United States
Environmental
Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-450/4-9C-007A
JUNE 1990
AIR
BE PA USER'S GUIDE FOR THE
URBAN AIRSHED MODEL
Volume I: User's Manual for UAM (CB-IV)
-------
EPA-450/4-90-007A
USER'S GUIDE FOR THE
URBAN AIRSHED MODEL
Volume I: User's Manual for UAM (CB-IV)
By
Ralph E. Morris
Thomas C. Myers
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, CA 94903
EPA Project Officer:
Richard D. Scheffe
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
JUNE 1990
U S. Environmental Protection Agenef
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
-------
Notice
This material has been funded wholly or in part by the United States Environmental
Protection Agency under contracts 68-02-4352 and 68D90066 to Systems Applica-
tions, Inc. It has been subject to the agency's review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
90008 1
-------
Preface
This user's guide for the Urban Airshed Model (UAM) is divided into five volumes as
follows:
Volume I-User's Manual for UAM(CB-IV)
Volume II— User's Manual for the UAM(CB-IV) Modeling System (Preprocessors)
Volume Ill—User's Manual for the Diagnostic Wind Model
Volume IV—User's Manual for the Emissions Preprocessor .System
Volume V—Description and Operation of the ROM-UAM Interface Program
System
Volume I provides historical background on the model and describes in general the
scientific basis for the model. It describes the structure of the required unformatted
(binary) files that are used directly as input to UAM. This volume also presents the
formats of the output files and information on how to run an actual UAM
simulation. For those user's that already possess a UAM modeling data base or have
prepared inputs without the use of the standard UAM preprocessors, this volume
should serve as a self-sufficient guide to running the model.
Volume II describes the file formats and software for each of the standard UAM
preprocessors that are part of the UAM modeling system. The preprocessor input
files are ASCII files that are generated from raw input data (meteorological, air
quality, emissions). The preprocessor input files are then read by individual
preprocessor programs to create the unformatted (binary) files that are read directly
by the UAM. Included in this volume is an example problem that illustrates how
inputs were created from measurement data for an application of the UAM in
Atlanta. The preprocessers available for generating wind fields and emission
inventories for the UAM are described separately in Volumes III and IV, respectively.
Volume III is the user's manual for the Diagnostic Wind Model (DWM). This model is
a stand-alone interpolative wind model that uses surface- and upper-level wind
observations at selected sites within the modeling domain of interest to provide
hourly, gridded, three-dimensional estimates of winds using objective techniques. It
provides one means of formulating wind field inputs to the UAM.
Volume IV describes in detail the Emission Preprocessor System (EPS). This software
package is used to process anthropogenic area and point source emissions for UAM
from countywide average total hydrocarbon, NOX, and carbon monoxide emissions
available from national emission inventories, such as the National Emissions Data
System or the National Acid Precipitation Assessment Program. An appendix to this
90008 1
-------
volume descrioes the Biogenic Emissions Inventory System (BEIS), wmch can oe used
to generate gridded, speciated biogenic emissions. Software for merging the
anthropogenic area, mobile, and biogenic emission files into UAM input format is
also described in this volume.
Volume V describes the ROM-UAM interface program system, a softare package that
can be used to generate UAM input files from inputs and outputs provided by the
EPA Regional Oxidant Model (ROM).
IV
90008 1
-------
Acknowledgements
Since its initial conception in the early 1970s, many individuals have contributed to
the development of the Urban Airshed Model. This document reflects the latest
methodology and software development and provides a guide for new user's of the
model. Based on the past efforts of the orginal developers of the UAM and the
authors of the original 1978 user's manual, the first four volumes were written by the
following individuals from Systems Applications, Inc.:
Volume I Ralph E. Morris, Thomas C. Myers, Jay L. Haney
Volume II Ralph E. Morris, Thomas C. Myers, Edward L. Carr, Marianne C.
Causley, Sharon G. Douglas, 3ay L. Haney
Volume III Sharon G. Douglas, Robert C. Kessler, Edward L. Carr
Volume IV Marianne C. Causley, Julie L. Fieber, Michele Jimenez, LuAnn
Gardner
Volume V, containing the ROM-UAM Interface Program Guide, as well as Appendix D
in Volume IV (Biogenics Emission Inventory System) were written by the following
individuals of Computer Sciences Corporation and EPA's Atmospheric Science
Modeling Division:
Volume V Ruen-Tai Tang, Susan C. Gerry, Joseph S. Newsom, Allan R. Van
Meter, and Richard A. Wayland (CSC); James M. Godowitch and
KenSchere (EPA)
The U.S. Environmental Protection Agency provided support for the preparation of
this document. We also acknowledge the support of the South Coast Air Quality
Management District for the initial documentation of the UAM (CB-IV). Richard D.
Scheffe, Ned Meyer, Dennis Doll, and Ellen Baidridge of the U.S. EPA's Office of Air
Quality Planning and Standards contributed to this document with their insightful
technical reviews. Henry Hogo and Tom Chico of the South Coast Air Quality-
Management District also reviewed the documents and provided their comments.
Others at Systems Applications that have contributed to the continued development
of the UAM in the last few years include Dr. Gary Whitten and Mr. Gary Moore. The
technical editing of this manual was performed by Mr. Howard Beckman. We would
like to acknowledge him for his excellent work in reviewing, editing, and clarifying
the text of this manual for easier readability. Finally, we would like to acknowledge
Rita Seacock, Jo Ann Moennighoff, and Cristi-Ann Griggs for their work in producing
the document.
-------
Contents
Preface iii
Acknowledgements v
List of Figures ix
List of Tables xi
1 INTRODUCTION 1
1.1 Conceptual Overview of the Model 2
1.2 History of the Development of the Model 3
1.3 Input Data Required by the Model 10
1.* Model Output 12
2 TECHNICAL FORMULATION 19
2.1 Overview of Model Concepts 20
2.2 Treatment of Atmospheric Chemistry 23
2.3 Treatment of Advective Pollutant Transport 25
2.4 Treatment of Turbulent Diffusion 25
2.5 Treatment of Surface Removal Processes 26
3 MODEL INPUT REQUIREMENTS 27
3.1 File Types Used in the UAM Modeling System 27
3.2 Overview of UAM Input Data Files 28
3.2.1 Header Records 32
3.2.2 Time-Invariant Data 34
3.2.3 Time-Varying Data 35
3.3 AIRQUALITY 37
3.4 BOUNDARY 41
3.5 CHEMPARAM 45
3.6 DIFFBREAK 49
3.7 EMISSIONS 53
3.8 METSCALARS 57
Vll
-------
3.9 PTSOURCE 61
3.10 REGIONTOP 65
3.11 SIMCONTROL 69
3.12 TEMPERATUR 73
3.13 TERRAIN 77
3.1* TOPCONC 81
3.15 WIND 85
4 MODEL OUTPUTS . 89
4.1 SIMULATION OUTPUT 89
4.2 AVERAGE 89
4.3 INSTANT 92
4.4 DEPOSITION 95
5 COMPUTER USER NOTES 97
5.i Adaptation of File Handing for Different
Computer Systems 97
5.1.1 IBM3CL 97
5.1.2 UNIX-Based Operating System 102
5.1.3 Location of File Manipulation Statements
in the UAM 102
5.2 Allocation of Memory 106
General References 109
References for Table 1-1 115
Historical Bibliography on the Urban Airshed Model 121
Appendix I: The Carbon-Bond IV Chemical Mechanism and
Implementation in the UAM 129
Part 1: A Photochemical Kinetics Mechanism for
Urban and Regional Scale Computer Modeling 133
Appendix A: Reactions and Species 179
Appendix B: Reactive Hydrocarbon Chemistry 195
Part 2: Implementation of the CB-IV in the UAM 243
Appendix II: Numerical Representation of Horizontal
Advection in the UAM 257
90008 1
-------
Figures
2-1 Schematic illustration of the use of the grid and treatment of
atmospheric processes in the Urban Airshed Model 21
2-2 Example of vertical cell distribution as a function of time or
space for a UAM region 22
3-1 UAM input file preparation process 29
3-2 Urban Airshed Model simulation program with input and
output files 30
Q n n n o i
IX
-------
Tables
1-1 Use of the Urban Airshed Model (UAM) in studies of air quality
in urban areas as of June 1990 13
2-1 Definition of the UAM (CB-IV) state species 24
-------
1 INTRODUCTION
This user's guide describes the latest version of the Urban Airshed Model (UAM),
which contains Version IV of the Carbon-Bond Chemical Mechanism (CB-IV). This
volume of the user's guide provides an overview of the history of the model's
conception and development, a list of the applications to date, a summary of the
model's technical formulation, details of the binary file formats for each of the input
files, a description of the outputs, and information on how to run the model.
Documentation for the 1978 version of the UAM, which contained the Carbon-Bond II
Mechanism, consisted of a user's manual and systems manual (Ames et al., 1978,
revised 1985). Over the last several years the UAM has undergone a number of major
modifications that are not documented in these earlier manuals, and some of the
information contained in these manuals is outdated. Furthermore, the earlier user's
guide is at best described as a reference manual; it does not provide adequate infor-
mation for first-time users.
It is anticipated that more federal, state and local agencies will use the UAM in
developing future ozone air quality plans for urban areas. This new guide (comprising
four volumes) describes the procedures needed to exercise the UAM(CB-IV). A fifth
volume describes the Regional Oxidant Model (ROM) - UAM interface program that
can be used to prepare UAM input files from ROM input and output data sets.
Although these manuals are intended to be self-contained and provide sufficient
information for operating the UAM, there still may be some information and pro-
cedures in the earlier manuals (Ames et al., 1985a,b) that may be of interest to
present-day users of the UAM.
90006 2
-------
1.1 CONCEPTUAL OVERVIEW OF THE MODEL
The UAM is a three-dimensional photochemical grid model designed to calculate the
concentrations of both inert and chemically reactive pollutants by simulating the
physical and chemical processes in the atmosphere that affect pollutant concentra-
tions. The basis for the UAM is the atmospheric diffusion or species continuity
equation. This equation represents a mass balance in which all of the relevant emis-
sions, transport, diffusion, chemical reactions, and removal processes are expressed
in mathematical terms. The model is usually applied to an 8- to 72- hour period
during which adverse meteorological conditions result in elevated pollutant concen-
trations of the chemical species of interest.
The major factors that affect photochemical air quality include:
The spatial and temporal distribution of emissions of NOX and VOC (both
anthropogenic and biogenic),
The composition of the emitted VOC and NOX,
The spatial and temporal variations in the wind fields,
The dynamics of the boundary layer, including stability and the level of mixing,
The chemical reactions involving VOC, NOX, and other important species,
The diurnal variations of solar insolation and temperature,
The loss of ozone and ozone precursors by dry deposition, and
The ambient background of VOC, NOX, and other species in, immediately
upwind, and above the region of study.
The UAM simulates these processes when it is used to calculate summertime ozone
concentrations. It can also be used to simulate wintertime carbon monoxide concen-
trations in an urban area, a simulation that involves no chemical reactions. In the
90008 2
-------
model, the species continuity equation is solved in four steps: (1) advection/diffusion
is solved in the x-direction, (2) advection/dif fusion is solved in the y-direction, (3)
emissions are injected and vertical advection/diffusion is solved, and (4) chemical
transformations are performed for reactive pollutants. The UAM performs this four-
step procedure during each time step. The maximum time step is a function of the
grid size and the maximum wind velocity. Typical time steps for urban-scale simula-
tions are on the order of 3 to 6 minutes.
Because the UAM accounts for spatial and temporal variations as well as differences
in the reactivity (speciation) of emissions, it is ideally suited for evaluating the
effects of emission control scenarios on urban air quality. This is accomplished by
first replicating a historical ozone episode to establish a base case simulation. Model
inputs are prepared from observed meteorological, emission, and air quality data for
a particular day or days. The model is then applied with these inputs and the results
are evaluated to determine its performance. Once the model results have been
evaluated and determined to perform within prescribed levels, the same meteoro-
logical inputs and a projected emission inventory can be used to simulate possible
future emission scenarios. That is, the model will calculate hourly ozone patterns
likely to occur under the same meteorological conditions as the base case.
1.2 HISTORY OF THE DEVELOPMENT OF THE MODEL
The UAM has been under continual development for over 20 years, involving more
than 100 person-years of technical effort. It has been supported by many organiza-
tions, with the U.S. Environmental Protection Agency providing most of the financial
support. Several other public and private organizations have contributed to the
substantial effort of demonstrating the utility of the UAM to investigate complex
ozone air quality management issues.
The history and development of mathematical photochemical models, particularly
the UAM, has been paced by advances along three fronts.
The scientific front, which is governed by the scientific community's accept-
ance of a suitable formulation, of supporting algorithms that represent perti-
90008 2
-------
nent physical and chemical processes, and of measurement methods and data
bases that support parameter estimation and model performance evaluations.
The regulatory front, which is governed by the relevance and practicality of
the UAM to evolving regulatory programs and by acceptance of decision-
makers.
The computing technology front, which is governed by the availability (to air
quality modelers) of computing systems capable of large-scale numerical
modeling, by the transportability of the UAM to those systems and by the UAM
being relatively "friendly" to users.
Since 1969, when the UAM was first conceived, substantial changes across all three
fronts have occurred. Some of these changes were anticipated, some changes were
completely outside the control of the developers even when they were anticipated,
and some were not anticipated.
1.2.1 Development and Scientific Evaluation
*
In the 1950s and 1960s many urban areas of the country became aware of periods of
elevated photochemical oxidants consisting mainly of ozone (also called smog),
although the problem of urban ozone was first recognized in the Los Angeles area in
the 1940s. The 1963 Clean Air Act provided limited federal power to abate air
pollution endangering health and welfare. It also instructed the Department of
Health, Education, and Welfare to develop criteria on the effects of air pollution and
its control. The 1967 Air Quality Act required states to establish air quality stan-
dards consistent with federal criteria, and established the National Air-Pollution
Control Administration (NAPCA), EPA's predecessor.
Many states failed to implement provisions of this act and there was little evidence
that air pollution was being reduced on a national scale. The Clean Air Amendments
of 1970 provided for uniform national air quality standards, set by the federal
government, and mandated that states develop abatement programs to meet the
standards. Prior to 1970 a model that could be used to assess urban ozone and to
90008 2
-------
evaluate control strategies for air quality planners did not exist, and the need for
such a tool was recognized. In 1969 the NAPCA contracted with Systems Applica-
tions, Inc. for the development of such a model, eventually to be called the Urban
Airshed Model (UAM).
Based on the time-dependent, three-dimensional eolations describing the conserva-
tion of mass of a chemically reactive pollutant in a turbulent atmosphere, the UAM
represented a significant departure from the steady-state, spatially homogeneous
Gaussian formulations used to evaluate air quality, since the latter could only treat
first-order chemical transformations. The solution of the system of coupled partial
differential equations required using numerical integration techniques. This in turn
required that the modeling region be partitioned into an array of three-dimensional
grid cells, each having fixed horizontal dimensions from 2 to 10 km and time-varying
vertical dimensions from 50 to several hundred meters. Selection of cell size and
number of cells is governed by consideration of reactive hydrocarbon and NOX
emission sources that contribute to ozone formation within and downwind of the
urban region and of computing resources. Documentation of the early development
of the UAM is contained in a major report issued in 1971.*
The model provides mathematical representations of precursor emissions, pollutant
transport, turbulent diffusion, chemical reactions, removal processes, and initial and
boundary conditions. Among the important features of the UAM formulation is that
it provides a framework within which to represent mathematically the most current
understanding of the physical and chemical processes that contribute to ozone for-
mation in and downwind of urban areas. Thus, as improved representation of atmor
spheric phenomena are developed, existing algorithms can be replaced or new pro-
cesses added by modifying the appropriate portions of the code.
* A historical bibliography, arranged chronologically, will be found at the end of
Volume 1.
90006 2
-------
The initial model development and evaluation phase, from 1969 through 1973, cul-
minated with the judgement by the EPA and its peer reviewers that the UAM
approach was feasible and practical. This led to further development and evaluation.
The period from 1973 through 1977, also heavily supported by the EPA, emphasized
algorithm development and evaluation. During this period the Carbon Bond Method
(CBM-l)~a method for condensing and simplifying complex explicit mechanisms
describing photochemical oxidant formation—was developed and tested; methods for
treating the effects of temperature and UV scattering by airborne particles on
reaction rate constants were identified, evaluated, and selected; a computationally
efficient approach for treating nighttime ozone chemistry was selected; alternative,
improved methods of numerical integration were examined, and the SHASTA (Sharp
and Smooth Transport Algorithm) procedure selected; improved representations of
vertical diffusivity, plume rise, and surface removal were evaluated and incorporated
in the UAM; a module was developed for treating the influence of subgrid-scale NOX
emissions from point and line sources on the grid-average ozone concentrations
output by the model; and the computer codes was modified to enable the input of a
fully three-dimensional wind field. A seven-volume series of reports issued in 1976
documents these development efforts (see bibliography).
The period 1977 to 1978 was devoted to improving the software and documentation
and to systematic evaluation of UAM performance using measurements obtained in
the St. Louis area as part of the EPA's Regional Air Pollution Study (RAPS). In
designing the software, consideration was given to those computer systems to which
the air pollution modeling community would likely have access in the 1979 to 1982
period. Because of EPA's desire that the UAM be exercised principally on the
UNIVAC and other systems of relatively modest computer power, the code design
was based on machines with an 8-bit architecture and small core memory without
"virtual" memory capability. To overcome the memory restrictions, special features
such as "overlay" and "grid segmentation" were adopted. The UAM was judged by
EPA's Office of Research and Development and Office of Air Quality Planning and
Standards to perform successfully for the 20-plus days chosen for examination from
the St. Louis data base. After further tests by EPA's staff, this version of the code
was made available to the public in 1980.
90006 2
-------
Since 1980, work has continued on improving chemical mechanisms for daylight and
nighttime conditions and for gaseous and liquid media, removal processes, numerical
methods for integrating parabolic equations, methods for improving gas-to-particle
conversion and particle dynamics, and improved methods for prescribing UAM inputs
and for evaluating UAM performance.
In 1984 the EPA's Office of Air Quality Planning and Standard proposed that the
UAM be a "recommended" (preferred) model for "photochemical pollutant modeling
applications involving entire urban areas." In finalizing its recommendation in 1986
the EPA noted that the UAM "is the most widely applied and evaluated photochemi-
cal dispersion model in existence."
The 1980 publically released version of the model contained the Carbon-Bond II
chemical mechanism. On the basis of experience with many different applications of
the UAM and recent technical advances, this version of the model was updated in
1988. The two main improvements to the UAM CB-II version was the incorporation
of the Carbon Bond IV chemical mechanism and the use of the Smolarkiewicz
numerical integration scheme to solve the advection equation.
1.2.2 Applications of the Urban Airshed Model
This section briefly describes some of the applications of the UAM. A comprehen-
sive list of UAM applications to date is provided in Table 1-1 (at the end of this
chapter).
UAM applications to areas other than Los Angeles were limited in the early 1970s.
Two early studies, both for Los Angeles, were adjuncts to ongoing UAM development
efforts. They examined alternative motor vehicle control strategies and two future-
year emission scenarios identified in the California State Implementation Plan of
1973.
The first serious UAM application was sought by the California Department of
Transportation and the California Air Resources Board, who jointly sponsored a
project in 1974 to expand the Los Angeles modeling region by approximately a factor
90008 2
-------
of four and to establish a set of model inputs for 3une 26-27, 1974, when some of the
highest ozone levels in the region occurred. The inputs developed at that time were
to be the subject of substantial review and scrutiny and the methodologies developed
were followed in several UAM applications over the next 10 years.
The first regulatory use and practical applications of the UAM were carried out for
the Denver area on behalf of the Colorado Division of Highways and EPA's Region
VII. The UAM was used to evaluate whether various transportation plans and pro-
grams were consistent with the SIP and to evaluate the effects on Denver's air
quality of urban growth and development that might result from the construction of
proposed wastewater treatment facilities. Three smoggy days in the years 1975 and
1976 were used in the study. This study demonstrated the UAM's ability to replicate
an ensemble of 1-hour average ozone observations in a manner comparable to the
uncertainty associated with routinely measured I-hour ozone concentrations.
Another comprehensive UAM performance evaluation and demonstration study was
performed for the Federal Highway Administration. Here, single day and multiday
simulations were performed for the June 26-27, 1974 episode in the Los Angeles
region, three days for the Denver area, and two days for Sacramento. Several sen-
sitivity runs were performed for a "hypothetical" city to examine the possible
influence of alternative highway location, design, and operation strategies on ozone
formation.
In the late 1970s EPA's OAQPS initiated a program to examine the applicability and
practicality of UAM in routine ozone attainment demonstrations required by the SIP
process. Data collection, emission inventory development, model performance
evaluation, and application were major elements of this nationwide program.
Building off the St. Louis UAM applications and an extensive series of UAM sensi-
tivity studies designed to provide guidance concerning the types and amounts of data
required to support UAM application, data for an application of the UAM, supported
by OAQPS, were collected in Tulsa, Philadelphia/New Jersey, Baltimore/Washington,
and New York. Because of funding limitations in the early 1980s, not all aspects of
the program have been completed.
90008 2
-------
Other regulatory and policy related DAM applications were performed for the
Southern California Association of Governments (SCAG) and the National Commis-
sion on Air Quality (NCAQ). The SCAG study examined the influence on calculated
ozone levels of six alternative population, housing, employment, and land use fore-
casts for 1990. The results supported the development of the 1982 AQMP for the Los
Angeles area. The NCAQ project examined the effect on ozone levels of several
alternative emission reduction strategies to address the attainability of the federal
and ozone standards in the Los Angeles area.
In the late 1970s NOX controls on stationary sources emerged as a potential way of
facilitating attainment of ozone standards and other air quality goals in Los
Angeles. Studies were performed for Southern California Edison, the Western States
Petroleum Association, and the California Air Resources Board to examine the
effectiveness of NOX controls on stationary sources in reducing ozone and NC>2
concentration levels. These and other studies in the early 1980s provided new
insights about model performance and provided data bases for new simulation periods
for the region, August 4, 197>and November 7-8, 1978. The studies also provided
insights about the potential for antagonistic effects on ozone concentrations and
population exposures when reductions in NOX emissions did not coincide with
comparable reductions in reactive hydrocarbon emissions.
In 1985 UAM applications were extended to address the potential impact of off-shore
oil and gas development on inland ozone levels in coastal regions of California. A
study known as the Joint Interagency Modeling Study—conducted for the EPA's
Region IX and the Department of Interior's Mineral Management Services, with the
participation of the California Air Resources Board and counties of Ventura, Santa
Barbara and San Louis Obispo — was carried out to assess the feasibility of using the
UAM as an alternative to EKMA in preparing a State Implementation Plan for
coastal counties. Based on evaluations of model performance for three two-day
periods, the UAM's performance was judged to be preferable for SIP purposes. In
1985 the model was also used to address ozone attainment issues in the Kern County,
California area, an area of the southern San Joaquin Valley that contains precursor
emissions from oil and gas development activities and other industries. This work
was performed for the California Air Resources Board.
90008 2
-------
Applications of the latest version of the UAM with CB-IV from 1988 to present
include: the Los Angeles area, as part of the 1988 Air Quality Management Plan; the
South Central Coast Air Basin in California, to assess the impacts of OCS sources;
Baton Rouge, Louisiana, for SIP development; the cities of Taipei and Kaohsiung in
Taiwan, for air quality planning activities;, and the cities of New York, St. Louis,
Atlanta, Dallas-Fort Worth and Philadelphia as part of an EPA-sponsored UAM
demonstration study that also evaluated the impacts on urban ozone of the use of
alternative fuels.
Routine use of the UAM is an emerging practice. Since its first use for air quality
planning in 1978, the UAM has been or is being used by the U.S. EPA, Exxon Corpora-
tion, British-Leyiand, the TNO in the Netherlands, Pacific Gas and Electric, Southern
California Edison Company, Arizona Department of Health, the South Coast
(California) Air Quality Management District, the New York Department of
Environmental Protection, the California Air Resources Board, the Connecticut
Department of Environmental Protection, the Georgia State Air Protection Branch,
and the National Institute of Environmental Studies (NIES) of Japan. In addition,
several consulting firms who provide air quality modeling services have installed and
operate the UAM. As part of this and other efforts, the UAM has been installed and
operated on UNIVAC, CDC, Cray Research, IBM, VAX, Prime, Amdahl, Data
General, and Multiflow Trace computer systems. Most recently, the UAM became
operational on IBM/XT systems with enhanced high-speed memory capabilities.
13 INPUT DATA REQUIRED BY THE MODEL
The UAM simulates the emission, advection, and dispersion of precursors and the
formation and deposition of ozone within every grid cell of the modeling domain (i.e.,
for the entire urban area). The successful and technically defensible simulation of
ozone formation and transport can be more easily accomplished with an enhanced
meteorological data base. However, recent applications of the UAM in five U.S.
cities demonstrated that using routine meteorological and air quality data for UAM
input is feasible in less complex airsheds (Morris et al., 1990a,b,c; Morris, Myers, and
Carr, 1990).
90008 2
10
-------
The use of UAM to adequately replicate the full three-dimensional structure of the
atmosphere during an ozone episode requires a day-specific data base for input
preparation. For UAM applications, the observed air quality data are used to esti-
mate the initial condition field for ozone, NOX, and volatile organic compounds
(VOC). These data are also used to evaluate model predictions. For urban applica-
tions, the UAM is usually used to simulate a two- or three-day episode, and the
simulation is started sometime during the early morning hours of the first day. This
procedure is followed so that the peak model calculations are not driven by the
prescribed (and possibly errant) initial condition field. Nitrogen oxides (NOX) are
important precursors in ozone formation, and the levels of NO and NC>2 calculated by
the UAM can be evaluated with NOX data from continuous samplers. It is desirable
to have data from a number of NOX sampler sites scattered throughout the modeling
domain. Concentrations of reactive hydrocarbons are not required to run the model,
however, in recent years measurements of reactive hydrocarbons have become more
and more desirable to check the modeled concentrations and the calculated hydro-
carbon to NOx ratios at various locations within the modeling domain.
The UAM requires hourly estimates for the height of the mixed layer. Because ozone
concentrations calculated by UAM are sensitive to mixing heights, day-specific
upper-air temperature and wind data are required at various times throughout the
day to adequately estimate the evolution of the nighttime and daytime mixed
layers. Other meteorological data required by the UAM include ambient tempera-
ture, water concentration (derived from relative humidity measurements),
atmospheric pressure, solar radiation, and cloud cover. In addition, the UAM
requires a fully three-dimensional wind field for each hour. Upper-level wind data
are used to estimate the flow field throughout and above the urban boundary layer
and surface measurements throughout the domain provide data for the surface wind
fields.
The UAM also requires hourly gridded emissions for NOX and VOC. For VOCs the
UAM can simulate the fate of emissions from anthropogenic and biogenic emission
sources. In addition, the VOC emissions must be speciated or classified into the
respective carbon-bond class because UAM employs the Carbon-Bond chemical
kinetics mechanism.
90008 2
11
-------
1.4 MODEL OUTPUT
The user can specify the output interval for which average concentrations are cal-
culated. This interval is usually one hour because most of the ambient air quality
standards are based on hourly averages. At each time step the model calculates
concentrations in all grid cells for all species. The DAM computes the average
concentrations for the user-specified time interval using these concentrations. The
UAM also provides the instantaneous concentrations for each species for each grid
cell at the beginning of the averaging period. The post processor program DISPLAY
is included with the UAM modeling system software package. This program provides
printouts of concentrations calculated by the model for each grid cell. The concen-
tration information that is output from UAM can also be displayed graphically.
90006 2
-------
TABLE 1-1. Use of the Urban Airshed Model (UAM) in studies of air quality in urban
areas as of June 1990. (References will be found in a separate list at the end
of this volume.)
Study Site
Study
Date
User
Sponsor and Reference
U.S. Applications
Atlanta
Baton Rouge
1989
(CB-IV)
1990
(CB-IV)
Central Valley 1978
Dallas-Fort Worth 1989
(CB-IV)
Denver
Denver
Denver
Denver
Denver
Denver
1973
1976
1976
1977
1977
1983
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc./
State of Colorado
Systems Appl., Inc.
Systems Appl.. Inc./
State of Colorado
Systems Appl., Inc.
Systems Appl., Inc.
National Center for
Atmospheric Research
EPA OAQPS
(Morris et al., 1990a)
Louisiana Ozone Task Force
Louisiana Dept. Env. Quality
(Haney et al., 1990)
Pacific Gas and Electric
(Roth et al., 1981)
EPA OAQPS
(Morris et al., 1990c)
Colorado Department of
Highways Contract 75-109
EPA Region VIII, Contract
68-01-4341 (Anderson et
al., 1977)
Colorado State Health Dept.
and Dept. of Highways
DOT, DOT-FH-11-8529
(Reynolds et al.,1979)
ERT/Colorado Interstate
Pipeline Co.
EPA, AD-49-F-0-167-0
(Dennis, Downton, and
Keil, 1983)
continued
90006 12
."13
-------
TABLE 1-1. Continued.
Study Site
Study
Date
User
Sponsor and Reference
Kern County 1984
Kern County 1985
Las Vegas 1975
Las Vegas 1982
Los Angeles 1973
Los Angeles 1973
Los Angeles 1973
Los Angeles 1977
Los Angeles 1977
Los Angeles 1977
Los Angeles 1980
California Air
Resources Board
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
and CalTrans
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
CARB and Western Oil and
Gas Association
(CARB, 1985)
Western Oil and Gas
Association (Whitten
et al., 1985)
EPA, 68-03-2174
(Liu et al., 1977)
Southern California Edison
Co. (Tesche, Oliver, and
Haney, 1982)
EPA, CPA 70-148
and 68-02-0339
(Reynolds, 1973)
California Department of
Transportation, Contract
K-7319
Texaco (Attaway et al.,
1975)
DOT, DOT-FH-11-8529
(Reynolds et al., 1979)
Southern California Edison
Company (Tesche and Burton,
1978)
Southern California Assoc.
of Governments
Systems Appl., Inc. Texaco (Doyle et al., 1980)
continued
90008 12
-------
TABLE 1-1. Continued.
Study Site
Study
Date
User '
Sponsor and Reference
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
Los Angeles
1978
1980
1981
1982
1982
1984
1986
1988
New York Metro. 1986
New York Metro. 1987
New York Metro. 1988
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc
New York Dept. of
Environ. Conservation
Connecticut Dept. of
Environ. Protection
Systems Appl., Inc.
Southern California Assoc.
of Governments and South
Coast Air Quality Mmgt.
District (Reynolds et al.,
1978)
Nat. Comm. on Air Quality
(Souten, Tesche, and
Oliver, 1981)
EPA, Contract 68-02-2870
(Tesche et al., 1981)
EPA, Contract RP-1375-1-2
(Seigneur et al., 1983)
Western Oil and Gas Assoc.
(Roth et al., 1984)
South Coast Air Quality
Mgmt. District
(Liu and Grisinger, 1986)
South Coast Air Quality
Mgmt. District
(Hogo and Yocke, 1987)
South Coast Air Quality
Mgmt. District
(Hogo et al. 1988)
NYDEC (Rao, 1987)
CTDEP (Wackter, 1988)
EPA OPPE, ORD, OMS, OAQPS,
(Morris et al., 1990b)
continued
90006 12
15
-------
TABLE 1-1. Continued.
Study Site
Study
Date
User
Sponsor and Reference
Philadelphia
Philadelphia
Philadelphia
Phoenix
Sacramento
San Diego
San Francisco
Bay Area
San Francisco
Bay Area
San Francisco
Bay Area
Santa Barbara-
Ventura
Santa Barbara-
Ventura
Santa Barbara-
Ventura
1983-85
1988
(CB-IV)
1989
(CB-IV)
1985-86
1978
1986
1987
1987
1988-90
1985
1985
1987
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl. Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Bay Area Air Quality
Management District
Radian Corporation
Systems Appl., Inc.
Systems Appl., Inc.
Applied Modeling,
Inc.
Radian Corporation
EPA, 68-02-3582 (Haney
and Braverman, 1985)
Electric Power Research
Institute (Haney, Whitten,
and Burton, 1988)
EPA, OAQPS (Morris, Myers,
and Carr, 1990)
Maricopa Cty., carbon
monoxide only (Haney,
1986)
DOT, DOT-FH-11-8529
(Reynolds et al., 1979)
Signal Environ. Systems,
Inc. (Yocke et al., 1987)
BAAQMD (BAAQMD, 1987)
California Energy Comm.
(Tesche et al., 1988)
BAAQMD - (Moore et al.,
1990)
EPA Region IX, CARB
(Haney et al., 1986)
State Lands Commission,
Santa Barbara Co.
(SLC, 1986)
EPA Region IX, Santa
Barbara Co. (Tesche et
al., 1988)
continued
90008 12
16
-------
TABLE 1-1. Continued.
Study Site
Study
Date
User
Sponsor and Reference
Santa Barbara- 1986
Ventura
Santa Barbara- 1988
Ventura
Santa Barbara- 1989-90
Ventura
Santa Maria DCS 1985
San Luis Obispo
St. Louis
St. Louis
St. Louis
Tulsa
1982
1982
1989
1982
Non-U.S. Applications
Athens, Greece 1989
Bonn/Cologne 1980
The Netherlands 1978
Taiwan (cities 1989
of Taipei and (CB-IV)
Kaohsiung
Systems Appl., Inc.
Calif. Air Resources
Board
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Systems Appl., Inc.
Freie Universitat/TNO
Systems Appl., Inc./
TWO
Systems Appl., Inc./
Freie Universitat
Systems Appl., Inc.
Texaco Project Proteus
Region IX, CARS (Mahoney
and Yocke, 1987)
Santa Barbara Co., CARB
(Wagner and Wheeler,
in press)
DOE MMS (Daly et al.f
1990)
Cities Services Oil and
Gas Corp. (Yocke et al.,
1985)
EPA, 68-02-2429
(Cole et al., 1983)
EPRI, RF-1375-1-2
(Seigneur et al., 1983)
EPA OPPE, ORD, QMS, OAQPS
(Morris, Myers, and Carr, 1990)
EPA 68-02-3370
(Reynolds et al., 1982)
SAI demonstration
Fed. Rep. of Germany
(Stern and Scherer, 1982)
TWO (Builtjes et al., 1982)
Republic of China EPA
(Carr and Haney, 1990)
continued
17
-------
TABLE 1-1. Concluded.
Study
Study Site Date User Sponsor and Reference
Tokyo Metro. 1985 National Institute of NIES-Japan (Wakamatsu
Environmental Studies et al., 1986, 1987)
Turin, Italy 1973 Systems Appl., .Inc./ British Leyland
British Leyland (Gaddo and Weaving, 1982)
18
-------
2 TECHNICAL FORMULATION
The Urban Airshed Model (UAM) is a three-dimensional grid (Eulerian) model
designed to calculate the concentrations of both inert and chemically reactive pol-
lutants by simulating the physical and chemical processes in the atmosphere that
affect pollutant concentrations. The basis for the UAM is the atmospheric diffusion
equation (also called the species continuity or advection/diffusion equation). This
equation represents a mass balance in which all of the relevant emissions, transport,
diffusion, chemical reactions, and removal processes are expressed in mathematical
terms as follows:
at
Time
Dependence
a(uc.) a(vc.)
ax ay
Advection
3Z
a / 3ci\ a / 3ci\ a / 3ci
o (w i.1 j. 2. \V i. I j_ ° \Y _
ax TH ax / * ay \*H ay / + az \Kv az
Turbulent Diffusion
R. H
Chemical
Reaction
> S. -
Emissions
" Li
Pollutant
Sinks
where c^ represents the pollutant concentration and is a function of space (x,y,z) and
time (t). The other terms in this equation are
u,v,w = horizontal and vertical wind speed components
KJ.J, Ky = horizontal and vertical turbulent diffusion coefficients
R^ = net rate of production of pollutant i by chemical reactions
S^ = emission rate of pollutant i
Lj = net rate of removal of pollutant i by surface uptake processes
19
90008 3
-------
2.1 OVERVIEW OF MODEL CONCEPTS
The UAM employs finite differencing numerical techniques for the solution of the
advection/diffusion equation. The region to be simulated is divided up into a three-
dimensional grid covering the region of interest (Figure 2-1). Horizontal grid cells
are rectangular with constant lengths in the x- and y-directions. Vertical layer
thicknesses are defined by the user based on the diffusion break, the top of the
region, the number of layers below and above the diffusion break, and the minimum
layer thickness. The diffusion break usually corresponds to the base of an inversion
layer, either an unstable convective layer during the day (i.e., the mixing height) or a
stable nocturnal layer at night. The region top is usually defined at or slightly above
the maximum daily diffusion break. An example of the vertical cell structure is
presented in Figure 2-2 for the case in which four vertical ceils are used: two cells
situated below the diffusion break and two cells above. The cell configuration may
change in space and time.
The UAM solves the advection/diffusion equation using the method of fractional
steps. At each integration time step, typically on the order of five minutes, the
terms in the equation that represent the different atmospheric processes (e.g.,
chemistry or diffusion) are solved separately in several steps using the most efficient
numerical integration technique for the given process. The order for solving the
equation is as follows.
Step 1—solve advection/diffusion in the x- (east-west) direction;
Step 2—solve advection/diffusion in the y- (north-south) direction;
Step 3—inject emissions and solve vertical advection/diffusion;
Step 4—perform chemical transformation of the pollutants.
The following subsections briefly describe each of the major processes of the UAM.
The treatment of these processes is described in more detail by Reynolds and others
(1976).
90008 3
20
-------
A. The Area to Be Modeled
B. Specification of the Grid
Trio sport
Transport
V)
k*
u Transport
E -
§
a
* Transport
••
\
Chemistry
~ Elevated Emissions ~
{Transport
Chemistry
— Elevated Emissions —
\ Transport
\
Chemistry
~ Elevated Emissions ~
( Transport
Ground-Level Emissions
Roadway Microscale Effects
Transport
Transport
Transport
INVERSION TOP
INVERSION BASE
GROUND SURFACE
Surface Removal
1 to 10 kilometers
C. Atmospheric Processes Treated in a Column of Grid Cells
FIGURE 2-1. Schematic illustration of the use of the grid and treatment of atmospheric processes in the
Urban Airshed Model.
;EE90008
21
-------
-^"mm-:
-------
2.2 TREATMENT OF ATMOSPHERIC CHEMISTRY
Ozone is formed in the atmosphere through chemical reactions between nitrogen
oxides (NOX) and volatile organic compounds (VOC). Hundreds of organic compounds
and thousands of reactions participate in the formation of ozone in the atmosphere.
The explicit treatment of all of these compounds and reactions would be prohibitive
in an Eulerian-based grid model such as the UAM. Thus most photochemical chem-
ical kinetic mechanisms treat organic compounds in groups, often on the basis of the
reactive functional groups they contain.
In general, two approaches are used to condense an explicit kinetic mechanism into a
usable number of species and reactions for use in .itmospheric simulation models:
the lumped approach and the carbon-bond approach. In the lumped approach organic
species with similar reactive products and reaction rates are combined into a single
lumped species. For example, propylene, butene, and 1-pentene, each of which con-
tain a carbon-carbon double bond, can be aggregated together into a single lumped
alkene species. In the carbon-bond approach the organic species are disaggregated
based on the carbon bonds of the organic compounds. For example propylene,
butene, and 1-pentene each have one olefinic carbon double bond but also have dif-
ferent numbers of single carbon bond. Thus in the carbon bond approach propylene,
butene, and i-pentene would be split into one olefinic bond (OLE) and one, two, and
three paraffinic bonds (PAR), respectively.
The UAM(CB-IV) employs version IV of the Carbon Bond Mechanism (CB-IV) for solv-
ing chemical kinetics (Gery, Whitten, and Killus, 1988). The CB-IV and its implemen-
tation in the UAM is described in Appendix A. As implemented in the UAM the CB-
IV contains over 80 reactions and over 30 species. The differential eolations that
describe the CB-IV are a "stiff" system, that is, the equations contain wide variations
in time (reaction rate) constants. Solving these equations with a "stiff numerical
integration scheme, such as the one developed by Gear (1971), would result in
prohibitively expensive computer time. Thus the solution of the CB-IV in the UAM
uses quasi-steady-state assumptions for the low-mass fast-reacting species (i.e., the
stiff species) and the more computationally efficient Crank-Nicholson algorithm for
the remainder of the state species. Table 2-1 lists the state species used in the
UAM(CB-IV).
90008 3
23
-------
TABLE 2-1. Definition of the UAM (CB-IV) state species.
UAM
Species
Species Name
NO
NO2
03
OLE
PAR
TOL
XYL
FORM
ALD2
ETH
CRES
MGLY
OPEN
PNA
NXOY
PAN
CO
MONO
H2O2
HNO3
MEOH
ETOH
ISOP
Nitric oxide
Nitrogen dioxide
Ozone
Olefinic carbon bond (C=C)
Paraffinic carbon bond (C-C)
Toluene (C6H5-CH3)
Xylene (C6H6-(CH3>2)
Formaldehyde (CH2=O)
High molecular weight aldehydes (RCHO, R > H)
Ethene (CH2=CH2)
Cresol and higher molecular weight phenols
Methyl giyoxal (CH3C(O)C(O)H)
Aromatic ring fragment acid
Peroxynitric acid (HO2NO2)
Total of nitrogen compounds (NO + NO2 + N2O^ + NO3)
Peroxyacyl nitrate (CH3C(O)O2NO2)
Carbon monoxide
Nitrous acid
Hydrogen peroxide
Nitric acid
Methanol (optional)
Ethanol (optional)
Isoprene (optional)
90008 6
24
-------
2.3 TREATMENT OF ADVECTIVE POLLUTANT TRANSPORT
Pollutants are transported primarily by advection, that is by the mean or bulk motion
of the air. Advection in the UAM advection is treated by specifying horizontal wind
fields (i.e. u and v wind components in each grid cell) for each vertical layer. The
vertical wind velocity in the UAM terrain-following coordinate system can then be
calculated from the conservation of mass equation. Proper specification of the
hourly and three-dimensional varying winds is one of the key steps to successful
application of the UAM. The winds influence how different emissions are mixed
together, advected downwind, and diluted. Sources of wind data for the UAM have
varied greatly, from prognostic meteorological model output (Daly et ai., 1990) to
use of constant winds (Rao, 1987). In most UAM applications the wind inputs are
defined by interpolating surface and upper-air observations or through the use of a
diagnostic wind model with interpoiative techniques.
