differences are hidden by the averaging process.
The absence of any significant differences between averaged observations
and predictions shows that the magnitude and phase of the predictions
are each quite good. It is particularly notable that the comparisons
seem about equally good on each of the three simulation days, although
the time variations are rather different on each day. Thus, no basic
model shortcomings are identified by this measure.
A more sensitive comparison of these averages is given in Figure II-5.
In this figure are shown differences between the pairs of data from
Figure II-4. These results seem to show no clear pattern. The variation
seems similar on two of the days but even on these days no simple inter-
pretable pattern is apparent. While this lack of pattern certainly does
not prove the validity of the Denver Model, it lends no support to any
inference of basic model error.
The cumulative frequency distributions of observed and predicted
concentration values are presented in Figure II-6. Since the occurrence
of concentrations of many pollutants has been observed to be log-normally
distributed (Larsen, 1971), the present data were plotted on log-normal
scales. It may be seen that although the distributions approach log-
normality (a straight line on these scales) at high ozone concentrations,
they certainly do not display this property over the whole range of values.
The curvature found indicates a more frequent incidence of low ozone
concentrations than would be consistent with a log-normal incidence of high
ozone concentrations. This finding supports the view that the concentration
distribution of ozone is strongly biased by background ozone not related
to the complex source distribution or to the random effects of wind
-------
O OJ
<_> O
c=
OJ O
C O
O
fsl _/ n
•r- O) -1
"O >
OJ i-
S- 01
O- en
-Q 9
c: o -2
-4 -
MEAN OF ALL STATIONS
O MEAN OF ALL STATIONS, 29 JULY 1975
D MEAN OF ALL STATIONS, 28 JULY 1976
0 MEAN OF ALL STATIONS, 3 AUGUST 1976
AVERAGE OF THE 3 DAYS
CO
0
10
__
11
r\_
12
ii
1
,. , f. , ,, in • n-j fstart hour!
Time of Day by Hourly Averaging Period St0p hour I
FIGURE II-5.
TIME VARIATION OF DIFFERENCES BETWEEN MEANS OF OBSERVED
AND PREDICTED OZONE CONCENTRATIONS
-------
40
30
20
I
Q-
Q-
C
O
£10
£ a
«= 7
0)
0 K
0 a
""* DDcnir
•^ rKtUIL
279 D;
-
-
~
-
-
1 I I f ' ' I
M« 99.1 •» •• 95
279 DATA PAIRS FROM 3 DAYS. 14 HOURS, 9 STATIONS
90 SO' 70 60 bO 40 JO 20 10 5 21 0,5 02 0 I
Probability of Exceedance of Given Ozone Concentration
FIGURE II-6. PROBABILITIES OF OZONE CONCENTRATION EXCEEDANCE
-------
60
speed and direction. Material that has been subjected to dispersive
effects for a long period of time becomes nearly uniformly distributed
in both space and time.
The probability-of-exceedance distribution for such material
approaches a step function: very high probability of exceeding lower
than background values, changing to very low probability of exceeding
higher values. If log-normally distributed concentrations of material
from more random sources are added to background concentrations, the net
distribution becomes qualitatively like that shown in Figure II-6.
A conclusion that might be drawn from this distribution is that although
background concentrations of ozone are not well-determined (low back-
ground concentrations are difficult to measure accurately), the higher
concentrations are more predictably distributed. The probability
appears to be quite high that the Denver Model is capable of reproducing
a representative distribution of the higher, and thus more important
ozone concentrations.
The correlation between observed and predicted ozone concentrations
may be seen in Figure II-7- The numerical correlogram in this figure
shows the number of occasions for which any particular combination of
observed and predicted concentrations were obtained. Observations and
predictions would be perfectly correlated if all points lay on the
diagonal through the origin. The points, in fact, lie generally along
the diagonal but with a substantial spread about that line.
The mean fractional deviation of points from the diagonal is given
in Figure II-8. This plot shows that aside from a small underprediction
at the lowest concentrations, the mean fractional deviations are both
modest and apparently random. These data do not confirm any finding
of systematic error in the simulations.
The distribution of points about the correlation line is presented
in Figure II-9. The data curve in this figure is a plot of sums of
numbers along diagonal lines in Figure II-7. In this plot the modal
-------
61
NAAQS
Limit
10
15
P=Predicted 0, Concentration (pphm)
FIGURE II-7.
MODEL PREDICTIONS CORRELATED WITH
INSTRUMENT OBSERVATIONS OF OZONE
(DATA FOR 3 DAYS, 9 STATIONS,
DAYLIGHT HOURS)
-------
62
•H.O.-
Root Mean Square Ozone Concentration (ppnm)
1/2
(Predicted)"|
2
(Observed)2 + (Predicted)2
FIGURE II-8. ESTIMATE OF BIAS IN MODEL PREDICTIONS AS A
FUNCTION OF OZONE CONCENTRATION
-------
0.18
0.16
O3QT
e£o:
CO ID
DEVIATION OF PREDICTED VERSUS OBSERVED POINTS
fROM PERFECT CORRELATION LINE (281 ONE-HOUR
AVERAGE DATA POINTS)
TRUE - INSTRUMENTAL)
EPA ACCEPTABLE MONITOR (MEAN BIAS = -3 PERCENT;
i 3 PPHM 0 95 PERCENT CONFIDENCE LEVEL)
(TRUE - INSTRUMENTAL)
.MAXIMUM PROBABLE ERROR (MEAN
BIAS = -8 PERCENT; ± 7 PPHM &
<>5 PERCENT CONFIDENCE LEVEL)
-8 -7 -6 -5 -4 -3-2-10 1 2
Difference (pphm)
en
CO
FIGURE II-9. MODEL PREDICTIONS COMPARED WITH ESTIMATES OF INSTRUMENT ERRORS
FOR OZONE (DATA FOR 3 DAYS, 9 STATIONS, DAYLIGHT HOURS)
-------
64
difference between prediction and observation is seen to be approximately
-1 pphm. The mean difference is, as noted before, -0.4 pphm. This is
~7 percent of the mean observed value of 5.6 pphm. The distribution is
roughly normal, although slightly skewed.
The data distribution in Figure II-9 is not a measure of prediction
"error," since the observations cannot be taken as perfect measures
of "truth." Burton et al. (1976) present an extensive analysis of
ozone monitoring techniques and measurement errors that might be
anticipated. Estimates from this analysis have been used to plot the
"true-instrumental" normal distribution curves also shown in Figure II-9.
The parameters used to define this normal distribution were:
> A bias of -8 percent (-0.4 pphm) because it is estimated
that "...The 1% and 2% neutral buffered potassium iodide
calibration methods produce results that exceed those by...
[absolute reference] methods, probably by a factor of
1.08 ± 0.08..." (Burton et al., 1976, p. 1-7).
> A standard deviation of 1.5 pphm for measurements made with
an EPA-acceptable monitor because "...for conditions of use
where a variety of ozone concentrations are measured, an
estimate of the maximum accepted measurement inaccuracy is
± 0.03 ppm..." This value of ± 3 pphm is taken as the 95
percent confidence limit, which is at twice the standard
deviation (ibid, p. 111-15).
> A standard deviation of 3.5 pphm for measurements made with
EPA-accepted monitors because "...the maximum uncertainty that
should be attributed to a value assigned to any one-hour-average
oxidant/ozone concentration — using any given instrument—in
the neighborhood of 0.08 ppm...[at] the 95% confidence limit
is...± 0.07 ppm." (ibid, p. 111-32).
> An area equal to the area under the predicted-observed
data curve.
-------
65
The sharper normal distribution curve in Figure II-9, which repre-
sents estimated performance consistent with EPA acceptance standards [40
C.F.R. §53.30 - §53.32 (1975], appears substantially equivalent to the
distribution of prediction/observation differences. These two curves
represent different, but related, parameters. The predictions and
observations each represent independent estimates of the true concentra-
tions. According to statistical theory, one would expect the error
distribution to be independent and, therefore, the distribution of
error differences to be the geometric sum of the individual error
distributions.
Put another way: we wish to know the expected "error" (i.e.,
difference from "truth") of the computed predictions. We do not know
the true concentrations. The observations are independent, but impre-
cise, measures of truth. The standard deviation of the difference
between the predicted and observed estimates of truth should be the
square root of the sum of the squares of the standard deviations of the
observational estimates and the predicted estimates. Since the 95
percentile point of the difference distribution is about ± 4.2 pphm
and twice its standard deviation is about 2.6 pphm, the difference
distribution is clearly not broader than the more optimistic instrument
error curve. These observations lead to the following possible
alternative explanations:
> The predicted results add nothing to the error, i.e., the
predictions are perfect.
> The observations are closer to the truth than estimated.
> The results are fortuitous; more data would show a wider
distribution of predicted/observed differences.
> Some combination of the above situations exists.
It would seem that the first alternative can be rejected out of
hand. Even though the difference distribution is no wider than the
instrument error distribution, the prediction error cannot be zero.
-------
66
While the second alternative cannot be rejected, it is highly unlikely to be
true since 2 pphm out of the assumed 3 pphm at the 95 percent limit is due
to the inherent precision and noise limits of the instrument. Drift and
interference errors have to be added to the 2 pphm and they cannot be
negligible.
The last two alternatives seem more plausible. A conclusion that
may be drawn is that the observational evidence is not precise enough to
establish confidence limits for the model predictions, but it is highly
unlikely that those limits would be wider than those on the observations;
in fact, it is likely that they would be considerably less. Note that if
the standard deviations of predicted and observed results (each from the
truth) are equal and independent, the standard deviation of the differ-
ences should be
-------
67
TABLE II-l. OCCURRENCE OF CORRESPONDENCE LEVELS OF PREDICTED
AND OBSERVED OZONE CONCENTRATIONS
Percent of Comparisons
Meeting Correspondence Level
Correspondence Level Both Predicted and
Between Predicted and Observed Pairs Comparisons Observed Cone. > 8 pphm
1) Factor of two (2P > 0 > P/2) 80%
2) Computed value is within ± twice
S.D. max. prob. inst. error
(95% level) of observed value 100
3) Computed value is within ± S.D.
of max. prob. inst. error
(95% level) of observed value 93
4) Computed value is within ± twice
S.D. of inst. errors by EPA std.
(95% level) of observed value 89
5) Computed value is within ± S.D.
of inst. errors by EPA std.
(95% level) of observed value 60
94%
100
90
77
37
3. Carbon Monoxide Comparisons
Predicted grid-cell CO concentrations are plotted along with station
observations in Figure 11-10. It shows much more variable correspondence
between predicted and observed results than was the case for ozone. In
Figure Il-lO(a), CAMP, Arvada, Overland, and Wei by show excellent corre-
spondence for central business district, suburban, and city residential
sites. N.J. Hospital, in a heavy traffic inner city zone, shows less good
correspondence and Northglenn, a suburban site, shows substantially no
correspondence. In the last case, the observed values varied by only
1 ppm during the daylight hours of 29 July 1975. While such measurements
could be valid, that possibility seems unlikely enough to at least suggest
instrument error.
-------
68
„. , Time of Day, By Hourly Interval ptart hour
— O Observed J J (.stop hour
A Predicted
(a) 29 July 1975
FIGURE 11-10. OBSERVED AND PREDICTED ONE-HOUR-AVERAGE CO CONCENTRATIONS (ppm)
-------
69
— -D
Time of Day, By Hourly Interval -r r
Observed Lst°P hour J
""'"'•*•"' (b) 28 July 1975
Predicted
FIGURE 11-10. (Continued)
-------
70
--O Observed
—A-— Predicted
Time of Day, By Hourly Interval "
(c) 3 August 1976
FIGURE 11-10. (Concluded)
-------
71
As Figure Il-lO(b) illustrates, C.A.R.I.H., Arvada, Overland, and
Welby showed observed CO concentrations of 2 ppm or less on the entire day.
These concentrations are so low that no meaningful analysis can be made of
them. The two inner city sites, CAMP and N.J. Hospital, showed substantial
observed values of CO that have very little correlation with the predicted
values. Figure II-lO(c) also shows both good and poor comparisons with
observed concentrations.
The more variable correspondence of predicted and observed CO concen-
trations, as compared to ozone, is not unexpected. Although CO and the
ozone precursors HC and NO have much the same source distribution,
A
ozone is not produced until substantial atmospheric mixing has occurred.
Ozone peaks in well-mixed clouds while CO concentrations decrease contin-
uously as the material moves away from the vehicle exhaust pipe. Thus,
CO concentrations tend to be quite variable on a scale comparable to
that by which streets and traffic lanes may be measured. This scale is
exceedingly small compared with the size of grid cells in a model simulating
photochemical effects.
Estimates of local variability of CO concentrations near urban
streets have been presented by Ludwig and Kealoha (1975). Figure 11-11,
reprinted by them from Georgii et al. (1967), shows that surface CO
concentrations in an urban street canyon may typically differ by 100
percent from one side of the street to the other. Figure 11-12, reprinted
from Johnson (1974), shows even greater side-to-side differences (up to
ten-to-one ratios) from dye tracer experiments; these results should be
representative of nonreactive pollutants.
Ludwiq and Kealoha (1975, pp. 83ff) present empirical expressions
indicatinq that the side-to-side CO ratio in a street canyon should be
approximately equal to the ratio of the street width to the offset distance
of the monitoring instrument from the street. From Eqs. (8a) and (8b)
of this reference, we can write
-------
72
40
30
20
I
52
UJ
I
10
LEEWARD SIDE
OF STREET
WINDWARD SIDE
OF STREET
\
\
\
\
10
CO — ppm
15
20
Source: Georgii et al. (1967).
FIGURE 11-11.
THE VERTICAL DISTRIBUTION OF CO CONCENTRATION
IN A STREET CANYON WITH TRAFFIC VOLUME OF
1500 VEHICLES PER HOUR
-------
73
RUN
NO.
0—0
o—a
11
is
19
140-)-
120
100
80
60
40 +
--20
100 80 60 40 20 0 20 40 60 80 100
DISTANCE FROM HIGHWAY CENTERLINE — meters
w— r —E
HIGHWAY 33
(a) Distributions for runs with cross-
roadway winds from the west (note
plotting error—points at 5 m
west should be at 15 m west)
100 30 60 40 20 0 20 40 60 80. 100
DISTANCE FROM HIGHWAY CENTEHLINE — meters
W urj — £
HIGHWAY 33
(b) Distributions for runs with very
light winds from the southwest
100 80 60 40 20 0 20 40 60 80 100
DISTANCE FROM HIGHWAY CENTERLINE — meters
W — V =
HIGHWAY 33
(c) Distributions for runs with winds
from the southeast sector
100 80 60 40 20 0 20 40 60 30 100
DISTANCE FROM HIGHWAY CSNTERUNE — meters
HIGHWAY 33
(d) Distributions for runs with very
light winds from the east
Source: Johnson (1974).
FIGURE 11-12. MEASURED TRACER CONCENTRATION DISTRIBUTIONS
-------
74
C d.w. = \2 + x2 + z2
C u.w. W
where C d.w. is concentration on downwind side (due to street canyon
sources), C u.w. is concentration on upwind side (due to street canyon
sources), x is horizontal distance from receptor to nearest lane of
traffic, z is elevation of receptor, and W is street width. Figure 11-13
(reprinted from Ludwig and Kealoha's Figure 29) shows that even an eight-
hour-average CO concentration at a point in a city may have one-half of
its value contributed by sources within a few meters (i.e., within the
street canyon under observation).
An inference that may be drawn from the CO data presented in Figure
11-9 is that correspondence between observations and predictions on some
occasions is not fortuitous—this should be the case when monitors are not
strongly affected by microscale meteorology—but that such correspondence
should not be expected on all occasions. Thus the Denver Model may do
as well at predicting CO as ozone on a grid-cell-average basis, but instru-
ments are less capable of confirming the CO predictions. Since no
adequate measure of true grid-scale CO concentrations is available, model
performance with ozone is probably the best measure of model performance
with CO.
4. Nitrogen Oxide Comparisons
Comparisons of observed and predicted NO are presented in Figure
11-14, and of N02 in Figure 11-15. Obviously the available observational
data were totally inadequate for validation of the Denver Model for these
species. NO was never high enough at Parker Road for the instruments to
register dependably. This was also true for N02 at Green Mountain on
28 July 1976.
-------
75
2.0
m
cr
t-
z
o
o
z
o
z
UJ
UJ
(T
K
V)
u.
o
<
cr
z
o
m
0.7
0.5
cr
UJ
T
0.3
0.2
I I I I I I I
10 30 50 70 90
PERCENTAGE GREATER THAN ORDINATE VALUE
Source: Ludwig and Kealoha (1975).
FIGURE 11-13.
FREQUENCY DISTRIBUTION OF THE RATIOS OF STREET
CONTRIBUTIONS TO CITYWIDE CONTRIBUTIONS FOR THE
HIGHEST 8-HOUR-AVERAGE CO CONCENTRATIONS
-------
76
5
3 AUGUST 1976
-j o PARKER ROAD
3 AUGUST 1976 _|
~~° CAMP
28 JULY 1976
o PARKER ROAD
A-
9
TO
10
rr
n
TI
1
T
O Observed
A Predicted
Time of Day, By Hourly Interval \start hour]
[stop hour J
FIGURE 11-14.
OBSERVED AND PREDICTED ONE-HOUR-AVERAGE
NO CONCENTRATIONS
-------
77
3 AUGUST 1976
o GREEN MTN
O Observed
___A Predicted
Time of Day, By Hourly Interval
FIGURE 11-15. OBSERVED AND PREDICTED ONE-HOUR-AVERAGE N0?
CONCENTRATIONS
-------
78
NO is susceptible to the same microscale meteorological effects
described above for CO. In addition,-predicted concentrations of
NO, NCL, and ozone are consistent with each other by accepted chemical
principles; that is, the rapid reactions between these species ensure
that they are near equilibrium. Observed concentrations of these species,
however, are often inconsistent; that is, they are in a local equilibrium
which may be quite far from equilibrium at the grid scale. Thus we con-
clude that validation should rest on the strongest evidence (ozone) and
that internal consistency of the model is more defensible than external
consistency with data which are not self-consistent.
D. COMPARISON OF THE DENVER MODEL WITH OTHER MODELS
Validation studies of other models in various stages of development
have been examined extensively by Roth et al. (1976) and will not be exan-
ined here. No photochemical model has been recommended by EPA, however, EPA
does recommend the use of rollback [40 C.F.R. §51.14(c)(4)(1975)] for ozone
control strategies. EPA is currently considering the recommendation of
alternatives in this context. Therefore, it will be informative to
contrast some performance measures for models recommended for use by
EPA with performance measures of the Denver Model.
The "Appendix J" rollback technique for assessing photochemical oxidant
control strategies has not been verified. It does not seem to be subject
to verification since verification would involve the trial of an oxidant
control strategy for an urban region. Years would be required before an
"observation" would be available to compare with a "prediction" and the
process would involve changes that might invalidate the prediction.
Roth et al. (1976) point out that:
In a critique of the Appendix J relationship, Dimitriades
(1975, p. 9) cited the following strengths of this approach:
-------
79
> The graphical relationship is based on actual atmospheric
data.
> The hydrocarbon-oxidant dependence qualitatively agrees with
smog chamber data....
> The relationship can be improved through the acquisition of
additional data.
Furthermore, the required calculations are relatively simple and
can be carried out graphically, in contrast to more sophisticated
relationships...whose governing equation(s) must be solved using a
digital computer.
Despite the advantages cited above, the Appendix J relation-
ship has been criticized for a number of reasons. In particular,
the following shortcomings have been identified:
> The relationship has not been validated.
> The effects of NO emissions changes are not explicitly
treated. x
> The relationship is based on a limited amount of historical
air quality data. The accuracy of the lower and upper portions
of the curve is limited because of experimental errors and the
scarcity of data points, respectively.
> The influence of local meteorological and emissions patterns
is not considered.
> The data base upon which the curve is based relies on
measurements of precursor and oxidant levels taken at the
same station.
> Background concentration levels are assumed to be negligible.
Each of the shortcomings listed above has particular relevance to
Denver, but the most important in terms of predicting Denver's future air
quality is the second. It is well known (Dimitriades, 1972, 1973, 1975;
Merz et al., 1972; Trijonis, 1972; Paskind and Kinosian,1974; Bailey, 1975)
that ozone production depends as strongly on NO as it does on hydrocarbons
A
The nature of this dependence is shown in Figure 11-16. The ozone
concentrations plotted there were computed using the chemistry portion
of the Denver Model. The simulations were of smog chamber experiments
and thus do not represent conditions in a real urban atmosphere. The
computations assume no wind, constant sunlight, and only initial charges
of the precursors HC and NO (i.e., no continuous emissions and no back-
J\
ground ozone).
-------
80
0.9 GRAMS/MOLE NO EMISSIONS STANDARD
EPA REGION VIII EMISSION FACTORS
EMISSIONS INVENTORY
FOR YEAR «-1976
0.6
0.5
-£0.4
0 Q-
C X
o <
CJ O
0.3
OJ
-(-»
g.0.2
U3 O
0.1
PREDICTED MAXIMUM
ONE-HOUR-AVERAGE
OZONE CONCENTRA-^.
TION (ppm)
(0.24)
ot-
0
AP42 EMISSION FACTORS
ISOPLETHS OF MAXIMUM OZONE CONCENTRA-
TIONS COMPUTED FROM SIMULATIONS OF
SMOG CHAMBER RUNS WITH THE CARBON-
BOND MECHANISM
POINTS COMPUTED WITH CARBON-BOND
MECHANISM IN DAQM SIMULATION RUNS
FOR 3 AUGUST 1976 METEOROLOGY
MAXIMUM OZONE CONCENTRATION
(ppm)
0.05 0.10 0.15 0.20 0.25
2000
(0.07),
(0.24)
1976 STAGE II HC CONTROLS
30% REDUCTION
IN 1976 HC AND
(0
0.5 1.0 1.5 2.0 2.5
6-9 a.m. Reactive Hydrocarbon Concentration (ppmC)
(computed near CAMP station by DAQM)
FIGURE 11-16.
MAXIMUM OZONE COMPUTED BY CARBON-BOND SIMULATIONS
OF SMOG CHAMBER EXPERIMENTS AND BY DENVER MODEL
FOR VARIOUS ASSUMED FUTURE EMISSIONS
-------
81
In spite of any nonrepresentativeness of such ozone isopleths, the
Denver Model simulation results are plotted in this figure. The 1976
Denver point was determined by the predicted 6 a.m. to 9 a.m. HC and NO
A
averaged at N.J. Hospital and at the downtown CAMP station. Points for
other years were determined similarly. To account for the different
conditions between the simulations used to derive the isopleths and the
model's point computation, the isopleths were scaled to give exact cor-
respondence to the 1976 point.
It will be seen that all other points correspond very well also.
The high NO points for the years 1985 and 2000 represent a relaxation of
/\
NO emission factor requirements from 0.4 grams per mile to 0.9 grams per
A
mile. The number next to each point is the maximum ozone concentration
for the year indicated as computed by DAQM. The low and high NO assump-
A
tions give quite different ozone predictions but rollback would have
predicted identical values. Note that the isopleths do not constitute a
prediction system since computed NO and HC are required to locate points.
/\
It is not practical to use ozone observations to locate points since the
maximum predicted ozone concentration was almost 50 percent greater than
any observed value. Use of the isopleths as an ozone prediction method
would give a 33 percent reduction (in this example) in ozone predictions.
EPA does recommend (even requires) the use of grid-based advection
dispersion models for the analysis of S0? effects on air quality. Two
of these, AQDM and COM are provided to any interested user by EPA. In arldi
tion, EPA (Turner et al., 1973) has done work on comparing the performance
of these models with "box" models (regional-averaged dilution models).
Turner's analysis compared the performance of the models for SO^ in New
York City on annual average predictions.
Even though the SAI Denver Model was used to generate one-hour
concentrations, we list in Table II-2 its performance measures along with
information from (Turner et al., 1973) On AQDM, COM, and two box models
-------
TABLE 11-2. VALIDATION MEASURES OF VARIOUS MODELS
[differences = (predicted-observed)/observed]
S09 Annual Averages*
AQDM
COM
Gifford (box)
Hanna (box)
Mean
GO Hourly Averages
SAI Denver Model
Number
of
Comparisons
75
75
75
75
RMS
Difference
0.90
0.39
0.61
2.44
Mean
Absolute
Difference
0.68
0.27
0.53
1.32
Lower
Limit
Difference
0.64
0.87
1 .30
1.07
Upper
Limit
Difference
2.30
1.23
0.22
9.13
Difference
Range
2.94
2.10
1.52
10.20
75
279
1.08
0.31
0.70
0.35
0.97
0.90
3.22
1.16
4.19
2.06
oo
* S0? model measures were computed from data in Turner et al. (1973).
-------
83
(Gifford's and Hanna's). This table shows that the Denver Model valida-
tion measures are based on many more data points than are the S02 valida-
tions (the Denver Model has one point per simulation hour at each station,
the others achieve the performance they do by averaging a year's data
into each comparison). Nevertheless the Denver Model displays more
favorable performance by almost every measure.
E. CONCLUSIONS
Conclusions that may be drawn from this validation study are as
follows:
> The Denver Model is a very good predictor of one-hour-
average ozone concentrations in grid cells in the Denver
reg i on.
> The model's ozone predictions at any given station probably
have at least as narrow an error distribution as do measure-
ments at the same station.
> If predicted and observed concentration estimates are
equally accurate, Denver Model predictions can be expected
to be within a factor of v^"of true concentrations 80 per-
cent of the time, and its predictions of exceedances (i.e.,
more than 8 pphm) could be expected to be within that
factor 94 percent of the time.
> The accuracy of predictions by the Denver Model of regional
maximum ozone concentrations should exceed the accuracy of
predictions of concentrations at specific stations.
> The Denver Model is likely to underpredict ozone concentra-
tions when ozone and/or precursor concentrations initially
or on an inflow boundary are unknown but higher than
expected.
-------
84
The model predicted maximum ozone concentrations that are
nearly 50 percent higher than any observed in the 1976
summer study, yet predictions for grid squares containing
monitoring stations are consistent with observations at
monitoring stations. It could thus be expected that obser-
vations up to 50 percent higher than in the 1976 study
could be made without justifying any inference of worse
air quality.
-------
85
III BASE CASE STUDIES
The major goal of this project was to predict the effect of the
proposed wastewater treatment facilities on future air quality in Denver.
The most important way that the treatment plants would have an effect on
air quality in Denver is by supporting the increase in pollutant emis-
sions associated with urban growth and development.
Since urban growth and development and the attendant increase of
pollutant emissions depend on factors in addition to the availability
of sewage systems, and, in particular, since urban development depends
on decisions not yet made, predictions of future growth are not simula-
tions of reality in the sense that physico-chemical models are. To
analyze Denver's future air quality then becomes an exercise in hypothe-
sis. The question is not so much "What will Denver's air quality be?"
but rather, "What will Denver's air quality be if ...?" Such a question
inevitably evokes the question "What will Denver's air quality be if
not ...?" Thus there is no single, unique analysis of future air quality.
The variation of air quality with changes in hypotheses is an important
aspect of the analysis, but to define the changes there must be an initial,
neutral, or "base" state from which to measure departure.
The base state of Denver's air quality for the future years under
study, 1985 and 2000, is taken to be a state that would possibly be
realized if growth patterns are not affected by the results of this
analysis. Since plans and projections for Denver's future have been
formulated, examined, and accepted in the absence of this analysis,
those plans became the "base" of the base case. The most questionable
aspect of the base case definition is the assumption that vehicle emis-
sions will be as described in EPA document AP-42 (Supplement 5). Recent
work by the staffs of EPA Region VIII and Colorado APCD led to the con-
clusion that emissions will be much higher than AP-42 values unless very
control measures are instituted (see discussion in Chapter V).
-------
86
The generation and features of future plans for Denver are given in
detail in Appendix C. Data files necessary to carry out the base case
simulations were provided by Colorado Division of Highways and the Air
Pollution Control Division of the Colorado Department of Health. Descrip-
tions of these files are given in Appendix B.
Air quality, on any given day, depends on detailed wind and turbu-
lence patterns as well as on emissions. Since these patterns cannot be
predicted, even in the sense that emissions can, it is necessary to define
representative meteorology. For the base case studies, the meteorology
for the two summer 1976 days that were selected for validation was used.
Similarly the COM simulations of annual average conditions used the same
STAR meteorological data as were used for 1976 simulations. Since the
1976 summer days chosen led to estimated worst-case ozone concentrations,
their meteorological conditions can be expected to be representative of
the worst conditions in the future years also.
A. BASE CASE OZONE CONCENTRATIONS PREDICTED BY THE DENVER MODEL
1. Station Comparisons
Ozone concentrations predicted by the Denver Model for grid squares
containing monitoring stations are plotted for each base case year in
Figure III-l. The area between successive concentration traces is shaded,
with hatching between 1976 and 1985 and cross-hatching between 1985 and 2000
The basic nonlinear character of the photochemical oxidant problem
is clear. Even though the wind and turbulence patterns are identical,
the 1985 concentrations are not a constant fraction of the 1976 concen-
trations. The ratio of those concentrations varies with time at a given
station and varies from station to station at a given time. The differ-
ences in concentration with a constant wind pattern occur because with
different emissions a parcel of air arrives over each pollutant source
area with different species concentrations, and thus the parcel has a
unique set of reactions along its trajectory for each year studied.
-------
87
9 10 11 12
To TT Tz: 1
Time of Day, By Hourly Interval
Reduction 1976-1985
198b-2000
(a) Meteorology for 28 July 1976 Assumed
FIGURE III-l. REDUCTION IN PREDICTED OZONE CONCENTRATIONS
(pphm) AT DENVER STATIONS DUE TO PREDICTED
FUTURE EMISSIONS CHANGES
-------
88
15
T1n,e of Day, By Hourly Interval
Reduction 1976-1985
ES3 Reduction 1985-2000
(b) Meteorology for 3 August 1976 Assumed
FIGURE III-l. (Concluded)
-------
89
In general, the absolute and fractional reductions are greater in
the 1976 to 1985 period than in the 1985 to 2000 period. Peak reductions
in the earlier period average 44 percent. Peak reductions in the latter
period only average 23 percent. Absolute reductions of the peaks are much
more substantial than are reductions of lower concentrations. As noted
above, the peak predicted concentrations at measurement stations are closer
to the peak observed concentrations than to the absolute peak concentra-
tions, which are predicted for locations other than measurement stations.
In the base case runs for 1976, the peak predicted ozone concentration
at any station was 16 pphm on 28 July 1976, the same as the observed peak.
The peak predicted anywhere was 22 pphm. On 3 August 1976 the station
peak predicted was 15 pphm; observed was 17 pphm. The concentration
predicted anywhere was 24 pphm.
Because of the nonlinearity of the photochemical process, there is
no assurance that a station might not experience an absolute peak on some
future day, but the ratio of the peak concentration predicted at a station
to the peak predicted anywhere is 0.77 in 1985. This is about the same
as the 1976 value of 0.70. In 2000 the ratio is higher (0.85), but it
is still significantly different from 1.0. The peak predicted ozone
concentrations at stations are 10 pphm in 1985 and 6 pphm in 2000.
2. Area Comparisons
Measures of the peak predicted ozone concentrations and the area!
extent of several levels of NAAQS exceedances are presented in Figure
III-2. In this figure, the area in which a given exceedance of the ozone
NAAQS is predicted is plotted as a function of time of day. Exceedance
levels plotted are 0 pphm (i.e., standard just met, no violation),
and even numbers of pphm above the standard.
In 1976 the NAAQS for ozone is first reached between 9 and 11 a.m.
each day. About one-half of the region experiences ozone concentrations
at the NAAQS or higher at a peak hour between 12 and 2 p.m. The area
of exceedance is greatly reduced or eliminated by late afternoon.
-------
Year 1976 Emissions
Year 1985 Emissions
Year 2000 Emissions
£ »[•
2*01-
zool-
120 t-
i
30 |-
COKEKTMTtOIC C » 00™"
0 10 pen
O 12 pom
OZOM CWUNIHA1 ION: D B piikii
0 10 H,h.
O 12 ii^ilwi
OZIml CWILIMIRAIICW- 1-1 H l'll'»
0 10 Vl'tm
S TTJ TT 1? T
2 3 J_ 5
TTT TT
1 10 II \2 12 14 56
TffTTTTTTTTS 6 7
Time of Day by Hourly Interval
(a) Meteorology for 28 July 1976 Assumed
FIGURE III-2. SIZE OF AREA IN WHICH PREDICTED OZONE CONCENTRATIONS EQUAL OR EXCEED
GIVEN VALUES FOR YEARS 1976, 1985, AND 2000
-------
Year 1976 Emissions
Year 1985 Emissions
Year 2000 Emissions
10 II 12
II I? I
2 1 4
) 4 S
mt
M
M
M
•
•
X
-A
1
480
MC
too
360
320
280
Z»0
200
160
120
ao
«0
5 e
I 7
OZONE COKCEHTHATIOH: Z 8 jpm
£. 10 ppnn
C 12 torn
-
\-
/^\
/ \
/^x^xl
/ ,/ \\
, .( /^\\\ , . ,
4B
441
404
J6H
3?(t
i?Bd
24U
eou
IM1
I
U(l
411
9 13 11 12 1 2 ; t 560
•mTTTTTrT * t JT(
-
_ 1 1 J.
9 10 II
i* IT U
Timo nf Dav/ hu Hmivlv/ Tntovwal Sta.rt nour
OZOK COIIC[NIIIMI(W. n 8 ff<"
I _.l J
3451
» 5 8
(b) Meteorology for 3 August T976 Assumed
FIGURE III-2. (Concluded)
-------
92
In 1985, not only is the peak predicted ozone concentration much
reduced from the 1976 level, but the area affected by 8 pphm ozone is
reduced by 40 to 60 percent and the area affected by 10 pphm ozone is
reduced by 70 to 75 percent. The area affected by 12 pphm is nearly
eliminated.
By 2000 predicted areas of exceedance are only about 1.5 percent of
the modeling region, and are predicted to occur for only a single one-hour
averaging period.
B. BASE CASE CARBON MONOXIDE CONCENTRATIONS PREDICTED BY THE DENVER MODEL
The most severe carbon monoxide episodes in Denver typically occur
in winter. Carbon monoxide is emitted mainly from traffic sources (see
Appendix B), and peak concentrations of CO are found very close to the
traffic lanes from which they are emitted. For these reasons, exceedances
of the CO standard tend to occur during the morning rush hour when the
mixing layer is very thin. Since sunrise does not occur until well into
the rush hour in winter, but much earlier in summer, there is less solar
heating to increase the thickness of the morning mixing layer in winter.
The confinement of CO in a thin mixing layer leads to high concentrations.
The winter high pollution day for which we were supplied meteorological
and air quality data was 15 November 1974. DAQM simulation runs were
made with meteorology from this date and emissions for 1976, 1985, and
2000. These runs were made with the photochemistry in the model
suppressed to save computer time since CO is essentially inert.
The minimum mixing depth corresponding to the meteorological
conditions was estimated to be only about 65 feet, much less than the
summer minimum of 165 feet. The CO concentration peak, however, occurs
in the region of densest traffic. In Denver the densest traffic is
-------
93
in the central business district, where high-rise construction ensures
the occurrence of eddies and turbulence such that the mixing layer can
hardly be thinner than the I6b-foot summer layer. The mixing layer was
thus assumed to be no thinner than 165 feet in the central business
district.
Both one-hour-average and eight-hour-average predicted maximum
CO concentrations were determined for comparison with the NAAQS of
35 ppm (one-hour-averaqe) and 9 ppm (einh^-hnnr-awpraqe). The neak con-
centrations, their times of occurrence, and the coordinates of the grid
cell in which they occurred are given in Table iii-i.
TABLE III-l. CO CONCENTRATIONS PREDICTED BY THE DENVER MODEL.
15 November 1974 Meteorology Assumed
Peak Concentration (ppm)
Year
1976
1985
2000
One-Hour-Average
26
17
4
Eight-Hour-Average
15
10
3
Time of
Occurrence
7-8a.m.
5a.m. -1p.m.
8-9a.m.
5a.m. -1p.m.
6-7a.m.
5a.m. -1p.m.
Grid Gen
(x,y Coordinates)
42,44
42,44
42,44
42,44
56,58
56,58
These results show that, even averaged over four square miles,
violations of the eight-hour CO standard now occur. By 1985 only a
minimal exceedance is predicted, and by 2000 the Denver atmosphere is
predicted to be in good compliance if the assumptions of the emissions
projections are correct. (Achievement of AP-42 emission factors is the
most important assumption.) Since CO peaks occur very close to the
sources, it may be presumed that both one-hour and eight-hour violations
that cannot be resolved on the simulation grid will occur.
-------
94
Both the 1976 and 1985 peaks are predicted to occur at grid cell
42,44, which is in the vicinity of the C.A.R.I.H. monitoring station.
In the year 2000 the peak is found at 56,58, in the vicinity of Barr
Lake along Interstate 805.
C. BASE CASE N02 AND PARTICULATE CONCENTRATIONS PREDICTED BY THE COM
The COM was run for the years 1974, 1980, 1985, 1990, and 2000.
The receptor network used was a 15 x 15 grid of receptors spaced two
miles apart. Isopleth maps were prepared by a computer plotting program
that interpolated contours between the receptors; isopleths of predicted
N02 concentrations are shown in Figure III-3. These maps may be compared
with Figures 22 through 31 of the Joint Regional Planning Program (JRPP)
Air Quality Assessment Statement (Colorado Division of Highways, 1976).
These isopleth maps are given in units of yg/m^, whereas in the JRPP
document N02 concentrations are given in ppm.
The isopleth maps show that the highest concentrations of both N02
and particulates are in the vicinity of major highways and the downtown
area. This observation leads to the conclusion that the major source of
these pollutants is traffic. Other features of the maps are the N02
peak at Stapleton Airport, and particulate peaks in the vicinity of
Golden and the intersection of U.S. 285 and State Highway 121.
The National Ambient Air Quality Standard (NAAQS) for N0£ is 0.05
ppm for an annual average, which converts to 82 yg/m^ at Denver's altitude
and annual average temperature. For annual average particulate concen-
trations the NAAQS is 75 yg/rn^, but the Colorado state air quality standard
3
of 45 yg/m is more stringent.
Figure III-4 shows how N02 concentrations predicted by COM change from
1974 to 2000. (The boundaries of the subregions listed in Figure III-4 are
shown in Figure IV-1.) In 1974 no exceedance of the NAAQS is calculated by
the model: the maximum annual average concentration given by the model is
-------
95
LU
SOUTH
(a) Year 1974
FIGURE III-3.
ISOPLETHS IN yg/m^ OF CALIBRATED ANNUAL AVERAGE N02 CONCENTRATIONS
PREDICTED BY COM FOR VARIOUS YEARS
-------
96
I l__l J L-!J_...... J__. A J_....L__1 I I L._.l __iL_ J 1 I I 1 .. J.... .1... J
(b) Year 1980
FIGURE III-3. (Continued)
-------
97
f, i ,~i r~i ~r i i
NUn i hi
(c) Year 1985
FIGURE III-3. (Continued)
C'.'1
cr
LtJ
-------
98
NORTH
10 20
>3 _
CO
—i—r —r—r—r—T r
._r__j—r_r—, 1 , r ,— r—1—) 1—T , f r
SOUTH
(d) Year 1990
FIGURE III-3. (Continued)
-------
99
NORTH
SOU I H
(e) Year 2000
FIGURE III-3. (Concluded)
-------
100
en
c
o
90
80
70
C
a;
o
J 60
OJ
o
8.
£ 50
O)
-------
101
CO
en
n
C
O
(O
70
60
o 50
OJ
O
Oj 40
an
to
s_
-------
102
79 ug/m3. This maximum is, in fact, determined by the one calibration
point available. It is not, therefore, a prediction in any real sense.
In 1980 some exceedances are predicted, and these persist through 2000.
Table 111-2 shows the number of square miles over exceedance of the NAAQS
is predicted. This table contains data from our use of COM from the
Air Quality Assessment Statement (CDH, 1976) and f»-om thP 1976 reoort
to the Public by the Colorado Department ot Heal Lh(1976b).
The agreement between the results of this COM study and those of
previous studies is fairly good. The discrepancies are probably the result
of the use of different arrays of receptors and the interpolation of
contours by hand in the previous studies.
The predicted exceedances of the NAAQS for N02 should be viewed
with skepticism in the light of the propensity of the COM to mispredict N0?
as discussed in Appendix A. The COM predictions are made assuming that all
NOX is in the form of N02- Since the emissions are mainly NO, and are
not all converted to N02, the calibration used to match the predictions
and observations can only be valid at the place and time for which the
calibration constant was determined. The Denver Model distinauishes between
NO and N02 and could predict annual average N02 concentrations, but extensive
computation would be required. Denver Model results for single days indi-
cate, however, that the assumption underlying COM that concentrations are
proportional to emissions is invalid for NOp- DAQM base case simulations
were examined for inferences as to N02 dependence on emissions. Time vari-
ations of N02 concentrations predicted by the Denver Model at CAMP, which
has the peak N02 concentration according to COM, are presented in Figure
II1-5 for 3 August 1976 and 2000. Also presented in this figure is the
1976 N09 concentration multiplied by the 2000/1976 NO regionwide average
L- X
ratio of NO emissions. This represents a prediction consistent with COM
X
assumptions if CAMP N02 depends on sources at a distance, but assuming all
N0x acts as N02_ If CAMP N02 depends more on local sources, the appropri-
ate 200/1976 emissions ratio would be that for the grid cell containing
the CAMP location. This ratio is about 1.015. Although this number is
very near unity, it is greater than one and its use would lead to a
-------
103
TABLE III-2. AREA OF EXCEEDANCE OF NAAQS FOR N02
(in square miles)
Predicted by COM
Assessment Statement* Report to Publict
2.1 1
11
19.1 16
14.7
23.0
*Colorado Division of Highways (1976).
tColorado Department of Health (1976).
Year
1974
1980
1985
1990
2000
in This Project
0
9
14
4
13
-------
1 -
D Results for 1976 NO Emissions
A
Results for 2000 NO,, Emissions
Results for 1976 NOX Emissions
Multiplied By Ratio of Regional
NOX Emissions—Year 2000/Year 1976
D 1
S
I
I
I
I
5_
6
6
7
8
9
9
10
li
11
12
12
T
Time of Day, by Hourly Interval
3 4
start hour
4_
5
E.
6
stop hour
FIGURE 111-5. N02 CONCENTRATIONS AT THE CAMP SITE IN DOWNTOWN DENVER PREDICTED
BY THE DENVER MODEL. 3 Auanst 1Q7fi mPtenrol ?r«" =
-------
105
prediction of increased NCL. It can be seen that the photochemical simu
lation leads to a prediction of a 22 percent decrease in peak N0? and an
18 percent decrease in average N02.
This analysis does not prove that annual average N02 concentrations
would decrease with an increase in NO emissions, but it certainly sug-
A
gests this possibility and lends credence to the skepticism noted above.
N02 13-hour-average predictions using 3 August computations for CAMP for
each year calibrated with present annual average at CAMP are presented
in the summary in Table III-3 at the end of this chapter.
Figure 111-4 shows the maximum and average N02 concentrations pre-
dicted by the COM for the city of Denver and various subregions around
Denver. (These subregions were chosen for study in connection with the
sensitivity analysis in Chapter IV.) These figures show that the COM
predicted maximum in the entire modeling region is 97 yg/m in 1985 in
central Denver. We believe that, considering the likely amount of over-
prediction of N02 by the COM, this result does not represent a firm pre-
diction of an exceedance of the NAAQS.
It may be seen from Figure III-4 that, while N02 concentrations
generally increase between 1974 and 2000, concentrations in some areas
decrease between 1985 and 1990. This effect is not related to total
population, as can be seen from Figure III-6, which shows the projected
populations of these areas. All areas show a continuous increase over
these years, with all except Denver and Jefferson County Urban having
a growth rate of about 2-1/2 percent per year.
The break in the N02 trends in 1985 most probably stems from one
of two causes:
> The emissions inventories were made up by different agencies
using different splits between single and multi-family
dwellings (although with the same total population projections).
-------
lUb
7
6
5
4
3 L
10
9
8
7
6
O DENVER
c? BROOMFI ELD/WESTMINSTER/ARVADA
D LAKEWOOD
O AURORA
0 S. METRO
0 NORTHGLENN/THORNTON
A JEFFCO URBAN
1975 1980 1985 1990
Year
1995
2000
FIGURE III-6. POPULATION PROJECTIONS FOR THE DENVER AREA, 197^ to 2000
-------
107
By 1985-1990, NOX emissions controls on light-duty vehicles
will have reached their maximum effectiveness. After this,
the increase of overall VMT will cause NOX emissions to rise
again.
Figure III-7 shows isopleths of particulate concentrations predicted
by the COM for 1974 through 2000. The state air quality standard is
exceeded over the entire area for all years. The NAAQS is exceeded
over 460 square miles in 1974 (51 percent of the region), 548 square
miles (61 percent) in 1980, 668 square miles (74 percent) in 1985,
716 square miles (80 percent) in 1990 and 724 square miles (80 percent)
in 2000.
Figure II 1-8 shows the maximum and average particulate concentra-
tions predicted for the various subregions by the COM. Both Lakewood
and Jefferson County Urban show large increases between 1985 and 1990.
The reason for this is not clear; it is not attributable to the increased
construction activity attendant on the area's rapid growth, since
Figure III-6 shows that population increases in all periods. Another
contributing factor could be the difference in population assumptions
used by the two groups who assembled the emissions inventories.
Another way of looking at these base case runs is to examine the
incremental changes in concentration levels which occur from one year
to another. Figures III-9 and 111-10 show contours of these incremental
changes for N0£ and particulates.
Examination of Figure III-9 shows that during 1974 to 1980 the larg-
est increases are forecast for Commerce City, Stapleton Airport, and S.E.
Denver. In 1980 to 1985, the largest increases are expected at Stapleton
Airport and in S.E. Denver again, and also in S.W. Denver. Between 1985
and 1990, it is predicted that concentrations in a large area surrounding
downtown Denver will actually decrease, while substantial increases are
expected in the northeast, around the Rocky Mountain Arsenal. In 1990 to
-------
108
NORTH
SOUTH
FIGURE III-7.
(a) Year 1974
ISOPLETHS IN pg/m3 OF CALIBRATED ANNUAL AVERAGE PARTICULATE
CONCENTRATIONS PREDICTED BY COM FOR VARIOUS YEARS
-------
109
NOPT
L._L I l__ J I
__| L_j 1 I J I
10
SOUTH
1 L.....L
20
^°
(b) Year 1980
FIGURE III-7. (Continued)
-------
no
L.. _..!. ___L 1_ „ I .. I II'!
SOU I'M
(c) Year 1985
FIGURE III-7. (Continued)
-------
in
NORTH
(d) Year 1990
FIGURE III-7. (Continued)
-------
112
SOUTH
(e) Year 2000
FIGURE III-7. (Concluded)
-------
113
240
220
200
180
160
cn
c
o
o
3
O
140
t-
(O
O-
S,
n3
O)
to 100
^
c
80
60
40
O DENVER
0 NORTHGLENN/THORNTON
._, RROOMFIELD/WESTMINSTER/
0 ARVADA
D LAKEWOOD
Q AURORA
5 S. METRO
A JEFFCO URBAN
1975
1980
1985
1990
1995
2000
Year
(a) Maximum Concentration in Any Grid Lell Within Subregion
FIGURE III-8. PARTICULATE CONCENTRATIONS PREDICTED BY COM FOR 1974 TO 2000
-------
114
180
en
o
•r- 160
O)
CJ
OJ
CT
(O
S-
O)
140
01
+J
NORTHGLENN/THORNTON
BROOMFI ELD/WESTMINSTER/ARVADA
• ENTIRE REGION
0 S. METRO
O AURORA
A JEFFCO URBAN
1975 1980 1985 1990
Year
1995
2000
(b) Average Concentration in Subregion
FIGURE III-8. (Concluded)
-------
115
SOU ! H
(a) 1974 to 1980
FIGURE III-9. INCREMENTAL CHANGES IN N02 CONCENTRATIONS PREDICTED BY COM. NO,
concentrations in yg/m3. i
-------
116
ORTH
I i I : 1
SOUTH
(b) 1980 to 1985
FIGURE III-9. (Continued)
-------
117
NOR
SOUTH
(c) 1985 to 1990
FIGURE III-9. (Continued)
-------
118
SOUTH
(d) 1990 to 2000
FIGURE III-9. (Concluded)
-------
119
OU i H
(a) 1974 to 1980
FIGURE 111-10. INCREMENTAL CHANGES IN PARTICULATE CONCENTRATIONS
PREDICTED BY COM. Participate concentrations
in yg/m
-------
120
JORTH
SOUTH
(b) 1980 to 1985
FIGURE 111-10. (Continued)
-------
121
NOR
n
10
20
~! i 1 1 i 1 T
-5
on
\&Y
J
SOU I H
(c) 1985 to 1990
FIGURE 111-10. (Continued)
-------
122
NORTH
SOU I H
(d) 1990 to 2000
FIGURE 111-10. (Concluded)
-------
123
2000, the largest increases are expected around the intersection of U.S.
285 and U.S. 85, and in the area south of Stapleton Airport. An area of
increase is also predicted along State Route 391 in the vicinity of the
Federal Center. Decreases are predicted in small areas to the north of
Broomfield and in Douglas County Urban.
There is no consistent pattern to the incremental changes in
N02 concentrations predicted by the COM, although the increases between
1974 and 1980 correspond to the major highways to a large degree. The
decrease predicted for downtown Denver between 1985 and 1990 might be an
effect due to the changing character of the automobile population, which
by 1990 should be meeting much more stringent NOX emission controls.
Figure 111-10 shows the incremental changes predicted by the COM
for particulate concentrations. The largest increases between 1974 and
1980 are expected in the area bounded by the freeways 1-25, 1-70, and 1-270
Between 1980 and 1985 the maximum changes are predicted west of downtown,
and some improvement is forecast along U.S. 285 south of Lowry AFB.
Jeffco Urban has the largest increases in 1985 to 1990, while the area
west of downtown improves slightly. Between 1990 and 2000, predicted
concentrations increase in Lakewood and in the area of Lowry AFB, whereas
decreases are forecast to the east and west of Lakewood.
The incremental changes predicted by COM for particulates are
not consistent from year to year. The increase in the urban portion of
Jefferson County could be ascribed to the increase in construction
activities required to accommodate the expected increase in population,
but reference to Figure III-6 shows that this activity should be occurring
in all years. It is hard to understand why particulate concentrations
should decrease during any time period, since no general controls on
particulates are expected. The reductions could occur because of the
phasing out of some activity that generates large amounts of particulates,
but more probably are an artifact of differences between the emissions
inventories used.
-------
124
D. CONCLUSIONS
The results of the base case studies of future air quality in the
Denver region in relation to national and state air quality standards are
summarized in Table 111-3.
-------
TABLE 111-3. SUMMARY OF BASE CASE STUDIES AND AIR QUALITY STANDARDS
Pollutant
°3
CO
N02
Particulates
Averaging
Period
1 hour
1 hour
8 hour
1 year
1 year
Air Quality Standard
Federal Colorado
0.08 ppm
35 ppm
9 ppm
82 pg/m3
75pg/m3 45(jg/m3
Observed Maximum
(in year)
0.18 ppm (1976)
51 ppm (1975)
24 ppm (1975)
79 wg/m3 (1974)
124ug/m3 (1975)
Model
Denver Model
Denver Model
Denver Model
Denver Model*
COM
Rollback
COM
Predicted
1976
0.24 ppm
26 ppm
15 ppm
791Jg/m3
79yg/m3
79,,g/m3
168ng/m3+
Concentration
1985
0.13 ppm
17 ppm
10 ppm
75ug/m3
97ug/m3
105pg/m3
217ug/m3
2000
0.09 ppm
4 ppm
3 ppm
74ug/m3
94|iq/m3
114,,g/m3
229ug/m3
*Predicted by ratio method--see text.
tFor 1974 rather than 1976.
-------
127
IV SENSITIVITY ANALYSIS
This chapter describes a study of the sensitivity of the air quality
predictions presented in Chapter II to changes in input variables. Sensi-
tivity analysis is a tool with which the response of a mathematical model
to changes in input parameters and variables can be explored. Variables
to which air quality predictions are sensitive fall into three main cate-
gories: those concerned with meteorological factors, those concerned
with the air quality data used for input to the model, and those which
affect the rate at which pollutants or pollutant precursors are emitted.
