EPA-R4-73-012c
March 1973
Environmental Monitoring Series
??8S^:^:;mrara:sffiS^^
&ff&£t^^
I
; :
1
::::::S:::::::::::::S:::::::::::::^
Illllllllllllllllllllllllll
-------
EPA-R4-73-012c
FIELD PROGRAM DESIGNS
FOR VERIFYING
PHOTOCHEMICAL DIFFUSION
MODELS
by
A.Q. Eschenroeder, G.W. Deley, andR.J. Wahl
General Research Corporation
P.O. Box 3587
Santa Barbara, California 93105
Contract No. 68-02-0336
Program Element No. 1A1009
EPA Project Officer: Ralph C. Sklarew
Meteorology Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
March 1973
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
11
-------
CONTENTS
SECTION PAGE
ABSTRACT i
1 INTRODUCTION 1
1.1 Motivation for Designing Experiments to Aid
Model Development 1
1.2 Objectives of Currently Planned Filed
Measurements 2
1.3 Approach of this Study 4
2 REVIEW OF OUR PREVIOUS RECOMMENDATIONS 5
2.1 Summary of Previously Suggested Elements of New
Experiments 5
2.2 Summary of Previously Recommended Auxiliary
Data Analyses • 7
3 CONCEPTUAL DESIGNS OF SOME FIELD EXPERIMENTS 10
3.1 Automotive Emissions Along Major Highways 10
3.2 Plan for Studying Photochemical Transformations 12
3.3 A Spatially Integrated Sampling Technique 14
4 A PROTOTYPE DATA MANAGEMENT PLAN (LARPP) 16
4.1 Advance Planning 16
4.2 Reports to the Program Director 16
4.3 Archives for Modelers , 23
5 BRIEF OVERVIEWS OF FUTURE FIELD PROGRAM OBJECTIVES 31
5.1 Definition of the Role~d£--'Aerosols in Atmos-
fr
pheric Chemistry 31
5.2 Further Understanding ftf^Gas Phase Processes ^ ;'
Affecting Reaction fetes^-- "r32
iii
-------
CONTENTS (Cont.)
SECTION PAGE
APPENDIX THE DEPENDENCE OF APPARENT REACTION RATES ON TURBULENT
MIXING PROCESSES ' 35
REFERENCES 43
IV
-------
ILLUSTRATIONS
NO. PAGE
1 CRT Display Terminal 17
2 Carbon Monoxide Tables (Figs. 2-7 fol-
low page 19)
3 Carbon Monoxide Profile
4 Carbon Monoxide Pattern
5 Flourescent Particle Tracer Gradient Pattern
6 Gas Chromatogram, Peaks 1-13
7 Gas Chromatogram, Peaks 23-32
8 Possible Turbulence Effects on the Ozone Quasi-
equilibrium for 1969 Ground Data at El Monte—
High NO Levels 42
-------
1 INTRODUCTION
1.1 MOTIVATION FOR DESIGNING EXPERIMENTS TO AID MODEL DEVELOPMENT
Computer simulation models adequate to predict air quality in terms
of emissions and meteorology must be improved in terms of fidelity if they
are to be useful in testing implementation strategy plans. Observational
evidence forms a substantial basis for the models we presently possess;
these existing models have each been tested (to a greater or lesser extent)
against the available body of measured data. Because the experimental
community now stands at the threshold of fielding elaborate measurements
programs, such as the Los Angeles Reactive Pollutant Program (LARPP) and
the Regional Air Pollution Study (RAPS), the time is at hand for modelers
to lay out specific data needs.
Recommendations for field programs must address themselves to achieving
specific design goals based on modeling needs. The first step in accomplish-
ing this is to identify gaps in the model calculations, recognizing that
gaps may exist in the input data, in the phenomenology underlying the
governing equations, or in the adequacy of the air quality, measurements.
Next, the variables to be measured must be designated and .any special
conditions of the experimental setting specified. As :a final step, details
of instrumentation, precision, sampling rate, and data displays must be
specified. Much of the current content of field programs embodies our
earlier recommendations; however, the present report refines and extends
these recommendations to address newly identified problems.
To be sure, we have not attempted to carry through all of these steps
in the present study. In some areas we lack answers either because we are
unable to conduct a meaningful experiment or because a reliable instrument
does not exist. In other areas, specifying a comprehensive program design
will depend on active collaboration with the experimenter in the planning
phases.
-------
1.2 OBJECTIVES OF CURRENTLY PLANNED FIELD MEASUREMENTS
1.2.1 Regional Air Pollution Study (RAPS/St. Louis)
The overall objective of RAPS is directed to the development of air
quality simulation models that are useful for predicting the effects of
emission controls on air quality in a region. In achieving this broad
objective, the program has been structured to achieve a set of complex and
interacting subgoals. The structuring of RAPS follows relatively funda-
mental lines to enhance our understanding of related phenomena. It also
has specific tasks for test and evaluation of the simulation models.
Among the fundamental objectives of RAPS are the characterization
of atmospheric motion in the urban airshed and the refinement .of chemical
mechanisms to describe transformation or losses. Laboratory experiments
and.rate constant measurements are considered as adjuncts to the successful
pursuit of the program. Tests of the assembled models as well as of their
components are listed among the RAPS/St. Louis subgoals.
In approaching this set of objectives, the RAPS planners envisioned
six broad work areas:
1. Emission Inventory
2. Atmospheric Transformations
3. Atmospheric Structure and Dispersion
4. Removal Processes
5. Mathematical Simulations
6. Control Economics
The philosophy of the RAPS approach recognizes the deficiency of
past programs due both to a lack of definite goals and to the fact that
data acquisition has been limited by the available measurement instruments.
Rather than relying on targets of opportunity as in the past, the regional
study plan recognizes the need to treat the problem on an extended scale,
both in geographical space and scientific complexity. In the final
-------
analysis RAPS will try to facilitate the development of control strategies
having general applicability. The stated implications touch on social and
economic considerations that must be understood in order to implement
actual control strategies.
1.2.2 Los Angeles Reactive Pollutant Program (LARPP)
The measurements phase of this program is scheduled to begin in
August 1973. The LARPP objectives are more limited than those of RAPS;
however, the orientation toward modeling is the same. For these reasons,
this discussion can be far more specific than that for RAPS. LARPP,
however, is by no means a simple program.
To describe the program directly it is useful to quote the first
2
sentence of the program plan:
Most simply stated, the objective of the Los
Angeles Reactive Pollutant Program (LARPP) is to
provide a data package suitable for developing
and subsequently validating computer models
which effectively simulate the photochemically
induced reactions causing smog.
This is in response to recommendation from a joint (CAPA-3/CAPA-7) committee
of the Coordinating Research Council, Inc. (CRC). In September 1972, CRC
set as a major goal of future field programs the task of acquiring a data
base for reactive pollutant models. They proposed to exploit; the moving
coordinate system concept by moving an instrument package along with an
air mass to the extent possible.
The embodiment of these objectives in LARPP is built around a helicopter/
*
tetroon system for air sampling and tracking. Gas species, tracer density,
and meteorological data will be measured by two helicopters that follow
tetroon clusters across the Los Angeles basin. Radar tracking, mobile van
*
A neutral-buoyancy, lighter-than-air vehicle which flies freely, presumably
following the path of air motion.
-------
sampling, and laboratory analyses constitute the main ground support
capabilities. Fixed wing aircraft missions will collect data on winds,
moisture, and temperature.
1.3 APPROACH OF THIS STUDY
In this paper we follow a chronological sequence in ordering the
major sections. First our recommendations developed in early 1970 for CRC
are reviewed and placed into the present context of test plans (Sec. 2).
The following section (Sec. 3) gives three specific conceptual plans that
can be pursued to shed further light on certain key questions that remain
unanswered. Using the LARPP design as a prototype, the next section
(Sec. 4) presents a data management plan that aids ongoing program direc-
tion in addition to providing an archive for modelers. We close (Sec. 5)
with a discussion of areas for future field programs. In contrast with
the specific conceptual plan section, this overview of the future is
heuristic in nature. Its incompleteness is unavoidable because of limita-
tions both in theory and instrumental technique.
-------
REVIEW OF OUR PREVIOUS RECOMMENDATIONS"
Some years ago, in an effort to improve the communication between
theoretical and field programs, we suggested new directions for future
field measurements of air pollution (under Coordinating Research Council
sponsorship). Although many of these suggestions have since been imple-
mented, there is a continued need for the fulfillment of their objectives.
It may be noted that LARPP will be a proving ground for some of the methods
proposed in the recommendations. The results from LARPP may help to achieve
an understanding of the physical and chemical processes that determine air
quality under specific emission source conditions and meteorological
settings using tested methodologies.