Previous versions of the UAM used the Sharp and Smooth Transport Algorithm
(SHASTA) (Boris and Book, 1973) to solve the advection/diffusion equation. When the
SHASTA method was implemented in the UAM, it was more accurate than methods
available (Killus et al., 1984). Since then, however, studies of several other
advection schemes have indicated that SHASTA may compute excessive numerical
diffusion (Schere, 1983; Chock and Dunker, 1983; Chock, 1985; Smolarkiewicz,
1983). Thus an improved numerical advection scheme (Smoiarkiewicz, 1983) was
implemented in the UAM(CB-IV). Appendix B discusses the implementation of this
advection scheme in the UAM(CB-IV).
2.4 TREATMENT OF TURBULENT DIFFUSION
The treatment of turbulent diffusion (dispersion) in the UAM(CB-IV) is the same as in
the UAM(CB-II). Dispersion of pollutants is assumed to be proportional to the rate of
change of concentration in space (i.e., the concentration gradient). The proportion-
ality factor is termed the eddy diffusivity coefficient (Kx, Kv, and Kz in the advec-
tion/diffusion equation). Because it has been difficult to obtain precise measure-
ments of the eddy diffusivity coefficients, theoretical estimates have been used.
25
90008 3
-------
Control theory techniques are employed in conjunction with the results of a
planetary boundary layer model to generate optimal diffusivity coefficients in the
(JAM. For further details, the reader is referred to the reports by Lamb (1976) and
Lamb and co-workers (1984).
2.5 TREATMENT OF SURFACE REMOVAL PROCESSES
Many types of pollutants, including nitrogen oxides, ozone, carbon monoxide and sul-
fur compounds, are removed from the surface layer by such features as vegetation
through the process of dry deposition. In the UAM dry deposition is assumed to occur
in a two-step process: the transfer of pollutants through the atmosphere to the sur-
face and the uptake of the pollutants by vegetation and other materials at the sur-
face. This process involves a resistance to mass transport and a resistance to sur-
face removal. The transport resistance is estimated from theoretical considerations
of turbulent transfer in the atmospheric boundary layer. The surface resistance is
obtained from experimental data on the uptake of pollutants by various surface
features. The UAM surface removal formulation is described in detail by Killus and
others (1984).
90008 3
26
-------
3 MODEL INPUT REQUIREMENTS
This chapter describes the unformatted (binary) input data files that are read
directly by the DAM. It provides enough detail on the structure of these files so that
interested users can prepare UAM input files or write utility programs to modify
existing input files. Many methods can be used to create UAM inputs. Preprocessing
programs for creating UAM input files are included with the UAM(CB-IV) modeling
system software (see Volume II of this guide). If these preprocessors are used, it is
not necessary to know the structure of the binary UAM input files. Complete
descriptions of the input files to the preprocessors and their formats are found in
Volume II of the guide. The various types of files that are used in the UAM modeling
system are described next.
3.1 FILE TYPES USED IN THE UAM MODELING SYSTEM
Two different file types are used in the process of creating UAM inputs. These
include files that were written using FORTRAN format statements and which can be
examined in a printout or at a computer terminal. For some computer systems these
files are referred to as text files or ASCII files. The second type of file is written
without format statements and is considered an unformatted file. In this guide
unformatted files are called binary files and formatted files are called ASCII files.
*
Raw meteorological, air quality, or emission data collected by state or local agencies
are usually in the form of a formatted ASCII file. To reduce computer file storage
requirements (disk space) and decrease the input/output (I/O) time needed to read
the files during the simulation, the files read directly by the UAM are binary files.
These files can be readily transferred between machines that use the same operating
system via magnetic tape or telephone transfer. However, binary files created under
one operating system usually cannot be read by (1) the same computer with a
90008 f
27
-------
different operating system, (2) a different computer with a different operating
system, or (3) a different computer with the same operating system. Therefore, to
transfer the UAM binary input files between machines under one of these three
conditions, the files must first be converted into a formatted file.
Figure 3-1 shows the flow diagram and files involved in processing ambient tempera-
ture data from the raw data files obtained from an agency such as the National
Weather Service into the binary UAM input files. The raw data are in a formatted
ASCII (or text) file. These data are used to create the preprocessor input file
TEMPERATUR.IN, which is also an ASCII (or text) file. This file is read by the
preprocessor program TMPRTR, which creates the binary UAM input file
TEMPERATUR.BIN. Two programs included with the UAM. modeling system soft-
ware, AIRASCI and AIRCONV, are used to convert the formats of files. As shown in
the figure, AIRASCI converts the binary UAM input file to an ASCII file
(TEMPERATUR.ASC), which can then be transferred to another machine. This file is
written in an entirely different format than the preprocessor input file. The program
AIRCONV can be used to convert any UAM ASCII file received from another
machine into binary format for use as UAM input.
3.2 OVERVIEW OF UAM INPUT DATA FILES
The UAM uses up to 13 input files. As illustrated in Figure 3-2, these files can be
broken down into meteorological, emission, initial and boundary condition, chemistry
parameters, and simulation control data files:
AIRQUALITY This file specifies the initial concentrations for each of the CB-IV
state species (see Table 2-1) in each grid cell at the start of the
simulation.
BOUNDARY This file contains the locations of the modeling domain boundaries
and the concentrations of each species used as boundary conditions
along each lateral boundary for each level.
CHEMPARAM This file contains information on the chemical species to be simula-
ted, including reaction rate constants, upper and lower bounds,
activation energy, reference temperature, and resistance to surface
sinks.
90008 f
28
-------
TEMPERATUR.ASC
(file received)
T ASCII version
of UAM
input file
I
AIRCONV
converts ASCH
to binary
Raw data:
f meteorological,
air quality
(ASCH file)
(NWS
temperature data)
Preprocessor
input file
(ASCDfile)
(TEMPERATUR.IN)
Preprocessor
program
TEMPERATUR.BIN
UAM
input file
(binary file)
UAM
TMPRTR
AIRASCI
converts binary
to ASCn file
ASCII version
of UAM
input file
TEMPERATUR.ASC
(file to be transferred)
FIGURE 3-1. UAM input file preparation process.
EEE90008
29
-------
Initial and Chemical Simulation
Meteorology Emissions Boundary Conditions Reaction Rates Control
DEFFBREAKf
REGIONTOP (
EMISSIONS
TERRAIN
(optional)
WIND
1PTSOURCE
(optional)
\
i
AiRQUALrrYf (CHEMPARAM( {SIMCONTROIJ
EMPERATURf
(optional) V
ETSCALARsf
k \
\
t i
f BOUNDARY f
f TOPCONC [
r \
1
UA
Prog)
r \
f
M
ation
•am
r 1
r
INSTANT* AVERAGE DEPOSITION
Execution
Trace
* Can be used as initial condition file to restart model (replaces AIRQU ALTTY).
FIGURE 3-2. Urban Airshed Model simulation program with input and output files.
EEE90008
30
-------
DIFFBREAK
EMISSIONS
METSCALARS
PTSOURCE
REGIONTOP
SIMCONTROL
TEMPERATUR
TERRAIN
TOPCONC
WIND
This file specifies the daytime mixing height and nighttime inver-
sion height for each column of cells at the beginning and end of
each hour of the simulation.
This file specifies the ground-level emissions of all emitted species
to be simulated for each grid cell and each hour of the simulation.
These species will usually be a subset of those listed in Table 2-1,
although any additional species not recognized by UAM will simply
be ignored by the UAM.
This file contains hourly values of meteorological parameters that
do not vary spatially. These scalars are the NC>2 photolysis rate
constant, the concentration of water vapor, the temperature
gradient above and below the diffusion break, the atmospheric pres-
sure, and the exposure class (a measure of the near ground-level
atmospheric stability due to surface heating or cooling).
This file contains point source information, including the stack
height, temperature, flow rate, the plume rise (effective stack
height), the grid cell that contains the stack, and emission rates for
all emitted CB-IV species for each point source for each hour of the
simulation.
This file specifies the height above ground of the top of the model-
ing region. It can vary both spatially and temporally, but the usual
practice is to set it to a constant.
This file contains simulation control information, including the
period of simulation, model options, and information on integration
time steps.
This file specifies the surface temperature for each hour and grid
cell.
This file contains values for surface roughness and deposition factor
for each grid cell.
This file specifies the concentration of each CB-IV state species
(Table 2-1) for the area above the top of the modeling domain.
This file specifies the x- and y-direction components of wind
velocity for every grid cell for each hour of the simulation. Also
contained in this file are the maximum wind speeds for the entire
domain and the average wind speed at each boundary for each hour
of the simulation.
90008
31
-------
All UAM input and output files have a similar overall file structure. The files are
divided into three types of records: header records, time invariant records, and time
varying records. Below we define each data item in terms of the record and word
number (words are 32 bits for most computer systems) in which the data item is con-
tained, the data value type (alphanumeric, integer, or real), and a description of the
data and units.
3.2.1 Header Records
Header Records contain up to four lines:
(1) The File Description Header Record contains 76 words:
1-10 A File name; must be 10 characters, upper-case, one character per
word (i.e., one character in each of 10 words rather than one
10-character variable) exactly matching one of the 13 UAM input
file names.
11-70 A File identifier; must be 60 characters, one character per word.
71 I Number of segments. The UAM is always run with one segment,
thus this number is always 1.
72 I Number of chemical species; usually set to 23 (see Table 2-2).
73 I Beginning date of file (Julian date in YYDDD, e.g., 3uly 13 1979 is
79195).
Ik R Beginning hour of file (2^-hour system, e.g., 1 pm is 13.0).
75 I Ending date of file (Julian).
76 R Ending hour of file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words refer to the reference origin:
1 R x-coordinate (meters). For most applications this value will
always be 0.
90008 k
32
-------
2 R y-coordinate (meters). For most applications this value will
always be 0.
3 I UTM zone.
The next two words define the location of the origin of the modeling domain
with respect to the reference origin. When the data values in words 1 and 2
above (the reference origin) are set to zero, the next two words, along with the
UTM zone in word 3, would define the origin of the modeling domain in terms
of UTM coordinates. (Note: Some UAM applications have used UAM modeling
domains that are tilted with respect to the UTM coordinate system (see for
example Whitten et al., 1985). Under this condition words 1 through 5 are all
set to zero and all station coordinates used in the UAM preprocessors (see
Volume II) will be specified in terms of meters from the origin of the modeling
domain.
4 R x-location (meters). This value will usually be the UTM easting
coordinate for the origin of the modeling domain.
5 R y-location (meters). This value will usually be the UTM northing
coordinate for the origin of the modeling domain.
The next two words define the grid cell size in the x- and y-directions, respec-
tively:
6 R Grid cell size in the x-direction (meters).
7 R Grid cell size in the y-direction (meters).
The next three words define the number of grid cells in the x- and y-directions
and the number of vertical layers:
8 I Number of grid cells in the x-direction.
9 I Number of grid cells in the y-direction.
10 I Number of grid cells in the z-direction.
The final five words describe the vertical distribution of the grid cells:
11 I Number of vertical cells (layers) between the ground and the diffu-
sion break.
12 I Number of vertical cells (layers) between the diffusion break and
the top of the modeling domain.
13 R Height of the surface layer (meters). Should be set to zero.
(Surface layer was used in the former UAM input file ROADWAY
and is no longer used.)
90008 »f
33
-------
14 R Minimum thickness of vertical grid cells (layers) for layers
between the ground and the diffusion break.
15 R Minimum thickness of vertical grid cells (layers) for layers
between the diffusion break and the region top.
(3) The Segment Description Header Record contains one group of four words for
each segment. The number of segments appears in the File Description Header
Record. The current version of the UAM does not support segmentation of a
simulation, thus only one segment is defined and it must correspond to the
entire modeling domain.
1 I x-location of segment origin with respect to origin of modeling
region (grid units). Must be 0.
2 I y-location of segment origin with respect to origin of modeling
region (grid units). Must be 0.
3 I Number of grid cells in segment in the x-direction. Must equal
number of grid cells in modeling domain.
4 I Number of grid cells in segment in the y-direction. Must equal
number of grid cells in modeling domain.
(4) The Species Description Header Record contains 10 words for each species (the
number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word, left-justi-
fied.
Information on the Region Description and Segment Description Header Records
must be the same for all UAM input files. Not all Description Header Records are
used in all files; for some files some of the Region, Segment, or Species Description
Header Records are omitted.
3.2.2 Time-Invariant Data
This category contains information for data that do not vary with time. The types of
data vary from file to file.
90008 •»
34
-------
3.2.3 Time-Varying Data
This category contains information for data that do vary with time. The types of
data vary from file to file.
In the remainder of this chapter we define the contents and formats of each of the
13 UAM input files.
90008 4
35
-------
. .ii i \jcwni_ i i i
3.3 AIRQUALITY
The AIRQUALITY file contains observed concentration values as a function of time
for each species over the region (x,y,z).
Header Records
The AIRQUALITY file begins with the four standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'AIRQUALITY1; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
7k R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90006
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 1 Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 I Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
each segment (the number of segments appears in the File Description Header
Record; for the CB-IV version of UAM, only one segment is allowed):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
^ I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record contains 10 words for each species
(the number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
Time-Invariant Data
The AIRQUALITY file contains no time-invariant data.
90006 H
38
-------
iVXUML.1 I I
Time-Varying Data
The AIRQUALITY file contains one set of the following records for each time inter-
val.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) For the one segment of the region the AIRQUALITY file contains a set of
Concentration Records for each species, and ordered within each species by
vertical level. The first 11 words of the record identify the segment and
species:
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character, per word.
The next series of words is the concentration array itself:
12+ R Concentrations (ppm, or yg/m for aerosols) for each cell in one
vertical level, varying by x-, then y-direction.
90008 »*
39
-------
1 I
3.4 BOUNDARY
The BOUNDARY file defines the external boundaries and internal segment interfaces
of the modeling region and the concentrations for each species at each external
boundary cell over the region (x,y,z).
Header Records
The BOUNDARY File begins with the four standard header records.
(1) The File Description Header Record contains 76 words.
1-10 A File name = 'BOUNDARY '; 10 characters, one character per
word.
11 -70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 H ,,
41
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I -Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments must be 1 in the File Description
Header Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-iocation of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
^ I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record contains 10 words for each species
(the number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
Time-Invariant Data
The BOUNDARY file contains defines the boundaries of the region, both the external
boundaries defined in the data preparation and the internal segment interfaces
created when the region is segmented.
42
90008 >f
-------
For each segment there are four Boundary Definition Records, one record for each
edge. The edges are defined as follows:
1 Left West lower limit column index for each row
2 Right East upper limit column index for each row
3 Bottom South lower limit row index for each column
4 Top North upper limit row index for each column
Each Boundary Definition Record defines the location of the boundary cells at an
edge of a segment. The first three words identify the edge and its dimensions:
1 I Segment number (must be 1).
2 I Edge number.
3 I Number of cells on edge (i.e., number of rows or columns).
The next four words define the boundary location for the grid index (row or column)
along each edge:
4 I Index, within the segment, of the cell at the edge of the region
modeled (i.e., the first or last cell simulated within the row or
column). If this number is 0, this row or column is to be omitted
from the simulation and the next three numbers are ignored.
5 I Segment number in which adjacent cell is located. If this number
is 0, the boundary is an external one, and the next two numbers are
ignored.
6 I x-index of adjacent cell within segment defined in word 5.
7 I y-index of adjacent cell within segment defined in word 5.
Time-Varying Data
The BOUNDARY file contains one set of the following records for each time inter-
val.
90008
43
-------
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
k R Ending time (hours).
(2) For each segment of the region, there is a set of Boundary Concentration
Records, grouped by species, with four records (one for each edge) for each
species. The first 12 words of the record identify the segment, species, and
edge:
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character per word.
12 I Edge number.
The next series of words is the boundary concentration array on the vertical
plane along the edge:
13+ R Boundary concentrations (ppm, or ug/m for aerosols) at each
vertical level for each cell along the edge. For rows or columns
that are not to be simulated and for edges that represent internal
segment boundaries, these numbers must be present but will be
ignored.
44
90008 >*
-------
UHtMKAKAM
3.5 CHEMPARAM
The CHEMPARAM. file specifies chemical species characteristics, reaction proper-
ties, and stoichiometric coefficients.
Header Records
The CHEMPARAM file begins with two standard header records.
(1) The File Description Header Record contains 72 words:
i-10 A File name ='CHEMPARAM '; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments. For the CHEMPARAM file, which does not
have spatially varying data, this number is 0.
72 I Number of chemical species.
The next four words describe the total time span contained on the file. Since
data in the CHEMPARAM file do not vary in time, these values are not used.
73 I Beginning date of the file (Julian).
7^ R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record is omitted for the CHEMPARAM file.
(3) The Segment Description Header Record is omitted for the CHEMPARAM
file, which does not have spatially varying data.
(4) The Species Description Header Record contains 10 words for each species;
(the number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
90008
45
-------
Time-Invariant Data
The CHEMPARAM File contains two to four time-invariant data records.
(1) The Chemistry Definition Record contains 62 words:
1-60 A Chemical kinetics mechanism identifier (e.g., 'Carbon Bond IV);
60 characters, one character per word.
61 I Number of reactions.
62 I Number of stoichiometric coefficients.
(2) The Species Parameters Record contains one group of 18 words for each
species (the number of species is defined in the File Definition Header
Record):
1-10 A Species name; 10 characters, one character per word.
11 L Reactive/unreactive flag (.TRUE. = reactive, .FALSE. =
unreactive).
12 L Steady state for initial concentrations (.TRUE. = compute steady-
state values, .FALSE. = use input values).
13 L Steady state for boundary concentrations (.TRUE. = compute
steady-state values, .FALSE. = use input values).
14 R Resistance to surface sinks (h/m).
15 R Lower bound for steady-state calculations (ppm).
16 R Upper bound for steady-state calculations (ppm).
17 R Lower bound for numerical integration (ppm, or yg/m for
aerosols).
18 R Upper bound for numerical integration (ppm, or yg/m for
aerosols).
(3) The Reaction Parameters Record contains one group of five words for each
reaction; the number of reactions is specified in the Chemistry Definition
Record (if the number of reactions is 0, this record is omitted):
1 R Rate constant.
2 L Photolysis flag (.TRUE. = photolysis reaction, .FALSE. = not a
photolysis reaction).
3 L Temperature-dependence flag (.TRUE. = temperature-dependent
reaction, .FALSE. = not a temperature-dependent reaction). If
.FALSE., the next two words will be ignored.
4 R Activation energy (K).
5 R Reference temperature for activation energy (K).
46
90008 i*
-------
CHEMPARAM
The Stoichiometric Coefficients Record contains one group of 11 words for
each stoichiometric coefficient; the number of these coefficients is specified
in the Chemistry Definition Record (if the number of coefficients is 0, this
record is omitted):
1-10 A Coefficient name; 10 characters, one character per word.
11 R Coefficient value.
Time-Varying Data
The CHEMPARAM File contains no time-varying data.
47
90008 ^
-------
DIFFBREAK
3.6 DIFFBREAK
The DIFFBREAK file specifies the height of the change of diffusion characteristics
(the "diffusion break") over the region (x,y).
Header Records
The DIFFBREAK File begins with three standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'DIFFBREAK '; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be i.
72 I Number of chemical species. Since the diffusion break does not
vary with species, this number is 0.
The next four words describe the total time span contained on the file.
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 i* 49
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since the
diffusion break does not vary vertically, these values are not used:
11 I Number of cells between surface layers and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
each segment (the number of segments appears in the File Description Header
Record; in the CB-IV version of UAM, only one segment is allowed):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
*
(4) The Species Description Header Record is omitted for the DIFFBREAK File,
since the diffusion break does not vary with species.
Time-Invariant Data
The DIFFBREAK File contains no time-invariant data.
90006 if
50
-------
Time-Varying Data
The DIFFBREAK File contains one set of the following records for each time inter-
val.
(1) The Time Interval Record contains four words:
1 I Beginning data (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) For the one segment of the region there is one Diffusion Break Record. The
first 11 words of the record identify segment and variable:
1 I Segment number (must be 1).
2-11 A Variable name = 'DIFFBREAK1; 10 characters, one character per
word
The next series of words is the diffusion break array itself:
12+ R Diffusion break (meters) at the beginning of the time interval at
each cell in the row for each row in the segment.
90006 «*
51
-------
tMISblUNS
3.7 EMISSIONS
The EMISSIONS file contains emissions values for each species for each grid cell
within the modeling domain (x,y). This file contains low-level anthropogenic and
biogenic emissions. The anthropogenic emissions may consist of mobile, area, and
minor point sources with small effective stack heights. Major point source emissions
with large effective stack heights are contained in the PTSOURCE file (Section 3.9).
Header Records
The EMISSIONS File begins with the four standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'EMISSIONS '; 10 characters, one character per word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 >* 53
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since data
in the EMISSIONS file do not vary vertically (all emissions are in the first
layer), these values are not used.
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
each segment (the number of segments appears in the File Description Header
Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
k I Number of grid cells in segment, y-directiori.
(4) The Species Description Header Record contains 10 words for each species (the
number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
90008 <*
54
-------
Time-Invariant Data
The EMISSIONS file contains no time-invariant data.
Time-Varying Data
The EMISSIONS file contains one set of the following two records for each time
interval.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) For the one segment of the region there is one Emissions Record for each
species. The first 11 words of the record identify segment and species:
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character per word.
The next series of words is the emissions array itself:
12+ R Ground-level emissions (gram-moles per hour, or grams per hour
for aerosols).
90008 t
55
-------
MtlbUALARS
3.8 METSCALARS
The METSCALARS file contains values for meteorological parameters that are
assumed to be constant over the entire region.
Header Records
The METSCALARS file begins with two standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'METSCALARS'; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments. Since the data in the METSCALARS file'do
not vary spatially, this number is 0.
72 I Number of chemical species. Since the data in the METSCALARS
file do not vary with species, this number is 0.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 4 "
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since the
METSCALARS file does not contain vertically varying data, these values are
not used:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record is omitted for the METSCALARS file,
which does not contain spatially varying data.
(4) The Species Description Header Record is omitted for the METSCALARS file,
which does not contain data that vary with species.
Time-Invariant Data
The METSCALARS file contains no time-invariant data.
Time-Varying Data
The METSCALARS file contains two records for each time interval.
90008 4
58
-------
MtlSUALAKS
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date Oulian).
4 R Ending time (hours).
(2) The Meteorological Scalars Record contains 66 words, or six name-number
pairs:
1-10 A Name = TGRADBELOW.
11 R Temperature gradient (K/m) below the diffusion break.
12-21 A Name = TGRADABOVE1; 10 characters, one character per word.
22 R Temperature gradient (K/m) above the diffusion break.
23-32 A Name = 'EXPCLASS '; 10 characters, one character per word.
33 R Exposure class.
34-43 A Name = 'RADFACTOR1; 10 characters, one character per word.
44 R Radiation factor at end of time interval.
45-54 A Name = 'CONCWATER1; 10 characters, one character per word.
55 R Concentration of r^O (ppm).
56-65 A Name = 'ATMOSPRESS1; 10 characters, one character per word.
66 R Atmospheric pressure (atm).
90008 H 59
-------
3.9 PTSOURCE
The PTSOURCE file contains the location of each point source and the total mass
emission for each species. Major point sources with large effective stack heights are
specified in this file. Minor point sources may be specified in the EMISSIONS file
(Section 3.7).
Header Records
The PTSOURCE file begins with the four standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'PTSOURCE '; 10 characters, one character per word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
The next four words describe the total time span contained on the file.
73 I Beginning date of the file Oulian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin: " •
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-iocation (meters).
90008 * 61
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
each segment (the number of segments appears in the File Description Header
Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
b I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record contains 10 words for each species (the
number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
90008 i+
62
-------
Time-Invariant Data
The PTSOURCE file contains the location and other fixed properties of each point
source. For the one segment there are two records.
(1) The Counter Record contains two words:
1 I Segment number (must be 1).
2 I Number of point sources in segment.
(2) The Point Source Definition Record contains the following group of six words
for each point source in the segment. If there are no point sources in the seg-
ment this record does not appear:
1 R x-coordinate of point source with respect to reference origin
(meters).
2 R y-coordinate of point source with respect to reference origin
(meters).
3 I x-index of grid cell within the segment.
i* I y-index of grid cell within the segment.
5 R Stack height (meters).
6 R Stack diameter (meters).
Time-Varying Data
The PTSOURCE file contains one set of the following records for each time interval.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
-------
i I Segment number (must be i).
2 I Number of point sources in the segment for this time interval. If
the number of point sources defined in the Counter Record for
time-invariant data is greater than zero, this number must also be
greater than zero.
The Point Source Location Record contains the following group of five words
for each point source in the segment. If there are no point sources in the seg-
ment, this record does not appear:
1 I x-index within segment of cell to receive emissions.
2 I y-index within segment of cell to receive emissions.
3 I z-index of cell to receive emissions.
4 R Flow rate (m3/h).
5 R Effective plume height (meters).
The Point Source Emissions Record contains the following group of words:
i I Segment number.
2-11 A Species name; 10 characters, one character per word.
12+ R Emissions (gram-moles per hour, or grams per hour for aerosols)
from each point source.
90008 t
64
-------
3.10 REGIONTOP
The REGIONTOP file specifies the height of the top of the region (x,y).
Header Records
The REGIONTOP file begins with three standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'REGIONTOP1; 10 characters, one character per word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species. Since the REGIONTOP file contains
no time-varying data, this number is 0.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
Ik R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 t
65
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since the
REGIONTOP file has no data that vary vertically, these values are not used:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the segment (the number of segments must be 1 in the File Description Header
Record):
1 I
2 I
x-location of segment origin with respect to origin of modeling
region (grid units).
y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
The Species Description Header Record is omitted for the REGIONTOP File,
since the data do not vary with species.
Time-Invariant Data
The REGIONTOP file contains no time-invariant data.
90008 V
66
-------
REGIONTOP
Time-Varying Data
The REGIONTOP file contains one set of the following records for each time inter-
val.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) The Top of Region Record contains 12 words:
1 I Segment number (must be 1).
2-11 A Variable name = 'REGIONTOP '; 10 characters, one character per
word.
The next series of words is the top of region array itself:
12+ R Top of region (meters) at the beginning of the time interval at
each cell in the row for each row in the segment. •
90008 i*
67
-------
SIMCONTROL
3.ii SIMCONTROL
The SIMCONTROL file contains calculation options, integration parameters, output
controls, and other parameters required by the simulation program.
Header Records
The SIMCONTROL file begins with one standard header record.
(1) The File Description Header Record contains 76 words:
1-10 A File name = 'SIMCONTROL1; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments. Since data in the SIMCONTROL file do not
vary spatially, this number is 0.
72 I Number of chemical species. Since data in the SIMCONTROL file
do not vary with species, this number is 0.
The next four words describe the total time span contained on the file. Since
the SIMCONTROL file has no time-varying data, these values are not used.
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record is omitted for the SIMCONTROL file,
which is independent of region.
(3) The Segment Description Header Record is omitted for the SIMCONTROL file,
since the file has no spatially varying data.
(4) The Species Description Header Record is omitted for the SIMCONTROL file,
since the file has no data that vary with species.
90008 »»
69
-------
Time-Invariant Data
The SIMCONTROL file contains one Simulation Controls Record that includes all
simulation control variables.
The Simulation Controls Record contains 87 words:
1-60 A Run identifier; 60 characters, one character per word.
61 I Beginning date of the simulation (Julian).
62 R Beginning time of the simulation (hours).
63 I Ending date of the simulation (Julian).
64 R Ending time of the simulation (hours).
65 L Restart flag.
66 L Surface sink flag.
67 L Point source flag.
68 L ROADWAY file flag (no longer used).
69 L TEMPERATUR file flag.
70 L TERRAIN file flag.
71 L Concentration variation flag.
72 R Default surface roughness (meters).
73 R Default vegetation factor (fraction of alfalfa).
7k R Maximum time slice size (hours).
75 I Maximum number of steps in a slice.
76 R Minimum step size (hours).
77 I Maximum number of iterations in chemistry.
78 R Relative error tolerance for convergence in chemistry.
79 R Darkness criterion (same units as RADFACTOR on the
METSCALARS file).
80 R Time interval for instantaneous concentrations output (hours).
81 R Time interval for averaging concentrations (hours).
82 I File history print flag.
83 .1 Core allocation print flag.
84 I Number of levels for instantaneous concentration print.
85 I Number of levels for average concentration print.
86 I Concentration variation print flag.
87 I Print option not implemented.
90008 V
70
-------
mm m • m
Time-Varying Data
The SIMCONTROL file contains no time-varying information.
90008 it
71
-------
ItMKtKAlUH
3.12 TEMPERATUR
The TEMPERATUR file contains ground-level temperatures over the region (x,y).
Header Records
The TEMPERATUR file begins with three standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = TEMPERATUR1; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be i.
72 I Number of chemical species. Since data in the TEMPERATUR file
do not vary with species, this number is 0.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
Ik R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 **
73
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:'
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since data
in the TEMPERATUR file do not vary vertically, these values are not used:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments must be 1 in the File Description
Header Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record is omitted for the TEMPERATUR file,
since no data in the file vary with species.
Time-Invariant Data
The TEMPERATUR file contains no time-invariant data.
90008 k
74
-------
ItMKtKAlUK
Time-Varying Data
The TEMPERATUR file contains one set of the following records for each time
interval.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) The Temperature Record contains 12 words:
1 I Segment number (must be 1).
2-11 A Variable name = TEMPERATUR1; 10 characters, one character per
word.
The next series of words is the temperature array itself:
12+ R Temperature (K) at ground level at each cell in the row for each
row in the segment.
90008 1+
75
-------
I tKKAIN
3.13 TERRAIN
The TERRAIN file contains surface roughness values and vegetation factors over the
region (x,y).
Header Records
The TERRAIN file begins with three standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = TERRAIN1; 10 characters, one character per word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be i.
72 I Number of chemical species. Since data in the TERRAIN file do
not vary with species, this number is 0.
The next four words describe the total time span continued on the file. Since
data in the TERRAIN file do not vary in time, these values are not used.
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
k R x-location (meters).
5 R y-location (meters).
90008 4
77
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters)
7 R Grid cell size, y-direction (meters)
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since the
data in the TERRAIN file do not vary vertically, these values are not used:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments must be i in the File Description
Header Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-iocation of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
(it) The Species Description Header Record is omitted for the TERRAIN file, since
data in the file do not vary with species.
Time-Invariant Data
The TERRAIN file contains two records for the one segment of the region.
90008 **
78
-------
I tHHMlIM
(1) The Surface Roughness Record begins with 11 words that identify the segment
and variable:
1 I Segment number (must be 1).
2-11 A Variable name = 'ROUGHNESS'; 10 characters, one character per
word.
The next series of words is the surface roughness array itself:
12+ R Surface roughness (meters) over the segment.
(2) The Vegetation Factor Record begins with 11 words that identify the segment
and variable:
1 I Segment number (must be 1).
2-11 A Variable name = 'VEGFACTOR1; 10 characters, one character per
word.
The next series of words is the vegetation factor array itself:
12+ R Vegetation factors over the segment (factors are based on fraction
of alfalfa gross absorption in each grid cell).
Time-Varying Data
The TERRAIN file contains no time-varying data.
90008 H
79
-------
IUKUUNU
3.14 TOPCONC
The TOPCONC file specifies the concentration values for each species at the top
boundary of the region (x,y).
Header Records
The TOPCONC file begins with the four standard header records.
(1) The File Description Header Record contains 76 words:
1-10 A File name = TOPCONC1; 10 characters, one character per word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R x-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 k
81
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells. Since data
in the TOPCONC file do not vary vertically, these values are not used:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments must be 1 in the File Description
Header Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record contains 10 words for each species (the
number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
90006 it
82
-------
Time-Invariant Data
The TOPCONC file contains no time-invariant data.
Time-Varying Data
The TOPCONC file contains one set of the following records for each time interval.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) The Top Concentration Record contains 12 words for each species within the
one segment of the region. The first 11 words of the record identify the seg-
ment and species:
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character per word.
The next series of words is the concentration array itself:
12+ R Concentration at top of region (ppm, or pg/m for aerosols) at each
cell in the row for each row in the segment.
90008 4
83
-------
WIND
3.15 WIND
The WIND file contains the x- and y-components of the wind vector at each cell in
the region (x,y,z).
Header Records
The WIND file begins with three standard header records.
(1) The File Description Header Record contains 76 words:
i-10 A File name = 'WIND '; 10 characters, one character per word.
11-70 A File identifier, 60 characters, one character per word.
7i I Number of segments: must be 1.
72 I Number of chemical species. Since data in the WIND file do not
vary with species, this number is 0.
The next four words describe the total time span contained on the file:
73 I Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeiing region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
90008 H
85
-------
The next two words define the size of each grid cell in the x- and y-directions:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in grid cells:
8 I Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells:
11 I Number of cells between surface layer and diffusion break.
12 I Number of cells between diffusion break and top of region.
13 R Height of surface layer (meters).
14 R Minimum height of cells between surface layer and diffusion break
(meters).
15 R Minimum height of cells between diffusion break and top of region
(meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments must be 1 in the File Description
Header Record):
1 I x-location of segment origin with respect to origin of modeling
region (grid units).
2 I y-location of segment origin with respect to origin of modeling
region (grid units).
3 I Number of grid cells in segment, x-direction.
4 I Number of grid cells in segment, y-direction.
(4) The Species Description Header Record is omitted for the WIND file, since
data in the file do not vary with species.
Time-Invariant Data
The WIND file contains no time-invariant data.
9000S *t
86
-------
WIND
Time-Varying Data
The WIND file contains one set of the following records for each time interval.
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) There is a set of three records for each vertical level; one for scalar values,
one containing the x- component of wind, and one containing the y-component.
The Wind Scalars Record contains eight words:
1 I Segment number (must be 1).
2 R Reference height of wind measuring stations (meters).
3 R Maximum absolute value of wind speed in x-direction (m/h).
4 R Maximum absolute value of wind speed in y-direction (m/h).
5 R Average wind speed in x-direction at west boundary (m/h).
6 R Average wind speed in x-direction at east boundary (m/h).
7 R Average wind speed in y-direction at south boundary (m/h).
8 R Average wind speed in y-direction at north boundary (m/h).
In the x-wind record the first 11 words identify the segment number and vari-
able name:
i I Segment number.
2-11 A Variable name ='WINDX '; 10 characters, one character per
word.
The next series of words is the X-wind array itself:
12+ R Wind in x-direction over the segment for one vertical level (m/h).
90008 4
87
-------
In the y-wind record the first 11 words identify the segment number and vari-
able name:
1 I Segment number.
2-11 A Variable name ='WINDY '; 10 characters, one character per
word.
The next series of words is the Y-wind array itself:
12+ R Wind in y-direction over the segment for one vertical level (m/h).
90008 ^
88
-------
MODEL OUTPUTS
The UAM generates up to four output files. Two unformatted (binary) files include
the time-averaged concentrations (AVERAGE) and instantaneous concentrations
(INSTANT) for all species in each grid cell. A formatted file includes the amounts of
deposited species, and a formatted file (simulation output) contains information
concerning execution trace. The data definitions and formats for these files are
described in this chapter.
4.1 SIMULATION OUTPUT
The UAM standard output file (execution trace) file is a formatted file that contains
information concerning which files were opened and closed, a summary of the con-
tents in the input files, the time steps used, the current status of the simulation,
mass balance summary, and any error messages. When the UAM stops prematurely,
this file along with the job control history file should be examined to determine the
reason. If the job terminates due to a system problem (e.g., misspelling of a file
name in the job control file), the job control history file should contain a system
error message. However, if the UAM terminates because of faulty input specifica-
tions (e.g., a wrong date in one of the input files), the standard output file will have
an error message. The information contained in this file is self-explanatory; Chapter
9 of Volume II of this user's guide contains an example.
4.2 AVERAGE
The AVERAGE file contains time-averaged concentrations for each species (see
Table 2-1) in each grid cell for the entire modeling region. The contents of the
unformatted AVERAGE file follows the structure for the UAM input and output files
given in Section 3.2 and is described below.
90008 5
89
-------
Header Records
The AVERAGE file begins with four standard header records.
(1) The File Description Header Record contains 76 words:
i-10 .A File name = 'AVERAGE '; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
73 R Beginning date of the file (Julian).
74 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
Words 73-76 describe the total time span contained on the file.
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R x-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
The next two words define the size of each grid cell:
6 R Grid cell size in the x-direction (meters).
7 R Grid cell size in the y-direction (meters).
The next three words define the size of the modeling region in terms of the
number of grid cells:
90008 5
90
-------
8 I Number of grid cells in the x-direction.
9 I Number of grid cells in the y-direction.
10 I Number of grid cells in the z-direction.
The last five words describe the vertical distribution of grid cells:
11 I Number of cells between the surface layer and diffusion break.
12 I Number of cells between the diffusion break and top of region.
13 R Height of surface layer (meters).
l*f R Minimum height of cells between the surface layer and diffusion
break (meters).
15 R Minimum height of cells between the diffusion break and top of
region (meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments is 1 in the File Description Header
Record):
1 I x-location of the segment origin with respect to origin of model-
ing region (grid units).
2 I y-location of the segment origin with respect to origin of model-
ing region (grid units).
3 I Number of grid cells in the segment, x-direction.
4 I Number of grid cells in the segment, y-direction.
(4) The Species Description Header Record contains 10 words for each species (the
number of species is defined in the File Description Header Record):
1-10 A Species name; 10 characters, one character per word.
Time-Invariant Data
The AVERAGE file contains no time-invariant data.
Time-Varying Data
The AVERAGE file contains one set of the following records for each time interval.
90008 5
91
-------
(1) The Time Interval Record contains four words:
1 I Beginning date (Julian).
2 R Beginning time (hours).
3 I Ending date (Julian).
4 R Ending time (hours).
(2) There is a set of Average Concentration Records for the one segment of the
region for each species, ordered within each species by vertical level. The first
11 words of the record identify segment and species
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character per word.
The next series of words is the concentration array itself:
o
12+ R Concentrations (ppm, or yg/m for aerosols) averaged over the
time interval for each cell in one vertical level. Values are
ordered by x and y location of the cells as follows:
«C(I,J),I=1,NX),J=1,NY)
where I specifies the x-cell and J specifies the y-cell.
4.3 INSTANT
The INSTANT file contains calculated instantaneous concentrations for each species
(Table 2-1) in each grid cell for the entire modeling region. The main purpose of this
file is to be able to restart the model when performing multi-day simulations or when
the model is terminated in the middle of a simulation because of insufficient disk
space or computer system failure. The INSTANT file follows the structure given in
Section 3.2 and is described below.
Header Records
The INSTANT file begins with four standard header records.
90008 5
92
-------
(i) The File Description Header Record contains 76 words:
1-10 A File name " 'AIRQUALITY1; 10 characters, one character per
word.
11-70 A File identifier; 60 characters, one character per word.
71 I Number of segments; must be 1.
72 I Number of chemical species.
73 I Beginning date of the file (Julian).
7>4 R Beginning time of the file (hours).
75 I Ending date of the file (Julian).
76 R Ending time of the file (hours).
Words 73-76 describe the total time span contained on the file.
(2) The Region Description Header Record contains 15 words. The first three
words define the reference origin:
1 R x-coordinate (UTM units).
2 R y-coordinate (UTM units).
3 I UTM zone.
The next two words define the location of the modeling region with respect to
the reference origin:
4 R x-location (meters).
5 R y-location (meters).
The next two words define the size of each grid cell:
6 R Grid cell size, x-direction (meters).
7 R Grid cell size, y-direction (meters).
The next three words define the size of the modeling region in terms of the
number of grid cells:
8 1 Number of grid cells, x-direction.
9 I Number of grid cells, y-direction.
10 I Number of grid cells, z-direction.
The last five words describe the vertical distribution of grid cells:
90008 5
93
-------
11 I Number of cells between the surface layer and diffusion break.
12 I Number of cells between the diffusion break and top of region.
13 R Height of surface layer (meters).
R R Minimum height of cells between the surface layer and diffusion
break (meters).
15 R Minimum height of cells between the diffusion break and top of
region (meters).
(3) The Segment Description Header Record contains one group of four words for
the one segment (the number of segments is 1 in the file Description Header
Record):
1 I x-location of the segment origin with respect to origin of model-
ing region (grid units).
2 I y-location of the segment origin with respsect to origin of
modeling region (grid units).
3 I Number of grid cells in the segment, x-direction.
4 I Number of grid cells in the segment, y-direction.
(
-------
(2) There is a set of Instantaneous Concentration Records for the one segment of
the region for each species, ordered within each species by vertical level. The
first 11 words of the record identify the segment and species:
1 I Segment number (must be 1).
2-11 A Species name; 10 characters, one character per word.
The next series of words is the concentration array itself:
12+ R Concentrations (ppm, or yg/m for aerosols) at the beginning of
the time interval for each cell in one vertical level. Values are
ordered by x and y location of the cells as follows:
((C(I,J),I=l,NX),J=i,NY)
where I specifies the x-cell and 3 specifies the y-cell.
4.4 DEPOSITION
The DEPOSITION file contains the amount of each deposited species on the ground in
each horizontal grid cell in the modeling region for the total time interval of the
simulation. The structure of this formatted file is described below.
The Time Interval Record contains four values in the format (110, F10.2, 110, F10.2):
1 1 Ending date of simulation (Julian).
2 R Ending time of simulation (hours).
"}> I Ending date of simulation (Julian, redundant).
4 R Ending time of simulation (hours, redundant).
The next line contains the first species name in the format (10A1).
1-10 A Species name; 10 characters.
Next is the array of depositions values in the format (10E12.6).
1+ R Total deposition of species named above (moles) for the span of
the simulation ordered by x and y location of the cells as
follows:
90008 5
95
-------
((C(I,3),I=1,NX),3=1,NY)
where I specifies the x-cell and J specifies the y-cell
The species name and deposition records then repeat for the remainder of the simu-
lated species.
90008 5
96
-------
5 COMPUTER USER NOTES
In order to run the UAM on a particular computer system for a particular
application, some modifications to the code may be necessary to increase model
storage arrays when simulating a domain requiring arrays larger than are presently
specified. Modifications must be made to various arrays contained in the model (see
Section 5.2).