We studied the effect of changes in variables in the first and third
categories. We determined that calibrated COM results depend directly on
measured air quality, but DAQM ozone predictions for Denver are relatively
insensitive to initial and boundary conditions determined from data (see
Appendix A).
A. SENSITIVITY OF PREDICTIONS OF THE DENVER MODEL
TO METEOROLOGICAL VARIABLES
The meteorological data input to the Denver Model, although defined
over every grid square, are extrapolated from fairly sparse data. There
are 24 wind stations taking hourly readings (although some data are miss-
ing). Each observation represents the air movement in the immediate
vicinity of the monitor, whereas the model treatment assumes the observa-
tions are averages over an extensive area. The mixing depths used are
inferred from the two temperature soundings per day and surface tempera-
ture measurements taken at Stapleton Airport. It is then assumed that
these mixing depths are typical of the entire modeling region.
From the above description, it is evident that the meteorological
data are by no means known with certainty. In addition, there is a form
of uncertainty associated with the use of a particular day's meteorology
-------
128
to run the model. While this gives an element of realism in that on the
days chosen ozone concentrations were high, the question arises as to
what might happen if the meteorological conditions are worse on some
occasion for which a prediction is required than they were on the days
chosen? How high might the ozone concentration be?
In order to obtain a measure of the sensitivity of the predictions
of the Denver Model to changes in meteorology, we ran three additional
cases:
> Case 1--3 August 1976 meteorology with all wind speeds
reduced by one-third.
> Case 2--3 August 1976 meteorology with the mixing depth
reduced by one-third.
> Case 3--3 August 1976 meteorology with both wind speeds
and mixing depth reduced by one-third.
These three cases were compared with the 3 August 1976 base case to eluci-
date the dependence of the predictions of the Denver Model on wind speed
and mixing depth, and to determine if their effects are compounded. The
one-third reductions represent plausible variations within the uncertainty
of a worst-case determination. In other words, the base case conditions
represent our best estimate of the most adverse days of 1976, but we cannot
exclude the possibility that one-third lower wind speeds or mixing depths
might occur, with attendant higher ozone concentrations.
Details of the predictions of the Denver Model for 3 August 1976, one
of the base case days, are presented in Chapter III. On that day, pre-
dicted ozone concentrations rose steadily to a peak one-hour-average con-
centration of 24 pphm, observed between 1 and 2 p.m. in the (42,32) grid
cell. The approximate trajectory of the ozone cloud, as shown by the
locations of successive one-hour-average ozone concentration maxima, is
presented in Figure IV-la. It may be seen that the concentration maxima
-------
129
NORTH
13 pphm
i I J
SOUTH
(a) Meteorology of 3 August 1976
FIGURE IV-1. LOCATIONS AND VALUES OF PREDICTED MAXIMUM ONE-HOUR-
AVERAGE OZONE CONCENTRATIONS FOR EACH HOUR
FROM 8 a.m. TO 6 p.m.
-------
130
NORTH
SOUTH
(b) Meteorology of 3 August 1976 with all wind speeds
reduced by one-third
FIGURE IV-1. (Concluded)
-------
131
are in the center of the region early in the day, are in the south and
register the highest values in the early afternoon, and are in the south-
west in the late afternoon.
Figure IV-lb shows a similar trajectory for Case 1, in which the
wind speeds were reduced by one-third. The locations of the maximum
concentrations of the entire day for the two cases are only about 3 miles
apart, although the distances between hourly trajectory increments are
much shorter for Case 1. The highest concentration observed for Case 1
was 25 pphm, and this concentration was observed for three consecutive
hours.
The trajectory for Case 2 was similar to that of the base case, and
that for Case 3 was similar to Case 1.
The maximum ozone concentrations predicted for each one-hour interval
throughout the day are presented in Figure IV-2. The differences between
different cases are small until about 11 a.m. Reducing the mixing depth
had no effect on the time at which the peak predicted concentration was
reached, but reducing the wind speed in both Cases 1 and 3 retarded the
attainment of the peak by one hour. This effect is possibly ascribable
to the slower movement of the air parcel under the influence of lighter
winds, so that it remains longer in an area of high pollutant emissions.
Thus a greater load of pollutants enters the parcel, with a concomitant
increase in reaction time.
The maximum predicted ozone concentrations are higher for each case
of more severe meteorological conditions. Recall that the base case
resulted in a maximum concentration of 24 pphm. Reducing wind speed by
one-third (Case 1) increased the predicted maximum to 25 pphm, an increase
of 4 percent. Reduction of the mixing depth (Case 2) had a larger effect,
increasing the maximum to 28 pphm, a 16 percent increase. When both con-
ditions were combined (Case 3), the maximum predicted concentration
-------
132
35
30
25
20
15
O Base case
A Case 1--2/3 wind speed
<) Case 2--2.3 mixing depth
D Case 3--2/3 wind speed and 2/3 mixing depth
Time of Day, by Hourly Interval Istart hour|
I stop hour i
FIGURE IV-2.
MAXIMUM ONE-HOUR-AVERAGE OZONE CONCENTRATIONS
PREDICTED BY THE DENVER MODEL FOR VARIOUS CASES.
3 August 1976 meteorology assumed.
-------
133
increased to 32 pphm, a 33 percent increase. Case 3 has considerably
more effect than the sum of Cases 1 and 2, indicating some synergism.
One might expect the effects to be multiplicative, since regional con-
centration is inversely proportional to wind speed and mixing depth
according to simple box model analyses (e.g., Hanna, 1971). In the
Denver Model results, the effect of Case 1 is less than 10 percent of
that suggested by a box model analysis, and the effect of Case 2 is one-
third of that suggested by a box model analysis. The combined effect
(Case 3) is approximately one-fourth of the box model prediction of a
100[(3/2)2 - 1] = 125 percent increase in ozone concentration.
Another important measure of air quality in the region is the degree
to which the NAAQS is violated. Figure IV-3 shows the number of square
miles over which the NAAQS (8 pphm one-hour-average ozone concentration)
is violated according to the predictions of the Denver Model for the base
case and Cases 1, 2, and 3. Also shown are the numbers of square miles in
which 10 pphm, 20 pphm, and 30 pphm one-hour-average ozone concentrations
are exceeded. In all cases, the NAAQS is exceeded during the 9 to 10 a.m.
period and this exceedance persists throughout the day. However, the
maximum area of exceedance is about 380 square miles in the base case,
but nearly 490 square miles in Case 3. Comparison of Figures IV-3a and
IV-3c shows that a thinner mixing layer does not result in substantially
greater exceedance areas. However, the areas having concentrations
greater than 10 and 20 pphm are somewhat greater. Reducing the wind
speed (Case 1) resulted in substantially greater areas of exceedance than
the base case for all concentration levels. Again, as for the maximum
concentrations reached, there is a synergism between Case 1 and Case 2
that results in their combined effect (Case 3) being greater than simply
additive, but less than multiplicative.
B. SENSITIVITY OF PREDICTIONS OF THE DENVER MODEL AND COM TO THE
SPATIAL DISTRIBUTION OF EMISSIONS
The base case studies for future years used projected emissions
inventories (described in Appendix B), which were supplied by various
-------
134
450 i-
Ozone Concentration: Q 8 pphm
Ll 12 pphm
A 20 pphm
Time of Day, by Hourly Interval ', start hour !
!stop hour
(a) Base Case
FIGURE IV-3.
AREA OF EXCEEDANCE OF GIVEN OZONE CONCENTRATION
FROM PREDICTIONS OF DENVER MODEL, FOR VARIOUS
CASES. 3 August 1976 meteorology assumed.
-------
135
450 -
400 -
350 _
= 300 l_
250 -
150 -
100 -
50 -
Ozone Concentration: Q 8 ppnm
D 10 pphm
0 20 pphm
Time of Day, by Hourly Interval
I start hour
I stop hour
(b) Case l--Wind speeds reduced by one-third
FIGURE IV-3, (Continued)
-------
136
450
400
350
o
5 300
=> 200
150
100
50
Ozone Concentration: Q 8 pphm
fj 10 pphm
Q 20 pphm
Ml I
9 10
10 11
/i I I i \
11 12 1 2
TI T 2 3
Time of Oav. bv Hourlv
I 1 1
3 4 5
4 T 6
TntPrval istart hour I
stop hour
(c) Case 2--Mixing depth reduced by one-third
FIGURE IV-3. (Continued)
-------
137
500
450
400
350
300
250
200
150
100
50
Ozone Concentration: Q 8 pphm
D 12 pphm
<} 20 pphm
A 30 pphm
9 10 11
TO TT T2
12
1
Time of Day, by Hourly Interval
4 5
5 6
start hour
stop hour
(d) Case 3--Wind speeds and mixing depth
reduced by one-third
FIGURE IV-3. (Concluded)
-------
138
public agencies. These inventories as supplied embody certain assumptions
regarding future events. In particular, assumptions about the distribu-
tion of population growth through the year 2000 are included. In this
section we describe our effort to ascertain the sensitivity of the air
quality predictions to population growth patterns. This is not a study
of alternative growth projections, only of whether growth patterns may
be at issue, in terms of air quality projections.
Since we did not have access to the computer programs necessary to
exactly reproduce the effects of different population distributions on
emissions, we proceeded by making uniform percentage changes to all emis-
sions within a study area, and comparing the results with a base case.
The base case chosen for study was the year 2000 with emissions
inventories supplied by the Colorado Division of Highways and the Colorado
Department of Health. These inventories are based on the following
scenario:
> JRPP Highway Plan.
> Rail rapid transit system in operation (73 miles in 2000).
> Improved bus system (925 extra buses).
> Federal Motor Vehicle Control Program, as embodied in AP-42
(Supplement 5), installed and fully effective.
> DRCOG-approved population growth and allocation.
Certain areas were considered by the DRCOG to be subject to relatively
high growth rates (see Appendix C, Figure C-22). From these we designated
six suburban areas for study. Table IV-1 lists projected population
increases for the areas chosen for study and Figure IV-4 shows their loca-
tions within the Denver region. These areas were chosen because they are dis-
tributed around the region and contain fast-growing communities. Assess-
ment of the effect of changing emissions rates in these areas around the Den-
ver metropolitan region will enable the elucidation of any special meteorolog-
ical-geographical effects on air quality. We also changed emissions rates
-------
139
within the Denver city limits to estimate the potential effect of zero
growth in activities there.
TABLE IV-1. POPULATION INCREASES PROJECTED FOR
AREAS AROUND DENVER
No.
1
2
3
4
5
6
7
Area
Broomf ield/Westmi nster/Arvadc
Northgl enn/Thornton
Aurora
South Metro
Jeffco Urban
Lakewood
City of Denver
Direction
from Denver
NW
N
E
S
SW
w
--
Projected
Population
Increase
(1975-2000)
92,000
45,000
107,000
57,200
116,700
90,600
114,400
To simulate a redistribution of population in different parts of the
Denver region, we made computer runs in which the emissions rates in each
of the first six areas described above were reduced by 25 percent, with
emissions in all other areas increased in proportion to their emissions
rates to maintain the same total emissions in the region. This procedure
approximates reducing the population in an area by 25 percent while main-
taining the same total population growth by redistributing it to other
communities in proportion to their populations. An additional model run
was made in which the emissions within the Denver city limits were reduced
by 17.5 percent (the expected 1975-2000 population increase is only 17.5
percent) with redistribution as above.
We thus have the following eight runs for the sensitivity analysis:
-------
140
1 . Broomfield/Westminster/Arvada
2. Northglenn/Thornton
3. Aurora
4. South Metropolitan
5. Jefferson County Urban
6. Lakewood
7. City of Denver
FIGURE IV-4. SUBREGIONS SELECTED FOR SENSITIVITY ANALYSIS
-------
141
1. Base case: year 2000 emissions (28 July 1976
meteorology used in calculating ozone concentrations)
2. Emissions in Broomfield/Westminster/Arvada reduced
25 percent.
3. Emissions in Northglenn/Thornton reduced 25 percent.
4. Emissions in Aurora reduced 25 percent.
5. Emissions in South Metro reduced 25 percent.
6. Emissions in Jefferson County Urban reduced 25 percent.
7. Emissions in Lakewood reduced 25 percent.
8. Emissions in Denver reduced 17.5 percent.
We used the Denver Model to evaluate ozone concentrations and the COM to
evaluate N02 and particulate concentrations. Reservations as to the appro-
priateness of COM for N02 studies expressed in previous chapters were not
judged to be important here, since only the effects of changes to the base
case are of interest.
1. Denver Model
The meteorology chosen for the sensitivity study was that of 28 July
1976. For that date and 1976 emissions, the Denver Model predicted a
region-wide maximum one-hour-average ozone concentration of 20 pphm between
3 and 4 p.m. For the emissions inventory for the year 2000 and the same
meteorology, the Denver Model predicted a maximum one-hour-average of 10
pphm between 2 and 3 p.m. For the seven sensitivity runs listed above, no
difference was found in either the location, time, or magnitude of this
maximum. In fact, the differences between the eight runs in terms of pre-
dicted ozone concentrations were confined at all times of day to at most
a difference of 1 pphm in one or two grid squares.
Concentration differences should depend on the location of the emis-
sions reduction area with respect to the region-wide emissions patterns
for any given mean wind direction. The maximum changes were too small,
-------
142
however, to identify any such effect. In other words, all seven pertur-
bations of the base case emissions inventory resulted in essentially
identical region-wide predictions for ozone concentrations. This result
is ascribed to the time factor in ozone production: by the time that
significant amounts of ozone have been formed, the emissions are too well
mixed to reflect their origins. Furthermore, the region-wide emissions
were not changed in this exercise and in spite of the rather drastic
imposition of growth control that the emissions changes would imply, no
more than 7 percent of the region-wide emissions were redistributed.
Thus, it is apparent that these changes in the spatial distribution of
emissions have no effect on ozone concentrations. From this one may
infer that land use controls that would reduce an area's population by
as much as 25 percent without changing the regional population would be
ineffective in terms of reducing ozone concentrations.
The sensitivity of one-hour-average NO^ concentrations predicted by
the Denver Model to the above emissions changes was not extensively anal-
yzed. We noted, however, that almost nndetectable changes were produced
by the emissions pattern changes. CO concentrations would be affected in
each area of reduced emissions but street-scale effects, which produce CO
peaks, are not treated in the Denver Model.
2. Climatological Dispersion Model (COM)
COM was used to evaluate annual average NOp and particulate concen-
trations for the eight cases listed above. As in the case of ozone con-
centrations, only slight differences in the calculated concentrations were
noted in the immediate vicinity of each large perturbation in emissions.
Table IV-2 lists the average concentrations and maximum concentrations pre-
dicted by COM within each area for the base case and also for the case in
which the total emissions in that area were reduced.
-------
143
TABLE IV-2. EFFECTS OF CHANGES IN EMISSIONS ON PREDICTED AVERAGE
AND MAXIMUM CONCENTRATIONS OF N0? AND PARTICULATES
IN VARIOUS AREAS
Area
Broomfield/Westminster/
Arvada
Northglenn/Thornton
Aurora
South Metro
Jeffco Urban
Lakewood
Denver
Concentration
NO,
Particulate
Base Case
Reduced
Base Case
Avg
32.6
Max
59.1
Avg
29.8
Max
56.5
Avg
118
Max
161
Reduced
Avg
112
Max
160
44.
22.
24.
23.
38.
60.
3
1
1
6
4
0
70.
53.
40.
31.
56.
94.
5
9
1
2
3
1
41.
19.
21.
20.
34.
54.
1
7
6
9
0
2
69.
50.
35.
28.
52.
83.
0
9
0
1
9
5
131
101
102
126
140
164
190
123
135
159
173
229
124
95
97
114
130
153
187
116
125
139
159
206
Table IV-2 shows that a 25 percent reduction in emissions does not
result in a 25 percent reduction in concentrations, but rather a 5-10
percent reduction only in the immediate area of the reduction in emis-
sions. In order to ascertain whether the lowering of predicted concen-
trations was any more than a local effect, we calculated differences
between the base case and each of the sensitivity runs in turn. These
differences are presented as isopleths in Figures IV-5 and IV-6. These
figures illustrate that the effect of large changes in emissions inven-
tories remains localized rather than spreading region-wide. Most of the
slight spreading that does occur is northward, which is reasonable in
light of the annual w"ind rose calculated from the COM data shown in
Figure IV-7. This figure shows that almost 20 percent of the year the
wind is from the south, and over 30 percent of the year it is from the
south to southwest.
-------
144
NORTH
10 20
—h--r~i—i—r—i--r-T—T-T—|-
_ _ i. _L i L-J.L... J
u.
LU
(a) Emissions Reduced by 25 Percent in
Broomfield/Westminster/Arvada
FIGURE IV-5.
ISOPLETHS (IN yg/mJ) OF DIFFERENCES BETWEEN ANNUAL AVERAGE NO
CONCENTRATIONS PREDICTED BY COM FOR AN EMISSIONS REDUCTION
IN A GIVEN AREA AND BASE CASE
-------
145
T~lI" •-[-—[—-r --J--— -,
C'O
t—
CO
LJ
L_J_. J I J 1_J LJ
..J. L
10
20
OUTH
I
rr,
a:
LU
(b) Emissions Reduced by 25 Percent in
Northglenn/Thornton
FIGURE IV-5. (Continued)
-------
146
b'UL I H
(c) Emissions Reduced by 25 Percent in
Aurora
FIGURE IV-5. (Continued)
-------
147
(d) Emissions Reduced by 25 Percent in
Littleton/Arapahoe County Urban
FIGURE IV-5. (Continued)
-------
148
ORTH
1- - J_. L_ J I.__J -L L L_. J.._ji_ J_ _L._. L_. i._
T"
i i i
-------
149
i i i L i j i i_..±i L
SOUTH
(f) Emissions Reduced by 25 Percent in
Lakewood
FIGURE IV-5. (Continued)
-------
150
! i .
8 0 U I H
(g) Emissions Reduced by 17.5 Percent in
City of Denver
FIGURE IV-5. (Concluded)
-------
151
K I t~ r-i -r | |
i l|j u n I n
SOU IN
(a) Emissions Reduced by 25 Percent in
Broomfi eld/Westmi nster/Arvada
FIGURE IV-6.
ISOPLETHS (IN yg/rri ) OF DIFFERENCES BETWEEN ANNUAL AVERAGE
PARTICULATE CONCENTRATIONS PREDICTED BY COM FOR AN
EMISSIONS REDUCTION IN A GIVEN AREA AND BASE CASE
-------
CO
152
NORTH
j 0 20
OJ
CO
UJ
L_ ..L_._l 1 ...1. _J ___!.__ J___L! j I. _.. I. _.!.....!. ...J L ..L...1-.L.
SOU I H
(b) Emissions Reduced by 25 Percent in
Northglenn/Thornton
O'l
AJ
h-
O"1
iT
LL\
FIGURE IV-6. (Continued)
-------
153
G).
fij
C)?'
LiJ
'^j_
y
""~T~
10
-T-H-T--
("1 C' "T I I
Jh i n
_1 !—!—r
20
L_.l 1...L -L _1...L.._ L _0._ .! .L_.iL._L J_ L
10
L_J...L
I I H
(c) Emissions Reduced by 25 Percent in
Aurora
FIGURE IV-6. (Continued)
-------
154
10
n
, i r- ,
i r
fj
U
\
UJ
(d) Emissions Reduced by 25 Percent in
Littleton/Arapahoe County Urban
HbUkb 1V-6. (Continued)
-------
155
_.. 1 . I... 1 -I- - [ ...J I
(e) Emissions Reduced by 25 Percent in
Jefferson County Urban
FIGURE IV-6. (Continued)
-------
156
Him i n
T—r
f--
i
—i
i
L . i
(f) Emissions Reduced by 25 Percent in
Lakewood
-IGURE IV-6. (Continued)
-------
157
NORTH
CO
&_.
cu
. -',•>
J.I
•5).
(g) Emissions Reduced by 17.5 Percent in
City of Denver
FIGURE IV-6. (Concluded)
-------
158
Source: 1971-1974 STAR data.
FIGURE IV-7. ANNUAL WIND ROSE FOR THE DENVER METROPOLITAN REGION
C. CONCLUSIONS
The main conclusion from the sensitivity study is that region-wide
control of pollutant concentrations is not achievable by large-scale
redistribution of emissions. For ozone, which is formed by reactions of
emissions, it was found that perturbations in emissions rates of 25 percent
in different communities had no effect on region-wide concentrations. For
species subject to annual average standards, only local reductions in con-
centrations could be observed, and the fractional changes in concentrations
were less than the fractional changes in emissions. Thus it would not be
fruitful to require emissions reductions in small areas in order to relieve
regional problems. It would seem that region-wide effectiveness will only
be achieved through region-wide controls. Furthermore, local changes in
emissions would have to be proportionally greater than the desired changes
in local pollutant concentrations, since local concentrations depend to
some extent on regional emissions.
-------
159
V MITIGATION CONSIDERATIONS
The base case results described in Chapter III show substantial
improvement in Denver's air quality between the present and the year
2000. Grid cell averages of peak ozone and CO concentrations are pre-
dicted by the Denver Model to decrease to the respective NAAQS or less.
"Hot spot" violations of the NAAQS for CO are predicted to decrease in
frequency, and NO^ is expected to stay within the NAAQS. (Predicted
annual-average 1^ concentrations have large uncertainties, as discussed
in Chapter III.) All of these predictions depend on certain assumptions,
which may or may not prove to be valid. The most significant assumptions
are that:
> Growth projections for the Denver region are accurate.
> Vehicle emissions will meet the standards of the Federal
Vehicle Emissions Control Program (FVECP).
In this chapter we discuss the effects of these assumptions on predicted
pollutant concentrations. We also briefly discuss the effects of some
mitigation measures.
A. GROWTH PROJECTIONS
Errors in growth projections can be either positive or negative,
although growth beyond projected levels would presumably require waste-
water treatment facilities in excess of those presently proposed. On
the other hand, a decision to not fund the proposed facilities would
probably limit growth to less than projected levels. It seems likely,
therefore, that Denver's growth will not greatly exceed projections,
and considering this factor alone, future air quality is conservatively
predicted by the base cases.
-------
160
B. VEHICLE EMISSION FACTORS
To study the effects of assuming that the emissions factors speci-
fied in the Federal Vehicle Emissions Control Program (FVECP) would be
met, particularly at Denver's altitude, we carried out simulations with
the Denver Model using other emission factors. The sets of vehicle
emission factors studied with the Denver Model are as follows:
> Set l--the AP-42 emission factors. The EPA publication
AP-42, Supplement 5, contains emission factors for an
average mix of vehicles for future years. Thus AP-42
emission factors are based on the assumptions that (1) the
emissions standards in the Federal Vehicle Emission Control
Program (FVECP) will be met, and (2) the mix of vehicles
of various ages, which are affected by different emission
standards, will be as specified in that document.
> Set 2--emissions factors selected by the staff of EPA
Region VIII to represent a vehicle flppf subject to
emission regulations not as severe as those implied by
AP-42. The degree of emission control with Set 2 emis-
sion factors was estimated to be in excess of that pro-
vided by presently proposed Colorado vehicle inspection
and maintenance programs.
> Set 3—the emission factors of Set 1 with a higher Federal
NO emission standard.
X
> Set 4--the emission factors of Set 2 with a higher Federal
NO emission standard.
The distribution of ozone concentrations with Set 2 emission factors we1
similar to the results obtained with Set 1 factors that graphical pre-
sentations are not given here. (The results of simulations with Set 1
of emission factors are given in Chapter III.)
The significance of some changed emission factors on peak ozone
concentrations may be seen in Figure V-l . In this figure peak ozone
-------
161
EMISSIONS INVENTORY
FOR YEAR -1976
0.6
0.5
c
o c
— o
C VI
0)
si 0-3
c_> o
X S-
O <*3
Z 01
= "O
• 01
ra ±j
en i.0.2
0.1
PREDICTED MAXIMUM
ONE-HOUR-AVERAGE
OZONE CONCENTRA-
TION (ppm)
(0.24)
SET 1 EMISSION FACTORS
SET 2 EMISSION FACTORS
SET 4 EMISSION FACTORS
ISOPLETHS OF MAXIMUM OZONE CONCENTRA-
TIONS COMPUTED FROM SIMULATIONS OF
SMOG CHAMBER RUNS WITH THE CARBON-
BOND MECHANISM
POINTS COMPUTED WITH CARBON-BOND
MECHANISM IN QAQM SIMULATION RUNS
FOR 3 AUGUST 1976 METEOROLOGY
MAXIMUM OZONE CONCENTRATION
(ppm)
0.05 0.10 0.15 0.20 0.25
2000
(0.07)
I
0.5 1.0 1.5 2.0 2.5
5-9 a.m. Reactive Hydrocarbon Concentration (ppmC)
(computed near CAMP station by DAQM)
FIGURE V-1. MAXIMUM OZONE COMPUTED BY CARBON-BOND SIMULATIONS
OF SMOG CHAMBER EXPERIMENTS AND BY DENVER MODEL
FOR VARIOUS ASSUMED SETS OF FUTURE EMISSIONS
-------
162
concentrations for 3 August 1976 are plotted as a function of precursor
concentrations. Points are plotted for Set 1, Set 2, and Set 4 of
emission factors. Set 2 leads to predicted ozone peaks about 2 pphm
higher than Set 1. This increase is sufficient to infer clear, but
small, NAAQS ozone violations in the year 2000. Set 2 also leads to
more severe violations in 1985 than Set 1.
Set 3 of emission factors was found to lead to ozone concentrations
that are substantially equivalent to those computed with Set 4 of emis-
sion factors. In either case, no violations of the NAAQS for ozone are
predicted to occur in 1985. It is very important to note, however, that
these conclusions apply only to ozone concentrations within the modeling
region. Ozone concentrations outside the modeling region and NO con-
A
centrations within or without the region might well be adversely affected
by relaxing the Federal standard for vehicular NO emissions.
X
Set 2 of emission factors also affected peak grid-cell-average CO
concentrations. The winter simulations showed peak one-hour-average
CO concentrations of 27 ppm in 1985 and 6 ppm in 2000. These figures
are both below the NAAQS for CO of 35 ppm. The calculated eight-hour-
average concentrations for Set 2 were 16 ppm in 1985 and 5 ppm in 2000.
Thus violation of the NAAQS of 9 ppm CO is predicted for 1985 but not
for 2000.
C. ADDITIONAL MITIGATION STUDIES
Given a population distribution, emissions may be affected by many
mitigation measures, but of the measures proposed (see Appendix C), the
most significant two (in addition to strict FVECP compliance, discussed
above) were judged to be
> Expansion of mass transit facilities
> Hydrocarbon vapor recovery at gasoline delivery nozzles.
-------
163
1. Mass Transit
Estimates of the effect of a proposed expansion of the Regional
Transportation District (RTD) bus fleet by approximately 1000 buses were
made. These estimates included consideration of anticipated ridership
and use of private automobiles between home and bus stop. The latter
consideration is especially important since by the year 2000 little
improvement is anticipated in cold start emissions, but warm trip emis-
sions are expected to decrease. The result of the study of RTD expan-
sion was that a 1 percent effect on emissions would be most likely and
that the effect is unlikely to exceed 2-1/2 percent. The reductions
should be widely distributed and thus a smaller fractional change is
expected in peak ozone concentrations for reasons discussed below.
2. Hydrocarbon Vapor Recovery
The effect of a Stage II vapor recovery system was studied. We
estimate that the effect of such a system would reduce Denver's peak
ozone concentrations by no more than 1 pphm. The analysis follows.
The emissions inventory supplied to SAI by the Air Pollution Control
Division of the Colorado Public Control Division of the Colorado Public
Health Department showed about 5 percent of the hydrocarbons emitted
into the regional atmosphere result from filling fuel tanks at gasoline
stations. Mr. David Joseph of EPA Region VIII expressed reservations
about this amount; he believes that gas station emissions might be as
much as 50 percent higher.
The California Air Resources Board (Venturini and Grandy, 1975)
estimates that uncontrolled gasoline stations emit about 0.0024 pounds
of hydrocarbons per supplied vehicle mile traveled. Of this about 40
percent may be controlled by submerged fill pipes, so about 0.0014
*
Stage II vapor recovery provides for double hose or equivalent systems
for recovery of hydrocarbon vapors at filling nozzles in gasoline
-------
164
(#HC/VMT supplied) would be ascribable to the vehicle tank filling
operation. In 1973 in California these emissions represented about 11
percent of all stationary source HC emissions and about 5 percent of all
HC emissions. These figures would confirm the reasonableness of 5 to 7
percent for Denver.
Several reports indicate the feasibility of 90 percent recovery of
vapor emitted during vehicle tank filling operations (Venturini and
Grandy, 1975; Radian Corporation, 1975; Scott Research Laboratories,
1974). This would suggest a potential reduction in present Denver total
HC emissions of 5 to 6 percent. If anticipated reductions in all other
categories of HC emissions are realized, the vapor recovery program would
represent a 6 to 8 percent reduction of the 1985 emissions inventory and
a 9 to 11 percent reduction of the year 2000 emissions inventory.
The photochemical simulation of a 30 percent reduction in all
emissions reported later in this chapter indicates a fractional reduc-
tion in peak ozone concentrations of roughly one-half the fractional
reduction in emissions. This would suggest that a reduction in ozone of
about 3 percent in 1976, 4 percent in 1985 and 5 percent in 2000 could
be achieved by Stage II vapor recovery. These numbers are clearly small
compared to the accuracy of analyses, observations, or emissions fore-
casts. However, Stage II vapor recovery would not affect NO emissions.
A
An estimate of ozone reduction at a constant NO level may be obtained
/\
from the ozone isopleth chart shown in Figure V-2. Here it is seen that
the ozone gradient is about (0.24-0.17)7(0.24-0.20) = 0.07/0.04 = 1.75
times as great for a reduction from 1976 HC at constant NO as it is for
/\
a reduction in both NO and HC by the same percentage. Thus a decrease
X
of about 1 pphm in peak ozone concentrations might be expected from imme-
diate institution of a Stage II program.
Venturini and Grandy (1975) estimate the cost-effectiveness of a
Stage II system to be about 15 to 50 cents per pound of HC controlled.
This is compared with the cost-effectiveness of other proposed control
techniques in Table V-l, taken from their paper.
-------
165
0.6
EMISSIONS FACTORS •* Set 1
EMISSIONS INVENTORY * 1985
MAXIMUM OZONE CON-
CENTRATION (ppm) ->• (0.09)
•ISOPLETHS OF MAXIMUM
ONE-HOUR-AVERAGE OZONE
CONCENTRATIONS FROM
SIMULATIONS OF SMOG
CHAMBER RUNS WITH THE
CARBON-BOND MECHANISM
MAXIMUM ONE-HOUR-AVERAGE
OZONE CONCENTRATIONS
PREDICTED BY THE DENVER
MODEL FOR VARIOUS EMIS-
SIONS INVENTORIES, EMIS-
SIONS FACTORS, AND 3
AUGUST 1976 METEOROLOGY
0.5
01
E S-
Q. CU
CL >
„ _ ^ ^
O)
c: Q
o
to
i- C.
•(-> O
c: -i-
(D +->
(J ro
o
E fO
• O)
CTi "O
I CL>
0
O)
CL
MAXIMUM ONE-HOUR-AVERAGE
OZONE CONCENTRATIONS—^0.05
(ppm)
0.10
0.15
0.20
0.25
0.4
0.3
0.2
0.1
A OZONE = 0.04 ppm
0.5
.0
.5
2.0
2.5
6-9 a.m. Reactive Hydrocarbon Concentration (ppmC)
(predicted near CAMP station by Denver Model)
FIGURE V-2. OZONE ISOPLETH DIAGRAM
-------
166
TABLE V-l. THE COST-EFFECTIVENESS OF VARIOUS METHODS
FOR REDUCING HYDROCARBON EMISSIONS
Method
Degreasing Controls
Dry Cleaning Controls
Painting Substitutions
Jet Engine Modification
Jet Aircraft Towing
Piston Aircraft Engine Mod.
LDMV* Minor Exhaust Retrofit
LDMV Catalytic Converter
LDMV Evaporative Controls
LDMV State Inspection/Maintenance
LDMV Maintenance to Minimum Pollution Cap
Modest Bus Improvement
Modest Mileage Surcharge
Heavy Mileage Surcharge
Major Bus Improvements
"LDMV = Light-Duty Motor Vehicles
Source: Venturini and Grandy (1975).
Dollars per Pound
Reduction in HC Emissions
Range for
Stage II
vapor recovery
-------
167
The conclusion is that hydrocarbon control approximates an optimally
directed strategy for improving Denver's ozone air quality. The double
hose vapor recovery system is a cost-effective technique for hydrocarbon
control, but has limited potential since only a small fraction of the
region's HC emissions can be controlled. The maximum effect on Denver's
peak ozone concentrations should be about 1 pphm.
In the early stages of this project we uncovered two errors in the
emissions inventories that had been used for some early runs of the
Denver Model. Although these runs give no useful information on air
quality in the future, the nature of the errors is such that they enable
us to ascertain the effects on air quality of certain types of mitigation
measures.
D. IMPLICATIONS FOR MITIGATION STRATEGIES OF ADDITIONAL SIMULATIONS
1. Effects of 30 Percent Reductions in All Emissions
The first error was the failure to include corrections to the emis-
sions inventory to allow for Denver's atmospheric temperatures and pres-
sure. This error resulted in an across-the-board reduction in all
emissions of 30 percent. Simulation runs made with this inventory thus
represent a scenario in which the imposition of emissions controls
results in such an emissions reduction.
Figure V-3 shows the maximum one-hour-average ozone concentrations
predicted over the entire region for two days in 1976 for both the cor-
rected emissions inventory and the one with a 30 percent error. Note
that for 28 July 1976 the reduction in the peak ozone concentration is
15 percent, from 20 pphm to 17 pphm. For 3 August 1976 the reduction
is 12.5 percent, from 24 to 21 pphm. On both days the buildup of ozone
with time is exactly the same for the two emissions inventories. Thus
one may conclude that a 30 percent reduction in emissions results in a
15 percent reduction in the peak ozone concentration. This illustrates
the nonlinearity of the ozone formation process, in that a given reduc-
tion in emissions results in only about one-half that reduction in the
peak ozone concentration. Figure V-4 shows the total area in the region
-------
BASE CASE
30 PERCENT REDUCTION
IN ALL EMISSIONS
3 AUGUST 1976 METEOROLOGY
25
CO
o
M
o
20
15
10
28 JULY 1976 METEOROLOGY
9
To
10
ii
ii
12
11
'1
Time of Day by Hourly Interval i-?^ hmjr
hour J
FIGURE V-3. THE EFFECT ON PREDICTED MAXIMUM OZONE CONCENTRATIONS OF A
30 PERCENT REDUCTION IN ALL EMISSIONS
-------
169
500
450
400
350
300
250
200
150
100
50
BASE CASE
30 PERCENT REDUCTION
IN ALL EMISSIONS
JL 10. 11 11
10 11 12 1
3
Time of Day by Hourly Interval start hour
[stop hour
(a) 28 July 1976 Meteorology
FIGURE V-4. AREA HAVING OZONE CONCENTRATIONS IN
VIOLATION OF THE NAAQS
-------
170
50 U
O BASE CASE
Q 30 PERCENT REDUCTION
IN ALL EMISSIONS
Time of Day by Hourly Interval
(b) 3 August 1976 Meteorology
FIGURE V-4. (Concluded)
-------
171
for which the average ozone concentration exceeds the NAAQS for ozone
(8 pphm) as a function of time of day. For this measure, the 30 percent
reduction in emissions results in about a 10 percent reduction in area
of exceedance.
2. Effects of Relaxation of the Federal NO Emissions Standard
A
The second error made in the emissions inventory was the use of an
incorrect NO emissions factor for trucks in 1985. This resulted in the
X
NO emissions from traffic being too high. For example, at 20 mph the
A
emissions factor used corresponded to automobile NO emissions of
X
0.9 g/mile instead of 0.4 g/mile. We calculated that this error resulted
in a spurious increase of 65 percent in the area-wide total NO emissions
X
The results of a model run made with this inventory may be used to obtain
an estimate of the effect on ozone concentrations of a slowdown in the
Federal timetable for reducing NO emissions from automobiles, while at
A
the same time staying on schedule for hydrocarbon emissions reductions.
Such a slowdown has been proposed because of problems with meeting the
current Federally mandated guidelines.
Figure V-5 shows the effect of this increase in NO emissions on
A
predicted peak ozone concentrations in the area in 1985. The figure
shows that exceedances of the NAAQS for ozone would be eliminated by
this change. The effect on peak concentrations for the 3 August 1976
meteorology is more pronounced than for the 28 July 1976 meteorology;
the reductions are from 13 pphm to 5 pphm (a drop of 61 percent) and
from 12 pphm to 8 pphm (a drop of 33 percent), respectively. A possible
explanation for this difference is that the predicted peak ozone con-
centrations using the 28 July 1976 meteorology may be lower than they
should be because of loss of material from the modeling region. Figure
V-6 shows isopleths of predicted ozone concentrations using 28 July 1976
meteorology and an emissions inventory for 1985 that was not corrected
for the error in NOV emissions factors. It appears that between 10 a.m.
A
and noon a considerable portion of the ozone cloud is outside the model-
ing region. Starting between noon and 1 p.m., the ozone cloud moves
northward back into the modeling region. However, the Denver Model has
-------
>,
c
ft
10
BASE CASE
HIGH N0x EMISSIONS
3 AUGUST 1976 METEOROLOGY
1985 EMISSIONS INVENTORY
28 JULY 1976 METEOROLOGY
1985 EMISSIONS INVENTORY
__
10
10
TV
1J
12
12
1
Time of Day by Hourly Interval
FIGURE V-5. THE EFFECT ON PREDICTED MAXIMUM OZONE CONCENTRATIONS OF AN
INCREASE IN THE N0x EMISSION STANDARD
-------
173
NORTH
30
-..= = --.i_" =a-*>-••-—" j^/S-rfc-'iJ.r^sfriJirieU—i* ««L.- -;rj
•--"?? -f2?. :-*ff^*^^^'t iMk »-f»ljLir -Tt i-.^ • -=•»=" 'i _ J *"--'-' • -T.--2Z T_"??"^
SOUTH
(a] 1000-1100 MST
FIGURE V-6. ISOPLETHS OF OZONE CONCENTRATIONS PREDICTED BY
THE DENVER MODEL FOR 28 JULY 1976 METEOROLOGY
AND 1985 EMISSIONS INVENTORY. Ozone isopleths
in units of pphm.
-------
174
NORTH
SOUTH
(b) 1100-1200 MST
FIGURE V-6. (Continued)
-------
175
NORTH
; i
60'
SOUTH
(c) 1200-1300 MST
FIGURE V-6. (Continued)
-------
176
NORTH
r\
SOUTH
(d) 1200-1300 MST
FIGURE V-6. (Concluded)
-------
177
no means to permit pollutants to reenter the modeling region. Thus the
predicted ozone concentrations are artificially reduced. This effect is
a possible explanation for the dip in the predicted peak ozone concentra-
tions seen in Figure V-5 starting at noon.
Two conclusions may be drawn from the runs described here:
> A 30 percent reduction in all emissions resulted in only
a 15 percent drop in the predicted peak ozone concentration.
> A relaxation of the NO emissions standard from 0.4 g/mile
A
to 0.9 g/mile is predicted to result in up to a 60 percent
reduction in the peak ozone concentration and to eliminate
violations of the NAAQS for ozone for the Denver region for
1985 and 2000.
These two conclusions have important ramifications in terms of
mitigation measures. The first shows that an across-the-board reduction
of emissions by a given percentage may reduce the peak ozone concentra-
tion by only one-half that percentage. The second seems to promise
great beneficial effects from a relaxation of NO emission standards.
A
If the NO emissions standards are relaxed, however, the urban plume
X
from Denver will contain more NO . Any downwind area releasing hydro-
A
carbons into this plume will experience larger ozone concentrations
because of reaction with the extra N0? from Denver. The predicted
decrease in ozone concentrations within the Denver metropolitan region
would, of course, be accompanied by increases in NO concentrations.
Although there is no Federal air quality standard for NO, it is not
clear that increased NO concentrations are desirable. NO oxidizes to
NOp, which is a regulated pollutant. Increased NO emissions do not
necessarily lead to increased NOp concentrations in the Denver region
for reasons similar to those explaining the ozone reduction. Increased
downwind N02 concentrations could, however, be expected. Thus relaxation
of the NO emission standard is not a panacea; if it is considered it
A
must be with full account of all downwind effects.
-------
179
VI EXPOSURE AND DOSAGE STUDIES
As indicated in Chapter III, simulations with the SAI Denver Model
show that a substantial decrease in ambient ozone concentrations may be
expected over the next 25 years. From the base case studies, peak ozone
concentrations in the Denver metropolitan region are expected to drop
from 24 pphm in 1976 to 13 pphm in 1985 and 9 pphm in 2000. The areal
extent of violations of the NAAQS is projected to decrease from about
350 sq. mi. in 1976 to about 150 sq. mi. in 1985 and to less than 40
sq. mi. in 2000. To complete the analysis of the impact of improved air
quality, we estimated the dosages and exposures that will be experienced
by the human population within the Denver region. This chapter indicates
briefly how ozone exposures and dosages were calculated, presents the
results of the calculations, and discusses their significance.
It is important to note that the exposure and dosage calculations
reported here depend on specific concentration distributions. The results
therefore are only representative, and are not predictions in the sense
that peak concentrations are. Other days with other wind patterns might
produce equivalent peak ozone concentrations and even equivalent areas of
exceedance, but if those peaks happened to occur at different locations,
exposures and dosages might well be quite different.
A. DOSAGE CALCULATIONS
In the calculations performed here, the cumulative dosage of ozone
is defined simply as the product of the population in a given area, the
ozone concentration to which that population is exposed, and the length
of time over which the exposure to that concentration occurs. The
dosage provides a measure of the total amount of ozone present in the
total volume of air that is inhaled by people over the time period of
concern. This may be illustrated as follows. Let the dosage, D, be in
units of pphm-person-hour. If the volume of air inhaled is V cubic
-------
180
meters per person-hour, the quantity of ozone (Q) present in the volume
V may be estimated as
Q - DV (1CT8 cubic meters)
If V is assumed to be a constant, then Q is proportional to D and the
dosage D provides a measure of Q. It may be noted that the dosage provides
no information as to the amount of ozone inhaled per person or the ambient
ozone concentrations or time duration of the dosage. Mathematically, the
dosage may be expressed as:
'
D(x,y) = / P(x,y,t)C(x,y,t)dt
where D is the dosage of ozone from time t-, to time t£, P is the population,
and C is the concentration of ozone.
The spatial variables x and y are used to indicate that D, P, and C
vary with location. The population P is assumed to be a function of time (t)
to account for the temporary redistribution of population that occurs during
the daytime working hours.
B. EXPOSURE CALCULATIONS
The exposure to concentrations above a certain level n is defined here
as the product of the number of people exposed and the time duration over
which the concentration n is exceeded. Exposure calculations provide a
measure of the number of people-hours in which the population is exposed
to concentrations higher than a specified level. By using values of n
equal to or greater than the NAAQS, it is possible to estimate the number
of people exposed to excessively high ozone concentrations. Exposure cal-
culations do not account for the impact of lower concentrations of ozone,
and dosage calculations give us no indication of the concentration to
which the individuals were exposed. Hence, exposure and dosage estimates
complement each other: Exposures provide information on short
-------
181
high concentrations and dosages indicate the impact of low or moderate
pollutant concentrations of relatively longer duration.
Mathematically, exposure may be defined by
E(x,y,n) = / P(x,y,t)6[C(x,y,t) - nj dt
\
where E(x,y,n) is the exposure corresponding to concentration n and the
delta function 6 equals 1 if the concentration at (x,y,t) is greater than
or equal to n and 0 otherwise.
C. INPUTS NECESSARY FOR EXPOSURE AND DOSAGE CALCULATIONS
From the definitions of exposure and dosage, it is clear that ozone
concentrations and population distributions must be estimated before ozone
exposures and dosages can be calculated. In this section we discuss how
ozone concentrations and population distributions were estimated and con-
sider certain characteristics of these distributions.
1. Ozone Concentrations
The ozone concentrations used here were obtained from simulations with
the SAI Denver Model for emissions in the years 1976, 1985, and 2000, using
the meteorological conditions for 28 July 1976 and 3 August 1976. The pre-
dicted ozone concentrations are in the form of one-hour-averages over two
mile square grid cells. (The results of these simulation runs are described
in detail in Chapter III.) Since the conclusions of this study are only
representative and since the results for the 28 July and 3 August meteoro-
logical conditions are qualitatively similar, results are presented here for
3 August only.
-------
182
Figure VI-1 displays the cumulative ozone concentrations in pphm-hours
for emissions in the years 1976, 1985, and 2000 and the 3 August 1976 meteo-
rology. This figure shows integrated concentrations over a period of several
hours, not hour-by-hour concentrations. It indicates the regions that
experience high exposure to ozone during the day but not the exact location
or time of occurrence of the peak one-hour-average ozone concentrations.
Relatively high cumulative ozone concentrations are predicted to occur in
the districts south of central Denver. The greatest cumulative concentra-
tions are expected in the Sheridan-Englewood-Cherry Hills area for 1976
emissions, but shift to an area southeast of Columbine Valley for 1985 emis-
sions and to an area south of Greenwood Village for 2000 emissions. Note
that although the peak one-hour-average ozone concentrations predicted by
the Denver Model for the base cases (Chapter III) drop sharply from 1976 to
1985 and again from 1985 to 2000, predicted reductions in the cumulative
ozone concentrations are much more modest.
2. Population Distributions
Several methods have been used to estimate the population in the Denver
metropolitan region; these are discussed in Appendix C. For the purposes
of estimating dosage and exposure, populations as specified in the Sub-area
Allocation Plan [SAP] (DRCOG, 1976) were used. The modeling region was
first divided into 27 districts as indicated in Figure VI-2. Estimates of
the population in each district in the years 1975, 1985, and 2000 were
obtained from the SAP, and the population was assumed to be uniformly dis-
tributed within each district. These population data are summarized in
Table VI-1. The population density in several suburbs will apparently
approach that of central Denver in the year 2000, presumably due to the
higher rate of growth in these areas. This is clearly brought out in
Figure VI-3, which shows the increases in population from 1975 to 1985 and
from 1985 to 2000. The southwestern suburbs of Lakewood and Jefferson
County-Urban and the northern suburbs of Broomfield and Thornton are anti-
cipated to be relatively high growth areas.
-------
1 2 3 4 S 6 7 O 9 10 11 12 10 14 18 16 17 IB 19 2O 21 22 23 24 23 26 27 28 29 3»
3O 40 4O 48 48 411 48 51 51 SU 53 37 57 53 65 52 52 54 04 53 B'J 46 46 37 37 33 33 32 32 29 29
29 48 48 4B 48 48 48 31 31 53 53 57 57 55 53 52 52 54 54 S3 53 46 46 37 37 33 33 32 32 29 29
2O 46 46 46 46 49 49 60 50 32 52 5H SO 58 5U 54 54 54 54 52 52 46 46 46 46 34 34 31 31 30 30
27 46 46 46 46 49 49 SO 30 32 32 50 50 58 58 54 54 54 34 62 62 46 46 4(1 40 34 34 31 31 30 30
BROOMF1ELD
26 45 45 50 50 31 51 54 64 02 52 57 57 57 57 52 62 52 52 49 49 43 43 41 41 36 36 32 32 29 29
NOHTUCLENN
23 45 45 00 50 51 51 54 54 52 02 37 57 37 67 52 52 52 52 49 49 43 43 41 41 36 36 32 32 29 29
24 48 48 B4 34 56 56 50 58 57 37 56 56 55 SS 31 51 50 50 45 45 43 43 38 30 36 36 36 36 29 29
WESTMINdTER THORNTON
23 48 48 54 54 06 36 38 58 57 37 56 36 33 03 51 51 50 00 45 43 43 43 30 3O 36 36 36 36 29 29
FDRL
22 34 84 58 58 61 61 66 66 68 6B 59 69 53 53 4O 4O 52 62 45 45 47 47 37 37 35 35 34 34 32 32
HCTS
21 54 34 50 SO 61 61 66 66 63 65 59 59 53 53 40 40 52 52 45 45 47 47 37 37 35 35 34 34 32 32
HCKY MNTN AHSNI.
20 53 S3 SO SO 63 65 69 69 70 70 67 67 37 37 50 60 53 53 BO 50 44 44 8B 30 37 37 34 34 32 32
ARVADA COMMERCE
19 53 53 50 50 65 65 69 69 70 70 67 67 37 57 50 30 53 33 60 30 44 44 30 30 37 37 34 34 32 32
18 31 51 SO 58 67 67 73 73 73 73 72 72 67 67 60 60 67 67 60 60 42 42 39 39 3O 30 36 36 33 33
17 51 51 SO 58 67 67 73 73 73 73 72 72 67 67 60 60 57 57 50 50 42 42 39 39 38 30 36 36 33 33
WI1T. RDC STAPL INTL
16 53 53 61 61 71 71 78 70 01 III 04 04 75 75 73 73 69 69 67 57 45 45 41 41 40 40 37 37 35 35
COLDEN E1ICEWATER
13 53 53 61 61 71 71 78 78 Ol 81 04 84 73 73 73 73 69 69 67 37 43 43 41 41 40 40 37 37 35 35
DENVER AURORA
14 34 34 63 63 71 71 Ol Ol O6 86 89 09 02 02 O3 O3 76 76 71 71 60 60 50 50 44 44 39 39 36 36
13 54 34 63 63 71 71 81 01 06 06 09 09 82 02 03 83 76 76 71 71 60 60 60 50 44 44 39 39 36 36 _,
12 51 51 62 62 73 73 05 OS 95 95 101 101 98 9O 94 94 07 07 02 82 73 73 59 59 52 52 43 43 39 39 u>
CLENDALE
II 51 31 62 62 73 73 83 03 93 95 101 101 9O 90 94 94 07 07 02 02 73 73 59 59 52 52 43 43 39 39
LAKEWOOB
10 49 49 58 58 72 72 80 00 97 97 105 105 109 109 109 109 96 96 O7 87 79 79 65 65 52 62 40 48 42 42
9 49 49 38 SB 72 72 00 00 97 97 103 105 109 109 109 109 96 96 87 07 79 79 65 65 52 62 40 40 42 42
MORRISON SHERIDAN ENCLEWOOD
0 46 46 55 35 68 6O 05 O5 95 95 105 105 114 114 115 115 106 106 III) U8 78 70 64 64 58 50 53 53 42 42
CHEIUIY 11 ILLS
7 46 46 55 63 60 68 OS iJ5 95 95 103 105 114 114 115 115 106 106 O8 88 78 70 64 64 68 50 53 53 42 42
6 3O 38 46 46 SB SO 76 76 OO OO 97 97 107 107 106 106 104 104 95 95 70 70 60 60 62 62 61 61 47 47
JEFF CO LITTLETON GREENWOOD VLC
3 3O 3O 46 46 38 08 76 76 OO 00 97 97 107 107 106 106 104 104 95 95 70 78 63 65 62 62 61 61 47 47
URBAN CLttUN VLV
4 41 41 42 42 48 40 60 60 O0 OO 97 97 106 106 99 99 100 100 94 94 75 78 5(1 50 57 57 63 63 46 46
3 41 41 42 42 40 40 fcO CO OO 00 97 97 106 1O6 99 99 100 100 94 94 75 75 50 50 57 07 63 63 46 46
I ,
2 3O 3O 43 43 49 49' 5O SO 70 7O 77 77 07 O7 77 77 01 Ol 78 70 62 62 51 5'l 52 52 49 49 4O 40
I 3O 30 43 43 49 49 5O 50 70 70 77 77 07 07 77 77 01 01 70 70 62 62 51 51 52 52 49 49 40 40
(a) Year 1976 Emissions
FIGURE VI-1. CUMULATIVE OZONE CONCENTRATIONS BETWEEN 500 AND 1300 HOURS PREDICTED BY THE
DENVER MODEL. Concentrations are given in pphm-hours within one-square-mile
grid cells for emissions in given year and 3 August 1976 meteorology; grid
numbers listed on the left side and top of figure.