2.1 SUMMARY OF PREVIOUSLY SUGGESTED ELEMENTS OF NEW EXPERIMENTS
The philosophy underlying the approaches initially suggested empha-
sized experimental design rather than after-the-fact analysis of available
data. Although the RAPS philosophy does not rule out new instrumental
development to achieve its goals, our current experimental design recommenda-
tions emphasize new uses and configurations of available measurement tech-
nology. In considering novel techniques it is profitable first to summarize
the main features of program innovations that were suggested earlier.
Considering the air over a city as a dispersion and reaction region
for air pollutants, we observe that the data from ground-based fixed moni-
toring stati9ns may be hampered by "wall effects." For assessing potential
damage to receptors, current ground level measurements are essential;
however, for gaining a deeper understanding of atmospheric processes,
ground-based data provide only a small part of the needed information.
Thus, in our recommendations we emphasized airborne measurements over a
few typical days at the possible expense of extensive statistics at a few
fixed stations in order to control program costs. Photochemical modeling
3
studies have shown that this shift in emphasis is essential for regional
studies.
*
Based on work sponsored by CRC-APRAC Project No. CAPA-7.
-------
Vertical profiles of temperature and concentration hold the key to
the ultimate effects of air pollution in an urban area and its vicinity.
A combination of modern monitoring instruments and available airborne
platforms can supply the needed measurements. To this end we asked that
helicopters or tethered balloons fitted with compact measuring equipment
be utilized more extensively than in prior programs. This approach permits
achieving two important goals.
1. Vertical profiles can be obtained over a meaningful height
interval near the ground (measurements at 1000 feet and above
usually give only the conditions at the upper edge of the
mixing layer).
2. Air masses can be tracked in the micrometeorological environ-
ment of an urban airshed (the wind seldom carried the pollu-
tants over just those ground-based stations where pertinent
quantities are being measured). To improve this aspect of
the work, we suggested improved ways to follow tetroons with
a helicopter, using tetroon and helicopter transponders,
along with a voice-link.
All of the advanced air pollution simulation models consider stability
effects on dispersion, horizontal advection, and the vertical spread of
pollutants after they are emitted. Many useful check points are obtained
by arranging for flights of instrument packages through the polluted layers
of air. Therefore, gathering data arrayed in three dimensions is essential
to the validation of the models.
Chemical concentrations of interest are carbon monoxide, hydrocarbons,
oxides of nitrogen, and ozone. Because of its slow reactions, CO acts as
a tracer and indicates the advective and diffusive spread of pollutants.
It is also rather specific to vehicular sources. For the same reasons,
acetylene should also be considered as a candidate for a tracer material.
For hydrocarbons we would recommend eliminating detailed analyses, which
resolve individual compounds, in favor of more frequent readings of
-------
a coarser chemical resolution. The breakdown only needs to include broad
reactivity classes (even if they have to be as coarse as total hydrocarbons
and methane).
Total oxides of nitrogen would help in solving the puzzle of the
nitrogen balance, which continues to be manifested in observations and
3
which we indicated in our early studies. Individual measurements of
oxides of nitrogen, of course, indicate the degree of conversion from NO
to N0_. This, combined with the ozone measurement, holds the key to
determining the progress of the main transformations involved in photo-
chemical smog. Among our suggestions on the chemical techniques were
tests of bag samples for aging effects that might shift the composition
and tests for rotor downwash interference with sampling from a helicopter.
We also offered recommendations on improved data communications.
The use of modern computer/telecommunications interfaces could provide a
real-time data base of "quick look" assessments of the measurements pro-
gram. Prescribed standardized data formats would allow analysts to design
input/output portions of computer models that would allow fast-response
feedback to the field program. This would greatly enhance rapid trans-
mission of the results to the community at large and encourage wider
participation in the study.
2.2 SUMMARY OF PREVIOUSLY RECOMMENDED AUXILIARY DATA ANALYSES
The analysis of data taken in previous field programs remains incom-
plete. Specifically, the measurements made by Scott Research Laboratories
in their Los Angeles Basin and New York City programs contain information
which could be of use in RAPS planning. Since this information has been
available for so long, it is not likely that its analysis will receive
much attention. There is valuable information yet to be gleaned from
these data. The ultraviolet statistics from both programs should be
studied in conjunction with weather records to determine if attenuation
of UV radiation due to smog aerosol is a significant feature that should
-------
be added to photochemical/diffusion models and to field programs designed
to validate them. This is especially true of the airborne ultraviolet
data (sparse as it is) .
Presumably the composition of the St. Louis atmosphere will be some-
where between that of Los Angeles and New York. (The former is charac-
terized more by photochemical oxidant, and the latter by sulfur dioxide
and particulates.) For this reason it seems essential to process the
(Scott) New York City data to at least the same extent as we did for the
4
1968 and 1969 Los Angeles Basin programs. This entails calculating diurnal
histories of hydrocarbon reactivity statistics. Profiles of reactivity
distribution would show the composition variation of the hydrocarbon as
it is attacked. Key ratios give important clues-regarding material
balances in the atmosphere. We examined CO/NO and C»H0/NO ratios
x 2 2 x
for Los Angeles; however, it would be of interest to add SO to the
X
ratio tests for New York City. Some surprising departures were observed
in the quasi-equilibrium test, leading to the hypothesis of turbulent
mixing interference explained in the appendix. The processing .codes fo-r
these analyses are still intact and operative from the previous work.
It is mainly a matter of identifying the work plan and modest resources
needed to complete these tasks.
These same codes can be integrated into the software for a quick-look
capability to back up either LARPP or RAPS data acquisition in a real-time
mode. With appropriate data links, the various contaminant concentration
ratios and reactivity indices can be fed back to the control center to
assist in planning future actions and in diagnosing any current problems.
Other pre-RAPS analyses should be directed toward a feasibility study for
running air quality simulation models nearly in real time, parallel to the
program of measurements. Side-by-side calculations using various available
models may enhance this aspect of the study by employing a tightly coupled
interaction between simulations and field data. The purpose would be to
-------
relate the future-time situation map to the present-time situation map
in order to deploy mobile sampling units and special test instruments. The
real-time operation would, therefore, provide guidance for testing certain
special aspects of a model (e.g., pollutant behavior at convergences or
division of flow around terrain features). Practically on-line operation
of a model is also useful in that it provides a timely assessment of the
adequacy of the data base to serve the model development efforts.
-------
3 CONCEPTUAL DESIGNS OF SOME FIELD EXPERIMENTS
3.1 AUTOMOTIVE EMISSIONS ALONG MAJOR HIGHWAYS
The determination of emission rates of the three pollutants CO, HC ,
and NO along major highways can be done in two steps: the first is
X
computational and the second is an experimental verification.
In the computational approach, certain basic measurements must be
used for inputs. These basic data are traffic statistics, vehicle emission
statistics by model year, and a vehicle age/mileage distribution. The
traffic statistics need not be finer than hourly vehicle counts and average
traffic speeds; the time resolution of meteorological data is seldom better
than this. Ten-minute resolution intervals would be warranted probably
for special tests to verify emissions experimentally.
Vehicle emission statistics are usually available in terms of a
standard driving cycle. The mean of a sample should be used for each
individual model year. Speed corrections- and removal of cold-start emissions
must be applied to the standard emission factors to approximate more nearly
the driving modes on a major highway. This procedure has been well estab-
lished in our recent EPA evaluation studies of photochemical/diffusion
modeling.
An aggregate emission factor is derived from the vehicle age/mileage
distribution by taking a weighted average of the mean model year factors.
The weighting coefficients are formed from a composite function of both
the percentage of the vehicles and the average annual mileage put on a
vehicle in the particular age bracket in question. This combination pro-
vides the likelihood of finding a vehicle within each age bracket and hence
a certain mean emission factor. Multiplying the aggregate emission factor
(corrected for cold-start and speed) by the traffic density gives the
roadway emission intensities.
10
-------
Refinements can be applied to this procedure, but the data volume
needed would be beyond the scope of most field programs. One of the
refinements involves using a traffic velocity distribution function to
account for the spread in emissions around that at the mean speed. Another
would be to use specially measured mode emission factors (e.g., accelera-
tion from 50 to 60 mph, cruise at 65 mph, and deceleration from 70 to 60 mph)
along with mode distributions for the traffic dynamics.