5.1 ADAPTATION OF FILE HANDLING FOR DIFFERENT COMPUTER SYSTEMS
The UAM is currently configured to accept file names and units from standard
input. The UAM opens all files internally using the FORTRAN 77 OPEN statement,
where the file names and units are supplied on the standard input. Some computer
systems, such as the EPA IBM/3090, require that all files be opened externally. Thus
in the following paragraphs we describe how to set up a UAM job control file using
IBM Job Control Language (JCL) and a UNIX-based operating system.
5.1.1 IBM JCL
Exhibit 5-1 shows an example IBM job control file written in IBM JCL for the two-
day Atlanta simulation. The first section of the IBM JCL is used to delete any
existing file with the same names as those that will be opened and used for the
current simulation. If these output files exist when the UAM starts running, the
model run will abort. The second section of the IBM JCL exercises the UAM
simulation preprocessor (SPREP). The final section of the IBM JCL runs the UAM
model for the first day (Exhibit 5-la) and second day (Exhibit 5-lb) of the
simulation. Note that all input and output files for running the UAM on the IBM
mainframe are opened externally.
90008 5
97
-------
o
o
o
o
o
o
o
[Jt,
CO
g
£
M
o
z
•
o
§
m
n
£
H
II
>•.
£
~
i
s? 3
(N <">
O Q «o
m J ii
-O en
XXX.
x -2
^ z u.
ffl 3
o a, g
^ "C
//XXXY
/* ROUTE
//*JOBP
* * •
5
at
H
en
z
tt
Q
U
X
en
a.
M
* * *
v\\\
•v\\\-
H
I
p-t
CO
u
cu
8
N
M
H
to
— » II
H tn
o x
•~r £J
Q tn
Z Cd
Ox • O
U II O >
* Q ^ ^
0 J « Z
(MO Z •
II X • «
a - «r co
CQ »3 oo c*\
Z << c% M
SO M »r
< o • «• n
Z J • T •
» II • • U
a H w u z
• tn cs z o
M Ed Cd O U
0 Q X U H
H - O CO
Eb < CO > Z
Ed II Q < M
SH Z • •
3 < O O
M o z -~t« e-i
II tO X Et, < <
X > O en • -
u tn u * M tn tn
D O Q X5 X
u o tn o M x x
U r- u-U.Ii.
x EIJ en tn
W t~t Oi Vi HH W
X Ed Z O. • -
Ed O. H M
H co z en Ed
J tn to M H c- ti.
Cd X > U. Ed Ed O
Q en * to o J J u * <
Cd
jjj .
t* ^
Cd
Q£
Cb
CO
a; ^«
Q tn D co
M || « *
• a, E< • >
M S IX
u d. <
Z Ed Z
O CK en w
M U Ed CD
O • ^ CNJ O > •
Ed CO in Cd O II E"
CS< fM fMM3 . Q<
«k « « IX J •
MM z ZOOH oco
< < 33 X <
U U. Z Z II El. X
ix to o O 6- tn co
CUM M MOCd OMOOM 3M||
CO II H i-103 .tOM
o a « j j tnao
a < .<
o a > > a a
U Q Q Cd Ed Q Q
Ed
X » »
coo Ocn ocr\in3CO--o- oo
MO KM MMMCClJ OOM OO
CL»JEt«>jH* H^VH< O wwEh
ajEdoeso^J • OIN
8M 3 Z M 3 Z XXXV,
tn 1-3 Ed en l-J Cd-vXXX
*
41
«
*
*
J
«
* -^ • ^
« cj a? 2
« tn Ed a.
« E- CM X
« cc cd a, x
« z x to o
« tn ^ ^- ~
* II v •«• «r
* CU O CO CO
* 1/1 IX to en cf* Cft
* 0> X MM M M
* vo en Q n ^ tn
« n II -<-> 1-3 ^
« u a. — •
« X en o E-I E-i E-
« M M > 3 3 3
<«HQ VOO O O
« - • Z Cd Cd Cd
z«:£o TO. a a
M«O< T a a. a.
< « o CJ ~ •
X«o3 H> > >
« in • 3 a £ a.
D*ll^ &4< < <
M«ZE£ HZ Z Z
tO*Mb.UEdOOB CD EQ
a * o • x > • •
M*EdtOIIII6-6H E- £-
« - . M J • •
Ed*xeooococn en tn
o * a i < < r 2 x x
Cd«OEhlll|EbCv U. Et.
x*tocoE-iE-ieoen to tn
Cd«MM33MM M M
« II II O O II II II II
«xztntnzz z z
«oto>xtnco tn to
^»'p.Qfnrf^^Q(V^vn^/^>iv
* XXX
« uooaaatnaeoatn
« Ed a Q Q Q a n a n a n
* x a. a. a.
* Ed tn to tO
* oaoooooosoo
*UMOOOO O O
*S.EdOOOO O O
««v"xX.vvvv
V\X\\\X\svV.-v\V.X
98
-------
o
>
VD
z
Tj.
00
OS
H
n
cj
Z
8
EH
CO
Z
fH
O
£
<
co
3
b<
CO
II
z
CO
Q K
O CO
Q II
O,
CO
rt HH
0 D
o
o
xx
XX
Ed
1
<
O
rt
b.
E* -
Z *~
w ^*
11
O Q
U Z
-m
Ed z
05 <*•"
w ^
X 00
0\
O rt
b] rt
32
fl. D
b] U
m a
OH
3*
§2
a z
M
Ed 02
b, <
EH CO
||
CO b.
b] CO
II
Z
co
0 K
Q Q CO
O Q II
ft.
rt cn
O 0 D
b. O
rt rt
S-E
xxx
xxx
*
,— ^
CM
X
CO
b.
Q
^-*
^
00
OS
41
1
O
K
a,
i
HI
ffi
EH
<
CO
3
i
co
ii
z
co
Q
Q
Q
0
O
(N
X
X
O
^
VD
x
rt
OQ
CO
Os
rt
|K
~
1
a
a,
z
a
z
M
ffl
co'
3
i
co
n
z
CO
ex Q a
X X
co a co
CO CO
a o a
o
rt
XXX
xxx
o
VD
2;
Ct
co
E
CO
OS
rt
O
,
g
s
a
04
rt
z
1
t-H
co
•
co
b.
CO
II
z
CO
a at
a co
a ii
a.
CO
0 Q
O
f
E
xx.
\x
Q
CO
g
z
^4 -*"•"
J £X
z a,
u z
§1
3 —
O *
-CO
U) OS
«^
f^
Cd rt
X T
Q EH
U 3
a< a.
Ed >
CO J
Q Z
J M
D 03
M •
J CO
M <
*** B
>• b.
J &3
vD CC frQ
Z^ u^
* r^
•vox
00 O J
OS rt ffl
rt » »
rt O vo
1-3 O fsl
• IT) O
o *•*<-
o «7
o 2 J
O EH U
< Ta
• U J
0 U -
• CO >
< < s
i > bi
CO CO K
11 EH II
Z W 03
CO Z CJ
Q 3 Q
Q
a
O
O
rt
EH
b.
X.'X'X
X.V.X,
^>
O b3
> CO
VO J
Z fli
9 /—
co o
OS O
rt rt
1-3 0
. fM
O to
Z *-*
o -
O bt
EH«
W EH
Z ^
/— • W
O O U
< • CO
U*i a"
Z tt V
^— co co
a. n H
CO Z W
M co z
o
a
o
o
(N
^*
EH
b.
X V.X.
\\\
h,
^^^
Z
o o
rt CU
§ 1
II •«?
bl CO
N OS
co rt
J 1
ffl EH
* 3
vD O
CN Cd
0 «
rt Oi
II >
J a
Ed Z
a H,
»o •
> < ^
II U CO
r *£
Ed Z fa
05 >_-CO
ii a n
ffl co z
U H CO
X
a co
Q£
CO
0 Q
O
rt
EH
b.
\\\\
X.X, V.V.
fc
r— _
H
1
a
^i
CO
rt
g
O
bl
EC
EX
1
w
n
co
b.
CO
II
z
co
o a
a co
a ii
a.
co
0 Q
o
b*
EH
bi
\x
\x
^
^^^
M
E
^
CO
OS
rt
H
i
£
1
H^
ffi
CO
3
b.
CO
II
Z
cn
0 K
Q CO
a 11
a,
CO
0 Q
o
b*
IO
EH
Ev,
XX
\\
«.
J— ^
Q
Q
Q
Z
3
^
CO
rt
D
g
as
a,
1
M
ffl
co
b.
CO
II
co
O K
o co
a ii
a.
CO
O Q
O
b«
EH
b,
x\
\X
X
CO
II
a.
co
Q
O
VO
z
rt
1-3
M
b4
H
D
0
H
CO
b,
CO
II
z
co
Q
Q
O
O
o
b«
E.
X"X
•x\
^
$
z)
1— 1
Si
o
(Tj
I
in
JJ
•H
a
•s
W
99
-------
o
o
o
o
o
o
o
cn
M
1!
u.
E-
O
z
-^
o
o
o
in
ii
W
£
M
EH
M
II
EH
cn
a
as
ITi
CN
O
ffi
( MMASAQMSP
HOLD
MS=8381
H &i O
O Z U.
ff_ Y
"i
£ <
M u a.
< SH D
U, D O
cn O ^
\* \>
\\\*
* *
a
E-I
Cft
z
ffi
Q
X
V)
a
v XXX-
vxxx-
iH
|!
cn
HH
&4
o
«
f»
i-H
^
cn
*"*• ii
H cn
J OS
o x
>«•• o
Q cn
ZCd
O >• »O
U II O >
-a > *o
O J \0 Z
(MO Z •
II X • v
OS - T CO
co J co cn
z < cn M
18 . S?
Z J • ^ •
X II • • CJ
a EH u u z
-cn as z o
M sa id o o
o as u EH
EH - U M
!*, < U) > Z
cd ii a < M
Bi 1 ~ss
u cn x ctT< <
X x O cn • •
u cn u « M cn cn
a o a x" J 2
u a cn a M £ x
U EH Cb u. b.
x cd cn cn
OS U Z Ok « -
cd a H M
H cn z cn u
jj cn cn M H EH u.
W x x tfci u td O
^X-X-X-X-£QIO'§X->
-
§
cc *•
EH ~
Cd
OS
cn
OS . ~
— X Cd «
o cn B oo
-O, EH M
•H cn m
x a 3
M < O
EH Q O U M O Cd
-.5 o cn o as
x ' «r J • a.
* O (M <
7 £ S
O as cn M
M CX Cd CO
O • IT! CS O X •
Cd CO ir> U O II H
»» «k «• a J •
Men z z co H ocn
< < 3 3 X <
cn £ as as -X
a. S < x as
Cd U. Z Z II U, S
as tn o o H cn cn
O.M M MOCd OMOOM 3M1I
CO II EH HOD • CM • Oil Q.
HZ < < 3 BMEH XOTM
uo« jj cn Q Q
a, < <
a a > . > a a
uaa u'cd aa
x » -»
coo o o\ oa.mcnco«'O- oo
MO K CO Q n rt en
II II -1-3 !-J 1-3
cd a. ~ • • •
X cn 0 EH EH EH
MM > 3 D 3
EH a vo o o o
- - Z Cd Cd Ct}
st a <-> as as as
o < 1-5 a, a. a.
o Q — » • • •
O J EH X X X
Z OS EH Z Z Z
OOcncn3M M M
MbCdCdOCO CO CQ
O • x x • •
W cn II II H EH EH H
Xcnoocncn cn cn
O Ci» II || a> h Ct. b
cn cn H H cn co cn cn
MMD3MM M M
II II O O II II II II
xzcncnzz z z
ocnxxcncn cn cn
XXX
w a a a a o 11 a n a 11
x. a, o. a,
cd cn cn cn
aeaooooaoaoQ
CdMOOOO O 0
ccuoooo o o
MHHHEHH EH H
XXXXXXXXXXXX
XXXXXXXXXXXX
c
o
TJ
Ul
(0
*0
(C
'a
Q
Q)
Cfl
Q)
5
0
0)
i—!
•H
M-(
P
w
_ll
1
•8
T™
1
1
2
•a
g
H
^
0)
c
"<
*
^
i
IA
•H
•H
100
-------
»
(AIRQUL)
^
CO
a
*H
.PREOOT. J3
SH
CC
z
CD
<
tn
5
b
en
M
II
Z
tn
a a
X
a en
a ii
0.
to
o a
0
u_
O
t
~^.^
^~^
Cd
|
O
S
b
EH *
Z — •
RE,COHME
( NBNDARY
Cd «•
r co
»
D -H
D BE PLACE
.PREOUT.J3
J >
I|
Cd CO
J •
M E-
b <
EH W
in
B4(EABIN<
O*
rH
•d1
1-1
,N3PREOUT.>
1
M
ffi
t-1
<
to
s
U.
en
n
z
en
at Q at
I S
eo a co
ii a n
Q, O.
to to
a o a
o
ti.
f*i
EH
b
XXX
XXX
-
a
Cd
tn
3
0 |
> Z
z > —
as J a:
en H cd
EH Z OH
a. cdx
"" * Q* W
•» 2 f
00 3 ^-'
9> U «•
-4 >CD
w M cr>
h-i r* —t
.N3PREOUT.i
PLACED HEI
•PREOUT.J3:
> Cd >
2§ "5
z a z
M -J HH
ffi 3 CC
O •
E- S EH
< <
• Cd •
tn J to
i 2 i
tn < en
» a n
z < z
en o tn
a « ce a a:
X X
a en a a to
a ii a a n
a a.
tn H en
o a o o a
o b o
V ,-H Is*
H b E-
U* * Uj
XXXXX
xxxxx
o
m
-o
^- n
» Cd fH
o en ii
> J Cd
vo a N
•7 -.1-1
4C.J41984.I
,(500,100),
L3026,BLKSJ
^ — •— i
Sv; n
» J
U EH O
> — Cd
< 1! IX
• w a ^
O U *O
H < to J
< CX 00 H
to > ii o
<< X *
* o t. «
. DD DSN=ISFM1
UNIT=SYSI
DCB=(REC1
DISP=(NEl
o
o
«-t
b
XXXX
xxxx
o
rt
-O
»-»o
> tn n
*£> J Cd
z ce t-a
)NC.J41984.
,(520,100) ,
. 3026,BLKS1
^t ~ ^
O X II
EH 2 J
sta
M II CC
• W J ^
O O -15
H < eo J
< CN flQ EH
• en > <
tn - ii o
f< X *
a b s
en o td
1*, x sa z
en en cc —
II EH || a
z M a to
to z o M
a 3 a Q
a
a
^
o
o
fV
^
cC
xxxx
xxxx
[TOPCON) ,
[REGNTP) ,
» v
CO CO
m &
PREOUT.J3]
PREOUT.J3J
> >•
S S
z z
(-4 M
ffi CO
E^ EH
*4 ^
to en
< i
b b
en tn
ii »
z z
tn en
a K a a;
X E
a tn a en
a n a n
0, 0.
to tn
o a o a
0 0
b b
n «
EH EH
xxxx
xxxx
'(rvaaax;
»
CO
o>
PREOUT . J 3 1
|
M
CC
sj
^.
tn
Cm
tn
n
en
a K
z;
a tn
a ii
0.
tn
o a
o
Eh
in
EH
b
X^
i
a
M
S
V
CO
Ov
PREOUT.J41
1
M
ffi
H
en
u.
tn
n
z
tn
a at
X
a en
a ii
0.
en
o a
o
b
VC
EH
bu
XX
XX
i
en
u
0.
en
M
a
•.
o
•>
•&
z
V
n
z"
M
Cv.
^4
O
CH
<
<
en
n
en
a
a
a
0
o
u.
i~
EH
U> «
XX
XX
oncluded.
O
.0
^-i
i
in
4J
•H
5
•H
*7/
(x3
101
-------
5.1.2 UNIX-Based Operating System
Exhibit 5-2 illustrates a UAM job control file as used on a computer system operating
with the UNIX operating system. The example presented here is for a two-day appli-
cation to Atlanta. (Volume II of the user's guide contains a complete description of
the files used in the Atlanta application.) The UAM job control file first runs the
simulation preprocessor SPREP, which defines the period of simulation and several
other control parameters (see Chapter 5 in Volume II). The UAM is then exercised
for the first day (June 3, Julian day 84155) with the input files read directly off the
job control file (standard input). The list of file names could also be read from an
external file using the '<' form of redirection instead of '«' (see below for further
explanation). The job control file continues for the second day (June 4, Julian day
84156) (Exhibit 5-2b) similar to the first day except that the UAM is initialized using
the instantaneous concentrations (INSTANT) at midnight from the first day rather
than the initial condition concentrations from an AIRQUALITY file. This change in
how the model is initialized is indicated by specifying a TRUE" for the restart flag
on line 9 of the SPREP input file.
5.1.3 Location of File Manipulation Statements in the UAM
All of the file input and output (I/O) manipulation statements in the UAM are located
in the subroutine OPENA. This routine reads from a unit specified as an asterisk
(*). On UNIX systems this is standard input. When using PRIMOS, the operating
system of PRIME computers, this is the terminal or the COMI (command input) file.
On most systems the asterisk will be equivalent to either unit 1 or 5. The routine
OPENA reads a series of file names or descriptors appropriate for use in a
FORTRAN 77 OPEN statement. The file names are the input and output file names
for the UAM. Each name is on a separate line and is preceded by 20 unused spaces
that may be left blank or used for comments. The order in which the file names will
be read is as follows:
90008 5
102
-------
I
01
5?
w
CO
c
3
•^
c
£
*J
2
H
; verbose
/u2/twoc1
* *
*
\
*
*
*
*
*
*
4
*
*
41
*
*
*
*
+
*
*
*
*
*
JH
41
*
*
•ft
*
41
•ft
*49««4«4i«««*¥4
{
*
+
•ft
*
1
o
s
S
f_
>,
^
I
<
1
CD
1
o
UJ
QC
<
1
S
2
C
Of
X
c
,
o
S s
u oo
g 5.
£ oo
0
o *
1 1
CM
*
C
c
o
u
0
15
Ol
10
c
^
4J
^C
c
s
a.
1
in
^4
m
o
1
1
OJ
U3
ffm
£
c
O
u
c
o
4J
o
VI
01
^
3
CO
z
oe
o
^
_i
— 1 OO
t— •
Z fO
0
<-> UJ
z z a
•— = z
3
ei
H-
u
c c?
o t-
•j;
•£
?J§3
t/i CM <
1
••
in u
•» rr oc
Z 00 H-
5
Sss
3 — « t—
_J
<
»
iU
<£ m" 00 u_
22
,73?
•« o
O
-
§•
•8
cn
a
I
•H
U-4
0)
CM
I
in
j-j
•H
,0
•H
f.
103
-------
o
u-u ie
"Tw^rvwrrv^rrr u
?OO OO 00 CO CD 00 CO CD ^* f^ f*l **^ ^^
CCCCCCCCO)>>>>
a •
^ CO Ch C3 ^^ Psj f*^ ^ ^*»
V
c
.
£
CL
§ §
I
"O "X3 <
Ol O> i
s °
0^00-
s
0^00
$
UJ
J-> O
^^ cc
S o
I CO Z 00 *O
1 Cft O Ch tf> I— - . ac
. H- «rr ac. < o
, 9 « V 00 V— U. O
T3
(U
O
O
z o X z
•H
X
104
-------
o
4-*
u;
E
CL
I
U
~ JO C C C •-
o <*>
uj •
V)
10 GO o. fn
& c o> >
1/^3 i- O)
i. "-) O. •
— • C >v C
1C (n -- •»- 3
'^ O) £3 1C •»•">
> > > >
O, o; cu (U
. — C C C C
p O 3 3 'S 3
Ol — O. '
— ^. U
U_ »^ *J CT14J C
LL. 0) 4-1
CM
I
in
•H
43
•H
J2
X
105
-------
Input
METSCALARS
SIMCONTROL
CHEMPARAM
AIRQUALITY
BOUNDARY
DIFFBREAK
EMISSIONS
PTSOURCE
TEMPERATUR
TOPCONC
REGIONTOP
TERRAIN
WIND
Output
AVERAGE
INSTANT
DEPOSITION
This routine may be rewritten or replaced to meet the requirements of specific com-
puter systems.
5.2 ALLOCATION OF MEMORY
Most of the data memory used by the (JAM is allocated dynamically from a large
scratch array AV in the common block AVAIL. If this array is exceeded, the UAM
will issue a message to that effect and stop. The user can increase the size of the
array AV and the variable MAXAV in the subroutine AAINIT. The value of MAXAV
and the dimension of the array AV must be the same. If it is impossible to increase
the size of this array due to system limitations, the user will have to decrease the
overall size of the application by reducing the number of vertical levels, the x and y
dimensions of the domain, or the number of species.
90008 5
-------
For some memory-limited systems the user will have to reduce the array size of the
scratch array AV in AVAIL, and correspondingly the value of the variable MAXAV in
subroutine AAINIT, in order for the UAM to run.
The maximum number of species is currently set to 31, one of which will always be
used by the tracer species CLBR. If more species are required, the dimensions in the
common blocks CHPARM, CHDATA, and NETDEP must be increased. The variable
NSMAX in BLOCK DATA should also be increased.
The maximum number of levels that can be used is currently limited to 20 by
the dimension of the array FLUX1 in the subroutine WINDZ. The dimension of
FLUX1 must be one more than the maximum number of levels to be used.
The maximum x and y dimensions are currently each limited to 100 by the size of the
array XXDEP in the common block NETDEP. These dimensions can be increased to
accomodate a larger region. Note that memory requirements of the UAM can be
reduced by decreasing the size of this array if a small region is being simulated.
The maximum number of point sources is limited to 5000 by the common block
PLHITE. To allow a larger number of point sources, increase the parameter
MAXPTS.
Finally, the maximum number of species that can appear on an input file is limited to
35 by the dimensions of arrays in the common block FILCON. The size of these
arrays can be increased, if necessary, to allow input files to contain more species.
The variable NSPMAX must also be increased in the BLOCK DATA.
90008 5
107
-------
References
Ames, 3., T.. C. Myers, L. E. Reid, D. C. Whitney, S. H. Golding, S. R. Hayes, and
S. D. Reynolds. 1985a. SAI Airshed Model Operations Manuals. Volume I—
User's Manual. U.S. Environmental Protection Agency (EPA-600/8-85-007a).
Ames, 3., S. R. Hayes, T. C. Myers, and D. C. Whitney. 1985b. SAI Airshed Model
Operations Manuals. Volume II—System's Manual. U.S. Environmental Protec-
tion Agency (EPA-600/8-85-007b).
Boris, 3. P., and D. L. Book. 1973. Flux-corrected transport: I. SHASTA, a fluid
transport algorithm that works. 3. Comp. Phys., 11:38-69.
Burton, C. S. 1988. Comments on "Ozone Air Quality Models." 3. Air Pollut. Con-
trol Assoc., 38(9):1119-1128.
Carr, E., and 3. Haney. 1990. "Application of the Urban Airshed Model in the
Republic of China (Taiwan)." Air & Waste Management Association's Inter-
national Specialty Conference on Tropospheric Ozone and the Environment, Los
Angeles, California.
Chameides, W. L., R. W. Lindsay, 3. Richardson, and C. S. Kiang. 1988. The role of
biogenic hydrocarbons in urban photochemical smog: Atlanta as a case study.
Science, 241:1473-1475.
Chock, D. P. 1985. A comparison of numerical methods for solving the advection
equation—II. Atmos. Environ., 19:571-586.
Chock, D. P., and A. M. Dunker. 1983. A comparison of numerical methods for
solving the advection equation. Atmos. Environ., 17:11-24.
Daly, C., et al. 1990. "Evaluation of PARIS Performance in the South Central Coast
Air Basin." Systems Applications, Inc., San Rafael, California (in progress).
EPA. 1986. Guideline on Air Quality Models (Revised). U.S. Environmental Protec-
tion Agency (EPA-450/2-78-027R).
Gear, C. W. 1971. Numerical Initial Value Problems in Ordinary Differential Equa-
tions. Prentice-Hall, Inc., Englewood Cliffs, New 3ersey.
90006 7
109
-------
Gery, M. W., G. Z. Whitten, and 3. P. Kilius. 1988. "Development and Testing of the
CBM-IV for Urban and Regional Modeling." Systems Applications, Inc., San
Rafael, California Volume VII; Mathematical
Modeling of Urban Aerosol Dynamics. U.S. Environmental Protection Agency
-------
Liu, M. K., and J. H. Seinfeld. 1975. On the validity of grid and trajectory models of
urban air pollution. Atmos. Environ., 9:555-574.
Liu, M. K., D. C. Whitney, and P. M. Roth. 1976. "Effects of atmospheric
parameters on the concentration of photochemical pollutants," 3. Appl.
Meteorol., 15:829-835.
Liu, M. K., D. C. Whitney, J. H. Seinfeld, and P. M. Roth. 1976. Continued Research
In Mesoscale Air Pollution Simulation Modeling—Vol. I; Analysis of Model
Validity and Sensitivity and Assessment of Prior Evaluation Studies. U.S.
Environmental Protection Agency (EPA-600/4-76-016a).
Morris, R. E., T. C. Myers, and E. L. Carr. 1990. Urban Airshed Model Study of Five
Cities; Evaluation of Base Case Model Performance for the City of St. Louis
Using Rich and Sparse Meteorological Inputs. U.S. Environmental Protection
Agency (EPA-450/4-90-006C).
Morris, R. E., T. C. Myers, M. C. Causley, L. Gardner, and E. L. Carr. 1990a. Urban
Airshed Model Study of Five Cities; Low-Cost Application of the Model to
Atlanta and Evaluation of the Effects of Biogenic Emissions on Emission
Control Strategies. U.S. Environmental Protection Agency (EPA-450/4-90-
006D).
Morris, R. E., T. C. Myers, H. Hogo, L. R. Chinkin, L. A. Gardner, and R. G.
Johnson. 1990b. A Low-Cost Application of the Urban Airshed Model to the
New York Metropolitan Area and the City of St. Louis. U.S. Environmental
Protection Agency (EPA-450/4-90-006E).
OTA. 1988a. "Urban Ozone and the Clean Air Act: Problems and Proposals for
Change." Office of Technology Assessment, Washington, D.C.
OTA. 1988b. "Ozone and the Clean Air Act: Summary of OTA Workshop with State
and Local Air Pollution Control Agency Officials." Office of Technology
Assessment, Washington, D.C.
OTA. 1988c. "Ozone and the Clean Air Act: A Summary of OTA Workshops on
Congressional Options to Address Nonattainment of the Ozone Standard."
Office of Technology Assessment, Washington, D.C.
Rao, S. T. 1987. Application of the Urban Airshed Model to the New York Metro-
politan Area. U.S. Environmental Protection Agency (EPA-450/4-87-011).
Reynolds, S. D. 1977. "The Systems Applications, Inc. Urban Airshed Model: An
Overview of Recent Developmental Work." In International Conference on
Photochemical Oxidant Pollution and Its Control. U.S. Environmental Protec-
tion Agency (EPA-600/3-77-001b).
90008 7
111
-------
Reynolds, S. D., 3. Ames, T. A. Hecht, 3. P. Meyer, D. C. Whitney, and M. A.
Yocke. 1976. Continued Research In Mesoscale Air Pollution Simulation
Modeling, Volume II; Refinements in the Treatments of Chemistry,
Meteorology, and Numerical Integration Procedures. U.S. Environmental Pro-
tection Agency (EPA-600/4-76-016b).
Reynolds, S. D., H. Hogo, W. R. Oliver, and L. E. Reid. 1982. "Application of the
SAI Airshed Model to the Tuisa Metropolitan Area." Systems Applications,
Inc., San Rafael, California (SYSAPP-82/004).
Reynolds, S. D., M. K. Liu, T. A. Hecht, P. M. Roth, and 3. H. Seinfeld. 1973.
"Further Development and Validation of a Simulation Model for Estimating
Ground Level Concentrations of Photochemical Pollutants." Systems Applica-
tions, Inc., San Rafael, California (SAI 73/19).
Reynolds, S. D., M. K. Liu, T. A. Hecht, P. M. Roth, and 3. H. Seinfeld. 1974.
Mathematical modeling of photochemical air pollution—III. Evaluation of the
model. Atmos. Environ., 8:563-596.
Reynolds, S. D., 3. H. Seinfeld, and P. M. Roth. 1973. "Mathematical modeling of
photochemical air pollution—I. Formulation of the model. Atmos. Environ.,
7:1033-1061.
Roth, P. M., S. D. Reynolds, P.3.W. Roberts, and 3. H. Steinfeld. 1971. "Develop-
ment of a Simulation Model for Estimating Ground-Level Concentrations of
Photochemical Pollutants," Systems Applications, Inc., San Rafael, California
(SAI 71/21).
Roth, P. M., P.3.W. Roberts, M. K. Liu, S. D. Reynolds, and 3. H. Seinfeld. 1974.
Mathematical modeling of photochemical air pollution—II: A model and inven-
tory of pollutant emissions. Atmos. Environ., 8:97-130.
Schere, K. L. 1983. An evaluation of several numerical advection schemes. Atmos.
Environ., 17:1897-1907.
Science. 1988. Rural and urban ozone. Editorial in Science, 241(4873): 1569.
Seigneur, C., T. W. Tesche, P. M. Roth, and M. K. Liu. 1983. On the treatment of
point source emissions in urban air quality modeling. Atmos. Environ., 17:1655-
1676.
Seinfeld, 3. H. 1988a. Ozone air quality models. A critical review. 3. Air Pollut.
Control Assoc., 38(5):616.
90008 7
112
-------
Smolarkiewicz, P. K. 1983. A simple positive definite advection scheme with small
implicit diffusion. Monthly Weather Review, 111:479-486.
SAL 1989. "User's Manual for Preparing Emission Files for Use in the Urban Airshed
Model." Systems Applications, Inc., San Rafael, California (SYSAPP-89/114).
Whitten, G. Z., J. P. Killus, and H. Hogo. 1980. Modeling of Simulated Photochemi-
cal Smog with Kinetic Mechanisms. Volume 1. U.S. Environmental Protection
Agency (EPA-600/3-80-028a).
Whitten, G. Z., T. C. Myers, C. Daly, L. R. Chinkin, S. D. Reynolds, N. M. Yonkow,
and B. S. Austin. 1985. "Application of the Urban Airshed Model to Kern
County." Systems Applications, Inc., San Rafael, California (SYSAPP-85/200).
Zimmerman, D., W. Tax, M. Smith, J. Demmy, and R. Battye. 1988. Anthropogenic
Emissions Data for the 1983 NAPAP Inventory. U.S. Environmental Protection
Agency (EPA-600/7-8S-022).
90008 7
113
-------
References for Table 1-1
Attaway, L. D., et al. 1975. "Maintenance Shutdown of Tail Gas Treating Unit: An
Assessment of Potential Sulfur Dioxide Concentrations and Related Health and
Welfare Effects." Systems Applications, Inc. and Greenfield, Attaway <5c Tyler,
Inc., San Rafael, California (SYSAPP-75/086)
Anderson, G. E., et al. 1977. Air Quality in the Denver Metropolitan Region: 1974-
2000. U.S. Environmetai Protection Agency (EPA-908/1-77-002).
BAAQMD. 1987. "Air Quality Planning in the Bay Area: 1987 Update." Bay Area
Air Quality Management District, San Francisco, California.
Builtjes, P.J.H., K. D. van den Hout, and S. D. Reynolds. 1982. "Evaluation of the
Performance of a Photochemical Dispersion Model in Practical Applications."
13th International Technical Meeting on Air Pollution Modeling and Its Applica-
tion, 14-17 September 1982, He des Embiez, France.
CARB. 1985. "Kern County Ozone Modeling for SIP Revision." California Air
Resources Board, Sacramento, California.
Carr, E. L., and J. L. Haney. 1990. "Application of the Urban Airshed Model in the
Republic of China (Taiwan)." Annual meeting Air and Waste Management
Association.
Cole, S. C., D. E. Layland, G. K. Moss, and C. F. Newberry. 1983. St. Louis Ozone
Modeling Project, U.S. Environmental Protection Agency (EPA-450/4-83-019).
Daly, C., S. G. Douglas, G. E. Moore, and T. C. Myers. 1990. "Evaluation of PARIS
Performance in the South Central Coast Air Basin." Systems Applications,
Inc., San Rafael, California (SYSAPP-90/051).
Dennis, R. L., M. W. Downton, and R. S. Keil. 1983. Evaluation of Performance
Measures for an Urban Photochemical Model. U.S. Environmental Protection
Agency (EPA-450/4-83/021).
Doyle, 3. R., et al. 1980. "An Assessment of Potential Sulfur Dioxide and Sulfate
Concentrations from Maintenance Shutdown of the Tail Gas Treating Unit."
Systems Applications, Inc., San Rafael, California. (SYSAPP-80/052R)
90006 13
115
-------
Gaddo, P. P., and 3. H. Weaving. 1982. "Urban Scale Mathematical Model for Pollu-
tants in a European City." Air Pollution by Nitrogen Oxides; Proc. of the US-
Dutch International Symposium (1982), T. Schneider and L Grant, eds. Studies
in Environmental Science 21, Elsevier, Amsterdam.
Haney, J. L. 1986. "Overview of the Urban Airshed Model Validation of 25-26
October 1985 for the Phoenix Carbon Monoxide (CO) Study." Systems Applica-
tions, Inc., San Rafael, California (SYSAPP-86/126).
Haney, 3. L., and T. N. Braverman. 1985. Evaluation and Application of the Urban
Airshed Model in the Philadelphia Air Quality Control Region. U.S. Environ-
mental Protection Agency (EPA-450/4-85-003).
Haney, 3. L., S. G. Douglas, L. A. Gardner, and C. S. Burton. 1990. "Application of
the Urban Airshed Model in the Baton Rouge Seven-Parish Study Area." Annual
meeting Air and Waste Management Association.
Haney, 3. L., D. R. Souten, T. W. Tesche, L. R. Chinkin, H. Hogo, and M. C. Dudik.
1986. "Evaluation and Application of the PARIS Photochemical Model in the
South Central Coast Air Basin." Systems Applications, Inc., San Rafael, Cali-
fornia (SYSAPP-86/065).
Haney, 3. L., G. Z. Whitten, and C. S. Burton. 1988. "Calculated Response of Urban
Ozone Levels to NOX and VOC Emission Changes: An Exploratory Study of the
Philadelphia Metropolitan Area." Systems Applications, Inc., San Rafael, Cali-
fornia (SYSAPP-88/117).
Hogo, H., and M. A. Yocke. 1987. "Development of a Photochemical Air Quality
Model with Extensions to Calculate Aerosol Dynamics and Visibility." Systems
Applications. Inc., San Rafael, California (SYSAPP-87/167).
Hogo, H., et al. 1988. "Urban Airshed Model Performance Evaluation for 5-7 3une
1985." Systems Applications, Inc., San Rafael, California (SYSAPP-88/138).
Liu, C. S., and 3. E. Grisinger. 1986. "Application of Urban Airshed Model for
Regulatory Analysis in the South Coast Air Basin." Draft staff report, Planning
Division, South Coast Air Quality Management District.
Liu, M. K., et al. 1977. "Application of a Methodology for the Design of a Carbon
Monoxide Monitoring Network to Las Vegas." Systems Applications, Inc., San
Rafael, California (SYSAPP-77/016).
Mahoney, L. A., and M. A. Yocke. 1987. "Air Quality Impact Assessment of Texaco
U.S.A., Inc. Exploratory Drilling on Lease P-0512." Systems Applications, Inc.,
San Rafael, California (SYSAPP-87/013).
90006 13
116
-------
Moore, G., G. Wilson, L. Gardner, and S. Douglas. 1990. "Modeling of Future Air
Quality Impacts in the Central Coast Counties Due to OCS Development
Scenarios." Systems Applications, Inc., San Rafael, California (SYSAPP-
90/035).
Morris, R. E., T. C. Myers, and E. L. Carr. 1990*. Urban Airshed Model Study of Five
Cities; Evaluation of Base Case Model Performance for the Cities of St. Louis
and Philadelphia Using Rich and Sparse Meteorological Inputs. U.S. Environ-
mental Protection Agency (EPA-450/4-90-006C).
Morris, R. E., T. C. Myers, M. C. Causley, L. Gardner, and E. L. Carr. 1990a. Urban
Airshed Model Study of Five Cities; Low-Cost Application of the Model to
Atlanta and Evaluation of the Effects of Biogenic Emissions on Emission Con-
trol Strategies. U.S. Environmental Protection Agency (EPA-450/4-90-006d).
Morris, R. E., T. C. Myers, H. Hogo, L. R. Chinkin, L. A. Gardner, and R. G.
Johnson. 1990b. A Low-Cost Application of the Urban Airshed Model to the
New York Metropolitan Area and the City of St. Louis. U.S. Environmental
Protection Agency (EPA-450/4-90-006e).
Morris, R. E., T. C. Myers, E. L. Carr, and M. C. Causley. I990c. Urban Airshed
Model Study of Five Cities; Demonstration of Low-Cost Application of the
Model to the City of Atlanta and the Dallas-Fort Worth Metroplex Region.
U.S. Environmental Protection Agency (EPA-450/4-90-006).
Rao, S. T. 1987. Application of the Urban Airshed Model to the New York Metro-
politan Area. U.S. Environmental Protection Agency (EPA-450/4-87-011).
Reynolds, S. D. 1973. "Further Development and Validation of a Simulation Model
for Estimating Ground Level Concentrations of Photochemical Pollutants."
Systems Applications, Inc., San Rafael, California (SYSAPP-72/026, 027;
73/015, 017, 018, 019, 027).
Reynolds, S. D., et al. 1978. "Application of the SAI Airshed Model to the Evalua-
tion of Alternative Population Growth Forecasts for the South Coast Air
Basin." Systems Applications, Inc., San Rafael, California (SYSAPP-78/124).
Reynolds, S. D., et al. 1979. "Photochemical Modeling of Transportation Control
Strategies—Vol. I. Model Development, Performance Evaluation, and Strategy
Assessment." Systems Applications, Inc., San Rafael, California (SYSAPP-
79/037R).
Reynolds, S. D., et al. 1982. "Application of the SAI Airshed Model to the Tuisa
Metropolitan Area." Systems Applications, Inc., San Rafael, California
(SYSAPP-82/004).
90008 13
117
-------
Roth, P. M., M. K. Liu, D. A. Stewart, R. E. Morris, and T. C. Myers. 1981.
"Evaluation of the Impacts of NOX Emissions from the Proposed Montezuma
Generating Stations on Oxidant Concentrations in an Arc Between Lodi and
Stockton." Systems Applications, Inc., San Rafael, California (SYSAPP-
81/065)
Roth, P. M., et al. 1984. An appraisal of emissions control requirements in the Cali-
fornia South Coast Air Basin. Environ. Int., 9:549-571.
Seigneur, C. S., et al. 1983. "Effect of Power Plant Emissions of NOX on Ozone
Levels in Three Different Environments." Systems Applications, Inc., San
Rafael, California (SYSAPP-83/023).
SLC. 1986. "Environmental Impact Report/Environmental Impact Statement for
Proposed ARCO Coal Oil Point Project." California State Lands Commission,
County of Santa Barbara, and U.S. Army Corps of Engineers L.A. District.
Souten, D. R., T. W. Tesche, and W. R. Oliver. 1981. "Use of the Airshed Grid Model
to Evaluate the Air Quality Impacts of Alternative Air Pollution Control
Policies on the South Coast Air Basin." Systems Applications, Inc., San Rafael,
California (SYSAPP-80/220).
Stern, R., and B. Scherer. 1982. "Simulation of a Photochemical Smog Episode in
the Rhine-Ruhr Area with a Three-Dimensional Grid Model." 13th Inter-
national Technical Meeting on Air Pollution and Its Application, 14-17
September 1982, lie des Embiez, France.
Tesche, T. W., and C. S. Burton. 1978. "Simulated Impact of Alternative Emissions
Control Strategies on Photochemical Oxidants in Los Angeles." Systems Appli-
cations, Inc., San Rafael, California (SYSAPP-78/022R).
Tesche, T. W., W. R. Oliver, and 3. L. Haney. 1982. "Photochemical Modeling of the
SCE Ivanpah Generating Station." Systems Applications, Inc., San Rafael, Cali-
fornia (SYSAPP-82/212).
Tesche, T. W., et al. 1988. "Photochemical Modeling of Two SCCCAMP 1984 Oxi-
dant Episodes. Volume II—Modeling Procedures and Evaluation Results."
Radian Corporation, Sacramento, California.
Tesche, T. W., et al., 1981. "Sensitivity of Complex Photochemical Model Estimates
to Detail in Input Information." Systems Applications, Inc., San Rafael, Cali-
fornia (SYSAPP-81/005).
90006 13
118
-------
Wackter. D. J. 198S. "Sensitivity Analysis of Ozone Predictions by the Urban Air-
shed Model in the Northeast." Connecticut Department of Environmental Pro-
tection, Hartford, Connecticut.
Wagner, K. K., and N. Wheeler. In preparation.
Wakamutsu, S., I. Uno, and K. L. Schere. 1987. "Application of a Three-Dimensional
Photochemical Smog Formation Model to the Tokyo Metropolitan Area." 16th
NATO-CCMS International Technical Meeting on Air Pollution Modeling and Its
Applications, Lindau, FRG, April 6-10, 1987.
Wakamutusu, S., et al. 1986. "A Study Using a Three-Dimensional Smog Formation
Model Under Conditions of Complex Flow." Proceedings of the 7th World
Clean Air Congress (Sydney, Australia, August 25-29, 1986).
Whitten, G. Z., et al. 1985. "Application of the Urban Airshed Model to Kern
County." Systems Applications, Inc., San Rafael, California. (SYSAPP-85/200)
Yocke, M. A., et al. 1985. "Analysis of the Air Quality Impacts of the San Miguel
Project." Systems Applications, Inc., San Rafael, California (SYSAPP-85/127).
Yocke, M. A., et al. 1987. "San Diego (SANDER) Energy Recovery Project Photo-
chemical Modeling Analysis." Systems Applications, Inc., San Rafael, Cali-
fornia (SYSAPP-87/029).
90008 13
119
-------
Historical Bibliography on the Urban Airshed Model
1971 "Development of a Simulation Model for Estimating Ground-Level Con-
centrations of Photochemical Pollutants." Systems Applications, Inc.
(SYSAPP-71/021), including the following appendixes (separately bound):
Appendix A: Contaminant Emissions in the Los Angeles Basin—Their
Sources, Rates and Distribution. P.J.W. Roberts, P. M. Roth, and C. L.