-------
1 2 3 4 8 6 7 a 9 10 11 12 13 14 16 16 17 IB 19 28 21 22 23 24 23 26 27 2B 29 38
i i '
30 44 44 43 43 41 41 42 42 43 43 46 46 44 44 44 44 49 49 49 49 44 44 36 36 32 32 32 32 29 29
29 44 44 43 43 41 41 42 42 43 43 46 46 44 44 44 44 49 49 49 49 44 44 36 36 32 32 32 32 29 29
28 43 43 41 41 42 42 41 41 42 42 46 46 47 47 45 45 49 49 4O 40 43 43 3O 38 33 33 36 38 29 29
27 43 43 41 41 42 42 41 41 42 42 46 46 47 47 46 45 49 49 40 4O 43 43 3B 38 33 33 38 38 29 29
BROOMFIELD
26 42 42 44 44 43 43 43 43 40 40 44 44 43 43 40 40 45 45 45 45 4O 4O 39 39 35 35 32 32 29 29
NORTHGLENN
25 42 42 44 44 43 43 43 43 48 40 44 44 43 43 40 40 45 45 45 45 40 40 39 39 36 35 32 32 29 29
24 43 43 46 46 43 49 49 43 42 42 39 39 48 40 37 37 41 41 39 39 40 48 36 36 33 33 33 35 29 29
WESTMINSTER THORNTON
23 43 43 46 46 43 45 43 43 42 42 39 39 40 40 37 37 41 41 39 39 48 48 36 36 39 35 35 35 29 29
FDRL
22 45 45 46 46 46 46 47 47 43 43 38 30 35 35 33 33 39 39 37 37 42 42 34 34 33 33 33 33 32 32
HCTS
21 43 45 46 46 46 46 47 47 43 43 30 38 35 35 33 33 39 39 37 37 42 42 34 34 33 33 33 33 32 32
RCKY mm ARSNL
20 42 42 43 43 43 45 45 48 42 42 39 39 33 33 29 29 35 35 30 38 30 30 35 33 35 35 33 33 32 32
ARVADA COMMERCE
19 42 42 43 43 45 45 45 45 42 42 39 39 33 33 29 29 35 35 38 30 30 30 35 35 35 35 33 33 32 32
10 30 30 40 40 44 44 42 42 48 40 30 3O 36 36 32 32 33 33 31 31 33 33 33 33 35 35 33 35 33 33
17 38 38 40 40 44 44 42 42 40 40 38 38 36 36 32 32 33 33 31 31 33 33 33 33 35 35 35 35 33 33
WHT. RDG STAPL INTL
16 39 39 41 41 44 44 45 45 44 44 43 43 37 37 36 36 38 30 34 34 35 35 35 35 37 37 36 36 34 34
GOLDEN EDCEWATER
13 39 39 41 41 44 44 45 45 44 44 43 43 37 37 36 36 30 30 34 34 33 35 35 35 37 37 36 36 34 34
DENVER AURORA
14 41 41 42 42 43 43 46 46 44 44 43 43 38 30 41 41 41 41 44 44 44 44 39 39 39 39 37 37 35 35 _i
CO
13 41 41 42 42 43 43 46 46 44 44 43 43 3O 38 41 41 41 41 44 44 44 44 39 39 39 39 37 37 35 35 -Pa
12 41 41 42 42 46 46 60 50 51 31 51 51 40 48 46 46 47 47 51 51 61 61 45 45 44 44 40 40 3O 3U
CLENDALE
II 41 41 42 42 46 46 58 30 31 31 51 31 40 40 46 46 47 47 61 51 51 51 43 43 44 44 40 40 38 38
LAKEWOOD
18 42 42 48 45 48 40 53 53 53 83 55 55 56 56 87 57 64 54 55 56 55 65 49 49 44 44 43 43 41 41
9 42 42 45 43 48 48 33 53 53 53 55 55 56 66 57 67 64 54 85 55 55 55 49 49 44 44 43 43 41 41
MORRISON SHERIDAN ENCLEWOOD
8 43 43 46 46 49 49 54 84 53 S3 59 59 62 62 66 66 64 64 56 56 55 55 51 81 49 49 47 47 39 39
CHERRY HILLS
7 43 43 46 46 49 49 54 54 55 35 59 59 62 62 66 66 64 64 56 56 55 S3 51 51 49 49 47 47 39 39
6 37 37 41 41 45 43 51 51 34 84 57 97 61 61 69 68 66 66 60 68 94 84 38 88 32 52 83 33 42 42
JEFF CO LITTLETON GREENWOOD VLG
3 37 37 41 41 43 43 51 51 34 54 37 57 61 61 65 66 66 66 60 60 64 54 50 SO 52 82 53 53 42 42
UHBAR CLMBR VLV
4 41 41 48 40 41 41 44 44 53 63 62 62 67 67 64 64 65 65 62 62 63 53 45 46 48 48 53 83 41 41
3 41 41 40 48 41 41 44 44 53 53 62 62 67 67 64 64 65 65 62 62 53 53 45 45 48 48 S3 53 41 41
2 38 38 42 42 44 44 48 40 53 S3 56 86 61 61 54 54 56 56 53 63 44 44 40 40 43 43 42 42 36 36
1 38 30 42 42 44 44 48 48 53 53 56 36 61 61 64 64 56 86 S3 53 44 44 48 40 43 43 42 42 36 36
(b) Year 1985 Emissions
FIGURE VI-1 (Continued)
-------
1 2 3 4 9 6 7 0 9 10 11 12 13 14 IB 16 17 1O 19 20 21 22 23 24 29 26 27 28 29 30
' j
36 44 44 42 42 41 41 43 43 44 44 47 47 46 46 45 45 49 49 4O 48 43 43 36 36 33 33 33 33 29 29
29 44 44 42 42 41 41 43 43 44 44 47 47 46 46 45 45 49 49 40 40 43 43 36 36 33 33 33 33 29 29
20 42 42 4O 4O 41 41 U9 39 40 40 44 44 44 44 42 42 47 47 47 47 42 42 30 3O 33 33 31 31 3O 3O
27 42 42 40 4O 41 41 39 39 40 40 44 44 44 44 42 42 47 47 47 47 42 42 3B 30 33 33 31 31 30 3O
BROOMF1ELD
26 40 40 42 42 40 40 39 39 37 37 39 39 39 39 37 37 42 42 43 43 39 39 3D 3O 35 30 33 32 29 29
NORT11CLENN
25 40 40 42 42 40 40 39 39 37 37 39 39 39 39 37 37 42 42 43 43 39 39 30 30 35 35 32 32 29 29
24 40 40 42 42 40 40 40 40 37 37 34 34 3S 35 34 34 3O 3O 36 36 37 37 34 34 35 35 36 36 29 29
WESTMINSTER THORNTON
23 40 40 42 42 40 4O 40 4O 37 37 34 34 35 35 34 34 30 30 36 36 37 37 34 34 30 35 36 36 29 29
FDRL
22 41 41 41 41 40 40 40 40 37 37 32 32 30 30 29 29 36 36 34 34 39 39 32 32 32 32 34 34 32 32
HCTS
21 41 41 41 41 40 40 40 40 37 37 32 32 30 30 29 29 36 36 34 34 39 39 32 32 32 32 34 34 32 32
IICKY MNTH ARSNL
20 39 39 30 38 30 30 37 37 36 36 33 33 29 29 25 25 31 31 35 35 35 35 33 33 34 34 33 33 32 32
AIWADA COMMERCE
19 39 39 3O 30 38 30 37 37 36 36 33 33 29 29 25 25 31 31 35 35 35 35 33 33 34 34 33 33 32 32
10 34 34 34 34 35 35 33 33 32 32 32 32 30 3O 27 27 29 29 27 27 31 31 31 31 34 34 34 34 33 33
17 34 34 04 34 35 35 3U 33 32 32 32 32 30 30 27 27 29 29 27 27 31 31 31 31 34 34 34 04 33 33
W11T. RDC STAPL INTL
16 35 35 34 34 33 33 34 34 35 33 34 34 2'i 29 29 29 31 31 20 20 31 31 32 32 33 35 35 35 34 34
GOLDEN EDCEKATEIl
15 35 35 34 34 35 35 34 34 39 35 34 34 29 29 29 29 31 31 20 20 31 31 32 32 35 30 30 33 34 34
DENVER AURORA _i
14 37 37 33 33 34 34 33 33 34 34 33 33 2O 2O 31 31 31 31 33 33 30 38 36 36 36 36 36 36 35 35 CO
cn
13 37 37 35 33 34 34 33 33 34 34 33 33 2O 2O 31 31 31 31 33 33 3O 38 36 36 36 36 36 36 35 35
12 30 38 37 37 37 37 36 36 39 39 30 30 35 35 33 33 35 35 39 39 42 42 39 39 41 41 39 39 37 37
CLENDALE
11 30 3O 37 37 37 37 36 36 39 39 30 30 35 33 33 33 35 35 39 39 42 42 39 39 41 41 39 39 37 37
LAXEWOOD
10 39 39 39 39 40 40 40 40 41 41 40 40 3O 30 39 39 39 39 42 42 45 43 43 43 40 4O 41 41 39 39
9 39 39 39 39 40 40 40 40 41 41 40 40 3U 30 39 39 39 39 42 42 45 43 43 43 40 40 41 41 39 39
I1ORRIBON SHERIDAN ENCLF.WOOD
O 41 41 41 41 40 40 42 42 43 43 43 43 41 41 47 47 47 47 43 43 46 46 43 45 43 40 44 44 36 36
CHERRY HILLS
7 41 41 41 41 40 40 42 42 43 43 43 43 41 41 47 47 47 47 43 43 46 46 45 45 45 45 44 44 36 36
6 36 36 3O 30 39 39 41 41 42 42 42 42 42 42 40 40 51 51 47 47 45 45 44 44 46 46 47 47 38 3O
JEFF CO LITTLETON GREENWOOD VLC
0 36 36 38 30 39 39 41 41 42 42 42 42 42 42 4O 4O 51 01 47 47 45 43 44 44 46 46 47 47 3D SO
URBAN CLMBN VLY
4 39 39 38 30 36 36 36 36 42 42 47 47 47 47 49 49 52 52 50 50 44 44 40 40 42 42 48 40 38 30
3 39 39 38 38 36 36 36 36 42 42 47 47 47 47 49 49 52 52 60 50 44 44 40 40 42 42 40 40 3O 30
2 38 38 40 40 41 41 42 42 42 42 43 43 47 47 44 44 46 46 46 46 39 39 36 36 39 39 38 38 33 33
I 3O 30 40 40 41 41 42 42 42 42 43 43 47 47 44 44 46 46 46 46 39 39 36 36 39 39 30 38 33 33
(c) Year 2000 Emissions
FIGURE VI-1 (Concluded)
-------
30
29
28
27
26
23
24
23
22
21
20
19
10
17
16
15
14
13
12
11
10
9
O
7
6
5
4
3
2
1 2 3
1 1 1
1 1 1
1 1 1
1 1 1
5 B 5
555
555
555
555
553
553
355
555
355
21 21 | 22
GOLTOtf
21 21 21
21 21 21
5 5 [jZ_l_
838
555
355
535
HORR
555
555
535
553
555
833
853
353
DISTRICTS USED FOR SPECIFICATION
•* 3 6 7 8 9 10 ll 12
1
1
1
5
5
5
5
3
3
3
3
5
3
22
22
^IJ.
5
3
8
20
20
5
3
5
3
3
5
5
3
1
1
1
5
3
5
3
5
3
3
24
22
22
22
22
3
5
5
20
20 1
8
3
5
5
5
5
8
3
1 1
1 1
8
3.
3
0
5
24
24
24
22
22
22
22
20
20
8
5
20
5
5
5
5
5
3
3
3
3
5
5
1 1
1 1
1 1
,
6
,1.
1 | 2
6 6
66666
BROpMF 1 P-l n
6 6
7
7 7
WESTMIRS
7
7 7
7
24 24 24 24
24 24 24 24
ARVADA
24 24 24 24
24 [" 23 ( 24
23 23 23
WHT. RDC
23 23 23
El
20 20 20
20 20 20
20 20 20
20 20 20
20 20 20
LAKF.WOOI
20 20 20
i~j 20 20
5 |T C
16 16
16 16
PLETON
10 10
10 13
in 10
09 9
TORHTON
999
25 1 10 10
10 10 10
19
2
2
2
2
2
O
9
9
10
10
10 111 11 11
COMMERCE
10 |l1 11 11
26 26 26
26 26 26
26 26 26
26 26 26
K V E R
26 26 26
26 26 26
26 26 r^J
GLEN DAL
26 26 26
26 26 26
lei 26 26
15 15 15
ffiRRY HILLS
15 15 15
111
26
S
26
26
26
26
26
E
26
26
26
26
26
14 14 14 14
CREEWTOOD-VLC
13 13 13 13
13 13 13 13
13 13 13 13
20
2
2
2
2
2
2
10
10
10
10
2
2
26
26
TAPL
26
26
26
26
26
26
26
26
26
26
14
13
13
13
21
2
2
2
2
2
10
10
10
2
2
26 |
22
2
2
2
2
2
1O
23 24
> • f
2 2
2 2
2 2
2 2
2 2
,r| 2
10 | 2 2
222
222
222
MHTW ARSNL
222
222
26 26
26 26
INTT
12
12
12
26
26
26
26
26
14
14
14
13
13
13
26 26
26 26
12 12 12
12 12 12
AURORA
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
3
3
13
13
13
13
3 12
3 ll 2
3 3
3 3
13 13
13 13
ZS 26 27 28
2222
2222
2222
2222
2222
2222
2222
2222
2222
2222
2222
2222
26 2 26
26 26 26
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
12 12 12
3 a] 12
333
333
333
26
26
12
12
12
12
12
3
3
3
3
3
3
3
3
29 30
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
2 2
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
3 3
00
FIGURE VI-2. POPULATION DISTRICTS SPECIFIED IN THE SUB-AREA ALLOCATION PLAN
-------
187
TABLE VI-]. POPULATION BY DISTRICT
(Thousands of Persons)
District
No. 1975 1985 _2000_
1 0.00 0.00 0.00
2 10.50 14 20 14.00
3 .70 .90 1.20
4 0.00 0.00 0.00
5 .20 6.00 12.00
6 16.40 29.70 44.50
7 31.60 40.70 58.90
8 31.50 36.70 38.60
9 27.60 39.50 65.50
10 62.50 71.40 83.10
11 18.60 19.50 20.00
12 121.70 157.00 228.70
13 35.90 39.90 79.30
14 3.50 5.20 10.00
15 5.40 8.30 10.70
16 35.90 40.90 45.90
17 6.00 8.10 13.20
18 34.10 38.60 47.90
19 21.20 68.20 137.90
20 123.00 163.70 213.60
21 14.50 18.60 26.00
22 9.50 11.60 15.20
23 40.60 46.40 53.70
24 80.80 94.80 117.40
25 7.60 9.90 12.90
26 532.30 566.10 646.70
27 3.70 8.70 8.70
Source: DRCOG (1976).
-------
30
29
20
27
26
25
24
23
22
21
20
19
111
17
1 2 3 4 S 6 7 O 9 10 11 12 13 14 15 16 17 IB 19
> j
0600000000022222222
0000000000022222222
000000000 no no no lie 19 19 2 2 2 2
o o e o e e o 110 no 110 no no 110 19 19 2 2 2 2
BROOMFIELD
NOHTHCLEHN
WESTMINSTER THORNTON
FDRL
HGTS
4 4 4 4 4 60 60 60 66 60 60 19 19 19 35 35 35 35 35
4 4 4 4 4 60 60 60 60 60 60 19 19 19 35 35 12 12 12
ARVADA COMMERCE
4 4 4 4 60 60 60 60 60 60 60 60 35 35 35 35 12 12 12
4 4 4 4 17 17 60 40 60 60 29 29 29 29 29 29 29 29 12
4 4 4 4 17 17 48 48 4(1 40 4B 29 29 29 29 29 29 29 29
20
2
2
2
2
2
2
35
35
35
35
]
2
2
29
29
WHT. RDC STAPL
16
13
14
13
12
11
10
9
U
7
6
5
4
3
2
1
41 41 17 17 17 17 4B 40 46 4fl 40 29 29 29 29 29 29 29 29
GOLDEN EDCEWATF.R
41 41 41 17 17 17 96 96 96 96 40 29 29 29 29 29 29 29 29
DENVER
41 41 41 41 17 96 96 96 96 96 96 29 29 29 29 29 29 29 29
4 4 41 4 4 96 96 96 96 96 96 29 29 29 29 29 29 29 29
4 4 4 4 4 4 96 96 96 96 96 29 29 29 29 29 29 499 29
CLENDALE
LAKE WOOD
4 4 4 96 96 96 96 96 96 96 96 29 29 29 29 29 29 29 29
4 4 4 96 96 4 4 96 96 96 29 29 29 62 62 62 29 29 29
MORRISON SHERIDAN ENGLEWOOD
CHEIUIY HILLS
JEFF CO LITTLETON GREENWOOD VLC
URBAN CLMDN VLY
29
29
21
2
2
2
2
2
33
35
35
35
2
RCKY
2
29
29
29
22 23
2
2
2
2
2
2
2
2
2
2
33 35
35
a
2
2
MNTH
2
2
29
29
0
2
2
2
Mil
2
2
29
29
24
2
2
2
2
2
2
2
2
2
2
3NL
2
2
29
29
25
2
2
2
2
2
2
2
2
2
2
2
2
29
29
26
2
2
2
2
2
2
2
2
2
2
2
2
2
29
27
2
2
2
2
2
2
2
2
2
2
2
2
29
29
2B 29 3O
2
2
2
2
2
2
2
2
2
2
2
2
29
29
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
INTL
34
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
2
2
2
2
AURORA
29
29
29
29
29
29
29
29
21
14
14
14
0
0
54
29
29
29
29
29
21
21
21
14
14
14
0
0
94
64
54
54
54
54
0
0
14
14
14
14
0
0
54
54
54
54
54
54
0
e
0
0
14
14
0
0
54
54
54
54
54
54
54
64
0
0
14
14
e
0
54
54
54
54
54
54
54
54
0
0
0
e
0
O
54
54
54
54
54
54
54
54
0
0
0
0
0
0
54
54
04
54
34
64
54
54
54
0
0
0
e
e
54
54
64
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
e
e
0
e
0
e
e
e
e
0
0
0 co
00
0
0
0
0
0
0
0
e
0
0
0
0
FIGURE VI-3.
(a) Increase from 1975 to 1985
INCREASE OF POPULATION (IN TENS OF PERSONS) IN ONE-SQUARE-MILE GRID CELLS IN
THE DENVER METROPOLITAN AREA. Grid numbers are listed on the left side and
top of figure.
-------
30
29
28
27
26
23
24
23
22
21
20
19
18
17
1 2 3 4 0 6 7 O 9 10 11 12 13 14 10 16 17 IB 19 20
e 0 0 0 0 ' 0 0 0 0 e 0 0 0 « 0 « 0 0 0 e
0000000000O0000O00O0
0000000OO 123 123 123 123 39 39 0 O 0 0 0
0000000 123 123 123 123 123 123 39 39 0 0 0 0 0
BROOMFIELD
WESTMINSTER THORNTON
FD11L
IICTS
AllVADA COtDIEHCE
4 4 4 4 9(1 98 98 98 98 98 98 98 46 46 46 46 7 7 7 0
4 4 4 4 30 30 98 60 98 98 69 69 69 69 69 69 69 69 7 69
4 4 4 4 30 30 60 60 60 60 60 69 69 69 69 69 69 69 69 69
21
0
0
0
0
0
46
46
46
46
0
RCKY
0
69
69
69
22
0
0
0
0
0
46
46
0
0
0
23
0
0
0
0
0
46
0
0
0
0
24
0
e
0
0
0
0
0
0
e
0
23
0
0
0
0
0
0
0
0
0
0
26
0
0
0
0
0
0
0
0
0
0
27
e
0
0
0
0
0
0
0
0
0
28
0
0
0
0
0
0
0
0
0
0
29
«
0
0
0
0
0
0
0
e
0
30
0
O
e
0
0
0
o
0
0
0
MNTH AHSNL
0
0
69
69
0
0
69
69
0
0
69
69
0
0
69
69
0
0
0
69
0
0
69
69
0
0
69
69
0
0
0
0
0
0
0
0
WIT. HI)C 8TAPL INTL
16
13
14
13
12
1 1
10
9
8
7
6
0
4
3
2
I
73 73 30 30 30 30 60 60 60 60 60 69 69 69 69 69 69 69 69 69
GOLDEN EUGKWATKIl
73 73 73 30 30 30 118 118 110 118 60 69 69 69 69 69 69 69 69 69
DENVER
73 73 73 73 30 118 118 118 118 118 118 69 69 69 69 69 69 69 69 69
4 4 73 4 4 1 18 1 18 1 18 1 18 1 18 1 18 69 69 69 69 69 69 69 69 69
GI.ENUA1.E
4 4 4 4 4 4 1 IB 1 18 1 18 1 18 I 18 69 69 69 69 69 69 69 69 69
LAKEKOOD
4 4 4 MB 1 18 1 18 1 IB 1 18 1 18 I 18 I 18 69 69 69 69 69 69 69 69 69
4 4 4 11(1 118 4 4 118 118 118 69 69 69 62 62 62 69 69 69 69
MOimlSON 8UE1IIDAN ENCI.EWOOO
CHERRY IIII.I.S
444444 248 248 248 241) 69 69 169 62 62 40 40 40 69 69
JEFF CO LITI'LETON GREENWOOD VI.G
UKI1AN Cl.rillN VLY
444444 248 2411 248 248 248 248 93 93 93 145 145 143 145 145
110
1 10
110
1 10
1 10
1 10
1 10
110
110
1 10
1 10
1 10
I 10
110
1 10
110
0
0
0
0
AURORA
1 10
69
69
69
69
69
60
60
60
143
143
145
0
0
1 10
1 10
I 10
1 10
1 10
1 10
0
0
143
143
143
145
0
0
110
1 10
1 10
1 10
1 10
1 10
0
0
0
0
143
143
0
0
110
1 10
1 10
110
1 10
1 10
1 10
1 10
0
0
143
140
0
0
110
110
1 10
110
110
1 10
110
110
0
0
0
0
0
0
1 10
1 10
1 10
110
1 10
1 10
1 10
1 10
0
0
0
0
0
0
1 10
1 10
1 10
1 10
1 10
1 10
1 10
1 10
1 10
0
0
0
0
0
I 10
110
1 10
0
0
0
0
0
0
0
0
O
0
0
0
0
0
0
0
e
0
0
0
0
0
0
0
0
0
0
e
0
0
0
0
0
0
0
0
0
0
0
(b) Increase from 1985 to 2000
FIGURE VI-3 (Concluded)
CXI
-------
190
3. Adjustments to Population Pistributions
The populations shown in Table VI-1 reflect the true population dis-
tributions only at times other than normal daytime working hours (0800 to
1500 MST). During normal working hours there is a substantial shift in
the population because employment centers are generally not located near
residential population centers. Since high ozone concentrations occur
during the daylight hours, this temporary dislocation of population must
be accounted for in estimating ozone dosages and exposures.
The adjustments required to estimate the population distribution
during working hours may be carried out as follows. Let the employment
in the i-th district be E- and the residential population P.. The total
population within the entire modeling region is
and the total employment is
Let a be a ratio defined by
A number of people
=E -aP
commute
will commute into the i-th district during working hours to fill the jobs
not filled by the local population. Hence the total population during
working hours is given by
, , , ,
1 w ' ' '
(P,) = P, +
1 w '
Note that the population adjustment may be positive, negative, or zero.
-------
191
In order to estimate the adjusted populations during work hours,
employment data are needed. Employment data are available by "super-
districts" as indicated in Appendix C of this report. The 24-super-
districts used for specification of employment are delineated in Figure
VI-4. Employment in areas lying outside these super-districts is assumed
to be negligible. Employment data are available for the years 1970 and
2000. The data for years 1975 and 1985 were obtained by linear interpo-
lation. The employment by super-district is summarized in Table VI-2.
Changes in employment between 1975 and 1985 and between 1985 and 2000 are
shown in Figure VI-5. Large increases can be seen to occur in the dis-
tricts to the south, the southeast, and the northeast of central Denver
and also in the vicinity of Glendale.
Figure VI-6, showing the difference between the adjusted and residential
population distributions, reveals the prevailing pattern whereby residents
from outlying suburbs commute to central Denver during working hours.
D. EXPOSURE AND DOSAGE ESTIMATES
Dosages and exposures were estimated using the simulated ozone concen-
trations for 3 August 1976 meteorology and the population distributions in
accordance with the Sub-area Allocation Plan (SAP). The population distri-
butions were adjusted as noted above for temporary changes that occur
during working hours.
Table VI-3 presents the estimated dosage and exposure over the Denver
metropolitan region as a function of time of day. These results are also
presented graphically in Figures VI-7 and VI-8. The dosage calculations
show how the rate of increase of ozone dosage tends to be low in the early
morning hours, increases to a maximum at mid-day and again gradually falls
off at night. The cumulative dosage during the day appears to decrease by
roughly 25 percent from 1976 to 1985 and then to increase by about 6.5 per-
cent from 1985 to 2000. Dosages tend to reflect the average regionwide
-------
1 2
30 0 0
29 00
20 00
27 00
26 11 11
25 ill 11
24 111 11
3466
On ft ' o
V \J v
0000
0000
0000
11 II 11 11 1
11 11 11 11 1
< a V iw 11
o o o o o
ooooo
ooooo
00000
BROOMF 1 ELD
1 11 11 11 11
.*
0
0
17
17
16
13
0
O
17
17
16
14
0
0
17
17
16
15
0
0
17
17
16
16
0
0
17
17
16
17 18
0 0
0
17
17 17
17 17
16 17
19
0
0
17
17
17
20
0
0
17
17
17 1
21
0
0
17
17
18
22
0 :
0 j
17
18
18
23 24
18 in
in in
18 in
10 10
10 18
25 26 27 28 29 30
in IB i o o o o
18 in 0 O 0 0
in IB, o o o o
18 IBJ 0 0 0 0
10 1O 0 0 0 0
NOHTHCLENN 1 :
1 11 11 11 11
11 It 11 11 11 11 11 11 II
16
16
16
16
16
16
i WESTTIINSTEIl
23 Oil
22 Oil
1
21 0 0 |
|
20 00
19 00,
IB 0 0
17 11 11
It 11 11
16 11 11
11 11 11 11 1
11 11 11 11 1
11 11 11 11 1
11 11 11 11 1
11 11 11 11 1
11 11 11 11 ]
11 If ff 11' 1
11 11 11 11 '
I 11 11 11 11 | 16
1 11 11 11 11
16
16
16
16
FDUL
16
16
16
16
16
16 17
16 16
17
1C
10
in
18
18
18
18
18 10
10 10
in 10 o o o o
in in o o o o
TIIOHHTON
16
16
16
16
16 16
16 16
19
19
19
19
19
19
19
19
19 19
19 19
19 ! in o o o o
19 20 0000
HGTS
1 12 12 12 12
1 12 12 12 12
AHVADA
2 12 12 12 12
2 12 12 12 12
3 13 13 13 13
WHT. RDG
11 11 11 11 13 13 13 13 13
GOLDEN
15 0 O
14 00
13 00
12 00
11 00
10 00
9 00
M
a 00
7 00
6 00
5 00
4 00
3 00
2 00
1 0 0
16
16
16
2
""
2
16
16
16
2
2
2
16
16
16
2
**
3
EDGEWATER _._.
11 II 11 11.13 13 13 13 13
14 1 1 11 13 13 13 13 13 13
14 14 14 14 14 14 14 14 14
14 14 14 14 14 14 14 14 14
0 0 14 14 14 14 14 14 14
LAKEWOOD
O 0 14 14 14 14 14 14 14
0 0 14 14 1
onnisoK
0 0 15 15 1
4 [ 15 14 9 9
5
5 15 13 1 9 9
0 0 0 0 15 15 15 15 15
00001
5 15 15 15 15
JEFF CO
0 0 0 0 15 15 15 15 15
2
10
10
10
9
,
^
9
ITEni
9
24
24
24
UIU3AN CLNBN
0 0 0 0 15 15 15 15 15 1 24l
0 0 0 0 j 13 15 15 15 15
0000
0000
0 1 15 15 13 0
00000
24
0
0
2
10
10
10
9
9
24
24
24
10
1
16
16
16
3
3
f 1
n i?
10 10
10
10
9
9
9
ENG
23
23
10
10
9
0
"23"
.F.W(]
16
16
16
3
3
3
N V
6
6
8
8
8
~23~
OD
16
16
19
3
3
3
r. p
6
6
8
OLE
0
8
0
19
19
19
19
RCKY
19
jraiE
19
4
4
5
5
5
19
ncE
19
1 O
1 y
19
19
STAPL
4
4
5
5
5
IDALE
7 7
7
7
23 j 22 22 22
CTTEnHY H I LLS
23 i 22 22 22
24 23 23
LITTLETON
23
VLY
23
23
0
0
23
23
23
0
0
23
23
23
0
0
1
23
(
23
23
23
0
0
7
7
7
22 1
22 22 22
HEEHVOOD VLG
22 22
22 22
22 22
0 0
0 0
22
22
22
0
0
4
4
5
5
6
3
7
7
7
7|
_J
22
22
22
22
0
0
19
19
19
19 19
19 20 0 0 0 0
MNTN AFISNL
19
19
19 19
19 19
INTL
20
20
21
6
5
5
5
5
7
22
22
22
22
22
0
0
20
20
— AUI
21
21
21
21
21
21
22
22
22
22
22
22
0
0
20 20
20 20
IOITA
21 21
21 21
21 21
21 21
21 21
21 21
21 21
21 21
22 21
22 21
22 22
22 22
0 0
0 0
19 2O 0 0 0 O
19 20 0 0 0 O
20 20 0 0 0 O
2O 20 0000
20 20 0 0 0 0
2O 20 20 20 0 0
. j
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
21 21 21 21 21 21
0 0 O 0 0 0
000000
FIGURE VI-4. SUPER-DISTRICTS USED FOR SPECIFICATION OF EMPLOYMENT
-------
193
TABLE VI-2. EMPLOYMENT BY SUPER-DISTRICTS
(Thousands of Persons)
Super-
District
No.
1975
Year
1985
2000
0
i
o
3
4
5
6
7
8
9
10
t 1
12
13
14
15
16
17
10
19
20
21
22
23
24
0.00
48. 10
21.60
62.70
28.90
36.00
32.00
21.70
24.40
13.40
54.00
20.80
10.80
26.30
20.60
10.50
37.40
2.00
3.70
12.30
14.80
17.50
16.30
31. 10
9.60
0.00
48.40
31.20
70.20
36.60
42.80
35.50
37.00
34. 10
22.70
56. 10
28.30
15.50
34.50
28.70
16.60
48.60
4.70
4.80
16. 10
28.20
28.60
37.60
39.60
19.40
0.00
48.90
45.60
81.40
48.00
53.00
40.60
60.00
46.50
36.60
59.20
39.60
22.40
40.00
40.90
25.70
63.40
8.70
6.50
21.70
44.50
45.20
69.40
52.40
34.00
Source: DRCOG (1973).
-------
30
29
28
27
26
23
24
23
22
21
20
19
IB
17
1
0
0
0
O
B
a
a
0
0
0
0
0
0
8
2
0
0
0
o
a
a
a
8
a
0
0
0
0
a
3
0
0
0
0
0
a
a
0
u
0
a
8
a
a
4
e
0
0
0
a
a
a
a
a
a
a
a
a
a
6
0
0
0
0
a
a
a
a
a
a
a
a
a
a
6
j 0
0
0
O
a
a
8
a
a
a
a
a
a
a
7
0
0
0
0
a
a
a
a
a
a
a
26
26
39
8 9 10 11 (2 13 14 15 16 17 IB 19 20 21
00000000000000
00000000009000
00009999999999
O0009999999999
DROOMFIELD
B B fl a 21 21 21 21 21 21 9 9 9 2
NORTHCLENN
a a a a 21 21 21 21 21 21 9 9 2 2
a a a a 21 21 21 21 21 21 21 21 22
VESTMl NSTER THORNTON
8 B a 8 21 21 21 21 21 21 21 9 9 9
FDRL
8 B B B 21 21 21 21 21 21 21 9 9 9
HCTS
26 26 26 26 21 21 21 21 21 21 9 9 9 9
RCKY
26 26 26 26 21 21 21 21 21 21 9 9 9 9
ARVADA COMMERCE
26 26 26 26 21 21 21 21 21 9 9 9 9 9
26 26 26 26 96 96 96 62 62 62 4O 9 40 4O
39 39 39 39 96 96 96 62 62 62 40 40 40 40
22 23
0
0
9
2
2
2
2
9
9
9
MNTN
9
9
40
40
2
2
2
2
2
2
2
9
9
9
24
2
2
2
2
2
2
2
9
9
9
28
2
2
2
2
2
2
2
9
9
9
26
2
2
2
2
2
2
2
2
60
60
27
0
0
0
O
0
0
0
0
0
0
28
0
0
O
0
0
0
0
0
0
0
29
0
0
0
O
0
0
0
0
0
0
30
6
0
0
0
0
0
0
o
o
e
AHSNL
9
9
40
40
9
9
40
40
9
9
60
60
60
60
60
60
0
0
0
0
0
0
0
0
0
0
0
O
0
0
0
0
WHT. HOC STAPL INTL
16
B
a
a
a
a
a
39
GOLDEN
15
14
13
12
1 1
10
9
0
0
0
0
0
0
0
0
0
0
0
0
o
0
a
21
21
21
0
0
0
a
e
21
21
0
o
0
a
a
21
21
21
21
21
a
39
21
21
21
21
21
39
39
21
21
21
21
2 1
MORRISON
B
7
6
0
4
3
2
1
0
0
o
0
o
e
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0
0
17
0
0
0
0
O
0
0
17
0
0
0
0
0
0
0
17
17
17
17
17
17
0
0
39 39 39 39 96 96 62 62 62 62 40 40 40 60
EDGEHATEH
39 39 39 39 96 96 16 30 62 62 40 40 40 60
DENVER
39 39 39 39 16 16 16 16 87 87 45 45 45 11
21 21 21 21 16 16 16 16 B7 87 45 45 45 45
21 21 21 21 16 16 16 16 138 138 45 45 45 45
CLENDALE
21 21 21 21 66 66 66 66 138 138 127 127 45 45
l.AKEWOOD
21 21 21 21 66 66 66 0 138 138 127 127 127 45
17 21 66 66 66 108 66 40 40 138 127 127 127 43
SHERIDAN ENCLEHOOD
17 17 66 66 66 108 40 40 53 63 53 127 127 127
CHEIUIY HILLS
17 17 17 17 108 108 4O 40 53 53 53 53 127 S3
17 17 17 17 108 108 40 40 40 53 53 53 53 53
JEFF CO LITTLETON GREENWOOD VLG
17 17 17 17 108 40 40 40 40 53' 53 53 53 53
URBAN CLMBN VLY
17 17 17 17 108 40 40 40 40 53 53 53 53 53
17 17 17 17 108 40 40 40 40 53 53 53 53 53
17 17 17 00000000000
00000000000000
60
60
60
60
60
60
60
60
60
60
0
60
0
60
0
0
0
0
AURORA
1 1
11
11
1 1
1 1
11
53
53
53
53
53
53
0
0
1 1
1 1
11
11
1 1
1 1
11
11
53
53
53
63
0
0
11
11
11
11
11
11
11
11
11
11
53
63
0
0
11
11
11
1 1
11
11
11
11
11
1 1
11
11
0
0
11
11
11
11
1 1
11
11
11
1 1
11
11
1 1
e
0
11
11
11
11
"
1 1
1 1
11
11
11
11
11
0
0
11
11
1 1
11
11
11
U
11
11
11
11
11
0
0
1 1
11
1 1
11
1 1
11
11
11
1 1
11
U
II
e
0
1 1
u
11
11
u
u
1 1
u
1 1
u
1 1
u
e
0
FIGURE VI-5.
(a) Between 1975 and 1985
INCREASE OF EMPLOYMENT (IN TENS OF PERSONS) IN ONE-SQUARE-MILE GRID CELLS IN
THE DENVER METROPOLITAN REGION. Grid numbers are listed on the left side and
top of figure.
-------
30
29
20
26
26
24
23
22
21
20
19
10
17
1
0
0
0
0
12
1 O
1 >£
12
0
0
0
0
0
0
12
2
0
0
0
0
1 O
1 <£
12
12
12
0
0
0
O
12
3
O
o
0
0
1 2
1 2
1 **
12
12
12
12
12
12
4
O
0
0
0
1 **
12
12
12
12
12
12
12
12
5
i
0 '
O
0
0
12
12
12
12
12
1 2
12
12
12
6
0
0
0
0
12
12
12
12
12
12
12
12
7
0
0
0
A
12
12
12
12
12
30
3O
26
8 t 10 11 12 13 14 15 16 17 1O 19 20
00OO000000000
0000000000 14 O0
0 0 O O 14 14 14 14 14 14 14 14 14
0000 14 14 14 14 14 14 14 14 14
imOOMFlELD
12 12 12 12 02 32 3*i 32 32 J2 14 14 (4
NOHTIIGLENN
12 12 12 12 32 32 32 32 32 32 14 14 4
12 12 12 12 32 32 32 32 32 32 32 32 4
WESTMINSTER THORNTON
12 12 12 12 32 32 32 32 32 32 32 14 14
FDRL
12 12 12 12 32 32 32 32 32 32 32 14 14
IICTS
30 3O 30 30 32 32 32 32 32 32 14 14 14
30 30 30 30 32 32 32 32 32 32 14 14 14
ARVADA COMMERCE
30 3(1 30 30 32 32 32 32 32 14 14 14 14
30 30 3O 30 143 143 143 93 93 93 59 14 59
26 26 26 26 143 143 143 93 93 93 69 59 59
21
0
0
14
I A
I t
4
14
14
14
HCKY
14
14
59
59
22
O
0
14
4
14
14
14
MNTK
14
14
59
59
23
4
4
4
4
14
14
14
24
4
4
4
4
14
14
14
25
r
4
4
4
4
14
14
14
26
4
4
4
4
4
74
74
27
0
0
0
0
0
0
0
0
0
0
28
0
0
0
0
0
0
0
0
29
O
0
0
0
0
0
0
0
0
30
0
0
0
0
0
0
0
0
0
0
ARSNL
14
14
59
59
14
14
59
59
14
14
74
74
74
74
74
74
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
WIIT. IIDC BTAI'L I NTL
16
12
12
12
12
12
12
26
COLDEN
15
14
13
12
1 1
10
9
0
0
0
0
0
0
0
0
0
0
0
0
0
O
12
32
32
32
0
0
0
12
12
32
32
0
0
0
12
12
32
32
32
32
32
12
26
32
32
32
32
32
26
26
32
32
32
32
32
MOIUUHON
0
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
26
0
0
0
0
0
0
0
26
0
0
0
0
O
O
0
26
26
26
26
26
26
0
0
26 26 26 26 143 143 93 93 93 93 59 59 59
EUCEWATER
26 26 26 26 143 143 23 50 93 93 59 59 59
DENVER
26 26 26 26 23 23 23 23 127 127 60 6O 60
32 32 32 32 23 23 23 23 127 127 6O 60 60
32 32 32 32 23 23 23 23 205 205 6(1 6O 60
CLENDALE
32 32 32 32 99 99 99 99 205 205 191 191 60
LAKEWOOD
32 32 32 32 99 99 99 0 205 205 191 191 191
26 32 99 99 99 162 99 60 60 205 191 191 191
SHERIDAN ENCLEWOOI)
26 26 99 99 99 162 60 60 79 79 79 191 191
CIIEIlllY II 11.1.8
26 26 26 26 162 162 60 60 79 79 79 79 191
20 26 26 26 162 162 60 60 60 79 79 79 79
JEEE CO LITTLETON CREENWOOD VLC
26 26 26 26 162 60 60 60 60 79 79 79 79
UIIHAN CLPlliN VLY
26 26 26 26 162 60 60 60 60 79 79 79 79
26 26 26 26 162 60 60 60 60 79 79 79 79
26 26 26 0 0 0 0 0 O 0
-------
30
29
2U
27
26
25
24
23
22
21
2O
19
IB
17
1 f\
1 O
1 3
14
13
1 I
10
9
1
0
0
0
0
2
2
2
0
0
0
0
0
0
2
-6
-6
0
0
0
0
2
0
0
O
0
2
2
2
2
2
0
0
0
0
2
CO!
— fi
O
-6
0
0
0
0
3
0
0
0
0
i
2
2
2
2
2
2
o
2
2
,!>EN
-4
O
0
0
0
0
4
0
0
0
0
2
2
2
2
2
2
2
2
2
2
~ 1
-4
5
0
— 13
-13
5
0
0
0
0
2
2
2
2
2
2
2
-13
-1
-1
~ 1
-1
5
3
-7
6
' 0
0
0
0
2
2
2
2
2
-13
-13
-13
-1
-I
~*
0
-7
5
3
7
0
0
0
0
2
2
0
0
0
-13
-13
-9
-9
-2
8 9 10 11 12 13 14 15 16 17 10 19
000000000000
000000O00000
0 0-6-6-5-5-2-2 0 0 0 0
-6 -6 -6 -6 -3-5-2-2 0 0 0 O
BROOMF I ELD
-3-3 0 0 4 4 4-5-5-5 -12 0
MOUTHCLENN
0 0 0 0 4 4 4-5-5-5 -12 -12
0000444-5-5 -10 -10 -10 -
WESTMINSTER THORNTON
0000444-4 -10 -10 -10 -14
FDHL
0 0 0 0 4 4 4 -4 -4 -3 -3 -B
1ICTS
-9 -9-9-9 4 4 4 -3 -3 -3 -8 -B
-9 -9-9-9 4 4 4 -3 -3 -4 -B -B
ARVADA COMMERCE
-9 -9 -9 -9 -fl -3 -3 -3 -3 -B -(1 -fl
-9 -9 -9 -14 0 O 0 31 31 31 -5 -0
-2 -2-2-2 0 0 0 31 31 31 -5 -5
20
0
0
0
0
0
0
10
-a
-a
-a
2
2
-5
-5
21 22 23
0
0
0
0
0
-10 -
-10 -
-8
-a
2
RCKY
2
-17
-5
-5
0
0
0
0
0
10 -
10
2
2
2
MNTN
2
2
-5
-5
WUT. RDO 8TAPL 1NTL
0
0
-7
-7
3
MORRISON
0
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
EDCEWATER
000 -2 0 0 20 460 31 31 —5 -5
DENVER
0 0 0 0 20 20 20 20 59 59 3 3
-7 -7 -7 -7 20 20 20 20 59 59 3 3
CLENDALE
-7 -7 -7 -7 -11 -11 -11 -II 14 14 -2 -2
LAKEWOOD
-10 -7 -3 -11 -11 -10 -10 -5 -5 14 -2 -2
SUERIDAN ENCLEWOOD
0 -10 -3 -3 -11 1 5-5 0 0 0-2
CHERRY HILLS
000 -17 -10 1-5-5 0 0 0 -16
000 -17 -4 -4 -5 -5 12 2 2 2
JEFF CO LITTLETON GREENWOOD VLC
000070000-1-1-1
URBAN CLMBN VLY
000070080-1-1-1
00007000B -1-1-1
30300000000O
000000000000
-5
3
3
3
— 2
-2
-2
-2
2
-1
-'
-1
0
0
0
0
0
0
0
10
0
2
2
2
24
0
0
0
0
0
0
0
2
2
2
25
f
0
0
0
0
0
0
0
2
2
2
26
0
0
0
0
0
0
0
0
6
6
27
0
0
0
O
0
0
0
0
0
0
28
0
0
0
0
0
0
0
0
0
0
29
0
0
0
0
0
0
0
0
0
0
30
0
0
0
0
0
0
0
0
0
0
ARSNL
2
2
-5
-5
- 1 - 1 — i
AURORA
-6
3
3
3
3
16
2
2
- i
-1
-1
0
0
-6
-6
£
O
-6
-6
-6
4
4
_,
-1
-'
-1
0
0
-6
-6
-6
—6
1
1
4
4
-1
-1
0
0
2
2
-5
-5
- 1
-6
-6
—6
-6
-6
-6
-6
-6
1
1
-1
-1
0
0
2
2
-14
-14
— 1
— 6
-6
— 6
-6
— 6
-6
-6
-6
1
1
1
1
0
0
6
6
6
-14
'
-6
-6
— 6
-6
— 6
-6
-6
-6
,
1
1
1
0
0
0
0
-20
-20
- 1
-6
-6
-6
-6
-6
-6
-6
-6
-6
1
1
1
0
0
0
0
-20
-20
Q
— 1
-6
-6
~6
i
,
j
i
i
i
i
i
0
0
O
0
0
0
0
1
'
,
1
1
1
1
1
j
1
1
0
0
0
0
0
0
Q
°
1
'
(
1
1
1
1
1
1
1
1
0
0
FIGURE VI-6.
(a) Year 1975
DIFFERENCE BETWEEN POPULATION DURING NORMAL WORKING HOURS AND TRUE RESIDENTIAL
POPULATION, BY ONE-SQUARE-MILE GRID CELL. In hundreds of persons; grid numbers
are listed on the left side and top of figure.
-------
213
Since not all areas will have ozone concentrations exceeding the NAAQS,
there appear to be sharp boundaries to the regions where significant
exposure occurs, and these boundaries are defined by the boundaries of
the ozone cloud rather than by the population distributions. Note that
the area in which ozone concentrations exceed the NAAQS not only shrinks
in size but moves southward away from the high population concentration
in central Denver. The latter is a major contributor to the sharp
decrease in exposure for future years.
E. SUMMARY AND CONCLUSIONS
Dosages and exposures to ozone were estimated for the Denver metropol-
itan region for three years--1975, 1985, and 2000. The ozone concentrations
were obtained from the SAI Denver Model simulations of 3 August 1976 meteo-
rology and 1976, 1985, and 2000 emissions. The population distribution wars
obtained from the Sub-area Allocation Plan and was adjusted to account for
a shift in the population during working hours.
Dosages tend to reflect a general reduction in the ambient ozone con-
centration over the entire day. The estimated dosages showed a 25 percent
drop from 1975 to 1985 but a 6.5 percent increase from 1985 to 2000. When
discounted for the increases in population between these years, the dosages
reflect an effective reduction in ambient ozone by 38 percent from 1975 to
1985 and 18 percent from 1985 to 2000.
Exposure is a measure of the impact of relatively brief but high ozone
concentrations. Exposure to ozone about the NAAQS is projected to decrease
80 percent from 1976 to 1985 and another 93 percent from 1985 to 2000.
These results are biased because for the 3 August 1976 meteorology used
here the peak ozone concentrations are located south of central Denver and
away from the regions of high population. Peak ozone concentrations may
occur over central Denver on other days, and in such a case the improvement
in exposure levels for future years may not be as great as that obtained
here.
-------
214
REFERENCES
Anderson, J. A., and D. L. Blumenthal (1976), "Characterization of Denver's
Urban Plume Using an Instrumented Aircraft," in "Denver Air Pollution
Study-1973," EPA-600/9-76-007a, Vol. I, Proc. of a Symposium,
P. A. Russell, ed., Environmental Protection Agency, Research Triangle
Park, North Carolina.
Bailey, B. S. (1975), "Oxidant-HC-NO Relationships from Aerometric Data--
L.A. Studies," Scientific SeminaY on Automotive Pollutants, EPA-600/
9-75-003, 10-12 February 1975, Washington, D.C.
Burton, C. S. , et al. (1976), "Oxidant/Ozone Ambient Measurement Methods:
An Assessment and Evaluation," EF76-111R, Systems Applications,
Incorporated, San Rafael, California.
Code of Federal Regulations [C.F.R.J (1975), Title 40, Section 51.14(c)(4).
Colorado Department of Health (1976a), data summary sheets, private communi-
cation to G. E. Anderson, Systems Applications, Incorporated, San
Rafael, California.
(1976b), "Report to the Pub!ic-1976," Air Pollution Control
Commission, Denver, Colorado.
Colorado Division of Highways [CDH] (1976), "JRPP Air Quality Assessment
Statement (CY1976)," Denver, Colorado.
de Nevers, N., and J. R. Morris (1975), "Rollback Modeling: Basic and
Modified," J. Air Poll. Contr. Assoc., Vol. 25, pp. 943-947.
Denver Regional Council of Governments [DRCOG] (1976), "Policy Population
Forecast: Subarea Population Allocation," Denver, Colorado.
Dimitriades, B. (1975), Conference on the State of the Art of Assessing
Transportation-Related Air Quality Impacts, Workshop I, Session I
(Chemistry), 22-24 October 1975, Washington, D.C.
(1973), "Photochemical Oxidants," Chemistry and Physics
Laboratory, Environmental Protection Agency, Research Triangle Park,
North Carolina.
(1972), "Effects of Hydrocarbon and Nitrogen Oxides on Photo-
chemical Smcg Formation," Environ. Sci. Techno!., Vol. 6, No. 3,
pp. 253-260. ~ ' '
-------
215
Georgii, H. W., E. Busch, and E. Weber (1967), "Investigation of the
Temporal and Spatial Distribution of the Immission [sic] Concentration
of Carbon Monoxide in Frankfort/Main," Report No. 11, Institute' for
Meteorology and Geophysics of the University of Frankfort/Main,
Germany.
Hanna, S. R. (1971), "Simple Methods of Calculating Dispersion from Urban
Area Sources," J. Air Poll. Contr. Assoc., Vol. 21, pp. 774-777.
Holzworth, G. C. (1972), "Mixing He-ights, Wind Speeds, and Potential for
Urban Air Pollution Throughout the Contiguous United States," Office
of Air Programs, Environmental Protection Agency, Research Triangle
Park, North Carolina.
Jerskey, T. N., et al. (1976), "Sources of Ozone: An Examination and
Assessment," Report EF76-110R to the American Petroleum Institute
by Systems Applications, Incorporated, San Rafael, California.
Johnson, W. B. (1974), "Field Study of Near-Roadway Diffusion Using a
Fluorescent Dye Tracer," Symposium on Atmospheric Diffusion and Air
Pollution, American Meteorological Society, 9-13 September 1974,
Santa Barbara, California.
Khanna, S. B. (1976), "Handbook for Unamap," Walden Corp., Cambridge,
Massachusetts.
Larsen, R. I. (1971), "A Mathematical Model for Relating Air Quality
Measurements to Air Quality Standards," Office of Air Programs,
Environmental Protection Agency, Research Triangle Park, North
Carolina.
Ludwig, F. L., and J.H.S. Kealoha (1975), "Selecting Sites for Carbon
Monoxide Monitoring," SRI Project 3515, Stanford Research Institute,
Menlo Park, California.
Merz, P. H., L. J. Painter, and P. R. Ryason (1972), "Aerometric Data
Analysis, Time Series Analysis and Forecast, and an Atmospheric Smog
Diagram," Atmos. Environ., Vol. 6, pp. 319-342.
Paskind, J. J., and J. R. Kinosian (1974), "Hydrocarbons, Oxides of Nitrogen
and Oxidant Pollutant Relationships in the Atmosphere over California
Cities," 67th Annual Meeting, Air Pollution Control Association,
9-13 June 1974, Denver, Colorado.
Radian Corporation (1975), "Study of Vapor Control Methods for Gasoline
Marketing Operations," Austin, Texas.
-------
216
Roth, P. M., et al. (1976), "An Evaluation of Methodologies for Assessing
the Impact of Oxidant Control Strategies," Report EF76-112R to the
American Petroleum Institute by Systems Applications, Incorporated,
San Rafael, California.
Schuck, E. A., and R. A. Papetti (1973), "Examination of the Photochemical
Air Pollution Problem in the Southern California Area," Appendix D
of "Technical Support Document for the Metropolitan Los Angeles
Intrastate Air Quality Control Region Transportation Control Plan
Final Promulgation," EPA Region IX, Environmental Protection Agency,
San Francisco, California.
Scott Research Laboratories (1974), "Performance of Service Station Vapor
Control Concepts," Project EF-14 Phase II Interim Report CEA-8 to
the American Petroleum Institute, Washington, D.C.
Trijonis, J. C. (1972), "An Economic Air Pollution Control Model Application:
Photochemical Smog in Los Angeles County in 1975," Ph.D. Thesis,
California Institute of Technology, Pasadena, California.
Turner, D. B., J. R. Zimmerman, and A. D. Busse (1973), "An Evaluation of
Some Climatological Dispersion Models," Appendix E of "User's Guide
for the Climatological Dispersion Model," EPA-R4-73-024, Environmental
Protection Agency, Research Triangle Park, North Carolina.