A spot check field program should be used to verify the computational
emissions model. This is best done along a stretch of major highway at
grade level, with few or no other emission sources nearby, and with a
direction nearly perpendicular to the prevailing wind. Simultaneous car
counts, traffic speed estimates, meteorological measurements, and air
quality measurements should be taken at the chosen site. The objective
of the experimental check is to run a mass balance between pollutants in
(from computations based on traffic dynamics) and pollutants out. Part
of the input will be pollutant advected in. Meteorological measurements
and air quality measurements will be taken around the periphery of an
Eulerian (ground-fixed) control volume aligned along the roadway. The
choice of crosswind conditions eliminates the difficulty of accounting
for advection in and out of the ends of this imaginary box.
Assessment of the computational approach will be expressed as a
percentage deviation between measured and computed pollutant fluxes. To
be sure, this is not an absolute test of the computational approach because
the experimental checkout results will also contain some assumptions (e.g.,
profile shape, quasi-steadiness, and the like). The assessments should be
made for a variety of sites, wind speeds, and traffic flows. The computa-
tional model may turn out to show 'systematise biases. For the purposes of
v
emission modeling, such biases may form the'bases of correction factors.
The use of the computational model for other roadways will, therefore,
eliminate the need for further extensive measurements.
11
-------
3.2 PLAN FOR STUDYING PHOTOCHEMICAL TRANSFORMATIONS
Since the Lagrangian coordinate frame (moving with the air mass) is
intrinsic to air pollution modeling, it seems natural to track an air mass
with an aircraft. Previous attempts at this have suffered from basic
logistical difficulties. Air traffic restrictions prevented flights at
altitudes and headings of interest in some cases. The use of tetroons
for marking air movement required visual tracking of the tetroons by a
helicopter, which was often impossible because of limited visibility.
Consequently, data in previous programs often had a series of air samples
being taken along a totally different flight path than that followed by
the air.
Remedies to these problems should be implemented in field programs.
Use of government aircraft may permit more altitude flexibility than
before because of exceptions to flight rules that may be made. Mounting
a transponder on the helicopter as well as on each of the tetroons will
allow simultaneous tracking by ground-based radar. Differences between
the tracking signals can be used to guide the helicopter by voice communi-
cation link.
Multiple tetroons flying simultaneously will enhance the certainty
as to the mass center location of the air parcel. Also, if a tetroon hangs
up on a building or hits the ground, the remaining tetroons can still com-
plete the mission. The auxiliary use of stratified multicolored particulate
tracers affords a means of checking out theonulti-tetroon approach. The
feasibility of these techniques will have been established by the Metronics,
Inc., tracer project now under way in the Los Angeles Basin under sponsor-
ship of the state of California.
Much ofxthis subsection stems from discussions with Dr. William A. Perkins,
Jr., on the experimental designs for LARPP. Consequently, many elements
of that program are found here since the documents" were written simultaneously.
12
-------
Although it may still be difficult, the airborne measurement of
carbon monoxide is an essential part of the program to follow air parcels.
The photochemical modeler uses this information to check out the diffu-
sion portions of his code. The rapidly reactive pollutants are also of
interest. It is difficult to list them in any order of importance, but
ozone, oxides of nitrogen, and nonmethane hydrocarbons (as an aggregate
group) are of interest.
An important adjunct to the flight program should be a highly
localized study of the balance among oxides of nitrogen in a three-
dimensional Eulerian control volume. This subprogram takes a form closely
resembling that of the CO-balance studies conducted by Stanford Research
Institute in San Jose, California. The ground area intercepted by the
control volume will be selected as one that has an excellent source-
emissions data base. The airspace of the "side walls" of the control
volume will be selected so that they are readily accessible for helicopter
transects. The thermal inputs and terrain near the control volume should
be uniform enough that wind data from a multiplicity of stations in the
area can be interpreted with some degree of confidence.
Because they may reveal the removal processes, nitric oxide, nitrogen
dioxide, nitrous acid, and organic nitrates should be measured in the gas
phase in helicopter samples and in samples collected from ground-based
fixed intakes near rooftop height. Particulate analyses and test surface
*
analyses must be made simultaneously to aid in ascertaining the fate of
oxides of nitrogen in the urban airshed. Determining this is not only
crucial from the standpoint of receptor dosages of certain of these oxides,
but is essential to understanding the role of surface reactions in causing
ozone buildup by removing nitric oxide (which would react rapidly with 0
to form NO- + 0 ).
*
Typical urban surface material samples could be exposed and analyzed for
NOX uptake in comparison with unexposed control samples. This could be
coordinated with laboratory studies.
13
-------
3.3 A SPATIALLY INTEGRATED SAMPLING TECHNIQUE
Three reasons present themselves immediately for testing spatially
averaged measurements of gas phase pollutants:
1. Photochemical/diffusion models have coarse space resolution;
hence, spatially averaged readings afford a more reliable
comparison than do highly localized readings.
2. Correlation with health effects demands some degree .of
averaging over space as well as time to build confidence
in the results.
3. Simultaneous collection and mixing from a spatially separated
array of sampling inlets can avoid the reaction rate errors
due to chemical times being shorter than turbulent mixing
times.
The advantages presented by the first two reasons are self-evident, but
the merit of the third is somewhat subtle.
Addressing the question of mixing interference with the NO + 0_
reaction, we can follow two sampling rules to avoid time averaging errors
in computing reaction rates (or conversely in comparing with model-derived
average concentrations). The first rule is to take simultaneous mixed
samples from points separated by more than the correlation length (integral
scale) of the local atmospheric turbulence. The second rule is that the
multiple samples must be rapidly mixed and allowed to equilibrate at the
ambient temperature and sunlight intensity conditions.
A device that should satisfy the two rules is the OASIS (Octopus
Air Sampler I_n S^itu) . The OASIS has eight horizontal sampling tubes, each
about one meter long, extending radially from a central mixing chamber.
The mixing chamber must be transparent to the ultraviolet N0? dissocia-
tion bands and will contain a nozzle cluster fed by the sampling tubes to
assure fast mixing. The volume of the OASIS mixing chamber must be
14
-------
consistent with the steady sampling flow in such a way that photochemical
equilibration will be nearly achieved in the typical detention time for a
gas element entering the chamber. Samples withdrawn from the mixing
chamber can be analyzed with available chemiluminescence detection
equipment.
The purpose of the OASIS design is to mix and react uncorrelated
samples of a polluted atmosphere. The analysis of the reacted gas mixture
will then correspond more nearly to the concentrations predicted by models.
If these steps are not taken, the product of time-averaged NO and 0
concentrations can correspond to much higher reaction rates than those
that actually occur (see Appendix).
15
-------
4.1 ADVANCE PLANNING
The efforts should begin with meetings with the test working group
whose responsibilities have been delineated. These meetings will permit
the group to plan in detail the test procedures, data collection procedures,
data formats, and to fill in any other details and gaps that may exist in
the preliminary program design. Each of the agencies that collect data
should be visited by the data manager so that the data collection plans
can be reviewed and understood. Tape formats must be planned down to the
bit level so that programs can be written to extract the data.
Detailed arrangements must be made to provide the program director
with quick-look data so that he can determine how .well the tests are going
and whether to institute changes in the procedures. Even though test pro-
cedure design may be conceptually complete, many finishing touches are
usually needed to coordinate data acquisition with data management require-
ments. The data manager must work actively during the planning phase to
establish standardized procedures for data acquisition and communication.
Moreover, he may provide further assistance in the test program planning
outside of-the area of data management.
It is extremely important that this task be allowed sufficient lead
time, since a great deal of development is required before implementation
of computer programs can begin. Many of the programs must be written and
completely validated before the beginning of data acquisition.
4.2 REPORTS TO THE PROGRAM DIRECTOR
During the conduct of a field pollutant measurement program, the
director must be able to know that optimum procedures have been followed
with regard to flight plans, sampling frequencies, and scheduling times.
He must also be able to assess the adequacy of sampling data including
measurement sensitivity, equipment reliability, and calibration. In
16
-------
ascertaining that each operation has met its specific test objectives, he
may find it necessary that certain changes be brought about in the test
procedures.
Essential to the achievement of these objectives is a systematic
compilation of test results in the form of "quick-look" readable displays
that can be provided to the program director very shortly after the collec-
tion of the data (preferably within 24-36 hours after each operation).
One effective means for displaying this information is an on-line terminal
located in the program director's office.
An example of such a terminal is the Control Data 711 Cathode Ray
Tube (CRT) Display Terminal. This terminal, shown in Fig. 1, is one of a
family of low-cost, stand-alone remote terminals designed for communica-
tions with a computer in a conversational mode over telephone lines. This
common-carrier compatibility of the 711 Terminal and its freedom from
Figure 1. CRT Display Terminal
17
-------
space-consuming external control units makes interactive computer-
communication easy to install wherever there is a telephone outlet.
The keyboard is electronic and has a standard typewriter layout with
a 10-key adding-machine-type cluster plus user's choice of additional eight
special-function keys for device and communications control.