Nelson. (SYSAPP-71/006).
A Vehicle Emissions Model for the Los Angeles Basin — Extensions and
Modifications. A Continuation of the Work Reported in Appendix A.
P.J.W. Roberts, M. K. Liu, and P. M. Roth. (SYSAPP-72/008)
Appendix B: A Kinetic Mechanism for Atmospheric Photochemical
Reactions. 3. H. Seinfeld, T. A. Hecht, and P. M. Roth. (SYSAPP-71/009)
Appendix C: The Treatment of Meteorological Variables. P. M. Roth,
S. D. Reynolds, and PJ.W. Roberts. (SYSAPP-71/017)
Appendix D: Numerical Integration of the Continuity Equations. S. D.
Reynolds. (SYSAPP-71/012)
Appendix E: Air Quality Data Used in Model Validation. P.J.W. Roberts
and P. M. Roth. (SYSAPP-71/007)
Appendix F: Description of the Computer Program. S. D. Reynolds.
(SYSAPP-71/026)
1972 Evaluation of a Diffusion Model for Photochemical Smog Simulation. A. Q.
Eschenroeder, J. R. Martinez, and R. A. Nordsieck. U.S. Environmental
Protection Agency (EPA-R4-73-012a).
Development and validation of a generalized mechanism for photochemical
smog. T. A. Hecht and J. H. Seinfeld. Environ. Sci. Technol., 6:47-57.
Further Development and Validation of a Simulation Model for Estimating
Ground Level Concentrations of Photochemical Pollutants. Systems
Applications, Inc. (SAI 73/19), with the following:
90008 t3
121
-------
Volume II: User's Guide and Description of the Computer Programs. S. D.
Reynolds. (SYSAPP-73/018)
Volume II, Appendix: User's Guide and Description of the Data
Preparation Programs. D. C. Whitney. (SYSAPP-73/027)
Appendix A: Extensions and Modifications of a Contaminant Emissions
Model and Inventory for Los Angeles. PJ.W. Roberts, M. K. Liu, S. D.
Reynolds, and P. M. Roth. (SYSAPP-73/015)
Appendix B: Further Validation of a Generalized Mechanism Suitable for
Describing Atmospheric Photochemical Reactions. T. A. Hecht.
(SYSAPP-72/026)
Appendix C: A Microscale Model for Describing the Contribution of Local
Vehicular Sources to Pollutant Concentrations Measured at Monitoring
Stations. M. K. Liu and P. M. Roth. (SYSAPP-72/027)
Appendix D: Numerical Integration of the Continuity Equations. S. D.
Reynolds. (SYSAPP-73/017)
1973 Controlled Evaluation on the Reactive Environmental Simulation Model
(REM), Volume I. Pacific Environmental Services. U.S. Environmental
Protection Agency (EPA-R4-73-013a).
"Mathematical modeling of photochemical air pollution—I. Formulation of the
model. S. D. Reynolds, J. H. Seinfeld, and P. M. Roth. Atmos. Environ.,
7:1033-1061.
"Staff Report on the Preparation of Data Files for the SAI Model." D. Ross.
California Department of Transportation, District 7, Environmental
Investigations Section.
1974 Mathematical modeling of photochemical air pollution—II: A model and
inventory of pollutant emissions. P. M. Roth, P.3.W. Roberts, M. K. Liu, S. D.
Reynolds, and 0. H. Seinfeld. Atmos. Environ., 8:97-130.
Mathematical modeling of photochemical air pollution—III. Evaluation of the
model. S. D. Reynolds, M. K. Liu, T. A. Hecht, P. M. Roth, and 3. H.
Seinfeld. Atmos. Environ., 8:563-596.
Further development of generalized kinetic mechanism for photochemical
smog. T. A. Hecht, J. H. Seinfeld, and M. C. Dodge. Environ. Sci. Tech.,
8:327ff.
90008 i*3
122
-------
1975 "Interim Evaluation of Strategies for Meeting Ambient Air Quality Standards
for Photochemical Oxidant." S. D. Reynolds and 3. H. Seinfeld. Environ. Sci.
Technol., 9:433-447.
Numerico-empirical analyses of atmospheric diffusion theories. R. G. Lamb,
W. H. Chen, and 3. H. Seinfeld. 3. Atmos. Sci., 32:1794-1807.
On the validity of grid and trajectory models of urban air pollution. M. K. Liu
and 3. H. Seinfeld. Atmos. Environ., 9:555-574.
1976 "Effects of atmospheric parameters on the concentration of photochemical
pollutants." M. K. Liu, D. C. Whitney, and P. M. Roth. 3. Appl. Meteorol.,
15:829-835.
1976 Continued Research in Mesoscale Air Pollution Simulation Modeling. US
Environmental Protection Agency (EPA-600/4-76-016), including:
Volume I: Analysis of Model Validity and Sensitivity and Assessment of
Prior Evaluation Studies," Liu, M. K., D. C. Whitney, 3. H. Seinfeld, and
P. M. Roth.
Volume II: Refinements in the Treatments of Chemistry, Meteorology, and
Numerical Integration Procedures. S. D. Reynolds, 3. Ames, T. A. Hecht,
3. P. Meyer, D. C. Whitney, and M. A. Yocke.
Volume III: Modeling of Microscale Phenomena. R. G. Lamb.
Volume IV: Examination of the Feasibility of Modeling PHotochemical
Aerosol Dynamics. T. N. 3erskey and 3. H. Seinfeld.
Volume V: Refinements in Numerical Analysis, Transport, Chemistry, and
Pollutant Removal. 3. P. Killus, 3. P. Myer, D. R. Durran, G. E. Anderson,
T. N. 3erskey, S. D. Reynolds, and 3. Ames.
Volume VI: Further Studies in the Modeling of Microscale Phenomena.
R. G. Lamb, W. R. Shu, D. R. Durran, 3. H. Seinfeld, and L. E. Reid.
Volume VII: Mathematical Modeling of Urban Aerosol Dynamics. T. N.
3erskey, 3. H. Seinfeld, F. Geibar, and L. E. Reid.
"Oxidant/Ozone Ambient Measurement Methods: An Assessment and
Evaluation." C. S. Burton, P. 3. Bekowies, R. I. Pollack, and P. Connell.
Systems Applications, Inc. (SYSAPP-76/111R).
"Air Quality Assessment Statement." Colorado Division of Highways, 3oint
Regional Planning Program, CY-1976.
90008
123
-------
"1-470 Detailed Assessment Report." Colorado Division of Highways.
1977 Air Quality in the Denver Metropolitan Region; 1974-2000. G. E. Anderson,
S. R. Hayes, M. 3. Hillyer, 3. P. Killus, and P. V. Mundkur. US Environmental
Protection Agency. (EPA-909/1-77-002).
"Analysis of Air Quality in the Denver Air Quality Maintenance Area."
Colorado Department of Health, Air Pollution Control Division.
"The Systems Applications, Inc. Urban Airshed Model: An Overview of Recent
Developmental Work." S. D. Reynolds. In International Conference on
Photochemical Oxidant Pollution and Its Control. U.S. Environmental
Protection Agency (EPA-600/3-77-001b).
Mathematical Modeling of Simulated Photochemical Smog. G. Z. Whitten and
H. Hogo. U.S. Environmental Protection Agency (EPA-600/3-77-011).
1978 "Oxidant Model Applications: Denver." D. E. Donnelly. 57th Annual
Transportation Research Board Meeting, Washington, D.C.
1979 "Photochemical Modeling of Transportation Control Strategies, Volume I:
Model Development, Performance Evaluation, and Strategy Assessment."
S. D. Reynolds, L. E. Reid, and M. 3. Hillyer. Systems Applications, Inc., San
Rafael, California (SAI 79/37R).
An air quality model performance assessment package. K. E. Bencala and
J.H.Seinfeld. Atmos. Environ., 13:1181-1185.
Performance Measures and Standards for Air Quality Simulation Models. S. R.
Hayes. US Environmental Protection Agency (EPA-450/4-79-032).
"Recent Verification Studies with the SAI Urban Airshed Model in the Sothy
Coast Air Basin." T. W. Tesche, C. S. Burton, and V. A. Mirabella. Proc. of
Fourth Symposium on Turbulence, Diffusion, and Air Pollution.
1980 The Carbon-Bond Mechanism: A condensed kinetic mechanism for
photochemical smog." G. Z. Whitten, H. Hogo, and 3. P. Killus. Environ. Sci.
Technol. 14:690-700.
Modeling of Simulated Photochemical Smog with Kinetic Mechanisms. G. Z.
Whitten, 3. P. Killus, and H. Hogo. U.S. Environmental Protection Agency
(EPA-600/3-80-028a).
Guideline for Applying the Airshed Model to Urban Area. U.S. Environmental
Protection Agency (EPA-450/4-80-020).
90008 13
124
-------
1981 Judging air quality model performance. D. G. Fox. Bull. Amer. Meteorol.
Soc., 62:599-609.
"Simulation of Impacts of Nitrogen Oxide Control Strategies Under Oxidant
Episode Conditions." 3. P. Killus, W. R. Oliver, P. D. Gutfreund, 3. E.
Langstaff, T. W. Tesche, H. 3. Su. Systems Applications, Inc. (SYSAPP-
81/165).
"Airshed Model Sensitivity to Emissions inventory Perturbations—26-27 3une
1974 South Coast Air Basin Oxidant Episode." T. W. Tesche, W. R. Oliver,
D. R. Souten, H. Hogo, and 3. L. Haney. Systems Applications, Inc. (SYSAPP-
81/054).
"Sensitivity of Complex Photochemical Model Estimates to Detail in Input
Information." T. W. Tesche, C. Seigneur, L. E. Reid, P. M. Roth, W. R. Oliver,
and 3. C. Cassmassi. Systems Applications, Inc. (SYSAPP-81/004, 005, 006)
"Evaluation of the Impacts of NOX Emissions from the Proposed Montezuma
Generating Stations on Oxidant Concentrations in an Arc between Lodi and
Stockton." P. M. Roth, M. K. Liu, D. A. Stewart, R. E. Morris, and T. C.
Myers. Systems Applications, Inc. (SYSAPP-81/065).
1982 A New Carbon-Bond Mechanism for Air Quality Simulation Modeling. 3. P.
Killus and G. Z. Whitten. U.S. Environmental Protection Agency (EPA 600/3-
82-041).
1983 Modeling the effects of emission controls in the Netherlands. P.3.H. Builtjes
and S. D. Reynolds. Environ. Inter., 9:573-580.
Evaluation of Performance Measures for an Urban Photochemical Model.
R. L. Dennis, M. W. Downton, and R. S. Keil. U.S. Environmental Protection
Agency (EPA-450/4-83-021).
A Review of Recent Applications of the SAI Urban Airshed Model. D. E.
Layland and H. S. Cole. U.S. Environmental Protection Agency (EPA-450/4-
84-004).
"Assessment of NOX Emission Control Retirements in the California South
Coast Air Basin", Systems Applications, Inc., including:
Volume 3: Results, Findings, and Implications for Control Requirements
Based on Short-term Modeling Studies. P. M. Roth, C. Seigneur, S. D.
Reynolds, and T. W. Tesche. (SYSAPP-83/157).
90008
125
-------
Volume >4: Appendix A—Performance Evaluation of the Systems
Applications Airshed Model for the 26-27 June 1974 O-j Episode in the
South Coast Air Basin. T. W. Tesche, W. R. Oliver, H. Hogo, P. Saxena,
and 3. L. Haney. (SYSAPP-83/037).
Volume 4: Appendix B—Performance Evaluation of the Systems
Applications Airshed Model for the 7-8 November 1978 NC>2 Episode in the
South Coast Air Basin. T. W. Tesche, W. R. Oliver, H. Hogo, P. Saxena,
and 3. L. Haney. (SYSAPP-83/038).
Volume 4: Appendix C, Emission Inventory Review and Development.
W. R. Oliver, H. Hogo, and P. Saxena. (SYSAPP-83/053).
Volume 4: Appendix D, Model Simulations of the Effect of ROG and NOX
Emission Reductions on Short-term O-j and NO2 Ambient Concentrations.
C. Seigneur, S. D. Reynolds, P. M. Roth, J. L. Haney, H. Hogo, M. C.
Dudik, W. R. Oliver, and T. W. Tesche. (SYSAPP-83/159).
On the treatment of point source emissions in urban air quality modeling.
C. Seigneur, T. W. Tesche, P. M. Roth, and M. K. Liu. Atmos. Environ.,
17:1655-1676.
Modeling of Simulated Photochemical Smog with Kinetic Mechanisms. G. Z.
Whitten, 3. P. Killus, and R. G. 3ohnson. U.S. Environmental Protection
Agency (EPA 600/3-82-043).
"Evaluation of Three First-Order Closure Models at a Plains She." S. D.
Reynolds, R. E. Morris, T. C. Myers, and M. K. Liu. Systems Applications,
Inc. (SYSAPP-83/173).
1984 An appraisal of emissions control requirements in the California South Coast
Air Basin. P. M. Roth, S. D. Reynolds, T. W. Tesche, P. D. Gutfreund, and
C. Seigneur. Environ. Inter., 9:549-571.
"Operational Validation of Gaussian and First-Order Closure Plume Models at
a Moderately Complex Terrain Site." S. D. Reynolds, T. C. Myers, 3. E.
Langstaff, M. K. Liu, G. E. Moore, and R. E. Morris. Systems Applications,
Inc. (SYSAPP-84/126).
1985 "Kern County Ozone Modeling for SIP Update." California Air Resources
Board, Technical Support Division, Sacramento, California.
Evaluation and Application of the Urban Airshed Model in the Philadelphia Air
Quality Control Region. 3. L. Haney and T. N. Braverman. U.S.
Environmental Protection Agency (EPA-450/4-85-003).
90008 43
126
-------
1986 "Evaluation and Application of the PARIS Photochemical Model in the South
Central Coast Air Basin, Volume I." 3. L. Haney, D. R. Souten, T. W. Tesche,
L. R. Chinkin, H. Hogo, and M. C. Dudik. Systems Applications, Inc.
(SYSAPP-86/065).
1987 "Evaluation of a Photochemical Air Quality Model with Extensions to
Calculate Aerosol Dynamics and Visibility." H. Hogo and M. A. Yocke.
Systems Applications, Inc. (SYSAPP-87/017).
1988 Development and Testing of the CBM-IV for Urban and Regional Modeling.
M. W. Gery and G. Z. Whitten. U.S. Environmental Protection Agency (EPA
600/3-88-012).
1990 Urban Airshed Model Studies of Five Cities. U. S. Environmental Protection
Agency (EPA-450/4-90-006). Includes the following volumes:
"Summary Report." (EPA-450/4-90-006a)
"Demonstration of Low-Cost Application of the Model to the City of Atlanta
and the Dallas-Fort Worth Metroplex Region." R. E. Morris, T. C. Myers, E.
L. Carr, and M. C. Causley. (EPA-450/f-90-006b)
"Evaluation of Base Case Model Performance for the Cities of St. Louis and
Philadelphia Using Rich and Sparse Meteorological Inputs." R. E. Morris, T.
C. Myers, and E. L. Carr. (EPA-450/
-------
Appendix I
THE CARBON BOND IV CHEMICAL MECHANISM
AND IMPLEMENTATION IN THE UAM
This appendix describes the Carbon Bond IV (CB-IV) chemical mechanism imple-
mented in the Urban Airshed Model (UAM). It contains two parts. The first part is a
description of the CB-IV that was published in the Journal of Geophysical Research,
Volume 94, 1989. This version of the Carbon Bond Mechanism was condensed down
into a usable number of reactions and species for implementation in a photocehmcial
air quality simulation model. The second part of the appendix describes how the CB-
IV was implemented in the Urban Airshed Model.
90008 38
-------
Part 1
A PHOTOCHEMICAL KINETICS MECHANISM FOR URBAN
AND REGIONAL SCALE COMPUTER MODELING
90008 38
-------
A PHOTOCHEMICAL KINETICS MECHANISM FOR URBAN
AND REGIONAL SCALE COMPUTER MODELING
16 June 19S9
Michael W. Gery
Atmospheric Research Associates
Boston, Massachusetts
Gary Z. Whitten
Systems Applications, Inc.
San Rafael, California
James P. Killus
433 Michigan Avenue
Berkeley, California
Marcia C. Dodge
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina
Paper submitted for publication in the Journal of Geophysical Research
510200 880t9r2
-------
A new chemical kinetics mechanism for simulating urban and regional photo-
chemistry has been developed and evaluated. The mechanism, called the Carbon
Bond Mechanism IV (CBM-IV), was derived by condensing a detailed mechanism that
included the most recent kinetic, mechanistic, and photolytic information. The
CBM-IV contains extensive improvements to earlier carbon bond mechanisms in the
chemical representations of aromatics, biogenic hydrocarbons, peroxyacetyl nitrates,
and formaldehyde. The performance of the CBM-IV was evaluated against data from
170 experiments conducted in three different smog chambers. These experiments
included NOx-air irradiations of individual organic compounds as well as a number of
simple and complex organic mixtures. The results of the evaluation indicate substan-
tial improvement in the ability of the CBM-IV to simulate aromatic and isoprene
systems with average overcalculation of ozone concentrations of 1% for the aro-
matic simulations and 6% for the isoprene simulations. The mechanism also per-
formed well in simulating organic mixture experiments. Maximum ozone concentra-
tions calculated for 68 of these experiments were approximately 2% greater than the
observed values while formaldehyde values were low by 9%.
INTRODUCTION
It has been known for decades that the formation of ozone and other oxidants results
from the complex chemical interaction of oxides of nitrogen (NOX) and organic
species in the presence of solar radiation [Haagen-Smit, 1952]. Photochemical
kinetics mechanisms that describe these processes are now used in atmospheric
models to estimate the amount and type of emission reductions needed to limit the
formation of urban ozone. These mechanisms are also used to study the formation of
regional ozone and oxidants involved in cloud water acidification. As new kinetic
and mechanistic information becomes available, mechanisms are updated or reformu-
lated to provide a more complete representation of tropospheric chemical pro-
cesses. In this paper we describe the development, condensation and evaluation of an
updated photochemical kinetics mechanism designed for use in urban and regional
scale air quality simulation models (AQSM).
*
Because large AQSMs are constrained by computing capacity, it is not feasible to
represent explicitly the chemistry of the hundreds of organic species present in the
troposphere; steps must be taken to organize the reactive organic chemistry into a
manageable, yet accurate, form. To achieve this, surrogate approximations are used
to classify the common chemistry of various reactive hydrocarbon groups. Because
there are a number of approaches for surrogate grouping, significant differences
between mechanisms usually occur in the organic representations, even though the
mechanisms may be based on common chemical kinetics information.
In this study we utilized the carbon bond approach for the lumping of organic species
[Whitten et al., 1980]. Organics are grouped according to bond type (for example, as
880»»9r2 m 135
-------
carbon single bonds, carbon double bonds, or carbonyl bonds). This structural-lump-
ing technique categorizes the reactions of similar carbon bonds, as opposed to the
molecular-lumping approach that groups the reactions of entire molecules. The
latter technique has been used successfully by a number of investigators [Atkinson et
al., 1982; Leone and Seinfeld, 1985; Lurmann et al., 1986; and National Center for
Atmospheric Research (NCAR), 1987], The main advantage of the structural-lump-
ing approach is that fewer surrogate categories are needed to represent bond groups,
whereas a larger number of molecular surrogates are generally needed to represent
the diversity of organic species emitted into the environment. Consequently,
lumped-structure mechanisms generally contain fewer species than lumped-molecule
mechanisms and therefore they are more easily implemented in large AQSMs.
Besides having the appropriate size for implementation in a large AQSM, an addi-
tional mechanism requirement is to reproduce experimental observations with
reasonable accuracy. Traditionally, tests are performed to evaluate the ability of a
mechanism to simulate ozone formation for closely controlled smog chamber experi-
ments. While smog chamber data are not free of experimental uncertainties, includ-
ing chamber-induced artifacts, simulation of these data is currently the best means
of evaluating the performance of a photochemical kinetics mechanism.
The starting point for our mechanism development was the Carbon Bond Mechanism
III [Killus and Whitten, 1982], The CBM-III was updated and expanded considerably in
this study to produce a detailed mechanism that we have named the CBM-EX
(expanded). This mechanism was tested against selected smog chamber data and
condensed to a level appropriate for use in large AQSMs. The condensed mechanism
(CBM-IV) was then evaluated using approximately 170 experiments conducted in
three different smog chamber facilities.
FORMULATION OF THE DETAILED
CARBON BOND MECHANISM (CBM-EX)
The CBM-EX is based on the most recent kinetic, mechanistic, and photolytic data
available. Much information was obtained from the comprehensive reviews of
Atkinson and Lloyd [1.984], Baulch et al. [1984], and DeMore et al. [1985]. Extensive
use was also made of more recent data available in the literature and from individual
investigators.
The CBM-EX treats the reactions of four different types of species: (1) inorganic
species; (2) organic species that, because of their unique chemistry or special impor-
tance in the environment, are treated explicitly; (3) organic species that are repre-
sented by carbon bond surrogates; and, (
-------
reactive and its oxidation chemistry is sufficiently different from that of other car-
bonyl species, it is represented explicitly. Ethene is also treated explicitly because
it (1) reacts significantly slower than other alkenes, (2) constitutes a large fraction
of hydrocarbon emissions, and (3) yields a high percentage of formaldehyde under
most conditions. Isoprene, the third and final organic represented explicitly, is the
most prominent of the biogenic hydrocarbons emitted into the rural environment.
Although isoprene is an alkene, its rate of reaction with OH and NOj radicals is
much faster than most other alkenes so that it warrants individual treatment.
Carbon bond surrogates are used to describe the chemistry of three different types
of carbon bonds commonly found as parts of larger organic molecules. The single-
bonded one-carbon-atom surrogate PAR is used to represent the chemistry of alkanes
and most of the alkyl groups found in other organics. The carbon bond surrogate OLE
(olef in), which contains two carbon atoms, is used to represent the carbon-carbon
double bonds that are found in 1-alkenes. A third surrogate, ALD2, which also con-
tains two carbon atoms, is used to represent the -CHO group and adjacent carbon
atom in acetaldehyde and higher aldehydes. It is also used to represent 2-alkenes
since these species react very rapidly in the environment to produce aldehyde pro-
ducts. Two molecular surrogates represent the chemistry of aromatic hydrocar-
bons. The surrogate TOL is a seven-carbon species used to categorize monoalkylben-
zene structures, and its chemistry is based on the reactions of toluene. XYL is an
eight-carbon surrogate used to represent dialkylbenzenes and trialkylbenzenes, and
its chemistry is based on m-xylene.
Figure 1 depicts the major species included in the CBM-EX in terms of their hier-
archical levels within the mechanism. The simplest species, in terms of molecular
complexity (NOX, HOX, CO and formaldehyde), occupy the lowest levels, while the
most complex species, in terms of their oxidation products, occupy the highest
levels. Table 1 provides examples of how selected organics in the CBM-EX are
partitioned among the organic groupings shown in Figure 1. The species n-butane,
which contains four alkyl carbons, is represented as 4 PAR; similarly, 2,2,5-tri-
methylpentane with its eight alkyl carbons is represented as 8 PAR. Ethene is cate-
gorized explicitly as ETH. The carbon-carbon double bond in the 1-alkene propene is
represented as I OLE, and the methyl group in propene is represented as I PAR. The
internal alkene trans-2-butene, because of its high reactivity, is treated as if it had
already reacted with OH radicals to form two molecules of acetaldehyde (2 ALD2).
Toluene and m-xylene are categorized explicitly as TOL and XYL. Ethylbenzene, an
8-carbon monoalkylbenzene, is represented as 1 TOL, with the remaining methyl
group represented as I PAR. Trimethylbenzene, which contains methyl groups at
three sites on the aromatic ring, is represented as 1 XYL and 1 PAR. Isoprene and
formaldehyde are categorized explicitly as ISOP and FORM. Propionaldehyde, which
contains three carbon atoms, is represented as 1 ALD2 and 1 PAR. As is evident
from the table, carbon atoms are conserved in the CBM-EX, which is one of the
advantages of the carbon bond formulation.
660»»9r2
-------
PAR
CUE
ISOP
TOL, XYL
\ /
FIGURE 1. Major species in the- CBM-EX and their
hierarchical relationship.
138
-------
TABLE 1. Representation of Selected
Organics in the CBM-EX
Organic Species Representation
n-Butane 4 PAR
2,2,5-Trimethylpentane 8 PAR
Ethene 1 ETH
Propene 1 OLE * 1 PAR
trans-2-Butene 2 ALD2
Toluene 1 TOL
n-Xylene 1 XYL
Ethylbenzene 1 TOL + 1 PAR
Trimethylbenzene 1 XYL + 1 PAR
Isoprene 1 ISOP
Formaldehye 1 FORM
Acetaldehyde 1 ALD2
Propionaldehye 1 ALD2 + 1 PAR
139
-------
The 204 reactions and rate constants that comprise the CBM-EX are shown in Table
Al. The 87 species contained in the mechanism are listed in Table A2 and explana-
tory notes detailing the sources of the kinetic and mechanistic data are found in
Appendixes A and B.1 The CBM-EX is described in detail by Gery et al. [198S] and
only a brief overview of the mechanism will be presented here.
INORGANIC CHEMISTRY
Inorganic chemistry is described by reactions (Al) - (A36) in Table Al. This group of
both photolytic and thermal reactions describes the chemistry of ozone, various NOX
species, hydrogen peroxide, OH and H©2 radicals, carbon monoxide, and nitric,
nitrous and peroxynitric acids.
Formaldehyde
Formaldehyde (FORM) chemistry is described by reactions (A37) - (A45). The photo-
lysis of formaldehyde ((A38) and (A39)) is the major source of radicals in both smog
chamber and atmospheric simulations. Unfortunately, the rates of these photolytic
reactions are somewhat uncertain because of conflicting absorption cross section
data [Bass et al., 1980; Moortgat et al., 1980, 1983]. In smog chamber simulations,
this uncertainty is amplified because formaldehyde is subject to heterogeneous, wall-
related processes. Since formaldehyde is one of the most important species in the
CBM-EX hierarchy, particular attention was paid |o simulating the formal-
dehyde/NOx chamber experiments. In this way we gained confidence that the repre-
sentation of chamber factors for formaldehyde photolysis, oxidation and wall-related
effects (discussed below) is adequate to describe the chemistry of the organics that
react to form formaldehyde. There is, however, no assurance that the formaldehyde
reactions remaining in the mechanism once the chamber-dependent reactions are
removed will accurately simulate atmospheric chemistry. This uncertainty will
remain a major weakness in all photochemical kinetics models until better data are
obtained from both the laboratory and smog chamber.
Appendix B is available with entire article on microfiche. Order from American
Geophysical Union, 2000 Florida Avenue, N.W., Washington, DC 2009. Document
D89-005; $2.50. Payment must accompany order.
$5.00
-------
Higher Molecular Weight Aldehydes
Higher molecular weight aldehydes were simulated using the carbon-bond surrogate
ALD2 (initial reactions represented by (A
-------
become competitive with peroxy-NO reactions and must be accurately represented
to predict both the peroxy radical concentrations and the formation rate of the
peroxide products. Only a few organic peroxy radical combination reactions have
been studied. Moreover, there is a discrepancy in the published reaction rates of
organic peroxy radicals with HC>2. The overall rate constants we use for reactions of
HO2 with C2O, ((A69) + (A70)) and MEO2 ((A67) + (A68)) are based on the studies of
Addison et al. [1980] and Cox and Tyndall [1980]. Although these values are greater
than other estimates [Kan et al., 1980; Niki et al., 1982] by large factors, they are
the only values based on peroxyacetyl and methylperoxy radical decay rates. The
lower rates from the other studies were based on the formation rates of observed
products (peroxides) and they may not be representative of the complete product
yield. Hence, we have based the peroxide-forming channel of those reactions ((A67)
and (A69)) on the lower, product-based rates and the overall reaction rates (the sum
of both channels) on the radical-decay-based rates. These choices for the overall
rates are now supported by recent measurements of the reaction rate of HO2 with
MEO2 [McAdam et al., 1987] and with C2H^O2 [Cattell et al., 1986], but we suggest
additional investigation.
Alkanes and Alkyl Groups
Alkanes and alkyl groups are represented in the carbon bond approach by the one-
carbon atom species PAR, which reacts with the general chemical characteristics of
the alkyl portion of the organic mixture under study. These reactions are presented
in Table Al as (A75) - (A92). Methane oxidation is represented by the pseudo first-
order reaction (A74), where a global background methane concentration of 1.85 ppm
has been incorporated into the rate constant. In the CBM-EX, the secondary and
tertiary peroxy radical products formed in alkane + OH reactions are lumped into the
species RO2R, and the primary radicals formed in the alkane reactions are lumped
into the species RO2. These peroxy radicals either oxidize NO to NO2 or assimilate
NO to form nitrate species. During NO-to-NO2 conversion, both RO2 and RO2R
form their respective distribution of ALD2, ketones, and the generalized alkoxy
radical, ROR. This radical can react to form nitrates, decompose, or isomerize to a
distribution of products, depending on the makeup of the organic mix being simu-
lated.
The surrogate representation of primary, secondary and tertiary alkyl carbon, with
respect to both kinetics and product distributions, is complex and discussed exten-
sively in Appendix B and by Gery et al. [1988]. The representation in the CBM-EX is
itself condensed from a much larger explicit representation, necessitating the inclu-
sion of two "counter" species that are used to balance carbon mass and radical func-
tion. The first such species, X, is used to remove a carbon atom from the system
(see (A86)). For instance, in (A77) a one-carbon peroxy radical (RO2) reacts to form
an aldehyde product. In order to conserve carbon mass, the aldehyde product is
represented as ALD2 + X, which is equivalent to ALD2 - PAR since ALD2 contains
two carbon atoms. The second "counter" species, D, converts the alkyl carbon that
880H9r2 "* 142
-------
would exist in a multicarbon alkoxy radical to the proper alkoxy radical decomposi-
tion products. These reactions are shown as (A87) - (A90). Again, the reader is
advised to consult Appendix B or Gery et al. [1988] for a more complete discussion of
this section of the carbon bond mechanism.
Ethene and Alkene
Ethene and alkene bond chemistry is represented in the CBM-EX by reactions (A93)
through (A 127). While ethene chemistry is explicitly represented using the species
ETH, the alkene carbon bond group in other 1-alkenes is categorized with the two-
carbon surrogate OLE. The mechanisms for OLE reactions with O, OH, O^ and NOj
are based mainly on propene and butene, which are the alkene species that have
received the greatest experimental attention. Information concerning sources of
reaction rates and product distributions is provided in Appendixes A and B.
Aromatic Hydrocarbons
Aromatic hydrocarbons are represented in the CBM-EX by the two surrogate species
TOL and XYL. These species react very rapidly with OH; hence, the most important
chemical features to treat in the surrogate chemistries are the OH reaction rate and
the secondary reaction scheme. Of these, appropriate representation of the
secondary reaction scheme is the most difficult to achieve because the chemistry is
complex and poorly understood. A large fraction of the reacted organic mass has yet
to be accounted for in laboratory studies. In this investigation we used smog
chamber data as an additional source of information when attempting to describe
aspects of the chemistry that have not yet been studied directly. These aspects were
then represented by a number of assumptions that were linked to all available labora-
tory kinetic data (always giving more weight to the known laboratory data). The
process is reported in Appendix B and by Gery et al. [1988]. Only the new and signi-
ficant features are summarized next.
One important difference between the chemistry of toluene and that of higher mole-
cular weight aromatic species is the inability of toluene to promptly form high yields
of reactive products (such as glyoxal and methyiglyoxal) when oxidized by OH radi-
cals [Shepson et al., 1984; Bandow et al., 1985; Gery et al., 1985; Tuazon et al.,
1986]. Another related phenomenon has been observed in toluene/NOx smog chamber
experiments. When NO is present in these systems, ozone formation is initially very
fast; but as NOX levels decrease, ozone formation rapidly terminates. Such
delineated dichotomous behavior is unique for a reactive hydrocarbon and may indi-
cate that, unlike alkenes and other aromatics, the products formed in the initial
stage of a toluene experiment may change to less reactive secondary products as the
experiment progresses.
-------
Most kinetics simulation models represent the addition of OH to toluene with a reac-
tion scheme similar to
TOL + OH (+O2) + TO2 (A130)
TO2 + NO - NO2 + OPEN + HO2 (A If 5)
where TO2 is a peroxy radical formed by the addition of molecular oxygen to the
toluene-OH adduct, and OPEN encompasses a mixture of reactive dicarbonyl pro-
ducts assumed to form after fragmentation of the aromatic ring. If this formulation
is used exclusively, the mechanism produces a highly reactive system when NO con-
centrations are high, but, in contrast to the smog chamber data, it also produces
extremely reactive products when NO concentrations are low. On the premise that
TO2 radicals do not have long tropospheric lifetimes and, therefore, must be lost by
an alternate reaction pathway when NO levels decrease, we considered possible
additional reaction pathways for the TO2 radical. The reaction we now utilize,
TO2 (+ O2) * HO2 + CRES (A 147)
does not involve the oxidation of NO to NO2 and produces a less reactive organic
surrogate product than (A 145). The mechanism is based on a discussion by Atkinson
and Lloyd [1984], combined with assumptions needed to simulate virtually all avail-
able toluene smog chamber data; however, it is still only an empirical representation
of an unknown process. The actual mechanism must be confirmed by laboratory
investigation under conditions more carefully controlled than those currently avail-
able.
Isoprene
Isoprene (ISOP) chemistry is included in the CBM-EX to provide an organic surrogate
exclusively for biogenic hydrocarbons. Our explicit isoprene mechanism is an update
of the mechanism developed by Killus and Whitten [1984]. In the CBM-EX isoprene
undergoes reaction with O, OH, O^ and NO? to form methylvinyl ketone (MVK),
methacrolein (MVK), and formaldehyde as the major products.
Evaluation of The CBM-EX
The performance of the CBM-EX was evaluated by simulating experimental data
from the University of North Carolina (UNO dual outdoor chambers [Jeffries et al.,
1982, 1985a, b], the University of California at Riverside evacuable chamber (UCR-
EC) [Pitts et al., 1979] and the UCR indoor Teflon chamber (UCR-ITC) [Carter et al.,
1985]. It is significant that, in this and most earlier studies, information from these
experiments was also used to develop chamber-dependent aspects of the
mechanisms. Because of this dual role for the experimental data, care must be taken
to delineate reactor-specific processes so that reactions depicting these processes
144
-------
can be included in the mechanism when simulating chamber data, but removed when
simulating the atmosphere. Ideally, such a procedure would prevent the uninten-
tional inclusion of reactor-specific characteristics in the mechanism. In reality,
however, the actual processes and magnitudes of the reactor-specific effects are not
well understood. Attempts to elucidate these processes and the uncertainty asso-
ciated with them have been discussed in detail elsewhere [Carter et al., 1982?
Lurmann et al., 1986; Gery et al., 1988] and only a brief summary is provided here.
Wall-loss rates of many species in smog chamber experiments are either unknown or
relatively low and we generally assume decay rates of zero. In the case of ozone,
however, decay rates have been measured, and we included this loss reaction in the
smoe chamber simulations with K = 1.8x10 j^ min for the UNC chambers, 1.4*10
min for the UCR-ITC, and 1.8xlO~3 min"1 for the UCR-EC. It should be noted,
however, that the decay rates for some oxidized species (i.e., HNO-j and r^C^) may
be large under certain conditions and in certain chambers.
Average dilution rates are provided for each experiment in the UCR-EC. This
chamber is diluted with clean air at rates that typically vary between 3x10 and
4x10 min . Because the UCR-ITC is a collapsible bag, no dilution is expected to
occur. At UNC the dilution rate is not reported, but the concentrations of tracer
species (chlorinated hydrocarbons) often are. These tracers can be used to calculate
dilution rates for the chambers, which typically vary from approximately 5x10
min~| at the start °* an exPer^ment>to 2x10 min after about 4 hours, and 3x10
min" after 8 hours. Replacement air entering the UNC chamber is from a relatively
clean rural environment. Key among the species entering the chamber at the
prescribed dilution rate are ozone (about 0.07 ppm over the course of the day), CO
(about 0.3 ppm), and nonmethane hydrocarbons (usually less than 0.05 ppm).
There are three complicated processes that may account for the chamber-related
formation of hydroxyl and other radicals: (1) emission of aldehydes from chamber
walls, (2) emission of nitrous acid from walls, and (3) heterogeneous conversion of gas
phase NC>2 to gas phase nitrous acid on the walls. Radicals result when the aldehydes
and HONO liberated in these processes photolyze to hydroxyl and other radicals. For
the UNC chambers, we generally use the reactions
NOx(wall) * HONO
FORM(wall) * FORM
with a HONO formationrate of 2.5x10 ppm min" and a formaldehyde off-gassing
rate of 5.8x10 e~ ' min , which is dependent on the formaldehyde mass
assumed to be on the chamber walls. The initial loading of this wall formaldehyde in
the UNC chambers does not usually exceed the equivalent of 40 ppb gas phase. The
third process, heterogeneous conversion of NO£ to HONO, was not included because
it appears to occur substantially only on Pyrex or quartz surfaces. For the UCR-ITC,
680t9r2 It
145
-------
a HONO formation rate of 3x10 ppm min and a formaldehyde off-gassing rate of
4x10 ppm min" were used.
The description of radical formation used for the UCR-EC chamber was taken from
the recent study of Lurmann et al. [1986], where wall off-gassing and photolysis were
combined to provide an OH source strength (Rj) for the two HONO-forming pro-
cesses. This radical flux is represented by
Rf = 0^(0.39 + 1.37[NO2]) ppb min"1
with an adjustable uncertainty parameter (a), which varies from 0.5 to 1.5 times the
NO2 photolysis rate (jj). The aldehyde emission rate of 3.0x10" ppm min" that was
included in the study by Lurmann et al. [1986] was also used in our simulations.
Initial gas phase HONO concentrations were not measured in any of the experiments
simulated in this work. Although HONO can affect the predicted oxidation rate
during the initial 30 to 60 min of a simulation, little effect is seen after that period
because other reactions are producing OH radicals. Hence, we selected initial HONO
conditions to match the reactivity of each experiment during the initial period. The
concentrations used never exceeded 2% of the initial gas phase NOX concentration
(as recommended by Lurmann et al. [1986]) and typically resulted in initial HONO
concentrations of only 1 to 10 ppb.
Besides the heterogeneous chemical processes, it is necessary to know the amount
and intensity of ultraviolet radiation within these chambers so that the rates of pho-
tolytic reactions can be accurately determined. This, like the chamber-dependent
processes, is an uncertain area that deserves further experimental investigation. The
UCR-EC uses an artificial light source and provides an average jj and a relative
spectral distribution for each experiment. For the UNC chambers, outdoor values of
integrated ultraviolet intensity are provided every minute. A function is used to
derive in-chamber spectral distributions from known characteristics of the solar
spectrum, chamber geometry and absorption by the Teflon film that covers the
chambers. These procedures were developed for both chambers over a decade ago
and are currently undergoing improvement to provide better resolution in the middle-
ultraviolet spectral region, where aldehyde photolysis and ozone photolysis are
important.
Our methodology for evaluating the CBM-EX is based on the hierarchical relation-
ships among species depicted in Figure 1. We develop and evaluate a mechanism by
starting at the lowest (simplest) level and proceed in a stepwise fashion to higher
(more complex) levels. When acceptable agreement between simulation results and
measurements is obtained at a lower level, no changes in rate constants or reaction
stoichiometry are allowed to the portions of the mechanism that describe that
chemistry. Simulations then proceed to the more complex organic species (higher
levels). If disagreement between predictions and measurements is found at a higher
level and all lower species have been successfully simulated, it is likely that the new
chemistry for the more complex species is in error.
146
-------
For mechanism evaluations using smog chamber data, the first testing levels in each
chamber were experiments with initial chamber loadings of (1) background air, (2)
background air plus NOX, (3) background air plus NOX and CO, and (<0 background air
plus NOX and formaldehyde. During the simulation of these systems, we formulated
mechanistic representations for the smog chamber processes (wall effects, photolytic
factors, dilution, and background conditions). After these processes were success-
fully simulated, experiments of increasingly complex organic species were tested,
including the higher molecular weight carbonyl species and the explicit and surrogate
categories shown in Figure 1. Only after the mechanisms for all species were evalu-
ated in this hierarchical manner were evaluations of the entire mechanism performed
using smog chamber experiments of complex organic mixtures.
The CBM-EX was evaluated using experimental data sets from the UNC and UCR
facilities. Table 2 provides a brief summary of the initial conditions and radical
source strengths used to simulate a subset of 21 experiments. These experiments
were selected to demonstrate CBM-EX performance for a variety of individual
organic species as well as for simple and complex organic mixtures. These same
experiments will later be used to demonstrate the condensation of the CBM-EX to
the CBM-IV. In the testing of any mechanism as large as the CBM-EX, performance
should be evaluated against as much time-resolved concentration data as are avail-
able. Data from both the UNC and UCR facilities typically include concentration-
time profiles for NO, NO2, O^, PAN, CO, and various organics. Although evaluations
were made for many individual species during the testing of the CBM-EX, we have
chosen to focus on a comparison of predicted and experimental results for three
parameters: (1) maximum ozone concentration, (2) maximum formaldehyde concen-
tration and (3) time to the maximum ozone concentration. These comparisons are
shown in Table 3. We note that the CBM-EX predictions show no strong bias in this
subset of experiments and that reasonable agreement is obtained for both of the
ozone parameters and for the formaldehyde concentrations.
FORMULATION AND TESTING OF THE
CONDENSED MECHANISM (CBM-IV)
The CBM-EX contains too many reactions and species to be practical for use in large
AQSMs. In this section, we describe the procedures that were followed to condense
the CBM-EX to a form that would minimize computer solution time without com-
promising predictive capabilities. Four types of techniques were used in this conden-
sation: (1) elimination of unimportant reactions and products, (2) creation of a
universal peroxy radical, (3) mathematical and algebraic manipulations to limit the
number of reactions, and (4) lumping of secondary reaction products.
The application of those techniques to each specific surrogate group is described by
Gery et al. [1988] and summarized in Appendix B. However, examples of each tech-
nique and discussion of the most important applications are included next.