Venturini, P. D., and D. M. Grandy (1975), "Background and Development of
the California Vapor Recovery Control Program for Service Stations,"
California Air Resources Board, presented at 4th Annual North American
Motor Vehicle Emissions Control Conference.
-------
217
APPENDIX A
DESCRIPTION OF THE MODELS
-------
218
APPENDIX A
DESCRIPTION OF THE MODELS
by James P. Kill us
Urban airshed models are mathematical representations of atmospheric
transport and chemical reaction processes which, when combined with a source
emissions inventory and appropriate meteorological data, may be used to
predict pollutant concentrations as a function of time and location in an
airshed. In the past several years, SAI has developed a number of airshed
models for specific research and applications tasks. Depending upon the
application, these models differ in the importance given to selected
physical and chemical phenomena, but all have certain features in common:
> A modeling region, divided into a number of horizontal and vertical
cells, and boundary conditions that define the concentrations of
pollutants in air entering the modeling region from the top and
sides.
> A source inventory which defines the emissions of pollutants into
each cell of the modeling region.
> A meteorological file which determines the advective transport of
pollutants within the region, and a diffusivity algorithm which
provides the rate of vertical diffusion of pollutants.
> A kinetic mechanism to predict the changes in pollutant concentra-
tions resulting from chemical reactions (for those modeling appli-
cations involving chemically reactive species).
In the next sections we discuss the Denver Air Quality Model and its
treatment of each of the above factors. The appendix closes with a much
briefer description of the well-known Climatological Dispersion Model (COM)
also used in this study.
-------
219
The Denver Air Quality Model--called simply the "Denver Model" here--is
*
an adaptation of several airshed models previously developed at SAI. The
starting point was the airshed model developed for the California ARB (Liu
et al., 1976a) to treat the advection and diffusion of S(>2 (or any linear
species) from power plant plumes. The model used SHASTA numerics, and had
the capability of treating winds (Reynolds et al., 1977) with vertical
variations. Because of its modular nature, this model was relatively easy
to modify in order to include:
> Treatment of 12 chemical species (the Carbon-Bond kinetic mechanism).
> A diffusivity algorithm developed specifically to aid in the treat-
ment of the onset of radiative inversions.
> An inversion breakup algorithm.
> Various input-output codes designed to facilitate treatment of the
grid conventions used by the Colorado Department of Highways.
1. THE MODELING REGION AND BOUNDARY CONDITIONS
The modeling region chosen was a 30 x 30 mile grid bounded at the bottom
by the ground and extending to the top of the mixing layer (or 2500 ft.,
whichever is less). Although it would be desirable to extend the modeling
region some distance into the inversion, for various reasons — including a
lack of pollutant measurements aloft—this extension is not yet within our
modeling capabilities.
The 30 x 30 mile region was chosen for the original CDH study in an
attempt to place the boundaries in areas where pollutant concentrations in
air flowing into the region could be expected to be near background levels.
The boundaries chosen are well beyond the areas of substantial emissions;
* The Colorado Department of Highways (CDH, 1976) used one of these models,
the Air Pollution Simulation Program (APSP), in an analysis of the 1-470
Highway proposal. The APSP differs from the Denver Model in some impor-
tant respects, notably the kinetic mechanism employed. The CDH/APSP
simulations gave markedly different results from the present study.
Factors that we discuss at some length below led us to the conclusion
that the ASPS is inappropriate for use in the Denver area, hence our
decision to develop the Denver Model.
-------
220
presumption of "clean" boundary inflows was thus quite acceptable except
when some of the pollutants that had been blown out of the modeling region
might have reentered it when the wind reversed. This possibility could
not be investigated in the present project because the region in question
(south of Denver) has uneven terrain and no data are available for character-
izing its wind field.
The 30 x 30 mile modeling region was divided horizontally into a
15 x 15 set of grid cells, with three stacked layers in the vertical direc-
tion. This reduction from the 30 x 30 x 5 cell modeling grid used in the
CDH/APSP study was made in the interest of reducing computing costs; the
Denver Model requires roughly four times the computing time of the
ASPS for a given simulation, owing to the doubling of the number of
chemical species. Several tests were made to see if the reduction in
spatial resolution would affect predictions. It was found that the results
from a 15 x 15 cell modeling region were nearly identical to an averaging of
the results from a 30 x 30 cell region. The change from five to three
vertical layers results in somewhat faster mixing, since each cell is assumed
to be a well-mixed box. However, this effect is minor because of the rapidity
of vertical diffusion in the Denver region.
2. THE EMISSIONS INVENTORY*
The starting point in preparing an emissions inventory suitable for the
Denver Model is the Emissions Data Preparation Program (EDPP) developed by
Reynolds (1973) for SAI's original APSP. The inputs to the EDPP include:
> Auto vehicle miles traveled (VMT) along major thoroughfares (in
link node format.).
> Auto area sources (traffic on side streets, etc.).
> Small point sources (industrial emissions, house heating, etc.).
A full description of the emissions data used in this project is given
in Appendix B.
-------
221
The output from the EDPP is a six species emissions inventory—NO,
N02, CO, reactive HC, unreactive hydrocarbons and a blank file for 0~
(since 0, is not a primary pollutant). The reactive hydrocarbon inventory
of the EDPP was then divided between the four categories required by the
Carbon-Bond mechanism. The assumed split was, by weight, 0.032 olefins,
0.65 paraffins, 0.05 aldehydes, and 0.205 aromatics (Reynolds et al., 1977);
the remainder was assumed to be unreactive and was disregarded, along with
the unreactive hydrocarbon emissions inventory in the EDPP. The resulting
inventory of the emissions of each of eight species in the 30 x 30 grid
was block-averaged into the 15 x 15 grid used in the Denver Model. Blank
files for HNOp, HUOp, peroxyacetylnitrate (PAN), and aerosols were also
added. Emissions from large point sources such as power plants were input
directly to the model at elevated locations, corresponding to risen chimney
plumes.
3. THE METEOROLOGICAL FILE AND DIFFUSIVITY ALGORITHM
The meteorological file used in the Denver Model is also a modification
of files prepared for the APSP. The Meteorological Data Preparation Program
(MDPP) interpolates surface wind measurements to generate a two-dimensional
wind field. The Denver Model requires input of a three-dimensional wind
field. Lacking both data aloft and the time for a detailed theoretical
estimation of winds aloft, in this project we simply extended the MDPP
surface winds aloft, effectively rendering the three-dimensional wind feature
of the Denver Model inoperative.
It is worthy of note that there is evidence of a breakup of the
inversion layer in Denver around midday on the days for which data were
available. (If there were no inversion present, the top of the modeling
region could not be the inversion base.) Under these circumstances we
chose to treat transport and dispersion explicitly in a limited modeling
region 2500 ft. high. At the top, pollutants are allowed to slowly
leak out of the modeling region at a rate based upon atmospheric dif-
fusivities and the concentration gradient within the region.
-------
222
The vertical diffusivity algorithm used in this study (Liu et al.,
1976b) was developed for application to Las Vegas, Nevada. This algorithm
was specifically developed to treat radiative inversions; communication
with the CDH suggested that the Lamb algorithm used in the CDH/APSP
(Lamb et al., 1975) had some difficulties in handling the onset of
radiative inversions. Comparison of the two algorithms suggests that
the Liu algorithm provides a better treatment of stable atmospheric
conditions, particularly at the onset of radiative stability, while
the Lamb algorithm is more accurate in its treatment of unstable condi-
tions. However, since the two algorithms yield only minor differences
in model predictions under unstable conditions (particularly with only
three vertical modeling layers.) but yield very different results under
stable conditions, we felt that the Liu algorithm was preferable for
this project.
We described above the features of th.e Denver Model that pertain
to emissions and atmospheric transport. In simplest terms, an airshed
model of reactive pollutants consists of three coupled mechanisms:
emissions, transport, and chemical reactions. Emissions and transport
(meteorology) are inputs to the SAI model. These processes are straight-
forward; our ability to account for them is restricted primarily by
the data available to us.
4. TREATMENT OF CHEMICAL REACTIONS
The photochemistry of urban smog is not as well understood. The
reactions known or postulated to occur in photochemical smog number
in the many hundreds. The number of reactions actually taking place
is probably many times that. Our ability to model the ozone formation
process rests upon the construction of various lumped kinetic schemes
that approximate the more complex reality.
A chemical kinetic mechanism is therefore the most critical fea-
ture of a photochemical airshed model. There are some general con-
siderations important in developing such a mechanism. First, the
-------
223
mathematical description of the mechanism (in terms of the number of
species included) must not he overly complex, or else computation times
for the model within which it is embedded are likely to be excessive.
On the other hand, an overly simplified mechanism may omit important
reaction steps and thus be inadequate to describe atmospheric reac-
tions over a range of conditions. A major requirement, then, is that
the mechanism include only a limited degree of detail, yet still pre-
dict the chemical behavior of a complex mixture of organic and inorganic
species. In particular, accurate prediction of the formation of ozone
is of prime importance, since ozone is the major constituent of photo-
chemical oxidant and has the largest data base for validation purposes.
In general, development of chemical kinetic mechanisms has three
stages: (1) laboratory experiments designed to elucidate specific
chemical reaction steps, their rate constants, and their products,
(2) smog chamber studies based upon rather simple analogs to the com-
plex atmospheric mixture, and (3) atmospheric studies, either fixed
ground-level observations of a changing air mass or airborne studies that
attempt to follow a single air mass (Calvert, 1976a, 1976b). The specific
reaction steps proposed at the first stage are linked together into a
specific mechanism (written for the photooxidation of a specific hydro-
carbon), which is validated against the smog chamber studies in the
second stage. These specific mechanisms are often quite complex, e.g.,
the Whitten and Hogo (1977) mechanism for propylene contains 65 reactions.
Any attempt to prepare and link together specific mechanisms for each
of the hydrocarbon species, in smog would be computationally intractable
even if sufficient theoretical backing for such a mechanism existed.
Instead, the insight gained in comparisons with smog chamber data
is condensed into a generalized lumped mechanism with the hydrocarbon
species grouped into one or more categories whose average behavior is
then represented by a single set of reactions for each species. Lumped
mechanisms are validated against both smog chamber data and, when
coupled with a transport model (either an airshed or trajectory model),
atmospheric data.
-------
224
Tables A-l through A-3 list the reactions used in the Hecht-Seinfeld
C1972), Hecht-Setnfeld-Dodge (Hecht et al., 1974a), and the original formula-
tion of the Carbon-Bond (Whitten and Hogo, 1977) mechanisms. These mechan-
isms were developed by SAI under EPA auspices in attempts to incorporate
state-of-the-art chemistry into a computationally efficient mechanism.
Each successive mechanism represents what we believe to be a major
advance over the previous mechanism. In the case of the Hecht-Seinfeld
mechanism, at least, subsequent advances (principally the addition of
aldehyde chemistry) have rendered it obsolete. This is conclusively
demonstrated by the present project, which represents the first side-
by-side comparison of different kinetic mechanisms applied to the
atmosphere.
The Carbon-Bond mechanism represents, an advance over the Hecht-
Seinfeld-Dodge mechanism that is more subtle than the difference between
the H-S and H-S-D mechanisms. In addition to realistic aromatic chemistry,
the Carbon-Bond mechanism offers an improved lumping scheme, mass-con-
servative hydrocarbon depletion, and an elimination of sensitive stoi-
chiometric parameters. The Carbon-Bond mechanism therefore requires
less tuning and is more user-oriented than the H-S-D mechanism.
The test of any simulation model is two-fold: (1) Does it obey the
requirements of accepted scientific theory? (2) Are its predictions in
acceptable agreement with observations? Below, we present the basic
details of atmospheric photochemistry as they apply to ozone formation
in photochemical smog and a discursive comparison of the utility of
the H-S, H-S-D, and C-B mechanisms. We then consider the predictions
of the Denver Model in comparison with the observed air quality in
Denver.
a. Urban Airshed Photochemistry and Its Simulation by Lumped Kinetic
Mechanisms
Ozone and other oxidants are formed within the troposphere primarily
by the chemical interaction of nitrogen oxides and hydrocarbons. This
-------
225
TABLE A-l. THE HECHT-SEINFELD MECHANISM
NO + hv ->• NO + O
0+0+M + O+M
O. + NO -»• NO + 0
•J £.4.
°3 + N°2 "* N°3 + °2
N°3 + N°2 H0 2
6
NO + N02 H+Q 2
HN02 + hv -»• OH + NO
CO + OH ^ C02 + HO
9
HO., + NO -*• OH + NO_
10
HO + NO -»• HNO + 0
11
HC + 0 -> aRO
12
HC + OH -> SRO
13
HC + 0, -
RO + NO ->• N02 + eOH
15
RO + N02 ->• PAN
Source: Hecht (1972).
-------
226
TABLE A-2. THE HECHT-SEINFELD-DODGE MECHANISM
N02 + hv - NO + 0
NO
>The N02-NO-03 Cycle
0 + NO + M -»• N02 + M
0 + N02 H. NO + 02
0 + N02 + M •>• N03 + M
Important Reactions of 0
with Inorganic Species
03 + N02 -*• N03
N03 + NO i 2N02
N03 + N02 ->• N£05
N0
2HN0
>The Chemistry of
'N00C, and HN00
25 3
NO + HNO.
HN02 + HN03
12
1 3
HN02 + N02
2N0
I Reactions of HN03 with
/ Inorganic Species
NO + N0
2HNO,
is
2KN02 -- NO + N02
16
HNO, + hv -»• OH + NO
Chemistry of HN02
-------
227
TABLE A-2 (Continued)
OH + NO,
17
18
HNO.
OH + NO + M + HNOo + M
OH + CO + (02) +' C02 + H02
Important Reactions of
OH with Inorganic Species
20
H02 + NO + OH + N02
idation of NO by HO
hv " 20H
'Photolysis of
22
HC, + 0 -*•" ROO + aRCOO + (l-a)HO?
I n c
0
HC, + 00 -»• RCOO + RO + HC,,
HC] + OH -* ROO + HC4
HC, + 0 -+ ROO + OH
i*
HC2 +OH -»• ROO + H20
27
HC, + 0 + ROO + OH
26
HC3 + OH •* ROO + H20
HC4 •«- hv " 6ROO + (2-B)H02
HC. + OH 3->° BRCOO+(1-B)HO?+ H?0
4 n *•£.
0
Organic Oxidation Reactions
HC] = Olefins
HC2 = Aromatics
HC3 = Paraffins
HC4 = Aldehydes
-------
228
TABLE A-2 (Concluded)
31
ROO + NO - RO + N02
32
RCOO + NO +(09) -* ROO + NO, + C0
HC. C.
3 3
RCOO + N0
RO
RO + NO,
31.
35
36
RCOONO,
n t
0
H02 + F
RONOo
RO + NO "^ RONO
Reactions of Organic
)Free Radicals with NO,
0, and 0
37
H02 + H02
H02 + ROO -^ RO + OH + 02
38
2ROO
39
2RO + 0,
Other Peroxy Radical
Reactions
Source: Hecht et al. (1974a).
-------
229
TABLE A-3. THE ORIGINAL FORMULATION OF THE
CARBON-BOND MECHANISM
Rate Constant
Reaction
N02 + hv -»• NO + 0-
0- + 02(+ M) -* 03 (+ M)
03 + NO -v N02 + 02
0- + N0£ -* NO + 02
03 + N02 •* N03 + 02
N03 + NO + N02 + N02
N03 + N02 + H20 -»- 2HN03
NO + N02 + H20 ->• 2HN02
HN02 + hv -»• NO + OH
N02 + OH- -v HN03
NQ + OH- -* HN02
CO + OH- H- C02 + HO^
°2
OLE + OH- •£ HCHO + CH^
°2
PAR + OH- 4 CH^ + H20
°2
ARO + OH- $• HCHO + CH^
202
OLE + 0- -4 HC(0)02- + CH3Oj
°2
PAR + 0- -i CH302 + OH-
20?
ARO + 0- -S Hl(0)02 + CH3Oj
°2
OLE + 03 ^ HC(0)0^ + HCHO + OH-
/ -1 • -1\
(ppm mm )
k*
Kl
2.08 x 10'5
25.2
1.34 x 104
5 x 10"2
1.3 x 104
1.66 x 10"3t
2.2 x 10"9t
k*
KHN02
9 x 103
9 x 103
2.06 x 102
3.8 x 104
1.3 x 103
8 x 103
5.3 x 103
20
37
0.01
-------
230
TABLE A-3 (Concluded)
Reaction
Rate Constant
(ppm" min~ )
ARO +
HCHO + OH.
0.002
OLE + 03 -»- ozonide
20,
HCHO + hv -* HC(0)02 + H02
0.005
k*
HCHO + radicals
HCHO + hv -*• CO + H,
^HCHO -> CO
HCHO + OH-
1 x 10H
+ NO + OH- + N0
2 x
+ NO -*• N02 + HCHO +
2 x 103
NO
2 x 10-
+ hv -* OH- + OH-
H2°2
HO
+ 0
4 x ID"
HC(0)0-
H3COOH
HC(0)OOH + 0
PAN
PAN ->• HC(0)0*2
ARO + N03 -^ PRODUCTS
4 x 10
1 x 104
150
0.02
50
20.
* Photolysis rate constants in units of min'1.
t Units of
Source: Whitten and Hogo (1977).
-------
231
interaction is driven by ultraviolet radiation from the sun. The
nitrogen oxides and hydrocarbons, the precursors for oxidant formation,
are emitted from both natural and anthropogenic sources. Their concen-
trations and the intensity of solar ultraviolet radiation are affected
by meteorological phenomena.
The interactions between pollutants that lead to ozone (oxidant)
formation are complex and nonlinear. However, many years of research
have established certain basic interactions, which are divided into
three areas for this discussion:
> Inorganic chemistry—the interactions between nitrogen com-
pounds and oxygen compounds.
> Organic oxidation chemistry—the primary paths by which hydro-
carbons and their oxygenated intermediates participate in
ozone (oxidant) formation.
> Induction effects—the sources of radicals that can start the
chain reactions involved in ozone (oxidant) formation.
1) Inorganic Chemistry of Nitrogen and Oxygen Compounds
Ozone, the major oxidant in photochemical smog, is formed when
oxygen atoms react with oxygen molecules in the presence of a third
body (usually a nitrogen molecule or another oxygen molecule):
0 + 02 + M->Oo + M - (A-l)
The major source of oxygen atoms in the troposphere is the photolysis
of nitrogen dioxide:
N02 + hv + NO + 0 , (A-2)
where hv represents the energy of a photon of light at the frequency
v. The quantum yield for this reaction, or the percentage of absorp-
tions of a photon that result in photolysis, is near unity for UV light
-------
232
with a wavelength of 40Q run (nanometers) or less. The quantum yield
decreases at higher wavelengths and is near zero at 430 nm. The rate
of photolysis depends on the number of photons, or the intensity of
the light, and on the concentration of N02-
The following reaction:
NO + 03 •*• N02 + 02 (A-3)
completes an ozone production and destruction cycle. In the troposphere,
Reaction (A-2) yields an oxygen atom that generates ozone by Reaction
(A-l). However, Reaction (A-2) also generates an NO molecule that can
combine with ozone by Reaction (A-3) to regenerate a molecule of N02-
These three reactions are quite rapid; thus NO, N02, and 03 are usually
near the equilibrium relationship given by:
k2[N02] = k3[NO][03J . (A-4)
Equation (A-4) is usually a good approximation (within 10 percent) of the
actual concentrations of those species in th.e atmosphere.
Clearly, the above reactions will result in no net ozone production
and buildup. High ozone concentrations in the troposphere are the
result of the introduction of hydrocarbons into the Ni /0-, cycle.
X *3
2) Organic Oxidation Chemistry
So far, the reactions of only oxygen and nitrogen compounds have
been discussed. If water vapor and hydrocarbons are included, however,
additional species--such as HO-, and H02, alkylperoxy radicals (R02),
HON02, and HONO--are formed from the interactions between H~0, NO ,
C. A
and the hydrocarbons. The most important reaction involving these
species is:
HO;, + NO + HO- + N02 - (A-5)
-------
233
Hydroxyl radicals (HO-) resulting from Reaction (A-5) are the main
radicals that begin the oxidation of organic compounds. A reaction
related to Reaction (A-5) involves alkylperoxy radicals:
R02 + NO •* RO- + N02 , (A-6)
where R is an organic radical such as methyl (CHA) or ethyl
Reactions (A-5) and (A-6) convert NO to NOp. Rewriting equation (A-4)
as
°3 ~ k3[NOJ ' (A~7^
we see that an NO to N02 conversion raises the steady-state concentra-
tion of ozone. Early in the smog formation process, while there is
still a low N02 to NO ratio, the efficiency of ozone production from
NO to NOo conversions is very low. During this period, known as the
induction period, any ozone formed tends to rereact with the abundant
NO available to form N02. It is only after most of the NO has been
oxidized to N02 that the efficiency of ozone production rises. After
the induction period, when the N02/N0 ratio is high, practically every
NO to N02 conversion results in the production of an ozone molecule.
The ability of organic compounds to supply peroxy radicals for the
conversion of NO to N02 by Reactions (A-5) and (A-6) is their most
important role in the production of photochemical oxidant. The
rate at which the peroxy radicals are produced depends on the reactiv-
ity of the hydrocarbons.
Thus, the basic ozone production cycle may be outlined as follows
Hydroxyl radicals (HO-) react with organic compounds (hydrocarbons,
oxygenated hydrocarbons such as aldehydes, and so on) in a series of
reactions to produce peroxy radicals. These peroxy radicals (ROA and
H02) convert NO to N02 by Reactions (A-5) and (A-6) and regenerate the
hydroxyl radicals by Reaction (A-5). When the sun is shining, the
ozone concentration increases as the N02/N0 ratio increases, as shown
in the steady-state relationship of Eq. (A-7).
-------
234
Support for this outline comes from numerous research studies
CNiki et al., 1972; Demerjian et al., 1974; Hecht et al., 1974a;
Nicolet, 1975; Calvert and McQuigg, 1975; Chameides and Walker, 1976).
The relative unimportance of water vapor in photochemical oxidant
production, for example, was shown by Demerjian et al. (1974). They
performed a computer simulation of addition of water vapor to a system
containing "clean" air, 0.1 ppm NO, and no N02- They found that water
vapor had almost no effect on ozone production. In contrast, water
vapor is important in polluted atmospheres, but within the normal dew
point range its concentration is not limiting to the kinetic mechanism
and therefore is not discussed here. The same workers studied the
effects of carbon monoxide, which is found in both polluted and unpol-
luted air. Carbon monoxide was once thought to be relatively unreac-
tive but is now known to enter into the chemistry of photochemical
smog by the following reactions;
HO- + CO -> C02 + H. , (A-8)
H- + 02 + H02 , (A-9)
H02 + NO + HO- + N02 . (A-5)
Since HO- radicals are destroyed in Reaction (A-8) and produced in
Reaction (A-9), the above sequence provides a chain reaction whereby
a single HO- radical could initiate the oxidation of many NO molecules.
To see the effect of this mechanism, Demerjian et al. (1974) added
10 ppm of CO to the computer simulation described above. After irradia-
tion for eight hours, the "clean" air, H20, NO, CO atmosphere "produced"
over 0.04 ppm of ozone, an increase of more than a factor of 100 over
the "clean" air, H20, and NO system. The importance of carbon monoxide
in smog formation is limited, however, by the low rate constant for
Reaction (A-8) as shown by Hampson and Garvin (1975). In actual pol-
luted atmospheres, the conversion of NO to N02 by oxygenated hydro-
carbons and other intermediate species found in photochemical smog
is so rapid that the effect of CO is obscured.
-------
235
Hundreds of hydrocarbons have been identified in the urban atmosphere,
yet only a few of them have been studied in detail. The discussion here
is therefore limited to those classes of reactions that are currently
thought to be important. As mentioned above, the main contribution
of organics is to oxidize NO to N02 through reactions with peroxy
radicals, both peroxyalkyl (R0£) and peroxyacyl (RCO^), where R is
hydrogen or part of a hydrocarbon molecule. These reactions are
written in the general form:
R0£ + NO -> RO- + N02 , (A-6)
0 0
I! I!
RCOO- + NO -> RCO- + N02 . (A-10)
The peroxyalkyl radical is formed when the alkyl radical, R-, reacts
with an oxygen molecule. Alkyl radicals are formed at various stages
when paraffins, olefins, or aromatics are oxidized in the smog mechanism.
The peroxyacyl radical is produced by a similar reaction scheme via
the acyl radical,
0
I!
RC
that is usually formed when an aldehyde, RCHO, is oxidized. Aldehydes
are introduced directly into the atmosphere from auto exhaust and other
combustion sources and are also formed as intermediates in photochemi-
cally active atmospheres. Computer simulations, done by many workers
indicate that the hydroxyl radical, HCN, is the dominant species that
attacks hydrocarbons to form alkyl and acyl radicals (Niki et a!., 1972;
Demerjian et al., 1974; Hecht et al., 1974a; Nicolet, 1975; Calvert
and McQuigg, 1975; Chameides and Walker, 1976).
The high rates of NO oxidation and subsequent ozone production
found in photochemical smog have been accounted for by a reaction chain
that begins with a hydrocarbon being attacked by an hydroxyl radical.
The peroxy radical formed then oxidizes NO, and HO- is eventually
-------
236
regenerated. This sequence can be summarized by the following set of
reactions:
RH + OH- -»• R- + H20 , (A-ll)
R. + 02 + R02 , (A-12)
R0'2 + NO -v RO- + N02 , (A-6)
RO- + 02 + R'CHO + HO^ , (A-13)
RO- -* HCHO + R- (A-14)
+ NO + OH- + N02 , (A-5)
and for aldehydes:
RCHO + OH- + 02 -* RCOO- + H20 , (A-15)
0 0
II II
RCOO- + NO -v RCO- + N02 , (A-10)
0
il
RCO- -> R- + C02 , (A-16)
followed by Reactions (A-12), (A-6), (A-13), and (A-5). These are two
HO- cycles wherein HO- starts the cycle and is returned at the end of
the cycle. The first cycle results in two NO molecules being oxidized
to N02 with the concomitant reduction of one carbon atom from the
organic (RH). The second cycle results in three NO molecules oxidized
when the organic is an aldehyde. The HO- cycles are terminated when
two free radicals combine. This overall process is responsible for
converting NO molecules to N0? and converting the reactive hydrocar-
bons to C02 and H20. Note that aldehydes resulting from Reactions
(A-13) and (A-14) can react by (A-15), (A-10), and other reactions.
Aldehydes can also photolyze and bring more radicals into the system:
-------
237
RCHO + hv •* HCO' + R- , (A-17)
HCO- + 02 + CO + H0£ , (A-18)
0
HCO« + 0
The number of NO molecules converted in the original two cycles is
complicated by Reaction (A-17) followed by either (A-18) or (A-19).
The two ratios of the rate constants of Reactions (A-13) to (A-14)
and (A-18) to (A-19) are not currently known with certainty (Demerjian
et al . , 1974). It can also be shown that reactions like (A-ll) can
proceed by addition:
OH
l
RH + OH- -> HR-
with the series of reactions beginning with Reaction (A-12), producing
similar results. For substituted aromatic compounds, another type of
reaction may occur:
+ OH- -> I 0 I + H- (A-20)
The ultimate fate of aromatic compounds in smog chemistry is not
certain at the present time, however. For example, the number of NO
conversions per cycle when the aromatic ring eventually breaks open
is not known. The reaction scheme [Reactions (A-6) and (A-10) through
(A-19)] does, however, include the dominant organic reactions that
convert NO to N02> which is the major step required to raise the ozone
concentration (via the steady-state relationship). Many other differ-
ent organic species and inorganic intermediate species are also formed
in photochemical smog (Demerjian et al., 1974, pp. 65-108). Although
the reactions involving these species are not well understood at this
-------
238
time, it is expected that their effect is only to modify and not sub-
stantially change the results predicted with regard to 03 production
using only the simplified reaction scheme described above.
Explanation of the formation of complex organic molecules requires
that some additional reactions be considered. The most important of
these compounds are the PAN (peroxyacetyl nitrate) types that cause
eye irritation and were first discovered in the Los Angeles atmosphere.
PAN-type compounds are formed from the reaction between the peroxyacyl
radicals and NO^:
RCOO- + N09 -> RCOONO- . (A-21)
II ^ II ^
0 0
This reaction competes with Reaction (A-22):
RCOO- + NO + NO, + RCO- . (A-22)
II z II
0 0
Reaction (A-21) will dominate when the NOo/NO ratio is high.
3) Induction Effects
As previously noted, before ozone can begin to form and accumulate
rapidly, a sizable fraction of precursor NO must be oxidized to N02
(90 to 99 percent of NOY emissions are initially in the form of NO).
A
The time required for this initial oxidation is called the induction
period.
In order to explain the formation of photochemical smog it is
necessary that there be a source of radicals sufficient to build and
maintain a radical concentration many times greater than that found
in clean air. At such concentrations, the induction period is shortened
to a few hours and ozone production proceeds efficiently during the
hours of brightest sunlight.
-------
239
The hydroxyl radicals that play an important part in the mechanism
to convert NO to N02 are, as discussed above, regenerated at various
stages in the mechanism, but the source of the hydroxyl radicals, an
important issue, was not discussed. Virtually any type of radical
can initiate the oxidant-producing mechanism by means of radical trans-
fer reactions. For example, Reaction (A-5) transforms peroxy radicals
to hydroxyl radicals.
Maintaining the total concentration of radicals in the mechanism
by carbonyl photolysis is necessary because of the numerous temporary
and permanent radical sinks. For example, both nitrous acid and
hydrogen peroxide are temporary radical sinks because they are formed
rapidly via reactions that consume radicals:
OH- + NO -> HN02 , (A-23)
H02 + H02 + H202 . (A-24)
However, both species can reproduce radicals at a later time by photo-
lysis. The most important permanent radical sink also appears to be
the most important sink of oxides of nitrogen:
HO- + NO •* HONO . (A-25)
Other radical sinks include reaction with surfaces and various radical-
radical reactions:
R02 + R02 -> RO + RO
Some early investigators thought that oxygen atoms from N02
photolysis might be the source of radicals that initiate the mechanism
of ozone formation (Leighton, 1961). However, more recent work has
shown this radical source to be insufficient in explaining the high
radical concentrations in photochemical smog (Hecht, et al., 1974b).
-------
240
Because most reactions that generate radicals are photolysis
reactions, the intensity of sunlight affects the induction period.
However, slight variations in trace quantities of nitrous acid, HN02,
(Chan et al., 1976) or aldehydes can easily be shown to affect induc-
tion time strongly (Calvert and McQuigg, 1975), which suggests that
either of these substances could be the source of initial radicals.
In smog chambers, the photolysis of nitrous acid,
HN02 + hv -»• HO- + NO , (A-26)
may be the most important initial source of radicals. Nitrous acid
has been detected in smog chambers in concentrations sufficient to
explain the observed induction time for smog chemistry, but the con-
centrations necessary to initiate smog chemistry in the atmosphere
are below the limits of measurement of most modern instruments.
Nitrous acid is produced in the absence of sunlight:
NO + N02 + H20 + HN02 + HN02 . (A-27)
This reaction proceeds rapidly on some surfaces and may also proceed
slowly in the gas phase (Calvert and McQuigg, 1975). High local con-
centrations of the three reactants could lead to nitrous acid forma-
tion in the real atmosphere, but the importance of this reaction
is not known at present (Chan et al., 1976). Since nitrous acid is
produced in photochemical smog by Reaction (A-23), some nitrous acid
might persist to initiate radicals and then smog formation on a sub-
sequent day. Ripperton et al. (1975) and Smith et al. (1976) have
suggested that the presence of these radicals may be responsible for
increased ozone formation on the second day.
Aldehyde photolysis is critical in maintaining the radical con-
centration in smog because it is a primary pollutant and it is also
generated from other hydrocarbons in the photochemical process
-------
241
[Reaction (A-14)]. Furthermore, the formation of an aldehyde does not
result in the net subtraction of any radicals from the radical pool
as does the formation of HNOp and ti-Q^.
Aldehydes are, however, very difficult to treat in atmospheric
models. Emissions, data concerning aldehydes are largely lacking and
there are few atmospheric measurements of aldehydes. Aldehydes are
also difficult to treat in kinetic mechanisms, owing to the variety
of their reaction pathways—photolysis that produces radicals [Reaction
(A-13)], photolysis that does not produce radicals [Reaction (A-28)],
and reaction with hydroxyl radicals [Reaction (A-15)] which leads to
oxidation of the carbonyl without net production of radicals:
RCHO + hv -> HCO- + R- (A-13)
RCHO + hv -> RH + CO (A-28)
RCHO + OH- + 02 + RCOO- + H20 . (A-15)
0
Recognition of the crucial role of aldehydes in smog chemistry is
a recent development; early attempts to devise photochemical kinetic
mechanisms omitted aldehydes, (Eschenroeder and Martinez, 1972; Wayne
et al., 1971; Hecht and Seinfeld, 1972). These mechanisms could be
made to approximate the simplified behavior of a smog chamber. How-
ever, they predict unrealistic behavior in the atmosphere, although
this fact has until recently been masked by the complex behavior of
the urban airshed models. In the next section, we discuss the inade-
quacies of the kinetic mechanism in the APSP, and our reasons for
choosing a more up-to-date mechanism for the Denver Model.
b. Critique of the APSP Incorporating the Hecht-Seinfeld Kinetic
Mechanism
The original airshed model developed by SAI in 1972, the first
photochemical airshed model, was recognized at the time as a first-
-------
242
generation effort. It was anticipated that subsequent research and
development would expand the base of knowledge and allow advances in
model conception and execution. The H-S kinetic mechanism in that
model reflected a knowledge of smog chemistry that was sketchy, but
the best available at that time. We will now consider that mechanism
in light of present knowledge to support the need for new formulations
that adequately model realistic atmospheric situations.
The original H-S mechanism contained six chemical species: NO, NQ2,
03, CO, and two lumped hydrocarbon species representing reactive and
unreactive hydrocarbons. Unreactive hydrocarbons did not enter the
photochemical process at all; CO entered only marginally, by slightly
enhancing the formation rate of ozone. However, the photochemistry
primarily depended on the species. NO, NO,.,, 0^, and reactive hydrocarbons
(RHC). Of these species, only N02 photolyzed to produce oxidizing
radicals, the sine qua non of photochemical smog formation. The nitrogen
oxides, NO and NOp, also were assumed to be a minor source of radicals
via the reactions NO + N02 + H20 + 2HN02 and HN02 + OH- + NO. Because
HN02 was assumed to be present in steady-state concentrations, that
species was eliminated from the computations and the net result was
the reduction of N02 to NO and the production of OH radicals.
It is now known (Hecht et al., 1974a') that the photolysis of
aldehydes is the major source of radicals during most of the photo-
chemical process. The H-S chemistry neglects this reaction, with two
major consequences. First is the major reduction in the reactivity
of an atmospheric mix, with a consequential increase in the induc-
tion period—the time required to oxidize significant amounts of NO
to N02- It is only after this induction period that ozone is efficiently
produced, as illustrated in Figure A-l.
Second, radical production in the H-S mechanism proceeds according
to the product of sunlight and N02 rather than the product of sunlight
and aldehydes as more recent work has shown to be the case. Aldehyde
-------
243
t - tat
r - tan
A •.!•
T
I
» • 0
9 9 <
II
QOO
3 9
• 32 «
1 3 9
19 00
2 x ooo
m«inira237mini oo
33 3 « » » » 90
3333 v*n a o o
3333 1 3 » 09
23333332333 rrTrTTTrrfrrr rrr TT
»oo i UTI rarrrrrrr r r TTT T i
0 OOOO O FT r PTTT 233 23 3
000090 r r r r r r xmn » 2 23 2 2 2
p TT7TT rrrTFTrrrrr nrrr r r FT s n B i sira JTH » » 5H n 222 2=3 ^ 2 233 2 :
IM.M 31*.«•
DDTOIl C3 S7CC1Z3 <13 flO V13 COACCTTQATI OK
(a) Carbon-Bond Mechanism
9 * 03
X * W
3 - wn
i
ill am x m i
3 i ^^
233 v it mm * i a
3 33
3 3
3 333
3 3333
33 3 23 3
OOO OO 9 OO 0 OOOOO 0
ib) Hecht-Seinfeld flechanistn
FIGURE A-l. RESULTS OF TWO BOX MODEL SIMULATIONS (DIURNAL SUNLIGHT
AND CONTINUOUS EMISSIONS TO REPRESENT URBAN TYPE CON-
DITIONS) FOR COMPARISON OF KINETIC MECHANISMS. The
Hecht-Seinfeld mechanism used rate constants employed
In the CDH study: note the negligible production of
-------
244
concentrations are not proportional to N0? concentrations. The differ-
6 c *
ent production and destruction mechanisms of the two species assure
that their peaks occur at different times. In order to provide for
adequate radical and subsequent ozone production with the H-S mechanism,
reaction rates or initial or inflow concentrations might be arbitrar-
ily increased, but such artificial increases would not reconcile the
H-S mechanism with the currently accepted principles (see Figure A-2).
In addition to new understanding of the chemical reactions involved
in ozone production, a significant improvement has been introduced
in the treatment of steady-state conditions at the modeling boundaries
or at the beginning of the simulation.
Making the steady-state assumption for any species requires assum-
ing that all transient effects have died out, and that the concentra-
tion of a species is equal to the production rate divided by the destruc-
tion rate. In the atmosphere, ozone is produced by the Reaction (A-29)
and consumed by the Reaction (A-30):
N02—*-03 + NO (/!
k3
NO + 03-*-N02 (A-30)
Thus the steady-state relationship for ozone is
k, N0?
°3 = iTW" (A~31)
where k-j and k3 are the rate constants in the reactions.
Invoking the steady-state assumption for the boundary conditions
of the SAI APSP model is necessary because of the speed of these two
reactions. Introducing material that is too far from steady state into
the modeling region through the boundaries results in difficulties
-------
a *a a • • •
urn m uiim u moan* m m m uwm m •»•
MJIVUI 1C* •rum M W •<>» O*PCUT«ATI*R MULI I
(a) Increase in H-S Ozone-Hydrocarbon Reaction
M •• • •• a «
•• !• iri oo
• a •• • •
a • • •
a ••• «
* • • • •
i aa • *
u s a •• •
» i • •
• ••••u rrrrrrrrrr
• • • s tasa r rrr rr
eo • » r r
• • ft rrr rr t • i
oooo r t rr u xa
«•• r r • • •>
• oo r r r ••• u a •
oo* rrrr •• a • • a
DO rrrr rr • •
rr rr r rr rr rr rr rr r r r • mm*
tm ^* • ••••*•»••••••
• !•! •• lit M »!•.*• 4=*.fl* I3I.M «a*.M
TIME (•iivrai
»»VU Cl WICIU M M lOS OOMCUnUTIOC KAIX rMHVB (•>•
rrr rrr r rr
ro
-p=»
01
uaunaaiaa
(c) C-B Mechanism
FIGURE A-2. RESULTS OF ATTEMPT TO "TUNE" H-S
MECHANISM TO PRODUCE OZONE EQUIVA-
LENT TO THE PREDICTIONS OF THE
CARBON-BOND MECHANISM IN FIGURE A-la.
Note that in both H-S cases the ozone
peak occurs much earlier than in the
C-B case. The C-B ozone peaks were
found to be in phase with observed
peaks in validation experiments on
the Denver Model.
• • us* • • taia t • • • a*a tuaaau
iiai mttatta ea ••
i *otL* rumm
(b) Increase in H-S Radical (Oxygen Atom)-Hydrocarbon Reaction
-------
246
for the finite difference method used in solving the chemical equa-
tions; at best it wastes computer time, at worst the model, unable to
follow large fractional changes in small concentrations, will not
function at all.
Using the steady-state assumption for ozone alone introduces a
major flaw: in reality, all three species are varying. If, for example,
one starts off with values of 0.01 ppm and 0.04 ppm for NO and N02 and
0.5/20 for k-j/kg, Eq. (A-31) gives a value of 0.1 ppm for Oj. But a
cursory glance at Eq. (A-29) immediately shows that 03 cannot exceed
N02. Obviously Eq. (A-31) is in error. In fact, the steady-state
assumption must be invoked for all three species. This error in SAI's
basic airshed model has been rectified only recently. Thus the 1969
Los Angeles validation studies were introducing as much as 0.15 ppm of
ozone via the boundaries, as well as significant quantities of NO
/\
in the form of NOp, already "primed" and ready to react with hydrocar-
bons, without the delay of the induction period.
The ozone concentrations computed in the APSP model runs made by
the Colorado Department of Highways, using the H-S kinetics, were
similarly enhanced by boundary condition values (a programming error
accentuated this effect). Also, the CDH increased the sunlight factors
somewhat above what could reasonably be expected in Denver, in order
to get a better fit to observed data. Under the circumstances, this
was perhaps the most reasonable thing for them to do.
In summation, then: the H-S kinetic mechanism contains flaws
that make it unsuitable for modeling urban atmospheres. These flaws
became obvious only with the advent of the second generation of models
such as the 12-species Denver Model. The atmospheric testing of second-
generation models is quite recent. Indeed, to a large extent, the pre-
sent project is the first comparison of the two generations. But now
that the comparison has been made, it is clear that the new chemistry,
including aldehydes, must be used if we are to attempt to model the
realities of urban atmospheres.
-------
247
c. Choice of a Kinetic Mechanism for the Denver Model
SAI's initial airshed modeling efforts, started in 1972 under EPA
Contract 68-02-0580, resulted in the development of the Hecht-Seinfeld-
Dodge (H-S-D) kinetic mechanism (Hecht et al., 1974a). This work is
described in three reports to the EPA: a detailed planning document
(Seinfeld et al., 1973) and two final reports (Hecht et al., 1973,
1974b).
The H-S-D mechanism struck a balance between compactness of form
and accuracy of prediction by treating the important inorganic smog
reactions explicitly and b.y treating hydrocarbons in four groups, or
"lumps": olefin, paraffin, aromatic, and aldehyde. It incorporated
the latest knowledge of smog chemistry, but because of gaps in that
knowledge it also incorporated some empirical parameters. The values
of these parameters were chosen by fitting predicted concentration-
time profiles to a set of smog chamber data.
The H-S-D mechanism was a significant improvement over the lumped
mechanisms then available. The results presented by Hecht et al.
(1974a) showed reasonable agreement with smog chamber data. In addition,
by combining the numerical sensitivity and estimated uncertainty of
each reaction and rate constant in the H-S-D mechanism, Hecht et al.
(1974b) were able to determine which reactions and rate constants pro-
duced the greatest uncertainties in predictions and were thus most
in need of further study. Some problems were encountered, however, in
attempting to apply the mechanism to situations other than smog chamber
experiments and to apply it outside the range of concentrations and
hydrocarbon mixes upon which it was based (Demerjian et al., 1974).
These problems stemmed largely from the use of parameters that
were based on smog chamber data, not on the fundamental chemistry.
Incorrect fundamental chemistry and chamber-dependent phenomena could
be compensated for or masked by these parameters. No single set of
parameters would fit all smog systems, and there was little theoretical
-------
248
guidance for adjusting the parameters for systems for which no experi-
mental data existed.
Meanwhile, technological progress in smog chamber construction
and analytical capabilities was making it possible to obtain more detailed
and more reliable data. At the same time, the application of advanced
computer hardware and software was reducing both the time and cost of
computer modeling. These developments made it possible, at least for
simple smog chamber experiments (one, two, or three hydrocarbons pre-
sent initially), to trace many of the major intermediate oxidation
products of carbon and hydrogen atoms from the initial hydrocarbon
to carbon dioxide and water. The combination of the need for a better
mechanism and the availability of more reliable and detailed data led
to a new phase in SAI's work—the development of explicit kinetic
mechanisms.
Whereas a lumped mechanism treats hydrocarbons in groups or "lumps,"
an explicit mechanism attempts to account for the fundamental or ele-
mentary reactions of every atom. For example, the explicit mechanism
for propylene and NO in a smog chamber contains over 60 reactions
X
involving 35 species. Explicit mechanisms are based on studies of the
chemical kinetics of individual reactions. The results of many of
these studies have been compiled and evaluated [e.g., Hampson and
Garvin (1975); Demerjian et al. (1974)]. When data for a particular
elementary reaction are insufficient, its rate constant and products
can often be estimated from analogous reactions using thermodynamic
principles (Benson, 1975). Because explicit mechanisms are based on
the fundamental chemistry, a poor fit between predictions and measure-
ments for a given species can sometimes be traced to uncertainties in
chemical reactions or inaccuracies in smog chamber experiments. For
example, poor fits between predictions and measurements for some
propylene/NO experiments in the evacuable smog chamber at the Univer-
A.
sity of California at Riverside (IICR) led Durbin et al. (1975) to
hypothesize that the intensity of light from the UV source in the chamber
was decreasing more rapidly at short wavelengths than at long
-------
249
wavelengths. Subsequent measurements on replacement light sources at
UCR were consistent with this hypothesis.
In the early stages of SAI's development of explicit mechanisms,
smog chamber data could be fitted more closely with the parameterized
H-S-D mechanism than with the explicit mechanisms. As knowledge of
smog chemistry increased, this situation was reversed. At present,
the explicit mechanisms predict smog chamber data better than the
H-S-D mechenism, and without any adjustment of parameters they fit a
much wider range of concentrations than does the H-S-D mechanism. They
provide more detailed insight into the smog formation process. Because
they are not as empirical, there is a theoretical justification for
applying them outside the range of concentrations and hydrocarbon
mixes used in smog chamber experiments.
One might question the usefulness of explicit mechanisms in a
regional model. Since the mechanism for propylene and NO alone con-
X
tains over 60 reactions, surely an explicit mechanism for urban smog
would contain too many reactions to be of practical use. A condensed
version of the explicit mechanisms would combine the advantages of a
basis in elementary chemical reactions and speed of computation. Some
months ago, SAI began development of a new condensed mechanism, the
Carbon-Bond mechanism.
Like the H-S-D mechanism, the Carbon-Bond mechanism employs
lumped chemical species, but there are important differences. The
H-S-D mechanism treats groups of molecules (e.g., aldehydes), whereas
the Carbon-Bond mechanism treats groups of similarly bonded carbon
atoms (e.g., carbonyl carbons). The Carbon-Bond mechanism is far easier
to use because it was designed to accept the types of data most commonly
reported. But the most important difference is the scientific basis:
whereas the H-S-D mechanism was of necessity somewhat empirical, the
Carbon-Bond mechanism is derived from the fundamental chemical reac-
tions in smog as representd by explicit mechanisms. Thus the Carbon-
Bond mechanism is a condensation of our understanding rather than a
parameterization of our uncertainty.
-------
250
Although the Carbon-Bond mechanism was developed recently, it has
been applied to some smog chamber experiments. Preliminary results
(Whitten and Hogo, 1977) indicate that it is far more accurate than
the H-S-D mechanism. Unquestionably, it is easier to apply, since it
does not require the difficult task of estimating average carbon numbers
and stoichiometric coefficients. Thus, the Carbon-Bond mechanism is
the mechanism of choice in the analysis and simulation of Denver air
quality. Information th.at follows indicates that its performance is
excellent.
5. PARAMETERIZATION OF THE SAI DENVER MODEL
Tables A-4 and A-5 show the parameters used in the Denver Model.
The form of the Carbon-Bond mechanism and the chemical reaction rates
used are given in Table A-6. Of these parameters, the least sensitive
are those for surface roughness and deposition velocities; the maximum
variability that one would expect in these quantities would have little
effect on predictions on the time scale of a day's simulation.
Similarly, while the N02 photolysis rate is an important model
parameter, we are confident of our estimates of its magnitude (to
within 10 percent). Thus we do not feel this to be a major source of
model uncertainty.
To the extent that pollutants leave the modeling region and are
then advected back into it later in the day, the assumed boundary condi-
tions are in error, since the concentrations at the boundaries are held
at background levels. However, this reentry only occurs in late after-
noon, if it occurs at all, and so it has little effect on the predicted
oxidant peak. Errors in the assumed background concentrations can be
expected to have a less than 1 pphm effect on oxidant predictions.
The sensitivity of the model to changes in meteorological and
total emissions factors is described elsewhere in this report.
-------
251
TABLE A-4. NON-TIME-DEPENDENT PARAMETERS USED
IN THE DENVER MODEL
Boundary Conditions
NOX* 3 ppb H202 0.01 ppb
03 2 pphm HN02t 0.1 ppb max.
Olefins 0.4 ppb
Paraffins 35 ppbC
Aldehydes 1 ppb
Aromatics and ethylene 0.8 ppb
CO 0.1 ppm
Surface Roughness
0.5 meters over entire modeling region
Deposition Velocities
All set to zero
Emissions Splits by Weight
Olefins 0.031
Paraffins 0.65
Aldehydes 0.05
Aromatics 0.205
All else unreactive.
* NO and N02 vary with sunlight; 03 remains essentially
unchanged; NO/NOX ratio varies from 0 to 0.6.
t HN02 varies according to steady-state relationship.
-------
252
TABLE A-5. TIME-DEPENDENT PARAMETERS USED IN
THE DENVER MODEL
(Midsummer)
Hour of
Day
0000
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
2400
k^mirf1)*
0
0
0
0
0
0
0.098
0.22
0.36
0.456
0.53
0.576
0.59
0.58
0.546
0.47
0.37
0.25
0.128
0
0
0
0
0
0
Exposure Class
-1
-1
-1
-1
-1
-1
0
1
1
2
2
3
3
3
3
2
2
1
-1
-1
-1
-1
-1
-1
-1
* See Table A-7 and Figure A-3 (k-, increased by 10% for
Denver altitude correction)
+ -1 = night
0 = twilight
1^3 = low, medium, and strong sunlight.
-------
TABLE A-6. THE CARBON-BOND MECHANISM USED IN THE DENVER MODEL
(a) Reactions
No. Reaction
1 N02 4 hv -»• NO 4 0
2 0 4 02 (+M) -» 03 (4M)
3 03 4 NO - N02 4 Q2
4 0 4 NO, «• NO 4 0,
d. C
5 0^ 4 NO,, -» NO, 4 Q2
6 N03 4 NO - N02 4 N02
7 N03 4 N02 4 H20 ->• 2HN03
8 HO;, 4 NO, -- HNO,
22 2
9 N02 4 OH- H- HN03
10 HN02 4 hv -* NO 4 OH-
11 NO 4 OH- H- HN02
°2
12 CO 4 OH- •* C02 4 H02
13 HO 2 4 NO - OH- 4 N02
No.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
H0^4
PAN -*
H2°2
OLE 4
OLE 4
OLE 4
PAR 4
PAR 4
HCHO
HCHO
ARO 4
ARO 4
ARO 4
ARD 4
Reaction No. Reaction
H02- - H202 4 02 28 CH?02 4 NO - N02 4 HCHO 4 H0?
HC(0)02 4 N02 29 HC(0)02 4 NO -> N02 4 C02 4 H02
4 hv - OH- 4 OH- 30 HC(0)02 + N02 •+ PAN
°2
OH- -> HCHO 4 CH.OA 31 CH,Oi 4 HOi -> CH,OOH + 0,
32 3223 2
20?
0 — ^ HC(0)02 4 CH302 32 HC(0)02 4 H02 •* HC(0)OOH 4 02
ro
0, en
034 HC(0)0^ + HCHO 4 OH- <^>
°2
OH- 4 CH3Oj 4 HO
0 -^ CH-.0- 4 OH-
3 2
°2
4 OH- 4 HC(0)02 4 H20
202
4 hv — >• uHC(0)Oj 4 aHO^
02
OH--i HCHO 4 CH Oj
20-
0 -^ HC(0)02 4 CH302
°2
034 HC(0)02 4 HCHO 4 OH-
NO, ' Products
-------
TABLE A-6 (Concluded)
(b) Rate Constants*
Reaction
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
Rate Constant
k/
2.08 x 10"5
2.52 x 101
1.34 x 104
5 x ID'2
1.3 x 104
2.0 x 10"3
3.0 x 101
9.0 x 103
1.9 x 10"1k1
9.0 x 103
2.06 x 102
2.0 x 103
Reaction
Number
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Reaction
Rate Constant Number
4.0 x 103 28
2.0 x 10"2 29
7.0 x 10-4k] 30
3.8 x 104 31
5.3 x 103 32
1.5 x 10'2
1.3 x 103
2.0 x 101
2.0 x 104
6.0 x 10*3k]
8.0 x 103
3.7 x 101
2.0 x 10"3
5.0 x 101
Rate Constant
2.0 x 103
2.0 x 103
1.5 x 102
4.0 x 103
4.0 x 103
* In units of ppm min~ except for photolysis reactions and Reaction 15 (units of min~ )
and Reaction 7 (units of ppm~2min~').