The CRT is a modified television display module, accepting EIA
Standard composite video images from the device's self-contained logic.
Characters are dot-formed and generated by TV raster scan. Input is via
control interface. Viewing area is 8 inches high by 10 inches wide and
displays 16-lines of 80 characters.
4.2.1 Quick-Look Data
If the central element in the program design is a body of helicopter
measurements, these data and radar tetroon tracking data should be avail-
able to the program director on the shortest possible lead time following
the collection of the information. Specifically, flight plan logs, the
profiles of tracer concentration, tetroon location with respect to tracer
material, carbon monoxide, ozone, nitric oxide, nitrogen dioxide, tempera-
ture, and ultra-violet intensity will be provided using data which have
undergone preliminary reduction. This means that nominal calibration fac-
tors should be used for quick conversion of the raw data into approximate
and unedited displays. At the same time, the data will be undergoing full
correction and editing. As those results become available, they should be
placed in the project director's data file to replace the approximate,
unedited values.
Examples of the kinds of displays that could be made available to
the program director are shown in the photographs of Figs. 2-7 and are
2
drawn from ideas for the Los Angeles Reactive Pollutant Program. These
examples, using fictitious numbers, were created by typing directly onto
the display screen.
18
-------
The following six pages (pre-printed, and thus unnumbered) contain
Figs. 2-7.
Figure 2 Carbon Monoxide Tables
Figure 3 Carbon Monoxide Profile
Figure 4 Carbon Monoxide Pattern
Figure 5 Flourescent Particle Tracer Gradient Pattern
Figure 6 Gas Chromatogram, Peaks 1-13
Figure 7 Gas Chromatogram, Peaks 23-32
19
-------
LOS RNGELES REflCTWE POLLUTE PMOW
OTBON HOHKIOE TRBLES - 9/14/73, K3&-1HB hours'
Prof i lei
0630 hrs.
480 2.0
290 6.0
8.1
Location
3391/1213
Fwy Stock
Profile2
9930 hrs.
Height Cone
ft Pf»
420 3.1
358 2.8
288 2.9
218 3.9
Location
3546/2011 •
SB F«y at Fwy 605
Profile^
1838 hrs/
Location'
4143/3W
Posad*"
Figure 2. Carbon Monoxide Tables
-------
LOS fWGELES REflCTWE POLUJTPHT matt
CH8QN MONOXIDE PMFILE-
Dote 9/M/7J-
Tiie 9938 trs."
Location 3546^811 (SB Ftiy at Fwy 685)'
Heists 218, 288, 358, 428 ft."
.3 4 5
Concentration - ff»
Figure 3. Carbon Monoxide Profile
-------
CO
to
LOS flNGELES SEPCTWE POillM PTOGRW
mm MONOXIDE PflTTtRN - 9/14/7T, 8938 tours, Flight poft 5'
Locution 3546/2811 (SB fy at F^6®
flveroge hei^it = 218 ft.'
4.3 Verticol = 348 (iea. heudina-
4.1 4.3
4.1 4.2
4.0 4.4
4.2 3.8
4.8 4.0
3.8 4-2
4:24,14.13.94,04.03.93.93.
Carbon Monoxide - ft*
Figure 4. Carbon Monoxide Pattern
-------
LOS BNGEIES REFCTWE POLLUTflKT. mm
FPT GRRDIEHT'PfiTTERN - 9/14/73, 9938 hours, Flight potti 5'
Location 3546/2811 (SB F*y at F*y MET
flveroge height = 218 ft."
-0.5 Vertical = 348 dq. heading-
-1.9-8.3
-1.2 -9.1
-2.5 8
-9.3 9
+9.2 9
-9.1 8
.
+9.4-0.3 9 -9.1 9 8 8 0 9
FPT Gradients
Figure 5. FPT Gradient Pattern
-------
10
COMPOUND
Methane .
Ethane
Ethylene
flcetylene
Propone
Propylene
Isobutone
Butane
Butenes
Isoperitane
n-Pentane
GELE5 R0CTWE fftiLU
m CHfflWDB
-
I -1-Butene *»*
Date 9/14/73
TiK 891 hrs.
Location 3546^81
Figure 6. Gas Chromatogram, Peaks 1-13
-------
LOS MELES REflCTWE FOLLUTiW mm
COMPOUND
3-Fteltiylpentcre
Utexene
n-Hexene
trons-3-Hexene
trons-2-Hexene
cis-3-Hexene
cis-1-Hexene
Methyl eye Iopentone
2,4-DimethylpenTcme
Benzene
%---mdrocarbons
Dcrte 9/H/73
lite 9938 nrs
Location 354M811
Hei^it 218 ft.
CsJ
On
Figure 7. Gas Chromatogram, Peaks 23-32
-------
Figure 2 is a photograph of an actual Control Data 711 CRT display
of three carbon monoxide profiles in tabular forms (from helicopter mea-
surements) . Headings show identification numbers and times. Below the
height-above-ground concentration entries are locational data in the form
of coordinates and key words from the helicopter log.
Figure 3 shows a plot option as applied to the second profile tabu-
lated in Fig. 2. Height station is the vertical coordinate and concentra-
tion is the horizontal coordinate. The points, each denoted by an "x",
are averages of the flight path around the tetroons at each height.
Figure 4 shows more detail for one of the flight paths averaged on
the previous figure. The triangular pattern indicates the shape of the
flight path (denoted by the identification number "5" in the heading) and
the numbers are locally averaged co-readings. If CO levels are recorded
every 5 seconds and a leg of the flight path takes 2 minutes, a smoothing
and interpolation operation is used to get ten concentration values per
leg, as shown. The vertical reference is 340 degrees on the compass for.
this particular pattern.
The display in Fig. 5 is very similar to that in Fig. 4 except gra-
dients of fluorescent particle tracer (FPT) are shown. The numbers show
that the helicopter flight pattern (which is presumably around the cen-
troid of the tetroons) is somewhat to the east of the apparently highest
concentration of the FPT cloud. This is clearly an important diagnostic
for the project director in planning future flight missions.
Figures 6 and 7 show tabulations of gas chromatographic data.
Because of laboratory requirements, this type of information will not be
available as rapidly as the helicopter data. Hand-recorded logs add to
the communication lag time.
The displays in Figs. 2-7 are given as examples and .are not repre-
sented as being complete or final. Changes are inevitable in the planning''
26
-------
phase of the work when the data manager interacts with the project director
to come up with the optimum forms of data presentation for the "quick-look"
feedback function.
4.2.2 Data System Procedures
The system proposed could function as follows: The quick-look raw
data will be picked up by the on-site data manager and transmitted to a
data center. There are several alternative means of doing this. One
simple approach is to take it by courier to the processing center.
Some of the quick-look data that are not in machine-readable form.
such as the plotted helicopter positions, certain calibration information
not on the tapes, and some log information, can be entered into the com-
puter through the on-line terminal by typing the data on the keyboard.
A goal of the data manager (providing the helicopter tapes are made
available to him immediately after the tests) should be to have all of the
quick-look data entered into the computer by midnight of the test day.
Programs can then be run that perform a rough data reduction (how rough
is to be determined—perhaps considerable machine editing can be performed
at this time), and a merging of data from different sources including data
entered via the terminal. The data can be stored in an on-line random-
access disk file for use by the program manager on the morning following
the test.
The program manager will access the file and obtain displays created
from the file by typing certain coded commands on the keyboard of the 711
terminal (which can be on or next to his desk) and sending the commands
(simply by pressing a key) to the computer. The computer will then access
the required data in the file (if available), format the requested display,
and transmit the display to the terminal over the phone line. The on-site
data manager can help the program manager as needed.
27
-------
Details in operations for obtaining the data for both the quick-
look file and the archival file must be worked out when more definite
plans are finalized by the agencies acquiring the data. In general, the
issuance of signed receipts should be required with endorsements any time
data change hands. Copies of these receipts will be retained by the data
manager, given to the data acquisition agency, and filed in the program
office. This will minimize misunderstandings regarding where certain
data are in the svstem.
Another data management function is the provision of project status
reports to the program director. These should be statistical in nature,
outlining "the number of hours flown for the week and for the program.
They will also provide information on the time lag between data acquisi-
tion and data reduction for all missions. This will highlight any serious
scheduling problems to the program director, and will provide him with an
up-to-date view of how much of -the resources in time and equipment have
been expended.
At least four months should be scheduled prior to initiation of the
field program to allow for program preparation and establishment of data
links for this rapid reporting system that will be provided to the program
director. An on-site representative of the data manager should be provided
at all times during the operation of the field program. It will be his
responsibility to obtain the data from the various agencies and to send
it to the data center. As the file of reports is refined and updated, it
will be placed in permanent storage so that it will form a part of the
modelers' archives.