8flO*9r2
-------
n
43
CO
•H
E
CO
0)
I
n
•o
§
CM
2
§
H
O
§
CO
(d
cc
£
So.
o.
o a
O .0
8*
is
xe
§5.
Q.
£8.
" D.
43 ^
C 43
IS
—H ^4
!_ CM
X C
u v
•o
o o>
§5
t. X
o in
in
in ir\
^- CM I CM I m CM
oS 10 too jo j o
d d ' d ' d d
in in in in
CM I CM CM CM
odd
in in
i CM i CM
i o i o
1 d ' d
i in i in i i in i in i 110101
i • i -ii • i -i ii • i • i
i o i o i i •- i o i i i »- i «- i
•- in «- in in
oo o
en en
ft 3f CO
a- ro CM
ooooooooooo oooooo
^^5 ^5 CO t^* CO
O^ ^^ ^5 (^? *™*
NO
O
O«-«-t\ir— ar^-
in in
CM
*
o
m
»- ro «- m cv
ouceoceouosocecj uuceoceo
9 CO
CM «-
o co
»- CM
§CM CM ON9 VO OO9 Sf 9 3 3
co sf co in c~- co m co NO co co co o co in co co co co CM
USCJJDOCJOOJtJJi-JZOSZCJf1"' S E^ U CJ
II I I
ce ce
B B
X
TH 43 43
•*3 EC C
0) 5 5
X CO Co O
& O'SO'XV&.B
SQ) QJ 0) —4 ^-( CP ^-t 3 O
CQ)0)CC43C43 43*vO
OOOt.ll4343t.t.O)IOOII>i >> >> CO CO
£
-------
0]
§
n
•t-i
u
(0
a
CO
u
03
g
0)
4)
o
s
*
o
is
•H -H
£- «M
0) —(
cx ^^
X C
tt 0)
CQ t*
O 3
O *i
i, X
•D 1-1
1 ^ O ^™ G> O CO CO *""" C\ CV G) ^3 ^? C^ ^~ G*
i dddddddddd oddddd
co
co
•
o
vO
d
co
•
o
I •- O «- O O CO ON
oo ^^ i co t^^ ^o ^^r ^^J
«- CO I O O O *- O
CO
co
oo
in
IOOOOOOOOOO IOOOOO
^^ *"" OJ C5
«- in «- co
co
in
co in co
rw in CM
O OJ
co in
co CM sr
CO
o
in
CJ
^-
\o
o in
^ co
in
t-
0
o
in
i
in
m
in
oo
in
in
co
S
coCT»cvcotn«-r-c*oint>- corr co ^ co o
ddddddddoodd oddddd
oo
•
o
oo
•
o
o
•
o
o
CO
> -
«— voooin
s
o'dddddd'-oooo oooodd
d
ffl QQ CH QQ 03 OS 09CQOQ OS OS
a- f\) ru CTN a- vo o o .=r ar ar
flfi flrt ^f rn \f\ p^ ffo f^ ffo v^> ^^ gQ f^ ^5 GO in *^v
8=>U3UUUo3u3 J Z O 2 0 H
^S CxJ ^S u3 ^j ^j (u ^^ tu ^^ *^J ^^ uu ^O Cu CO
i i i i i t i i i i i i i i i i i
uuccuosucjeuceu UUOSCJBU
p^ *^ r *» -^ r \ "^ *^ r *s «^ r > "^ ^ ^» r ^ ^» p ^ ^»
CO
co
03
O
in
d
t—
d
a
co
CO
rsj
CO
•
o
C\J
«•
C\J
0>
0} S 0)
33 C
x o
o>
(M
C.
J-
e
X CO
-* JG
O 8 o x o B
0) 0) 0) 0) i1^ *^ Q) **H
Q>Q)CCO)0)CC^C-t> -U C
Or Q/ Q/ CM Or C
*O XI "O C C S ^vv/fcMWwwiuh^'w ** **
^HrHi-HO(Q-UO) OJCCOJOJOJrtOQOJ
Afc. ca)0)^^j=jQj=^£
B . * i » J^ «1 Ml At rtl ^ #^ rt .^ -4 ^ «w *k t\ t L* -1 it
.
O
D 0) O03B*f_ — — _ — . . . -
OUL.I I *J ±> S- t. ffj I O O I I ^~
"B
149
-------
The first procedure was used to eliminate CBM-EX reactions (A42) - (A45), which
describe the oxidation of formaldehyde by HO2 radicals. This oxidation is an equi-
librium process and the rate of the reverse reaction (A43) is so rapid that (A42) -
(A45) are unimportant under most conditions. Methoxy radical (MEO) reactions were
also eliminated since the primary fate of this radical is reaction with O2 to yield
FORM + HO2 (reaction (A59). The alternative sinks for MEO ((A56) - (A58)) form
products that either regenerate the initial reactants or yield the products of (A59).
Hence, in the CBM-IV, MEO has been replaced by FORM + HO2« Several products
that do not appear as reactants in the CBM-EX were also eliminated in the condensa-
tion. These species included H2, CO2, several oxygenated organics (for example,
PROX, FACD, ACAC, and EPOX), and some organic nitrates (for example, MEN^,
NTR, and PBZN).
In the second condensation technique, a universal peroxy radical was created to
eliminate many of the organic peroxy radicals in the CBM-EX. These radicals all
react with NO to produce NO2 and a variety of organic products. In the CBM-IV,
most of these peroxy radicals have been eliminated by substituting the specific
organic products that they form during the NO-to-NO2 conversion and adding an
operator that converts an NO to NO2. For example the reaction sequence
AONE + OH * ANO2
ANO2 f NO * NO2 + C2O3 •*• FORM
is represented in the CBM-IV by
AONE + OH * XO2 + C2O3 + FORM
and the universal peroxy radical (XO2) oxidizes a molecule of NO:
XO2
This approximation was used to eliminate MEO2, the RO2 and RO2R radicals pro-
duced during the oxidation of PAR, and several other peroxy radicals. The self-
reaction (R80) of XO2 is used to represent peroxy-peroxy radical reactions, which
become important at low NO levels. Another operator (XO2N) is used to describe
nitrate formation resulting from the addition of NO to organic peroxy radicals.
Mathematical manipulation, the third condensation technique, was used to combine
several reactions into one. For example, the four O^ + OLE reactions in the CBM-
EX ((A98) - (A101)) have been combined into one reaction by using nonunity
stoichiometric coefficients for the products. Similar combinations were made for
multiple reactions of PAR, TOL, and XYL.
8BOt9r2
-------
The fourth procedure (lumping of secondary products) was used most extensively to
condense the CBM-EX isoprene chemistry. The condensation involved reducing the
complex reaction scheme to a few reactions whose products would be surrogate
species (such as FORM, ALD2, and OLE) that were already included in the CBM-IV.
The alkene bonds of the largest product species, methacrolein and methylvinyl
ketone, were simulated using the combination of ETH and OLE that was found to
give the proper reactivity with OH and O^. This representation also provides a high
yield of formaldehyde from the secondary oxidation of ETH and OLE, resembling the
formaldehyde yield of isoprene. The formation of PAN and PAN-like compounds was
simulated by assuming that both ALD2 and C2O? are formed during isoprene oxida-
tion, and the radical yield from the larger products was simulated by the formation
(and photolysis) of MGLY. The product yields were adjusted to give the best fits to
the midrange of isoprene-to-NOx ratio experiments where double ozone peaks were
observed. The double peaks in these experiments are caused by ozone reaction with
isoprene and isoprene products and by continued ozone formation due to PAN decom-
position at elevated temperatures. Successful simulation of the double peaks pro-
vided additional demonstration that a reasonable balance was achieved for the mul-
tiple processes occurring in isoprene oxidation.
The CBM-IV is listed in Table *f and the species contained in the mechanism are given
in Table 5. Notes related to this mechanism can be found in Appendix A. To demon-
strate the validity of the condensation for known conditions, we show in Table 6 the
results obtained when the CBM-IV was used to simulate the experiments listed in
Table 2. The results can be compared with the results for the detailed mechanism
given in Table 3. There appears to be no specific bias included in the CBM-IV for any
species or surrogate group. The results for reaction schemes requiring extensive
condensation, such as the aromatic and isoprene chemistries, are particularly
gratifying.
EVALUATION OF THE CBM-IV
Approximately 170 experimental data sets from the UNC and UCR smog chamber
facilities were simulated to evaluate the performance of the CBM-IV. These
experiments were selected because they represent the best available data, in terms
of light, chamber dryness, number of species monitored, data density and quality.
The data consisted of NOx-air irradiations of individual organic compounds as well as
a number of simple and complex organic mixtures. The individual organics simulated
were aldehydes (including formaldehyde), alkenes (ethene and larger alkenes), aro-
matics (toluene and xylenes), and biogenic hydrocarbons (o-pinene and isoprene).
Although simulation results for a few alkane/NOx experiments are shown in Tables 3
and 6, we did not extend that analysis to include an evaluation of the remaining UNC
and UCR alkane/NOx data base. This is because there are only a few alkane data
sets (mostly butane) and most of those experiments were performed at extremely
high hydrocarbon/NOx ratios to "force" the relatively unreactive alkanes to produce
measurable levels of ozone. Under these conditions, the chemistry is highly sensitive
880»»9r2
-------
TABLE 4. The Carbon Bond Mechanism IV
Reaction
No.
(R1)
(R2)
(R3)
(R4)
(R5)
(R6)
(R7)
(R8)
(R9)
(R10)
(R11)
(R12)
(R13)
(R14)
(R15)
(R16)
(R17)
(R18)
(R19)
(R20)
(R21)
(R22)
(R23)
(R24)
(R25)
Reaction
Inorganic
N02 * hv * NO + 0
02,M
0 03
03 + NO - N02
0 + N02 * NO
M
0 + N02 - N03
M
0* NO » N02
03 + N02 - N03
03 + hv * 0
03 + hv * 01D
M
01D •> 0
<
O^D + H20 + 2 OH
03 + OH * H02
03 + H02 * OH
N03 + hv * 0.89 N02 +
0.89 0 + 0.11 NO
N03 + NO * 2 N02
N03 + N02 * NO + N02
M
N03-»- N02 •> N205
N205 + H20 - 2 HN03
M
OP
NO + NO — £+ 2 N02
NO + N02 -t- H20 * 2 HONO
M
OH * NO HONO
HONO + hv * OH + NO
OH + HONO * N02
HONO+ HONO * NO + N0?
Rate Constant (k),
car molecule" s
Reactions
radiation dependent
1 ll v 1<"3 .» ' 175/T
i .H x iu e
1.8x 10-12 e"1370/T
9.3 x 10'12
1.6 i 10-13e68^/T
2.2 x ID'" e^2/T
m'^ -2450/T
radiation dependent
radiation dependent
1.9x 108e390/T
1 n
2.2 x 10'1U
12 e-940/T
1.4 x 10'14 e-580/T
radiation dependent
4 o v in~1^ -250/T
i .3 x iu e
2.?xio e
5.3 x 10-" e256/T
1.3 x 10"21
3.5 x 10111 e-1o897/T
1.8 x 10-20 e530/T
4.4 x 10-40
4.5 x 10-" e8o6/T
radiation dependent
6.6 x 10-12
1.0 x 10'20
Notes*
1
2
3,38
3,38
2
2
3,38
1,38
1,38
4
3
3,38
3,38
1,38
3
5,38
2
5
3
5
6
2
1
5,38
6,38
152
880>»9r2 11
-------
TABLE 4. (continued)
Reaction
No.
(R26)
(R27)
(R28)
(R29)
(R30)
(R3D
(R32)
(R33)
(R34)
(R35)
(R36)
Reaction
M
OH + N02 * HN03
M
OH + HN03 — N03
H02 + NO * OH + N02
M
H02 + N02 •> PNA
M
PNA * H02 * N02
OH + PNA * N02
H02 * H02 * H202
H02 + H02 + H20 * H202
H202 * hv * 2 OH
OH + H202 - H02
°2
OH + CO H02
Rate Constant (k),
cnr molecule' s~
1.0i ID'12 e713/T
5.1 x 1CT15 e1000/T
3.7 x 10-12 e^0/T
1.2x 10-13e749/T
4.8 x 1013 e-10121/T
1.3x 10-12e380/T
5.9 x ID'1* e1^0/T
2.2 x 10-38 e5800/T
radiation dependent
3.1 x 10-12e-l87/T
2.2 x 10"13
Notes*
2
7,38
3
2
3
3,38
5,38
5,38
1
3,38
3,38
Formaldehyde Reactions
(R37)
(R38)
(R39)
(R40)
(R41)
(R42)
(R43)
(R44)
°2
FORM + OH * H02 + CO
202
FORK + hv 2 H02 + CO
FORM + hv * CO
FORM + 0 * OH + H02 + CO
Op
FORM + N03 — ^* HN03 * H02 + CO
Higher Molecular Weight
°2
ALD2 + 0 C203 * OH
ALD2 + OH * C203
°2
ALD2 + N03 * C203 + HN03
1.0 x 10"11
radiation dependent
radiation dependent
3.0 x 10-11 e-1550/T
6.3 x 10'16
Aldehyde Reactions
1.2 x ID'11 e-986/T
7.0 x 10-12 e250/T
2.5 x 10'15
3,38
1
1,38
3
8
5
11,38
12
202
153
880>f9r2 11
-------
TABLE 4. (continued)
Reaction Rate Constant (k),
No. Reaction car molecule" s Notes
(R45) ALD2 + hv * FORM + radiation dependent 1
X02 + CO + 2 H02
°2
(R46) C203 + NO * FORM + 5.4 x 10"12 e250/T 13,38
X02 + H02 + N02
(R47) CP07 + N09 * PAN 8.0 x 10'20 e5500/T 13
(R48) PAH* C203 * N02 9.4 x 1016 e'14000/1 5
(R49) C203 + C203 * 2 FORM * 2.x 10~12 15
2 X02 + 2 H02
(R50) C203 + H02 * 0.79 FORM +
0.79 X02 + 0.79 H02 + 0.79 OH 6.5 x 10~12 15
Alkane Reactions
(R51) OH * FORM + X02 + H02 1.1 x 102 e'1710/T 18
(R52) PAR + OH * 0.87 X02 + 0.13 X02N + 8.1 x 10"13 17
0.11 H02 + 0.11 ALD2 + 0.76 ROR
- 0.11 PAR
(R53) ROR * 1.10 ALD2 + 0.96 X02 + 1.0 x 1015 e'8000/T 17
0.94 H02 + 0.04 X02N + 0.02 ROR
- 2.10 PAR
(R54) ROR * H02 1.6 x 103 17,38
(R55) ROR + N02 * 1.5 x 10"11 17,38
Alkene Reactions
(R56) 0 + OLE * 0.63 ALD2 + 1.2 x 10"11 e'324/T 5
0.38 H02 + 0.28 X02 + 0.30 CO +
0.20 FORM + 0.02 X02N + 0.22 PAR
+ 0.20 OH
(R57) OH + OLE * FORM + ALD2 * 5.2 x 10'12 e504/T 19
X02 + H02 - PAR
(R58) 03 + OLE * 0.50 ALD2 «• 1.4 x 10'14 e"2105/T 20,38
0.74 FORM + 0.33 CO + 0.44 H02 +
0.22 X02 + 0.10 OH - PAR
(R59) N03 + OLE * 0.91 X02 + 7.7 x 10~15 21
FORM + ALD2 +0.09 X02N + N02
(R60) 0 + ETH * FORM + 0.70 X02 + 1.0 x 10~11 e'792/T 22
CO + 1.70 H02 + 0.30 OH
880H9r2
154
-------
TABLE 4. (continued)
Reaction Rate Constant (k),
No. Reaction cm3 molecule" s
(R61)
(R62)
(R63)
(R64)
(R65)
(R66)
(R67)
(R68)
(R69)
(R70)
(R7D
(R72)
(R73)
(R74)
(R75)
(R76)
(R77)
OH + ETH * X02 + 1.56 FORM + 2.0 x 10'12 e4l1/T
H02 +0.22 ALD2
03 + ETH - FORM + 0.42 CO + 1.3 x 10'14 e'2633/T
0.12 H02
Aromatic Reactions
OH + TOL - 0.08 X02 +
0.36 CRES + 0.44 H02 + 0.56 T02 2.1 x 10"12 e322/T
T02 + NO * 0.90 N02 + 8.1 x 10'12
0.90 OPEN + 0.90 H02
T02 - H02 + CRES 4.2
OH + CRES * 0.40 CRO + 0.60 4.1 x 10"11
X02 + 0.60 H02 + 0.30 OPEN
N03 + CRES * CRO + HN03 2.2 x 10~11
CRO + N02 * 1.4 x 10~11
OH + XYL * 0.70 H02 + 1.7 x 10"11 e1l6/T
0.50 X02 + 0.20 CRES + 0.80 MGLY +
1.10 PAR + 0.30 T02
OH + OPEN * X02 + C203 + 3.0 x 10'11
2 H02 + 2 CO + FORM
OPEN + hv * C203 + CO radiation dependent
+ H02
03 + OPEN * 0.03 ALD2 + 5.4 x 10'17 e'500/T
0.62 C203 + C.70 FORM + 0.03 X02
+ 0.69 CO + 0.08 OH + 0.76 H02 +
0.20 MGLY
OH + MGLY * X02 + C203 1.7 x 10'11
MGLY + hv * C203 + CO + H02 radiation dependent
Isoprene Reactions
0 + ISOP * 0.60 H02 + 0.80 ALD2 1.8 x 10~11
+ 0.55 OLE + 0.50 X02 + 0.50 CO
+ 0.45 ETH + 0.90 PAR
OH + ISOP * FORM + X02 + 9.6 x 10"11
0.67 H02 + 0.40 MGLY + 0.20 C203
+ ETH + 0.20 ALD2 + 0.13 X02N
03 + ISOP * FORM* 0.40 ALD2
+ 0.55 ETH + 0.20 MGLY + 0.06 CO +
0.10 PAR + 0.44 H02 + 0.10 OH 1.2 x 10"17
*
Notes
23
20,38
26,38
29,38
29
19
12
27,38
26
31
1,31
31
28
1,31
32,39
33,39
33,39
34,39
34,39
880i*9r2 11 -,,.[.
-------
TABLE 4. (concluded)
Reaction Rate Constant (k),
No. Reaction cm3 molecule* s Notes
(R78) N03 + ISOP * X02N 3.2 x 10~13 35,39
Operator Reactions
(R79) X02 + NO * N02 " 8.1 x 10'12 40
(R80) X02 + X02 * 1.7 x 10'14 e1300/T 38,40
(R81) X02N * NO * 6.8 x 10~13 38,40
* See Appendix A for notes.
156
-------
TABLE 5. Chemical Species in the CBM-IV
Species Name Representation
Nitric oxide NO
Nitrogen dioxide N02
Nitrogen trioxide (nitrate radical) N03
Dinitrogen pentoxide N205
Nitrous acid MONO
Nitric acid HN03
Peroxynitric acid (H02N02) PNA
Oxygen atom (singlet) 01D
Oxygen atom (triplet) 0
Hydroxyl radical OH
Water H20
Ozone 03
Hydroperoxy radical H02
Hydrogen peroxide H202
Carbon monoxide CO
Formaldehyde (CH2=0) FORM
High molecular weight aldehydes (RCHO, R>H) ALD2
Peroxyacyl radical (CH3C(0)00*) C203
Peroxyacyl nitrate (CH,C(0)OON02) PAN
Paraffin carbon bond (C-C) PAR
Secondary organic oxy radical ROR
Olefinic carbon bond (C=C) OLE
Ethene (CH2=CH2) ETH
Toluene (C6H5-CH3) TOL
Cresol and higher molecular weight phenols CRES
Toluene-hydroxyl radical adduct T02
Methylphenoxy radical CRO
High molecular weight aromatic oxidation
ring fragment OPEN
Xylene (C6H4-(CH3)2) XYL
Methylglyoxal (CH3C(0)C(0)H) MGLY
Isoprene ISOP
NO-to-N02 operation X02
NO-to-nitrate operation X02N
Total 33
880H9r2 9
157
-------
•o
§
u
0)
5
B
n
c
(*
i
n
o
•^
4J
2 •
4)
Ou
0
e
*>
§
O>
§
n
a
"8
0>
"8
c -u
4) (Q
B O
t. C_
8.3
& S
•o
O 0)
c &
SH
&4 09
O 3
O -U
u X
•O «H
i »^o«-^oomco«-cMmoooo»-o
i oooddddddd dddddd
co
CM
oo
•
o
co
•
o
I O\ Sf ^"
I «- O «-
ocouti
so
5«
IOOOOOOOOOOIOOOOO
s s
• •
o o
ininmNO.J.tlina.tla-in
CT* co ro ro
CM in CM in
rorg
in
o
in
n")
inso
OCM in roc^
^S^o^^^tCm
vlVin'-inONOt7i
. incvj .a- in
a-v£> o
mm in
,
o
.- ,- "I «> «-
mo>cMO\in»-vocr>oin^ ma- m t- a- o
ddddddddcodd dodoo'o
in
m T
CO O
I
O =t
o
CO
oe-soaroeocoav
ma\r\iosa-«— v£>o
xo
mo
I
in
o
ooooooo«-cooo oooooo
d
CB CQ CS.CQCQ BE CD CD CQ CC OS
»• (M CM cjNa1 NO OOar s; a1
cocoa-coine-comcoNOcococoocoinco
o\a'inaroc^ina>m»-vo ini—a-t^a-m
ocMCMCMmoo'-»-CM'- csjcumcMmo
uoeuccoocecjcco uuccuccu
m
oo
CD
S
S
in
o
CO
s
§
0\
•
o
0)
•S -o •§ "S
>> fe >» 2 m
Ain)n){rta)C Q)C) 0)V *^ ~" **^ C) •*•!
T3 T> T3 C C 3 QJO)CCa»a»CC4JC-0 *>
33
x
-------
to chamber radical and NO)C emissions and an unambiguous evaluation of alkane
chemistry is not possible. Therefore, we utilized complex mixture experiments with
high alkane loadings to demonstrate the performance of the CBM-IV alkane
chemistry.
In a more extended analysis [Gery et al., 1988] we documented the ability of the
CBM-IV to simulate timing and reactivity by tabulating and plotting results for each
experiment for a large number of species. Our demonstrations in this paper focus on
a comparison of predicted and measured maximum concentrations for ozone and
formaldehyde. These comparisons are shown in Figures 2 through 7. Simulated and
observed maximum ozone and formaldehyde concentrations are compared in Figures
2 and 3 for the single-organic-component experiments containing aldehydes and
alkenes. Similar scatterplots for the aromatic and biogenic species are found in
Figures
-------
4)
s
N
o
•o
41
T5
U
0.00 020 0.40 0.60
1.00 120 1.40
Measured Ozone (ppm)
Fig. 2 Comparison of CBM-IV predicted and measured
maximum ozone concentrations from hydrocarbon/He^
smog chamber experiments for formaldehyde, higher
aldehyde, ethene and. alkene systems. Error bounds
are 25 percent.
160
-------
L.
I
V
15
U
D
1
•
ALD2
D
ETH
100 020 0.40 * 0.60 0.80 IjOO
Meosired FORM (ppm)
Fig. 3. Comparison of CBM-IV predicted and measured
maximum formaldehyde concentrations from hydrocarbon/NOx
smog chamber experiments for higher aldehyde, ethene and
alkene systems. Error bounds are 25 percent.
161
-------
0.00 020 0.40 0.60
1JOO
Measured Ozone (ppm)
Fig. A. Comparison of CBM-IV predicted and measured
maximum ozone concentrations from hydrocarbon/NOx smog
chamber experiments for aromatic and biogenic systems.
Error bounds are 25 percent.
162
-------
cr
0
U.
o
U
Msoared FORM (ppm)
Fig. 5a. Comparison of CBM-IV predicted and measured
maximum formaldehyde concentrations from hydrocarbon/NO
smog chamber experiments for aromatics and biogenic systems.
Error bounds are 25 percent. The area within the dotted
lines is shown in better detail in Fig. 5-b.
163
-------
o
U.
I
O
0.25
0.20
0.15
0.10
0.05
01
D
TOL
SOP
*•
o-PN
25% BOUhDS
1.00 0.05 0.10 0.15 020 025
Measured FORM (ppm)
Fig. 5b. Detail of area in Fig. 5a.
164
-------
s
N
o
•o
0
I
*
O
D
SYN
•
EC
B
fTC
*
REAC
+
OTffR
25%BOUMS
020 0.40 0.60 050 1.00
Measured Ozone (ppm)
Fig. 6. Comparison of CBM-IV predicted and measured
maximum ozone concentrations from different hydrocarbon-
mixture/NOx smog chamber experiments. Error bounds are
25 percent.
165
-------
o
LL
I
£
U
D
SYN
•
EC
E
(TC
*•
REAC
+
OTHER
25%BOUK)S
OJ
020 0.40 0.60 0.80 1.00 120
Meaared FORM (ppm)
Fig. 7a. Comparison of CBM-IV predicted and measured
maximum formaldehyde concentrations from various hydrocarbon-
mixture/NOjc smog chamber experiments. Error bounda are
25 percent. The area within the dotted lines is shown in
better detail in Fig. 7b.
166
-------
L.
I
u
D
SYN
•
EC
B
rrc
*>
REAC
+
OMR
25%BON)S
0.00 0.05 0.10 0.15 020 025 0.30
Meoared FORM (ppm)
Fig. 7b. Detail of are in Fig. 7a.
167
-------
day. This could have caused the delay in reactivity seen during the early morning
hours and the rapid acceleration in reactivity observed near midday.
For the UNC runs, the average maximum formaldehyde concentration was underpre-
dicted by 23% for the ethene experiments and by 31% for the larger alkenes. In
contrast, formaldehyde concentrations for the UCR-EC experiments were overpre-
dicted by a sizeable amount. These results suggest there are uncertainties
associated with either the formaldehyde data or the characterization of chamber
wall effects. More accurate formaldehyde chamber data and further clarification of
wall-related processes are needed before the reliability of the alkene mechanisms
can be assessed.
We simulated all usable toluene and xylene experiments from the UNC and UCR-EC
chambers, as well as a number of UNC organic mixture experiments in which
selected aromatic compounds were substituted for various fractions of the initial
organic mix. As noted earlier, these data were also used to develop the mechanisms
for the TOL and XYL surrogates. The predictions and measurements for ozone and
formaldehyde are shown in Figures *f and 5 for the individual aromatic hydrocarbon
experiments. For the toluene/NOx systems, the average maximum ozone concentra-
tion was overpredicted by *f ± 8% for the UNC runs and underpredicted by 5 ± 8% for
the UCR-EC runs. This is remarkable agreement, considering that the ozone predic-
tions obtained with an earlier version of the CBM for a subset of these experiments
were high by nearly a factor of 2. We attribute this improvement to the new product
chemistry of the toluene oxidation mechanism. An example of the goodness of fits
obtained is shown in Figure 8 for a typical UNC toluene/NOx experiment. Equally
good agreement was obtained for the eight m- and o-xylene experiments simulated in
this study. Maximum ozone was overpredicted by 4 ± 16% and formaldehyde was
high by 5 ± M%. Formaldehyde yields for the toluene experiments were overpre-
dicted (especially for the UCR-EC) by about 80 ± 130%, but the experimental
formaldehyde yields were always low and usually near the detection limit of the
measurement technique. An example of a typical m-xylene/NOx simulation from the
UCR-EC is shown in Figure 9. As the aromatic profiles show, the new mechanism is
able to simulate the timing and reactivity of these systems in two different smog
chambers.
The new condensed mechanism .for isoprene and the carbon bond surrogate speciation
for o-pinene were also tested against UNC chamber data. For the 12 isoprene exper-
iments simulated in this study, the average maximum ozone concentrations were
overpredicted by 6 ± 23%. We consider this good evidence that the condensed
mechanism successfully represents isoprene chemistry. Because isoprene produces
high yields of formaldehyde and PAN-like species, we attempted to develop a con-
densed isoprene representation that would predict these concentrations as accurately
as possible. For these experiments the average maximum PAN concentration was
overpredicted by 10 ± 22% and formaldehyde by 1 ± 18%. For simulations of the a-
pinene/NOx system, we selected a carbon bond representation of 0.5 OLE, 1.5 ALD2,
and 6.0 PAR. This representation maximized agreement between simulated and
-------
1
t
§
!!!!! 3 55 i
uicid
5 ! 3 2 S S I
169
-------
8
I
6?
(N
: ! ! S ! S i! 9 S *
uidd
2 ! 2 3 5 2 5 5 i !
-------
observed ozone concentrations and resulted in adequate PAN predictions. Ozone and
formaldehyde were overpredicted by about 20% for nine experiments, although
formaldehyde data upon which to base a comparison were sparse.
To demonstrate the applicability of the CBM-IV to a range of reactive hydrocarbon
mixtures, we simulated selected organic/NOx mixture experiments conducted in the
UNC and UCR smog chambers. These experimental data sets reflect attempts to
create organic mixtures representative of atmospheric conditions. The mixtures
simulated consisted of the UNC synthetic urban and automobile exhaust mixtures
(SYN), the UCR-EC seven-component urban surrogate (EC), the UCR-ITC multi-day
surrogate (ITC), and the UNC substitution experiments where varying amounts of
aromatics and other organics were substituted for the urban surrogate (REAC). A
more complete description of the characteristics of these mixtures and the condi-
tions used to simulate them is found in Gery et al. [1988]. We simulated 68 mixture
experiments that varied in hydrocarbon composition, initial hydrocarbon/NOx ratio,
and duration (the UCR-ITC experiments were multiday tests). Ozone and formal-
dehyde results for these experiments are shown in Figures 6 and 7. The average
ozone overprediction was 2% with a 22% standard deviation. The maximum formal-
dehyde concentration was underpredicted by 9 ± 34%. Given the uncertainties in
both chemistry and chamber characterization, we consider this agreement to be
especially good.
CONCLUSIONS
Within the evaluation methods available, we find the CBM-IV to be an improved and
well-performing mechanism. It provides new capabilities for simulating the complex
photochemistry of the urban and regional troposphere. However, there are a number
of uncertainties associated with the development and evaluation of this mechanism
that make assessment of its overall accuracy difficult. These uncertainties exist
because (0 the basic information used to develop the mechanism is sparse in certain
key areas, and (2) the smog chamber data used to test the mechanism contain defici-
encies. In general, more kinetic information for peroxy radical reactions would
provide a basis for an improved secondary product reaction scheme. Of particular
importance is the need for kinetic data at temperatures below room temperature.
The lack of this information for many peroxy radical reactions (such as PAN forma-
tion) renders application of kinetic mechanisms uncertain at low atmospheric
temperatures. In addition, agreement on formaldehyde absorption coefficients would
allow clearer focus on other formaldehyde-related issues. With regard to the
chemistry of specific surrogate groups, we are most concerned with uncertainties
that exist in the alkene and aromatic representations. The successful predictions of
the aromatic mechanism were gratifying, but since aromatic oxidation kinetics is
still poorly understood, we cannot be sure that the predictions are founded on the
actual chemistry. Therefore, the mechanism could display anomalies outside the
range of conditions under which it was tested. Also, although many of the basic
reaction kinetics for the alkene mechanisms have supposedly been identified, these
88049r2
-------
mechanisms often perform poorly in simulating smog chamber measurements.
Whether this is due to inconsistencies in the mechanism or uncertainties in the
description of the smog chamber environment should be investigated.
r.
Uncertainties associated with smog chamber experiments, particularly regarding the
effects of chamber surfaces and spectral distribution of in-chamber light, are cur-
rently being reviewed by a number of groups. It is imperative that those performing
the experiments and those simulating the data merge their findings and views so that
we can reduce uncertainties related to radical and NOX sources, heterogeneous
chemistry, and in-chamber light intensity. Such collaboration would not only enable
us to define key experiments for further mechanism testing, but should also provide
information that would allow the utilization of many experimental data sets now
considered marginal.
172
-------
References
Addison, M. C., 3. P. Burrows, R. A. Cox, and R. Patrick, Absorption spectrum and
kinetics of the acetylperoxy radical, Chem. Phys. Lett., 73, 283, 1980.
Atkinson, R., Kinetics and mechanisms of the gas-phase reactions of the hydroxyl
radical with organic compounds under atmospheric conditions, Chem. Rev., 86,
69, 1986.
Atkinson, R., and S. M. Aschmann, Rate constants for the reaction of OH radicals
with a series of alkenes and dialkenes at 295 + 1 K, Int. 3. Chem. Kinet., 16, 1175,
1984.
Atkinson, R., and W. P. L. Carter, Kinetics and mechanisms of the gas-phase
reactions of ozone with organic compounds under atmospheric conditions, Chem.
Rev.. 84, 437, 1984.
Atkinson, R., and A. C. Lloyd, Evaluation of kinetic and mechanistic data for
modeling of photochemical smog, 3. Phys. Chem. Ref. Data, 13, 315, 1984.
Atkinson, R., and 3. N. Pitts, 3r., Absolute rate constants for the reaction of CX P)
atoms with a series of olefins over the temperature range 298-439 K, 2. Phys.
Chem., 67, 38, 1977.
Atkinson, R., A. C. Lloyd, and L. Winges, An updated chemical mechanism for
hydrocarbon/NOx/SC>2 photooxidations suitable for inclusion in atmospheric
simulation models, Atmos. Environ., ^6, 1341, 1982.
Atkinson, R., S. M. Aschmann, A. M. Winer, and 3. N. Pitts, 3r., Kinetics of the gas-
phase reactions of NO^ radicals with a series of dialkenes, cycloalkenes and
monoterpenes at 295 ± 1 K, Environ. Sci. Technol., 18, 370, 1984.
Bandow, H., N. Washida, and H. Akimoto, Ring-cleavage reactions of aromatic
hydrocarbons studied by FT-IR spectroscopy, I, Photooxidation of toluene and
benzene in the NOx-air system, Bull. Chem. Soc. 3pn., 58, 2531, 1985.
Barnes, I., V. Bastian, K. H. Becker, E. H. Fink, and F. Zabel, Pressure dependence of
the reaction of OH with HO2NO2, Chem. Phys. Lett., 123, 28, 1985.
Bass, A., personal communication with Arnold M. Bass, Nat. Bureau of Stand.,
Gaithersburg, Maryland, 1985.
880>»9r2 15
173
-------
Bass, A. M., L. C. Glasgow, C. Miller, 3. P. 3esson, and D. L. Filkin, Temperature
dependent absorption cross sections for formaldehyde (Ch^O): The effect of
formaldehyde on stratospheric chlorine chemistry, Planet. Space Sci., 28, 675,
1980.
Baulch, D. L., R. A. Cox, R. F. Hampson, Jr., 3. A. Kerr, 3. Troe, and R. T. Watson,
Evaluated kinetic and photochemical data for atmospheric chemistry:
Supplement II, CODATA Task Group on gas phase chemical kinetics, 3. Phys.
Chem. Ref. Data, 13, 1259, 1984.
Calvert, 3. G., The homogeneous chemistry of formaldehyde generation and
destruction within the atmosphere, Proceedings of the NATO Advanced Study
Institute on Atmospheric Ozone, Off. Environ, and Energy, Fed. Aviat. Admin.,
Washington, D.C., 1980.
Cantrell, C. A., W. R. Stockwell, L. G. Anderson, K. L. Busarow, D. Perner, A.
Schmeltekopf, 3. G. Calvert, and H. S. Johnston, Kinetic study of the NO3-CH2O
reaction and its possible role in nighttime tropospheric chemistry, 3. Phys.
Chem., 89, 139, 1985.
Carter, W. P. L., R. Atkinson, A. M. Winer, and 3. N. Pitts, 3r., An experimental
investigation of chamber-dependent radical sources, Int. 3. Chem. Kinet., J^,
1071, 1982.
Carter, W. P. L., M. C. Dodd, W. D. Long, and R. Atkinson, Outdoor chamber study
to test multi-day effects, Rep. EPA-600/3-8
-------
Gery, M. W., D. L. Fox, H. E. Jeffries, L. Stockburger, and W. S. Weathers, A
continuous stirred tank reactor investigation of the gas-phase reaction of
hydroxyl radicals and toluene, Int. J. Chem. Kinet., ]^» 931, 1985.
Gery, M. W., D. L. Fox, R. M. Kamens, and L. Stockburger, Investigation of hydroxyl
radical reactions with o-xylene and m-xylene in a continuous stirred tank reactor,
Environ. Sci. Technol., 21^, 339, 1987.
Gery, M. W., G. Z. Whitten, and J. P. Killus, Development and testing of the CBM-IV
for urban and regional modeling, Rep. EPA-600/3-88-012, U.S. Environ. Prot.
Agency, Research Triangle Park, N.C., 1988.
Haagen-Smit, A. J., Chemistry and physiology of Los Angeles smog, Ind. Eng. Chem.,
W, 1342, 1952.
Jeffries, H. E., R. M. Kamens, K. G. Sexton, and A. A. Gerhardt, Outdoor smog
chamber experiments to test photochemical models, Rep. EPA-600/3-82-016, U.S.
Environ. Prot. Agency, Research Triangle Park, N.C., 1982.
Jeffries, H. E., K. G. Sexton, R. M. Kamens and M. S. Holleman, Outdoor smog
chamber experiments to test photochemical models: Phase II, Rep. EPA-600/3-
85-029, U.S. Environ. Prot. Agency, Research Triangle Park, N.C., 1985a.
Jeffries, H. E., K. G. Sexton, T. P. Morris, M. Jackson, R. G. Goodman, R. M.
Kamens, and M. S. Holleman, Outdoor smog chamber experiments using
automobile exhaust, Rep. EPA-600/3-85-032, U.S. Environ. Prot. Agency,
Research Triangle Park, N.C., 1985b.
Jeffries, H. E., K. G. Sexton, J. R. Arnold, and T. L. Kale, Validation testing of new
mechanisms with outdoor chamber data, vol. 3, Calculation of Photochemical
reaction photolysis rates in the UNC ourdoor chamber, report, U.S. Environ. Prot.
Agency, Research Triangle Park, N.C., 1989.
Kaiser, E. W., and C. H. Wu, A kinetic study of the gas phase formation and
decomposition reactions of nitrous acid, J. Phys. Chem., 81, 1701, 1977.
Kan, C. S., J. C. Calvert, and J. H. .Shaw, Reactive channels of the
reaction, report, Dep. Chem. and Dep. of Phys., Ohio State Univ., Columbus,
1980.
Kerr, J. A., and J. G. Calvert, Chemical transformation modules for Eulerian acid
deposition models, vol. I, The gas-phase chemistry, Rep. EPA-600/3-85-015, U.S.
Environ. Prot. Agency, Research Triangle Park, N.C., 1985.
880U9p2 15 175
-------
Killus, J. P., and G. Z. Whitten, A new carbon-bond mechanism for air quality
simulation modeling, Rep. EPA-600/3-82-041, U.S. Environ. Prot. Agency,
Research Triangle Park, N.C., 1982.
Killus, 3. P., and G. Z. Whitten, Isoprene: A photochemical kinetic mechanism,
Environ. Sci. Technol., UJ, 142, 1984.
Kleindienst, T. E., G. W. Harris, and 3. N. Pitts, 3r., Rates and temperature
dependencies of the reaction of OH with isoprene, its oxidation products, and
selected terpenes, Environ. Sci. Technol., L6, 844, 1982.
Leone, J. A., and 3. H. Seinfeld, Comparative analysis of chemical reaction
mechanisms for photochemical smog, Atmos. Environ., ^9, 437, 1985.
Lurmann, F. W., A. C. Lloyd, and R. Atkinson, A chemical mechanism for use in
long-range transport/acid deposition computer modeling, 3. Geophys. Res., 9_1,
10905, 1986.
Malko, M. W., and 3. Troe, Analysis of the unimolecular reaction N2O^ + M * NO2 +
NO3 + M, Int. 3. Chem. Kinet., 14, 399, 1982.
McAdam, K., B. Veyret, and R. Lesclaux, UV absorption spectra of HO2 and CH3O2
radicals and the kinetics of their mutual reactions at 298 K, Chem. Phys. Lett.,
133. 39, 1987.
Molina, L. T., and M. 3. Molina", Absolute absorption cross sections of ozone in the
185- to 350-nm wavelength range, 3. Geophys. Res., 9J_, 14501, 1986.
Moortgat, G. K., W. Klippel, K. H. Mobus, W. Seller, and P. Warneck, Laboratory
measurements of photolytic parameters for formaldehyde, Rep. FAA-AA-EE-80-
47, Off. of Environ, and Energy, Fed. Aviat. Admin., Washington, D.C., 1980.
Moortgat, G. K., W. Seller, and P. Warneck, Photodissociation of HCHO in air: CO
and H2 quantum yields at 220 and 300 K, 3. Chem. Phys., 78, 1185, 1983.
National Center for Atmospheric Research (NCAR), Development and
implementation of chemical mechanisms for the regional acid deposition model
(RADM), internal rep., U.S. Environ. Prot. Agency, Research Triangle Park, N.C.,
1987.
Niki, H., P. D. Maker, C. M. Savage, and L. P. Breitenbach, A FTIR study of a
transitory product in the gas-phase ozone-ethylene reaction, 3. Phys. Chem., 85,
1024, 1981.
880H9r2 15
-------
Niki, H., P. D. Maker, C. M. Savage, and L. P. Brietenbach, Fourier transform
infrared studies of the self-reaction of CH3O2 radicals, Phys. Chem., 85, 877,
1982.
Pitts, 3. N., 3r., K. Darnall, W. P. L. Carter, A. M. Winer, and R. Atkinson,
Mechanisms of photochemical reactions in urban air, Rep. EPA-600/3-79-110,
U.S. Environ. Prot. Agency, Research Triangle Park, N.C., 1979.
Shepson, P. B., E. O. Edney, and E. W. Corse, Ring fragmentation reactions on the
photooxidations of toluene and o-xyiene, 3. Phys. Chem., 88, 4122, 1984.
Smalley, J. F., F. L. Nesbitt, and R. B. Klemm, Branching ratio for the hydrogen
atom product channel in the reaction of ground-state atomic oxygen with
ethylene, 3. Phys. Chem., 90, 4-91, 1986.
Su, F., 3. G. Caivert, 3. H. Shaw, H. Niki, P. D. Maker, C. M. Savage, and L. D.
Breitenbach, Spectroscopic and kinetic studies of a new metastable species in the
photooxidation of gaseous formaldehyde, Chem. Phys. Lett., 65, 221, 1979a.