The rate constant for N02 photolysis, k-, , depends on the UV i
Model k-| was determined by using the solar zenith angles in
ntensity. In the Denver
Table A-7 and the values
of k-, in Figure A-3.
-------
255
o
4J
o
JZ
0.
0.2 -
0.1 -
10
20
30 40 50 60
Solar Zenith Angle (degrees)
70
80
90
Source; Killus et al. (1977)
FIGURE A-3. ESTIMATED GROUND-LEVEL N02 PHOTOLYSIS RATE CONSTANT
AS A FUNCTION OF SOLAR ZENITH ANGLE
-------
256
TABLE A-7. SOLAR ZENITH ANGLES FOR DENVER
(39.75° latitude, 105° longitude)
29 July
Hour
500.00
530.00
600.00
630.00
700.00
730.00
800.00
830.00
900.00
930.00
1000.00
1030.00
1100.00
1130.00
1200.00
1230.00
1300.00
1330.00
1400.00
1430.00
1500.00
1530.00
1600.00
1630.00
1700.00
1730.00
1800.00
1830.00
1900.00
1930.00
2000.00
Angle
110.14
104.46
98.88
93.43
88.15
83.09
78.29
73.82
69.75
66.16
63.14
60.77
59.14
58.30
59.14
60.77
63.14
66.17
69.76
73.83
78.29
83.09
88.16
93.44
98.89
104.48
110.16
115.90
121.66
127.42
133.11
15 November
Hour
500.00
530.00
600.00
630.00
700.00
730.00
800.00
830.00
900.00
930.00
1000.00
1030.00
1100.00
1130.00
1200.00
1230.00
1300.00
1330.00
1400.00
1430.00
1500.00
1530.00
1600.00
1630.00
1700.00
1730.00
Angle
90.08
84.76
79.28
73.68
67.98
62.24
56.48
50.74
45.08
39.57
34.32
29.50
25.38
22.38
21.01
21.57
23.93
27.63
32.20
37.29
42.71
48.32
54.04
59.80
65.56
71.29
-------
257
We are thus left with those factors pertaining to photochemical
reactivity: splits of hydrocarbon emissions among classes and chem-
ical rate constants. Ideally, these parameters would be directly
available from data or theoretical considerations. Some of the chem-
ical rate constants are in fact obtained this way. However, many of
these factors including several of critical importance, are, of neces-
sity, estimates. Thus, any simulation exercise that agrees well with
observations in its initial trials must be regarded as containing
significant amounts of luck. This is very nearly what happened in
this study.
Figures A-4, A-5, and A-6 represent results of a trajectory
analysis of three air parcels which passed over Department of Health
air quality stations 2, 3, and 4 at 12 noon, and 10 and 8 a.m. respec-
tively on 29 July 1975 (see Table A-8). The two sets of simulations
represent aldehyde photolysis rates of 0.005 x k-, and 0.006 x k-,.
TABLE A-8. OBSERVATIONS AT COLORADO DEPT. OF HEALTH MONITORING
STATIONS AND TRAJECTORY MODEL PREDICTIONS FOR OZONE
Station 2 Station 3 Station 4
Observations
Predictions 1
2
The aldehyde photolysis rate is the main adjustable reactivity parameter
because it is only known to within about 50 percent accuracy and
because it varies naturally on the order of 20 percent.
On the basis of these six simulations we chose the aldehyde photo-
lysis rate to be 0.006 x k-j. This is the only parameter adjustment
made in the kinetic mechanism in this project. The exceptionally
good agreement between station observations and predictions must be
viewed as a combination of good science and luck.
8 pphm
4 pphm
4.5 pphm
10 pphm
10 pphm
12 pphm
11 pphm
8 pphm
9 pphm
-------
C
0
N
C
E
N
T
R
A 0.30
T
1
0
N
O.lb
ro
en
CO
O.OJ »-
00 U 00
U U U U
000
»
600
75
ISO
225
300 375
MME (MINUTES)
525
FIGURE . DOT TtST
SPECIkS Oi
CONCENTHAT1UN SCALt FACTOR
(a) 0.005 Aldehyde Factor
FIGURE A-4. TRAJECTORY ANALYSIS FOR HEALTH STATION 2
-------
O.bOt
o.v;
c
n
N
c
c
N
T
R
A 0.30
T
I
0
ro
en
0.15
CJ 0
00 00
o n
oo ooo
75
150
225 30J 375
TIME (MINUTESI
525
600
FIGURE . DOT TrST . SPFCIES 03
CONCENTRATION SCALE FACTOR 1
(b) 0.006 Aldehyde Factor
FIGURE A-4 (Concluded)
-------
0.60*
0.45
C
0
N
C
e
N
T
R
A 0.30
T
I
n
N
0.15
ro
cr>
o
0.00+-
oo oo ooo
75
150
J25
3JO 375
TIME (MINUTES!
525
FICURt . DOT TEST
SPTCIES 03
CONCENTRATION SCALE FACTOR
600
(a) 0.005 Aldehyde Factor
FIGURE A-5. TRAJECTORY ANALYSIS FOR HEALTH STATION 3
-------
O.bOl
0.45
C
0
N
C
t
N
T
P
A 0.30
T
I
C
N
ro
O1
0.15
0 U
o.oo*-
0 U 0 00 QUO 00
75
FIGURE
150
001 TEST
22'j
300
TIME (MINUTtSI
60U
.REGIES C3
CONCtNHAriL'N SCALE F4CTUR
(b) 0.006 Aldehyde Factor
FIGURE A-5 (Concluded)
-------
0.60*
0.45
0.30
CTi
ro
0.00*-
Q
0
75
0 C C
150
FIGURE . DOT TEST
225
SPECIES 03
300 375
TIME (MINUTES!
450
525
600
CONCENTRATION SCALE FACTOR
(a) 0.005 Aldehyde Factor
FIGURE A-6. TRAJECTORY ANALYSIS FOR HEALTH STATION 4
-------
O.b J>
c
0
N
C
E
N
T
R
A
T
I
0
N
F
P
H
U.30
ro
cri
oo
0 00
o.oo*-
75
C 0 0
i —
150
FIGURE . OUT TEST
225
300 375
TIME (MINUTES!
525
iPECItS 03
CONCENTRATION SCALE FACTOR
*
600
(b) 0.006 Aldehyde Factor
FIGURE A-6 (Concluded)
-------
264
6. THE CLIMATOLOGICAL DISPERSION MODEL (COM)
The COM is used for calculating long term, e.g., seasonal or
annual, average concentrations for quasi-stable pollutant species.
Input to the model includes a detailed specification of the magnitude
and distribution of pollutant emissions from point and area sources,
together with the frequency of occurrence of various meteorological
conditions for the time span under consideration.
The basis for computing the dispersion of pollutants in COM
is that the steady-state distribution of concentrations for a homo-
geneous atmosphere from a continuously emitting source is given by a
Gaussian plume formula. The pollutant concentrations at each recep-
tor are calculated by adding contributions of pollutants from each
upwind source (within a —- radian sector). This summation is taken
1 6
over each combination of wind speed, direction, and atmospheric
stability class, weighted by the estimated probability of occurrence
of that combination. These probability functions are derived from
historical meteorological records.
The major strong points of the COM are;
> It is relatively inexpensive to run.
> It includes a good description of the emissions inventory.
Major point sources and their characteristics can be identi-
fied and included, while dispersed sources can also be accounted
for.
> It includes a basic description of the climatology of the
modeling region—enough to characterize average meteorology
for the time span being considered. Since COM considers the
distribution of wind speeds and directions over a period, on
the average it does reproduce correctly the dilution of
emissions due to wind. It also takes average mixing heights.
into consideration.
-------
265
> The COM allows data on pollutants that are subject to annual
air quality standards to be assessed. Models that calculate
one-hour-average concentrations of pollutants for a single
study day cannot easily be used to predict annual averages.
The major assumptions included in the COM are:
(1) As pollutants are carried downwind from their sources, thev
^ n
spread horizontally across sectors (— radians) and disperse
vertically according to a Gaussian plume formula. This formula
is modified to account for the existence of a mixing layer
with finite depth. For receptors at large distances downwind
from the source a uniform vertical distribution of concen-
trations is assumed. For short distances the Gaussian formula
is used, and for intermediate distances an interpolation formula
is provided.
(2) Meteorological conditions can be described by a distribution
function of wind speed, direction, and atmospheric stability
class. This function is usually derived from records kept at a
local airport and is assumed to be invariant over the modeling
region. Allowance is made for vertical variation of wind
velocity by a power law relationship that is a function of
stability class.
(3) The effective height of area sources is constant over the model-
ing region. All sources that cannot be identified individually
as point sources are gridded into area sources.
(4) Source strengths do not vary with time, or if they do, such
variation is not correlated with meteorological variables.
This assumption conforms to the characteristics of most
emissions inventories, in which long-term temporal variations
of source strengths are unaccounted for However, in reality,
such temporal variations are common. For example, there
are daily and seasonal patterns in power demand, which in turn
affect power station emissions. Use of the formulas in the
COM, with its steady-state assumption, requires that emissions
-------
266
are evenly spread among the possible combinations of meteoro-
logical conditions.
(5) No material is lost from the plume to the ground and there is
no gravitational settling within the plume.
(6) The diffusion coefficents and depth of the mixing layer may
be parameterized in terms of atomspheric stability classes.
(7) Pollutants are carried in straight lines at constant speed
from source to receptor. This treatment does not account
for air parcel trajectories changing direction when the wind
shifts, nor for spatial variations due to topography or local
heating.
(8) Only sources within the modeling region contribute to observed
pollution levels. This means that pollutants blown into the
region from outside sources are not accounted for.
Before interpreting the results from the COM, it should be noted
that the NOo concentrations obtained from the COM are calculated consid-
ering all NO emissions as being N09, even though at their sources they
X L.
are predominantly NO. This treatment correctly accounts for the eventual
conversion of NO to NOp in the atmosphere. Because the conversion is
not complete except in the presence of sufficient ozone (which is not
available during the early morning, when a large amount of the daily
NO is emitted), the pollutant plumes will have a high N0/N09 ratio
X C.
for long travel distance during morning commuter rush hours. Therefore
the use of NOX as a surrogate for N02 will result in serious over-
predictions of N02 concentrations, particularly near the sources.
Farther downwind, where the plumes are more diffuse, the NO will be
mostly N02- Thus, for a given receptor, without calibration the COM
will overpredict N02 concentrations in plumes from nearby sources and
predict more correctly for sources at some larger distance. The result
will be a net overprediction of N02 concentrations over the entire
modeling region but the peak concentrations, being most influenced by
nearby sources, will exhibit the greatest over-prediction.
-------
267
The particulate concentrations predicted by the COM are based,
as was pointed out above, on the assumption that there is no gravita-
tional settling in the plume or loss of material from the plume to
the ground. This assumption seems inconsistent with the expected
large particle size of some types of emissions, for example, those
from street sanding.
A further potential problem in interpreting the results from the
COM arises because it is a calibrated model. By this we mean that a
linear transformation is made to the calculated concentrations to
obtain the predicted ones. The experimental data used to obtain the
transformation are taken from monitors that are often placed at points
where large concentrations of pollutants are expected, such as beside
a highway. Since the concentrations at these points are much higher
than the average concentrations over the area representative of the
spatial resolution of the model (on the order of a square mile), an
upward bias is imparted to all of the predicted concentrations reported.
Calibration cannot remove difficulties in interpreting NO emissions
as N0?. NO air quality data are available for COM calibration only
£• /\
at the downtown CAMP station. Thus calibration might approximately
correct the COM overprediction at this site, but would then apply
this scaled reduction at remote sites where substantial conversion
has occurred. Most importantly, the calibration derived for one year
cannot be valid for subsequent years if the mix of photochemical
precursors change. Earlier in this report, Denver Model results
are described that show decreases of NOo when NO emissions are increased.
7. REFERENCES FOR APPENDIX A
Benson, S. W. (1975), "Current Status of Methods for the Estimation of Rate
Parameters," Int. J. Chem. Kinetics Symposium No. 1, pp. 359-378.
Calvert, J. G. (1976a), "Test of Theory of Ozone Generation in Los Angeles
Atmosphere," Environ. Sci. Technol., Vol. 10, No. 3, March 1976.
(1976b), "Hydrocarbon Involvement in Photochemical Smog Forma-
"in Los Angeles Atmosphere," Environ. Sci. Technol., Vol. 10,
tion
p. 256.
-------
268
Calvert, J. G., and R. D. McQuigg (1975), "The Computer Simulation of the
Rates and Mechanisms of Photochemical Smog Formation," Int. J. Chem.
Kinet. Symposium No. 1, pp. 113-154.
Chameides, W. L., and J.C.G. Walker (1976), "A Time-Dependent Photochemical
Model for Ozone Near the Ground," J. Geophys. Res., Vol. 81, No. 3,
pp. 413-420.
Chan, W. H., et al. (1976), "Kinetic Study of HONO Formation and Decay
Reactions in Gaseous Mixtures of HONO, NO, N02> H20, and N2>"
Environ. Sci. Techno!., Vol. 10, pp. 674-682.
Colorado Division of Highways [CDH] (1976), "JRPP [Joint Regional Planning
Program] Air Quality Assessment Statement (Calendar Year 1976),"
Denver, Colorado.
Demerjian, K. L., J. A. Kerr, and J. G. Calvert (1974), "The Mechanism of
Photochemical Smog Formation," in Advances in Environmental Science and
Technology, J. N. Pitts and R. L. Metcalf, eds. (John Wiley & Sons,
New York, New York).
Durbin, P. A., T. A. Hecht, and G. Z. Whitten (1975), "Mathematical Modeling
of Simulated Photochemical Smog," EPA-650/4-75-026, Systems Applications,
Incorporated, San Rafael, California.
Eschenroeder, A. Q., and J. R. Martinez, (1972), Advan. Chem., Vol. 113, p. 101
Hampson, R. F., and D. Garvin (1975), "Chemical Kinetic and Photochemical
Data for Modelling Atmospheric Chemistry," NBS Technical Note 866,
National Bureau of Standards, Washington, D.C.
Hecht, T. A. (1972), "Further Validation of a Generalized Mechanism Suitable
for Describing Atmospheric Photochemical Dynamics," Report 72-SAI-26,
Systems Applications, Incorporated, San Rafael, California.
Hecht, T. A., and J. H. Seinfeld (1972), "Development and Validation of a
Generalized Mechanism for Photochemical Smog," Environ. Sci. Technol.,
Vol. 6, p. 47.
Hecht, T. A., P. M. Roth, and J. H. Seinfeld (1973), "Mathematical
Simulation of Atmospheric Photochemical Reactions: Model Develop-
ment, Validation, and Application," R73-28, Systems Applications,
Incorporated, San Rafael, California.
Hecht, T. A., J. H. Seinfeld, and M. C. Dodge (1974a), "Further Develop-
ment of a Generalized Kinetic Mechanism for Photochemical Smog,"
Environ. Sci. Techno!., Vol. 8, pp. 327-339.
Hecht, T. A., M. K. Liu, and D. C. Whitney (!974b), "Mathematical Simula-
tion of Smog Chamber Photochemical Experiments," EPA-68-02-0580,
Environmental Protection Agency, Research Triangle Park, North
Carolina.
-------
269
Killus, J. P., et al. (1977) "Continued Research in Mesoscale Air Pollution
Simulation Modeling: Volume V--Refinements in Numerical Analysis,
Transport, Chemistry, and Pollutant Removal," EF77-142, Draft Final
Report, Systems Applications, Incorporated, San Rafael, California.
Lamb, R. G., et al. (1975), "Numerico-Empirical Analyses of Atmospheric
Diffusion Theories," J. Atmos. Sci., Vol. 32, pp. 1794-1807.
Leighton, P. A. (1961), Photochemistry of Air Pollution (Academic Press,
New York, New York).
Liu, M. K., et al. (1976a), "The Chemistry, Dispersion, and Transport of
Air Pollutants Emitted from Fossil Fuel Power Plants in California:
Data Analysis and Emissions Impacts Model," ER76-18, Systems
Applications, Incorporated, San Rafael, California.
Liu, M. K., et al. (1976b), "Development of a Methodology for the Design
of a Carbon Monoxide Monitoring Network," report under Contract 68-03-2399
to the Environmental Protection Agency, Las Vegas, Nevada, by Systems
Applications, Incorporated, San Rafael, California.
Nicolet, M. (1975), "Stratospheric Ozone: An Introduction to Its Study,"
Rev. Geophys. and Space Phys., Vol. 13, No. 5, pp. 593-636.
Niki, H., E. E. Daby, and B. Weinstock (1972), Photochemical Smog and
Ozone Reactions, Advances in Chemistry Series 113 (American Chemical
Society, Washington, D.C.).
Reynolds, S. D. (1973), "Further Development and Validation of a Simulation
Model for Estimating Ground Level Concentrations of Photochemical
Pollutants," Volume II, "User's Guide and Description of the Computer
Programs," R73-18, Systems Applications, Incorporated, San Rafael,
California.
Reynolds, S. D., et al. (1977), "Continued Research in Mesoscale Air
Pollution Simulation Modeling," Vol. V, EF76-142, Systems Applications,
Incorporated, San Rafael, California.
(1976), "Continued Research in Mesoscale Air Pollution Simula-
tion Modeling: Volume II--Refinements in the Treatment of Chemistry,
Meteorology, and Numerical Integration Procedures," EPA-600/4-76-016b,
Environmental Protection Agency, Research Triangle Park, North
Carolina.
Ripperton, L. A., et al. (1975), "Urban Photochemistry and Rural Transport,"
presented at meeting of Mid-Atlantic States Section, Air Pollution
Control Association, Newark, New Jersey (16 October).
Seinfeld, J. H., T. A. Hecht, and P. M. Roth (1973), "Existing Needs in
the Experimental and Observational Study of Atmospheric Chemical
Reactions: A Recommendations Report," EPA-R4-73-031, Systems
Applications, Incorporated, San Rafael, California.
-------
270
Smith, T. B., et al. (1976), "Analysis of Data from the Three-Dimensional
Gradient Study," MRI 75FR-1395 and SAI EF75-84, Meteorology Research,
Incorporated, Altadena, California, and Systems Applications, Incor-
porated, San Rafael, California.
Wayne, L. G., M. Weisburd, R. Danchick, and A. Kokin (1971), "Final Report-
Development of a Simulation Model for Estimating Ground Level Concen-
trations of Photochemical Pollutants," System Development Corp., Santa
Monica, California.
Whitten, G. Z., and H. Hogo (1977), "Mathematical Modeling of Simulated
Photochemical Smog," EPA-600/3-77-011, Environmental Protection Agency,
Research Triangle Park, North Carolina.
-------
271
APPENDIX B
EMISSIONS ANALYSIS
-------
272
APPENDIX B
EMISSIONS ANALYSIS
by Pravin V. Mundkur
For the purpose of evaluating and modeling the present and future
air quality in the Denver metropolitan region, SAI obtained a number of
emissions files of estimated present and future emissions of various
air pollutants including carbon monoxide, particulates, oxides of nitro-
gen, and hydrocarbons. These emissions data were compiled primarily
by the staff of the Air Pollution Control Division of the Colorado
Department of Health and by the Colorado Division of Highways. Various
other groups interacted with the above two state departments in this
estimation process; they include the Environmental Protection Agency
(EPA), the Denver Regional Council of Governments (DRCOG), and the
Regional Transportation District (RTD).
SAI used the emissions data files primarily as inputs for the
Denver Model and the Climatological Dispersion Model (COM). However,
substantial information can be gained from these files without the use
of any models. SAI wrote separate computer programs to bring out the
following information:
> Year-to-year variations in overall pollutant loading levels.
> Year-to-year variations in pollutant loading from subregions
of the entire region selected for the modeling effort. The
subregions are the same as those chosen for the sensitivity
study.
> Relative contributions of various source categories to the
overall pollutant loading.
Such an analysis of these emissions files provides a preliminary,
although crude, assessment of the trends in air quality that is free
-------
273
from the limitations of the models used. Further, the study of the
emissions from the various subregions is useful in assessing their
relative impacts on air quality (as in the sensitivity study). Finally,
the identification of the relative contributions of various source cate-
gories to the overall pollutant burden should be of great help in plan-
ning the most effective mitigation strategies. For these reasons,
this appendix has been prepared to describe the results of SAI's
in-depth analysis of the air pollutant emissions in the Denver metro-
politan region.
This presentation is organized into a number of sections. In the
first we summarize the present and potential emissions problem in the
Denver metropolitan region. In Section 2 we take a brief look at fed-
eral and state emissions regulations as they apply to the emissions from
highway and other traffic in this region. Section 3 presents an over-
view of the methodology used to arrive at the emissions estimates.
Section 4 presents an analysis of the emissions files used as inputs
to COM and Section 5 presents an analysis of the files input to the
Denver Model.
1. OVERVIEW OF PRESENT EMISSIONS IN THE DENVER REGION
The air pollutants that constitute the main problem in the Denver
metropolitan region at present are carbon monoxide, particulates, reac-
tive and unreactive hydrocarbons, and nitrogen oxides. Traffic, both
highway and nonhighway, accounts for a major fraction of the emissions
of pollutants: roughly 80 percent of the hydrocarbons, 30 to 40
percent of the nitrogen oxides, and about 90 percent of the carbon
monoxide. While automobiles and other vehicles do not directly contri-
bute a high proportion of the particulate emissions in the Denver
metropolitan region, street sanding operations, which are related to
the operation of these vehicles, contribute from 45 to 60 percent of
the total particulate emissions.
Next to automobiles and other traffic, point sources are the most
significant contributors to total emissions. Emissions from large
-------
274
point sources account for as. much as 50 percent of the total emissions
of nitrogen oxides.
As may be expected, a high proportion of the total emissions of
any pollutant occurs in the Central Denver (Metro) subregion. The
emissions flux is from 30 to 100 percent higher here than in any other
subregion. A suprisingly high proportion (more than 50 percent) of
the nitrogen oxides emissions lies outside any of the seven subregions
delineated for the sensitivity study. An apparent reason for this is
that more than 90 percent of the emissions from large point sources lie
outside these subregions and large point sources account for a major
portion (about 50 percent) of the emissions of nitrogen oxides.
By the year 2000, the emissions of particulates are projected
to rise to roughly 180 percent of their 1974 level and the nitrogen
oxides emissions to about 150 percent of that level, if events proceed
in accordance with the JRPP plan. In contrast, both carbon monoxide
and hydrocarbon emissions are expected to fall to roughly 55 percent
of their 1974 levels by the year 2000. Over this same period, the
emissions appear to rise most rapidly in the Jeffco-Urban and South
Metro subregions, where the most rapid development is projected to
take place.
2. EMISSIONS STANDARDS AND REGULATIONS APPLICABLE TO HIGHWAY AND
OTHER TRAFFIC
The Federal Motor Vehicles Control Program, which is being imple-
mented as a consequence of the Clean Air Act of 1970, requires a phased
reduction in automotive emissions of hydrocarbons, carbon monoxide,
and oxides of nitrogen to specified fractions of the 1971 emission levels.
The original deadlines for achieving the ultimate lowered emission rates
were 1975 for hydrocarbons and carbon monoxide and 1976 for oxides of
nitrogen. Difficulties associated with meeting these deadlines have
resulted in two one-year postponements by the Administrator of the
Environmental Protection Agency and a one-year delay from Congressional
action. Thus the present deadline for meeting the original statutory
-------
275
standards is 1978. Table 8-1 summarizes the allowable emissions rates
from automobiles manufactured in different years according to the
current Federal Motor Vehicle Control Program. Figure B-l is a grapni-
cal presentation of these Federal standards for cars and light-duty
trucks.
The emissions factors used in this study were based on the 1978
deadline for the Federal Motor Vehicles Control Program. Pending
amendments to the Clean Air Act could affect the estimates of emission
factors, and hence could alter the results of this study. For instance,
further delays in the deadline by a year or two would probably result
in higher emissions in 1980 and 1985 than estimated here, but emissions
in 1990 and 2000 would be largely unaffected since emissions are expected
to be stabilized by then. Any amendments that permanently alter the
statutory emissions standards would certainly affect the emissions
estimates for future years.
One of the pending Clean Air amendments would relax the original
nitrogen oxides standard of 0.4 grams per mile to 1.0 grams per mile.
It is possible to estimate roughly the impact of such a relaxation on
total nitrogen oxides emissions in the year 2000. In that year the
estimated NO emissions are roughly 270 tons per day (as N0?). Of
J\ *—
these emissions about 120 tons per day (as NC^) are due to automobiles.
If the auto emissions in the year 2000 are stabilized and the emissions
standard is relaxed from 0.4 to 1.0 grams per mile, the auto emissions
may be expected to be larger by a factor of 1.0/0.4, or 300 tons per day
rather than 120. Including the estimated 150 tons of NOX per day from
nonauto sources in the year 2000, the total estimated emissions are 450
tons per day, which is an increase of 67 percent.
3. METHODOLOGIES USED FOR ESTIMATION AND PROJECTION OF EMISSIONS
A comprehensive discussion of the methodology used by the Colorado
Department of Health is contained in the Denver AQMA report for the
years 1974, 1980, and 1985. A discussion of the methodology used by
-------
276
TABLE B-l. FEDERAL AND CALIFORNIA NEW VEHICLE EMISSIONS STANDARDS
(a) Passenger Cars and Light-Duty Trucks
YEAR
Prior to controls
1966-1967
1968-1969
1970
1971
1972
1973
1974
1975 **pc
**PC
**LDT
**LDT
1976 **PC
**PC
**LDT
**LDT
977f **PC and LOT
STANDARD
Calif.
Calif, s
Federal
Calif. &
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif.
Federal
Calif
COLD START
TEST
7-raode
7-mode
7-roode
7-mode
50-100 CID
101-140 CID
over-140 CID
7.-mode
7-mode
7-mode
7-mode or
CVS-1
CVS-1
CVS-1
CVS-1
CVS-1
CVS-1
CVS-2
CVS-2
CVS-2
CVS-2
CVS-2
CVS-2
CVS-2
CVS-2
CVS-2
HYDROCARBONS
850 ppra
(11 gin/mi)
275 ppcn
410 ppm
350 ppm
275 ppm
2.2 gin/mi
2.2 gm/mi
2 . 2 gm/mi
1 . 5 gm/m i
3 . 2 gm/mi
3 . 4 gm/mi
3 . 2 gm/mi
3 . 4 gm/mi
3.2 gm/mi
3,4 gm/mi
0.9 gm/mi
1.5 gm/mi
2.0 gm/mi
2.0 gm/mi
0. 9 gm/mi
1.5 gn/mi
0.9 gm/mi
2 . 0 gm/mi
0.41 gm/mi
CARBON
MONOXIDE
3.4%
(80 gm/mi)
1.5%
2.3%
2.0%
1.5%
23 gm/mi
23 gm/mi
23 gm/mi
23 gm/mi
39 gm/mi
39 gm/mi
39 gm/mi
39 gm/mi
39 gm/mi
39 gm/mi
9 gm/mi
15 gm/mi
20 gm/mi
20 gm/mi
9 gra/mi
15 gm/mi
17 gm/mi
20 gm/mi
9 gra/mi
OXIDES OF
NITROGEN
1000 ppm
(4 gm/mi)
no std.
no std.
no std .
no std .
no std.
4 gm/mi
-
3 gm/mi
*3.2 gm/mi
-
3 gm/mi
3 gm/mi
2 gm/mi
3 gm/mi
2.0 gm/mi
3 . 1 gm/mi
2.0 gm/mi
3.1 gm/mi
2 . 0 gm/mi
3 . 1 gm/mi
2 . 0 gm/mi
3 . 1 gm/mi
1 . 5 qm/mi
Federal CVS-2 1.5 gin/mi 15 /mi 2.0 gm/mi
1978
**PC and LOT Federal
CVS-2
0.41 gm/mi
3.4 gm/mi
* 7-Mode Hot Start
**PC-Passenger Cars LDT-Light Duty Trucks
t The 1977 NOx standards have not been officially adopted at this time
0. 4 gm/mi
-------
277
TABLE B-l. (Concluded)
(b) Heavy-Duty Vehicles (Over 6,000 Pounds)
YEAR
*1969-1971
1972
1973-1974
1975-1976
1977
STANDARD
State-gasoline
State-gasoline
State-gasoline
& diesel
State-gasoline
& diesel
State-gasoline
S diesel
HYDRO-
CARBONS
275 ppm
100 ppm
CARBON
MONOXIDE
1.5%
1.0%
HC + NOx » 16 gm/BMP
CO - 40 gm/BHP
HC + NOx =10 gm/BHP
CO =30 gm/DHP
HC + NOx " 5 gm/BHP
CO =25 gm/BHP
OXIDES OF
NITROGEN
no std .
no 3td.
hr.
hr.
hr.
hr.
hr.
hr.
gm/BHP hr. grams per brake horsepower-hour
* Federal standards remained at this level through 1973. The Federal
Government adopted standards for heavy-duty gasoline and diesel
vehicles for 1974 and subsequent model years which are identical to
California's 1973-74 standards.
State Smoke Standards for Heavy-Duty Vehicles
1971 and later vehicles may discharge smoke no darker than Ringelraan 1 or 20 percent
opacity for up to 10 seconds.
Vehicles sold before 1971 may discharge smoke no darker than Ringelman 2 or 10 percent
opacity for up to 10 seconds.
Crankcase Emissions
On all new vehicles manufactured for sale in California after January 1, 1964, crank-
case emissions are virtually zero. Coaparable Federal standards became effective in
1968 for light-duty vehicles, and 1970 for heavy-duty vehicles.
Evaporative Emissions
Evaporative emissions of hydrocarbons have been 6 gras/test for light-duty vehicles
since 1970, and 2 gns/tast since 1972. Heavy-duty gasoline-powered vehicles are
2 gms/test, effective 1973.
Source: California Air Resources Board (1974).
-------
grams/mile
-c
Emissions
o o
en
O
E
J
ro p.
(3.1)
(3.0)
(2.0)
No Federal NO Controls
— A
(0.4)
» . i i 1 i 1
CVS-I
1
1,
t 1
,
r\tf T T trto
1 1 1
1966 67 68 69 70 71 72 73 74
Model Year
(a) Oxides of Nitrogen
IX)
co
75 76 77 78 79 80 81
FIGURE B-l. FEDERAL NEW VEHICLE EMISSIONS STANDARDS FOR LIGHT-DUTY VEHICLES
-------
12r-
10
OJ
r—
i
ro
s_
C1:
c
o
00
1/1
c
o
O
O
E
X
(11, 7-Mode)
Jl-li.
TaTf)"
Jlfl
TOT
SH«fr-
PC - PASSENGER CAR
LOT = LIGHT DUTY TRUCK
-•=$••=
_I_.
No Controls
(0.41)
-P-H»-
7-Mode
1 i t^t
LOT
PC
CVS-I
•*!
"E
- CVS-II-
1966 67
68
69 70
71 72 73
Model Year
74
75 76
77
78
79 80
81
(b) Hydrocarbons
FIGURE B-l. (Continued)
-------
80
D
!| 64
E
IV
i.
Ol
l/t
c
0
* 48
I/I
E
UJ
T3
X
o
C 00
0 Ji
i.
c
o
_a
u
«O
I 16
•r-
X
rt)
s:
0
(80, 7 -Mode)
-
—
(39)
(33}
(23)
PC - PASSENGER CAR
LOT = LIGHT DUTY TRUCK
12P1 _[
415) J _j
No Controls
7-Mode
LOT
PC
. cv.sr' .
13.4L. ' ' P " '!" '1
i i
L 1 1
,
i PVSII tiifc *"
1 T . 1 PH*- , ,
CO
O
1966 67 68 69 70 71 72
Model Year
73 74 75 76
77
78
79
80 51
(c) Carbon Monoxide
FIGURE B-l. (Concluded)
-------
281
the Colorado Division of Highways is contained in the Detailed Assess-
ment Report 1-470 for the years 1985 and 2000. The source categories
considered in these estimates were:
Point sources (all pollutants)
Auto emissions (all pollutants)
Gasoline service (hydrocarbons)
Solvent users (hydrocarbons)
Oil-based paint use (all pollutants)
Space heating (all pollutants)
Incinerators (all pollutants)
Airports [all pollutants)
Street sanding (particulates)
Construction (particulates).
Automobile and point source emissions are major emissions sources
in the Denver metropolitan region. Automobiles account for roughly
80 percent of the hydrocarbon emissions, 30 to 40 percent of the nitro-
gen oxides emissions, 90 percent of the carbon monoxide emissions, and
10 to 15 percent of particulate emissions. Point sources account for
roughly 15 to 20 percent of the hydrocarbon emissions, 40 to 50 per-
cent of the nitrogen oxides emissions, and 10 to 15 percent of particu-
late emissions. Particulate emissions are largely from street sanding
(45 to 60 percent) and construction (25 percent).
a. Automobile Emissions
Automobile emissions were estimated by the Colorado Department
of Health and the Division of Highways in consultation with the EPA.
Two separate files are used to estimate total automobile emissions.
The Auto Link file accounts for automobile traffic that crosses traffic
zone boundaries and the Auto Area (or Intrazonal VMT) file accounts
for traffic that remains within traffic zone boundaries. The Link
file alone accounts for more than 99 percent of the total automobile
emissions for any pollutant.
-------
282
The Link file was created by the Colorado Division of Highways
using their transportation models. For the years 1974 and 1975 the
input to these models was derived from traffic count data for the
year 1971. For future years the inputs were derived from the year
2000 JRPP land use plan and the I-470 study. The DRCOG Empiric Activ-
ity Allocation Model was also used. Automobile emissions factors were
estimated using AP-42, Supplement 5. Table B-2 summarizes the assump-
tions that were needed to estimate the emissions factors. Figure B-2
displays the variation, for winter and summer, in average CO emissions
with the speed of operation of the "average" highway vehicle.
Since the emissions of pollutants from automobiles and other traffic
are dependent on the typical cycles of operation, which vary with loca-
tion and type of roadway, the Division of Highways classified the links
in the Denver transportation system under eight roadway types and four
area types. For the years 1985 to 2000, however, only five different
roadway types were used in this classification system. Depending on
whether the operating conditions are peak Crush-hour) or off-peak, the
Division of Highways estimated average vehicle operating speeds for
each of the roadway types within each area type. These estimates are
summarized in Table B-3. Figure B-3 shows the estimated diurnal vari-
ation in traffic flow for the Denver metropolitan region and also
indicates the hours of the day that are considered to be "peak" commute
hours. In the summer, daylight savings time shifts the diurnal pattern
one hour to the left.
Using the link node positions and lengths (inputs to the trans-
portation models), the estimated average daily traffic (output of the
transportation models), the estimated emission factors (as in Figure B-2)
and the speed tables for the various link types (as in Table B-3), the
Colorado Division of Highways estimated the Auto Link emissions as a func-
tion of the time of day and location within the Denver Highway Planning
coordinate system. This Auto Link emissions file was created with the
help of a special program, SAIEMIS, written by that division. The
Auto Link emissions file, which incorporates hourly and seasonal
-------
283
TABLE B-2. 1-470 AIR QUALITY ANALYSIS
70% of vehicles are light-duty gas vehicles.
20% of vehicles are light-duty gas trucks.
0% of vehicles are light-duty diesel vehicles.
7% of vehicles are heavy-duty gas vehicles.
3% of vehicles are heavy-duty diesel vehicles.
0% of vehicles are motor cycles.
20% of the light-duty gas trucks and vehicles are operating from a
cold start.*
27% of the light-duty gas trucks and vehicles are operating from a
hot start.*
All heavy duty trucks and vehicles are operating hot.
Summer emission rates are based on 24° C (75° F).
Winter emission rates are based on -4° C (25° F).
VehicleAge
Percent
4.9
2.2
2.8
3.9
5.4
6.6
6.9
8.7
10.3
10.6
11.9
15.4
10.4
Age, Years
12 or more
11
10
9
8
7
6
5
4
3
2
1
0
Hourly Traffic
Volume Factors, f(ADT)
Hour
1
2
3
4
5
6
7
8
9
10
11
12
Factor
.012
.006
.005
.004
.005
.013
.049
.105
.076
.057
.052
.054
Hour
13
14
15
16
17
18
19
20
21
22
23
24
Factor
.053
.058
.064
.078
.101
.090
.058
.038
.030
.030
.027
.020
* The remainder of the vehicles (53%) are operating
in a hot stabilized condition.
Source: Colorado Division of Highways (1976)
-------
500
10
15
20 25 30
Average Vehicle Speed (mph)
D 1974
A 1985
O 1990-2000
40
45
50
CO
(a) Winter (Ambient temperature = 25° F.)
FIGURE B-2. VEHICLE EMISSIONS FACTORS USED IN THE DENVER MODEL
-------
500
400
300
200
IOO
0
D 1975
A 1985
O 1990-2000
(
V,
1
^
X
==,
•— — ,
£~" 1
^ *
*~ — -^
\ .
" N
1 1
*> =1
}- .-
1 {
> «
3
:) =
a c
^) (.
5 10 15 20 25 30 35 40 45 5(
Average Vehicle Speed (mph)
Source: Colorado Division of Highways (1976).
(b) Summer (Ambient temperature = 75° F.)
00
en
FIGURE B-2. (Concluded)
-------
TABLE B-3. ESTIMATED OPERATING SPEEDS FOR TRAFFIC IN THE DENVER METROPOLITAN AREA
BY ROADWAY AND AREA TYPES
(miles per hour)
1 J pC U I
Roadwav
Freeway
Expressway
Principal arterial
Major arterial
Minor arterial
Collector
Centroid connector
Ramp
Freeway
Expressway
All arterial s
Collector
Centroid connector
Freeway
Expressway
All arterials
Collector
Centroid connector
CBD*
--
15
15
15
10
10
10
--
--
10
10
10
--
—
10
10
10
Fringe1
50
35
20
20
20
15
10
15
30
--
20
10
10
25
--
20
10
10
Residential5 Rural
55 55
40
25
25
25
20
15
20
50
25
20
10
15
50
--
25
15
15
45
35
35
35
30
15
30
50
35
25
15
15
50
--
30
20
15
CBD Fringe Residential Rural
35 45 45
—
10
10
10
10
10
10
--
--
10
10
10
--
--
10
10
10
25
15
15
15
15
10
—
20
--
15
10
10
20
--
15
10
10
30
20
20
20
20
15
--
40
20
15
10
15
40
--
20
15
15
45
35
35
35
30
15
--
50
35
25
15
15
50
--
30
20
15
Year
1975
1935
(used
in
1-470
study)
2000
(JRPP
plan)
* CBD = Central business district.
f Fringe = Fringe and outlying business area.
5 Residential = Residential and suburban.
Source: Mr. Rich Griffins, Colorado Division of Highways.
ro
00
-------
0.10
0.08
0.06
0.04
0.02
PEAK HOURS-
OFF-PEAK
1 1 1 1 u
OFF-PEAK
/\ >
-PEAK HOURS
OFF-PEAK
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22
Hour of the Day (MST)
23 24
oo
FIGURE B-3. DIURNAL TRAFFIC FLOW VARIATION IN THE DENVER METROPOLITAN REGION
-------
288
emission variations, may be input to the Denver Model. To create an
input file for COM, the inputs to SAIEMIS were first adjusted to reflect
yearly average emission factors rather than seasonal (summer or winter)
emissions; the output of SAIEMIS was further modified by a post-
processor program to obtain average daily emissions that do not reflect
the diurnal variations in traffic flow.
In order to present graphically some of the changes in land use
anticipated between the years 1971 and 2000, we plotted the links that
fall within each area type classification for both these years. These
plots are presented in Figures B-4 through B-7. Figures B-4(a) and
(b) show that the Central Business District comprises only a very small
portion of the total area and increases only very slightly from 1971
to 2000. A comparison of Figures B-5(a) and (b) shows that the fringe
and outlying business area are expected to develop significantly in
the period from 1971 to 2000. Finally, Figures B-6(a) through B-7(b)
show that a significant southward spreading of residential suburbs is
expected at the expense of the rural area there. This development
will also occur--although to a somewhat lesser extent—northward,
westward, and eastward of Denver.
The procedure for determining the Auto Area emissions is outlined
in the AQMA report of the Department of Health. Since these emissions
account for less than one percent of the total auto emissions, we do
not discuss this procedure here.
In treating auto emissions, the following assumptions were made:
> By weight, 80 percent of the hydrocarbons emitted are reactive
and 20 percent are unreactive.
> Nitrogen oxides are 85 percent by weight nitric oxide and 15
percent by weight nitrogen dioxide.
-------
289
(a) Year 1971
FIGURE B-4. CENTRAL BUSINESS DISTRICT IN YEARS 1971 and 2000-
ALL DENVER LINKS. The grid is of 1-mile squares.
-------
290
(b) Year 2000
FIGURE B-4. (Concluded)
-------
291
(a) Year 1971
FIGURE B-5. FRINGE AREA IN YEARS 1971 AND 2000—ALL DENVER
LINKS. The grid is of 1-mile squares.
-------
292
(b) Year 2000
FIGURE B-5. (Concluded)
-------
293
(a) Year 1971
FIGURE B-6. SUBURBAN AREA IN YEARS 1971 and 2000--ALL
DENVER LINKS. The grid is of 1-mile squares
-------
294
ri /j f ^m
I i/»H * I I "'
(b) Year 2000
FIGURE B-6. (Concluded)
-------
295
[a) Year 1971
FIGURE B-7. RURAL AREA IN YEARS 1971 and 2000--ALL
DENVER LINKS. The grid is of 1-mile squares.
-------
296
!b) Year 2000
FIGURE B-7. (Concluded)
-------
297
b. Emissions from Other Sources
As indicated above, point sources and such sources as street sand-
ing and construction also contribute significantly to the overall
emissions of pollutants in Denver. The major assumptions made in esti-
mating the emissions from these and other sources are summarized in
Table B-4.
4. ANALYSIS OF EMISSIONS INPUT TO COM
The Climatological Dispersion Model accepts both point source
emissions and area source emissions as inputs. The input emission rates
must reflect annual averages. Inputs to the model include certain
weighting factors to account for the differences in day/night emissions.
There were 171 point sources input to COM and a number of area
sources:
> Auto area
> Auto link
> Ai rport
> Construction
> Street sanding
> Space heating
> Small point.
The emissions files used as inputs to COM in this study were ana-
lyzed with the help of a specially written program to determine the
trends in overall pollutant loading, the pollutant loading due to each
source category, and the pollutant loading within each subregion
designated for the sensitivity study.
Table B-5 presents the total emissions of particulates and nitrogen
oxides (as nitrogen dioxide) for all modeling years considered here;
-------
TABLE B-4. SUMMARY INFORMATION ON NONVEHICULAR SOURCES
Source
Category
Point
Sources
(general)
Gasoline
Service
Stations
Incinerators
Assumptions and Comments
With the exception of emissions from the Cherokee Power Plant, emissions from existing
point sources are assumed to remain unchanged from 1974 emissions levels.
Emissions from the Cherokee Power Plant for the years 1980 and beyond are assumed to
be one-half of the 1974 emissions levels.
Hydrocarbon emissions are assumed to be 75 percent reactive and 25 percent nonreactive.
Nitrogen oxide emissions are assumed to be 85 percent by weight nitric oxide and 15
percent nitrogen dioxide.
Point source projections are made on the assumption that increases in point source
emissions are directly proportional to increases in the area of industrially zoned
land.
There are a total of 171 point sources. All point sources are input to COM, but for
the Denver Model the largest sources are aggregated to form 46 sources. Plume rise is
computed using modified forms of Holland's equation.
Emissions for 1974 were computed in proportion to the number of pumps.
Future emissions (1980, 1985) are assumed to increase in proportion to the increase in
gasoline usage.
Phase I vapor recovery is assumed for future years (1980, 1985).
Sources of
Information
Emissions Inventory
Subsystem (EIS).
AP-42.
DRCOG empiric four-
cycle zonal allocations
State Department of
Labor and Employment,
Oil Inspection
Section.
AP-42.
Source testing.
Air Contaminant
Emission Notice
File.
ro
10
CO
Solvent
Users
Solvent use changes are assumed to be proportional to population changes.
Sources are primarily dry cleaners. Two types of solvents are considered--Perk and
Stoddard.
List of solvent
users compiled by
Enforcement Unit.
DRCOG empiric four-
cycle zonal allocations
-------
TABLE B-4. (Continued)
Source
Category
Oil-based
Paint Use
Space
Heating
Assumptions and Comments
Airports
Consumption of oil paint is assumed to be proportional to population.
Oil-based paint use in 1980 is assumed to be 50 percent of 1974 usage.
No oil-based paint usage in 1985 and beyond.
Residential gas use is directly proportional to population.
Commercial gas use is directly proportional to the area of commercially zoned land.
Industrial gas use by small industrial users is directly proportional to the area of
industrially zoned land.
Industrial gas use by large industrial users is unchanged through the years.
The airports are Stapleton International, Arapahoe County, Jefferson County, and
Buckley Air National Guard Base.
Sources of
Information
Colorado Paint Manu-
facturers Association.
DRCOG four-cycle
empiric zonal
allocations.
Public Service Company
of Colorado.
ORCOG four-cycle
empiric zonal
allocations.
Environmental impact
statements.
FAA Master Records,
(1974, 1975).
AP-42.
Traffic Control Center
(Stapleton).
Aviation Forecasts
Fiscal Years 1975-1986.
FAA Report.
Military Air Traffic
Forecasts Fiscal Years
1975-1986, FAA report.
-------
TABLE B-4. (Concluded)
Source
Category
Street
Sanding
Construction
Assumptions and Comments
Sources of
Information
0.17 lb/(vehicle-mile) emission factor (particulates)
On 18 days per year dry sanded street conditions exist throughout the metropolitan
area.
Seventy percent of the total VMT is on sanded streets.
Ninety percent of the average daily VMT occurs on the 18 "sand days.
Projections of street sanding emissions are based on growth in auto link VMT.
Six months per year of active construction.
Emissions of particulates are 0.8 tons/acre/month of active construction.
Rate of construction is proportional to the rate of change of total acres of land used
for residential, commercial, and industrial purposes.
Investigation of
Fugitive Dust:
Sources, Emissions
and Control for
Attainment of Ambient
Air Quality Standards
Colorado, PEDCO
report.
Investigation of
Fugitive Dust:
Sources, Emissions
and Control for
Attainment of Ambient
Air Quality Standards
Colorado,"PEDCO
report.
CO
o
o
-------
301
TABLE B-5. TOTAL PARTICULATE AND NITROGEN OXIDE
EMISSIONS INPUT TO THE COM MODEL
Year
Emissions 1974 1980 1985 1990 2000
Participate emissions
Tons per day 88 106 130 152.6 157.6
Increase over 1974 emissions -- 20% 48% 73% 79%
Nitrogen oxide emissions
(as N02)
Tons per day 180 213 241 245 262
Increase over 1974 emissions -- 18% 34% 36% 46%
-------
302
the table shows that particulate emissions tend to rise relatively
faster than do the nitrogen oxide emissions, over the period from 1971
to 2000. Figure B~8 shows the changes in emissions over this period;
the yearly increase of nitrogen oxide emissions is much lower in the
period from 1985 to 200Q than in the period from 1974 to 1985. There
is a similar lowering in the rate of increase of particulate emissions
over the period from 1990 to 2000.
Table B-6 presents the particulate emissions from the point sources
and the various source categories for which separate files were
created. In Figure B-9 the year-to-year variation in total emissions
from each source category is plotted and in Figure B-10 the percentage
contribution from each source category is plotted. Street sanding and
construction can be seen to be major sources of particulates; large
point sources and auto link emissions account for most of the remaining
particulate emissions. It must be noted that there appears to be a
rather large and sudden increase in emissions from the "space heating
plus construction" category in the period from 1985 to 1990. The
estimates of emissions up to and including the year 1985 were made by
the Department of Health (Air Pollution Control Division), and the
estimates for the years 1990 and 2000 were made by the Colorado
Division of Highways. The abrupt change in estimated construction
and space heating emissions could be due to some difference in the
estimation methodologies used by these independent groups. A some-
what similar but less abrupt change in street sanding emissions is
also evident.
The nitrogen oxides emissions were also analyzed by source cate-
gory; these data are presented in Table B-7 and Figures B-ll and B-12.
Auto link and large point emissions account for the major part of the
emissions. Space heating and small point sources also emit consider-
able quantities of nitrogen oxides. There again appears to be a rather
marked change in the year-to-year trend in emissions over the period
from 1985 to 1990 for the space heating and auto link emissions
categories. As before, this change could result from differences
-------
303
300
c
o
250
200
150
100
50
A Nitrogen Oxides (as N02)
O Particulates
-O
1974
1980
1985
1990
2000
Mode linn Year
FIGURE B-8. TOTAL EMISSIONS OF N02 AND PARTICULATES INPUT
TO THE CLIMATOLOGICAL DISPERSION MODEL
-------
TABLE B-6. PARTICULATE EMISSIONS FROM SOURCE CATEGORIES AS INPUT TO THE COM MODEL.
A = tons per day, B = percent of total.
1974
Source Category
Auto area
Auto link
Airport
Street sanding
Space heating
Construction
Space heating
plus construction
Small point
Large point
Total
A
0.0823
10.4
0.156
39.4
1.27
21.7
23.0
0.233
14.6
88
B
0.094%
11.8
0.18
44.8
1.4
24.7
26.1
0.26
16.6
99.9
1980
A
0.0377
11.1
0.326
56.4
1.57
21.7
23.3
2.27
12.3
106
B
0.036%
10.5
0.31
53.2
1.5
20.5
22.0
2.1
11.6
99.7
0.
13.
0.
81.
1.
17.
19.
3.
12.
130
1985
A
0591
9
188
2
69
4
1
68
3
B
0.045?;
10.7
0.14
62.5
1.3
13.4
14.7
2.8
9.5
100.4
1990
A
0.0591
14.1
0.22
80.6
*
*
41.9
3.68
12.3
152.6
B
0.039%
9.2
0.14
52.8
*
*
27.5
2.4
8.1
100.1
2000
A
0.0539
15.1
0.253
86.2
*
*
40.5
3.68
12.3
157.6
B
0.034%
9.5
0.16
54.7
*
*
25.7
2.3
7.8
100.3
* For the years 1990 and 2000, space heating and construction emissions were lumped together; their total
is therefore also shown for prior modeling years.
OJ
o
-------
305
100
10
O Street Sanding
D Space Heating plus Construction
Q Construction
^ Large Point
A Auto Link
•^ Small Point
O Space Heating
0 Airport
.1
_JL_L
1974
1980
1985
Modeling Year
1990
2000
FIGURE B-9. PARTICULATE EMISSIONS FROM SOURCE CATEGORIES
INPUT TO THE COM MODEL
-------
306
100(
-o
-a
O Street Sanding
D Space Heatinn plus Construction
0 Construction
A Large Point
/\ Auto Link
o Small Point
O Space Heating
0 Airport
1974
1980
1985
Modeling Year
1990
2000
FIGURE B-10. PERCENTAGE CONTRIBUTION OF SOURCE CATEGORIES
TO PARTICULATE EMISSIONS
-------
TABLE B-7. NITROGEN OXIDE EMISSIONS (AS N02) FROM SOURCE CATEGORIES AS INPUT
TO THE COM MODEL. A = tons per day, B = percent of total.
1974
Source Category
Auto area
Auto link
Airport
Street sanding
Space heating
Construction
Small point
Large point
total
A
0.523
78.2
3.86
0
13.3
0
1.216
84
180
B
0.29%
43.4
2.1
0
7.4
0
0.12
46.7
100
1980
A
0.313
99.9
4.57
0
16
0
8.58
84
213
B
0.15%
46.9
2.1
0
7.5
0
4.0
39.4
100
1985
A
0.476
120
5.23
0
17.2
0
14.4
84
241
B
0.20%
49.8
2.2
0
7.1
0
6.0
34.0
100.2
1990
A
0.476
10.5
6.12
0
35.5
0
14.4
84
245
B
0.19%
42.9
2.5
0
14.5
0
5.9
34.3
100.3
2000
A
0.351
114
7.04
0
42.1
0
14.4
84
252
B
0.13%
43.5
2.7
0
16.1
0
5.5
32.1
100
-------
308
Space Heating
Small Point
1980
1985 1990
Modeling Year
2000
FIGURE B-ll
NITROGEN OXIDE EMISSIONS FROM SOURCE CATEGORIES
INPUT TO THE COM MODEL
-------
309
Large Point
Auto Link
Space Heating
Small Point
Airport
Auto Area
1974
1980
1985
Modelinq Year
1990
2000
FIGURE B-12.