4.2.3 Computer Program Development
As a part of this task, the data manager must specify, design, code
and validate the computer programs that read-in the quick-look data, merge
data from various sources, create the quick-look on-line file, and, in
response to commands from the terminal, format displays. A telecommunica-
tions package will be needed to allow interactive communication with the
28
-------
terminal. The same displays that will be sent to the terminal screen can
also be printed on either the printers in the data center or one of the
printers attached to various remote terminals in the test area as backup.
4.3 ARCHIVES FOR MODELERS
.Investigators who develop simulation models for prediction of air
quality generally use two types of data:
1. Input data that specify the needed meteorological parameters.
2. Test data which consist of air quality measurements that the
model output must be checked against.
LARPP .will be collecting both types of data for the purpose of broadening
the base for simulation model development. In performing this task to
establish archives for the modelers, machine readable records should be
utilized to minimize errors throughout. In this way numerous key punch
errors in transcription from printed books will be avoided. Standard mag-
netic tape files should be used and detailed documentation in catalogs
must be provided. As part of this documentation, the measurement methods,
the data reliability, the limitations in its use, and calibration informa-
tion that might affect the application of these results to modeling tests
all must be described. The catalog part of the documentation will first
provide the modeler with some indication of the types of days and the types
of flight programs conducted. Summary graphs and tabulations will provide
a guide to the selection of the appropriate tape files. Codes and formats
for the tapes will be given in full detail in the documentation so that,
using the catalog, the modeler can order the magnetic tapes that are best
suited to his purposes.
A number of measurements not represented in the quick-look reports
to the program director should be included in the archives files. The
aircraft wind field and meteorological data from the fixed-wing aircraft
equipped with an INS or doppler system will be included in this file.
This will round out the data sample and provide additional information
29
-------
for those modelers who do not utilize a moving coordinate system concept.
Also, greater details in the hydrocarbon analysis will be entered in the
final version of the data archives. Finally, other data sources may be
utilized such as the five CHESS stations and the Los Angeles Air Pollu-
tion Control District stations. Lidar information will give further
checks on the mixing layer depth and structure if it becomes available.
Similar information might be provided by acoustic sounder data. One of
the data manager's responsibilities is the editing of these peripheral
bits of information as to their relevance and utility to the modeling
effort.
The data manager will act as custodian for these archives during the
life of the proposed program. At the conclusion of -the program the com-
plete documentation and data files that have resulted from the program
should be delivered and made available to users.
4.3.1 Data Sources and Volume
In order to size the data management problem we have .estimated the
total number of data points to be recorded on the archival files. We do
not presently have enough information to know how accurate our estimate
is, but the estimate serves as a reasonable point of departure.
The estimates given in Table 4.1 show approximately 13 million data
points. Assuming an average of 10 BCD characters per data point (a con-
t
servatively high estimate), the archival file will contain approximately
130 million characters.
A single 2400-foot reel Of magnetic tape recorded at a normal 556
bpi with record gaps can hold about 12 million characters. Therefore the
entire file will fit on about 11 reels of magnetic tape. This is the
most convenient method of storing archival files. However, if desired,
each day can be put on a separate small reel for certain applications.
30
-------
TABLE 4.1
ESTIMATE OF TOTAL DATA POINTS PER DAY IN LARPP
Data Source
Helicopters
Radar
Fixed Wing
Van
Bag Samples
Other Sources
Totals per day
Totals per 30
days
Rate
1800/hr*
120/hr
8400/day
200/hr
30/day
Duration
6.4 hr
6 hr
6 hr
Quantity
2 helicopters
3 targets
1 plane
1 van
Records
23,000
2,160
-8,400
1,200
30
1,000
35,790
1,000,000
Variables per Record
~15
5
6 (t, 6, z, V, T, H-O)
14 (t, x, y, FPT, UV.
03, NO, N02, GO, HC, S02,
nephelometer, T, H20)
37 (t, x, y, z, 33 peaks)
10 .
Data Points
-345,000
10,800
-50,000
-16,000
1,140
10,000
433,000
13,000,000
Data every 2 seconds for 20-minute profile, up to three profiles per hour.
Assumes 700 miles per sortie, one sortie per day, 12 measurements per mile.
Includes one extra point per record for record type and index information.
l/i
-------
4.3.2 Archival File Format
The archival file should be on magnetic tape. The retrieval pro-
gram, described later, could generate records for users on seven-track
tape with a variety of selection and formatting options.
One possible format of the file is as follows: The file will be a
serial file ordered by time. Each day's file will consist of (1) a header
record giving the date and other pertinent information, (2) a time-ordered
series of data records from all sources, followed by (3) a series of log
entry records.
The data records from' each source will have this format:
Time
Record
Type
Indexes to
Log Entries
Variable 1
Variable 2
•
Variable
n
The Record Type indicates the source of the data and the data record type.
For example, a record from the radar could be designated as Record Type
No. 1, and could have as Variable 1 the object the data is on (e.g., tet-
roon 1), Variables 2, 3, and 4 could be the position. A different record
type would contain data from the helicopters. For fast processing, each
record type has a fixed number of variables in a fixed format. (If it
turns out to be required, in some cases a variable-length format can be
used.) Variables for which data is missing are filled in with a "no data"
character.
The Indexes to Log Entries point to appropriate records containing
the third type of information stored on each day's file, the log entries.
These records contain text from the various logs that were produced during
the day. For convenience in data processing, it is recommended that all
log entries have a prescribed format consisting of (1) time of entry, (2)
text of any kind, (3) a code that indicates which type of data records
are affected (e.g., an "HI" to indicate that the records from helicopter
32
-------
1 are affected), (4) a list of the variables that might be affected, and
(5) the beginning and end times during which the log entry affects data.
The reason for the above information is so the data entries can be
automatically keyed to the log entries that might affect the data. Dur-
ing retrieval of data for a particular modeler, the appropriate footnotes
(in the form of log-entry text) can then be automatically retrieved for
just the data that the modeler requires.
The proposed variable-length format for the log entries is:
Log Entry
Number
Time
Ll
Text
Tl
T2
R
Vl
V/
/Nv
V
n
$
where Log Entry Number is a unique number assigned by the computer
to each log entry for the day
Time is the time of the log entry
LI is the variable length of the text
TI is the first time this log entry applies
T is the last time this lag entry applies
R is the record, type number to which this log
entry applies
V , ..., V are :the variables to which this record applies
R is a delimiter
Additional sets of record types and variables can follow the delimiter
(represented here by $), each set separated by a delimiter.
The Log Entry Number is the same number that appears in the data
records called "Indexes to Log Entries." These indexes are generated
automatically by the computer from the trailing information in the log
entries.
33
-------
The reasoning behind the proposed organization of the archival file
will become more evident under the later discussion on data retrieval.
4.3.3 Supplementary Data Files
Two additional archival data files should be constructed along with
the principal file described in the previous subsection. Both of these
files will be in essentially the same format and can be created at rela-
tively small additional expense. One of the files will contain all raw
data from the field measurement program; the other will be a much abbre-
viated file containing the smoothed (and possibly interpolated) data that
many modelers will require.
Following most field measurements programs, the raw data are set
aside and frequently lost unless some conscious effort is made to preserve
and catalog these files. There is no better time for this activity than
during and following the program itself. Usually the experimenter is so
busy with the operational and logistical aspects of the program that this
detail is overlooked. It is for these reasons that the data manager should
create and maintain a raw data bank. This raw data bank will provide a
source of basic information to those who would investigate special aspects
of the field program at some future date. For example, it may be dis-
covered that a certain type of interference occurred in one piece of
instrumentation. With the raw data and all of the independent measure-
ments, it may be possible to correct and refine, even further the informa-
tion that was obtained during the program. Also, in the course of evaluat-
ing the data for modeling purposes, it is sometimes useful to be able to
go back to the original raw data. The accomplishment of this task would
begin by preparing and designing a record structure that would contain
the various types of data that are collected and recorded. The informa-
tion included may, for practical purposes, be limited to the information
recorded on magnetic tape or on punched paper tape.
34
-------
It might be anticipated that since the retrieval system can smooth
certain data, different modelers will need the same smoothed data.
Rather than perform the same smoothing operations over and over, or re-
quiring each modeler to smooth his own data, there should be a much
shortened file of selected smoothed data.
4.3.4 Retrieval of Archival Data
An archival data bank is useful in direct proportion to the flexi-
bility and quality of the retrieval programs. The above file and record
structure is designed with convenient and versatile retrieval as the
uppermost criteria.