Su, F., 3. G. Caivert, and 3. H. Shaw, Mechanism of the photooxidation of gaseous
formaldehyde, 3. Phys. Chem., 83, 3185, 1979b.
Tuazon, E. C., H. MacLeod, R. Atkinson, and W. P. L. Carter, o-Dicarbonyl yields
from the NOx-air photooxidations of a series of aromatic hydrocarbons in air,
Environ. Sci. Technol., 20, 383, 1986.
Veyret, B., 3. Rayez, and R. Lesclaux, Mechanism of the photooxidation of
formaldehyde studied by flash photolysis of Cr^O-C^-NO mixtures, 3. Phys.
Chem., 86, 3424, 1982.
Whitten, G. Z., H. Hogo, and 3. P. Killus, The carbon-bond mechanism: A condensed
kinetic mechanism for photochemical smog, Environ. Sci. Technol., J.4, 690, 1980.
15
-------
APPENDIX A
Tables Al and A2 give the reactions and species in CBM-EX. Following are the notes
for Tables k and Al.
1. Variable photolysis rates depend on light intensity and spectral distribution (which
changes with angle of solar elevation). In smog chamber simulations, other factors,
such as Teflon absorption and reflection, and floor and exterior structure reflection,
also affect these rates. The molecular data referenced in Table A3 were used to
calculate some important rates for the simulation of smog chamber experiments.
Structural effects were estimated based on data from each laboratory [see Gary et
al., 1988; Jeffries et al., 1989], Photolysis rates for species with only partial
molecular data were estimated based on those data and relevant smog chamber
results [eery et al., 1988]. Rates designated as "times k^" are ratios to the photoly-
sis rate of reaction n (n usually equals 1 for NC>2 photolysis).
2. These pressure- and temperature-dependent rate constants are determined from
the following formula:
k0(T)[M]
iC — -Ji--—• _--__
• 0.6(1 +
where
mo
kQ(T) = kQ (T/300) u
m
kJT) = k^ (T/300) "
Atmospheric pressure is assumed in all calculations. See Table A4 for parameter
values.
3. Kinetic data from DeMore et al. [1985].
4. Combined kinetic expressions for N2 and ©2 [DeMore et al., 1985]:
6BO>»9r2 16
179
-------
O1D + N2-O + N2 k= 1.8 * 10'11 el07/T
0*0 + O2 + O + O2 k = 3.2x 10"Ue67/T
5. Atkinson and Lloyd [1984],
6. Kaiser and Wu [1977].
7. The pressure dependence and temperature dependence were fit by combining a
low-pressure (bimolecular) limit with a Lindemann-Hinshelwood expression for
pressure dependence [DeMore et al., 1985].
k =
where
k0 = 7.2 x 10"15 e785/1
k2 = 4.1 x ID'16 e1WO/
k, = 1.9 x 10"33 e725/T
8. Cantrell et al. [1985].
9. The reaction rate constants for (A42) and (A43) have been measured by su et al.
[1979a, b] as 1.0 x 10 * cm3 molecule" s"* and 1.5 s . The values were also
measured by Veyret et al. [1982] as 7.5 x 10 cm3 molecule"* s" and 30 s" .
Barnes et al. [1985] reported k^2 as 1.1 x 10 cm molecule" s" with no value
for kA
-------
14. Baulch et al. [1984].
15. Addison et al. [1980].
16. See text. Rates based on kA67 + kA6g of cox and Tyndail [1980].
17. Surrogate alkane chemistry is discussed in the text and in depth by Gery et al.
[1988].
if' U17T§/TA74 = kOH+CH'[CH]' with t°H = L85 PPm and
12 -i/io/ r et ale
19. Based on pressure and temperature dependence [Atkinson, 1986] described by
Gery et al. [1988].
20. Based on pressure and temperature dependence [Atkinson and Carter, 1984]
described by Gery et al. [1988].
21. Kinetic data for the reaction of NOj with the OLE group are given by Atkinson
and Lloyd [1984], but were calculated using the N2O^ equilibrium constant of Malko
and Troe [1982]. We now believe the use of this value, compared to the data
presented earlier in the CBM-EX, leads to an undercalculation of ^231 ^Y * factor of
1.83. Therefore, we have converted the rate suggested by Atkinson and Lloyd by
that factor to yield a+k for NO^ + OLE. A 9% nitrate yield is assumed for this
reaction.
22. Kinetic data from Atkinson and Lloyd [1984]. Product distributions from that
review and the more recent work of Smalley et al. [1986].
23. Kinetic data from Atkinson [1986]. Product distributions from that work based
on the finding of glycolaldehyde as a stable product [rriki et al., 1981].
Giycolaldehyde is represented as
24. Decomposition of intermediate formic and acetic acid products follows the frac-
tions proposed by Dodge and Arnts [1979].
25. Biradical reaction rates were taken from the review of Kerr and Calvert [1985],
ignoring the relatively slow reaction with
26. Kinetic expressions from Atkinson [1986]. Initial toluene product distributions
from Gery et al. [1985] and initial xylene product distributions from Gery et al.
[1987].
27. Kinetics and mechanism based on discussion and references in the work by Gery
et al. [1985].
880«f9P2 16
181
-------
28. Kinetic expression from Atkinson [1986].
29. Rates and ratios discussed in text and by eery et al. [1988].
30. Based on discussion and references in the work by Gery et al. [1987].
31. The chemistry of aromatic ring decomposition products is based on smog
chamber simulations as discussed by Gery et ai. [1988].
32. O atom rate assumed to be the sum at each of the alkene bonds of isoprene.
Kinetics from Atkinson and Pitts [1977] as explained by Gery et al. [1988].
33. Rate constant derived from average of values found in the works by Atkinson
and Aschaann [1984] and Kleindienst et al. [1982].
34. Rate constant is within the uncertainty calculated by Atkinson and Carter
[1984].
35. AtJtinson et al. [1984].
36. Killus and Whitten [1984].
37. AtJcinson and Carter [1984].
38. Untracked products in the CBM-IV (for example, CC>2» C>2» r^O, methanol,
ozonides, organic acids, and nitrates) are not included in the reaction listing. Con-
sult the CBM-EX listing (Table Al) for the complete products.
39. Condensed isoprene product distribution is discussed by Gery et al. [1988].
40. XO2 and XO2N are chemical operators used to condense the CBM-EX to the
CBM-IV. A description of their function is found in the text.
Acknowledgment. The research described in this article has been supported by the
United States Environmental Protection Agency (EPA) through contract 68-02-4136
to Systems Applications, Inc. This article has been reviewed by the Atmospheric
Research and Exposure Assessment Laboratory of the EPA and approved for publica-
tion. Approval does not signify that these contents necessarily reflect the views and
policies of the Agency and no official endorsement should be inferred.
182
-------
TABLE A1. The Carbon Bond Mechanism EX
Reaction
No.
(A1)
(A2)
(A3)
(A4)
(A5)
(A6)
(A7)
(A8)
(A9)
(A10)
(A11)
(A12)
(A13)
(A14)
(A15)
(A16)
(A17)
(A18)
(A19)
(A20)
(A21)
(A22)
(A23)
(A24)
(A25)
(A26)
(A27)
(A28)
(A29)
(A30)
N02
05
0 -£
°3 *
0%
0 +
0 +
o3 *
Oo +
o3 +
01D
01D
o3**
Il33
•f
NO,
j
NO,
N2°5
N2°5
NO +
NO +
OH +
HONO
OH +
HONO
OH +
OH +
H02
H02
PNA
Reaction
Inorganic
+ hv * NO * 0
,M
— •• o3
NO * N02 + 02
N02 * NO + 02
N02 -51* N03
NO -51* N02
N02 * N03 + 02
hv * 0 + 02
hv * 01D •»• 02
-51* o
+ H20 * 2 OH
OH * H02 + 02
H02 * OH + 2 02
+ hv * 0.89 N02
0.89 0 * 0.11 NO + 0.11
* NO * 2 N02
+ N02 * NO + N02 + 02
+ N02 -51* N20c
+ H20 * 2HN03
-51* N03 + N02
0,
NO -&* 2 N02
N02 + H20 * 2 HONO
NO -51* HONO
+ hv * OH + NO
HONO * N02 * H20
•c HONO * NO + N02 + H20
N02 -51* HN03
HN03 -51* N03 * H20
+ NO * OH + N02
^ Uf\ j. DMA
* N02 — - PNA
-51* H02 + N02
Rate Constant (k).
cm3 molecule" s~
Reactions
radiation dependent
1.4 x 1o3e"75/T
1.8 x 10-12 e-1370/T
9.3 x 10'12
1.6 x 10-13e687/T
2.2 x 10"13 e602/T
1 o « m~13 _-2450/T
1 .c x iu e
radiation dependent
radiation dependent
1 o v ir>8 ~390/T
i .y x lu e
2.2 x 10~10
l.OXlO 6
I • M X 1 0 €
radiation dependent
°2
1.3x10 e
2c ^ 1A~ 14 _— 1230/T
.D x iu e
55 v in""13 —256/T
o x iu e
1.3 x 10'21
3.5 x 10 e
tr\ rr->rt /T
1 R v m~2U 0D30/T
I .O X IU €
4.4 x 10"40
4.5 x 10-13 e8o6/T
radiation dependent
6.6 x 10"12
1.0 x 10'20
i.oxio e
51 v in~15 «1000/T
. i x iu e
3.7 x 10-12 e240/T
1.2x10 €
4.8 x 1013 e-10121/T
Notes
1
2
3
3
2
2
3
1
1
4
3
3
3
1
3
5
2
5
3
5
6
2
1
5
6
2
7
3
2
3
continued
133
-------
TABLE A1. Continued.
Reaction
No.
(A3D
(A32)
(A33)
(A34)
(A35)
(A36)
Reaction
OH + PNA * N02 + H20 + 02
H02 + H02 - H202 + 02
02 + H02 + H20 * H202 + 02 + -H20
H202 •»• hv - 2 OH
OH + H202 - H02 -t- H20
OP
OH + CO -2- H02 + C02
Rate Constant (k).
o 11
cm0 molecule' s
I • j X 1 U 6
I'.l x !°-38 ^5800/T
radiation dependent
3.1 x lO-12e-l87/T
2.2 x 10"13
Notes
3
5
5
1
3
3
Formaldehyde Reactions
(A37)
(A38)
(A39)
(A40)
(A41)
(A42)
(A43)
(A44)
(A45)
(A46)
(A47)
(A48)
(A49)
(A50)
(A5D
(A52)
(A53)
(A54)
(A55)
(A56)
Op
FORM + OH -£* H02 + CO + H20
20p
FORM + hv — ±-+ 2 H02 + CO
FORM + hv -» CO * H2
FORM + 0 - OH + H02 + CO
Op
FORM + N03 -£* HN03 + H02 + CO
FORM + H02 - FROX
FROX * H02 + FORM
FROX + H02 * PROX
FROX + NO * N02 + H02 + FACD
Higher Molecular Weight
Op
ALD2 + 0 -i+ C203 + OH
ALD2 + OH * C203 + H20
AT no , \rn 2^ r ft j. HWH
HJL*L/& T nUo -•••'^ W^WQ * tlWU-j
205
ALD2 + hv — £+ ME02 + H02 + CO
Op
C203 -t- NO -£•* N02 + ME02 •»• C02
C203 + N02 * PAN
PAN * C203 + N02
ME02 t- N02 * MPNA
MPNA * ME02 + N02
ME02 + NO * MEO + N02
MEO + NO * MNIT
1.0 x 10'11
radiation dependent
radiation dependent
3.0 x 10-11 e-1550/T
6.3 x 10-Jjj
1.0 x 10'iq
1.5
6.5 x 10'12
7.0 x 10'12
Aldehyde Reactions
1.2x10 €
7.0 x ID'12 e250/T
2.5 x 10-15
radiation dependent
5.4 x 10'12 e250/T
8.0 x 10-20 e5500/T
9.4 x 1016 e-^000/T
2.6 x 10-13 e*^^
1:1 x !o-i2<>o/T
1.5 x 10-11 e200/T
3
1
1
3
8
9
9
10
9
5
11
12
1
13
13
5
2
3
3
5
continued
184
-------
TABLE A1. Continued.
Reaction
No. Reaction
(A57)
(A58)
(A59)
(A60)
(A61)
(A62)
(A63)
(A64)
(A65)
(A66)
(A67)
(A68)
(A69)
(A70)
(A7D
(A72)
(A73)
MEO + NO -^ FORM + H02 + NO
MEO + N02 * MEN3
MEO -^ FORM + H02
MEN3 + OH -> FORM + N02 + H20
MNIT + OH - FORM + NO + H20
MNIT + hv - MEO + NO
ME02 + ME02 - 2 MEO + 02
ME02 + ME02 * FORM + MEOH + 02
ME02 + C203 * ME02 * MEO + 02
C2°3 * C2°3 * 2 ME02 * °2
ME02 + H02 * PROX + 02
ME02 + H02 * MEO + OH + 02
C203 + H02 * PROX + 02
C2°3 * H02 * ME02 + OH * °2
20?
AONE + hv — £-* ME02 + C203
AONE «• OH * AN02
AN02 •»• NO * C203 * FORM * N02
Rate Constant (k).
crsr molecule" s~
1.3 x 10-]2
1.5 x 10"11
8.4 x 10-14 e"1200/T
3.0 x 10~13
2.0 x 10"13
0.3 x k1
1.3 x 10-13
2.4 x 10'13
3.0 x 10"12
2.5 x 10'12
5^8 x !o"11* el300/T
1.4 x 1C"12
1 O
5.1 x 10'12
4.0 x 10"5 x k1
3.9 x 10"13
8.1 x 10'12
Notes
5
5
3
5
5
1
14
14
15
15
16
16
15
15
1
11,17
17
Alkane Reactions
(A74)
(A75)
(A76)
(A77)
(A78)
(A79)
(A80)
(A81)
(A82)
(A83)
(A84)
(A85)
(A86)
(A87)
(A88)
(A89)
(A90)
(A9D
(A92)
OH * ME02
PAR + OH - R02
PAR + OH * R02R
R02 * NO * N02 t- H02 + ALD2 + X
R02 + NO * NTR
R02R + NO * N02 + ROR
R02R + NO * NTR
ROR + N02 * NTR
ROR .* KET + H02
ROR * KET + D
ROR * ALD2 + D + X
ROR * AONE + D + 2 X
X * PAR *
D + PAR * R02
D + PAR * A02 * 2 X
D + PAR * R02R
D + KET * C203 + X
A02 + NO * N02 + AONE + H02
KET * hv * C?0, + R02 + 2 X
1.1 x 102e-1710/T
9.2 x 10~14
7.2 x 10'13
7.7 x 1CT12
4.4 x 10"11 e"l4oo/T
7.0 x 10'12
1.2x 10-«> e-l400/T
1.5 x 10"11
1.6 x 103
2.1 x 1014 e-8000/T
4.0 x 1014 e-8ooo/T
4.4 x 1011* e-8ooo/T
6.8 x 10"12
5.1 x 10'12
1.5 x 10'12
1.7 x 10"13
6.8 x 10'12
8.1 x 10'12
3.0 x 10~4 x k!
18
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
17
1
continued
880t9r2 3
185
-------
TABLE A1. Continued.
Reaction
No.
Reaction
Rate Constant (k).
cm3 molecule" s"
Notes
Alkene Reactions
(A93)
(A94)
(A95)
(A96)
0 4- OLE * 2 PAR
0 t- OLE * ALD2
0 + OLE * H02 4- CO + R02
0 + OLE * R02 4- X + CO
4.1 x
4.1 x
1.2 x
2.4 x
10-12 e-324/T
10-12 e-324/T
m-12 0-324/T
lu e
\r> ."Spll/T
4A '^ A j£"' ^
i(j e
5
5
5
5
+ FORM 4- OH
(A97)
(A98)
(A99)
(A100)
(A101)
(A102)
(A103)
(A104)
OH 4-
°3 +
°3 +
0, +
°3 +
N03 4
PN02
PN02
OLE - ME02 4.
OLE * ALD2 4.
OLE * FORM 4.
OLE -* ALD2 +
OLE - FORM 4-
• OLE * PN02
* NO * DNIT
4. NO - FORM H
ALD2 4- X
CRIG 4- X
MCRG 4- X
HOTA 4- X
HTMA 4- X
- ALD2
5.2 x
_ 1 p cnll /T
4 f\~ l£ ^jVJ*t/ 1
iu e
2.8 x 10"1!> e'ai0!)/T
2.ox10 6
i c ^11 r\f /m
4
1.
6
6
.3 x
.3 x
.7 x
.8 x
.8 x
10'
- iy Q-& IU3/ i
10-15 e-2105/T
10-1*
10"13
ID"12
19
20
20
20
20
21
21
21
4. X + 2 NO-,
(A105)
(A106)
(A107)
(A108)
(A109)
(A110)
(A111)
(A112)
(A113)
(A114)
(A115)
(A116)
(A117)
(A118)
(A119)
(A120)
(A121)
(A122)
(A123)
(A124)
(A125)
(A126)
(A127)
0 4- ETH * ME02 + }
i00 4- CO
0 + ETH ~ FORM + HO^ * CO f OH
OH 4-
ET02
ET02
°3 +
°3 +
HOTA
HOTA
HOTA
HTMA
HTMA
HTMA
HTMA
HTMA
CRIG
CRIG
CRIG
CRIG
MCRG
MCRG
MCRG
MCRG
ETH * ET02
4- NO * N02 4-
4- NO - N02 4-
ETH * HCHO 4.
ETH * HCHO 4-
-> C02 4- H2
* CO'* H20
f.
2 FORM 4- H02
ALD2 4- H02
CRIG
HOTA
* 2 H02 4- C02
- CH4 + C02
* ME02 + CO •
+ ME02 + H02
* FORM + CO •
* ME02 * H02
+ NO * N02 +
•»• H20 * FACD
4- FORM * OZD
+ ALD2 * OZD
* NO * N02 4-
4- H20 * ACAC
4- FORM * OZD
4- ALD2 * OZD
i- OH
4- C02
t- 2 H02
+ co2
FORM
+ H20
ALD2
4- HP0
b
7
3
2
6
1
5
7
0
0
0
0
0
0
0
0
7
4
2
2
7
4
2
2
.3 x
.1 x
.0 x
.3 x
.8 x
.0 x
.5 x
.20
.70
.10
.20
.32
.32
.08
.08
.0 x
.0 x
.0 x
.0 x
.0 x
.0 x
.0 x
.0 x
1 5 *7QO /T
1 A~ '£ &~ lj£' 1
IU €
1 0 7Q5 /T
1A~lt .,— (•Jfc/i
iu e
10-12 6411/T
10"12
1 o
10'12
10-15 e-2633/T
1C O^^O /T
10
10
10
10
10
10
10
10
" 3 e" JJ
-12
-16
1 o
- ie
-12
-12
-16
-12
-12
22
22
23
23
23
20
20
24
24
24
24
24
24
24
24
25
25
25
25
25
25
25
25
continued
880t9r2 3
186
-------
TABLE A1. Continued.
Reaction
No.
(A128)
(A129)
(A130)
(A13D
(A132)
(A133)
(A134)
(A135)
(A136)
(A137)
(A138)
(A139)
(A140)
(A141)
(A142)
(A143)
(A144)
(A145)
(A146)
(A147)
(A148)
(A149)
(A150)
(A15D
(A152)
(A153)
(A154)
(A155)
(A156)
(A157)
(A158)
(A159)
(A160)
OH +
OH +
OH t-
B02
BZA
OH +
BZ02
BZ02
PBZN
PH02
PHO
OH +
OH +
N03
CRO
CR02
CR02
T02
T02
T02
OH +
OH +
OH +
OH +
XL02
XINT
+
OH +
MGPX
MGLY
OH +
+
OPPX
OPEN
°3 +
*J
•»•
TOL *
TOL *
TOL *
+ NO *
+• hv •*•
BZA *
+ NO *
•»• N02
- BZ02
•i-NO *
+ N02 -
CRES *
CRES *
+ CRES
+ N02 -
+ NO *
+ NO -
+ NO *
+ NO *
* H02 +
XYL *
XYL *
XYL *
XYL *
+ NO •»
+ NO +
2 PAR
MGLY *
+ NO *
+ hv •*
OPEN -
CO
+ NO *
+ hv *
OPEN ->
CO
Reaction
B02
CRES +
T02
N02 +
BZ02
N02 +
* PBZN
+ N02
N02 +
NPHN
CRO
CR02
* CRO
NCRE
N02 +
N02 +
N02 *
NTR
CRES
XL02
CRES +
T02
XINT
N02 *
N02 *
MGPX
N02 +
c2o3
OPPX
N02 *
c2o3
ALD2
Rate Constant (k).
cm^ molecule" s~
Aromatic Reactions
1.7 x
H09 7.6 x
c
BZA +
PH02
PHO
* HNO,
j
OPEN
ACID
OPEN
PAR +
H02 +
H02 +
C2°3
+ CO +
+ c2o3
FORM
+ CO +
+ MGPX
H02
+ CO
* H02
+ H02
+ HO,
c.
H0?
c.
BZA * PAR
2 MGLY
H02
*H02
+ H02 + CO
H02
* FORM
1
8
4
1
2
1
9
8
1
1
2
2
1
4
4
7
8
4
1
3
5
6
8
8
1
8
.2
.1
.0
.3
.5
.7
.4
.1
.4
.6
.5
.2
.4
.0
.0
.3
.1
.2
.7
.3
.0
.6
.1
.1
.7
.1
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
10-13 e322/T
10-13 e322/T
10-12 e322/T
io-12
10~3 x k.
io-11 1
io-12
io-12
1016 e-l4000/T
io-12
io-11
io-11
io-11
io-11
io-11
io-12
* *\
1C'12
4 r*
10'12
10'13
,0-12 6116/T
10-12 6116/T
10-12 6116/T
10-12 6116/T
io-12
m ^\
io-12
io-11
1C'12
radiation dependent
3
8
.0
.1
X
X
10'"
1C'12
radiation dependent
1
.6
X
,0-1B e-500/T
«
Notes
26
26
26
27
1,27
28
27
27
27
27
27
19
19
12
27
27
27
29
29
29
26
26
26
26
30
30
28
31
1,
31
31
1,
31
,21
31
31
continued
880H9r2 3
187
-------
TABLE A1. Continued.
Reaction
No.
(A161)
(A162)
(A163)
0 +
OPEN -
+ 2H02
0, + OPEN *
0 .
OPEN -*
Reaction
FORM + CO + OH
MGLY
c2o3 •
«• FORM + H02
Rate Constant (k).
cur molecule' s
4.3
1
.1
3.2
X
X
X
+ CO
(A164)
(A165)
(A166)
(A167)
(A168)
(A169)
(A170)
(A171)
(A172)
(A173)
(A174)
(A175)
(A176)
(A177)
(A178)
(A179)
(A180)
(A181)
(A182)
(A183)
(A184)
(A185)
(A186)
(A187)
(A188)
(A189)
(A190)
(A191)
°3 +
0 +
0 t-
0 +
0 +
0 +
OH +
OH +
0, +
j
OQ +
o3 +
NO-
IS01
IS02
IS03
•*•
IS03
IS03
IS04
4.
IS04
IS04
OPEN *
ISOP *
ISOP *
ISOP *
ISOP *
ISOP *
ISOP *
ISOP -
ISOP *
ISOP -
ISOP *
ISOP -»
+ ISOP
+ NO -»
+ NO •»
+ NO -
MVK
* NO *
+ H02
+ NO -
MACR
+ NO -
+ HO,
ISNT + N0&-
0, +
3
OH +
OH +
MVK
MVK -
MVK -
MVK *
MVK *
+ hv *
0, + MACR -
EPOX
Isoprene
OLE + ALD2 1- PAR
IS01 +
IS02 +
ME02 •»•
IS03
IS04
FORM
FORM
CO * H02
CO + H02
C^O^ +• OLE
* MACR
+ MVK
FORM + OZD + CO
FORM
* ISNT
• N02 +
• N02 *
• N02 +
• ISN
* PROX
• N02 +
• ISN
* PROX
• DISK
MGLY +
MV1
MV2
c2o3 +
• MGLY
* OZD + CO
H02 + MVK
H02 * MACR
H02 + FORM
H02 + FORM
FORM
ETH + H02
+ FORM
5
.4
X
10-18 e-500/T
10-17 e-500/T
1Q_17 e-500/T
IO-18 e-500/T
Notes
31
31
31
31
Reactions
7
5
2
1
9
6
3
5
2
7
3
3
8
8
6
1
5
6
1
5
5
2
2
1
2
1
5
.3
.5
.7
.8
.1
.4
.2
.5
.4
.3
.5
.2
.1
.1
.8
.0
.4
.8
.0
.4
.4
.4
.4
.1
.2
.0
.6
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
,0-12
10'12
io-12
10'12
10
10'11
ID'11
io-18
io-18
10~19
10~18
10'^
io-12
10'12
<• o
io-12
10"12
10'12
10'12
10'12
io-12
10'12
10'^8
io-18
10-12 6514/T
10-12 Q514/T
10"1* x k1
10"19
32
32
32
32
32
33
33
34
34
34
34
35
36
36
36
36
36
36
36
36
36
36
36
28
28
1,
36
,37
,37
,36
,36
36
,37
continued
188
-------
TABLE A1. Concluded.
Reaction
No.
(A192)
(A193)
(A194)
(A195)
(A196)
(A197)
(A198)
(A199)
(A200)
(A201)
(A202)
(A203)
(A204)
0, +
08 +
OH +
MACR
MV1
4.
MV1
MV1
MV2
MV2
MAC1
+
MAC1
MAC2
+
MAC2
MACR
MACR
MACR
•*• hv
+ NO *
H02
+ NO +
+ H02
+ NO *
+ H02
* NO
co2
+ H02
+ NO
H02
+ H02
Reaction
^
- MAC1
* MAC2
- ME02 * ETH + H02 +
N02 + FORM + MGLY
MVNT
* PROX
N02 + C203 + ALD2
*. PROX
* N02 + ETH + ME02
* PROX
* N02 * FORM -t- MGLY
* PROX
Rate Constant (k).
cm^ molecule" s
5
1
4
CO 1
6
1
5
8
5
8
5
8
5
.6
.5
.9
.7
.8
.0
tn
.1
.n
.1
.4
.1
.4
X
X
X
X
X
X
X
X
X
X
X
X
X
10
10
10
10
10
10
10
10
10
10
10
10
-19
-11 e134/T
-12 e134/T
x k1
-12
-12
-12
-12
-12
-12
-12
-12
10"12
Notes*
36
28
28
1,
36
36
36
36
36
36
36
36
36
,37
,36
,36
36
* See text of this appendix for notes.
880>*9r2 3
189
-------
TABLE A2. Chemical Species in the CBM EX
Species Name
Representation
Nitric oxide
Nitrogen dioxide
Nitrogen trioxide (nitrate radical)
Dinitrogen pentoxide
Nitrous acid
Nitric acid
Peroxynitric acid (H02N02)
Oxygen atom (singlet)
Oxygen atom (triplet)
Hydroxyl radical
Water
Ozone
Hydroperoxy radical
Hydrogen peroxide
Carbon monoxide
Formaldehyde (CH2=0)
Hydroxymethylperoxy radical (HOCHg
Organic peroxide (ROOM)
Formic acid (HCOOH)
High Molecular weight aldehydes
(RCHO, R>H)
Peroxyacyl radical (CH3C(0)00')
Peroxyacyl nitrate (CH3C(0)OON02)
Methylperoxy radical (CH
Methylperoxy nitric acid
Methoxy radical (CHgO')
Methyl nitrite (CHjONO)
Methyl nitrate (CH3ON02)
Metnanol (CHjOH)
Acetone (CH3C(0)CH3)
Acetylmethylperoxy radical
(CH3C(0)CH200')
Paraffin carbon bond (C-C)
Primary organic peroxy radical
Secondary organic peroxy radical
Organic nitrate
Secondary organic oxy radical
-------
TABLE A2. Continued.
Species Name Representation
Ketone carbonyl group (-C(O)-) KET
Dimethyl secondary organic peroxide A02
radical
Olefinic carbon bond (C=C) OLE
Criegee biradical (H2COO') CRIG
Methyl criegee biradical (CH3(H)COO') MCRG
"Excited" formic acid HOTA
"Excited" acetic acid HTMA
Nitrated organic peroxy radical PN02
(-CH(ON02)-CH(00')-)
C2 dinitrate group DNIT
Ethene (CH2=CK2) ETH
Ethanol peroxide radical ET02
(CH2OH-CH200')
Ozonide and further products OZD
Acetic acid (CH3COOH) ACAC
Toluene (C6H5-CH3) TOL
Benzylperoxy radical B02
Cresol and higher molecular weight CRES
phenols
Toluene-hydroxyl radical adduct T02
Benzaldehyde BZA
Peroxybenzoyl radical BZ02
Peroxybenzoyl nitrate PBZN
Phenylperoxy radical PH02
Phenoxy radical PHO
Nitrophenol NPHN
Methylphenylperoxy radical CR02
Methylphenoxy radical CRO
Nitrocresol NCRE
High molecular weight aromatic OPEN
oxidation ring fragment
Aromatic ring fragment acid ACID
Xylene (Cgfy-fCH^g) XYL
Methylbenzylperoxy radical XL02
Xylene-Hydroxyl radical adduct XINT
Methylglyoxal (CH^C(0)C(0)H) MGLY
continued
-------
TABLE A2. Concluded.
Species Name Representation
Peroxide radical of MGLY MGPX
(CH3C(0)C(0)00')
Peroxide radical of OPEN OPPX
Isoprene ISOP
Isoprene epoxide product EPOX
Isoprene 0-adduct IS01
Isoprene 0-adduct IS02
Isoprene OH-adduct IS03
Dinitrate of isoprene DISN
Methylvinyl ketone OH-adduct MV1
Methylvinyl ketone OH-adduct MV2
Methacrolein OH-adduct MAC1
Methacrolein OH-adduct MAC2
Methylvinyl ketone nitrate MVNT
Isoprene OH-adduct IS04
Methacrolein MACR
Methylvinyl ketone MVK
Nitrate of isoprene ISNT
Nitrate of isoprene ISN
Paraffin loss operator (-PAR) X
Paraffin-to-peroxy radical operator D
Total ' 87
88049r2 >f
192
-------
TABLE A3. Sources of Molecular Data Used
to Calculate Photolysis Rates
Species
N02
Oo (to 0)
Oo (to (TD)
N03
HONO
H202
FORM
(to 2H02)
FORM
(to H2+CO)
ALD2
TABLE A4.
Reaction
No.
(R2)
(R5)
(R6)
(R17)
(R22)
(R26)
(R29)
(R53)«
Quantum
Yield
DeMore et al. [1985]
AtJtinson and Lloyd [1984]
DeMore et al. [1985]
Atkinson and Lloyd [1984]
Baulch et al. [1984]
assumed to be 1.0
Calvert [1980]
Calvert [1980]
Baulch et al. [1984]
Fall-Off
Absorption Cross Section
DeMore et al. [1985]
Molina and Molina [1987]
Bass [1985]
Atkinson and Lloyd [1984]
DeMore et al. [1985]
flaulch et al. [1984]
£*ss et al. [1980]
Bass et al. [1980]
Baulch et al. [1984]
Parameters for Pressure- and Temperature-Dependent
Rate Constants
k0 *o k. n.
6.0
9.0
9.0
2.2
6.7
2.6
2.3
1.5
X
X
X
X
X
X
X
X
10-3*
10-*
10-*
10-30
10-31
10-30
10-31
10-3°
-2.3
-2.0
-1.5
-4.3
-3.3
-3.2
-4.6
-4.0
2
2
3
1
3
2
4
6
.8 x
.2 x
.0 x
.5 x
.0 x
.4 x
.2 x
.5 x
io-12
io-11
io-11
io-12
10-11
10-1 1
io-12
io-12
0.0
0.0
0.0
-0.5
-1.0
-1.3
0.2
-2.0
Reference
DeMore
Baulch
DeMore
DeMore
DeMore
et al.
et al.
et al.
et al.
et al.
Atkinson and
DeMore
et al.
Atkinson and
DeMore
et al.
[1985],
[1984]
[1985]
[1985]
[1985]
Lloyd [1984]
[1985]
Lloyd [1984]
[1985]
* Not included in CBM-IV.
880»»9r2 5
193
-------
Appendix B
KEACTIVE HYDROCARBON CHEMISTRY
(microfiche version only)
12
-------
Appendix B
REACTIVE HYDROCARBON CHEMISTRY
This appendix provides an expanded description of reactive hydrocarbon
chemistry for alkanes, alkenes, aromatics, and biogenic species. We also
discuss specific steps and assumptions used in the condensation of the
CBM-EX to the CBM-IV. The information is intended for chemists, modelers,
and mechanism developers. For a more expanded discussion, see Gery et al.
[1988].
The chemical mechanisms for the above reactive hydrocarbon groups
employ both explicit structures and surrogates as the entities of lumping
and reaction; therefore, a single molecule can be represented by several
reactive structural groups. For instance, a complex molecule with both
aromatic and alkene structures might be represented with a combination of
TOL, OLE, and PAR surrogates. Also, some of the rates and stoichiometries
discussed next will depend on the atmospheric composition of reacting
hydrocarbons. To derive these parameters we used the hydrocarbon profile
given in Table B1, representing an average of 23 samples taken in Los
Angeles [Grosjean et al., 1981].
Alkane and Alkyl Group Chemistry
Basic Alkane Chemistry
Photooxidation of alkyl carbon in urban smog results almost exclusively
from hydrogen abstraction by OH in the general form:
R
H-C + -OH — * -C - R + H 0
. R ,22
I 2 |
R R
1 1
When all the variable functional groups (R) are hydrogen, the organic
reactant in this reaction is methane. Methane oxidation is of interest
880**9r2 12
-------
TABLE B-l. Average concentration (ppb) of each hydrocarbon
species identified in ambient air by the ERT and WSU methods
for 23 common analyses. (Source: Grosjean et al.f 1981.)
in
i
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
27
28
29
30
31
32
33
34
35
Species
Ethane
Ethylene
Acetylene
Propane
Propene
Propyne
Propadiene
Isobutane
Butane
1-Butene
Isobutene
Trans-2-Butene
C1s-2-Butene
Isopentane
Pentane
3-Methyl-l-Butene
1,3-Butadiene
1-Pentene
Isoprene
Trans-2-Pentene
C1s-2-Pentene
2-Methyl-2-Butene
2, 2-D1methyl butane
Cyclopentene
Cyclopentane
2,3-Dimethylbutane
2-Methylpentane
C1 s-4-Methy 1 -2-Pentene
2-Methylpentane
2-Methyl-l-Pentene
Hexane
Trans-2-Hexene
2-Methyl-2-Pentene
C1s-2-Hexene
Methyl cyclopentane
Ave. ERT
(ppb}
76.68
40.20
43.89
38.53
10.65
.00
.00
15.89
32.66
.00
.00
.00
.00
34.77
15.30
.00
.00
1.24
.00
.00
.00
.00
.00
.00
.00
3.51
13.82
.00
8.27
.20
10.39
.00
.00
.00
7.48
Ave. WSU
(PDb}
13.82
31.38
39.71
45.10
14.69
.00
.00
23.50
53.84
2.26
4.06
3.80
.34
28.72
18.94
.85
.00
.34
1.50
1.54
1.58
.42
.12
.76
2.81
3.31
11.57
.03
8.11
1.05
8.60
.51
.49
.28
9.62
Ratio1
ERT/WSU
5.55
1.28
1.11
.85
.72
-99.00
-99.00
.68
.61
-99.00
-99.00
-99.00
-99.00
1.21
.81
-99.00
-99.00
3.67
-99.00
-99.00
-99.00
-99.00
-99.00
-99.00
-99.00
1.06
1.19
-99.00
1.02
.19
1.21
-99.00
-99.00
-99.00
.78
-------
TABLE B-l. (Continued)
ID
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
Species
2 ,2 ,3-Trimethyl butane
2,4-Dimethylpentane
1-Methylcyclopentene
Benzene
Cyclohexane
2-Methylhexane
2,3-Dimethylpentane
3-Methylhexane
Dimethyl cyclopentane
2,2,4-Trimethylpentane
Heptane
Methylcyclohexane
Ethylcyclopentane
2,5-Diir-jthylhexane
2,4-Dimethylhexane
2,3,4-Trimethylpentane
Toluene
2,3-Dimethylhexane
2-Methyl heptane
3-Methyl heptane
2,2, 5-Tr Imethy 1 hexane
Dimethylcyclohexane
Octane
Ethylcyclohexane
Ethyl benzene
P- & M-Xylene
Styrene
0-Xylene
Nonane
I sopropyl benzene
Propyl benzene
P-Ethyltoluene
M-Ethyltoluene
1, 3, 5-Trimethyl benzene
0-Ethyltoluene
Tert-Butyl benzene
1,2,4-Trimethylbenzene
Sec-Butyl benzene
1,2,3-Trimethylbenzene
Ave. ERT
(ppb)
.39
2.83
.00
3.13
15.43
.00
4.37
6.47
1.90
7.56
4.63
5.51
.00
.84
.00
2.09
33.92
.00
.00
2.39
1.18
.06
1.76
.00
6.3?
18.06
.00
6.83
1.13
.08
1.17
5.47
.00
2.46
1.52
.00
5.63
.24
1.40
Ave. WSU
(ppb)
.00
3.98
.00
13.94
35.25
5.29
5.30
6.94
4.97
7.55
6.02
6.37
.97
.14
2.87
1.76
40.37
1.28
4.76
2.49
1.87
.00
2.67
.99
7.08
20.40
1.21
9.36
1.54
.64
2.56
5.98
2.72
2.06
1.76
.00
7.14
.65
1.71
Ratio
ERT/WSU
-99.00
.71
-99.00
.22
.44
-99.00
.83
.93
.38
1.00
.77
.86
-99.00
5.90
-99.00
1.19
.84
-99.00
-99.00
.96
.63
-99.00
.66
-99.00
.90
.89
-99.00
.73
.73
.12
.46
.92
-99.00
1.20
.86
-99.00
.79
.37
.82
(Continued)
199
-------
TABLE B-l. (Concluded)
Ave. ERT Ave. WSU Ratio
ID Species (ppb) (ppb) ERT/WSU
75 Decane 2.93 1.05 2.79
76 Methylstyrene .00 1.21 -99.00
77 1,3-Diethylbenzene .00 1.17 -99.00
78 1,4-Diethylbenzene .00 1.25 -99.00
79 1,2-Diethylbenzene .00 .00 -99.00
80 Undecane .00 .00 -99.00
81 Dodecane .00 .00 -99.00
82 N-Butylbenzene .20 .00 -99.00
83 .00 .00 -99.00
84 .00 .00 -99.00
85 Unidentified 80.12 52.62 1.52
*Ratio of -99.00 Indicates ratio cannot be calculated.
200
-------
mainly due to its involvement with the global carbon cycle, and is
essentially unreactive on the time scale of urban smog. We treat the
background reactivity of methane in the CBM by assuming a constant methane
concentration of 1.85 ppm (see Table 4 and notes).
OH Abstraction of Primary C - H Bonds. When only one of the R groups
in the hydrocarbon reactant is an alkyl chain (R-CH?) abstraction of one
of the three remaining hydrogen atoms by the hydroxyl radical proceeds at
a molecular rate of about 1.9 x 10"'^ cnr molec~' s [Atkinson, 1986;
Atkinson and Lloyd, 1984]. In the presence of tropospheric oxygen, the
resulting RCH- radical rapidly forms a peroxy radical (RCH 0-) that can
react with NO:
RCH 0- * NO * NO + RCH 0-
A minor fractional channel for this reaction results in nitrates:
RCH20^ + NO - RCH2ON02
with a yield that increases with increasing alkyl chain length (maximum
yield » 30 percent). We discuss treatment of ROx + NO * nitrate later in
this section.
For R < CH3(CH2)2-, or for any molecule in which the maximum carbon
chain length is less than 4, the RCH-0* radical reacts mainly with 02 to
form an aldehyde and a hydroperoxy radical:
RCH20- + 02 * RCHO + HO^
Additional intramolecular reactions can occur for alkoxyl radicals with a
carbon chain length of C^ or greater. Both decomposition and "tailbiting"
isomerization reactions are possible [Hendry and Kenley, 1979].
Decomposition reactions follow the general form:
RCH 0- - * RO- + HCHO
2 2
in which a carbon-carbon bond is severed to form a stable carbonyl species
and a peroxy radical.
880t9r2 12
201
-------
The isomerization sequence proceeds as follows:
H-CRCHgO- —-o -02H2CRCH2OH
•02H2CRCH2OH * NO * N02
•OH2CRCH2OH * HOH CRCH(OH)
The a-hydroxy radical could react with 02 via abstraction of hydrogen from
the hydroxy group to yield an aldehyde [Carter et al., 1979]:
HOH CRCH(OH) + 0 -> HO- + HOH CRCHO
The overall effect of the isomerization reaction, therefore, is to
delay the formation of the RCHO and H02 products of the nonisomerization
reactions by inserting extra alcohol-forming steps and extra NO-to-N02
conversions. The importance of these reactions in alkane chemistry is
uncertain since these compounds have not been directly observed, nor have
their reaction kinetics been studied, and the yield of secondary alcohols
in high molecular weight alkane systems has not been quantified. The
recent mechanism developed by Carter et al. [1986] assumes an
isomerization alcohol yield of about 0.2, which has only a minor effect
upon oxidant formation. Until isomerization phenomena are better
quantified, we believe that inclusion of alcohol groups in this mechanism
is not required.
OH Abstraction of Secondary C-H Bonds. The reactivity of OH for
secondary C-H hydrogen abstraction is about 1.15 cnr molec s at 298 K,
and has only a small activation energy [Atkinson, 1986]. In compounds
with an alkyl carbon chain length of four carbons or less (butane and
methyl substituted butanes), hydroxyl abstraction of a secondary C-H bond
and reaction of the alkylperoxy' radical with NO yields an alkoxy
radical. This radical can follow two main reaction sequences (again,
neglecting nitrate formation until later):
B80H9r2 12
202
-------
R -CH-R -*• R CHO + R-
121 2
Competition between these alkoxy reactions (hydrogen abstraction by 02
and decomposition) is largely governed by variations in the decomposition
reaction rate due to molecular size and temperature. For most of the
hydrocarbons in the atmospheric mix, the R^HO formed in decomposition
will be acetaldehyde (R=CH_). Since the average carbon number of alkanes
in the atmosphere is about 5.5, the alkyl group (R2) formed in the reac-
tion will be approximately C^-C^. In the review by Atkinson and Lloyd
[1984] the competition between alkoxy radical abstraction by 02 and
decomposition was shown to dramatically change with the number of carbon
atoms in the alkoxy radical. For the series sec-propoxyl , sec-butoxyl,
sec-pentoxyl and longer sec-radicals the competition can be expected to go
from essentially abstraction only to decomposition only.