PERCENTAGE CONTRIBUTION OF SOURCE CATEGORIES
TO NITROGEN OXIDE EMISSIONS
-------
310
in the estimation methodologies used by the Health Department and the
Division of Highways.
Both the participate emissions and the nitrogen oxides emissions
were further analyzed to determine how much of each of these pollutants
is emitted within each of the seven subregions delineated in the sensi-
tivity study and also in the surrounding area. There results are
presented in Tables B-8 and B-9 and Figures B-13 through B-16. The
majority of pollutants are emitted in Central Denver and in the "outer"
subregion (due to the large total area of this subregion). In the case
of the oxides of nitrogen, more than 50 percent of the total emissions
are emitted outside the seven subregions chosen for the sensitivity study.
This is largely due to the fact that 90 percent of the large point
emissions lie outside the seven subregions. Finally, note that even
though the emissions in the relatively small subregions such as South
Metro and Jeffco-Urban rise comparatively rapidly, they still account
for less than a few percent of the total emissions even in the year
2000.
Another simple set of calculations for studying how the emissions
of pollutants vary from year to year within each of the subregions is
to study the density of emissions (or emissions flux) within each
subregion. Table B-10 indicates the total land area covered by each
subregion. Using this table and data on total emissions within each
subregion, the emissions fluxes were computed and are presented in
Tables B-ll and B-12 and Figures B-17 and B-18. Emissions fluxes
increase in all the subregions but the rate of increase is relatively
rapid in some of the currently less developed areas, such as Jeffco-
Urban and Metro. Furthermore, there appears to be some decrease in
the overall range of emissions fluxes as development of land spreads
outward from Central Denver.
5. ANALYSIS OF EMISSIONS INPUT TO THE DENVER MODEL
Since the Denver Model predicts concentrations of reactive pollu-
tants and is used for short-term averages, its emissions input
-------
TABLE B-8. PARTICULATE EMISSIONS IN SUGREGIONS AS INPUT TO THE COM MODEL.
A = tons per day, B = percent of total.
1974
1
2
3
4
5
6
7
8
Subregion
Outer
Broomfield-Westminster-
Arvada
Northglenn-Thornton
Aurora
South Metro
Jeffco-Urban
Lakewood
Central Denver
Total
A
32.6
6.18
2.84
4.91
3.4
1.18
6.62
30.2
88
B
37 . 0%
7.0
3.2
5.6
3.9
1.3
7.5
34.3
99.8
1980
A
37.1
7.88
3.81
6.13
4.37
1.51
8.08
36.9
106
B
35.0%
7.4
3.6
5.8
4.1
1.4
7.6
34.8
99.7
1985
A
47.1
9.65
4.58
7.68
6.58
3.53
9.43
41.9
130
B
36.2%
7.4
3.5
,5.9
5.1
2.7
7.3
32.2
100.3
1990
A
51.6
10.8
5.99
9.37
8.13
8.73
12.3
45.9
153
B
33.7%
7.1
3.9
6.1
5.3
5.7
8.0
30
99.8
2000
A
51.2
12.2
6.14
10.5
7.91
9.49
12.5
48.2
158
B
32.4%
7.7
3.9
6.6
5.0
6.0
7.9
30.5
100
-------
TABLE B-9. NITROGEN OXIDE EMISSIONS (AS N0~) IN SUBREGIONS AS INPUT TO THE COM
MODEL. A = tons per day, B = percent of total.
1974 1980 1985 1990 2000
Subreglon
1 Outer
2 Broomfield-Westminster-
Arvada
3 Northglenn-Thornton
4 Aurora
5 South Metro
6 Jeffco-Urban
7 Lakewood
8 Central Denver
Total
A
103
7.41
3.44
4.14
2.39
0.631
7.93
51.6
180
B
57.2%
4.1
1.9
2.3
1.3
0.4
4.4
28.7
100.3
A
115
9.79
5.01
6.14
4.7
1.29
10.2
60.9
213
B
54.0%
4.6
2.4
2.9
2.2
0.6
4.8
28.6
100.1
A
126
12.2
6.13
8.37
8.42
3.88
11.8
63.8
241
B
52.3%
5.1
2.5
3.5
3.5
1.6
4.9
26.5
99.9
A
131
12.9
6.64
9.02
8.19
4.39
12
61.3
245
B
53.5%
5.3
2.7
3.7
3.3
1.8
4.9
25
100.2
A
132
14.5
7.4
10.3
8.77
6.12
14.4
68.3
262
B
50.
5.
2.
3.
3.
2.
5.
26.
99.
4%
5
8
9
3
3
5
1
8
OJ
ro
-------
313
TOO
D Central Denver (Metro)
O Outer Areas
0 Lakewood
O Broomfield-Westminster-Arvada
A Aurora
o South Metro
/\ Northglenn-Thornton
0 Jeffco-Urban
2 10
1974
1980
T995~
Model ina Year
2000
FIGURE B-13.
PARTICULATE EMISSIONS IN SUBREGIONS
INPUT TO THE COM MODEL
-------
314
lOOr
-a
10
Outer Areas
Central Denver (Metro)
Lakewood
<3> Broomfield-Westminster-Arvada
A Aurora
c? South Metro
A Northglenn-Thornton
0 Jeffco-Urban
1974
1980 1985 1990
Modelinq Year
2000
FIGURE B-14. PERCENTAGE CONTRIBUTION OF SUBREGIONS TO
PARTICULATE EMISSIONS
-------
1000
315
O Outer Areas
Q Central Denver (Metro)
<} Lakewood
Broomfield-Westminster-Arvada
A Aurora
A Northglenn-Thornton
<37 South Metro
Q Jeffco-Urban
100
-O
10
1974
1980
1985 1990
Modeling Year
2000
FIGURE B-15.
NITROGEN OXIDE EMISSIONS IN SUBREGIONS
INPUT TO THE COM MODEL
-------
316
lOOr
O
•O
10
-D
c
O
O Outer Areas
D Central Denver (Metro)
0 Lakewood
O Broomfield-Westminster-Arvada
A Aurora
^ Northglenn-Thornton
^ South Metro
Q Jeffco-Urban
1974
1980 1985 1990
Modeling Year
2000
FIGURE B-16.
PERCENTAGE CONTRIBUTION OF SUBREGIONS
TO NITROGEN OXIDE EMISSIONS
-------
317
TABLE B-10. SIZE OF THE SUBREGIONS SELECTED
FOR SENSITIVITY STUDY
Subregion Number of Grid Squares*
1 Outer 509
2 Broomfield-Westminster-Arvada 75
3 Northglenn-Thornton 31
4 Aurora 64
5 South Metro 41
6 Jeffco-Urban 27
7 Lakewood 41
8 Central Denver 112
* Note: each grid unit is one square mile.
-------
318
TABLE B-ll. PARTICIPATE EMISSIONS DENSITY
BY SUBREGION (kg/day/sq. mi.)
Year
1
2
3
4
5
6
7
8
Subregion
Outer
Broomfield-
Westminster-
Arvada
Northglen
Thornton
Aurora
South Metro
Jeffco-Urban
Lakewood
Central Denver
Average
1974
64.0
82.4
91.6
76.8
82.8
43.6
162
270
98
1980
72.8
105
123
95.6
106
56
197
330
118
1985
92.4
129
147
120
160
129
230
374
144
1990
101
144
193
146
198
323
300
410
170
2000
100
163
198
164
193
352
305
430
176
-------
319
TABLE B-12. NITROGEN OXIDE EMISSIONS (AS N02) DENSITY
BY SUBREGIONS (kg/day/sq. mi.)
1
2
3
4
5
6
7
8
Subregion
Outer
Broomf i el d-Westmi nster-Arvada
Northglenn^-Thornton
Aurora
South Metro
Jeffco-Urban
Lakewood
Central Denver
1974
202
98.8
no
64.8
58.4
23.2
194
461
1980
226
130
162
96
115
47.6
249
544
Year
1985
248
163
198
131
205
144
288
570
1990
257
172
214
141
200
162
293
547
2000
259
193
239
161
214
227
351
610
Average 200 237 268 272 291
-------
320
1000
E
ty
iooh
O
D Central Denver (Metro)
Q Lakewood
d Northglenn-Thornton
^3 Soutn Metro
Q Broomfield-Westminster-Arvada
^ Aurora
O Outer Areas
0 Jeffco-Urban
10*-
1974
1980
1985 1990
Model inn Year
2000
FIGURE B-17. PARTICULATE EMISSIONS FLUX IN THE SUBREGIONS
-------
321
lOOOr
s
o-
TOO
Central Denver (Metro)
O Outer Areas
Q Lakewood
^ Northglenn-Thornton
O Broomfield-Westminster-Arvada
A Aurora
o South Metro
0 Jeffco-Urban
101-
±
1974
1980 1985 1990
Modeling Year
2000
FIGURE B-18. NITROGEN OXIDES EMISSIONS (AS N09) IN THE SUBREGIONS
-------
322
requirements differ significantly from the input requirements of COM.
In the first place, the Denver Model requires emissions of reactive
hydrocarbons and nitrogen oxides to determine the concentrations of
photochemical oxidants. If carbon monoxide concentrations are to be
predicted, then CO emissions must also be input. The reactive hydro-
carbon emissions are further split into emissions of aldehydes, olefins,
aromatics, and paraffins. In the second place, since short-term
averages are desired, short-term variations in the emission rates,
such as diurnal and seasonal variations, must be accounted for.
The emissions used as inputs to the Denver Model were analyzed by
means of a special computer program to determine the contributions
arising from each source category. These analyses for the summer and
winter months of 1975, 1985, and 2000 are presented in Tables B-13
and B-14, and in Figures B-19 through B-22. An examination of these
tables and figures reveals the following;
> Winter emissions exceed summer emissions for reactive hydro-
carbons, nitrogen oxides, and carbon monoxide. This is appar-
ently due to an increase in auto emissions and "other area"
emissions in the winter. The most significant differences
between winter and summer emissions are for carbon monoxide.
> Emissions of carbon monoxide and reactive hydrocarbons decrease
in future years, apparently due to a reduction in automobile
emissions as a result of the Federal Motor Vehicle Control
Program.
> Emissions of nitrogen oxides increase in future years, apparently
due to an increase in emissions from automobiles and a few
other sources. The fact that total auto emissions increase
in spite of the Federal program is probably due to the increas-
ing flow of traffic. "Other area" emissions of nitrogen oxides
also appear to rise significantly in future years.
> Automobiles appear to be the prime source of pollution, con-
tributing about 75 percent of the reactive hydrocarbons, 45
percent of the nitrogen oxides, and 85 percent of the carbon
monoxi de.
-------
TABLE B-13. TOTAL EMISSIONS INPUTS TO THE DENVER MODEL—SUMMER.
A - tons per day, B = percent of total.
(a) Year 1975
Total NOV
RHC
Airport
Auto area
Auto link
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources*
Total
A
4.30
1.35
148.1
153.7
1.90
13.1
168.8
13.2
182.0
B
2.4%
0.74
81.4
84.5
1.0
7.2
92.7
7.3
100
URHC
A
1.07
0.34
49.9
51.3
0.48
3.28
55.0
4.4
59.4
B
1.8%
0.57
84.0
86.4
0.81
5.5
92.6
7.4
100
NO
A
2.27
0.23
49.1
51.6
3.16
1.38
56.1
66.6
122.7
B
1.9%
0.2
40.0
42.1
2.6
1.1
45.7
54.3
100
NO,
A
0.39
0.039
8.30
8.72
0.54
0.23
9.50
14.7
24.2
B
1.6 Z
0.16
34.3
36.0
2.2
1.0
39.3
60.7
100
as N00
A
3.87
0.39
83.6
87.8
5.4
2.35
95.5
116.8
212.3
B
1.8%
0.18
39.4
41.4
2.5
1.1
45.0
55.0
100
CO
A
21.1
17.9
1730
1769
1.93
2.56
1774
190.9
1965
B
1.1%
0.91
88.0
90.0
0.098
0.13
90.3
9.9
100
*Hydrocarbon emissions from large point sources are assumed to be 75 percent reactive hydrocarbons (RHC)
and 25 percent unreactive hydrocarbons (URHC).
OJ
IND
CO
-------
TABLE B-14 (Continued)
(b) Year 1985
RHC
Airport
Auto area
Auto link
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources*
Total
A
5.28
0.56
106.7
112.6
0.43
12.0
125.0
13.2
138.2
B
3.8%
0.41
77.2
81.4
0.31
8.7
90.4
9.6
100
URHC
A
1.32
0.14
36.8
38.3
0.108
3.01
41.4
4.4
45.8
B
2.9%
0.31
80.3
83.6
0.24
6.6
90.4
9.6
100
NO
A
3.08
0.84
71.4
75.3
4.03
9.68
89.1
66.6
155.7
B
2.0%
0.5
45.9
48.4
2.6
6.2
57.2
42.8
100
NO.,
A
0.52
0.14
12.1
12.8
0.69
1.65
15.1
14.7
29.8
B
1.7%
0.5
40.6
43.0
2.3
5.5
50.7
49.3
100
Total NOX
as N00
A
5.24
1.43
121.6
128.3
6.9
16.5
151.7
116.8
268.5
«-
B
2.0%
0.53
45.3
47.8
2.5
6.1
56.5
43.5
100
CO
A
25.2
8.71
1278
1312
1.35
3.88
1317
190.9
1508
B
1.7%
0.58
CO
84.7 r\D
87.0
0.09
0.26
87.3
12.7
100
-------
TABLE B-14 (Concluded)
(c) Year 2000
RHC
Airport
Auto area
Auto link
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources*
Total
A
5.28
0.21
73.6
79.1
0.986
12.0
92.1
13.2
105.3
B
5.0%
0.20
69.9
75.1
0.94
11.4
87.5
12.5
100
URHC
A
1.32
0.053
24.3
25.7
0.247
3.01
29.0
4.4
33.4
B
4.0%
0.16
72.8
76.9
0.74
9.0
86.8
13.2
100
NO
A
3.08
0.14
71.2
74.4
9.91
9.68
94.0
66.6
160.6
B
1.9%
0.1
44.3
46.3
6.2
6.0
58.5
41.5
100
N00
A
0.52
0.024
12.1
12.7
1.69
1.65
16.0
14.7
30.7
B
1.7%
0.1
39.4
41.4
5.5
5.4
52.1
47.9
100
Total NOX
as N00
A
5.24
0.24
121.3
126.8
16.9
16.5
160.1
116.8
276.9
L.
B
1.9%
0.087
43.8
45.8
6.1
6.0
57.8
42.2
100
CO
A
25.2
3.47
866.2
894.9
3.08
3.88
901.8
190.9
1093
2.3%
0.32
79.2
81.9
0.28
0.35
82.5
17.5
100
CO
ro
en
-------
TABLE B-14. TOTAL EMISSIONS INPUTS TO THE DENVER MODEL—WINTER.
A = tons per day, B = percent of total.
(a) Year 1974
RHC
Airport
Auto area
Auto link
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources*
Total
A
4.30
1.93
206.8
213.0
6.18
13.1
232.3
13.2
245.5
B
1.8%
0.79
84.2
86.8
2.5
5.3
94.6
5.4
100
URHC
A
1.07
0.48
69.5
71.0
1.55
3.28
75.8
4.4
80.2
B
1 .3%
0.60
86.7
88.5
1.9
4.1
94.5
5.5
100
NO
A
2.27
0.33
53.9
56.5
10.3
1.35
68.1
66.6
134.7
B
1.7%
0.2
40.0
41.9
7.6
1.0
50.6
49.4
100
NO,
A
0.39
0.056
9.19
9.63
1.75
0.23
11.6
14.7
26.3
B
1.5%
0.2
34.9
36.6
6.7
0.9
44.1
55.9
100
Total NOX
as N00
A
3.87
0.56
91.8
96.3
17.5
2.3
116.0
116.8
232.8
t-
B
1.7%
0.24
39.4
41.4
7.5
1.0
49.8
50.2
100
CO
A
21.1
25.5
2793
2840
6.28
2.56
2849
190.9
3040
B
0.69%
0.83
91.9
93.4
0.21
0.084
93.7
6.3
100
CO
* Large point hydrocarbon emissions are assumed to be 75 percent RHC and 25 percent URHC.
-------
TABLE B-14 (Continued)
(b) Year 1985
RHC
Airport
Auto area
Auto link
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources
Total
A
5.28
0.75
138.4
144.4
1.4
12.0
157.9
13.2
171.1
B
3.1%
0.44
80.9
84.4
0.82
7.0
92.3
7.7
100
URHC NO NO,
A
1.32
0.19
46.5
48.0
0.35
3.01
51.4
4.4
55.8
Total NOX
as N00 CO
B A B* A B* A B* A
2.4% 3.08 0.52
0.34 0.92 0.16
83.3 — * --*
86.0 — * — *
0.63 13.1 2.23
5.4 9.63 1.65
92.1 --* --*
7.9 66.6 14.7
100 — * — *
5.24 25.2
1.57 14.6
— * 2034
--* 2074
22.3 4.38
16.5 3.88
--* 2082
116.8 190.9
--* 2273
B
1.1%
0.64
89.5
91.2
0.19
0.17
91.6
8.4
100
* The emissions factors used as inputs to compute auto link NO emissions were in error.
t Large point hydrocarbon emissions are assumed to be 75 percent RHC and 25 percent URHC.
OJ
ro
-------
TABLE B-14 (Concluded)
(c) Year 2000
RHC
Airport
Auto area
Auto link*
Subtotal
Other area
Small
point sources
Subtotal
Large
point sources
Total
A
5.28
0.38
82.7
88.4
3.2
12.0
103.6
13.2
116.8
B
4.5%
0.33
70.8
75.7
2.7
10.3
88.7
11.3
100
URHC
A
1.32
0.095
28.4
29.5
0.80
3.0
33.6
4.4
38.0
B
3.5%
0.25
74.7
78.4
2.1
7.9
88.4
11.6
100
NO
A
3.08
0.23
79.9
83.2
32.2
9.7
125.1
66.6
191.7
B
1.6%
0.1
41.7
43.4
16.8
5.1
65.3
34.7
100
NO,,
A
0.52
0.040
13.4
14.0
5.48
1.6
21.1
14.7
35.8
B
1.5%
0.1
37.4
39.1
15.3
4.5
58.9
41.1
100
Total NOX
as NO,
A
5.24
0.39
135.9
141.6
54.9
16.5
212.9
116.8
329.7
B
1.6%
0.12
41.2
42.9
16.7
5.0
64.6
35.4
100
CO
A
25.2
6.1
983.4
1015
10.0
3.9
1029
190.9
1220
B
2.1%
0.5
80.6
83.2
0.82
0.32
84.3
15.6
99.9
* The emissions factors used to compute auto link emissions were in error.
t Large point emissions are assumed to be 75 percent RHC and 25 percent URHC.
oo
r\D
oo
-------
329
5000
O-.
— — Winter
————— Summer
O CO
Q Total N0x as N02
A Reactive Hydrocarbons
D NO Only
0 NO, Only
1000
100
--0
0
0
10
1975
1985
2000
Modelino Year
FIGURE B-19. TOTAL EMISSIONS INPUTS TO THE DENVER MODEL
-------
330
TOO
__ Winter
Summer
A Auto Link
A Large Point
0 Airport
• Auto Area
• Other Area
o Small Point
10
E
LjJ
O
c
-------
100
331
A Large Point Emissions
A Auto Link
• Other Area
0 Airport
^ Small Point
• Auto Area
10
.1
1975
1985
Modeling Year
2000
FIGURE B-21. PERCENTAGE CONTRIBUTION OF SOURCE CATEGORIES
TO NITROGEN OXIDE EMISSIONS—SUMMER
-------
332
1001
4
Summer
Winter
A Auto Link
A Large Point
^ Small Point
0 Airport
• Other Area
• Auto Area
10
.1
1975
1985
Modeling Year
2000
FIGURE B-22. PERCENTAGE CONTRIBUTION OF SOURCE CATEGORIES
TO REACTIVE HYDROCARBON EMISSIONS
-------
333
> Point sources are the second largest source of pollutants, con-
tributing about 20 percent of the reactive hydrocarbons, 50
percent of the nitrogen oxides, and 10 percent of the carbon
monoxide.
As in the case of COM emissions inputs, the Denver Model emissions
inputs were analyzed to determine the total emissions from each of the
subregions delineated for the sensitivity study. These results are
presented in Table B-15 and Figures B-23 through B-28. Central Denver
has by far the largest emissions of the seven regions chosen for study.
The "outer" area not included in the sensitivity study accounts for a
suprisingly large portion of the total emissions, especially for the
oxides of nitrogen (over 55 percent). The emissions in some of the
southern subregions such as South Metro and Jeffco-Urban either do not
decrease as rapidly as the emissions in the other regions in the cases
of reactive hydrocarbons and carbon monoxide or increase relatively
rapidly Cin the case of nitrogen oxides). However, these subregions
still only contribute a few percent of the total pollutant loading for
the entire modeling region.
-------
TABLE B-15. EMISSIONS IN SUBREGIONS AS INPUT TO THE DENVER MODEL—SUMMER.
A = tons per day, B = percent of total.
1975
1985
2000
RHC
1
2
3
4
5
6
7
8
Subregion
Outer
Brooinf ield-
Westminster-
Arvada
Northglenn-
Thornton
Aurora
South Metro
Jeffco-Urhan
Lakewood
Centra]
Denver
Total
A
47.5
11.9
5.15
9.73
4.78
0.984
13.3
88.7
182.1
B
26.1%
6.5
2.8
5.3
2.6
0.54
7.3
48.7
99.8
NOX
as NO
A
132.5
7.06
3.15
4.20
2.39
0.674
7.72
54.6
212.3
B
62.4%
3.3
1.5
2.0
1.1
0.32
3.6
25.7
99.9
CO
A
577.9
128.2
53.1
85.1
51.3
10.9
144.)
913.8
1964.4
B
29.4%
6.5
2.7
4.3
2.6
0.55
7.3
46.5
99.9
RHC
A
41.9
9. 72
4.22
8.76
5.75
3.32
9.21
55.4
138.3
B
30.3%
7.0
3.1
6.3
4.2
2.4
6.7
40.1
100.1
as N00
A
154.2
11.7
5.66
8.08
8.43
4.04
11.42
64.9
268.5
c •
B
57.4%
4.4
2.1
3.0
3.1
1.5
4.3
24.2
100
CO
A
519
109
44.3
81.2
63.7
42.0
100.5
548.6
1508
B
34.4%
7.2
2.9
5.4
4.2
2.8
6.7
36.4
100
RHC
A
31.5
7.02
3.23
7.19
4.04
2.61
7.15
42.6
105.4
B
29.9%
6.7
3.1
6.8
3.8
2.5
6.8
40.4
100
NOX
as NO,,
A
154.5
12.7
6.27
8.78
8.48
5.16
13.59
67.45
276.9
B
55. 8X
4.6
2.3
3.2
3.1
1.9
4.9
34.4
100.2
CO
A
395.4
74.8
32.4
62.3
42.1
30.2
71.1
384.3
1093
B
36.2
6.8
oo
oo
3.0 -^
5.7
3.9
2.8
6.5
35.2
100.1
-------
335
100
10
c
o
n Central Denver (Metro)
O Outer Areas
Q Lakewood
-------
100
336
O Outer Areas
Q Central Denver (Metro)
Q Lakewood
Q Broomfield-Westminster-Arvada
A Aurora
/\ Northglenn-Thornton
^7 South Metro
Jeffco-Urban
10
•O
-a
1975
1985
Modeling Year
2000
FIGURE B-24. NITROGEN OXIDE EMISSIONS IN SUBREGIONS-
SUMMER
-------
337
lOOOr
100
1C
G Central Denver (Metro)
Q Outer Areas
0 Lakewood
Q Broomfield-Westminster-Arvada
A Aurora
/] Northglenn
^37 South Metro
0 Jeffco-Urban
1985
Modeling Year
2000
FIGURE B-25. CARBON MONOXIDE EMISSIONS IN SUBREGIONS—SUMMER
-------
338
100
-o
10
[] Central Denver (Metro)
O Outer Areas
Q Lakewood
Q Broomfield-Westminster-Arvada
^ Aurora
/] Northglenn-Thornton
^3 South Metro
0 Jeffco-Urban
.1
1975
1985
Modeling Year
2000
FIGURE B-26.
PERCENTAGE CONTRIBUTION OF SUBREGIONS TO
REACTIVE HYDROCARBON EMISSIONS — SUMMER
-------
339
100
D-
-G-
10
1 —
f
_L
O Outer Areas
Q Central Denver (Metro)
() Lakewood
Q Broomfield-Westminster-Arvada
A Aurora
A Northglenn-Thornton
^ South Metro
<,^ Jeffco-Urban
1975
1985
Modelinq Year
2000
FIGURE B-27. PERCENTAGE CONTRIBUTION OF SUBREGIONS TO
NITROGEN OXIDE EMISSIONS—SUMMER
-------
340
100
D Central Denver (Metro)
O Outer Areas
0 Lakewood
O Broomfield-Westminster-Arvada
Aurora
A
Northglenn-Thornton
South Metro
0 Jeffco-Urban
1975
1985
Modeling Year
2000
FIGURE B-28. PERCENTAGE CONTRIBUTIONS OF SUBREGIONS
TO CARBON MONOXIDE EMISSIONS—SUMMER
-------
341
APPENDIX C
DENVER REGIONAL PLANNING OVERVIEW
(AS OF 24 NOVEMBER 1976)
-------
342
AUTHOR'S NOTE
The reliability of an airshed model simulation is heavily dependent
on the quality and consistency of the data supplied to it. Because several
agencies separately provided SAI with portions of the data we used, and
because different data vintages existed within each agency, it became
essential early in the Overview EIS project to develop a full perspective
on the Denver planning process. This was particularly necessary for those
planning aspects having direct impact on air quality: population level
and allocation, land use, and transportation.
This appendix was completed in November 1976, early enough to provide
SAI with the perspective needed to guide its analysis efforts. We believe
its portrayal of Denver regional planning is accurate as of its publication
date. Subsequent events, however, are not reflected in the text. We urge
the reader to consider this fact when evaluating its contents.
The most significant change occurring since publication concerns the
role of the Denver Regional Council of Governments (DRCOG). This appendix
refers to the Joint Regional Planning Program (JRPP) as the designated
Metropolitan Planning Organization (MPO), with the DRCOG serving as the
policy body of the JRPP. That relationship has since been altered. As a
result of a three-party agreement between the DRCOG, the Regional Trans-
portation District (RTD) and the Colorado Division of Highways (CDH)--an
accord officially sanctioned by the Governor--the DRCOG has now been
designated as the MPO in the Denver region. In this new role DRCOG sup-
plants the JRPP. Both the RTD and the CDH, previously members of the JRPP,
are now no longer a part of the MPO.
-------
343
APPENDIX C
DENVER REGIONAL PLANNING OVERVIEW
(AS OF 24 NOVEMBER 1976)
by Stanley R. Hayes
1. INTRODUCTION
a. Task Objective
Systems Applications, Incorporated, under contract to the Environ-
mental Protection Agency, Region VIII, is preparing the air quality
element for the Denver Metropolitan Wastewater Facilities Overview EIS.
The broad geographical scope of this study emphasizes the importance
of comprehensive regional planning. To address regional issues while bal-
ancing contending local interests, any such process must be a dynamic one.
Sensitivity and responsiveness require continual improvement and reexamin-
ation. As a result, planning must be viewed as evolutionary.
The analysis of primary and secondary air quality impacts, however,
is acutely sensitive to the assumptions imbedded in regional plans. As an
area characterized by extensive ownership and usage of private automobiles,
Metropolitan Denver is particularly subject to elevated levels of carbon
monoxide (CO) and hydrocarbons (HC). The distribution and intensity of the
resulting automobile-produced mixture is strongly dependent on population,
land use, and the transportation network. To the extent that uncertainties
exist in any of these areas, air quality must likewise be considered in
question. These issues are further complicated by the nonlinear relation-
ship of pollutant concentrations to vehicle-miles-traveled (VMT).
In order to perform properly the analysis required by the Overview
EIS, SAI must consider all those uncertainties having a direct bearing on
air quality. SAI has no intention of intruding on the planning process or
of considering planning alternatives other than those that have already
been the subject of analyses by agencies within the Denver region.
-------
344
Guided by the above caveat it is the purpose of this task report to
consider the planning process in the Denver Metropolitan region, as to
both institutional framework and substantive policy issues. The role of
specific agencies will be explored, along with their particular perspec-
tives on population, land use and transportation network alternatives.
Because the analyses to be performed by SAI will be so heavily
driven by the assumptions underlying the input data supplied to it and
because several permutations of that data exist within the agencies pro-
viding it, a precise, detailed examination of the relevant planning
issues is an essential preliminary to the study process.
The results of this task report will be drawn upon heavily in struc-
turing SAI's analysis efforts. Specifically, a base case regional devel-
opment scenario must be identified along with a complete catalog of the
important uncertainties associated with it. A carefully constructed
sensitivity analysis must then be conducted to examine the air quality
effects of as broad a range of these uncertainties as possible. This
report, through its examination in some depth of the critical elements of
regional planning, will assure the broadest perspective in SAI's efforts.
b. Report Organization
This report has been partitioned into three broad areas: organiza-
tional perspectives, planning issues, and final observations. The first
of these, in Section 2, will examine the role in the planning process
played by the Joint Regional Planning Program (JRPP), the Denver Regional
Council of Governments (DRCOG), the Colorado Division of Highways (CHD),
the Regional Transportation District (RTD), and the Air Pollution Control
Division of the Colorado Department of Health.
The second area of discussion, an overview of planning issues, will
be presented in Section 3. Those issues having the most direct bearing
-------
345
on SAI's air quality modeling will be explored in detail. Both the nature
of the issues themselves and their institutionally endorsed variants will
be considered.
Finally, in Section 4, planning options considered by the various
regional and state agencies mentioned above will be summarized. The im-
pacts of any potential planning uncertainties on the air quality study
performed by SAI will be assessed and preliminary recommendations stated.
The references for this appendix are listed at its end.
-------
346
2. ORGANIZATIONAL PERSPECTIVES
a. Insti tuti ona1 Overview
Since it is the transportation process that exacts so noticeable a
price on air quality in the Denver metropolitan area, this institutional
overview will concentrate on agencies having direct planning responsi-
bilities in areas bearing most heavily on transportation. These areas
include population level and its allocation, land use, transportation
planning (highways and rapid transit), and transportation modeling.
1) Air Quality Regulations
A brief historical perspective is helpful at this point. In 1970
the Congress enacted certain amendments to the Clean Air Act (42 U.S.C.
1857). In addition to creating the Environmental Protection Agency (EPA),
the Act in Section 110 mandated the preparation and approval of State
Implementation Plans (SIP). It was required of each SIP that it set
forth specific plans for the attainment and maintenance of the National
Ambient Air Quality Standards (NAAQS) to be promulgated by the EPA as
part of its statutory responsibilities under Section 109.
For the purposes of air quality evaluation the Administrator of the
EPA was empowered by Section 108 of the Clean Air Act to designate bound-
aries for interstate or intrastate regions called Air Quality Control
Regions (AQCR). The State of Colorado was divided into eight regions,
as shown in Figure C-l, from Reference 1.
In 1972 AQCRs were classified Priority I or Priority III for the
transportation-related pollutants, carbon monoxide (CO), photochemical
oxidants (Ox), and nitrogen dioxide (N02). These classifications were
based on existing air quality measurements, if available, or urban popu-
lation size, if suitable measurement data were unavailable. The purpose
-------
• MOUTT \ JACKSON I LAHIMt*
I «_
'PAWNEE
\ \1 ADAMS\I
sssrk \ 1 METRO
V soul otNVErt i i L. i M w
XDH i DENVER
.{--v—i-
«. N-" ARAPAt
DENVER
I C 0 M A N C H E I
f "T'StHT [pROWtHS •
—I I I |
_L \
BACA I
Sij**MMi*f / /t
rTrinida|r
/ FOUR CjORNERS
CO
-ti.
—4
FIGURE C-l. COLORADO AIR QUALITY CONTROL REGIONS AND STATE AIR POLLUTION CONTROL DESIGNATED AREAS
-------
348
of this action was to establish the levels of analysis necessary in the
preparation of SIPs. Priority I designation was assigned where pollutant
concentrations at certain levels above the NAAQS were measured. In the
absence of such data, AQCRs containing urban centers with 1970 populations
exceeding 200,000 were also included. For those pollutants designated
Priority III, no analysis was required for the SIP element from that AQCR.
Within the State of Colorado, only Metropolitan Denver Intrastate AQCR was
designated a Priority I. Specific pollutants so labeled were CO and QX-
In 31 metropolitan areas throughout the U.S., implementation of strin-
gent stationary source and vehicle emissions controls will produce reduc-
tions insufficient to attain the NAAQS. In those areas Transportation
Control Plans (TCP) are required as part of the SIP. The Metropolitan
Denver AQCR has been identified by EPA as subject to this requirement.
In further satisfaction of its charter, EPA is also charged with iden-
tifying geographical areas where the potential exists for violation of the
NAAQS in the years following initial attainment. For these areas, desig-
nated as Air Quality Maintenance Areas (AQMA), states must include within
their air quality implementation plans additional measures to insure con-
tinued adherence to the NAAQS. These additional measures comprise the Air
Quality Maintenance Plan (AQMP). The Metropolitan Denver AQCR is also
subject to this requirement for transportation-related pollutants, CO, QX,
and N02-
2) Transportation Regulations
In 1973 the Federal-Aid Highway Act added Section 109(j) to 23 U.S.C.
134, which directed the Department of Transportation to develop guidelines
to assure that highways constructed with Federal funds are consistent with
any approved plan for attainment of the NAAQS. On 26 November 1974, the
FHWA published final regulations setting forth these procedures.
Also pursuant to the 1973 Federal-Aid Highway Act, within each geo-
graphical area an agency must be designated as responsible for conducting
the "continuing, comprehensive and cooperative" transportation planning
-------
349
process. Such an agency is referred to in generic terms as a Metropolitan
Planning Organization (MPO).
In Colorado the Governor was responsible for the designation of MPOs
throughout the state. In the Metropolitan Denver region, the Joint
Regional Planning Program (JRPP) was appointed the MPO. A more detailed
discussion of the JRPP will be presented later in this section.
In order to assure conformity of transportation planning to federal
highway and air quality standards, the FHWA, through its 1974 guidelines,
requires that local highway agencies in cooperation with local planning
agencies establish a continuing review procedure with the local air pollu-
tion control agency. The purpose of this review procedure is to assess
the consistency of the transportation plan with the approved SIP. In the
Metropolitan Denver region, the Colorado Division of Highways (CDH) is the
local highway agency, the JRPP, as the MPO, is the local planning agency,
and the Air Pollution Control Division (APCD) of the Colorado Department
of Health is the local air pollution control agency.
The CDH is required to request the MPO (JRPP) to determine annually
the consistency of the transportation plan to the SIP. After review by
the APCD, the Regional Federal Highway Administrator, in consultation with
the Regional Administrator of the EPA, will annually assess the degree of
coordination in the planning process between transportation and air qual-
ity and review the determination on consistency between the transportation
plan and the SIP. The full review process is outlined diagrammatically in
Figure C-2, from Reference 1.
b. Joint Regional Planning Program (JRPP)
As a result of the 1973 Federal-Aid Highway Act, it became necessary
for the Governor of Colorado to designate for each region in the state a
Metropolitan Planning Organization (MPO). It was to be the responsibility
of these agencies to conduct the "continuing, comprehensive and coopera-
tive" transportation planning process mandated by that law (23 U.S.C. 134).
-------
\
Vtl
fiawroni to tht
Approval Land UM
ml Transportation
t
Incorporate
Raviiiont into
th» Approved
No
Updatad Land
UM and T rani-
port* t ion Plant
V
No
thiSIP
Rniwd Stata
Implamanuiion
Yn
.
Incorpi
Chanfl.
SIP
f
oral*
Tranwnit to MPO
tht Comment! (or
Thtir R*viiw »nd
Ev*luction
"^
Tranimit to
FHWA tor
Thtir Rtcordi
CO
en
o
FIGURE C-2. AIR QUALITY REVIEW PROCEDURE FLOW CHART
-------
351
Throughout the state, then-Governor John Vanderhoof, who had just
succeeded Governor John Love upon his resignation to assume federal re-
sponsibilities in Washington, D. C., appointed MPOs in conformity with the
guidelines provided. In all regions, with the exception of the Metropolitan
Denver region, he named as MPO the local council of governments. In the
Denver/Boulder region, however, he designated as MPO a hybrid agency called
the Joint Regional Planning Program (JRPP), consisting of the Denver Regional
Council of Governments (DRCOG), which he specified as the Policy Body, the
Colorado Division of Highways (CDH), and the Regional Transportation
District (RTD).
The structure of the JRPP was outlined in a Memorandum of Agreement with
working responsibilities included in an Operations Plan. Under federal guide-
lines, decisionmaking authority was vested in the Regional Director, consist-
ing of those individuals or individual as outlined in the Memorandum of
Agreement. In the JRPP the Regional Director contained representatives from
all three member agencies. Originally, membership consisted of the executive
directors of each agency. However, in February of 1976, an amendment to
the Memorandum of Agreement revised the membership, expanding it by four
members—the chairman of the state highway commission, the RTD chairman,
and two DRCOG officials who are also local elected officials. The latter
two officials were DRCOG Chairman Don DeDecker, a Lakewood city councilman,
and DRCOG Vice-Chairman James Nolan, a Denver city councilman. Since that
time, the final Regional Director configuration has consisted of seven
members, four from state agencies and three from the DRCOG.
The JRPP had been in existence prior to its designation as the Metro-
politan Denver MPO. It began formally on 16 April 1971, and the original
objectives of its program were: (1) to integrate regional transportation
planning with other elements of the comprehensive planning process; (2) to
meet the planning requirements of the 1962 Federal Aid Highway Act and the
1964 Urban Mass Transit Act; (3) to develop, update and adopt transportation
and land use plans; (4) to translate plans into priority programs; and (5) to
develop a continuing long range comprehensive and transportation planning
-------
352
capability within the DRCOG, to serve the future planning needs of both the
CDH and the RTD.
In October of 1973, the JRPP published a draft summary report prepared
by the staff of the DRCOG which set forth concepts for land use, highways,
and public transportation. These are presented, for the JRPP planning
area shown in Figure C-3, in Figures C-4 through C-6.
Having outlined above the general structure of the JRPP, it is now
important to examine its infrastructure as it pertains to air quality.
The FHWA, through its 1974 guidelines, requires for the Metropolitan Denver
region coordination in the transportation planning process between the CDH
and the JRPP, as the MPO. Major responsibility is placed on the CDH for
incorporating air quality considerations into transportation planning. In
cooperation with the MPO, CDH must insure that land use and transportation
planning pursuant to 23 U.S.C. 134 is coordinated with air quality planning
conducted pursuant to 42 U.S.C. 1857 (Section 108 of the Clean Air Act of
1970). A continual review procedure is also required to assess the consis-
tency of transportation plans with the State Implementation Plan (SIP).
In order to implement this review procedure, an Air Quality Technical
Advisory Committee (AQTAC) has been formed within the JRPP. This is possible
because CDH is itself a member of the JRPP. The committee consists of
representatives from EPA, FHWA, APCD, and the Regional Review Team (RRT).
The AQTAC should be regarded as an advisory committee to the RRT. How the
AQTAC fits into the JRPP organizational structure is shown in Figure C-7,
from Keferpnre 1. Functionally, the AQTAC serves in an advisory capacity to
the RRT. As such, its comments and advice need be considered by the RTT
as only informal and advisory.
The final air quality output of the JRPP transportation review is an
Air Quality Assessment Statement (AQAS). A number of criteria of consistency
with the SIP, all detailed in joint FHWA/EPA guidelines (Reference 2), must
-------
353
FIGURE C-3. JRPP PLANNING AREA
-------
354
2>
'-"") ;;.) '%
.:- ]( /-^••^_ "^ ' \£
(7 / •" \ , ;--.. " . „„/:
V.' ; . -i \ V 1 x ' • "i- * • "V .'
ft • !-•$ -^ -:y^
/>U .., "x . i -
j IOW DENSITY URBANIZATION
1 H'tiH DENSITY URBANIZATION
1 EMPLOYMENT OR TRANSPORTATION
ACTIVITY CENTER
~j PRIVATE & PUBLIC ACCCSS
•jj PRIMARY 6 SECONDARY PRESCRIPTION
3 NON-URBAN I'hE 2000
>V
FIGURE C-4. JRPP LAND USE PLAN
? 3'.L,_' ^K'ir3-^-^;:- ::;€?^C-^^:-^
:. !- 1- *'.,':•: c y .' £&. . >v,;-"-"V^.: . / -\ />-j^',w.--. • . o-;--,
-=--=• ,«»;«;,'.'• •;-'-. n & -trltf -v (--- • - •••'' .--.^ < • ' -•.'•'-^'\ >^
,--. • --W •,- • „ j U ,!-, -' , l r : .. , •-' •. j-x" •- ". •• - *-".••-.' i.,*
.-::N
-------
LIMITED ACCESS FACILITY
PRINCIPAL ARTERIAL
ADOPTED
DENVER REGIONAL CQUMCIl flf fiOVERMMTS
COLORADO OErARTMENT OF HIGHWAY!
RESIDUAL TflAtSPORTATIOI OIJ1RICT
-------
RAPID TRANSIT FACILITY
BUS ROUTE
I I BUS SERVICE AREA
•US ROUTES ARE SHOWN FOR SCHEMATIC
PURPOSES ma. THEY ARE nor PART OF THE HAD.
ADOPTED
DENVER REGIONAL COI/KCU OF GOVERNMT1
COLORADO DEPARTMENT OF HIGHWAYS
REGIONAL TRANSPORTATION DISTRICT
FIGURE C-6. JRPP PUBLIC TRANSPORTATION PLAN
rW.-V-'W— ——t*i /--""-cj
-------
357
Denver Regional
Council of
Governments
Transportation
Advisory
Committee
Regional
Director
Citizens
Advisory
Committee
Env i ronment a1
Protection
Agency
Federal
Highway
Administration
Air Pollution
Control
Division
Air Quality
Technical Advisory
Committee
Regional
Review
Team
Direct line of Communication
Advisory line of Communication
Boulder
Technical
Committee
FIGURE C-7. JRPP ORGANIZATIONAL STRUCTURE—AIR QUALITY
-------
358
be addressed by the JRPP, as the MPO. These include SIP investigation
strategies, such as stationary hydrocarbon control, gasoline limitation,
bus/carpool lanes, parking limitation, and mass transit improvements. The
final form of the AQAS, arrived at by the review procedure detailed in
Figure C-2, is subject to review by the. APCD.
c.
Denver Regional Council of Government (p_RCOG_)_
When the JRPP was named by the Governor as the MPO in the Metropolitan
Denver region, pursuant to the 1973 Federal-Aid Highway Act, he designated
the Denver Regional Council of Governments as its Policy Body. As such
it has considerable influence on the transportation planning process. In
practical terms, because of the size of its staff and the resources they
command in technical training and access to computing facilities and math-
ematical modeling tools, the DRCOG plays a central role in all JRPP planning
efforts. Due to the intent of its original charter in 1971, the JRPP itself
has few formal staff members not simultaneously serving in other capacities
within member agencies.
The composition of the DRCOG underscores the uniqueness of its position.
Among those agencies comprising the JRPP, only the DRCOG has the direct,
majority participation of locally elected officals. Members of the DRCOG,
all elected officials in their own right, are appointed to the Council from
the city councils and boards of supervisors representing the cities and
counties comprising the Metropolitan Denver region.
Numerical composition is so structured that no individual city can
dominate voting on the Council. Similarities both in attitude and life-
style, however, would suggest certain regional communities of interest.
Issues of importance to the downtown, highly urbanized portion of Denver,
for instance, may not be emphasized to the same degree in the smaller, more
surburban areas surrounding the city. Members from the outlying suburbs,
together constituting a majority on the Council, could be expected to
-------
359
represent actively the interest of their constituents. To find such an
emphasis in the planning process conducted by the DRCOG would not be
unexpected.
The DRCOG is headed by a chairman and vice-chairman, as well as a six-
person executive committee. Currently, Don DeDecker, a city councilman from
Lakewoocl, is serving as Chairman, and James Nolan, a Denver city councilman
is the Vice-Chairman. Robert Farley, as Executive Director, heads the DRCOG
staff, assisted by David Klotz, the Deputy Executive Director. Meetings,
held monthly, are attended by a number of Council members. For the meeting
held on 20 October, for example, more than twenty members were present.
The part played by the DRCOG in the regional planning process has
been institutionally endorsed not only by its own Articles of Association
but also by federal and state law as well as by contractual agreements
with the U.S. Department of Housing and Urban Development, the Department
of Transportation, and the Environmental Protection Agency. In its planning
role the DRCOG has addressed the forecasting of regional population and its
allocation, land use, and transportation network planning. Though no
specific sanctions exist requiring the conformance of local communities to
DRCOG adopted plans, considerable indirect leverage towards compliance does
exist. These same forecasts are used by transportation, water supply, and
wastewater treatment agencies in their facilities and service planning.
An individual community in search of state or federal assistance, whose
planning is not consistent with DRCOG-approved plans, may find such fund-
ing denied. The level to which DRCOG regional planning extends is indicated
in Figure C-8, taken from Reference 7.
To date the DRCQG has adopted a number of the component parts of its
regional development plan. Since population projections serve as the basis
for nearly all piajor planning decisions, initial attention was turned to a
regional population forecast. Developed in 1971-72 and subsequently adopted
on January 17, 1973, such a forecast of population level was used as the basis
for land use, highway and transit plans. Adoption of the forecast marked
official endorsement of a year 2000 population of 2,350,000 for the five-county
-------
Appraise
Population
Policy
Forecast
Define Local
Growth
Alternative
Define
Regional
Growth
Alternative
Local Review
of Alternatives
Establish &
Agree on
1975 Pop'n
Estimates
Water Quality
Assessment
of Alt Plans
Analyze Reg'l
Growth
Allocations
Stage Local
Dev Plans
Develop &
Review
Staged
Subarea Plans
AUG 75
SEP/NOV 75
NOV 75/JAN 76
JAN/APR 76
APR/JUN 76
FIGURE C-8. SUBAREA POPULATION ALLOCATION PROCESS
-------
361
Denver Metropolitan area. Having agreed to a population level, efforts
by the Council to agree on where that population will live have been
conducted. Several iterations on an allocation plan have been performed
by the DRCOG staff since the initial Council directive in October of 1975.
Agreement, however, on a subarea allocation plan was finally achieved with
Resolution 18, 1976, adopted by the full Council on 18 August. The specifics
of these population level and allocation plans will be discussed in detail
in Section 3.
The population level forecast adopted in early 1973 was subsequently
used to develop regional transit and land use policy. Incorporating the
forecast, the DRCOG formally adopted on 19 December 1973 the Regional
Land Use, Highway and Public Transportation Plans. The land use and high-
way plans became JRPP policy when, in January 1974, they were endorsed by
both the Colorado Division of Highways and the Regional Transportation
District.
Insofar as air quality planning is concerned, DRCOG, acting as the
policy arm of the JRPP, plays a major role in the joint CDH/MPO assessment
of transportation plan consistency with the SIP. The review procedure,
described in the discussion of the JRPP presented earlier in this section,
draws heavily upon the staff resources of the DRCOG. Among their responsi-
bilities, DRCOG staff include the development and application of computer-
ized transportation models. To enhance and refine the accuracy of the level
and distribution of vehicle-miles-traveled (VMT) projections has required
continuing evolution in the modeling process. A detailed discussion of
transportation modeling progress has been deferred until Section 3. It is
sufficient at this point simply to recognize that any air quality assess-
ment must be regarded as acutely sensitive to the certainty with which VMT
projections are viewed.
)ijmij^
The Colorado Department of Highways, Division of Highways (CDH), is
another of the three member agencies comprising the JRPP. The Department
of Highways is currently headed by its Executive Director, Jack Kintslinger.
-------
362
As a basic charter it is charged with the primary responsibility for
initiating and carrying forward the construction or reconstruction
of highway sections. The Federal-Aid Highway Act of 1973 as implemented
in FHWA guidelines established the additional requirement that the CDH
take steps to assume consistency with the SIP for any highway construction
involving federal funding. Furthermore, in cooperation with the JRPP,
as MPO, the CDH must insure this planning compatibility by entering
into an annual review procudure with the Air Pollution Control District.
In fulfilling their air quality responsibilities, the staff of the
CDH has also acquired and supports the use of the SAI urban airshed
modeling program. Most recently, the SAI model was used on a regional
basis, in conjunction with the California Line Source Model (CALINE2)
for microscale analysis, to develop the air quality assessment component
of the Detailed Assessment Report for the Interstate 470 Project, a pro-
posed circumferential highway around Southwest Metropolitan Denver.
The 1-470 Project has long been a part of CDH plans. In 1972 a Final
Environmental Impact Statement was submitted for FHWA review. The con-
clusions of this review (September 1974) were that significant deficiencies
and questions remained regarding air quality, alternative modes of trans-
portation, alternative highway alignments, and effects on land use. A
revision of the EIS was required by the FHWA. An 1-470 Ad Hoc Commission
was set up by the Governor to study and make recommendations on the 1-470
question.
The first major phase of that review, the Preliminary Screening Process,
was completed on 30 January 1976. The Report evaluated eleven highway
alternatives based on 23 environmental and transportation assessment criteria
The second major phase, the Detailed Assessment, then analyzed the remaining
five feasible highway alternatives. Originally scheduled to be completed
in May, 1976, the final date was extended to August by the passage of the
1976 Federal-Aid Highway Act, which enabled the consideration of federal
funding for other road system alternatives.
-------
363
Similar reviews must be undertaken by CDH for all major highway
projects. The air quality assessment component can be expected to play
a similarly significant role. As a consequence the CDH staff will main-
tain air quality analysis capability.
e- Regional Transportation District (RTD)
The Regional Transportation District is one of the three member
agencies comprising the JRPP. Its enabling legislation provides that RTD
shall develop, operate, and maintain a mass transportation system for the
District, which presently is comprised of the City and County of Denver,
Boulder and Jefferson Counties, the western portions of Adams and Arapahoe
Counties, and the northeastern portion of Douglas County. Jurisdictional
boundaries are as shown in Figure C-9, from Reference 12.
The enablina statute provided for the development and adoption of a
comprehensive transit plan. Also, in addition to providing for the col-
lection of certain tax revenues in support of transit activities, it called
for a referendum on bond issuance authority. Such a referendum was held on
September 7, 1973, with the voters of the region authorizing the issuance
of up to $425,000,000 in sales tax revenue bonds. They also approved a
ballot issue which specified the development of a multi-modal mass transport-
ation system for the region. The Public Transportation Plan for the Year
2000, as adopted and amended by the JRPP, provided the framework within
which RTD efforts were directed.
The long-range plans of the RTD contemplate a high level of bus service
between communities within the district, and community level circulator
service designed to provide a collection and distribution complement to
the rapid transit portion of an integrated multi-modal system. The bus
fleet presently operated by RTD consists of 517 vehicles. Accounting for
the orderly retirement and replacement of vehicles, the RTD plans to expand
the fleet by approximately 28 vehicles each year, bringing the total to 799
by 1985. Funding assistance has been sought from the federal government
through UMTA.
-------
North-South Rapid Transit Corridor
Preferential Treatment Existing
Preferential Treatment Proposed
RTD park-n-Ride Sites 1976
RTD park-n-Ride Sites 1977-78
Joint Use park-n-Ride Sites
FIGURE C-9. RTD JURISDICTION BOUNDARIES
-------
365
In addition to bus service, the RTD has been considering for several
years various alternative fixed rapid transit networks. On promising
heavy traffic corridors, a number of candidate systems were investigated,
including light rail transit (LRT), conventional rail transit, automated
rapid transit (ART), anc' personalized rapid transit (PRT). The RTD board on
12 February 1976 acted to endorse an LRT plan on the most promising guide-
way corridor. The selected network, shown in Figure C-10 from Reference 13,
consisted of 14 stations extending over a north-south alignment 22 miles long.
The system cost for the RTD-selected alignment, shown in Table C-l from
Reference 12, totals $492,000,000 in December 1974 dollars. Of this amount
the fixed facility portion is $391,000,000. Federal funding assistance
was sought by RTD through UMTA.