Retrieval programs are needed to permit the modeler to access only
those parts of the data required in the format desired. After consulting
the archival-file documentation, the modeler will specify (1) the day(s)
of interest, (2) the start and end times during the day(s) of interest,
(3) the record types of interest, and (4) the variables of interest. Then
a tape will be created using one of the options described below that
contains exactly the data requested. Because of the data log indexing
method described above, the log entries that pertain to the data requested,
and only the data requested, will also be included on the modeler's tape.
Several options should be made available to the modeler so that the
data are in a form most useful to him. These options include (1) within
the requested time span get all of the actual data in the selected records
every time it appears in the data bank, (2) get actual data at the recorded
time nearest to a desired time (e.g., request data every 10 minutes, get
actual data at a closest time available to 10-minute rate), (3) get inter-
polated data at exact time increment specified, (4) get interpolated data
that has also been smoothed.
4.3.5 Documentation
The documentation of the archival tapes should describe each test
day in detail, preceded by an appropriate summary. This summary describes
35
-------
the experimental design objectives and details the types of data obtained.
A discussion of experimental data accuracy aids the modeler in using the
archives to the best advantage. A data catalog shows tetroon trajector-
ies, helicopter positions, and coarse concentration data. It includes
log comments on the day's operations and certain meteorological data.
The data available on the archival tapes are described in detail, includ-
ing types of files on the tapes.
A separate section (or volume) describes the formats of the archi-
val files in detail, and emphasizes a treatment of the available retrieval
capabilities and the simple steps that modelers need to follow to extract
the required data.
4.3.6 Data Security
Whenever a data bank is being created and stored, procedures must
be instituted to ensure that the data are not irretrievably lost by any of
the many ways data banks are destroyed:, fire, riot, sabotage, stray mag-
netic fields, operator error, etc. The data manager should make two copies
of each source data and store the extra copy in a secure place. All lost
tapes can then be recreated from the copy. If the data manager is per-
mitted to keep and store the original tapes, then these should be stored
in the secured area rather than copies. In no cases should original data
tapes be used for other than making copies.
4.3.7 Computer Program Development
The data manager will specify (in cooperation with appropriate
agencies), design, code and validate all of the software required to per-
form this task, including reading and formatting software for the various
sources, creation of the data files, and software to perform the:retrie-
val functions.
36
-------
5 BRIEF OVERVIEWS OF FUTURE FIELD PROGRAM OBJECTIVES
5.1 DEFINITION OF THE ROLE OF AEROSOLS IN ATMOSPHERIC CHEMISTRY
Those modelers who have been engaged in photochemical validation
tests over the p.ast several years have noted deficits in the gas-phase
oxides of nitrogen. A worthwhile goal of any field program is a systema-
tic search for the pathways whereby NO leaves the gas-phase system.
X
Certainly this has received attention in the smog chamber work, and we
have now reached the point where field programs can no longer ignore this
problem. Two gas-solid reaction possibilities immediately come to mind
regarding the fate of NO . One involves the aerosol and the other sur-
X
faces on or near the ground. Rough calculations of nitrate in aerosol
suggest that these effects cannot account for much of the loss. Urban
surfaces, however, may take up much of the oxides of nitrogen before they
mix upward.
Aerosol formation causing reduced visibility-is probably the most
obvious manifestation of photochemical smog. The physicochemical pheno-
mena governing these reactions should be investigated in sufficient detail
at least to give a bulk reaction rate to put in air quality simulation
models. A broad recommendation in this area was recently drafted and
adopted by the Photochemistry and Transformation Modeling Panel at the
Third AEC/EPA Chemist Meteorologist Workshop (Ft. Lauderdale, Florida,
January 15-19, 1973). The recommendation is:
It is recommended that models be tested by field
measurements. For aerosols these measurements
should include (a) light scattering (b) concen-
tration in terms of particle number density (c)
concentration in terms of mass per unit volume
(d) particle size distribution, and (e) composi-
tion (significant for heterogeneous catalysis,
health effects, or as needed to test models).
For gases the measurements should include (a)
identification of chemical compounds and (b)
concentrations. The meteorological measurements
will include (a) temperature (b) turbulence (c)
relative humidity and (d) wind vector. Vertical
37
-------
distributions should be obtained. The measure-
ments will be made with respect to both a ground-
fixed and air-fixed coordinate system. It is
important that the measurements attempted in any
given field program be limited to the needs of
the program and be sufficient to test the model
involved.
Another particulate interaction of potential significance is attenua-
tion of ultraviolet radiation. The attenuation goes up with decreasing
wavelength so that even though total solar input is only slightly reduced,
shorter wavelengths in the NO -dissociation band may be cut down. Cer-
tainly some data exist on this from previous Los Angeles basin programs.
They must be analyzed prior to designing new field experiments.
5.2 FURTHER UNDERSTANDING OF GAS PHASE PROCESSES AFFECTING REACTION
RATES
Nonuniformities in concentrations arise because of incomplete mix-
ing in the atmosphere. Non-zero correlations between fluctuations of
reactant species-pairs lead to errors in chemical calculations if the
reaction-time is smaller than (or equal to) the mixing time. The large
effects of this interference were exhibited by the 1969 Los Angeles data
4
and were documented in our 1970 data analysis report. It must be an
objective of any second-generation field program (such as LARPP or RAPS)
either to investigate this effect in detail or to circumvent it by devis-
ing physical arrangements that average properly over space or time (like
the OASIS sampling technique described in Sec. 3). The 1973 AEC Chemist-
Meteorologist Workshop drafted and adopted the following recommendation
that was suggested by presentations by G. Hilst and by A. Eschenroeder:
Turbulent mixing corrections to reaction rate terms
should be evaluated. Spatial nonuniformity in the
gas phase concentration may lower the reaction rates
of species that are being mixed together; this beha-
vior is expected at plume boundaries. In models this
must be reflected as a "mixedness" correction.
38
-------
As developments in basic theory become available,
it will be necessary to approximate certain terms.
For example, in a second-order closure scheme for
two reactive species, there can be nine partial
differential equations to be solved. .For a ten-
species mechanism, more representative of atmos-
pheric chemistry, there can be 120 such simultan-
eous differential equations. A laboratory program
should be conducted in parallel with the model
improvements to test the validity of turbulent
reaction theory.
Efforts should be directed toward the development
of new transport and diffusion formulations that
are suited for reacting gas flow.
The possibility of using wind tunnels to physically
model diffusion and the behavior of chemically reac-
tive systems should be thoroughly investigated.
39
-------
40
-------
APPENDIX
THE DEPENDENCE OF APPARENT REACTION RATES ON TURBULENT MIXING
PROCESSES
A.I INTROCUTION
Heterogeneities appear in the atmosphere as sources emit gases that
take some time to mix on a microscopic scale. Large blobs of newly emitted
gas break up into smaller blobs by the turbulent cascade mechanism. When
small enough scales are attained, molecular scale diffusion smears out the
nonuniformity. If the emitted pollutant is a potential atmospheric reactant,
this final microscopic mixing must take place before the reaction can be-
gin. If the reaction is fast compared with mixing, then the rate of
reaction is controlled by the turbulent cascade process. Consequently,
if concentrations of reactants are averaged over many turublent blobs
(as they are in most models), the reaction rates calculated from these
average values can be greatly overestimated. The overestimation occurs
because the calculation assumes that the reactants are homogeneously mixed.
In this appendix we assess the significance of the inhibition of reactions
by turbulent mixing process in atmospheric photochemical calculations.
Our specific approach is the calculation of the parameter called
r which is a ratio of reeaction time to mixing time. (This is simply
the reciprocal of the first Damkohler number for the convective and
diffusive structure of the turbulence.) If r is smaller.than unity,
mixing can be the rate controlling process and inhomogeneities can cause
errors in the reaction rate calculation. If the reverse is true, then
the reaction is not hindered seriously by mixing (the greater r is than
unity, the less is the hindrance).
41
-------
A.1.1 Atmospheric Kinetics and Concentrations Typical of Photochemical Smog
Recent extensions of our earlier work on photochemical diffusion
simulation suggest that the lumped-parameter mechanism in Table A.I gives
a high degree of consistency for atmospheric modeling. The main improve-
ment is the addition of hydroxyl radical reactions which are now thought
to dominate the organic chains.
Q
Although Donaldson and Hilst assign concentrations of 1 ppm to each
species, it is preferable to use observed and calculated values for a
polluted atmosphere. The levels in Table A.2 will be used to evaluate
r in this appendix.
A. 1.2 Calculation of Turbulent Mixing Time
Mixing time consists of two components: one for the convective
breakup of large inhomogeneities into fine structure; and another for
diffusive smearing of fine structure into homogeneity. The first process
needs to occur before the second one can proceed. We will show here
which one is rate-controlling and how long the total mixing process takes.