The discussion above indicates that, for sec-alkoxyl radicals of four
or less carbons, abstraction is the exclusive or a major pathway. Also,
since ambient air contains an average carbon number between 5 and 6, the
first alkoxyl radical formed after reaction with OH will most often
decompose. However, the peroxy radical formed from such a decomposition
will most often have fewer than four carbons and the subsequent alkoxyl
radical will therefore undergo abstraction only. Symbolically, these
conclusions lead to the following sequence for secondary attack by OH:
°2
OH + alkane -£» sec-RO^ + HgO
sec-ROi + NO * NO- + sec-RO-
sec-RO- * ALD2 + R'O
R'0;j + NO * N02 + R'O-
R'O- - HO^ + ALD2 or KET
OH Abstraction of Tertiary C-H Bonds. Hydrogen abstraction by OH from
tertiary C-H bonds occurs at a rate of about 2.1 cnP molec s
[Atkinson, 1986]. The alkoxyl radical formed in this oxidation sequence
is devoid of hydrogen atoms and therefore cannot undergo abstraction (as
in the aforementioned scheme) by 02. Hence
B80t9r2 12
203
-------
R 0
I3 I
R2-C-0- * I^-C-Rg * R^
R1
A fraction of tertiary alkoxyl decomposition may yield ketones (e.g.,
3-methyl pentane could form methylethyl ketone). For a specific
hydrocarbon, the ketone production ratio from tertiary allcoxyl
decomposition can be calculated from its structure. For example, 2,3,4-
trimethylpentane has three tertiary carbons. Reaction and decomposition
at the 2-carbon site will yield acetone:
CH3 CH3 CH3 0 CH. CH.
III I i3 i3
CH.- C - C - C - CH. * CH--C-CH- + -C - C -CH.
3 I I I 3 3 3 II3
0 H H H H
*
while reaction at the 3-carbon site will yield a higher molecular weight
ketone (methylisopropyl ketone):
CH— CH— CH— CH— ,-,1, .-lit
333 3 CH_ LMo
ill I I3 I3
CH.- C - C - C - CH. - CH--C- * C - C - CH
3 | | | 3 3 , , |
H 0 H H OH
Thus, the overall ketone yield would be 2 acetones to 1 larger ketone if
the three sites are equally reactive. From our hydrocarbon profile (Table
B-1), we can estimate the ketone production ratio from tert-alkyl carbons
as 2.5 acetone to 1 higher ketone. We also note the existence of three
compounds containing quartenary carbons: 2,2,3-trimethylbutane, 2,2,4-
trimethylpentane, and 2,2,5-trimethylhexane. All three could be expected
to yield acetone if decomposition affects the quartenary carbon.
Quartenary carbon atoms can probably be neglected since compounds
containing them appear to be rare in the atmosphere (see Table B1).
880t9r2 12
204
-------
Ketones. The ketone structure
0
I
appears to be essentially unreactive to hydroxyl chemistry at the carbonyl
carbon as might be expected from the lack of labile hydrogens. The
primary reaction route involves OH and occurs via decomposition resulting
from reactions in the alkyl side chains, e.g., for methylethyl ketone [Cox
et al., 1981]:
OH 0 00-
I | +0 I |
CH.-C-CH-CH3+-OH —^ H20+CH.-C-CH-CH3
0 00- 00-
' I I I
CH3-C-CH-CH3+NO*N02+CH3-C-CH-CH3
00 00
I | +0 I I
CH -C-CH-CH * CH -COO- * HC-CH
Note that the presence of the carbonyl structure creates a peroxyacyl
product, whereas an alkyl group would yield an alkylperoxy product.
Acetone is sufficiently dissimilar to larger ketones in its rate of
photolysis to warrant explicit treatment. Acetone is formed directly from
propane oxidation but the principal formation pathway in the urban
atmosphere may be from decomposition of tertiary alkoxy radicals:
0- 0
I «
R, -C-R-+R, -C-R.J+R- •
1 I J 1 * 3
R2
12
205
-------
As previously noted, acetone is by far the most significant product of
such reactions because methyl side chains are more common than lengthier
alkyl branching groups, and also because the C-C bond strength is weaker
for carbon chains greater than C2. Therefore, the decomposition tends to
occur preferentially for alkyl side chains over methyl side groups.
Nitrate Formation from Alkanes. Nitrate formation from alkane/NOx
systems has been extensively discussed by Atkinson et al. [1982b, 1983,
and 1987]. We briefly summarize their observations here. Alkyl nitrate
yields in alkane-NOv photooxidation indicate that
A
R02- + NO + RON02
is an important source of alkyl nitrates when R contains three or more
carbon atoms [Darnall et al., 1976], In their experimental investigation
of these reactions, Atkinson et al. [1982b; 1983] showed that the alkyl
nitrate yields from the reaction of NO with the peroxy radical from a
specific alkane were the same regardless of (1) the time scale of the
experiment, (2) whether the experiment was conducted in the SAPRC
evacuable chamber, large teflon bag chambers, or small (- 100 liter)
teflon bag chambers, and (3) whether the peroxy radicals were generated
from N0x-air, CH^ONO-NOj-air, or C12-NOx-air photolyses. These findings
indicate that heterogeneous formation of alkyl nitrates is extremely
unlikely. Furthermore, there was no observable induction period for their
formation, indicating that they are closely associated with primary
products from the initial attack by OH on the alkane.
The alkyl nitrate yields at 740 torr and 300 K have been determined for
the n-alkanes (ethane through n-octane), and monotonically increase from
less than about 0.014 for ethane to 0.33 for n-octane [Atkinson et al.,
1982b]. A plot of the yield against carbon number [Atkinson et al.,
1982b] indicates a limiting nitrate yield of about 0.35 for the higher
alkanes. Assuming that all the n-alkanes have the same limiting high-
pressure alkyl nitrate yields, these data, along with temperature and
pressure effect data for pentyl and heptyl nitrates [Atkinson et al.,
19831, can be fit with an empirical, Troe-type fall-off curve. The best
fit to the data results in a maximum nitrate yield of 0.38.
The increase in nitrate yield with increasing carbon number is
consistent with increases in nitrate from secondary alkylperoxy reaction
with NO. Tertiary alkylperoxy radicals show a low nitrate yield of about
0.04 [Atkinson and Carter, 1987] that does not seem to increase with car-
bon number. Nitrate from primary radicals is also low and increases more
slowly than nitrate from secondary alkylperoxy radicals. The limiting
nitrate yields for primary alkyl peroxy radicals may be as low as 5
percent.
12
206
-------
The reduction in nitrate formation for primary and tertiary alkyperoxy
radicals implies that chain branching has a profound effect on nitrate
yields for an alkyl compound. Table B2 lists the calculated nitrate
yields for alkanes, based on the yields from primary, secondary, and
tertiary alkylperoxy yields, weighted by the fraction of hydroxyl reaction
at each carbon site. The measured yields for n-alkanes are shown for
comparison. It is apparent from the table that variations of as much as a
factor of 4 can occur in alkyl nitrate yields for compounds with the same
carbon number. A triple branched compound such as 2,3,4-trimethylpentane
might have a nitrate formation yield as low as 0.04, whereas the
corresponding n-alkane (n-octane) could have a nitrate yield eight times
as great. This phenomenon presents great difficulty for any molecular
lumping scheme. To lump alkanes with similar carbon numbers it would be
necessary to average greatly dissimilar nitrate yield molecules, whereas
to lump compounds with similar nitrate yields it would be necessary to
average a wide range of rate constants. We do not regard the averaging of
disparate nitrate yields to be a major problem, as we demonstrate in a
subsequent discussion, this can be easily resolved by use of a reactive-
structure lumping methodology.
Alkyl nitrate yields also show a temperature dependence that is, within
experimental error, the same for all alkoxy radicals yet studied. Using
data for six hydrocarbons given in Atkinson et al. [1987] and Atkinson
et al. [19831, we calculated an activation energy for the nitrate reaction
of 1400 K (a standard deviation of 300 K).
Development of Alkane Chemistry for the CBM-X and CBM-IV
When added to standard inorganic formaldehyde, acetaldehyde, and PAN
chemical reactions, a reaction scheme as described above will yield a
mechanism for treating the chemistry of alkanes. However, such a
mechanism would have at least three alkyl carbon groups and numerous
product species (e.g., multiple aldehyde and ketone species). In our view
the current requirements of atmospheric modeling does not require, nor
does current knowledge support, the use of such a detailed mechanism.
Therefore, the following simplifications were made to yield the CBM-EX.
Treatment of Decomposition in Carbon-Bond Groups. The alkyl radicals
formed from the decomposition reactions of secondary and tertiary alkoxy
radicals are similar to, but smaller than, the alkyl radicals formed by
the initial hydrogen abstraction reaction of the parent molecules. If an
R group is a methyl group, then a methyl radical may be formed. A methyl
radical is expected to react with 02 to form a methyl peroxy radical under
tropospheric conditions.
12
-------
TABLE 6-2. Calculated and measured nitrate yields
for various alkanes.
Carbon
Number
2
3
4
5
6
7
8
Calculated*
0.078
0.126
0.195
0.27
0.32
n -alkanes
0.01
0.036
0.077
0.128
0.22
0.31
0.33
Single
Branch
Alkanes
ll_ m
—
0.042
0.07
0.118
0.178
0.23
Double
Branch
Alkanes
^^
—
—
0.05
0.042
0.087
0.136
» Assuming 0.04 yield from tertiary and 0.05 yield
from primary carbons (Carter and Atkinson, 1985);
secondary carbon yields vary with carbon number
(Atkinson et al., 1987); weighted by OH reactivity
taken fromn Atkinson and Lloyd (1984) and the urban
hydrocarbon mixture of Table F-1.
860H9r2 12
208
-------
RP 0
1+0, «
R c-0- —£+ R.-C-R- + CH-O
1 I 12 3
CH3
Note that all primary carbons are methyl groups by definition. Therefore,
methyl peroxy radicals are formed at a rate relative to the overall number
of primary carbon atoms. (The bond strength relative to the secondary and
tertiary C-C bonds will also play a role.)
Similarly, if a secondary (-Cf^R) group is next to the decomposition
site, the alkyl radical formed is equivalent to that formed from the OH
abstraction from a primary carbon.
R
R2
Also, if a branched or tertiary group is involved, then a secondary
radical is formed:
*
OH 0
The forementioned reactions can actually be represented by the single
reaction:.
R--C-0- — * R..-C-R,j(carbonyl)
e. I 1 c.
R1
If R^ is hydrogen the carbonyl product is an aldehyde, otherwise the
carbonyl product is a ketone. The nature of the carbon atoms adjacent to
the site of decomposition determine the form within the R,OA product. If
12
209
-------
the general CBM assumption of independently reacting carbon-bonded
structures holds, then the distribution of products for R-O* will be
related to the distribution of carbon-bonded structures such as primary,
secondary, tertiary, and quartenary carbon atoms in the generalized pool
of alkyl carbon atoms and other structures such as ketones, etc. Although
we will later assume a generalized distribution within a single structure
species for alkyl carbon atoms, we can presently assume that four single-
bonded carbon atom structure species might be used in some extended form
of the CBM.
The generalized decomposition pathway would form a distribution
of R~OA radicals at rates related to the relative abundance of the carbon
structure groups with perhaps some minor modification by local bonding
effects. The minor modifications would also be a function of the distri-
bution of the carbon-bonded structures groups. One way of relating the
distribution of products for the generalized peroxy radical product
ILOi is the use of a virtual species or operator we will call D. This
species D would replace R_0i as a product in the generalized decomposition
reaction and then would rapidly "react" with the structure species.
R. 0
I3 o2 i
R.-C-0- -**• R,-C-R0 + D
2 I 1 2
R1
Primary Carbon + D •* MeO*
Secondary Carbon + D * RCH-O*
Tertiary Carbon + D —^ R R'CHO^
The relative rate constants would reflect the minor modifications. As
noted, quartenary carbon is rare (about 1 percent of typical atmospheric
mix, see Table B1) and could be neglected. The C-C bond strength of
primary carbon is apparently greater than that of secondary or tertiary
carbon, so the rate of the first reaction is relatively low.
For ketones, the carbonyl group adjacent to decomposition will result
in a peroxyacyl radical, as previously noted. Thus
B80k9r2 12 210
-------
Ketone + D —*~+ RCO» .
Upon later parameterization of these reactions to stoichiometric yields
(parameterization accomplished by assuming that intramolecular reactions
take place instantaneously), the "D" species will drop out of the
calculations and standard reaction stoichiometrics will result.
Higher aldehydes. The RCHO species that represent aldehydes with
carbon chain length greater than Co have be eliminated by using
acetaldehyde as a surrogate species. Our analysis of altcoxyl
decomposition pathways for the atmospheric mix of hydrocarbons given in
Table B1 yields a 45/55 split between RCHO aldehydes and acetaldehyde, but
also shows that aldehydes from alkenes, especially acetaldehyde from
propene and 2-alkenes, will dominate the aldehyde production from alkanes.
Alkyl Nitrate Products. The amount of carbon lost to alkyl nitrates is
apparently small. Although the molecular yields for highcarbon-number
n-alkanes can be as high as 38 percent, this yield involves only a single
carbon atom per molecule because the alkyl carbons in the nitrate are
still available for reaction. Decomposition is apparently the predominate
fate of a high-carbon-number alkane subsequent to hydroxyl abstraction,
and decomposition rapidly reduces the carbon number of the alkyl group,
making further alkyl nitrate formation unlikely. Studies of the kinetics
of alkyl nitrates [Atkinson et al., 1982b] suggest that the C-H bond on
the same carbon as a nitrate group has a low reactivity to OH
abstraction. Thus, nitrated carbon can be treated as an unreactive
product.
Lumping of Primary, Secondary, and Tertiary Carbon. The reactivity of
OH differs greatly for primary, secondary, and tertiary carbon
respectively [Atkinson, 1986]. Usually, such a broad range of rates
precludes lumping and rate-constant averaging. However, for a given
carbon number in a molecule, the rate constant for hydroxyl reaction is
similar (except when quartenary carbons are involved). The difference
between a normal alkane and a branched isomer is the loss of two secondary
carbons and the gain of one primary carbon and one tertiary carbon; twice
the secondary rate constant (2 x 1.15 x 10"'* cnr molec"^ s"1) is
approximately the sum of the primary and tertiary rate constants (2.1 x
10"12 + 1.9 x 10-3 cm3 molec"' s"1}. Because the tertiary carbon oxidizes
preferentially, one might expect the oxidation process to cause a change
in the relative abundances of tertiary, secondary, and primary carbon,
thus causing problems in a carbon-structure lumping scheme. However,
tertiary alkoxy radicals also preferentially decompose, with acetone as a
12
211
-------
common product. The formation of acetone removes less reactive primary
carbon from the alkyl pool. The formation of acetone removes a total of
three carbon atoms from the overall alkyl carbon pool. For the
atmospheric mix of alkanes given in Table B1, the relative proportions of
primary, secondary, and tertiary carbon are 0.47, 0.39, and 0.13. This
gives an overall lumped OH reactivity per carbon atom of
801 x 10 3 cnjS moiec-1 s-11 The three carbon loss from acetone formation
would therefore remove an equivalent to 2.43 x 10"^ cnH molec s . If
the alkyl carbon were divided into primary, secondary, and tertiary
groups, the acetone formation would remove one tertiary carbon and two
primary carbons. The rate constant total for this combination is
1 ? 3 11
2.48 x 10 cnr molec" s . Hence, for a typical atmospheric mix, the
average rate constant of the alkyl carbon pool would not change
appreciably if tertiary carbons rapidly were depleted, and acetone were
the major product of tertiary carbon reaction.
Figure B1 gives a representations of the product ratios and reaction
pathways of alkyl carbon subsequent to reaction with hydroxyl in the
CBM-EX. This reaction sequence lumps secondary and tertiary alkylperoxyl
(R02R) and alkoxyl radicals (ROR) in order to reduce the number of radical
species required. The nitrate formation pathway for primary carbon is
separated out, since the primary alkoxy (R02) radical is followed
independently. Nitrate formation for R02R is for lumped radical secondary
and tertiary nitrate formation weighted according to the reaction
proportions of each. The fraction thus derived, 0.14, can be compared to
the 0.16 value computed by weighting on the basis of the fractions of
secondary and tertiary carbon only. The lower fraction is preferred
because the subsequent reaction of reaction products involves various
species that often have lower molecular weights due to decomposition. The
average nitrate formation rate of product species will therefore be
lower. Thus, the initial nitrate formation rate of 0.14 is, in fact, an
upper limit. The ratios of radical production from unimolecular
decomposition of the alkoxy radical ("D + PAR") are derived from the ratio
of secondary and tertiary carbon and from the ratio of tertiary carbon
expected to"yield acetone and tertiary carbon involving molecular
fragments of C^ or greater in our assumed atmospheric mix. The sum of
alkyl decomposition involvement is the same as the ketone rate since the
alkyl carbon represents several combined pathways.
Condensation of the CBM-EX Alkvl Chemistry for CBM-IV
The CBM-EX alkyl reaction sequence presented in Table A1 is still too
large for inclusion in a condensed mechanism such as the CBM-IV. We have
212
880»»9p2 12
-------
H02 + ALD
(ALD2 + x)
AID* D
(ALD2 * x)
Dependent
B-l. septic ^presentation of the PAR re&ctSan
-------
therefore performed a number of condensation steps ranging from simple
bookkeeping to formulations requiring rather extensive conditional testing
to verify their impacts:
(1) Elimination of species X through the use of negative
stoichiometry. Since X was a mass balance counter that rapidly
removed PAR to conserve carbon, the production of X, as in
ROR * AONE + D + 2 X ,
can be changed to
ROR * AONE + D - 2 PAR .
(2) The mechanism can be abbreviated through the use of fixed
stoichiometry reaction combinations; e.g., the temperature-
dependent ROR reactions can be merged to
ROR -> D + 0.198 KET + 0.383 ALD2 + 0.419 AONE
- 1.222 PAR ,
with a combined decomposition rate at the same temperature
dependence (-E/R) of -8000 K.
(3) We created a second operator (X02N) similar to X02 to account
for the formation of a nitrate from the reaction of NO with a
peroxy radical. For example, the reaction set of
PAR * OH * R02R
R02R •«• NO + N02 + ROR k = 7.0 x 10~12 cm3 molec"1 s"1
R02R + NO - NTR k29g = 1.13 x 10'12 cm3 molec'1 s'1
can be written using the earlier X02 = NO*N02 operator or as
PAR + OH * 0.86 X02 + 0.86 ROR + 0.1M X02N ,
where X02N is defined through the reaction
X02N + NO * (NTR) .
Thus, using some previous condensations (X = -PAR, X02 =
NO * N02, and X02N = NO •»• NTR), and an assumption that the alkyl
nitrate yields are approximated by the 298 K ratios, we can
reduce the initial six reactions of Table A1 to one:
-------
PAR + OH * 0.871 X02 + 0.107 H02 + 0.107 ALD2
•»• 0.129 X02N + 0.764 ROR - 0.107 PAR ,
with a combined kQH of 8.1 x 10~" cnr molec s .
(4) Acetone (AONE) and ketone (KET) are slowly formed products of
alkane chemistry. In addition, the photolytic and reaction
losses (with OH) are relatively slow (Atkinson and Lloyd,
1984). For acetone, the central carbon is unable to enter into
hydroxyl reaction; therefore, we find the representation of the
molecule as 2 PAR to be satisfactory. This results in
alteration of product representations; for instance, the ROR
reaction Just discussed as an example of X elimination
ROR * AONE + D - 2PAR ,
is now
ROR * D ,
and
A02 + NO * N02 + H02 + AONE
is now
A02 + NO * N02 + H02 + 2PAR .
The production of A02 in the reaction
D ••• PAR * A02 + 2X
can be changed (using the X = -PAR and X02 = NO * N02 plus
products condensations) to
D * PAR * X02 + H02 .
For ketones, recall that KET represents the carbonyl carbon that
forms when a secondary or tertiary alkoxy radical stabilizes
through loss of a radical. The forementioned treatment of
acetone assumes no reactivity for this carbon. Simulation
results indicate that the mass throughput rates for the two KET
reactions are small; i.e., always at least two orders of
magnitude less than the D + PAR reactions. Therefore, we
eliminate the product KET and eliminate KET photolysis and OH
reactions from the CBM-IV.
215
880<*9r2 12
-------
(5) We can now eliminate D since the elimination of KET allows D to
be an operator on PAR only through three non-temperature-
dependent reactions:
D 4. PAR + R02 ,
D f PAR * X02 4- H02 ,
D 4- PAR * R02R .
We have already used the X02 and X02N condensations for R02 and
R02R in condensing the initial PAR reactions, and we can use
similar product splits to combine these reactions into
D * PAR * 0.959 X02 + 0.938 H02 + 0.713 ALD2
+ 0.041 X02N 4- 0.022 ROR - 0.713 PAR .
Since D is an operator with an instantaneous reaction rate, the
formation of D removes a PAR and instantly forms the D + PAR
products. Hence, the "product" D can be replaced by the D + PAR
products minus an additional PAR.
D = 0.959 X02 + 0.938 H02 + 0.713 ALD2
4. 0.041 X02N 4- 0.022 ROR - 1.713 PAR .
When substituted into the CBM-EX reaction scheme, these
condensations result in the CBM-IV alkane mechanism:
. - PAR + OH * 0.87 X02 + 0.13 X02N + 0.11 H02
+ 0.11 ALD2 4- 0.76 ROR - 0.11 PAR
ROR * 1.10 ALD2 4- 0.96 X02 4- 0.94 H02
4. 0.04 X02N 4- 0.02 ROR - 2.10 PAR
ROR * H02
ROR 4- N02 *
The temperature-dependent reaction rates are given in Table 4.
Alkene and Ethene Chemistry
Alkenes are a rather reactive class of hydrocarbons found in the
atmosphere. These species are generally involved in atmospheric chemistry
during all periods of the day since their structure allows attack by at
880U9r2 12
216
-------
least four important oxidizing species (0^, OH, 0, and NOq). As noted in
the main text, we treat 1-alkenes (OLE) as a reactive hydrocarbon
surrogate, often using propene chemistry as a model. Ethene (ETH) is
treated explicitly in the CBM-EX, and very reactive alkenes with two or
more alkyl substitutions are treated as surrogate aldehyde and ketones.
From a theoretical standpoint, the advantage of the reactive structure
approach for alkenes is mass conservation. For the 1-alkenes additional
alkyl groups in alkenes of three carbons and greater is accounted for with
the PAR surrogate. Next, we discuss the treatment of 1-alkenes, followed
by the explicit ethene chemistry.
0(3P)-OLE Chemistry
The gas-phase reactions of 0(^P)(0) and 1-alkenes have been studied for
over 30 years, yet their chemistry under atmospheric conditions is still
somewhat uncertain. Basis on their review of earlier data, Atkinson and
Lloyd [1984] suggest the following scheme for propene:
l=Un2
CH3CH-CH2 (30*)
CH3CH2CHO (30*)
(HCO + CH3CH2') (20%)
(CH2=CHO' * CH3') (20*)
The 1-butene yields were found to be 44, 39, 0, and 17 percent, with an
uncertainty of up to about 20 percent. In our formulation of the OLE + 0
reaction scheme, we use the following splits:
OLE
2 PAR
ALD2
H02 + CO + R02
R02 + X * CO + FORM
OH
(35*)
(35*)
(10*)
(20*)
880«*9r2 12
217
-------
The epoxide product is represented by the reactivity of 2 PAR (based on
the kgH calculations of Winer et al. [1978] cited in Atkinson and Lloyd.
[1984]). ALD2 is used to represent higher aldehydes, and the products of
the third reaction are simply represented in the CBM by assuming exclusive
reaction of the formyl radical with 02. We assume that the vinoxy radical
(CH2=CHO*) reacts exclusively with 02 to yield the products shown, with
R02 + X representing the formation of a peroxy radical from the remaining
carbon in the chain. In their review of kinetic data, Atkinson and Lloyd
[1984] recommend
k = 1.18 x 1CT11 exp(-324/T) cm3 molec"1 s'1
k298 = ^*° x 10 ^ molec" s~1 .
For the CBM-EX, these rates were partitioned into the above reactions.
OH-OLE Chemistry
The reaction scheme for OH addition alkenes was first clearly explained
through the FTIR work of Niki et al. [1978]. Briefly, OH is assumed to
add to either of the double-bonded carbons in the OLE group. Abstraction
of non-OLE hydrogens (as described by Atkinson and Lloyd [1984]) is
handled in the PAR chemistry. Since terminal alkenes have two different
carbon groups, two diffent reaction sets can be followed depending on the
initial addition site:
OH 00. OH
. I *°2 II
OH + R1R2CH=CH2 * R1R2CH—CH2 > R1R2CH—CH2
or
+°2
* R1R2CH--CH2 > R^gCH—CH2
OH OH 00'
Both peroxy radicals are expected to react with atmospheric NO to form N02
and an oxy radical. Both radicals decompose through the loss of the
hydroxyl hydrogen (forming H02*) and the severing of the original alkene
sigma bond, forming two aldehyde groups at the point of cleavage. For
these molecules, the products resulting after NO*N02 conversion,
regardless of the initial OH addition site, are
HCHO + H02' .
BBO>f9r2 12 218
-------
Therefore, the CBM-EX formulation for OH reaction with a terminal alkene
group is
OH + OLE * ME02 + ALD2 + X ,
ALD2 •*• X describes the -CHO end of a longer chain hydrocarbon whose
chemistry is represented by other carbon bond groups, and ME02 oxidizes NO
to N02 to form H02 and FORM.
The kinetic expressions used in the CBM-EX are
k = 5.24 x 10"12 exp(504/T) cm3 molec"1 s~1 ,
k298 = 2.84 x 10~11 cm3 molec"1 s"1 .
The temperature dependence used is from the review of Atkinson [1986]. We
use a slightly higher k2gg value than is suggested therein (2.6 x
tO"' cm3 molec" s~ ), which is easily within the stated uncertainty of
15 percent.
0^-OLE Chemistry
Alkenes are unique in their atmospheric reactivity to ozone since other
common organics do not react with ozone at significant rates. The
chemistry of the ozone-alkene system has been extensively investigated
over the past decade, but has yet to be clearly delineated. Although the
secondary reactions are not as uncertain as those of aromatic
hydrocarbons, there is much speculation about the tropospheric chemistry
of Criegee (biradical) species formed during intermediate reactions. For
a more complete discussion, the reader is advised to consult the reviews
of Atkinson and Lloyd [1984], Atkinson and Carter [1984], and Kerr and
Calvert [1985], keeping in mind that the OLE species represents only
terminal alkene groups.
The reaction rate data used for 03 plus OLE in the CBM-EX are •
k = 1.42 x 10'1Z| exp(-2105/T) cm3 molec"1 s"1 ,
= 1.22 x 10"1^ cm3 molec"1 s~1 .
The temperature dependence is from the review of Atkinson and Carter
[1984] and the k2qg value is well within their stated uncertainty.
880»t9r2 12
219
-------
Ozone adds to terminal alkenes at the pi bond, initially forming a
3-oxygen bridge known as a molozonide, which rapidly rearranges to an
ozonide and decomposes. Ozonide decomposition apparently results in
formation of an aldehyde and a biradical product. Following the recommen-
dations of Atkinson and Lloyd, we assume an equal split of ozonide cleav-
age channels:
CH CH--CH
3 2
A
CH CH CH
0-0
HCHO
[CH COO-]
[H COO-] + CH CHO
(50%)
(50%}
(propene molozonide)(propylene ozonide)
Initially biradical species probably form with excess energy. Following
the work of Herron and Huie [1977], Niki et al., [1978], Su et al. [1980],
and Kan et al. [1981], we assume that approximately 40 percent of the
biradical products are thermalized in the atmosphere while the remaining
60 percent may rearrange and decompose to smaller products. In the CBM-
EX, we describe these pathways with
03 + OLE
ALD2 + CRIG + X
k FORM + MCRG + X
ALD2 + HOTA + X
FORM + HTMA + X
(20%)
(20%)
(30%)
(30%)
The first two reactions form the stabilized biradicals (CRIG = H-COO- and
MCRG = CH^COO-) and the lower two reactions represent the formation of
intermediate "hot" formic and acetic acids, which rapidly decompose via
[Dodge and Arnts, 1979]:
HOTA
(C02 + H2)
CO (+H20)
2 H02 (+ C02)
(20%)
(70%)
(10*)
HTMA
(CH4 + C02)
ME02 + CO + OH
i* ME02 + H02 (+ C02)
2 H02 + CO + FORM
ME02 + H02 (+ C02)
(20*)
(32*)
(32*)
(8*)
(8*)
220
8BOif9r2 12
-------
The most likely reactions of the stabilized biradicais, ignoring the
possible reaction with N02 (^298 * 1-° x 10~13 cm3 molec"' s"1), are [Kerr
and Calvert 19851:
CRIG + NO * N02 + FORM k = 7.0 x 10'12
CRIG + H20 * FACD + H20 k2gg = 4.0 x 10'16
CRIG + FORM * OZD k2gg = 2.0 x 10~12
CRIG + ALD2 * OZD k2gg = 2.0 x 10"12
and
MCRG + NO * N02 + ALD2 k2gg = 7.0 x 10'16
MCRG + H20 - ACAC «• H20 k = 4.0 x 10'16
MCRG * FORM -> OZD k2gg = 2.0 x 10'12
MCRG + ALD2 * OZD k2gg = 2.0 x 10"12
(k2gg units are cnP molec" s~ ).
NO^-OLE Chemistry
The kinetic data for the reaction of NOg plus the OLE group are given
in Atkinson and Lloyd [1984], but were calculated using the N^
equilibrium constant of Malko and Troe [1982]. We now believe the use of
this value, when compared to the data presented in the ICRS section, leads
to an undercalculation of k2«g by a factor of 1.83. Therefore, we have
converted the rate suggested by Atkinson and Lloyd by that factor to yield
the temperature-independent rate of
k(59) = 7.68 x 10'15 cm3 molec'1 s'1 .
Experimental evidence indicates the addition of NO? to the OLE bond,
primarily at the 1-position. Following that channel for discussion
purposes, oxygen should add to the initial adduct to form a peroxy
racical:
00-
12
221
-------
Reaction of this peroxy radical with NO to form N02 or a
disproportionate reaction with H02 should lead to formation of an oxy
radical structure under atmospheric conditions. However, formation of a
dinitrate species can also occur:
00- 0'
I I
R1R2C-CH2ON02 + NO f* N02 * RjR2C-CH2ON02
ON02
I
R1R2C-CH2ON02 (minor pathway)
Decomposition of the oxy radical could lead to the lower molecular weight
products observed by Bandow and coworkers (as referenced in Atkinson and
Lloyd [1984] in their propene study (R1 = CH3 and R2 = H):
o-
CH3CH-CH2ON02 * CH3CHO + HCHO + N02 .
It can be shown that N03 addition to the 2-position would yield the same
products; therefore, in the CBM-EX these reactions are represented by
N03 * OLE * PN02
PN02 f NO * DNIT (9%)
+ FORM + ALD2 -t- X + 2 N02 (91%)
Ethene (ETH) chemistry is explicitly represented in the CBM because ethene
is generally the most abundant alkene, and its chemistry is sufficiently
unlike generalized OLE chemistry to warrant individual consideration. The
main differences are that its initial oxidation reactions are much slower
than higher molecular weight alkenes and, because of its C2 structure, it
will not form PAN as an oxidation product. The specific chemistry of
ethene is presented in reactions A105 through A111 of the CBM-EX (Table
A1).
0(3P)-ETH Chemistry
Unlike the higher molecular weight alkenes, ethene does not appear to form
stable epoxide products. The assumed product distribution used in the
222
8 8 0*9*. 2 12
-------
CBM-EX is based on the review by Atkinson and Lloyd [1984] and the recent
data of Smalley et al. [1986]. We use
(HCO + CH3') (10%)
(CH2=CHO- + H-) (30%)
The kinetic data for this reaction are taken from Atkinson and Lloyd
[1984] and partitioned by the above fractions:
k = 1.04 x 1CT11 exp(-792/T) cm3 molec'1 s"1
= 7.3 x 10~13 Cm3 molec~1 s~1 .
Assuming rapid reaction of H* with 02 and the product chemistry discussed
in the 0 plus OLE section, we represent this chemistry in the CBM-EX by
0 * ETH f ME02 + CO + H02 (70%)
H02 + OH + FORM + CO (30%)
OH-ETH Chemistry
The kinetic expression for the hydroxyl radical-ethene reaction has
been updated to one given by Atkinson [1986]:
k s 2.03 x 10~12 exp(4l1/T) cm3 molec'1 s"1
k29g r 8.06 x 10~12 cm3 molec'1 s'1 .
These values are very close to those of earlier CBM expressions.
The detection of glyocaldehyde as a product of the OH-ethene reaction
[Niki et al., 1981] has led to a redevelopment of the reaction mechanism
to include that species. Initially, OH is expected to add to ETH, forming
a peroxy radical in the atmosphere in a mechanism analogous to that
described for OLE:
HO HO 00+
OH + H2C=CH2 * H2C——CHp ———-> H2C»—CH2 .
In the case of ethene, however, the oxy radical formed from oxidation of
NO to N02 (for example) can decompose to form two sets of products:
12
223
-------
HO 00' HO 0-
II II
H2C~CH2 + NO + N02 + H2C—CH2
followed by
HO 0-
! I
H2C~CH2 * 02 2 HCHO + H02' (78%)
HO 0
I II
H2C—CH + H02' (22%)
This set of reactions is represented in the CBM-EX by
OH + ETH * ET02
ET02 * NO f* N02 + 2 FORM * H02 (78%)
N02 + ALD2 + H02 (22%)
O^-ETH Chemistry
The rate expression used in the CBM-EX for the ozone plus ethene
reaction is the sum of reactions A110 plus A111 of Table A1 :
k = 1.26 x 10"11* exp(-2633/T) cm3 molec'1 s"1
k2gg = 1.8 x 10"18 cm3 molec"1 s"1 .
The temperature dependence used is from Atkinson and Carter [ 1 984 ] ; the
value is within the experimental error determination of that review.
The reaction mechanism has received much experimental attention in
recent years. These data are summarized by Atkinson and Lloyd [1984].
The overall CBM mechanism follows the reaction scheme given for 0^ plus
OLE, except that the symmetry of ethene simplifies the secondary reaction
pathways. Therefore, we have
03 + H2C=CH2 * HCHO + [CH200-] ,
680t9r2 12
224
-------
where we agair assume a 40:60 split between the thermalized biradical and
the decomposing species. In the CBM-EX this is given by
03 + ETH r* HCHO + CRIG
U HCHO + HOTA (60?)
The reactions of the CRIG and HOTA species have been given previously.
Condensation of the CBM-EX OLE and ETH Mechanisms to CBM-IV
The condensation of these CBM-EX reactions to a CBM-IV scheme primarily
involves a simple algebraic combination of reactions to obtain fewer
reactions with fractional stoichiometries. Some steady-state assumptions,
such as use of the X02 and X02N operators, are employed.
The 0 plus OLE reactions are easily combined using X = -PAR to yield
0 * OLE * 0.35ALD2 + 0.50PAR + 0.30CO
+ 0.20FORM + 0.200H + 0.10H02
+ 0.30R02
Assuming the steady-state R02 assumption used for PAR condensation,
0 + OLE + 0.63ALD2 * 0.22PAR + 0.30CO
+ 0.20FORM + 0.200H + 0.38H02
+ 0.28X02 * 0.02X02*N
The OH plus OLE reaction is condensed using the ME02 steady-state assump-
tion (ME02 = X02 + FORM * H02) and -PAR for X:
OH + OLE * ALD2 - PAR + FORM t- X02 + H02
The Og plus OLE reactions can initially be combined using X = -PAR to:
03 + OLE * 0.50ALD2 + 0.50FORM - PAR
+ 0.20CRIG + 0.20MCRG
* 0.30HOTA + 0.30HTMA
Under most atmospheric conditions, the transformation of thermalized
biradicals (CRIG and MCRG) to organic acids in the presence of water vapor
should be the dominant reaction. The rate constants given in Table A1 are
very uncertain, however, and better measurements are needed to add
certainty to such an assumption. Since this set of reactions appears to
be dominant, the biradical products can be replaced by their respective
acids; and because these acids are rather stable in the atmosphere, we do
BB049r2 12
225
-------
not treat their chemistry in the CBM-IV. The "hot" rearranged biradicals
are assumed to decompose in the proportions given in reactions A112
through A119 of Table A1. These assumptions lead to
03 + OLE * 0.50ALD2 + 0.74FORM - PAR
+ 0.44H02 + 0.33CO + 0.22X02
+ 0.100H
The NOo plus OLE reaction is condensed using the X02 and X02N operators to
eliminate PN02. Also, -PAR is substituted for X and DNIT is assumed to be
stable and thus is not followed in the CBM-IV. The reaction is
N03 + OLE - 0.91 X02 + 0.09X02N + FORM * ALD2 + N02 - PAR .
If the assumption that ME02 = FORM + H02 + X02 is made, the 0 plus ETH
reaction can be condensed to the sum of reactions A105 and A106 of the
CBM-EX:
0 + ETH * 1.70H02 * CO + FORM + 0.70X02 + 0.300H .
ET02 can be eliminated from the OH plus ETH reaction using the X02
assumption:
OH t- ETH * X02 + 1.56FORM + 0.22 ALD2 + H02 .
Finally, the Oo plus ETH reaction can be condensed using the same
assumptions as in the 0? plus OLE condensation:
03 + ETH * FORM + 0.42CO + 0.12H02 .
Aromatic Chemistry
Although most aromatic oxidation is through reaction with OH, the
mechanisms for these reactions are highly uncertain and have been the
focus of experiments and modeling investigations for over a decade
[Atkinson et al., 1980; Killus and Whitten, 1982; Leone and Seinfeld, '
1984]. Because of recent advances in analytical equipment, considerable
effort has been directed toward identification of aromatic ring
fragmentation species, especially the larger, conjugated a- and y-
dicarbonyl products [Dumdei and O'Brien, 1984; Shepson et al., 1984].
However, experimental studies that attempted to track all of the reacted
aromatic carbon were unable to account for more than a fraction of the
reacted carbon atoms in dicarbonyl products. This is especially true for
toluene. Hence, a large portion of the secondary product mass has yet to
226
-------
be confidently identified, probably because of the many branching and
isomeric pathways involved in the complex photooxidation of aromatics and
secondary products. Representation of this missing mass has caused a
dilema for atmospheric chemists, who have generally proposed greater
yields of the detected dicarbonyl products in place of the unidentified
products. Judging from the relatively poor agreement of simulation
results with smog chamber data, however, use of the highly reactive
dicarbonyl species to account for all of the carbon known to have reacted
appears to provide overly reactive mechanisms. Such an inconsistency is
especially obvious if the undetected species have significantly different
structures, reactivity, or chemical functions than do the surrogate a- and
y-dicarbonyl species in the models (as appears to be the case for
toluene).
Because basic kinetic data are sparse, smog chamber data were used to
evaluate the effectiveness of mechanism approximations (such as the
secondary product chemistry) because these systems provide conditions more
similar to those encountered in the atmosphere than the conditions of
laboratory chemical kinetics studies. On the other hand, the smog chamber
can rarely yield specific data on chemical kinetics or mechanism
development; this type of information is most often obtained in a
laboratory. One connection between the more complex smog chamber system
and the individual chemical kinetics experiment has been the ability of
the former (given enough well-conceived experiments) to indicate
functional trends related to possible atmospheric reactions of secondary
species not yet studied individually in the laboratory. That is, though
the chamber results cannot point out the exact structure or chemistry of
the missing aromatic product species, the chemical nature of potentially
missing products is inherent in the smog chamber data. Therefore, our
overall mechanism development included both available laboratory kinetic
and mechanistic information and systematic or conditional trends
ascertained from smog chamber data. As just described, we used the smog
chamber data to indicate the presence of chemical reactions that have not
yet been studied in kinetic experiments.
CBM-IV mechanism development also involved a number of application-
based requirements. For instance, the current generation of'photochemical
kinetic mechanisms may be used to study formaldehyde chemistry; therefore,
the formaldehyde mechanism should be explicit. Formaldehyde was
previously used as a surrogate for glyoxal (an aromatic ring fragment).
However, since formaldehyde is a less significant primary product than is
glyoxal [Bandow et al., 1985; Bandow and Washida, 1985; Gery et al., 1985,
1987], that association has now been removed and formaldehyde is
explicitly described in the CBM-IV. Other constraints, such as the
limitation of the number of species in a condensed mechanism, have led us
12
-------
to develop a secondary reaction scheme for toluene and xylenes that is
less oriented toward the chemistry of explicit species and is more func-
tionally based. This shift is mandatory for all aromatic mechanisms to
some extent because of the large number of possible reaction branch points
and isooers.
Choice of Aromatic Surrogate Species
Two aromatic surrogate species and their respective chemistries are
used to represent the chemistry of all aromatics: TOL for toluene (mono-
alkylbenzenes) and XYL for xylene (multi-alkylbenzenes). These species
were selected because the surrogate chemistries must represent the widest
possible range of aromatic reactivities (in both OH reaction rate
constants and secondary chemistries) if they are to provide the most
appropriate overall representation.
OH is the primary oxidizer of simple aromatic hydrocarbons under
atmospheric conditions. Ohta and Ohyama [1985] have shown that the OH-
plus-aromatics reaction rates correlate well with structural
configurations of simple aromatic species. Therefore, we currently use
the OH-plus-toluene rate constant for the OH plus monoalkylbenzene
reaction
k = 2.10 x 10'12 exp(322/T) cm3 molec'1 s~1
or
k298 = 6'19 x 10~12 ^ nolec"1 s"1 •
The XYL rate constant is set to that of m-xylene:
k = 1.66 x 10~11 exp(1l6/T) cm3 molec'1 s~1 •
or
k2gg = 2.45 x 10'11 cm3 molec'1 s'1 .