TABLE C-l. RTD-SELECTED RAPID TRANSIT PROJECT COST
DOLLARS
(MILLIONS)
Guideway 153
Trackwork 28
Stations 64
Station Parking 9
Control System 23
Electrification 28
Yard & Shops 34
Adm. & Control Facility 10
Right-of-Way 42
Sub-Total 391
Vehicles 35
System Verification Test 11
Program Management 55
Sub-Total 101
(December 1974 Dollars) TOTAL 492
After review of the RTD proposal, on 29 June 1976 in a letter to
John Crowley, Chairman of the RTD, Robert Patricelli, the Administrator
for UMTA, informed the RTD of UMTA's decision concerning federal funding
-------
& Storage Yard
% i| I tisjr
North South Rapid Transit Corridor—
{ Selected February, 1976)
Stations^)
Rapid Transit Alignment
Figure 1.2
FIGURE C-10. RTD-SELECTED RAPID TRANSIT PLAN
-------
367
of the 22-mile light-rail line. The conclusions were threefold: first,
that it was premature for the Federal Government to provide engineering
or construction funds for the development of the proposed line; second,
that UMTA was prepared to loan advance right-of-way acquisition funding
to allow RTD to hold such property until the need for a fixed guideway
system might become more pronounced; and, three, that UMTA was willing to
consider substantial funding for future bus system improvments. The action
detailed in this letter can be expected to cause a substantial reassess-
ment by the RTD of their regional transportation plans. A reorientation
towards an all-bus transit system seems a likely course of action. The
amount of funding made available under the third point above has been said
to be in the neighborhood of $200,000,000 or less than half the RTD request
for the light-rail system.
A few words are in order about RTD planning efforts having either
direct or indirect effect on regional air quality. As will be explained
in more detail in Section 3, several iterations of population allocation,
land use, and transportation plans exist within various agencies, each
an evolutionary extension of their predecessor. In conducting planning
efforts, however, planners have no choice but to use those versions currently
most accepted at the time the study is conducted. These planning efforts,
consequently, are accurate only to the extent their assumptions still agree
with new population, land use, and transportation plan assumptions.
System planning by the RTD has been done using a relatively early
population allocation version. Of the three major population allocation
versions, Cycle 3 and Cycle 4 (both generated using a computerized popula-
tion allocation model called EMPIRIC) and the Subarea Allocation Plan
recently considered by the JRPP/DRCOG, the version known as Cycle 3 was
used by the RTD. Changes in allocation have since occurred, particularly
in the southwestern portion of the Denver region (southeastern Jefferson
County).
The design of any regional transit system and its ameliorating effects
on air quality is sensitive to the highway system on which that transit
-------
368
system must operate. The JRPP has adopted a Year 2000 Transportation Plan
which details their projected roadway network. However, in the event
that the JRPP plan is substantially overbuilt and that the funding program
necessary for its construction is an unrealistically ambitious one, RTD
planning has been conducted using a more modest highway system known as
the Existing plus Committed (E+C) network. Basically, it consists of
those highway elements already built or for which funds have already been
committed. This is an inconsistency with CDH highway planning, which,
though also assuming Cycle 3 population allocation, assumes the Year 2000
JRPP highway plan. The extent of these differences on plan compatibility
remains unanswered in any definitive way. It was, in fact, cited by the
FHWA as a factor in their recent alteration of JRPP certification as MPO.
The ultimate effects on air quality projections for the Denver region
also remain unassessed.
f. Air Pollution Control Division (APCD)
Under the 1974 federal guidelines established by the FHWA pursuant to
the 1973 Federal-Aid Highway Act, the Air Pollution Control Division plays
an important role in verifying the consistency of regional transportation
plans to the State Implementation Plan for air quality attainment and
maintenance. It is charged with reviewing on a continual basis the trans-
portation plans of the CDH and the JRPP, as MPO.
In terms of jurisdictional responsibilities the APCD, a part of the
Colorado Department of Health, is responsible for enforcement of the regul-
ations and standards as developed and adopted by the Air Pollution Control
Commission (APCC). The full regulatory framework of Colorado's air quality
program was established by the Colorado Air Pollution Control Act of 1970.
The administrative structure set up by the Act consists of the APCC, with
broad rulemaking authority, the Air Pollution Variance Board, with rule
suspension authority, and the APCD, with enforcement capability. The 1970
enabling legislation organized all three within the Colorado Department of
Health.
-------
369
Among their responsibilties, it is the function of the APCD to enforce
stationary source controls; to reduce or plan for the eventual reduction
of mobile source emissions; to monitor pollutants; and to allocate air
quality program resources and plan for the attainment of the NAAQS. The
APCD is provided an engineering staff to support the enforcement effort.
The surveillance of statewide air quality conducted by the APCD repre-
sents the rtajor source of air quality data in the Metropolitan Denver region.
Because of their Priority I designation in the Denver AQCR, emphasis is
placed on carbon monoxide (CO), oxidant (03), and particulates. The basic
monitoring network maintained by the APCD is shown in Figures C-ll and C^-12,
from Reference 16.
Over fifty percent of the population in Colorado resides within the
Denver AQCR. The high vehicle usage, combined with three major power
plants, two oil refineries and a number of other medium sized to large
industrial sources, results in a serious air quality problem. A breakdown
of both the size and source of pollutants within the Denver AQCR is shown
in Table C-2, from Reference 16.
TABLE C-2. DENVER AQCR - SIGNIFICANT POINT SOURCES
SOURCE TYPE
Agricultural
Mineral Processing
Energy
Industrial
Miscellaneous
NUMBER
9
46
15
38
7
115
TOTAL EMISSIONS, EACH POLLUTANT
TONS/YEAR
Particulate
SO2
CO
NOX
HC
7,066
38,852
61,171
49,112
5,193
The responsibility has also been assigned by the Air Pollution Control
*•
Commission to the APCD to develop and maintain the capability to model
observed air quality readings. It was in this context that the APCD
acquired the SAI urban airshed modeling program. It has been most recently
-------
1. UELBY
78th and Steele
2. OVERLAND PARK
South Huron and Evans
3. ARVADA
57th and Garrison
4. C.A.R.I.H.
21st and Julian
5. NATIONAL JEWISH HOSP.
Colfax and Colorado
6. C.A.M.P.
21st and Broadway
DENVER AIR MONITORING NETWORK
012 miles
oo
~j
o
FIGURE C-11. APCD DENVER AIR QUALITY MONITORING NETWORK
-------
371
0 CArlC S t ri t i on
I Hul 1 Photo
2 Sl.ltc llr.il tn Bldq.
3 Gdlcs Rubber Ca.
k Scwor Plant
5 Schnul Admin. 3ldg
7 Aurora
3 Ad.-mi c i ly
9 EnqIcKood
I I Cherry Creek 0*m
15 Arv.-ula
16 Coldrn
19 Snuldcr
22 tomjinonl
56 Rotky Flats
S3 3ri
-------
372
used by the APCJ staff in an analysis of the Denver-Boulder Air Quality
Maintence Area. The purpose of the study was to define the existing and
projected air quality in the Metropolitan Denver region, and to determine
whether the NAAQS will be attained in Denver by 1985. This assessment
was performed for three years: 1974 (the baseline year), 1980 and 1985.
For the latter two years, results of the SAI model were used for short-term
averaging (i.e., 1 hour and 8 hours), and the Climatic Dispersion Model
(COM) was used to evaluate long-term averages (i.e., annual).
The pollutants specifically addressed were CO, 03, N02, and particu-
lates. The NAAQS for these and other pollutants are given in Annex C-2
of this report. The major assumptions in the study are outlined in Table
C-3, with the final results shown in Tables C-4 and C-5. All of these are
taken from Reference 16.
TABLE C-3. MAJOR ASSUMPTIONS IN APCD STUDY
Federal Motor Vehicle Control Program Assumed to be on
schedule and to be Effective at this Altitude (AP-42)
Automotive Inspection/Maintenance not Assumed
City of Boulder not Included in Analysis Concentrations
Calculated
Ozone from Set of Reactions Derived for Los Angeles, Calif-
ornia
Highway and DRCOG Traffic and Population Inputs are Appro-
priate
No Small Scale Localized Analysis will be Performed
Phase II Vapor Recovery Strategy not Assumed
The emissions inventory developed by the APCD for use in the SAI and
COM models is presented in Table C-6, also from Reference 16. Both daily
and annual figures are shown.
-------
373
TABLE C-4. REGIONWIDE DENVER AIR QUALITY
ATTAINMENT IN APCD STUDY
Pollutant
Status to 1985
Carbon Monoxide: 1-hour
8-hour
Ozone
Particulate: 24-hour
1-year
N02
Non-attainment
Non-attainment
Non-attainment
Non-attainment
Non-attainment
Maintenance Problem
TABLE C-5. DENVER AQMA AREA OF VIOLATION
IN APCD STUDY
SQUARE MILES OF VIOLATIONS+
POLLUTANT
1975
1980
1985
Carbon Monoxide : 1-hour*
8-hour
Ozone : l-hour++
Particulates : 24-hour
Annual
NO2 : Annual
38
170
740
776
675
1
41
179
662
900
900
11
6
26
582
900
900
16
+Maximum Hour Violations
*Average of 0700 and 0800 Violations
++Study Day Violations
(.Total Study Area 900 Square Miles)
-------
374
TABLE C-6. DENVER DAILY AND ANNUAL SOURCE CONTRIBUTIONS IN 1974
(a) Daily
SOURCE CATEGORY
Carbon Monoxide : *
Automotive
Space Heat
Point Sources
Hydrocarbons : **
Automotive
Space Heat
Oil Paint
Gas Stations
Cleaners
Incinerators
Point Sources
Airports
TOTAL TONS /DAY
2925.0
2.9
167.6
3095.5
199.18
0.48
2.08
11.67
4.94
0.11
13.14
5.92
237.52
% CONTRIBUTIONS
94.5
0.1
5.4
100.0
83.9
0.2
0.9
4.9
2.1
0.0
5.5
2.5
100.0
*Based on winter emissions distributions
**Based on summer emissions distributions
(b) Annual
SOURCE CATEGORY
Parti culate:
Space Heat
Automotive
Airports
Construction
Street Sanding
Point Sources
N02:
Space Heat
Small Points
Automotive
Airports
Point Sources
TOTAL TONS/YEAR
511
4,213
63
8,730
15,857
5,981
35,355
5,348
87
31,646
1,552
33,773
72,406
% CONTRIBUTIONS
1.4
11.9
0.0
24.7
44.9
16.7
99.6
7.4
0.0
43.7
2.1
46.6
99.8
-------
375
A final word about the APCD study assumptions should be injected.
As mentioned earlier, several iterations have been developed for popu-
lation allocation. It appears that the APCD staff used EMPIRIC Cyle .4
projections to obtain their population data. This is at variance with
most transportation modeling efforts done by the CDH and the DRCOG. Done
at earlier times, these studies incorporated EMPIRIC Cycle 3 population
allocation forecasts.
-------
376
3. OVERVIEW OF PLANNING ISSUES
a. Planning Issues and Planning Agencies
It is not difficult to secure from most planning agencies agreement
on the identity of the fundamental issues of planning. Population and
land use forecasts intertwine to form the core of the planning process.
Implied in each are basic decisions about human priorities. Growth--
how much and where it will be distributed—dominates as a precursor to
any planning study. The construction of facilities to provide public
services, such as transportation, water supply, and sewerage treatment,
both paces and responds to the level, distribution and staging of popu-
lation growth. Sound planning can encourage the development of a sound
regional ecomomic base which, by accommodating demands for new jobs, pro-
ducts, and services, creates a tax base healthy enough to support strong
and responsible community programs.
The difficulty in any planning process, however, lies not in assess-
ing the broad issues but rather in obtaining a consistent set of planning
assumptions about the specific ones. The problem is aggravated by the
evolutionary nature of planning. In response to public reexamination,
planning is usually iterative. Specific studies, however, must draw on
the results of predecessor studies. Planners often have no choice but
to accept the best set of assumptions available at the time of the study.
Consequently, the results can be regarded as accurate only to the extent
the study's assumptions still agree with new population, land use and
transportation plans. An ongoing assessment of consistency is thus
essential in any area in which planning is subject to volatility.
In part due to the above reasons, there is a proliferation of planning
alternatives, each of varying vintage, within the agencies responsible
for planning in the Metropolitan Denver region. Though the DRCOG has
agreed since January 1973 on a year 2000 population forecast of 2,350,000
for the five-county Denver area, several attempts have been made to allo-
cate the new growth to individual communities. In 1972 the EMPIRIC
-------
377
Activity Allocation Model was developed for the Denver region to aid in
making these allocations. The two most recent applications of the model,
referred to as Cycle 3 and Cycle 4, have been used by different agencies
for different purposes.
Of the three population allocation forecasts, Cycle 3, Cycle 4 and
the Subarea Allocation Plan recently considered by the JRPP/DRCOG, the
earliest of these, Cycle 3, was used by both the CHD and the RTD for their
transportation planning, although each agency used its own version.
Within the Health Department, the APCD has used Cycle 4 for their trans-
portation/air quality studies. The area source emissions inventory used
by the APCD seems to have been based on Cycle 4. This appears to have
been the chief if not sole formal use so far of Cycle 4. Since Cycle 3
was not available for transportation modeling prior to the adoption in
early 1974 of the JRPP transportation plan, it has never been formally
agreed to by the JRPP or the DRCOG, acting as the JRPP Policy Body. The
DRCOG, however, acting on its own, will ultimately use Cycle 4 for a
revisiting of the JRPP transportation plan.
Water quality planning done pursuant to the 1972 Clean Water Act
is required to address planning at the regional, basin and facilities
level. The designation of these programs are 208 (regional), 303 (basin),
and 201 (facilities). It appears that Cycle 4 population allocation
forecasts were used for 208 program planning. Cycle 3, however, was used
in the preparation of the Interim Regional Water Quality Management Plan.
Land use planning is intertwined with population level and allocation.
The JRPP land plan, adopted in early 1974, underlies both Cycle 3 and
Cycle 4 through specification of land availability and development para-
meters. The DRCOG, however, is undertaking a review of land use allo-
cation, seeking a more specific restatement in Hght of new population
allocation versions. It seems unlikely that such review will lead to a
new land use plan in the time frame of this Overview EIS.
-------
378
Transportation planning is subject to two additional uncertainties
beyond population and land use considerations. The first includes the
difficulties of forecasting the highway network at a particular time in
the future. The second refers to the limitations of the transportation
modeling process. In early 1974 the JRPP adopted a Year 2000 transportation
plan. This plan became embedded in the transportation planning efforts
conducted by both the DRCOG and the CDH. In the event that such a net-
work proves substantially overbuilt and that the funding program necessary
for its construction is an unrealistically ambitious one, RTD planning
has been conducted using a more modest highway system referred to as the
Existing plus Committed (E+C) network. Basically, it consists of those
highway sections already built or for which funds have already been
committed. Though RTD and CDH transportation planning both assume Cycle 3
population allocation, a basic inconsistency is introduced through their
differing assumptions about highway network configuration.
As mentioned above transportation planning is also subject to uncer-
tanties introduced through the limitations of transportation modeling.
These models can be viewed in four segments: network configuration, trip
generation, trip distribution and modal split (i.e., auto vs. rapid
transit), and traffic assignment to highway links. Assumptions in any of
these areas cascade through the modeling process, affecting the level and
distribution of VMT projections. There have been three generations of
transportation models in use by various agencies at different times. An
idea of the magnitude of impact of changing modeling assumptions can b'e
obtained by comparing total regional daily VMT projections in the year
2000. The second generation model projected a daily VMT of 38,000,000;
the third generation model predicted about 30,000,000. This represents a
20 percent reduction in VMT due simply to changing behavioral assumptions
about trip distribution and capacity restraint.
Transportation planning done by the DRCOG and the CDH was accomplished
usinq second generation transportation models. Very recently, however,
-------
379
the DRCOG staff has been applying the third generation model using it
in their efforts to revisit the JRPP transit plan. The RTD has also
used the second generation model, although the trip table differs from
the one used by the DRCOG/CDH because of their use of the E+C highway
network rather than the JRPP plan. A transportation modeling run using
the third generation model and the JRPP plan trip table does not appear
to have been undertaken.
Underlying the above planning permutations are assumptions of fun-
damental importance in the planning process. A careful analysis of their
impact on regional air quality requires an understanding of their basic
form as well as detailed examination of the mutual consistency of agency
plans. In order to further such understanding, the remainder of this
section will be spent in detailing population (level and allocation),
land use, and transportation (network configuration and modeling) planning.
b. Population Planning
Assumptions about growth, both as to its size and distribution, are
central to the planning process. The projections that result serve as
the basis for nearly all major planning decisions. As measures of the
size, staging, and distribution of regional population, they determine
the level of demand for urban services, such as highways, rapid transit,
water services, and sewerage treatment. Growth in population provides
the underlying justification for facilities construction or expansion.
It will be the purpose here to explore various aspects of the popu-
lation forecasting and allocation process. Forecasts of population level
will be explored first. Then the discussion will address the problem of
allocating to the various communities the incremental gain in population
due to growth.
1) Population Level
Recognizing its basic importance, the staff of the DRCOG has invested
a sizable effort over the past five years in developing for the Metropolitan
-------
380
Denver region a population level forecast through the year 2000. It is
upon this forecast level of 2,350,000 that all subsequent planning in
other agencies has been based.
Several forecasting techniques were considered as candidates by
the DRCOG. Direct forecasts are based on historical data, such as fertil-
ity rates, birth and death rates, and in-migration levels. The method
considers the annual growth in population to be the sum of two components,
one proportional to the current level of population, i.e., the net of
births over deaths, and the other equal to the net migration into the area.
Indirect forecasts usually relate population growth to other economic,
social, and political indicators. Econometric methods are frequently
used to relate past growth to several key explanatory indices, such as
employment or personal income.
Summation of individual local forecasts is another method occasionally
used. Such a method rarely maintains a regional perspective. Local com-
munities tend to be optimistic when estimating their share of future
growth, usually underplaying or ignoring the effects of competition from
other communities for that same growth. As a result such forecasts tend
to be higher than with other methods. The sum of local forecasts in the
Denver region, for instance, originally totaled over 4,000,000 for the
year 2000.
Saturation forecasting is still another method of projecting popula-
tion. It is frequently applied by planning engineers to provide facil-
ities designs sufficiently large to accomodate growth in the event it
exceeds more realistic forecasts. Use of such methods is based on the
assumption that it is less costly to overbuild initially than to extend
capacity later in the event original forecasts prove conservative. In
the past, however, these methods have led to overbuilding and underutil-
ization of resources, as well as to inducing even higher growth than would
otherwise have occurred.
-------
381
After consideration of the above methods, the staff of the DRCOG
chose the cohort-survival method, a direct forecasting technique. Cohort-
survival decomposes growth into two components: (1) net natural increases,
which is the net of births over deaths, and (2) the net of in-migration
over out-migration. Assumptions are made concerning future birth/death
rates and migration levels for each 5-year age group of the population.
The method can in this way account for differences in the various age
groups. Death rates, for example, would be much higher in the 65-70 age
group than in those 30-35.
An application of the process can be outlined by the following
steps, as indicated in Reference 17. First, the base year population,
in 5-year age groups, is multiplied by the assumed death rates for each
age group. These deaths are then subtracted from the base year population.
Second, the net migrants for the time interval, by age group, are added
to or subtracted from the surviving population. Finally, births are
determined by considering males and females in the 15-44 age groups as
potential parents. This population fraction is multiplied by an assumed
birth rate for each age group, producing the expected number of births
during the 5-year period. These births are then added to the surviving
population and in-migrants to produce the total population at the end
of the interval. Presumably, surviving members of each age group graduate
to successive age groups as each iteration progresses.
It is clear from the above description that application of the cohort-
survival forecasting technique embodies an extensive set of assumptions
about the stratification of birth, death and migration variables. To
simplify discussion, however, these can all be distilled down into two
basic input variables: a single composite rate of net natural increase,
and a single general in-migration level.
It was these two composite parameters that the DRCOG staff used as
independent variables in preparing alternative population forecasts. The
eight alternative cases considered in their analysis are outlined in
Table C-7, from Reference 17, along with relevant assumptions and results.
-------
TABLE C- 7. DRCOG POPULATION FORECASTING ALTERNATIVES
ALTERNATIVE POPULATION FORECASTS
Case
I
II
in
IV
V
VI
VII
VIII
Annual 1970-1980 Increase
Net Natural Migration
Current Rate 1960's Level
in 1975-80
Current Rate Current Level
Current Rage Current Level
Current Rate Current Level
Current Rate Current Level
Current Rate Los Angeles
1943-1953
Current Rate Current Rate
Declining to 1960's Level
zero by 2050 in 1975-80
Annual 1980-2000 Increase
Net Natural Migration
Current Rate 1940-1970 Level
Current Rate 1960's Level
Current Rate 1960's Level
Weighted by
U.S. Pop.
Current Rate Current Level
Historical Rate Current Level
Current Rate Los Angeles
1953-1970
Current Rate Current Rate
Declining to 1940-1970 Level
zero by 2050
Population
(millions)
1980 2000
1.6 2.2
1.8 2.4
1.8 2.5
1.8 3.0
1.8 3.3
2.0 3,9
1.8 4.3
1.6 2.1
Simple
1970-2000
Growth Rate
2.5%
3.2%
3 . 5%
k
4.9%
5.5%
7.2%
8.4%
2.4%
Net Natural: Current Rate (8.5/1,000 persons) Migration: 1960's Level (15,700 persons)
Historical (1900-1970) Rate (16.3/1,000 persons) 1940-1970 Level (15,100 persons)
Current (1970-1973) Level (42,000 persons)
Current Rate (34/1,000 persons)
SOURCE: Denver Regional Council of Governments
-------
383
Alternative forecasts for the year 2000 are shown in Figure C-13, also
from that reference.
Of these alternatives, Case I was chosen by the DRCOG to represent
the baseline forecast. It assumes that the high level of migration
experienced by the Denver region in the period 1971-73 will not continue
through the year 2000, but that migration will approach the levels of
the 1960's (15,700 annually) during 1975-1980 and the levels of the
30-year (1940-1970) average (15,100 annually) during 1980-2000. Case I
also assumes that the net natural rate of population increase will con-
tinue at the 1972 rate (8.5 persons/1000 total population annually)
through the year 2000.
These assumptions lead to a population forecast of 2,175,000 in the
year 2000. Because of the sensitivity of this result to the above stated
assumptions, it is instructional to compare the net natural increase rate
and in-migration level with historical data available both before the
1972-73 study period and after (1973-1975). In Table C-8, the historical
net natural increase rates are presented, with the birth and death rates
shown graphically as in Figure C-14. Net migration levels are shown in
Table C-9. All of these tables and the figure are from Reference 17.
Two interesting trends have exhibited themselves since the DRCOG
study was done. First, as shown in Table C-8, the net natural increase
rate declined from 8.5 per 1000 in 1972 to 6.3 per 1000 in 1974. This
is consistent with a downward national trend in fertility rates. It is
unclear, however, whether this is a true long-term phenomenon or simply
a temporary dip due to the increasing trend of families towards later
child-bearing. To the extent this is a true decline in birth rate, the
DRCOG forecast may be proven to be hiqh.
The second trend is the observed decrease in migration from the
elevated levels of the early 1970's. From a high of 61,400 in 1972,
migration into the Denver region has decreased. Although not noted in
-------
YEAR 2000 SMSA POPULATION
(IN MILLIONS)
_i ro co .tk
• • • •
o o o o
)
1
co
00
-P.
II
III
IV
VI
VII
VIII
ALTERNATIVE CASES
SOURCE: DENVER REGIONAL COUNCIL OF GOVERNMENTS
FIGURE C-13. DRCOG ALTERNATIVE POPULATION FORECASTS FOR THE YEAR 2000
-------
385
TABLE C-8. HISTORICAL NET NATURAL INCREASE RATES IN POPULATION
(rate per 1,000 population)
DENVER UNITED
YEAR REGION* COLORADO STATES
1960 17.2 15.8 14.1
1965 11.4 10.5 10.3
1970 11.0 10.9 8.9
1971 9.6 9.9 8.0
1972 8.5 8.6 5.7
1973 7.5 8.0 5.2
1974 6.3 7.1 4.7
* Rates for 1960 and 1965 are for the 5-
county urbanized area, rates for 1970-
74 are for the 8-county region.
SOURCES: Colorado Department of Health
U.S. Bureau of the Census
Denver Regional Council of
Governments
-------
386
•0
CO
X
I-
tr
m
O
H
O
Q.
CO
tr
UJ
Q.
30 r
25
20
15
10
1960
BIRTHS
DEATHS
1965
i
1970
1974
YEAR
NOTE: 1960 & 1965 FIGURES FOR 5-COUNTY URBAN AREA
1970-74 FIGURES FOR 8-COUNTY REGION
SOURCES: COLORADO STATE HEALTH DEPARTMENT
DENVER REGIONAL COUNCIL OF GOVERNMENTS
FIGURE .G-14. HISTORICAL BIRTH AND DEATH RATES IN THE DENVER REGION
-------
387
TABLE C-9. HISTORICAL NET MIGRATION IN THE DENVER REGION
PERIOD
AVERAGE
ANNUAL
IN-MIGRATION
RATE PER
1,000
POPULATION
PERCENT
OF TOTAL
POPULATION
CHANGE
1940-1950
1950-1960
1960-1970
1970-71
1971-72
1972-73
1973-74
1974-75
10,700
18,800
15,700
29,500 23.5
44,800 34.3
61,400 44.8
47,600 33.2
32,500 21.8
64%
59%
53%
68%
78%
84%
82%
77%
SOURCE: Denver Regional Council of Governments
-------
388
Table C-9, preliminary data for 1975-76 suggest that migration may have
dropped as low as 16,000, much more in line with historical levels and
the DRCOG staff assumptions in Case I. A determination of the perma-
nence of this trend, however, is premature. More data are required to
verify or refute Case I migration assumptions.
In order to approximate the sensitivity of Year 2000 population to
changes in forecasting assumptions, the author developed the simplified
forecasting analysis described in Annex C-l. According to that analysis,
for Case I, a decrease in the rate of net natural increase of 1.0, e.g.,
from 8.5 per 1000 to 7.5, can be expected to decrease Year 2000 popula-
tion by about 56,000. Similarly, an increase in the annual in-migration
level of 1000 persons would lead to an increase in Year 2000 population
of about 34,000.
In late 1972 the DRCOG staff recommended to the Council the adoption
of Case I (2,175,000) as the accepted population forecast for planning
purposes. The recommendation was initially reviewed and accepted. How-
ever, a period of extensive discussion ensued. By negotiation, another
175,000 (8 percent) was added to the recommeded Year 2000 total from
Case I, giving a final total of 2,350,000. Most of the addition was attri-
buted to the southeastern part of Jefferson County (southwest of Denver).
By formal action on January 17, 1973, the Council of the DRCOG adopted
the amended population forecast of 2,350,000.
The action of the Council in negotiating an addition of 175,000 to
the staff-recommended Case I has created a change in the implied assump-
tions about migration. While Case I adhered well to historical levels,
the migration level required to produce a Year 2000 population of 2,350,000
is about 22,000 person annually. This has been stated in several DRCOG
publications (see References 17 and 18) and is confirmed by the author's
analysis in Annex C-l. Critics have observed that such a level is
-------
389
appreciably higher than historical levels, except for those in the early
1970's. To assume so high a level (in historical terms) will persist
over a 25 year period seems unrealistic, such critics have maintained.
However, without tying migration to causal relationships, such as employ-
ment, income and other local economic indicators, the forecast of migration
must be viewed as inherently speculative.
Without devoting too much space to the discussion, it is worth
observing that other forecasts of population exist for the Metropolitan
Denver region. Those totals for the year 2000 and the agencies respon-
sible for them are presented in Table C-10 and shown graphically in
Figure C-15, both from Reference 17.
The OBERS projection seems to be the only one tied to economic indi-
cators, assuming that migration will tend towards areas of economic oppor-
tunity and away from declining areas. It also is based on "Series E"
projections of the national population, i.e., the total fertility rate
(the total number of children a woman can be expected to have in her life-
time) will be 2.1 children. The chief criticism of the OBERS projection
has been due to their assumption of saturation in migration, i.e., that
the gradients in national economic forces that trigger migration will
tend to even out over time.
As for the other studies, the Colorado Land Use Commission study
done in 1972 simply relied on Case I projections. The 1975 Denver Research
Institute study was based on a survey of the various population project-
tions available at the time. DRI reasoned that a range of 2,350,000 to
3,000,000 was justifiable, accounting for uncertainties in the forecasting
process. For convenience, they quoted the midpoint of that range as their
expected projection. The Colorado Division of Planning projection was
based on a 1974 survey of local government forecasts by county. Their
result was simply the sum of the county totals.' They also conducted a
similar survey among municipal officials by city. They subsequently
-------
390
TABLE C-10. ALTERNATIVE POPULATION FORECASTS FOR THE DENVER
REGION IN THE YEAR 2000
1.. U.S. Commerce Department, Office of Business 1,981,000
and Economic Research Service (OBERS)
2. Colorado Land Use Commission 2,175,000
3. DRCOG Policy Forecast 2,350,000
4. Denver Research Institute (DRI) 2,675,000
5. Colorado Division of Planning (County Total) 2,886,000
6. Colorado Division of Planning (City Total) 2,892,000
7. Metropolitan Denver Water Study Committee 3,000,000
8. Colorado Division of Planning (Adjusted City Total) 3,399,000
SOURCES: "Denver Regional Council of Governments
U.S. Department of Commerce
Colorado Land Use Commission
Denver Research Institute
Colorado Division of Planning
Metropolitan Denver Water Study Committee
-------
3,000,000
2,500,000
2 2,000,000
O
h-
Q.
O
D.
1,500,000
1,000,000
500,000
WATER STUDY
COMMITTEE
jj) DRCOG
^ LAND USE
COMMISSION
1950
1960
1970
1980
1990
2000
SOURCE: DENVER REGIONAL COUNCIL OF GOVERNMENTS
FIGURE C-15. ALTERNATIVE PROJECTIONS OF POPULATION GROWTH FOR THE DENVER REGION
-------
392
issued an adjusted total forecast in which they substituted county fore-
casts when the sum of municipal projection was less than the county pro-
jection. The Metropolitan Denver Water Study Committee established in
1974 a projection of 3,000,000 based on the reasoning that an engineering
safety factor was needed over the DRCOG-approved forecast. No other
technical support for this choice seems to exist.
In summary while its embedded assumptions have at times been chal-
lenged, the DRCOG population forecast of 2,350,000 in the year 2000
is the Denver region planning standard by all practical terms. Whether
its net natural increase rate and migration level are both too high
must remain an unanswered question until current trends are either rein-
forced or refuted in the coming years.
2) Population Allocation
Even though the DRCOG has been able to agree on a population level
for the Denver region, efforts at allocation of that population to
individual communities have been subject to intense discussion. The lack
of formal adoption of an allocation plan has not prevented the staff of
the DRCOG from generating a series of iterations in population allocation,
each subsequent version designed to account for deficiencies in the previous
one. These allocation plans as they evolved became embedded in the plans
of a number of other agencies, including the CDH, the RTD, and APCD.
In 1972 the EMPIRIC Activity Allocation Model was developed for the
Denver region to assist in making these allocations. The model distri-
butes to subareas of the region, by several subcategories, a wide range
of characteristics, such as households, population and employment. This
allocation is done by the simultaneous solution of a set of linear
equations relating changes in population to changes in the characteristic
indicators.
Of the iterations produced using EMPIRIC the last two applications,
referred to as Cycle 3 and Cycle 4, have the most current relevance to
-------
393
the planning process. Those agencies using each of these have been
noted earlier in this section. Since a number of differences exist
in the policy scenarios examined by Cycle 3 and Cycle 4, it is instruc-
tive to outline the more important of these differences, as well as
their similarities.
Direct policy inputs for both allocations were essentially the same.
Both use the JRPP adopted Year 2000 highway and public transportation
plans. Cycle 4 used an adjusted 1974 Metropolitan Water Study rather
than the 1968 Study used in Cycle 3. A larger water service area was
used in Cycle 4, and sewer service in Cycle 4 is based on the adopted
Interim Regional Water Quality Plan which was only available in prelim-
inary form when Cycle 3 was developed.
Indirect policy inputs vary considerably between Cycle 3 and Cycle 4.
The DRCOG staff stated their desire to incorporate in Cycle 4 the eight
adopted regional plan policys which were not available in the Cycle 3
application. For instance, while Year 2000 regional population totals
2,350,000 in both cycles, Cycle 4 did not specify the subregional controls
that had been explicitly stated in Cycle 3 for county population equiva-
lents. The concept of activity centers was more strongly implied in
Cycle 4 through the use of direct minimum population and employment levels
Activity centers have represented in the planning process the focused con-
centration at a number of points in the Denver region of high levels of
inward and outward movement, e.g., shopping centers, employment complexes,
schools and universities, hospitals, to name just a few. Activity cor-
ridors, however, were not specified to the same degree in Cycle 4 as in
Cycle 3. Policies for contiguous development, unique urban areas, and
employment areas were developed by different means.
Land availability and development parameters were basically the same
in both cycles. Land lying within 100-year flcrodplains or which has a
slope greater than 15 percent was considered undevelopable. Remaining
available space is assumed within EMPIRIC to be consumed at rates based
-------
394
upon "change densities," or the amount of land needed by increases in
population. By decade planning districts are assigned to one of four
types of development based on the extent of development in the initial,
or base, year of the period. Densities are then applied to allocations
of households to compute land consumption. In Cycle 3 it was assumed
that half of all new development would be in single-family households
while the remaining half would be in multi-family households. Cycle 4,
however, assumed that 57 percent would be in single-family households
with the remaining 43 percent in multi-family dwellings. The densities
used to allocate population growth to land use consumption are presented
in Table C-ll, from Reference 18.
Based on the above listed variations in assumptions between Cycle 3
and Cycle 4, population allocation projections differed between the two.
The extent of those differences are illustrated in Tables C-12 and C-13. A
map of the county areas comprising the five-county Metropolitan Denver
region is shown in Figure C-16. The above tables and figures are from
Reference 18.
While both cycles have similar five-county population totals for all
three forecast years, 1980, 1990 and 2000, there are substantial differ-
ences in the individual county figures, both in level and staging. Com-
paring Cycle 4 with Cycle 3, the principal differences in the year 2000
are slower growth in Denver and Jefferson counties with Cycle 4, with a
compensating higher growth in Adams and Arapahoe counties. Also, in
Cycle 4, almost no growth is projected to occur in Denver for the 1980-
1990 period.
A third population allocation has been done for the Denver region.
Referred to as the Subarea Allocation Plan, its importance in comparison
with Cycle 3 and Cycle 4 is that the Council of the DRCOG on 18 August
formally adopted it as DRCOG policy. Neither Cycle 3 nor Cycle 4 were
ever formally adopted. It is important to recognize that the Subarea
Allocation Plan (SAP), however, is not simply an iterative extension of
-------
395
TABLE C-ll. DENSITIES USED TO ALLOCATE GROWTH
TO LAND CONSUMPTION
(dwelling units per net residential acre)
Development Area
Mature Rapidly
Category Core Developed Urbanizing Satellite
Area Area Area
Single-Family 7.0* 6.88 5.68 2.97
Multi-Family 53.6 31.40 19.40 7.89
*3/C EMPIRIC used 10.0 units/acre as a maximum.
TABLE C-12. PERCENT CHANGE IN POPULATION FOR
EMPIRIC CYCLE 3 AND CYCLE 4
1970-1980 1980-1990 1990-2000
Geo-County
3/C 4/C 3/C 4/C 3/C 4/C
Adams
Arapahoe
Boulder
Denver
Jefferson
35.1
61.0
48.7
9.1
62.6
43.5
73.0
37.7
7.9
56.5
17.3
30.8
9.7
9.6
37.9
29.5
37.4
17.5
0.4
28.2
14.5
20.3
14.0
13.8
25.4
20.4
29.5
20.6
8.5
22.9
Total 34.2 34.2 20.6 19.3 18.1 19.4
-------
TABLE C-13. COUNTY POPULATION ALLOCATIONS FOR EMPIRIC CYCLE 3 AND CYCLE 4
EMPIRIC
County*
Adams
Arapahoe
Boulder
Denver
Jefferson
Total
1980
3/C 4/C
251,000
254,500
196,000
565,500
383,000
1,650,000
266,600
273,500
181,500
559,300
368,800
1,650,200
1990
3/C 4/C
294,400
332,800
215,000
619,700
528,100
1,990,000
345,300
375,800
213,300
561,500
472,800
1,968,800
2000
3/C 4/C
337,200
400,400
245,000
705,200
662,200
2,350,000
415,600
486,600
257,200
609,000
581,300
2,350,000
CO
UD
cr>
* Approximation of 1970 County jurisdictional area.
-------
397
GEO-COUNTIES
FIGURE C-16. COUNTY AREAS IN DENVER REGION
-------
398
Cycle 4. It is based much more directly on the local development plans
proposed by each of the communities within the region.
The aggregate of the local plans totaled in the year 2000 to over
4,000,000 in the version first offered several years ago. In the time
since the study was authorized in 1975, the DRCOG staff seems to have met
with local planning staffs in an effort to evaluate the realism of their
forecasts. Water supply and sewerage treatment capacity formed an impor-
tant determinant of reasonability. After the completion of these discus-
sions, the aggregate of local development plans had dropped to 2,739,200
in the year 2000.
Forecasting for the SAP was based on the concept of the Urban Service
Area (USA). These are defined to be those parts of the region in which
local governments plan to support urban development and redevelopment
with urban services. In rough terms the boundaries of the USAs corres-
pond to community boundaries as foreseen in the year 2000. Certain high
growth areas not formally incorporated have also been defined as USAs.
The division of the Denver region into USAs is shown in Figure C-17, from
Reference 18.
Given the local development plans, the DRCOG staff still had to con-
strain the total Year 2000 regional population to the DRCOG-adopted level
of 2,350,000. It was here that a major assumption was made. The staff
assumed that although the sum of the local plans was too high by nearly
400,000 the relative percentages of the growth attributable to each com-
munity were in proper proportion. Based on this assumption, the local
plans in each Urban Service Area were all scaled down by the same factor,
one chosen to reduce the 2,739,200 total forecast to 2,350,000. Staging was
kept the same as that forecast by the individual local development plans.
Those areas considered by DRCOG to be subject to relatively high
growth rates are indicated in Figure C-18. In order to make more graphic
the distribution and staging of these allocations, two series of regional
-------
FIGURE C-17. URBAN SERVICE AREA
BOUNDARIES
URBAN SERVICE AREAS
COMMUNITY SERVICE AREAS
,____.^-.
\ o* ,-. , /-\ . • •
-JEF£cSTCPf>beweQD PLEASANT
-------
400
FUTURE GROWTH AREAS
(SUBAREA ALLOCATION
PLAN)
-------
401
maps are presented. In the first, Figures C-19(a) through C-19(e), the pop-
ulations are shown at periodic intervals for each Urban Service Area in the
Metropolitan Denver region. In the second series, Figures C-20(a) through
C-20(e), subarea allocations by Urban Service Area in high growth areas are
presented for consecutive time periods through the year 2000.
The issue of most basic interest to regional planners, however, is that
of compatibility between the three current population allocation plans:
Cycle 3, Cycle 4, and the Subarea Allocation Plan. The differences between
the first two of these have been discussed earlier and are tabulated in
Table C-13. The data in that table can be considered with the geo-county
projections in Table C-14 to form a complete comparative set.
TABLE C-14. ALLOCATION PLAN COMPARISON (YEAR 2000)
Subarea 3/C 4/C SPA
Adams
Arapahoe
Bbulder
Denver
Jefferson
5 Co-Region
337,200
400,400
245,000
705,200
662,200
2,350,000
415.600
486,600
257,200
609,000
531,400
2,350,000
392,400
418,700
287,900
603,700
647,300
2,350,000
Notes: Subarea - Approximation of 1970 county juris-
diction
3/C - Third Cycle EMPIRIC Application
4/C - Fourth Cycle EMPIRIC Application
SPA - Subarea Population Allocation Study
A more detailed breakdown of the Subarea Allocation Plan by County and
Urban Service Area are presented in Tables C-15 and C-16. The first two of
these three tables are from Reference 18. The last table is from the text
of Resolution No. 18, 1976, passed by the DRCO£i (Reference 9).
The basic differences between the three projections are as follows.
The Subarea Allocation Plan assigns to Denver a level in the year 2000
(603,700) slightly below the Cycle 4 allocation (609,000) and well below
-------
FIGURE C-19. POPULATION BY URBAN
SERVICE AREA (SUBAREA
ALLOCATION PLAN)
(a) Year 1975
(Note: Compound growth rates over
previous 5-years are in
parentheses.)
DENVER
ANDVICINITY
O
ro
-------
FIGURE C-19 (Continued)
(b) Year 1980
(Note: Compound growth rates over
previous 5-years are in
parentheses.)
o
CO
-------
FIGURE C-19 (Continued)
(c) Year 1985
(Note: Compound growth rates over
previous 5-years are in
parentheses.)
DENVER
ANDVICINITY
O
-p.
-------
FIGURE C-19 (Continued)
(d) Year 1990
(Note: Compound growth rates over
previous 5-years are in
parentheses.)
o
tn
-------
FIGURE C-19 (Concluded)
(e) Year 2000
(Note: Compound growth rates over
previous 5-years are in
parentheses.)
-------
DENVER
ANDVICINITY
FIGURE C-20. SUBAREA ALLOCATIONS OF POPULATION
IN GROWTH AREAS BY TIME PERIOD
(a) Period 1970-75
(Note: Percents of total regional allocation
are in parentheses.)
-------
DENVER
AND VICINITY
FIGURE C-20. (Continued)
(b) Period 1975-80
(Note: Percents of total regional
allocation are in parentheses.)
o
co
-------
FIGURE C-20. (Continued)
(c) Period 1980-85
(Note: Percents of total regional
allocation are in parentheses.)
-p.
o
-------
FIGURE C-20. (Continued)
(d) Period 1985-90
(Note: Percents of total regional
allocation are in parentheses.)
-------
FIGURE C-20. (Concluded)
(e) Period 1990-2000
(Note: Percents of total regional
allocation are in parentheses.)
-------
TABLE C-15. SUBAREA ALLOCATION PLAN BY COUNTY (1970-2000)
County
Turlsdictional*
Adams
Arapahoe
Boulder
Jefferson
Denver
Region
Geb-Countv**
Adams
Arapahoe
Boulder
Jefferson
Denver
Region
Base
1970
185,800
162,100
131,900
235.300
715,100
514.700
1,229,800
185,800
158,100
131,900
235,600
711,400
518,400
1,229,800
Current
1975
230,700
223,400
169,900
322.700
944,500
527,100
1,473,800
230,700
219,400
169,900
323.400
943,400
530,400
1,473,800
1980
767,400
254,700
194,100
395,100
1,111,300
538.900
1,650,200
267,400
247,700
193,600
397.100
1,105,800
544,400
1,650,200
Short Term
1985
288,800
278,900
217,000
456,300
1,241,000
561,500
1,802,500
294,400
274,200
217,400
459.000
1,245,000
557,500
1,802,500
1990
308,300
313.600
244,500
518,000
1,384,400
584,400
1,968,800
319,500
310,900
245,900
521.600
1,397,900
570,900
1,968,800
Long Term
2000
360,800
420,300
286,400
625,600
1,693,100
656.900
2,350,000
392,400
418,700
287,900
647.300
1,746,300
603,700
2,350,000
* Current Jurisdictional area .
** Aggregation of Transportation Study Districts approximating 1970 county Jurisdictlonai area.
-------
TABLE C-16. SUBAREA ALLOCATION PLAN BY SERVICE AREA (1970-2000)
Service Area
Mountain
Boulder Co Mtn
Evergreen*
Jefferson Co Mtn.
Lyons*
Nederland*
Plains
Bennett*
Box Elder (Ad)
Box Elder (Ar)
.Byers*
Deer Trail-*
East Plains (Ad)
East Plains (Ar)
Strasburg*
Base
1970
24,8"00
5,700
5,700
11,900
1,000
500
4,000
700
400
400
400
400
100
600
1,000
Current
1975
29,600
5,900
7,800
14,200
1,000
700
4,400
900
400
400
400
400
200
700
1,000
Short Term
1980
34,600
6,600
9,900
16,900
1,300
900
5,800
1,400
600
500
500
400
800
600
1,000
1985
40,000
6,700
12,000
18,700
1,600
1,100
6,800
1,700
700
500
700
400
600
700
1,500
1990
46,600
6,700
14,000
22,500
2,000
1,400
7,800
2,000
900
500
800
400
300
900
2,000
Long Term
2000
55,500
7,900
16,800
26,800
2,500
1,500
9,900
2,300
1,100
600
1,000
400
100
1,400
3,000
CO
*Community Service Area
-------
TABLE C-16 (Continued)
Service Area
Adams Co (Uninc)
Adams Co (Urban)
Arap Co (Uninc)
Arap Co (Urban)
Arvada
Aurora
Boulder City
Boulder Co (E Uninc)
Bow Mar
Brighton
Broomfield
Cherry Hills Village
Columbine Valley
Commerce City
Denver
Edgewater
Englewood
Erie**
Federal Heights
Base
1970
1,201,000
9,600
54,200
600
28,800
58,300
78,200
78,400
8,700
1,000
8,300
7,700
4,600
500
17,400
520,200
4,900
33,700
-
1,500
Current
1975
1,439,800
10,500
62,500
700
35,900
80,800
121,700
92,200
13,300
1,100
12,900
16,400
5,400
600
18,600
532,300
5,500
35,900
100
7,600
1980
1,609,800
12,000
70,000
800
37,400
89,600
140,800
98,800
14,700
1,300
16,000
24,200
7,500
600
19,500
543,900
5,500
38,900
100
9,100
Short Term
1985
1,755,700
14,200
71,400
900
39,900
94,800
157,000
106,700
17,300
1,300
18,000
29,700
8,300
600
19,500
566,100
5,600
40,900
300
9,900
1990
1,914,400
14,400
74,900
1,000
48,100
100,500
175,200
115,500
19,600
1,300
20,000
35,600
9,100
600
19,800
588,600
5,600
42,900
500
10,800
Long Term
2000
2,284,600
14,800
83,100
1,200
79,300
117,400
228,700
130,100
24,700
1,300
25,000
44,500
10,700
600
20,000
646,700
5,900
45,900
800
12,900
**Allocation for part in Boulder County only; total service area forecasts are:
1975 - 1,300 1990 - 2,000
1980 - 1,800 2000 - 2,200
1985 - 1,900
-------
TABLE C-16 (Concluded)
Service Area
Valley (Cont'd)
Glendale
Golden
Greenwood Village
Jefferson (A-P)***
Jefferson (SE Urban)
Jefferson (Uninc)
Lafayette
Lakewood
Lakeside
Littleton
Longmont
Louisville
Morrison*
Mountain View
Northglenn
Sheridan
Superior
Thornton
Westminster
Wheat Ridge
5-County Region
Base
1970
800
9,800
2/600
8,200
9,100
100
4,100
95,iuO
** Ar*
27,700
23,200
2,800
400
700
27,900
4,800
200
15,500
21,600
29,800
Current
1975
3,700
14,500
3,500
9,500
21,200
200
5,800
123,000
****
33,500
34,200
3,700
500
800
31,500
6,000
300
27,600
31,600
34,300
1980
Short Term
1985
1990
8,700
16,200
4,200
10,600
40,900
4,000
8,500
148,900
****
36,300
42,100
6,200
500
800
34,300
6,600
300
35,500
36,500
38,000
8,700
18,600
5,200
11,600
68,200
6,000
10,200
163,700
* ***
38,000
46,300
10,000
600
800
36,700
8,100
400
39,500
40,700
40,000
8,700
21,000
6,500
12,700
96,600
8,000
11,900
176,700
****
40,100
53,200
13,600
700
800
37,800
9,500
400
44,800
45,400
42,000
Long Term
2000
8,700
26,000
10,000
15,200
137,900
12,000
14,000
213,600
*•*•**
47,300
67,400
13,600
800
800
38,600
13,200
500
65,500
58,900
47,000
1,229,800 1,473,800 1,650,200 1,802,500 1,968,800 2,350,000
-p.
U1
***A-P - Applewood-Pleasantview.
****Less than 50.
-------
416
the Cycle 3 figure (705,200). Analysis of local and regional develop-
ment plans indicated that both Cycle 3 and Cycle 4 underallocated growth
in Boulder County. The SAP allocation of 287,900 to Boulder is well above
the earlier allocations (2:45,000 and 257,000). In Adams and Arapahoe Coun-
ties, the SAP allocation is closer to that for Cycle 4. In Jefferson
County the reverse is true: the SAP is closer to Cycle 3. This latter
point is helpful from the point of view of air quality analysis, since
most available traffic data is based on Cycle 3. The prevailing south-
west to northeast wind during the morning rush-hour period also empha-
sizes the importance of emissions in the southeast part of Jefferson
County (southwest of Denver).
The above discussion details both the development and detailed results
of the population allocation efforts most currently in use in the Denver
regional planning process. If the SAP represents the institutionally
endorsed plan while available emissions data is based on Cycle 3 or Cycle
4, careful attention must be paid to both the staging and level of differ-
ences. While the above paragraph indicates that in Jefferson County the SAP,
much closer to Cycle 3, may be compatible with transportation planning,
this may not be true in Denver County. The SAP in that county is closer
to Cycle 4. This would suggest Cycle 3 might have higher emissions in
the downtown area. Without a more detailed breakdown of the differences
in population allocation, however, and some idea of their work-leisure
driving patterns, any statement of definitive conclusions is premature.
It is, nevertheless, helpful that all three allocation plans sum to the
same year 2000 regional population.
The staging of population growth is also an important factor in plan-
ning analyses. Much of the difference in Denver County between Cycle 3
and Cycle 4/SAP, for instance, comes in the decade 1980-1990. Growth
in Arapahoe County, on the other hand, comes primarily in the decade
1990-2000. Growth in southeast Jefferson County is high in the begin-
ning and middle years of the study period, tapering off toward the end.
-------
417
c. Land Use Planning
Land use planning is intertwined with population planning, pacing
the level, distribution and staging of growth. Its importance as an
issue in planning, however, well exceeds the size of the discussion
devoted to it in this report. This is. due primarily to the apparent
uniformity of adherence throughout regional planning agencies to the land-
use plan adopted by the JRPP/DRCOG in January of 1974. Although in the
intervening time since then a number of issues have emerged which may
cause the DRCOG to reexamine its adopted plan, such action will probably
not occur during the time-frame of this Overview EIS. It is thus impor-
tant to understand the policy issues underlying the approved plan.
As described in Reference 10, seven policy statements were enumerated
at the time of the adoption of the land use plan by the DRCOG, acting on
behalf of the JRPP. First, a Year 2000 population below 2,350,000 should
be encouraged. Second, the Central Business District (CBD) of Denver
should be considered the "major high density core of business, cultural,
governmental, commercial and residential activity," with major high density
corridors running east, west and south from the CBD. Third, the develop-
ment of several major activity centers should be encouraged in the Metro-
politan Area. Fourth, new, low density residential growth should be
encouraged only in locations contiguous to existing urban areas. Fifth,
new industrial development should be encouraged only to the extent suit-
able locations can be found where environmental hazard potential can be
effectively controlled or minimized. Sixth, major unique urban areas
should be encouraged to remain distinct, i.e., "not to ultimately be allowed
to expand and run together into a nondescript, low density urban sprawl."
Seventh, major areas of "ecological, environmental, agricultural, historic
and archeological significance" should remain in a natural open or low-
density, non-urban condition.
The DRCOG stated that the purpose of the land use plan was not to
replace local plans but to provide a framework of guidelines within which
-------
418
local decisions could be made. It expressed, however, concern that the
surge in in-migration and residential construction encountered in the
early 1970's had resulted in the conversion to single and multiple family
residential use large parcels of formerly agricultural or vacant land.
The DRCOG noted that this had generated concern over the undesirable
effects of urban sprawl.
A major planning concept advanced by the JRPP adopted plan was that
of activity centers. These were designed to be aggregations of urban act-
ivity great enough to be primary generators of large internal activity.
The DRCOG considered them to be "a key tool by which the Regional Plan
can be achieved." A map of the JRPP Land Use Plan, presented earlier in
the report, is repeated for convenience in Figure C-21. The activity
centers are denoted on it by circles.