The convective breakup time is easily estimated by using turbulence
cascade time T given by
T = k-3/2E-1/2 (A.I)
where k is turbulence wave number and E is the turbulence energy
spectrum function. The highest intensity of large scale inhomogeneities
is at the spectral peak. Using the peak properties we can relate E and
k to bulk variables of the turbulence. A turbulence spectrum solution
gives the following values at the peak:
k = 2/A and E = u'2A/5 (A. 2)
42
-------
TABLE A.I
PHOTOCHEMICAL KINETIC MECHANISM FOR
ATMOSPHERIC POLLUTION
Rate constants are in ppm -sec except where noted
Rate Constants Used in Model
Reaction
1. hv + N02 •* NO + 0
la. 0+02+M+03+M
2. NO + 0., •* N02 + 0
3. 0 + HC •* 8RO,
4. OH + HC -c 8R02
5. 03 + HC •* R02
6. R02 + NO ->• N02 + (1/8)OH
7. R02 + N02 •* PAN
8. -OH + NO * HONO
9. OH + N02 -»• HN03
10. hv + HONO •* OH + NO
H20
11. NO + N02 -> 2HONO
12. N02 + 03 + N03 + 02
13. N03 + N02 + N205
14 . NO - NO + NO
Early Time Late Time
8. 3 (-4)*** 5.0 (-3) sec'1
2.2(-7)ppm sec
4.4(-l)
4.6(0)
1.7 (+2)
6.7(-5)
1.7 (+3)
3.3(0)
2.5(+l)
5.0(+1)
_i
3.K-6) 1.9(-5)sec L
1.7(-5)
8.3(-5)
7.5(+l)
2.3(-l)
Sources
Ref. 10
Ref. 11
Ref. 12
**
**
**
**
**
Ref. 13
Ref. 14
**
**
Ref. 10
**
Ref. 10
H20
15.
1.0(0) sec
-1
16. N02 + particulates -»• products
-
Early time refers to 0630-0700 LST and late time to V1300 LST. Photochemical rate constants
are continuously varied with the solar zenith angle; therefore, values given for photolysis
rates are typical for the times shi>wn only.
**
Where values are either uncertain or unreported, estimates are made based on smog chamber
results.
Parentheses after a number are defined as follows (a) = * 10 , i.e., 8.3(-4) = 8.3 * 10
-4
43
-------
TABLE A.2
MEASURED AND MODELED POLLUTANT CONCENTRATIONS IN THE LOS ANGELES
BASIN
Species
NO
N02
HC
RO,
OH
HN02
N03
N2°5
°3
Typical Concentrat
Early Time (0630-0700 LST)
K-l) to 5(-l)
2(-l) to 4(-l)
5(-D
K-8)
K-8)
K-2)
K-8)
K-8)
2 (-3)
:ions in ppm
Late Time (1300 LST)
K-2)
K-l)
K-l) to 2(-l)
K-6)
K-7)
K-2)
K-7)
K-6)
2(-l) to 4(-l)
where A is the longitudinal integral scale and u
fluctuation velocity. Therefore
,2
is the mean square
T * 4A/5u'
c
(A.3)
In the atmospheric boundary layer, the RMS fluctuation velocity is 20%
or 30% greater than the friction velocity, and the scale for energy-con-
taining eddies is approximately 20% of the height above ground:
u' - 5u*/4
(A.4)
For the lower atmosphere, an RMS fluctuation of 40 cm-s is typical so
that
T = 4 s at a height of 10 meters.
44
-------
The time T, for diffusive smearing of fine structure may be
d
estimated as one random-walk time scale over a dissipation length
T, = X/2D (A. 5)
d d
where A, is a dissipation length scale and D is the molecular diffusion
17 18
coefficient. Priestley points out that A, is about fifteen length
scales so that
- 15v3/V1/4 (A.6)
where v is kinematic viscosity and e is the turbulence dissipation
3 -3/4
rate. Since t; « u' and A, <* u' , we can derive a velocity
scaling law to get Priestley's empirical curve for a 5 ms wind speed
down to 1 ms which is more typical for a day with high air pollution.
2 -1
Because air Schmidt numbers are near unity, D = v - 15 cm «s . At
10 meters above the ground this gives us T, - 63.5 s.
A. 13 Calculation of. Chemical Time
In calculating r for a multicomponent mixture, each species can
be assigned a chemical time that can be defined by the concentration
Q
divided by absolute value of the reaction rate. Donaldson and Hilst
applied their binary mixture model to reactions occurring in a multicom
ponent system. They thereby had an r for each reaction. The species
actually mix and react (rather than the reactions); therefore, it seems
physically meaningful to define an r for each species.
The use of species chemical times requires that we compute only
the binary or tertiary reaction terms since those are the ones that can
be affected by inhomogeneities. For stationary state species, both forward
and reverse rates are large but their difference is small. In these cases,
we used averages between absolute values of forward and reverse rates.
It should be noted that the expected stationary state balances are not
precisely satisfied by using the concentrations in Table A.2 in the
45
-------
mechanism in Table A.I. This is to be expected because the concentrations
were assigned only rough order-of-magnitude values in the cases of highly
active species.
A.1.4 Discussion of Values of r
Table A.3 gives values of the r-parameter. Only hydrocarbon, nitro-
gen dioxide, and nitrous acid vapor are relatively unaffected. It is
significant that Reaction 2 can affect either nitric oxide or ozone with
respect to hindrance by insufficient mixing. This is in contrast to the
findings of Donaldson and Hilst who employed a rate constant a few orders
of magnitude smaller.
This finding has considerable consequences for predicting pollutants
in photochemical smog because Reactions 1 and 2 are the dominant inorganic
TABLE A. 3
RATIOS OF CHEMICAL TIMES TO MIXING TIMES
AT 10 meters HEIGHT IN A 1 m'S~l WINDFIELD
Early Time (0630-0700 LST) Late Time (1300 LST)
NO
*
0_
3
0
HC
RO
2
OH
HONO
V5
1.6(+1)
5.0(+2)
1.2(-1)
3.K-7)
8. 3 (+3)
4.4(-5)
8.K-3)
1.9 (+3)
l.K-3)
l.Z(-l)
2. 6 (+2)
3.3(0)
3.2(-7)
4.0(+2)
2.2(-4)
2.6(-2)
1.3 (+3)
8.9(-4)
*
Only binary reactions were considered in chemical times.
46
-------
cycle. Indeed, we have =cited atmospheric observations made several years
ago in the Los Angeles basin which tend to bear this out. Figure A.I shows
*
the logarithm of the rate ratio of Reaction 1 to Reaction 2. It is
plotted versus ozone level. Photochemical quasiequilibrium theory would
say that the ratio should be near unity (i.e., the logarithm should be
near zero); however, we have surmised that inhomogeneous effects inhibit
Reaction 2 enough to drive the points distinctly negative at high ozone
levels. This could occur because of fresh NO being introduced into an
atmosphere high in ozone, but no reaction takes place until they are
intimately mixed. It is described in our paper cited above how Mast
meter inaccuracies, NO inaccuracies at low NO-levels, reaction interfer-
ences, and sampling tube dark reactions can all be eliminated as causes
of the large departures.' Although positive evidence is not yet available,
the small values of r(i.e., 1.2 x 10~ ) for either NO or 0, suggest
that turbulent inhomogeneities are responsible for the apparently anomalous
behavior.
Reaction la is so fast that an ozone molecule forms almost immediately
following a photolysis of a nitrogen dioxide molecule.
47
-------
1.00
0.75
0.50
0.25
0.00
o
O -0.25
-0.50
-0.75
•1.00
-1.25
-l-
I
POINTS WITH [NO] > 1 pphm
+
-H-
LOCUS OF POINTS SATISFYING
QUASI-EQUILBRIUM HYPOTHESIS
4 + +
+
I I
I I
12 16
OZONE, pphm
20 24 28
Figure 8. Possible Turbulence Effects on the Ozone Quasiequilibrium
for 1969 Ground Data at El Monte—High NO Levels
48
-------
REFERENCES
1. Regional Air Pollution Study for St. Louis, Missouri, Environmental
Protection Agency RAPS Series No. I Study Plan (Fifth Draft, October
27, 1973).
2. W. A. Perkins, The Los Angeles Reactive Pollutant Program, Metronics
Associates Inc. Report, September 1972.
3. A. Q. Eschenroeder and J. R. Martinez, A Modeling Study to Charac-
terize Photochemical Atmospheric Reactions to the Los Angeles Basin
Area, General Research Corporation CR-1-152, November 1969.