Both sets of kinetic data are from Atkinson [1986]. Because of the
present uncertainty in secondary products and their chemistry for higher
molecular weight alkylbenzenes, a third aromatic surrogate is not
warranted. If more data become available, however, and the requirement of
a limited number of species is lifted for future photochemical kinetics
mechanisms, a good argument for expansion of this scheme to a third (high
molecular weight) surrogate could be made on the basis of the high OH
reaction rate for that group.
An important issue that is less obvious than the OH reaction rate
groupings involves the secondary reaction scheme for different aromatic
species. Two different levels of product formation are involved here.
880H9r2 12 228
-------
First, we assume from limited evidence that aromatic ring reactions are
strongly dependent on the number and location of functional groups, but
only weakly affected by variations in types of alkyl groups. This
assumption allows the use of data for simple alkylbenzenes to describe the
initial oxidation reactions and ring cleavage schemes for higher molecular
weight species. Second, the yields of the very reactive (and
photolytically reactive) a- and conjugated y-dicarbonyl products demand
scrutiny. As noted by many researchers [Shepson et al., 1984; Bandow et
al., 1985; Bandow and Washida, 1985; Gery et al., 1985, 1987; Tuazon et
al., 1986], the yield of dicarbonyl species from the OH-plus-toluene
reaction is relatively low compared to yields from the xylenes and
trimethylbenzenes. For whatever reason, the prompt formation of
dicarbonyl species after OH-aromatic reaction appears to be a definite
feature of multi-alkylbenzene/NOx smog chamber experiments, but is far
less evident in toluene systems. Such products are stable but extremely
thermally and photolytically reactive. Therefore, their presence
perpetuates the reactive nature of the system even after the initial
hydrocarbon is exhausted.
Available smog chamber data indicate that toluene systems rapidly
consume NO. Radical flux and ozone formation in these toluene experiments
appears to be very reactive in this initial stage, but then also rapidly
terminates and often actually consumes ozone throughout the second part of
an experiment. This dichotomous smog chamber behavior is unique for a
"reactive" hydrocarbon and may indicate that, unlike other aromatics and
olefins, the product mixture formed through the initial (relatively
reactive) portion of these experiments may develop into less reactive
species as experimental conditions (availibility of NO) change. Toluene
appears to be very sensitive to these reactivity parameters, which results
in the dichotomous nature of these smog chamber experiments. Xylenes and
higher molecular weight aromatics react much faster with OH and produce
higher yields of reactive products. This faster reaction rate decreases
their sensitivity to NO concentrations and results in rapid production of
reactive species that can be more correctly approximated by the assumption
of prompt dicarbonyl (reactive product) formation.
TOL and XYL.Chemical Mechanisms
The reaction mechanisms of toluene and o- and m-xylene were
specifically studied because of the availability of smog chamber data and
the relative abundance of these species in urban atmospheric samples and
automobile exhaust. The initial product distributions for toluene and
xylene oxidation by OH are taken from the limited set of data available in
the literature. OH reacts with toluene by either abstracting a methyl
hydrogen to form H20, or by adding to the aromatic ring. Reactions for
12
229
-------
the xylenes are analogous. We assume approximately 8 percent hydrogen
abstraction for toluene and 10 percent for m-xylene based on the
experimental results of Gery et al. [1985, 19871. The abstraction pathway
is minor and is not discussed here. Reactions of the alkylbenzylperoxy
radicals (B02, the methylbenzylperoxy radical; and XL02, the xylene
analogue) and subsequent species are included in the CBM-EX (Table A1) as
reactions A131 through A138 and A152. For a discussion of the abstraction
pathway and specific sources of reaction rate constants for toluene see
Gery et al. [1985], and for xylenes see Gery et al. [1987].
Most products are formed by the addition of OH to the aromatic ring
structure. This OH adduct can either reversibly add an oxygen molecule to
form a peroxy radical (T02 and XINT), or loose a hydrogen atom to form an
alkylphenol (CRES for cresol) and H02*. The approximate fractions of
phenolic yields (fn^) were again taken from Gery and co-workers. For
toluene, fQH * 0.36; for m-xylene, fQH « 0.20. In the CBM-EX the overall
OH-plus-aromatic fractions (including the hydrogen abstraction pathway)
for toluene are:
OH + TOL * B02 (8*)
OH + TOL - CRES + H02 (36*)
OH + TOL * T02 (56*)
and for xylene:
OH + XYL * XL02 (10*)
OH + XYL * CRES + PAR + H02 (20*)
OH + XYL * T02 (30*)
OH + XYL -> XINT (40*)
where XINT represents the mechanism intermediate from which dicarbonyls
(MGLY) may rapidly form as a result of ring decomposition. Also, because
T02 is an intermediate of one less carbon than the parent xylene, a
nonreactive carbon is formed. This carbon atom is nonreactive because the
formation of 1 PAR would be incorrect in this case.
Beyond this point, the description of continuing oxidation, especially
of the chemistry of the OH-02 radical adducts, becomes very speculative.
All schemes to date show either the initial OH-02 radical adduct or a
bridged adduct reacting with NO to form N02 and products (generally,
reactive ring fragments). The functional description of most mechanisms
is similar up to this point. In terms of species, these reactions in the
CBM-X are:
880t9r2 12
230
-------
T02 + NO p N02 + OPEN + H02 (90%)
I TNTH (10*)
XINR * NO * N02 +• 2MGLY * 2PAR + H02
For the first two reactions, we use an overall reaction rate constant
of 8.1 x 10~12 cm^ molec"1 s~1. The same rate is used for the XINT plus
NO reaction. This reaction represents the fraction of XYL molecules that
promptly form dicarbonyl products. Later we describe the chemistry of
CRES, a general phenolic product, and MGLY and OPEN, which are dicarbonyl
species. However, we note here that our approach to the formation and
chemistry of dicarbonyls is to lump all of the dicarbonyl reactions into
one general surrogate species (OPEN) and one other explicit species,
methylglyoxal (MGLY). We believe this is an appropriate approach because
(1) the uncertainty of products and their chemistry subsequent to
dicarbonyl formation via ring decomposition is large, and (2) the timing
of variations between data and model results appears to result from poor
descriptions of mechanistic branching points prior to dicarbonyl
formation.
The greatest area of uncertainty in aromatic mechanisms lies in the
radical reactions following ring fragmentation. One assumption currently
used in many photochemical models of aromatic oxidation is the idea of
"prompt" product formation. In models such as the CBM-IV, this idea can
be specifically instituted when the (generalized) reaction set,
RH + OH * R+(+ H20)
R- •»• 02 * RO^
R0£ -K NO * N02 + products,
is simplified to
RH •»• OH * X02 + products (02 constant),
with the X02 operator defined as
X02 ^ NO * N02.
12
231
-------
In the main text this concept was introduced to minimize the number of
species and reactions in a chemical mechanism by grouping the common NO-
to-N02 conversion function of all R0£ species into the X02 operator,
thereby eliminating explicit description of specific R02 species in favor
of their "promptly" formed products. This approximation drifts from the
explicit scheme when another reaction competes with the NO reaction (even
at relatively high NO concentrations) and produces different products. In
such a case the X02 operation must be eliminated and the R0£ reactions
explicitly represented. On the basis of the smog chamber data and the
laboratory product yield data for toluene, we believe that this is
necessary for toluene. One possibility that has not yet been considered
is alternate reactions of the OH-02 adduct radicals (T02) to form products
that are less reactive than dicarbonyls. The T02 intermediates probably
do not have long reactive half-lives, and if they cannot react with NO,
must be lost by an alternate reaction pathway. Our methodology in
developing this most recent mechanism has been, in part, to develop the
mechanism and kinetics for an alternate reaction of the T02 radical. Such
a reaction probably becomes important only at low NO concentrations, and
produces less reactive products than the NO-to-N02 conversion process.
Through an extensive set of smog chamber simulations and sensitivity
tests, we have estimated a T02 decomposition of the form
T02 * H02 + CRES ( + 02).
The unimolecular rate of this reaction was estimated to be:
k = 4.2 s'1
We use cresol here only because it is already available in the CBM and
describes a secondary aromatic product that is less reactive than
dicarbonyls and that removes nitrogen from the system.
Aromatic Product Species
The species, CRES, represents alkylphenolic products formed after OH
addition to the aromatic ring. In the case of toluene, the CRES species
formed is cresol, typically o-cresol (o-methylphenol). The CRES chemical
reactions in the CBM are formulated for that compound because the
chemistry of higher molecular weight alkylphenol homologues and other
cresol isomers has been far less studied. For the other cresol isomers,
we assume the same chemistry as o-cresol, and for the higher molecular
weight homologues, we include alkane surrogates (PAR) groups upon
formation. Thus, the dimethyIphenol compounds formed from xylene
oxidation by OH are represented by CRES plus PAR.
12
232
-------
The chemical reactions of cresol and larger alkylphenols are summarized
in Atkinson and Lloyd [1984]; additional information is found in Gery et
al. [1987 and 1985] and Kerr and Calvert [1985]. The reader is referred
to these works for a complete description of the current state of
knowledge with respect to alkylphenol chemistry. We represent this
chemistry in the CBM-EX in reactions A139 to A144 of Table Al. Our
overall rate of hydroxyl reaction with CRES is from Atkinson and Lloyd
[1984]:
k = 4.1 x 10"11 cm3 molec'1 s'1 ,
with an assumed OH addition fraction of 60 percent and a 40 percent
abstraction of the phenol hydrogen by OH to yield CRO, a methylphenoxy
radical. The remaining reactions are very speculative, and are based on
the review by Atkinson and Lloyd [1984]. The ACID species represents
organic acids formed from aromatic ring decomposition of the alkylphenol
peroxy radical (the 02 adduct formed after OH addition to CRES).
Methylglyoxal (CH^CCOJCHO) is a dicarbonyl ring fragment found in
aromatic systems containing ring methyl groups. It is very reactive with
the hydroxyl radical and also photolyzes to radical products. These
reactions are represented in the CBM-EX by
OH + MGLY (+ 02) * MGPX (+ H20)
MGPX + NO - N02 + C203 (+C02)
MGLY ^ C203 + CO + H02
where MGPX represents a peroxy radical formed after hydroxyl abstraction
of the aldehydic hydrogen.
The kinetic information used in the CBM-X for the OH reaction is from
Atkinson and Lloyd [1984]:
k s 1.7 x 10'n cm3 molec'1 s~1 .
The MGPX is assumed to react with NO at
k =- 8.1 x 10"12 cm3 molec"1 s"1 .
Photolysis of MGLY has been studied [Plum et al., 1983], though the
uncertainty of the quantum yield data causes calculated j-values to be
very uncertain in the lower wavelength region of the surface solar
spectrum. On the basis of smog chamber data, we use the following rate
expression in the CBM-EX:
12
233
-------
JMGLY = 9-614 x ^HCHOP s" »
where J^c^Or ^s ^ne J"va^ue f°r formaldehyde photolysis to radicals.
The species OPEN represents a combination of possible dicarbonyl
species with up to seven carbons. We felt that, given the limited
information concerning dicarbonly chemistry, speculation about the
specific structure and reactions of individual compounds was pointless.
Instead, we combined the expected kinetics of the anticipated dicarbonyl
product structures into one general species that could represent the mass
of all products. Hence, the final rates used for these species are based
on the goodness-of-fit for a large set of smog chamber simulations. In
addition, since photochemical kinetics mechanisms used in large air
quality simulation models require small numbers of species, we felt it was
best to implement the improvements to the T02 chemistry instead of using
additional dicarbonyl species (all of which are very reactive). The
photolysis of OPEN and the OH reaction in the CBM-EX are
OPEN ^ C203 ••- H02 + CO
OH + OPEN * OPPX + C203 + H02 •«• CO
NO + OPPX * N02 f FORM + H02 + CO
The kinetic expression used for OPEN photolysis is similar to that of
MGLY:
a"1 .
We assumed the rate for OH oxidation of OPEN to be
k = 3.0 x 10"11 cm3 molec"1 s"1
and that of the NO plus OPPX reaction to be:
k s 8.1 x 10"12 cm3 molec"1 s"1 .
The product distributions are very uncertain. Many of the larger
assumed products were 5 or 6 carbon o- and y-dicarbonyls with conjugated
alkene groups. We assumed photolysis of only o-dicarbonyl products.
Hydroxyl reaction was even more speculative because of the ability of OH
to react rapidly with both the olefinic and the aldehydic carbons of a-
dicarbonyls.
Ozone reaction with the alkene bonds of larger OPEN species was assumed
to proceed along the lines pointed out by Dodge and Arnts [1979] for
simpler alkenes. We used a hypothetical product distribution to derive
12
234
-------
product stoichiometries for these reactions. These results are snowr; in
the CBM-EX as reactions A-160 through A- 164 (Table A1). For the reaction
of ozone with OPEN, we use the overall rate expression
k = 5.1*3 x 10'17 exp(-500/T) cm3 molec"1 s"1 ,
k298 = 1-° x 10~17 cm3 molec"1 s~1 .
The chemical reactions now used to describe the photooxidation of
aromatic hydrocarbons provide simulation results for smog chamber data
that are much improved over those of earlier models. On the other hand,
because fundamental information on aromatic systems is limited, this
mechanism does not rest on as strong a basis as other sections of the
CBM. Rather, it is a combination of the available fundamental information
and an empirical description of additional information available from more
complex smog chamber experiments.
Condensation of the Reaction Schemes
As noted, the reaction scheme for xylenes Just given should ideally
include the reactions of xylene carbon product species. Because that
chemistry is extremely uncertain, however, we assume the chemistry of B02»
T02 and CRES as a logical approximation. Condensation of the extended
chemistry proceeds as follows: For the initial reactions of OH with
toluene and xylene, all species except the abstraction products are
accounted for by direct combination of reaction stoichiometry and rate
constants (see Table 4). The abstraction products are represented by
B02 = X02 + H02
and
XL02 = X02 ••- H02 + PAR ,
where X02 is generally the NO-to-N02 conversion function (as previously
described). Hence, the reaction of B02 with NO to form N02, H02, and
benzaldehyde is represented by the X02 operator and product H02> with the
BZA product omitted. A similar representation is made for xylenes, except
that PAR is used to represent the additional carbon produced from the
reaction of XL02 with NO. After combination of these reactions, the
resulting reactions are
OH + TOL - 0.08 X02 + 0.36 CRES + 0.44 H02 + 0.56 T02
OH + XYL - 0.50 X02 + 0.20 CRES •»• 0.70 H02 + 0.30 T02
+ 0.80 MGLY +1.10 PAR
12
235
-------
where XINT is eliminated with an X02 operator but T02 chemistry is
explicit, as noted:
T02 + NO -> 0.90 N02 + 0.90 H02 + 0.90 OPEN
T02 * H02 + CRES .
The reactions of cresol are combined to
OH + CRES * O.HO CRO + 0.60 X02 + 0.60 H02 + 0.30 OPEN
NOj + CRES * CRO f HN03
CRO + N02 *
The OH and ozone oxidation reactions of MGLY and OPEN are condensed
algebraically and through the elimination of MGPX and OPPX via the X02
operator, to yield:
OH + MGLY - X02 + C203
OH + OPEN * X02 «• 2.0CO * 2.0H02 + C203 + FORM
03 * OPEN * 0.03ALD2 + 0.62C203 * 0.70FORM + 0.03X02
+ 0.69CO * 0.080H + 0.76H02 + 0.20MGLY
Biogenic Hydrocarbons
o-Pinene
Pinene is representative of the general class of terpenes, having a
reactivity in outdoor bag experiments similar to other compounds such as
d-limonene, terpinolene, and delta-3-carene [Arnts and Gay, 1979].
Although a-pinene itself reacts rapidly with OH, Oj, and NOj, its reaction
products seem to be similar in reactivity to an urban mix, as indicated in
hydrocarbon substitution experiments. Because product data for a-pinene
and related biogenics are lacking, it was decided to limit the
representation of a-pinene in the CBM-IV to a surrogate mix of species
whose CBM-IV behavior would mimic a-pinene in smog chamber simulations.
It was found that mixture consisting of 0.5 OLE, 1.5 ALD2, and 8 PAR
would reasonably reproduce the nitric oxide conversion and ozone formation
of a given quantity of a-pinene. This mixture also preserves some rough
equivalence to expected a-pinene products, e.g., PAN and formaldehyde are
12
236
-------
both expected to be secondary rather than primary products. Moreover,
this mix provided a replication of the oxidizing potential of an
equivalent amount of o-pinene pretreated with ozone prior to the
introduction of nitric oxide and subsequent irradiation. Thus, we have at
least a crude simulation of the fate of pinene which reacts with ozone
prior to interaction with an urban plume. We suggest that this is the
most likely scenario of biogenic impact on photochemical oxidant
formation.
Specific simulation results of individual experiments for pinene may be
found in Gery et al. [1988].
Isoprene
Isoprene may be an important factor in regional oxidant formation due
to its emissions from deciduous forests which, while diffuse, can account
for large total mass in a regional emission budget. Isoprene is also a
highly reactive compound, having both a rapid initial oxidation by OH, and
products that are considerably more reactive than an urban hydrocarbon
mix. Thus, even low concentrations of the compound might significantly
contribute to the inititation of oxidant photochemistry.
In our task of developing the condensed isoprene mechanism, we wished
to limit the product species considered to those already included in the
CBM-IV. The first step in implementing this explicit isoprene mechanism
was the development of a good kinetic representation of the initial
isoprene oxidation reactions. These oxidation reactions occur with 0,
NOo, OH and 0?; the last two are usually most important in daytime
atmospheric chemistry. We have reviewed the reaction rate constants
published in recent studies, and revised our earlier explicit mechanism
rates [Killus and Whitten, 1984], when appropriate.
Since a rate constant for the 0 atom reaction with isoprene has not
been well studied, it was assumed to be the sum of 0 atom additions to
each of the alkene bonds in the isoprene molecule. The dialkylated bond
characteristics were estimated to be similar to those of isobutene
[Atkinson and Pitts, 1977], and the monoalkylated bond was estimated to be
equal to the general 1-alkene (OLE) bond in the CBM (e.g., propene). The
resulting rate constant is
k(75) = 1.8 x 10-11 cm3 molec'1 a"1.
The OH rate constant for reaction with isoprene was estimated by averaging
the rates of Atkinson and Aschmann [1984] and Kleindienst et al. [1982]:
860t9r2 12
237
-------
k(76) s 1.03 x 10-10 cm^ molec"' a"*'.
The reaction rate with Oo is taken as
k(77) = 1.22 x 10-17 cm3 molec'1 a'1,
the same as the rate constant for 1-alkenes (OLE) in the CBM and within
the error bounds of the studies reviewed by Atkinson and Carter (1984).
As noted in Killus and Whitten [1984], the reaction of 0? with isoprene
shows evidence of being a complex process, dissimilar in the nature of
secondary reactions to both monoalkenes and other dialkenes (e.g.,
butadiene). This is reflected in our product surrogate choices.
Finally, the rate of reaction with NOq is derived from the rate
constant presented by Atkinson et al. [1984]:
k(78) = 3.18 x 10-13 cm3 moleo'1 a'1.
The foregoing reactions, if implemented with an atmospheric model having
the correct concentrations of 0?, OH, 0, and NOo, appropriate atmospheric
dispersion characteristics and correct emissions for isoprene, should
yield the correct description of isopr ene concentration behavior.
Conversely, comparisons of measured isoprene concentrations with modeled
predictions may give some indication of the correctness of a model's
representation of isoprene emission and dispersion.
An explicit isoprene mechanism was initially developed by Killus and
Whitten [1984]. In the development of that mechanism we found the
following significant phenomena to be associated with isoprene oxidation:
Highly reactive products, including methacrolein and methyl vinyl
ketone. Some product reaction with ozone was also indicated, as
would be expected from the olefinic character of these two products.
A high rate of radical production, presumably from photolysis of iso-
prene products.
A high yield of PAN as well as a PAN-like compound, probably from
methacrolein.
A high yield of formaldehyde.
Because the complex secondary chemistry of isoprene and its oxidation
products (methacrolein, methylvinyl ketone, and possibly methylglyoxal)
precludes a direct condensation to a few representative species, these
880**9r2 12
238
-------
reactions were represented in condensed CBM-IV format by the use of
existing carbon-bond species. Our intent was not to algebraically reduce
the explicit reaction scheme, but to simulate the major chemical features
(i.e., ozone production, formaldehyde yields, and the temporal
characteristics of radical and PAN-type compounds). This was done by
simulating the UNC data set of isoprene/NOx experiments. The results for
these simulations were very similar to those obtained in Killus and
Whitten [1984] and may be seen in Gery et al. [1988].
Our general methodology for the formulation of the condensed isoprene
mechanism was based on the following rationale: The olefinic nature of
isoprene products would best be simulated by ethene, since the alkene
bonds of methacrolein and methylvinyl ketone are partly deactivated and
resemble ethene in their reactivity to OH and Oj. This representation
also gives a high yield of formaldehyde from the secondary oxidation of
ethene, which resembles the formaldehyde yield of isoprene. The formation
of PAN and PAN-like compounds from isoprene was simulated with a primary
yield of both acetaldehyde and acetylperoxy radical. Inevitably, this can
result in an overprediction of measured PAN, since simulated PAN includes
the PAN analogues as well. The radical yield for products was simulated
by the formation (and photolysis) of methylglyoxal. These separate
product yields were adjusted to give the best fits to the mid-range of
hydrocarbon-to-NOx-ratio isoprene experiments, in which notable double
ozone peaks become evident.
The double ozone peaks in isoprene/NOx experiments are caused by NOX
limitation by PAN formation, ozone reaction with isoprene and isoprene
products, and continued ozone formation by PAN decomposition at elevated
temperatures. Simulation of the double peak effect gives some indication
that a reasonable balance has been achieved for the multiple processes
occurring in isoprene oxidation. The condensed isoprene mechanism
resulting from simulations of the UNC isoprene/NOx data set is shown as
reactions R75 through R78 in Table 4.
12
239
-------
REFERENCES FOR APPENDIX E ^not found in main text)
Arnts, R. R., and B. W. Gay, Jr. 1979. Photochemistry of Some Naturally
Emitted Hydrocarbons. U.S. Environmental Protection Agency (EPA-
600/3-79-081).
Atkinson, R., S. M. Aschman, W.P.L. Carter, A. M. Winer, and J. N. Pitts,
Jr. 1982b. Alkyl nitrate formation from the NO—air photooxida-
tions of C2-Cg n-alkanes. J. Phys. Chem.. 86:4563-4569.
Atkinson, R., S. M. Aschman, and A. M. Winer. 1987. Alkyl nitrate forma-
tion from the reaction of a series of branched R02 radicals with NO
as a function of temperature and pressure. J. Atings. Chem.. 5:91-
102.
Atkinson, R., and W.P.L. Carter. 1987. Atmospheric Chemistry of
Alkanes. J. Atmos. Chem.. 3:377-405.
Atkinson, R., W.P.L. Carter, K. R. Darnall, A. M. Winer, and J. N. Pitts,
Jr. 1980. A smog chamber and modeling study of the gas phase NOX-
air photooxidation of toluene and the cresols. Int. J. Chem. Kinet..
12:779-836.
Atkinson, R., W.P.L. Carter, and A. M. Winer. 1983. Effects of tempera-
ture and pressure on alkyl nitrate yields in the NOX photooxidations
of n-pentane and n-heptane. J. Phys. Chem.. 87:2012-2018
Bandow, H., and N. Washida. 1985. Ring-cleavage reactions of aromatic
hydrocarbons studied by FT-IR spectroscopy. II. Photooxidation of
o-, m-, and p-xylenes in the N0x-air system. Bull, Chem. Soc. Jpn..
58:2541-2548.
Bandow, H., N. Washida, and H. Akimoto. 1985. Ring-cleavage reactions of
aromatic hydrocarbons studied by FT-IR spectroscopy. I. Photooxida-
tion of toluene and benzene in the N0x-air system. Bull. Chem. Soc.
Jpn.. 58:2531-2540.
Carter, W.P.L., A. C. Lloyd, J. L. Sprung, and J. N. Pitts, Jr. 1979.
Computer modeling of smog chamber data: Progress in validation of a
detailed mechanism for the photooxidation of propene and n-butane in
photochemical smog. Int. J. Chem. Kinet.. 11:45-103.
Carter, W.P.L., F. W. Lurmann, R. Atkinson, and A. Lloyd. 1986.
"Development and Testing of a Surrogate Species Chemical Reaction
Mechanism." U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina (EPA/600/3-86/031).
880H9r2 12
240
-------
Cox, R. A., K. F. Patrick, and S. A. Chant. 1981. Mechanisms of atmo-
spheric photooxidation of organic compounds. Reactions of alkoxy
radicals in oxidation of n-butane and simple ketones. Environ. Sci.
Technol.. 15:587.
Dumdei, B. E., and R. J. O'Brien. 1984. Toluene degradation products in
simulated atmospheric conditions. Nature, 311:248-250.
Gery, M. W., D. L. Fox, H. E. Jeffries, L. Stockburger, and W. S.
Weathers. 1985. A continuous stirred tank reactor investigation of
the gas-phase reaction of hydroxyl and toluene. Int. J. Chem.
Kinet.. 17:931-955.
Gery, M. W., D. L. Fox, R. M. Kamens, and L. Stockburger. 1987. Investi-
gation of hydroxyl radical reactions with o-xylene and m-xylene in a
continuous stirred tank reactor. Environ. Sci. Technol.. 21:339.
Grosjean, D., R. Countess, K. Fung, K. Ganesan, A. Lloyd, and
F. Lurmann. 1981. "Deriving Empirical Kinetic Modeling Approach
Isopleths from Experimental Data: The Los Angeles Captive-Air
Study." Environmental Research and Technology, Inc., Westlake
Village, California.
Hendry, D. G., and R. A. Kenley. 1979. "Atmospheric Reaction Products of
Organic Compounds." U.S. Environmental Protection Agency, Office of
Toxic Substances, Washington, D.C. (EPA-560/12-79-001).
Herron, J. T., and R. E. Huie. 1977. Stopped-flow studies of the
mechanisms of ozone-alkene reactions in the gas phase. Ethylene. J^
Amer. Chem. Soc.. 99(16): 5430-5435.
Kan, C. S., F. Su, J. G. Calvert, and J. H. Shaw. 1981. Mechanism of the
ozone—ethene reaction in dilute ^A^ mixtures near 1-atm pres-
sure. J. Phys. Chem.. 85:2359-2363.
Leone, J. A., and J. H. Seinfeld. 1984. .Updated chemical mechanism for
atmospheric photooxidation of toluene. Int. J. Chem. Kinet.. 16:159-
193.
Niki, H., P. D. Maker, C. M. Savage, and L. P. Breitenbach. 1977. Fourier
transform ir spectroscopic observation of prophylene ozonide in the
gas phase reaction of ozone-cis-2-butane-formaldehyde. Chem. Phys.
Lets.. 46:327-330
880"49r2 12
241
-------
Niki, H., P. D. Maker, C. M. Savage, and L. P. Breitenbach. 1978.
Mechanism for hydroxyl radical initiated oxidation of olefin-nitric
oxide mixtures in parts per million concentrations. J. Phys. Chem.,
82(2):135-137.
Ohta, T., and T. Ohyama. 1985. A set of rate constants for the reactions
of OH radicals with aromatic hydrocarbons. Bull. Chem. Soc. Jpn.,
58:3029-3030.
Plum, C. N., E. Sanhueza, R. Atkinson, W.P.L. Carter, and J. N.
Pitts, Jr. 1983. OH radical rate constants and photolysis rates
of a-dicarbonyls. Environ. Sci. Technol.. 17(8):479-48H.
Su, F., J. G. Calvert, and J. H. Shaw. 1980. A FTIR spectorscopic study
of the ozone-ethene reaction mechanism in 02-rieh mixtures. J. Phys.
Chem.. 8M:239-246.
Tuazon, E. C., H. MacLeod, R. Atkinson, and W.P.L. Carter. 1986.
a-Dicarbonyl yields frm the N0v-air photooxidations of a series of
o
aromatic hydrocarbons in air. Environ. Sci. Technol., 20(4):383.
Winer, A. M., K. R. Darnall, R. Atkinson, and J. N. Pitts, Jr. 1978.
J. Phys. Chem. Ref. Data, 13(2):198M.
880t9r2 12
242
-------
Part 2
IMPLEMENTATION OF THE CB-IV IN THE UAM
As with any new merger of chemistry into a large air quality simulation model
(AQSM), the predictive capabilities and solution speed of the new computer code for
solving the CB-IV required optimization and evaluation. This process is even more
important for the recent generation of gas-phase chemical kinetics mechanisms
(SAPRC, RADM, and the CB-IV), which are larger than previous mechanisms and
therefore require significantly more computing time. Using internal resources, Sys-
tems Applications, Inc. implemented the CBM-IV in the Urban Airshed Model. In the
implementation of the CB-IV in the UAM two minor modifications were made to the
CB-IV reactions (shown in Table 4) in order to add ethanol (ETOH) and methanol
(MEOH) as explicit species:
ETOH + OH * ALD2 + HO2 with a rate constant of 4,300 ppnrf1 min'1
at 298 K.
MEOH + OH * FORM + HO2 with a rate constant of 1,380 ppm"1 min'1
at 298 K.
Both numerical and chemical improvements were made to the CB-IV representation
so that computational speed could be increased and solution uncertainty diminished. •
Evaluation of the numerical technique used to solve a chemistry package should be
performed during implementation of all new mechanisms because of the conflicting
constraints of high computing costs for large AQSMs and somewhat limited accuracy
of faster solution techniques. These constraints result in an operational trade-off
between the time required to perform chemistry calculations and the accuracy of the
predicted concentrations. For any given mechanism and solution technique, either
speed or accuracy is inevitably sacrificed. However, when we implemented the
CBM-IV in the UAM, we were able to slightly modify the solution technique to mini-
mize the trade-off between speed and accuracy. In some cases we were able to
improve computer solution time without loss of solution accuracy; in others we
actually improved the chemistry representation and thus obtained more accurate
concentrations.
90008 38 243
-------
Uncertainty in the Chemistry Solution Routine
In most AQSMs a number of new or updated chemistry solution techniques are cur-
rently being employed. The solution techniques include those used in the EPA
Regional Oxidant Model (ROM), the UAM as developed by Systems Applications, Inc.
(SAI), various versions of the California Air Resources Board's UAM and CALGRID,
the SAI Regional Oxidant Model (the RTM-III), and the Regional Acid Deposition
Model (RADM). Some new algorithms have also been suggested but have not yet
been employed for atmospheric photochemistry systems. Most people outside the
modeling field assume that the same mechanism implemented with two different
techniques should give identical predictions. Although this is the ideal, it is not what
actually results. For a number of years we have advocated comparison of individual
techniques, both with each other and with a reasonably accurate standard solution
method, such as that of Gear. Comparisons with the Gear routine were performed
for the CBM-IV implemented in the UAM and should be performed for all solution
schemes currently implemented in complex chemistry AQSMs. To our knowledge,
this is rarely done (or at least rarely reported).
To implement the CBM-IV in the UAM we used a modified Crank-Nicholson
algorithm for the simultaneous solution of the differential equations that represent
each species. Because this algorithm utilizes both the rate of change for each
species and a Jacobian matrix relating the changes in each species to all others, the
technique can be slow under certain conditions. However, .it was chosen because,
when combined with our steady-state approximations, it calculates concentrations of
non-steady-state species that are very similar to those of the Gear algorithm. (Some
typical results of single-cell simulations with diurnally varying ultraviolet radiation
are shown for a few non-steady-state species in Figures 1 and 2.) Thus, we have
sacrificed some speed for more accurate predictions (in relation to Gear). As dis-
cussed later, we have devised some methods for regaining speed without the loss of
accuracy.
Further research is required to implement more efficient numerical integration
schemes for complex chemical mechanisms, such as the CB-IV, that will provide
computational speed as well accuracy that are comparable to Gear's. This research
should consider investigation into the speed and accuracy of as many solution
packages as possible in an effort to (1) identify the state of the science regarding
speed and uncertainty, (2) establish testing standards, and (3) weed out solution
packages that do not respond well in the rather unique environment of diurnally vary-
ing atmospheric chemistry solutions. At the least, we should verify that all the
major AQSMs can predict nearly the same results (similar to those of Gear) for
simple and representative atmospheric tests. For the CBM-IV implementation, for
instance, we used a trajectory model with a five-layer vertical resolution and a base
in an Eulerian emissions grid for much of the comparison work (this work could even
be performed with a single box model). Establishing, performing, and reporting such
a set of tests for other solution packages is in the best interest of the general photo-
chemical kinetics modeling community.
90006 36 244
-------
OJsoo
•MUUtnON TIKt (tnln)
FIGURE 1. NO, N02 and 03 concentrations for Example 2 (sunset 1s at 680
minutes); lines are Gear simulation results and symbols are Crank-
Nicholson predictions.
-------
«.DO
•-DO -
7.00-
«.oo -
W t.DO -
I:
. { 4 JO -
4.00 -
2.00 -
tJOO
0.00
0.2
TMIt
-------
Implementation of Speed Enhancement Procedures
Related to the solution algorithms are the approximations used to represent the
chemical features of a mechanism and, concurrently, enhance the solution speed. An
explicit mechanistic representation of a chemical mechanism, which assumes few or
no steady-state approximations, most accurately accounts for the mass of each
species (provided numerical errors do not arise because of the large range of concen-
tration variation between species); however, this technique may be computationally
slower than is necessary. The speed of a photochemical solution can be improved in
a number of ways without sacrificing accuracy. As noted, implementation of appro-
priate steady-state approximations is probably the best known. We have successfully
used steady-state approximations in the Carbon-Bond chemistry solution packages,
and see no reason why at least some set of steady-state calculations cannot be per-
formed in ail models requiring some improvements in speed.
An algorithm that solves higher-order steady-state relationships has recently been
successfully formulated. The fourth- and sometimes fifth-order algebraic solutions
for the OH, €20^, and HC>2 radical species can be rapidly derived. Results for
steady-state species using the higher-order relationships and Gear predictions are
shown in Figures 3 through 6 for the same conditions as in the non-steady-state
examples. Since these simulations were 2^-hour tests with diurnaily varying insola-
tion, the performance is good for all parts of the day. We caution, however, that
steady-state approximations can be misleadingly simple. Some aspects of their use,
such as numerical roundoff errors and mass balance considerations, must be investi-
gated carefully to ensure confidence in implementation. As noted, we pay particular
attention to comparisons with the Gear predictions, proceeding through a number of
different test cases at each logical step of a steady-state approximation.
This type of analysis has recently resulted in another enhancement of the photo-
chemical mechanism. One of the most difficult and potentially stiff chemistry loops
is the NK^-NO^-^O^-HNO-j cycle in which the intermediate species (NO-j and
^Oc) are potential steady-state species (especially during the day), whereas NO2
and HNO? are necessarily non-steady-state compounds because of their high concen-
trations. Under certain nighttime conditions, however, NO3 and N2O^ concentra-
tions can be relatively large, thus eliminating the possibility of steady-state treat-
ment. The dilemma is that steady-state representation tends to cause nitrogen mass
to be artificially lost, while non-steady-state representation is still stiff in many cir-
cumstances. We have recently tested a technique that allows us to move between
state and steady-state representations with the inclusion of only one additional non-
steady-state species, represented by the species NXOY. This technique ensures
nitrogen mass balance but also allows steady-state representation when it is prefer-
able. Recent test results of the new representation for the nitrogen cycle have been
compared to those produced by the Gear algorithm, using the same conditions as in
90006 38 247
-------
r
M
O.2 O.4 C-»
{ThMMMfc)
•HUUTION HUE (mln)
QJDQ
TIMt (tnln)
FIGURE 3. OH concentrations (top) and H02 concentrations (bottom) for
Example 2 (sunset 1s at 680 minutes); lines are Gear simulation results
and symbols are Crank-Nicholson predictions.
248
-------
•.00
7.00
«.oo -
H
f 2
|f
UDQ-
9JK) -
1.00-
C203
*J§
(•ntn)
FIGURE 4. C203 concentrations for Example 1 (sunset 1s at 680 minutes);
lines are Gear simulation results and symbols are Crank-Nicholson
predictions.
-------
e.oo
•IMUUTION H«E (tnte)
FIGURE 5:V C203 concentrations for Example 2 (sunset is at 680 minutes);
lines are Gear simulation results and symbols are Crank-Nicholson
predictions.
250
-------
OJCOOKO
OJKMO
fcOOSO -
0.OO4D -
6.DOSO-
B-OO20 -
C.001C -
C.BCDD •<
O.2
0.4 0.«
fn»n»n*)
•IMUUKTIOM TIMt (mln)
FIGURE ^. N03 concentrations (top) and N205 concentrations (bottom) for
Example 1 (sunset 1s at 680 minutes); lines are Gear simulation results
and symbols are Crank-Nicholson predictions.
-------
Figures i and 2, and are shown in Figure /. We are now considering this technique
for the SO2 oxidation cycle, which will be better represented with a "guaranteed"
sulfur balance.
Numerical Scheme Accuracy
The solution of state species (e.g., Oy NO, NO2, HNO^, VOC species) using the
CB-IV steady-state chemistry and the Crank-Nicholson scheme agree with the results
produced by the Gear algorithm to within two percent over a wide range of atmo-
spheric conditions. There has also been very close agreement between the solution
for steady-state species (e.g., OH, h^) using the higher-order relationships and the
solution for these species produced by the explicit treatment in Gear. Although the
solutions for the steady-state species using the higher-order relationships are within
5 percent of the Gear solutions most of the time, for some short periods of time the
agreement may be only within 10 to 15 percent (see Figure 3).
Photolysis Rate Calculations
There are fundamental differences in the calculations of photolysis rate in CB-II and
CB-IV. In the CB-II all photolysis reaction rates are calculated via a direct
porportionality relationship with the N©2 photolysis reaction rate. However, some
photolysis reactions (e.g., ozone, formaldehyde, and higher aldehyde photolysis rates)
are activated at spectrums of ultraviolet radiation different from those of ^^2'
Thus CB-IV requires separate photolysis reaction rates for NO?* formaldehyde,
higher aldehydes, and ozone. On the basis of the input N©2 photolysis rate, the
UAM(CB-IV) calculates the solar zenith angle required to produce the NO2 photolysis
rates and then calculates the photolysis rates for the other photolytic reactions on
the basis of the intensity of the solar radiation.
90008 36
-------
10.00
TIKt (into)
OJDOli
O.DDO16 "
O.ODD1* -
6.O0012 ~
I
0.0001 -
ejoooo» -
0X000? -
e
0.00005 -
O.ODOD4 -
B JD003 -
0-ODOD2 -
OJ0001 -
0
N205
0.2
TIMC
FIGURE 7. N03 concentrations (top) and N205 concentrations (bottom) for
Example 2 (sunset 1s at 680 minutes); lines are Gear simulation results
and symbols are Crank-Nicholson predictions.
-------
Appendix II
NUMERICAL REPRESENTATION OF HORIZONTAL
ADVECTION IN THE UAM
90006 6
-------
Appendix II
NUMERICAL REPRESENTATION OF HORIZONTAL
ADVECTION IN THE UAM
The Urban Airshed Model utilizes a numerical scheme formulated by Smolarkiewicz
(1983) to approximate the horizontal advection of chemical species. The
Smolarkiewicz advection scheme was adopted for the following reasons:
The scheme exhibits minimal artificial horizontal diffusion.
The scheme is positive definite, i.e., it cannot generate negative
concentrations.
The scheme is computationally efficient.
The Smolarkiewicz scheme is essentially a variant of the standard "upstream" finite-
difference advection scheme. It has been shown by many investigators that the
upstream advection scheme contains an undesirable degree of numerical diffusion,
which would preclude accurate representation of strong chemical concentration
gradients in a Eulerian air quality model such as the UAM. Smolarkiewicz (1983)
proposes removal of most (but not all) of this numerical diffusion through addition of
a "corrective" or "anti-diffusion" step to the "upstream" scheme. The corrective
step takes the same numerical form as the "upstream" scheme, with the "actual"
velocity replaced by an "anti-diffusion" velocity of opposite sign and considerably
smaller magnitude.
In the UAM advection is calculated separately in the x- and y- directions, as detailed
below:
90006 6
•5C-7
-------
in tnis discussion, superscripts ir,? n * 1) represent tne "time level" of a variable.
while subscripts (i, i + 1) represent horizontal grid index. We define
C. = concentration at center of cell i, at time step n
hj = vertical thickness of cell i
Uj = velocity component in advection direction region (x- or y-) at center
of cell i
Ax = horizontal grid spacing
At - time step
We define
»; . h, c;
as the variable to be advected.
"Upstream" Step
The advection scheme is posed in "flux" form. The "upstream" step is evaluated thus
A* = An - (F. - F. ) ^
•
where
represents a "flux" of material through a cell wall. Note that only one of the terms
in the bracket is non-zero.
90008 8
258
-------
"Corrective Step"
The "corrective" step is evaluated thus
An+1 - A* - (F - F ) —
•i ' fli ui i-1; Ax
where
F. = 0.5 [(u. + |u.| AJ + (u. - |u.|) AJ+1]
II * » A 1_\ / ft Jft AM
U. Ax - u. At\/A* - - A*
u. = - l! * V 1+1 x
and E is a small number (e.g., 10" ) which ensures that u^ = 0 if A* = A^^ = 0.
The derivation of u^ is detailed by Smolarkiewicz (1983).
Note that the "corrective" step can be repeated one or more times if desired. Based
on tests reported by Smolarkiewicz , the corrective step is exercised twice per time
step in the UAM.
90006 6
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before comnletinpi
REPOR" NC
SPA-450/4-90-G07A
13. RECIPIENT'S ACCESSION NO
i
4. TITLE AND SUBTITLE
USER'S GUIDE FOR THE URBAN AIRSHED MODEL
Volume I: User's Manual for UAM (CB-IV)
5. REPORT DATE
June 1990
6. PERFORMING ORGANIZATION CODE
7. AUTHORS) Ra-|pn £. |\ff0rri S
Thomas C. Myers
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael, CA 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, N. C. 277711
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document is a manual for operating the central Urban Airshed Model.
It also includes an overview of the model.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Ozone
Urban Airshed Model
Photochemistry
3. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (Tliis Report)
20. SECURITY CLASS (This page)
c. COS AT I Field/Group
'
21. NO. OF PAGES
22. PRICE
'A Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOUETE
-------
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Bouievatd. 12th Floor
Chicago, IL 60604-3590
------- |