It was conceived that the activity centers would create distinct nodes
of higher level activity and services throughout the region. By orderly
location of that activity, it was hoped some development pressure could
be relieved on the urban fringes and also provide impetus for redevelop-
ment of currently depressed areas. The provision of activity centers would
also allow a large portion of the individuals living within the areas to
live and work there. Each activity center was considered to extend through
an area of about one mile in radius. A list of the activity centers recom-
meded by staff to the DRCOG is presented in Table C-17, from Reference 10.
The plan also recommends two high intensity land use corridors to
be served by a fixed guideway public transportation system. As noted in a
previous section, such a system was recommended by the RTD, but funding
was denied by UMTA. The first corridor was proposed to lie east and west
of the CBD along Col fax Avenue from Kipling Avenue on the west to 1-225
on the east. The other corridor was to extend southward from the CBD to
Englewood along a route paralleling Broadway. They ranged in width from
one-quarter to one-half mile.
-------
IOW DENSITY URBANIZATION
HIGH DENSITY URBANIZATION
EMPLOYMENT OR TRANSPORTATION
ACTIVITY CENTER
PRIVATE & PUBLIC ACCESS
PRIMARY & SECONDARY PRESERVATION
NON- URBAN PRE 2000
FIGURE C-21
JRPP LAND USE PLAN
-------
420
TABLE C-17. ACTIVITY CENTERS FOR JRPP LAND USE PLAN
Center
Boulder
Northglenn
Arvada
Federal Center
Villa
Alameda
Denver CBD
Medical Center
Cherry Creek
South Colorado
Englewood
Technological Center
General Location
University of Colorado/Crossroads Shopping Area/
CBD
Northglenn Shopping Area/We stern Electric Front
Range Denver Area
Old CBD/Tri-Center Area
Federal Center/Community College/Westland
Areas
Villa Italia Vicinity
Alameda/I-225 Area
CBD/Auraria/Skyline
CU Medical Center/Colorado-Colfax Area
General Cherry Creek Shopping and Residential
Area
Colorado Blvd./I-25 Area
Englewood CBD/Cinderella City Shopping Center
General Area of Denver Technological Center/
Greenwood Plaza
Littleton
Littleton CBD Area
-------
421
The plan does not explicitly address future employment, leaving that
subject to local governments. It does attempt to identify those areas
most desirable from a regional perspective. Public parks and recrea-
tional areas are shown on the map in Figure C-21, forming a component of
the Open Space Plan.
From an air quality perspective, two features of the JRPP Lane Use
Plan have particular significance, the first explicit, the second implicit.
The first concern is the identification of those land areas in which
fixed-source emissions can be expected to occur. In the map in Figure C-21,
these regions are designated as "High Density Urbanization" and "Employ-
ment or Transportation." The second feature of the plan of concern is
the activity center and corridor element. Areas of such intense activity
have a sizeable indirect impact on the frequency and distribution of vehicle
travel. The resulting highway loading parameters are strong determinants of
air quality. A more explicit description of the twelve activity centers
and the three travel corridors in the JRPP Land Use Plan is provided by
Figure C-22 and Table C-18, both from Reference 19.
It should be kept in mind, however, that the activity center concept
has been relaxed in recent planning efforts. Transportation planners
noted, for instance, that using the concept in attempts to model the traffic
network leads to distortions in the traffic assignment process by which trips
are loaded onto roadway links. Also, since the UMTA decision not to fund
a fixed guideway transit system, the RTD has had to reorient its plan to
incorporate an all-bus Year 2000 transit plan. This is certain to be a
factor in a revisiting of the JRPP Land Use Plan expected to occur in the
next year.
-------
422
FIGURE C-22. ACTIVITY CENTER AND CORRIDOR LOCATIONS
(JRPP LAND USE PLAN)
-------
TABLE C-18. ACTIVITY CENTERS AND TRAVEL CORRIDORS (JRPP LAND USE PLAN)
Current Current 2000 2000
Activity Centers Popu- Employ- Employ- Persons 2000
and Corridors Primary Functions lotion 2000 ment ment DU/G.A. DU/G.A.
A. Boulder
B. Northglenn
C. Arvada
D. Denver CBD-Aurarla
E. Federal Center-
Westland
F. Villa Italia
G. C.U. Medical
Center
H. Cherry Creek
I. New Aurora
I. Writers Manor-
University Hills
K. Englewood
L. Tech Center
M. East Colfax
Corridor
N. W. Colfax Corridor
O. S. Broadway
Corridor
Shopping/employment/educational/
housing
Shopping/employment/housing
Shopping/employment/housing
Employment/shopping/educational/
cultural/institutional/housing
Employment/shopping/housing/
Institutional
Shopplng/employment/housing
Institutional/employment/housing
Shopping/employment/housing
Shopping/employment/institutional/
housing
Employment/shopping/housing
Shopping/employment/institutional/
housing
Employment/shopping/housing
Employment/shopping/housing
Employ ment/s hopping/housing
Employment/shopping/housing
32,000
23,000
28,000
27,000
14,000
15,000
25,000
12,000
10,000
35,000
16,000
2,000
47,000
29,000
45,000
39,000
54,000
47,000
40,000
27,000
26,000
42,000
24,000
16,000
49,000
27,000
33,000
81,000
64,000
64,000
20,000
5,000
5,000
78,000
13,000
4,000
14,000
8,000
500
11,000
10,000
2,000
15,000
12,000
18,000
26,000
21,000
16,000
85,000
20,000
16,000
25,000
14,000
5,000
32,000
19,000
36,000
39,000
30,000
23,000
15. 3
16.8
21.9
22.5
9.3
20.1
30.9
23.7
13.5
12.0
22.2
26.1
43.2
27.6
35.1
5.1
5.6
7.3
7.5
3.1
6.7
10.3
7.9
4.5
4.0
7.4
8.7
14.4
9.2
11.7
-------
424
d. Transportation Planning
It is the results of the transportation planning process that have
the most direct influence on regional air quality. Agreed upon regional
population level, distribution and staging lead to design of a highway
network, supplemented by a rapid transit system, able to accommodate
anticipated travel levels and distribution, both temporal and spatial.
Both the vehicle-miles-traveled (VMT) and the average speed along each
highway link determine the amounts of carbon monoxide (CO) and hydro-
carbons (HC) emitted into the surrounding air. The problems resulting
from highway design are particularly acute because of the extensive
ownership and use of private automobiles in the Metropolitan Denver
region.
In this section the transportation planning process will be explored
in some detail. The discussion will be presented in two parts. The first
will deal with the physical transportation network alternatives. Two
such plans exist: the DRCOG-adopted JRPP Highway Plan used by the DRCOG,
the CDH, and the APCD; and the Existing plus Committed (E+C) used by the
RTD as an alternative in planning their rapid transit network. The second
part of the transportation planning discussion will consider the problems
inherent in transportation modeling and the uncertainties introduced by
the various modeling iterations, principally within the DRCOG staff.
1} Transportation Networks
The JRPP Highway Plan, repeated here as Figure C-23 from Reference
10, was originally adopted in December 1973 by the DRCOG acting as the
Policy Body for the JRPP. The other member agencies, CDH and RTD,
approved it in January 1974. It has been incorporated along with Cycle
3 population allocation into the transportation planning efforts of the
DRCOG and the CDH. It has also been used by the APCD, although in con-
junction with Cycle 4 population allocation.
-------
LIMITED ACCESS FACILITY
PRINCIPAL ARTERIAL
ADOPTED
OIHVEII MGIOHAt COUNCIL Of COVEIMMTI
COLORADO OErAnTMEKT Of HICHWMI
IEEIOIIAL TB»«SPORT»TIO» DISTRICT
FIGURE C-23. JRPP HIGHWAY PLAN
-. r-f
--
-------
426
The highway plan was conceived as a systems plan, identifying general
corridors rather than specifying detailed construction plans. It was
designed as a package to be presented for approval at the same time as the
Regional Land Use and Public Transportation Plans. The three were designed
to complement each other. The Highway Plan, with its grid of principal
arterials, spaced one to two miles apart, and its grid of limited access
arterials, spaced four to five miles apart, was to serve the lower density
developments, while the public transportation system was to serve the
higher density areas. The Highway Plan provided for the addition of 80
miles to the limited access facilities. Major elements of that addition
are detailed in Table C-19, as abstracted from Reference 10. Some of the
routes on the principal arterial system were identified as needing exten-
sive widening or construction.
In developing the above-mentioned complementary land use, highway,
and public transit plans, the staff for the DRCOG, acting on behalf of
the JRPP, considered six alternatives for the distribution of urban
activities within the Denver region. The six were: (1) Continue Exist-
ing Trends, (2) Reinforce Central Denver, (3) Reinforce Metropolitan
Centers, (4) Reinforce Regional Centers, (5) Stimulate Development of
New Towns, and (6) Dispersal. According to Reference 19, it was decided
not to use computer oriented or extensive simulation techniques in the
evaluation and analysis of the six alternatives. Such efforts were
judged inappropriate at the level of concept specification.
For each of the alternatives, a transportation system was conceived.
Each was then evaluated to determine its impacts on the region. These
impacts were measured against parameters formulated from an analysis of
local community goals and objectives and regional goals adopted by the
DRCOG. The resulting generalized conclusions about the success or failure
of the six alternatives in meeting local and regional objectives produced
a synthesis of the six into one land development and transportation concept.
An iterative process refined this concept into specific regional land use
and transportation plans. This process included discussions with local
officals and groups.
-------
427
TABLE C-19. NEW LIMITED ACCESS FACILITIES (JRPP HIGHWAY PLAN)*
Interstate 470: This facility will serve the anticipated growth in
Southern Jefferson County, Arapahoe and Douglas Counties.
Interstate 225: This facility will be completed from Parker Road to
1-25.
Interstate 25: This is, and will continue to be, the most heavily
traveled freeway in the region.
Interstate 805: This facility will connect Interstate 25 with Interstate
70 and will follow the Clear Creek flood plain.
Hampden Avenue: This facility will service the east-west travel in the
southern portion of the Denver Metropolitan Area . A detailed location
study will be required between Santa Fe Drive and Interstate 25 before
the facility can finally be located.
.Santa Fe Drive: This facility, long discussed as the "Columbine
Freeway," is needed to serve the travel needs of the developing south
and south-western portions of the Metropolitan Area. The facilities
implementation will have to be coordinated with that of the construction
of the rapid transit line which uses the same corridor.
KiPlinq: This facility is being planned as an arterial by the City of
Lakewood. The City of Wheat Ridge strongly opposes its construction.
The expansion of this arterial to a limited access facility will have to *
carefully consider the highway's economic and environmental benefits
and costs.
Quebec: This facility, from 1-70 to Parker Road, will require extensive
citizen participation and environmental design before it will meet the
travel and community needs of the area affected. This element is placed
on the regional plan pending more detailed analysis that must be done
by the Joint Regional Planning Program and the Colorado Division of
Highways.
Longmont Diagonal: This facility will connect Longmont with 1-25 and the
City of Boulder.
* From Reference 10.
-------
428
The transportation concept arising out of the above process took
specific form through analysis of regional travel demand corridors. Com-
puterized traffic assignment techniques and transportation models were
used to evaluate the resulting highway and public transportation alter-
natives. Nine computer tests of these alternatives were conducted and
used in the plan development work. The tests were referred to as 1971,
2000 Base Case, 2000 Concept, A-l, A-2, A-3, A-4, A-5, and 2000 Compara-
tive Case. The 1971 test was used to calibrate the model against observed
data. The Base and Comparative Cases were tested for comparisons with
the plan test. The 2000 Concept represented the initial plan but the
"new town" concept was subsequently eliminated. The tests denoted A-l,
A-2, A-3, and A-5 were then used to select the preliminary plans. The
A-4 test with an updated "new towns" component was used as a comparison
against the preliminary plans. The assumptions embedded in each of these
tests are presented in Table C-20 from Reference 19.
As mentioned earlier in this section, a basic feature of the land
use plan resulting from the DRCOG/JRPP analysis was the concept of act-
ivity centers. These were viewed as localized centers of high-intensity
urbanized activity. They were included with the high density travel
corridors and other intense centers of development to catalog a list of
major traffic generators. A map of these major traffic generators is
shown in Figure C-24, with a list showing projected trip-ends presented
in Table C-21. Both are from Reference 19.
-------
TABLE C-20. ASSUMPTIONS OF POLICIES AND CONSTRAINTS ASSOCIATED WITH EACH CASE
Base Case
Comparative
A-l,A-2,A-3,A-5
Concept- New
Towns
Concept A-4
Concept- No
New Towns
*Alternatlve
Urban Form Sprawl
Environment
Development Policy
Activity Parameters
1. Population
2. Employment
Highway System
Transit System
Sprawl
Activity centers,
activity corridors,
new towns
Activity Centers,
activity corridors
As close to activity
centers & corridors
as possible
Limited
2,097
1, 104
Existing
Level of
>
control
,100
,400
priorities
Service
C
Limited Control
2,
1,
097
104
,100
,400
LAF, principals
Level of Service
S: C
Strong control
"radical" new
legislation
1,899
848
,900
,800
LAF, principals
Level of Service
•< c
Strong control
2
1
,097
,049
,100
,000
LAF, principals
Level of Service
< C
Strong control
2,
1 ,
097, 100
049,000
LAF, principals
Level of Service
< C
ro
ID
Extended Denver
Metro Transit
None
Financial Restraints
l: Transit
2. Highway
Political-Public None
LAF - Limited Access Facility
PRT - Personal Rapid Transit
Extended Denver
Metro Transit
None
None
None
PRT
1.0 Billion
None
No LAF Central
No LAF on Wads-
worth, few new
LAF's
PRT
1.0 Billion
None
No LAF Central
No LAF on Wads-
worth, few new
LAF's
AH bus-but new
technologies &
concepts and/or
limited PRT
1.0 Billion
None
No LAF Central Denver
NO LAF on Wadsworth
few new LAF's
'Proposed
-------
430
FIGURE C-24. MAJOR TRAFFIC GENERATORS (JRPP HIGHWAY PLAN ANALYSIS)
-------
431
TABLE C-21. LOCATIONS OF MAJOR TRAFFIC GENERATORS (JRPP HIGHWAY PLAN ANALYSIS)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Longmont CBD
IBM at Niwot
Broomfield CBD
Brighton CBD
North Valley S. C. /Valley View Hospital
Westminster Plaza
Golden CBD/Coors
Lakeside S. C.
Montbello
Stapleton International Airport
Fitzsimmons Hospital
Lowry Air Force Base
Buckingham Square S. C.
Denver University
Sheridan & 88th S. C.
Bear Valley S.C.
Fort Logan Hospital
Littleton CBD
Johns-Manville
Southglenn S.C.
Gates Industrial Park
Martin Marietta
1970 Trip Ends
17,700
14,200
25,600
8,900
13,800
16,200
35,900
18,600
40,800
30,800
30,000
58,400
0
35,000
14,000
8,700
8,300
33,100
0
7,100
900
24,800
2000 Trip Ends
26,800
35,000
49,400
20,000
52,500
35,800
39,700
22,800
126,200
71,400
21,300
55,000
31,600
35,000
47,100
22,800
20,000
89,500
111,000
18,700
20,000
42,800
-------
432
In the JRPP analyses the Denver region was divided into districts
for transportation planning purposes. A map of the individual districts
as conceived in 1972 is presented in Figure C-25 from Reference 19. Aggre-
gation into super-districts was also performed. Boundaries of those super-
districts are shown in Figure C-26 with population and employment informa-
tion provided in Table C-22, also from Reference 19. Year 2000 population
and employment density on a regional basis are shown in Figures C-27 and
C-28, from Reference 21 as prepared for the RTD.
It appears that the plan ultimately adopted by the DRCOG on behalf of
the JRPP corresponds fairly closely with the alternative referred to as
A-5 and shown in Figure C-29. It seems to have been used in an analysis of
rapid transit networks performed by the DRCOG staff in 1972-73. Prelimin-
ary estimates indicated a construction cost for the highway plan of about
$960,000,000. New construction areas are indicated in Figure C-30.
A second highway plan has since emerged from within the Regional Trans-
portation District. Much of the rationale for the new plan came from air
quality analyses carried out for RTD by Systems Management Contractor.
During the latter phase of that work in early 1975, the E+C network was
introduced for comparison with the JRPP plan, presumably in the event that
the JRPP highway construction program proved unrealistically optimistic
from a funding point of view. This was in response to the urging of UMTA
representatives reviewing RTD alternatives. The E+C network so defined
consisted of existing roadways and those new construction projects for
which funding had already been committed.
Several principal features of the JRPP plan have already run into
difficulties. Opposition arose some time ago to the 1-470 circumferential
highway. At best federal funding seems delayed. A detailed assessment
report on 1-470 was not filed until September of 1976. Also, UMTA recently
has denied federal funding for the fixed guideway rapid transit system
proposed by the RTD. The impact this will have on highway planning has
not yet been assessed, though it is possible that higher levels of auto-
mobile traffic might result.
-------
433
1971 DISIKICT MAP
FIGURE C-25. TRANSPORTATION ANALYSIS DISTRICTS
(JRPP HIGHWAY PLAN)
-------
434
J ^
FIGURE C-26. TRANSPORTATION ANALYSIS SUPER-DISTRICTS
(JRPP HIGHWAY PLAN)
-------
435
TABLE C-22. POPULATION AND EMPLOYMENT BY SUPERDISTRICT
1970 Compar. 2000 Plan
Super
Dist.#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
29
33
34
35
36
37
38
39
40
al DBL
Pop
x 1000
3.1
64.7
61.4
40.8
51.8
62.6
42.9
63.2
52.9
74.5
26.2
48.5
74.1
55.5
8.3
112.5
1.2
11.1
28.3
28.9
49.0
20.4
67.5
19.4
.3
8.4
.7
36.6
1.1
6.5
17.7
18.5
25.6
9.4
.4
1,194.0
Emp .
xlOOO
47.9
16.8
58.9
25.1
32.6
30.3
14.0
19.6
8.8
52.9
17.0
8.5
23.5
16.5
7.4
31.8
.7
3.1
10.4
8.9
12.0
5.7
26.8
4.7
.'1
2.9
.3
19.6
4.0
5.2
2.2
2.8
8.5
.7
—
530.2
Pop.
3.6
62.4
58.9.
57.0
74.4
64.9
60.1
63.1
54.5
72.2
94.0
77.6
94.3
126.7
104.0
211.8
21.0
25.0
30.1
64.3
149.4
132.4
87.6
32.9
1.7
19.0
14.0
57.6
5.2
24.2
52.0
38.0
45.0
35.0
1.0
2,114.9
Emp .
46.2
43.7
85.1
48.7
72.1
40.0
37.8
47.7
40.3
61.4
43.. 2
23.2
39.4
36.7
26.8
68.6
9.7
7.8
21.9
24.4
48.7
62.4
52.8
18.0
.3
13.8
4.7
26.9
5.7
11.9
11.1
5.4
14.3
9.2
.5
1,110.4
POD.
x 1000
7.0
67.2
78.6
70.7
90.9
98.8
80.4
85.1
68.2
97.8
77.0
78.0
122.7
106.2
62.5
210.5
14.2
15.0
23.0
54.4
135.4
103.5
113.0
26.0
1.7
18.5
2.8
47.4
8.0
17.0
27.5
26.1
42.7
16.3
3.0
2,097.1
Emo .
x 1000
50.2
46.3
75.5
46.0
32.6
40.0
39.5
48.0
32.0
55.5
35.0
21.0
39.5
44.0
24.0
60.5
7.5
5.0
21.0
19.5
40.5
74.4
50.5
15.5
.5
12.0
1.7
29.5
5.6
12.3
8.1
8.3
20.5
4.6
.2
1,026.8
-------
436
FIGURE C-27. YEAR 2000 POPULATION DENSITY
-------
437
FIGURE C-28. YEAR 2000 EMPLOYMENT DENSITY
-------
438
FIGURE C-29. PLAN A-5 IN JRPP HIGHWAY ANALYSIS
-------
439
•;'•' /,J~ IK . •srdVro"-*-. x -rf M r
FIGURE C-30. NEW CONSTRUCTION FOR PLAN A-5 IN
JRPP HIGHWAY ANALYSIS
-------
440
2) Transportation Modeling
Even once population (level, allocation and staging), land use,
and highway network configuration have been agreed upon, the transpor-
tation planning process is not complete. The vehicle trips--their number
and distribution, both spatial and temporal--taken by that population
must be forecast and the resulting vehicle-miles-traveled (VMT) assigned
to links on the roadway network. The important parameters from an air
quality standpoint are both the VMT and the average roadway speed on each
link of the highway system for each hour of the day. Using these quanti-
ties, vehicle emissions--CO, NOX and HC--can be determined.
A brief word about the modeling process itself is helpful prior to
reviewing the status of specific model results. In an effort to quantify
the gradients of "travel desire" that lead to trip generation, the geo-
graphical region of interest is devided into a gridwork of traffic zones.
There are 654 such zones in the Metropolitan Denver region, for example.
Each of these has been paired with every other to establish a relative
ranking of attractiveness. The result is a matrix of interzonal attrac-
tions used as weighting factors in assigning total forecast VMT in indi-
vidual areas and highway links.
Having modeled the relative attractiveness of individual areas, it
is necessary also to determine the number and type of trips generated
between zones. This is done by first stratifying trips by type. Then
the number of trips of each type are determined by relating each to the
various socio-economic characteristics of the people taking them. This
is frequently done through regression analysis relating trip type and
level to such variables as land use, household type and size, population
and family size, and employment. Typical trip types are the following:
home based (work, shopping, other); non-home based; truck; external -
internal (to and from outside the modeled region); and external-external
(through traffic). Also, reliance is often place on Origin and Destina-
tion (O&D) surveys, the last one of which was completed in Denver in 1971.
-------
441
The distribution of the trips generated by the above methods is
determined by examining the propensity of people to travel. Since travel
time is perceived as a cost, most people prefer to travel as short a time
as possible in performing a given trip function. Their disincentive
with increasing distance or time can be expressed in probabalistic terms
by a trip length distribution curve incorporating the concept of average
trip length. Also, in determining the pattern of trip-taking, modal split
must be considered, i.e., the relative proportions of trips taken by
private vehicles and public transit. Since costs are perceived somewhat
differently by the users of both modes, different trip length distributions
are modeled.
The final component of the modeling process is the trip assignment
process. As a preliminary to this, it is necessary to determine minimum
time or distance paths between each of the traffic zones. Travelers
desiring trips between any two traffic zones are then assumed to prefer
these optimal paths. Some modeling algorithms simply assign all such trips
to the same path up to the capacity of the roadway links along that path.
Excess trips are then off-loaded to next-best paths. More sophisticated
models attempt to account for congestion effects by assuming that at some
point prior to traffic reaching full capacity drivers begin to voluntarily
avoid that path, off-loading onto secondary routes. When that point will
occur and how far drivers will be willing to detour in distance or time
again depends on their trip length distribution function.
After validation and calibration against known roadway loading data
(trip counts), a model comprised of the previously described four compon-
ents is ready for use, the trip generation, trip distribution, modal split
and trip assignment segments all functioning properly. Among the planning
agencies in the Denver region, three evolutionary versions of such models
have been used, referred to as the first, second, and third generation
transportation models.
-------
442
Early modeling efforts within the JRPP were conducted within legal
time constraints imposed upon the RTD. As a result the initial version
of the model was completed before results of the regional 1971 Origin
and Destination survey were processed. Two distinct generations of
modeling thus were used, the first (1st Generation) based on the results
of a 1959 O&D survey and the second (2nd Generation) incorporating 1971
data as it became available. The 2nd Generation Model also incoporated
more sophisticated algorithms.
It is the differences between the 2nd and 3rd Generation Models that
have direct significance to the current planning process, however. Trans-
portation modeling results upon which most planning has been based within
the DRCOG, the CDH, and the RTD were generated by the 2nd Generation Model.
Emissions data available from those agencies for air quality analysis is
thus 2nd Generation vintage. To the extent the 3rd Generation Model
produces different results (link VMT and speed), the air quality analysis
may also differ, although the kinetics of air pollution chemistry can
exhibit pronounced non-linear behavior. Also, a summary of the differences
between the 2nd and 3rd Generation models is particulary important because
recent DRCOG transportation studies have begun to incorporate the 3rd
Generation .Model. Significant difference seems to exist between total daily
VMT projections in the year 2000: 38,000,000 VMT per day with the 2nd
Generation Model and just under 30,000,000 VMT per day with the 3rd Gene-
ration Model. This is a gross difference of over twenty percent.
According to a recent letter from David Pampu, the Assistant Director
for Planning for the DRCOG, to William Geise, Jr., Chief of the Environ-
mental Evaluation Branch of the EPA, Region VIII, the differences between
the two transportation models can be summarized as follows. The theoretical
bases of the two models were developed independently. The 2nd Generation
Model was based on the Pratt Marginal Utility Model. This 3rd Generation
Model, also known as the Unified Travel Patterns Model, was based on the
n-Dimensional Logit Model.
-------
443
Network building procedures are similar in both models, each using
UMTA discrete network link coding methods. The chief difference in this
area lies in the means of accounting for the disfavor in which the public
holds mass transit. The 2nd Generation Model explicitly accounts for
this through disutility weighting factors. The arbitrariness introduced
through the choice of these factors was removed by eliminating them in the
3rd Generation Model.
Trip generation is similar in both models. The only difference lies
in the elimination in the 3rd Generation Model of a separate trip category
for airline passengers. Revised estimates of airline traffic have greatly
reduced the regional significance of this trip category.
Even though both generations use a gravity model, the algorithm used
for trip distribution represents the most significant difference between
the 2nd and 3rd generations. The 2nd Generation Model distributes trips
by trip purpose (but not by income) independently over the highway and
transit system. It then uses a balancing technique after modal split to
combine the trip tables for both. The 3rd Generation Model, on the other
hand, utilizes a single composite network representing the three available
modes: auto driver, auto group and transit passenger. A compound impedance
is formed by summing the inverse of the inverses of the individual mode
impedances, i.e., summing link impedances in parallel. Trips are then
distributed over this theoretical network by trip purpose and by income.
Differences exist in modal split methods as well. The 2nd Generation
Model uses manual curve fitting technique to relate transit ridership to
a logarithmic form equation in which transit travel time and costs are
subtracted from auto totals. The 3rd Generation Model utilizes a part-
ially disaggregate modal choice model (used to account for travel time
cost). The probability of using a mode is then assumed equal to the indivi-
dual impedances, raised to the base "e," divided by the sum of the impe-
dances, each raised to the base "e."
-------
444
Traffic assignment is also performed differently, with important
variances arising. Although both models utilize a capacity restraint
function, the 2nd Generation Model simply reloads the entire trip table
three times and then averages the results. As a consequence, the process
tends to overload the freeway system. The speed table input was chosen
to represent calibrated speed for 1971. Because the model could only
lower speeds and not raise them, successive program iteration resulted in
speeds that were considered by the DRCOG to be "very low." The 3rd Gene-
ration Model, however, uses an initial Level of Service speed table and
an incremental loading process. With this method ten percent of the trip
table is loaded onto the network, speeds are recalculated, and then the
next ten percent of the trip table is loaded. Successive application of
this process results in both less overloading of the freeway network and
a more accurate representation of speed.
In air quality terms the most significant differences are the daily
link VMT and speed histories. As mentioned earlier the forecast of total
VMT in the year 2000 is 38,000,000 for the 2nd Generation Model and about
30,000,000 for the 3rd Generation Model. It seems that the differences
in the models distill down to shorter trips and higher overall speeds
(for the same highway and transit network). This would suggest that
use of results from the 2nd Generation Model might be somewhat more con-
servative from an air quality perspective, since greater mileage and lower
speeds both can have detrimental effects on pollutant emissions.
It also seems clear that future DRCOG transportation modeling efforts
will employ the 3rd Generation Model. A Year 1975 model validation run
is nearly complete. A run incorporating the Year 2000 JRPP Plans is
planned next. The first iteration has been performed, with completion
expected in early December. Intermediate year runs can be expected in
the Spring of 1977.
-------
445
4. SUMMARY AND RECOMMENDATIONS
a. Summary
This task report has been intended as an overview of the planning
process in the Metropolitan Denver region, specifically considering
those planning issues and agencies having crucial bearing on air quality.
Because the analyses to be performed by SAI will be so heavily driven by
the assumptions underlying the input data supplied to it and because
several permutations of that data exist within the agencies providing it,
a precise, detailed examination of the relevant planning issues is an
essential preliminary to the study process.
The results of this task report will be drawn upon heavily in struc-
turing SAI's analysis efforts. Specifically, a base case regional develop-
ment scenario must be identified along wi.th a complete catalog of the
important uncertainties associated with it. A carefully constructed
sensitivity analysis must then be conducted to examine the air quality
effects of as broad a range of these uncertainties as possible. This
report, through its examination in some depth of the critical elements of
regional planning, will assure the broadest perspective in SAI's efforts.
Brevity in this section is as important as attention to detail was
in the last. Adhering to that principle, it is sufficient to outline
by use of a table the status of planning within each agency having input
to the air quality analysis. In Table C-23 such a summary is provided.
Also, in Figure C-31, a schematic is presented illustrating those plannina
alternatives that have a bearing on air quality analysis.
b. Re c omme n d a t i o n s
Since the pollutant emissions files available for input to the SAI
model are provided by the Colorado Division of Highways, and because
-------
TABLE C-23. SUMMARY OF TRANSPORTATION PLANNING ASSUMPTIONS
Transportation
Agency
Denver
Regional Council
of Governments
(DRCOG)
Colorado
Division of
Highways
(CDH)
Regional
Transportation
District
(RTD)
Air Pollution
Control Division,
Health Department
(APCD)
Study Vintage
1973 JRPP Plan
analysis
Current
Current and
Last Several
Years
Early Transit
Study (1974)
Later Transit
Study (1975)
and Current
Denver AQMA
Study (1975)
Current
Population*
Cycle 3
Subarea Allocation
Plan
Cycle 3
Cycle 3
Cycle 3
Cycle 3
Cycle 4
Land Use
JRPP Plan
(Cycle 3)
JRPP Plan
JRPP Plan
(Cycle 3)
JRPP Plan
(Cycle 3)
JRPP Plan
(Cycle 3)
JRPP Plan
(Cycle 3)
JRPP Plan
(Cycle 4)
Plan
JRPP Plan
JRPP Plan
JRPP Plan
JRPP Plan
Existing Plus
Committed
(E+C)
JRPP Plan
JRPP Plan
Model
2nd Generation
3rd Generation
2nd Generation
2nd Generation
System Management
Contractor (SMC)
n-Dimensional
Logit
2nd Generation
2nd Generation
01
* All population allocation plans shown here total 2,350,000 in the year 2000.
-------
447
POPULATION PLAN
* CYCLE 3 - TRANSPORTATION
PLANNING BY DRCOG AND RTD
* CYCLE 4 - AREA SOURCE EMISSIONS,
AIR QUALITY ANALYSIS, AND
208 ALLOCATIONS BY APCD
(HEALTH DEPARTMENT)
* SUBAREA ALLOCATION PLAN
RECENTLY APPROVED BY DRCOG;
ALLOCATION BY URBAN SERVICE AREA
LAND USE PLAN
* ADOPTED JRPP PLAN - USED BY
ALL AGENCIES; CYCLE 3 AND 4
CONSUMPTIONS AS INDICATED FOR
POPULATION PLAN
* NEW PLAN - MAY BE DEVELOPED
BY DRCOG IN SPRING 1977
TRANSPORTATION PLAN
* ADOPTED JRPP PLAN - USED BY
DRCOG AND CDH
* EXISTING PLUS COMMITTED (E+C) -
USED BY RTD
FINAL AIR
QUALITY MODEL
ASSUMPTIONS
TRANSPORTATION MODEL
* 2ND GENERATION - USED BY
DRCOG AND CDH; FORECASTS
A YEAR 2000 DAILY VMT OF
38 MILLION
* 3RD GENERATION - CURRENTLY BEING
USED BY DRCOG TO REVISIT
TRANSPORTATION PLANS; FORECASTS
A YEAR 2000 DAILY VMT OF JUST
UNDER 30 MILLION
FIGURE C-31.
PLANNING ALTERNATIVES HAVING A BEARING
ON AIR QUALITY
-------
448
neither time nor funds exist for alternative transportation modeling runs,
in practical terms it is their assumptions that define a base case. As
shown in Table C-23, data from the CDH has embedded in it Cycle 3 popula-
tion allocation, the adopted JRPP land use and transportation plans, and
the use of the 2nd generation transportation model. It is important that
the sensitivity analysis conducted about this base case consider varia-
tions wide-ranging enough to encompass the important alternatives to
this set of assumptions.
It appears that the differences in population allocation are most
pronounced in the high growth areas, e.g., southeast Jefferson County,
Lakewood, Aurora, and the urban portion of Arapahoe County. In reviewing
the magnitude of those differences a perspective should be maintained about
the allocation alternatives. Cycle 4, as an extension of Cycle 3, can be
assumed to have corrected some of the faults of that version. Its use
would thus seem more current. However, only the Subarea Allocation Plan
(SAP) has been formally approved by the DRCOG, and it was recommended
by their staff for use in the Denver Regional Clean Water Program for the
final Water Quality Management Plan. This latter point suggests that
participants in the water quality elements of the Overview EIS may find
necessary the use of the Subarea Allocation Plan figures.
Fortunately, from an air quality point of view, Cycle 3 and the SAP
are fairly close in Jefferson County. Although the information reviewed
thus far presents this conclusion only on an aggregate county-wide basis,
if the comparison between the two holds uniformly throughout the county,
southeast Jefferson County, Lakewood, Arvada, and parts of Westminster
can be assumed comparable between Cycle 3 and the SAP. This is particu-
larly helpful in southeast Jefferson County, since the prevailing
southwest-to-northeast flow of wind during the morning rush hour can
emphasize the effect of emissions in that region.
Cycle 3 and the SAP do not agree well in Denver County, where
the SAP allocation more closely approximates Cycle 4. The chief dif-
ference seems to lie in the assumption of low growth in the county
-------
449
in Cycle 4 and the SAP. A sensitivity analysis about a base case using
Cycle 3 might therefore consider varying population related emissions
in Denver County. If, however, the allocation difference can be ideal-
ized as distributed uniformly throughout the county, localized effects
could be considerably diminished due to the county's relatively large sur-
face area. With such a simplifying assumption, incorporation of such
differences in the sensitivity analysis might not be necessary.
Allocation differences in Adams and Arapahoe Counties might be con-
sidered in the sensitivity analysis. However, staging in the high growth
area of Arapahoe County indicates such growth will not occur until 1990-
2000, well after attainment of Federal air quality standards will be
required. Also, growth in Adams County seems to be most likely in Aurora,
particularly in the southeast part of the city. Due to the prevailing
winds, emissions from this area would be expected to have different effects
than would emissions closer to the Denver CBD.
It should be noted that each of the three allocations sum to 2,350,000
in the year 2000, since this figure is a cornerstone in DRCOG/JRPP planning
To the extent this figure may be too high, due to declining birth rate or
inmigration levels, emissions may also be too high, though not necessarily
by any linear relationship. Long-term historical trends suggest that both
these population effects seem possible, although it may still be premature
to so conclude. Some accounting, however, for uncertainties in the popu-
lation level could be made in the sensitivity analysis by adjusting popu-
lation by some uniform factor on a regional basis.
A serious obstacle lies in the way of a direct examination of the
effects of these allocation differences, just as it does for the effects
of differences in land use and transportation plans. The most pronounced
effects of growth on air quality lie in automobile usage patterns and
VMT levels. Additional population in an area,' however, may produce its
increased auto emissions in some other area. This is particularly true
during rush-hour commutes. Varying emissions in the growth areas then
-------
450
may not consider all the detrimental effects on air quality that growth
might produce. The only analytically consistent way to model such effects
is to use a full transportation model to reload the altered population
level onto the roadway system. Since the Overview EIS has neither time
nor money to perform large scale variations using a complete transportation
model, a simpler surrogate for such models must be sought.
The highway network assumed for the planning process can provide clues
for inferring the highway usage patterns of incremental population amounts.
Growth in outlying suburbs, for instance, might be expected to travel
inward during the morning rush hour towards either the Denver CBD or one
of the other employment centers. Since relatively long distances are
involved, it might further be anticipated that such travel would occur on
freeways or principal arterials. The link-node representation of the high-
way network could be used to isolate those links over which travel might
occur. Emissions along these routes could be adjusted in a sensitivity
analysis, thus avoiding a major reloading of the transportation model.
Clearly differences in the highway network can have pronounced effects
on pollutant emissions, as the traffic redistributes itself on the road-
way system. Implied in the base case identified earlier is the use of
the JRPP Highway Plan. Objections to this plan were raised by UMTA during
the RTD planning process. They suggested that the E+C network might
represent a more realistic roadway system. Whether or not this is true,
the extensive transportation modeling analysis required to make such
modifications seems prohibitive in the time frame of the Overview EIS.
Air quality results then will have to be judged with the highway network
as an implied caveat.
Other things being equal, it would seem that loading a given VMT onto
a smaller roadway system, such as the E+C network, could increase congestion
considerably and elevate pollutant levels much more than proportionally.
The base case would seem from this standpoint an optimistic one. Air
quality results might well be more favorable, although the interdependence
of travel patterns, VMT and congestion could alter considerably the level
and distribution of growth.
-------
451
Another significant area of uncertainty is that due to the trans-
portation modeling process. Projections of total daily VMT in the year
2000 differ by as much as twenty percent between the 2nd and 3rd generation
models, a considerable amount. Once again the extensive modeling expense,
time and effort needed to consider these differences in a sensitivity
sense seem beyond the resources of the Overview EIS. Preliminary results
of the 3rd Generation Model, however, suggest for the JRPP highway net-
work that trips are shorter, a conclusion supported by the lower Year 2000
daily VMT totals. This could lead to an easing of roadway congestion and
higher link speeds. The error involved in using the 2nd Generation Model
over the 3rd Generation Model, however, is in the opposite direction from
that resulting from use of the JRPP highway network over the E+C network.
The two errors seem countervailing, although not at all necessarily
compensating.
This final discussion has been intended to raise and examine a number
of the issues that must be considered in structuring the air quality analy-
sis. Though much remains to be resolved, it is only the level of detail
in this task report that permits the subtleties involved in the planning
process to be recognized. It should also be obvious that many of the
issues raised in the discussion of the sensitivity analysis are also appli-
cable to the analysis of mitigation measures. The important uncertainties
must be accounted for in both. The issues, however, remain the same. It
is hoped that this report has shed a useful illumination upon them.
-------
452
ANNEX C-l
A SIMPLIFIED POPULATION FORECASTING METHOD
a. Introduction
Growth—how much and where it will be distributed—is at the heart
of the planning process, serving as a key determinant of the level and
staging of demand for public services. Because of this importance, con-
siderable effort is usually expended in its determination. Detailed
models are frequently used to forecast population level and cross-section,
stratifying individuals by age group and applying in a probabalistic sense
the appropriate age-group fertility and death rates as well as migration
levels. Such models divide population into two components: net
natural increase (the net of births over deaths), which is proportional
to the current population; and migration, which is an annual level,
usually historical and assumed independent of current population. This
type of population level forecasting is often referred to as a "direct"
method, and an example of it is the "cohort-survival" technique used as
the basis of DRCOG efforts in the Metropolitan Denver region.
Because of the complex set of assumptions embedded in these models,
their projections are often difficult to assess. For this reason it is
helpful to distill the full set of assumptions down into two more concep-
tually convenient parameters: net natural increase and migration. All of
of the more complex assumptions about fertility and death rates by age-
groups can be collapsed into a single composite net natural increase
rate which can be applied directly to the aggregate current population.
Migration can similarly be reduced to a single aggregate level.
The above procedure allows the development of a simplified popula-
tion projection method, useful for forecasting total population levels.
Though not as accurate as a fully stratified "cohort-survival" model,
-------
453
results of the simplified procedure developed in this discussion compare
favorably with the aggregate population levels forecast by the DRCOG using
the more complex model. It also provides a convenient means to evaluate
analytically sensitivities of population forecast to the net natural
increase rate or the migration level.
In the following pages, the simplified method will first be derived,
sensitivities will be determined, and finally a brief review of the DRCOG
final forecast population will be presented.
b. Derivation of the_Simplified Method
The simplified method relies for its final closed-form solution on
the decomposition of the population growth problem into two parts. The
first, the net natural increase rate, is assumed to be proportional to
the current population. The second, the migration level, is assumed to
be a fixed annual figure independent of the current population. In rela-
tional terms the differential equation incorporating these assumptions
can be written as
p * rP + m (i)
*
where P is the instantaneo.us rate of change of the total population,
P is the current population, K* is the net natural increase rate, and
W is the migration level.
This equation can be directly integrated to solve for the time-
varying form followed by total population. To do so, it must be recogn-
ized that the solution to Eq. (1) has two components, one a solution to
its homogeneous form and the other a particular solution. The homogeneous
form of Eq. (1) is
which has as its solution
-------
454
where /I is constant whose value is determined by the problem's boundary
conditions. By inspection, the particular solution to Eq. (1) has the
form
P * - J2- (4)
*f> r
Applying the superposition principle the homogeneous and particular
solution to Eq. (1) can be summed to give the full solution, i.e.,
The initial boundary condition, i.e. , ^i» » /» , can be applied to
solve for the unknown constant, & . The value of A thus determined is
/>-(/5
where rc is the population in the initial year of the simulation period.
Substituting Eq. (6) into Eq. (5) yields the final form of the
solution, i.e.,
It is this equation that expresses the growth of total population over time.
Results of this equation can be verified by comparing them with
several alternatives considered by the DRCOG. In their "Appraisal of
the DRCOG Policy Population Forecast" (August 1975), the staff of the
DRCOG detailed the eight alternative population growth cases that they
considered in their studies. Case IV, for instance, assumed a net natural
increase rate between 1970 and 2000 of 8.5/1000 persons (1972 rate) and
-------
455
a migration leva! of 42,000 persons annually (the 1970-73 average level).
The resultant 1980 and 2000 populations were 1.8 and 3.0 million, res-
pectively. The simplified form in Eq. (7), using the same assumptions,
reproduced these values to the same level of significance.
As another example, Case I, which represented the DRCOG "preferred"
alternative, also assumed a net natural increase rate of 8.5/1000 persons,
but a 1960's level migration (15,700 persons annually) through 1980 and a
1940-1970 thirty-year average level (15,100 persons annually) from 1980-
2000. The resultant 1980-2000 populations were 1.6 and 2.2 million. The
simplified method projected 1.5 and 2.1 million, values not quite the same.
If 1975 is used as the base year, using the actual population in that year,
the 1980 and 2000 forecasts become 1.6 and 2.2 million.
c. Population Sens it jyjtie s
Using the closed-form expression for total population growth derived
in the previous section it is possible to determine analytically the
sensitivity of forecast population to the two basic input assumptions,
the net natural increase rate and the level of migration. This is done
simply by taking the partial derivatives of Eq. (7) with respect to each
of the two independent parameters.
The analytic sensitivities thus determined are expressed in the follow-
ing form:
:VP / r_ *•#-*•} .7
(8)
£K» V>" rvv~ -vj\~ - r*jc. -ij (9)
where these are the sensitivities of total population to migration level
and net natural increase rate, respectively.
-------
456
d. Review of DRCQG Forecast
The population growth alternative recommended by the DRCOG staff
was Case I, with a Year 2000 population level of 2,175,000. This figure
was subsequently modified by the Council of the DRCOG to add 175,000,
principally in southeast Jefferson County. The final approved total was
2,350,000. The assumptions implicit in this figure differ from those in
Case I. Although the net natural increase rate of 8.5/1000 persons
remained the same, the migration level is higher in the approved forecast.
The migration level embedded in a population forecast may be deter-i
mined approximately by inverting the simplified relation in Eq. (7). The
migration level thus determined can be expressed as
(10)
where P is the population in a future year f and ^ is the population
in the year £0 , with r being the net annual natural increase rate.
This relation can be used to determine the migration level necessary
for the Denver regional population to reach 2,350,000 in the year 2000.
Using the population in 1970 as a base and the net natural increase rate
assumed in both Case I and the approved forecast, Eq. (10) predicts that
a migration level of 22,328 persons annually would be required to reach
the approved Year 2000 figure.
-------
457
ANNEX C-2
NATIONAL AMBIENT AIR QUALITY STANDARDS
Pollutant
Carbon Monoxide
Hydrocarbons
(non-methane )
Type of
Standard
Federal
Federal
Nitrogen Dioxide Federal
Ozone
(Oxidants)
Sulfur Dioxide*
Particulates
Federal
Federal
Primary
Secondary
Federal
Primary
Secondary
State
Non-Designated
Areas
Designated
Time
Interval
1 hour
8 hour
3 hour
(6-9 a.m. only)
1 year
(arith.)
1 hour
24 hour
1 yr . (arith. )
3 hour
24 hour
1 yr. (geo. )
24 hour
1 yr. (geo. )
24 hour
1 yr. (arith.)
24 hour
1 yr. (arith. )
Effective
Year
1977
1977
(see
ozone)
(undetermined)
1977
1975
1975
1975
1975
1975
1975
1975
1970
1970
1973
1976
1980
1973
1976
1980
Concentration
ug/m3 PPM
40,000
10,000
160
100
160
365
80
1,300
260
75
150
60**
150
45
200
180
150
70
55
45
35
9
0.24
0.05
0.08
0.14
0.03
0.5
—
—
__
—
—
—
—
—
—
—
* Refer to table titled "Sulfur Dioxide Ambient Air Standards for the State of
Colorado"
** Federal guideline only
-------
458
Annex C-2 (Concluded)
SULFUR DIOXIDE (SO_) AMBIENT AIR STANDARDS FOR THE STATE OF COLORADO
Micrograms Per Cubic Meter
Maximum Allowable Increments
Over Baseline ug/m3
Maximum Allowable
Concentrat ions
ug/m3
Sulfur Dioxide
Category I
(formerly
non-designated areas)
Sulfur Dioxide
Category II
Sulfur Dioxide
Category III
(formerly
designated areas)
Annual Mean
24-hr, max.
3-hr. max.
3
(.001)
15
(.005)
75
(.026)
15
(.005)
100
(-035)
700
(.245)
60
(.021)
260*
(.091)
1300*
(-455)
( ) = equivalent values in parts per million (1 ppm=2860 ug/m3 at 0°C and
760 mm Hg (Torr)).
* Not to be exceeded more than once in a twelve-month period.
-------
459
REFERENCES FOR APPENDIX C
1. "Continuing Air Quality Review Procedure," Joint Regional Planning
Program, Technical Memorandum 8/03, March 1976 (Prepared by Colorado
Division of Highways).
2. "Air Quality Guidelines," Federal Highway Administration, U. S. Depart-
ment of Transportation, Federal-Aid Highway Program Manual, Vol. 7,
Chap. 7, Sect. 9, Transmittal 105, 26 November 1974.
3. "Guidelines for Analysis of Consistency Between Transportation and Air
Quality Plans and Programs," Federal Highway Administration and En-
vironmental Protection Agency (prepared jointly), April 1975.
4. News Articles, Denver Post, 12, 14 and 22 October 1976.
5. "Denver/Boulder Urban Transportation Planning Process—Certification
Determination," Attachment to Letter to Mr. David Pampu, Executive
Secretary, JRPP and DRCOG, from Theodore G. Weigle, Jr., Regional
Director, Urban Mass Transportation Administration, Department of
Transportation, Letter dated 1 October 1976, and Attachment dated
24 September 1976.
6. Personal Notes, Denver Regional Council of Governments Meeting on
20 October 1976.
7. "Policy Population Forecast: Subarea Population Allocation," Denver
Regional Council of Governments, August 1976.
8. "Appraisal of the DRCOG Policy Population Forecast," Denver Regional
Council of Governments, August 1975.
9. "A Resolution Reaffirming Population Policy and Approving a Policy
for Allocation of the Regional Population Forecast to Subareas of
the Denver Region," Denver Regional Council of Governments, Resolution
No. 18, 1976, Passed on 18 August 1976.
10. "Regional Land Use, Highway and Public Transportation Plans," Denver
Regional Council of Governments, Joint Regional Planning Program,
Draft Summary Report dated 17 October 1973.
11. "Detailed Assessment Report, 1-470," Colorado Division of Highways,
September 1976.
-------
460
12. "North-South Rapid Transit Project, A Synopsis of the Project and
Supplement to the Draft Environmental Impact Analysis of December,
1975," Regional Transportation District, March 1976.
13. "Transit Network Analysis, Summary Report," System Management Contrac-
tor, Prepared for the Regional Transportation District, 24 June 1975.
14. Letter to John Crowley, Chairman of the Regional Transportation District,
from Robert Patricelli, Administrator of the Urban Mass Transportation
Authority, communicating the UMTA funding decision on the RTD-proposed
light-rail rapid transit system, dated 29 June 1976.
15. "Air Pollution Control Act, 1970," State of Colorado, Published by the
Air Pollution Control Division, Colorado Department of Health, 1975.
16. "Report to the Public, 1976, "Air Pollution Control Commission, Colorado
Department of Health, 1976.
17. "Appraisal of the DRCOG Policy Population Forecast," Denver Regional
Council of Governments, August 1975.
18. "Policy Population Forecast, Subarea Population Allocation," Denver
Regional Council of Governments, August 1976.
19. "Transportation System Report," Denver Regional Council of Governments,
Joint Planning Program, 11 June 1973.
20. "Development, Calibration, and Documentation of the Second Generation
Transportation Models," Denver Regional Council of Governments and the
Colorado Division of Highways, Joint Regional Planning Program,
February 1975.
21. "Transit Network Analysis Summary Report," System Management Contractor,
Prepared for the Regional Transportation District, 24 June 1973.
22. "Development, Calibration, and Documentation of the Second Generation
Transportation Models," Denver Regional Council of Governments, Joint
Regional Planning Program, February 1975.
23. Letter to J. William Geise, Jr., Chief of the Environmental Evaluation
Branch, U.S. Environmental Protection Agency, Region VIII, from David
Pampu, Assistant Director for Planning, Denver Regional Council of
Governments, communicating the differences between the 2nd and 3rd
generation transportation models, dated 16 November 1976.
-------
1. REPORT NO. [2 "
EPA-908/1 -77-002
4. TITLE AND SUBTITLE
Air Quality in the Denver Metropolitan Region 1974-2000
7. AUTHOR(S)
G. E. Anderson, S. R. Hayes, M. J. Hillyer, J. P. Kill us,
and P. V. Mundkur
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Incorporated
950 Northgate Drive
San Rafael, California 94903
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency, Region VIII
1860 Lincoln Street
Denver, Colorado 80203
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
May 1977
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
EF77-222
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-4341
13. TYPE Of REPORT AND PERIOD COVERED
Final Report: 10/76 - 4/77
14. SPONSORING AGENCY CODE
TECHNICAL REPORT DATA
(Please read Inuructions on the reverse before completing)
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes an air quality analysis for the Denver metropolitan region for the
years 1976, 1985, and 2000. The analysis was carried out to provide background informa-
tion as to the environmental impact of the urban growth that might be associated with
the availability of new wastewater treatment facilities. Generally improving air qual-
ity is forecast, although exceedances of some air quality standards are projected.
These results are based on physico-chemical computer simulations, using pollutant
emissions forecasts.
Projections of photochemical oxidant concentrations, exposures, and dosages were
obtained with the Denver Air Quality Model. DAQM, developed during this program, is
based on previous Systems Applications, Inc. (SAI) photochemical models. A validation
study -showed that DAQM, without calibration, does not consistently under- or over-
predict peak oxidant concentrations. At least 80 percent of the predictions were
within a factor of two of the observations. Air quality projections were found to be
negligibly affected by major changes in projected land use and less than proportionately
affected by large changes in atmospheric dispersion.
Measures proposed to mitigate adverse air quality were examined. The only measure
identified as having significant mitigation potential was the control of vehicle emis-
sion factors. The effects of that measure were found to be large, complex, and not
predictable by simple methods.
17.
KEY WORDS AND DOCUMENT ANALYSIS
1. DESCRIPTORS
18. DISTRIBUTION STATEMENT
Release to Public
b.lDENTIFIERS/OPEN ENDED TERMS
Unclassified
20. SECURITY CLASS (This page)
Unclassified
c. COSATI Field/Group
21. NO. OF PAGES
460
22. PRICE
EPA Form 2220-1 (9-73)
*U.S. Government Printing Office: 1977-780-080/267 Region 8
-------