4. A. Q. Eschenroeder and J. R. Martinez, Analysis of Los Angeles
Atmospheric Reaction Data from 1968 and 1969, General Research
Corporation CR-1-170, July 1970.
5. W. B. Johnson, F. L. Ludwig, W. F. Dabberdt, R. J. Allen, "The
Urban Diffusion Simulation Model for Carbon Monoxide," Proceedings
1972 Summer Computer Simulation Conference, San Diego, California,
June 1972, pp. 1062-1076.
6. A. Q. Eschenroeder, J. R. Martinez, and R. A. Nordsieck, "A View of
Future Problems in Air Pollution Modeling" Proceedings of Second
Summer Simulation Conference, Simulation Councils, Inc., June 1972,
pp. 1013-1031 (also General Research Corporation TM-1631, March 1972).
7. A. Eschenroeder, J. Martinez, R. Nordsieck, "Concepts and Applications
of Photochemical Smog Modeling," "Advances in Chemistry," Series
No. 113 entitled Photochemical Smog and Ozone Reactions, (American
Chemical Society, Washington, December 1972, pp. 101-168.
8. C. Donaldson and G. Hilst, "The Effect of Inhomogeneous Mixing on
Atmospheric Photochemical Reactions," Environmental Science and
Technology, Vol. 6, No. 9, September 1972, pp. 812-816.
9. L. Onsager, "Statistical Hydrodynamics," Nuovo Cimento, Vol. 6,
No. 2, 1949, p. 279.
10. P. A. Leighton, Photochemistry of Air Pollution, N.Y.: Academic
Press, 1961.
11. E. A. Sutton, "Chemistry of Electrons in Pure-Air Hypersonic Wakes,"
AIAA Journal, Vol. 6, No. 10, October 1968, pp. 1873-1882.
49
-------
REFERENCES (Cont.)
12. K. Schofield, "An Evaluation of Kinetic Rate Data for Reactions of
Neutrals of Atmospheric Interest," Planetary and Space Sciences,
Vol. 15, 1967, pp. 643-670.
13. F. Stuhl, private communication, Ford Motor Co., Scientific
Research Staff, June 1, 1972.
14. J. Anderson, private communication, University of Pittsburgh,
June 7, 1972.
15. A. Q. Eschenroeder, "Solution for the Inertial Energy Spectrum of
Isotropic Turbulence," Physics of Fluids, Vol. 8, No. 4, April 1965.
16. J. L. Lumley and H. A. Pano.fsky, The Structure of Atmospheric
Turbulence, Interscience Publishers/John Wiley and Sons, New York,
1964, p. 133.
17. C. H. B. Priestley, Turbulent Transfer in the Lower Atmosphere,
University of Chicago Press, 1959, p. 58.
18. A. N. Kolmogoroff, The Local Structure of Turbulence in Incompressible
Viscous Fluid for Very Large Reynolds Numbers," C. R. Acad. Sci.
(USSR), Vol. 30, 1941, p. 301.
50
-------
BIBLIOGRAPHIC DATA
SHEET
1. Report No.
EPA-R4-73-012C
3. Recipient's Accession No.
4. Title and Subtitle
Field Program Designs for Verifying Photochemical
Diffusion Models
5- Report Date
March 1973
6.
7. Author(s)
A.Q. Eschenroeder, G.W. Deley, R.J. Wahl
8' Performing Organization Kept.
No'CR-3-273
9. Performing Organization Name and Address
General Research Corporation
P.O. Box 3587
Santa Barbara, California 93105
10. Project/Task/Work Unit No.
11. Contract/Grant No.
68-02-0336
12. Sponsoring Organization Name and Address
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina 27711
13. Type of Report & Petiod
Covered
14.
15. Supplementary Notes
Carefully designed experimental measurement programs are necessary to the task of collec
ting the extensive data base needed for testing air pollution models. In this paper a
series of recommendations is offered as input for fielvd measurement programs. These rec-
ommendations update and refine suggestions made earlier. Some specific techniques and
instrumentation are called for in certain cases, but in others only conceptual objective
are set forth. The report presents three specific conceptual plans. A description is
presented of a data management system using the Los Angeles Reactive Pollutant Program
"LARPP) as a prototype. It is concluded that future field programs must be structured
to meet specific goals. The specific goals discussed here include verifying emissions
nodels for vehicle populations, following photochemical transformations in a moving air
nass, searching for turbulent reaction inhomogeneities in the atmosphere, and charac-
terizing the interaction of gas phase species with aerosols.
16. Abstracts
17. Key Words and Document Analysis. 17a. Descriptors
Air pollution
Photochemical reactions
Diffusion
Mathematical models
Field tests
Design
Data processing
17b. Identifiers/Open-Ended Terms
Los Angeles Reactive Pollutant Program (LARPP)
17e. COSATI Field/Group
13B
18. Availability Statement
Unlimited
19.. Security Class (This
Report)(
UNCLASSIFIED
20. Security Class
Page
. XINCLASSIFIED
21. No. of Pages
58
22. Price
FORM NTIS-33 (REV. 3-72)
USCOMM-DC MOS2-P72
-------
INSTRUCTIONS FOR COMPLETING FORM NTIS-35( 10-70) (Bibliographic Data Sheet based on COSATI
Guidelines to Format Standards for Scientific and Technical Reports Prepared by or for die Federal Government,
PB-180 600).
1. Report Number. Each individually bound report shall carry a unique alphanumeric designation selected by the performing
organization or provided by the sponsoring organization. Use uppercase letters and Arabic numerals only. Examples
FASEB-NS-67 and FAA-RD-68-09.
2. Leave blank.
"3. Recipient's Accession Number. Reserved for use by each report recipient.
\
4. Title and Subtitle. Title should indicate clearly and briefly the subject coverage of the report, and be displayed promi-
nently. Set subtitle, if used, in smaller type or otherwise subordinate it to main title. When a report is prepared in more
than one volume, repeat the primary title, add volume number and include subtitle for the specific volume.
5. Report Dote. Each report shall'carry a date indicating at least month and year. Indicate the basis on which it was selected
(e.g., date of issue, date of approval, date of preparation.
6. Performing Organisation Code. Leave blank.
7. Authors). Give name(s) in.conventional order (e.g., John R. Doe, or J.Robert Doe). List author's affiliation if it differs
from the performing organization.
8- Performing Organ!cation .Report Number. Insert if performing organization wishes to assign this number.
9. Performing Organization Name and Address. Give name, street, city, state, and zip code. List no more than two levels of
an organizational hierarchy. Display the pame of the organization exactly as it should appear in Government indexes such
as USCRDR-I.
10. Proiect/Tosk/Work Unit Number. Use the project, task and work unit numbers under which the report was prepared.
11. Contract/Grant Number. Insert contract or grant number under which report was prepared.
12. Sponsoring Agency Nome and Address. Include zip code.
13* Type of Report and Period Covered. Indicate interim, final, etc., and, if applicable, dates covered.
14. 'Sponsoring Agency Code. Leave blank.
IS. Supplementary Note*. Enter information not included elsewhere but useful, such as: Prepared in cooperation with . . .
Translation of ... Presented at conference of ... To be published in ... Supersedes . . . Supplements . . .
16. Abstract. Include a brief (200 words or less) factual summary of the most significant information contained in the report.
If the report contains a significant bibliography or literature survey, mention it here.
17. Key Word* and Document Analysis, (a). Descriptors. Select from the Thesaurus'of Engineering and Scientific Terms the
proper authorized terms that identify the major concept of the research and are sufficiently specific and precise to be used
as index entries for cataloging.
(b). Identifiers and Open-Ended Terms. Use identifiers for project names, code names, equipment designators, etc. Use
open-ended terms written in descriptor form for those subjects for which no descriptor exists.
(c). COSATI Field/Croup. Field and Group assignments are to be taken from the 196) COSATI Subject Category List.
Since the majority of documents are multidisciplinary in nature, the primary Field/Group assignment(s) will be the specific
discipline, area of human endeavor, or type of physical object. The application(s) will be cross-referenced with secondary
Field/Group assignments that will follow the primary posting(s).
IS. Distribution Statement. Denote releasability to the public or limitation for.reasons other than security for example "Re-
lease unlimited". Cite any availability to the public, with address and price.
19 & 20. Security Classification. Do not submit classified reports to the National Technical
21. Number of Page*. Insert the total number of pages, including this one and unnumbered pages, but excluding distribution
list-, if any.
22. Price. Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
FORM NTIS-SS (REV. S-T2) ' USCOMM-DC I4OT2-P72
£ G. P. O. 1973 — 746-772 / 4199. REGION NO. 4
------- |